Hereditary Systemic Amyloidoses — Approach and Non-TTR Forms

Fibrinogen Aα (AFib), apolipoprotein A-I (AApoAI), apolipoprotein A-II (AApoAII), lysozyme (ALys), gelsolin (AGel) and β2-microglobulin (Aβ2M)

Literature and regulatory information updated: 1 November 2026.

0. Initial Reference Table

GeneReviews: there is no dedicated text for any of the six non-TTR entities. The only hereditary amyloidosis chapter available is Hereditary Transthyretin Amyloidosis (NBK1194), which covers ATTRv.

European Reference Network: there is no network dedicated to the hereditary amyloidoses. The thematic groups of ERKNet, the European network for rare kidney diseases, do not include amyloidosis [27]. In practice, amyloid polyneuropathy falls within ERN EURO-NMD and amyloid cardiomyopathy within ERN GUARD-Heart, and the European Society of Cardiology position statement advocates a networked model in which local centres refer complex decisions and selected procedures to regional or national reference centres [28].

1. Introduction

The amyloidoses are diseases of extracellular deposition of insoluble, highly ordered, degradation-resistant protein fibrils, which accumulate in tissues and lead to progressive organ failure [29]. What unites them is a common physical mechanism, and what separates them is the identity of the precursor protein: each type of amyloidosis is defined by the protein that forms the fibril, and it is this identity that determines prognosis and treatment [30]. The Nomenclature Committee of the International Society of Amyloidosis (ISA), in the revision approved at the XIX International Symposium on Amyloidosis held in Rochester, Minnesota, on 27 May 2024, counts 42 human amyloid fibril proteins: 14 occur only in systemic deposits, 23 only in localised deposits and 5 in both forms (AL, AH, APrP, ACal and AIL1RAP, the last reclassified after two cases of systemic amyloidosis following prolonged treatment with anakinra), so that 19 are associated with systemic deposition; the article's abstract gives 4 proteins with localised or systemic deposits, but the body of the text and Table 1 list five [31]. The formal designation of each type is the letter A, for amyloid, followed by the abbreviation of the precursor protein: AL for monoclonal immunoglobulin light chains, AA for serum amyloid A protein, ATTR for transthyretin, AFib for the fibrinogen A alpha chain, and so on [32, 33]. The abbreviation designates the protein and not the disease, so the disease name is the protein followed by the word amyloidosis — AFib amyloidosis, AL amyloidosis —, with clinically useful specifications such as ATTR cardiomyopathy or AL neuropathy admitted; variant status is indicated by a suffix, “wt” for wild type and “v” for the variant form, a suffix the ISA expressly prefers to “m”, for mutant, and to “h”, for hereditary, the latter because it is confused with human [31].

This article covers the class of hereditary systemic amyloidoses and the approach common to all systemic amyloidoses, and develops in full the six hereditary forms not mediated by transthyretin: AFib, due to FGA variants; AApoAI, due to APOA1 variants; AApoAII, due to APOA2 stop codon variants; ALys, due to LYZ variants; AGel, due to GSN variants; and Aβ2M, due to B2M variants. Hereditary transthyretin amyloidosis, the most frequent hereditary form worldwide and the most relevant in Portugal and Brazil, appears here only in the classification, in the class table and in the differential diagnosis.

1.1 The Ostertag type and the discovery of the fibril proteins

The clinical history of this class begins with Benno Ostertag, who in 1932 and 1950 described a family with visceral amyloidosis in which a woman, three of her children and a grandchild had chronic nephropathy, arterial hypertension and hepatosplenomegaly, with albuminuria, haematuria and pitting oedema as early signs, variable age at onset and death about ten years after the appearance of symptoms; visceral involvement was extensive at autopsy [29]. The forms of hereditary systemic amyloidosis that do not have peripheral neuropathy as part of the syndrome came to be called “Ostertag type”, in reference to that German family [29, 34]. The label served for half a century as a clinical category of exclusion, grouping entities that, as is now known, are caused by different proteins.

The molecular dissection of this group took place between 1988 and 2012. In 1988, Nichols and colleagues identified variant apolipoprotein A-I as the major constituent of the amyloid fibrils of a member of the five-generation family of English, Scottish and Irish ancestry that Van Allen had described in 1969, and named the disease familial amyloidotic polyneuropathy type III, or Iowa type; in 1990 they confirmed, in DNA extracted from paraffin-embedded tissues of three members who died in the 1960s, the substitution of glycine by arginine at position 26 of the mature protein [6]. In 1990, Maury and Haltia showed that the amyloid of Finnish-type amyloidosis, described by Jouko Meretoja in his 1973 national survey, is antigenically and structurally related to gelsolin, and identified the 654G>A transition, which replaces aspartic acid 187 by asparagine, in the five unrelated patients studied and in none of the 45 controls [20]. In 1993, Pepys and colleagues showed that, in two unrelated English families with non-neuropathic hereditary systemic amyloidosis in which the deposits did not contain apolipoprotein A-I — despite an earlier report to the contrary in one of them —, the fibril protein was lysozyme, and that the affected individuals were heterozygous for substitutions of highly conserved residues, threonine for isoleucine at position 56 in one family and histidine for aspartic acid at residue 67 in the other [34, 35]. In the same year, Benson and colleagues identified, in a three-generation Peruvian family in which three members died of renal amyloidosis, a fibrinogen alpha-chain variant, proven by amino acid sequencing of amyloid isolated from the kidney post mortem: this was the first demonstration of hereditary amyloidosis associated with a variant fibrinogen [29, 36]. In 2001, Benson and colleagues described apolipoprotein A-II with a 21-residue peptide extension at the carboxyl terminus as the fibril protein in a family with previously uncharacterised hereditary renal amyloidosis, the first mutation identified in this protein and the first report of apoA-II amyloid in humans [37]. Finally, in 2012, Valleix and colleagues identified, in a three-generation French family, the β2-microglobulin variant Asp76Asn as the fibril protein of a hereditary systemic amyloidosis, distinct from dialysis-related amyloidosis because it occurs with normal renal function and normal circulating concentrations of the protein [24].

For fibrinogen amyloidosis, the correspondence with the historical description is more than a curiosity. Benson, in 2005, considered that the form caused by FGA variants, because it is the most common in Europe and presents with hypertension and proteinuria, is a strong candidate for being the amyloidosis originally described by Ostertag [29]. Among the intermediate descriptions are the three siblings of Maxwell and Kimbell, who died of visceral, mainly renal, amyloidosis in the fifth decade, whose follow-up by McKusick in 1974 documented a nephew with episodes of unexplained fever in childhood, proteinuria at 35 years, nephrotic syndrome for two years, uraemia over the following two years and death at 39 years, with renal, adrenal and splenic amyloidosis at autopsy [29]; the three Irish-American siblings of Mornaghi, with biopsy-proven rapidly progressive renal amyloidosis, without plasma cell dyscrasia, without monoclonal protein and with immunoperoxidase negative for amyloid A, light chains and prealbumin [29]; and the two American families of Irish ancestry of Uemichi, in whom the disease presented in the fifth to seventh decade with non-neuropathic amyloid nephropathy [29].

1.2 Hereditary, acquired and localised

The distinction between hereditary and acquired amyloidosis is the first branch point of clinical reasoning and also the one most often got wrong. Most systemic amyloidoses are caused by protein variants that promote misfolding, but in AA and ATTR the deposits more often consist of the wild-type protein or its fragments [31]. AL amyloidosis results from a clonal plasma cell dyscrasia producing a structurally unstable free light chain; AA amyloidosis results from high and prolonged concentrations of serum amyloid A protein in chronic inflammatory disease; wild-type ATTR amyloidosis results from deposition of normal transthyretin with age; and dialysis-related β2-microglobulin amyloidosis results from high and sustained plasma concentrations of normal β2-microglobulin in chronic renal failure treated by long-term haemodialysis [24, 38, 39]. None of these has germline genetic significance. Localised forms, in turn, are confined to one organ or territory and have no systemic expression: isolated atrial amyloid, derived from atrial natriuretic peptide (AANP), frequently occurs in the atrium and makes it necessary to require conclusive evidence of ventricular amyloid before diagnosing systemic cardiac amyloidosis [32]. In the 2022 revision, the ISA Nomenclature Committee added six proteins to the list, three of them polypeptide hormones and two peptide drugs in use, bringing to four the known iatrogenic forms, all appearing as subcutaneous nodules at the injection site; in 2024, one of these peptide drugs was shown to induce rare cases of systemic deposition [31, 33].

The hereditary amyloidoses are rare overall: they account for less than 5% of cases of renal amyloidosis [1], and the non-TTR types together make up a minimal fraction in the heart — AL and ATTR together account for more than 98% of diagnosed cardiac amyloidoses, and of the more than 30 proteins capable of forming amyloid only nine accumulate in the myocardium in amounts that cause significant disease [28]. Rarity, however, does not make the distinction dispensable, because the consequence of error is large in both directions. In the study by Lachmann and colleagues, 350 patients with systemic amyloidosis in whom the AL type had been suggested by clinical and laboratory findings and by the absence of a family history underwent testing for amyloidogenic mutations: 34 (9.7%) had them, most frequently in the fibrinogen A alpha-chain gene (18 patients) and in the transthyretin gene (13 patients), and in all 34 the diagnosis of hereditary amyloidosis was confirmed by further investigations; in 8 of these 34 (24%) there was a low-grade monoclonal gammopathy that would have falsely supported the diagnosis of AL [40]. The operational conclusion is that a genetic cause should be sought in all patients with amyloidosis that is not of the reactive AA type and in whom confirmation of the AL type cannot be obtained [40].

1.3 Nomenclature and recent reclassifications

The OMIM classification of these entities was reorganised in December 2023, with the creation of a dedicated phenotypic series. Series PS105210, “Amyloidosis, hereditary systemic”, has six entries: hereditary systemic amyloidosis 1 (AMYLD1, 105210, TTR), 2 (AMYLD2, 105200, FGA), 3 (AMYLD3, 620657, APOA1), Finnish-type amyloidosis, whose alternative titles include hereditary systemic amyloidosis 4 (AMYLD4, 105120, GSN), 5 (AMYLD5, 620658, LYZ) and 6 (AMYLD6, 620659, B2M), all with autosomal dominant inheritance and known molecular basis [20, 41]. Apolipoprotein A-II amyloidosis is not part of this series and has no phenotype MIM number: the only entries associated with APOA2 are apolipoprotein A-II deficiency, recorded as a provisional phenotype (621417, autosomal recessive), and the role of the gene as a modifier of familial hypercholesterolaemia (143890) [9]. The clinical entity exists, is documented in independent families and has its own Orphanet code (ORPHA:238269), but has not yet found a place in the OMIM phenotype nomenclature.

Orphanet organises the three predominantly renal forms — AApoAI, ALys and AFib — as clinical subtypes of a parent entity, “hereditary amyloidosis with primary renal involvement” (ORPHA:85450), and treats AGel (ORPHA:85448) and variant Aβ2M (ORPHA:314652) as separate diseases [2, 13, 17, 22, 42, 43]. The Orphanet OMIM mappings for AApoAI and ALys still point to 105200, a situation predating the creation of numbers 620657 and 620658, so that, in case of conflict, the OMIM assignment prevails [13, 41, 43].

Variant nomenclature is a second persistent source of confusion, because the historical literature numbers the residues of the mature or plasma protein and HGVS nomenclature numbers the precursor protein, including the signal peptide. The ISA Nomenclature Committee recommends, for amyloid protein variants, numbering based on the mature protein sequence, without leader sequence or propeptides, and likewise recommends that this classical numbering be used in parallel with standard protein nomenclature counted from the initiation codon, as in V30M (p.V50M) in transthyretin [31]. This is the convention followed in this article. The most common variant of fibrinogen amyloidosis is referred to as E526V in almost all of the clinical literature and as p.Glu545Val in the nomenclature that includes the 19-amino-acid signal peptide [1, 3]. The classical gelsolin variant is D187N in the mature protein and p.Asp214Asn in the precursor, and is also referred to as D214N [20]. The index β2-microglobulin variant was published as D76N and corresponds to p.Asp96Asn, a difference the authors themselves made explicit [23]. The Portuguese variant of the same protein is P32L in the mature protein and P52L in the precursor [23, 25]. Any conversion between the two numberings should be confirmed in a normalisation tool, not calculated, because the length of the signal peptide differs between proteins.

2. Molecular Genetics and Pathophysiology

2.1 Common mechanism

All forms of amyloidosis share a mechanistic aetiology: a protein adopts a misfolded, generally insoluble conformation; the dense, insoluble molecules aggregate into fibrils; and the fibrils accumulate in tissues throughout the body [32]. Precursor proteins normally exist in a steady state that favours the most stable form, but conditions that promote misfolding — production of a monoclonal free light chain at high concentration, oxidative stress or a destabilising variant — facilitate the initiation and propagation of the process, and, once an unstable subset has developed, its tissue deposition is as a rule irreversible [32]. Irreversibility has a central therapeutic consequence: once an amyloid niche has formed, progression may continue even after the source of the abnormal protein has been removed, as seen in patients with hereditary ATTR in whom liver transplantation removes production of the mutant protein but the disease progresses with deposition of wild-type transthyretin [32].

In the hereditary amyloidoses, the defect is not a loss of protein function but a gain of amyloidogenic property. Variant lysozyme fully retains bacteriolytic activity and carriers have no immunodeficiency [44]; variant gelsolin does not cause defective actin regulation [21]; and variant fibrinogen retains haemostatic function, so fibrinogen amyloidosis does not present with any coagulation abnormality [45]. In two of these entities it has been shown that only the mutated protein makes up the deposits: in AApoAI, laser microdissection with tandem mass spectrometry showed that the amyloid contains exclusively mutated apolipoprotein A-I, and not the wild-type protein [7]; in Aβ2M due to the P32L variant, with almost equal amounts of variant and wild-type protein in the circulation, the deposits were composed exclusively of the variant [25].

2.2 Table of all hereditary systemic amyloidoses and of the forms that enter the same reasoning

Note on nomenclature and divergence: the 2021 European Society of Cardiology position statement classifies β2-microglobulin amyloidosis as non-hereditary and associated with long-term dialysis, and apolipoprotein A-IV amyloidosis as non-hereditary [28]. Both classifications have been superseded. The hereditary form of Aβ2M has been established since 2012, with two B2M variants described in independent families, has had its own OMIM number since December 2023 and appears as a separate, systemic and hereditary entry in the ISA nomenclature table, alongside the row for the acquired wild-type form [23–25, 31]. As for apolipoprotein A-IV, the 2024 revision of the nomenclature added to the table a hereditary systemic form, dependent on a p.D33N missense variant, alongside the acquired wild-type form that until 2022 was the only one listed [31]. In both cases, the divergence is resolved in favour of the primary sources and the current nomenclature.

An additional note in the 2024 nomenclature has direct diagnostic relevance. Amyloid deposits contain, besides the main fibril protein, components present in almost all deposits — serum amyloid P component and heparan sulphate proteoglycans, both important for fibril stability and, in the case of the proteoglycans, for their genesis —, apolipoprotein E possibly belonging to this group and apolipoprotein A-IV being regularly bound to the fibril; these are the so-called amyloid signature proteins, used in mass spectrometry as additional proof that the material extracted from the tissue is amyloid [31]. The practical consequence is that detection of apolipoprotein A-IV in a mass spectrum may mean signature rather than fibril protein, a distinction that requires judgement about relative abundance and that underlies the diagnostic pitfall described in section 6 [31, 46].

2.3 Variant nomenclature: dual numbering and its pitfalls

The literature on these diseases numbers the residues of the mature or plasma protein, and HGVS nomenclature numbers the precursor protein, including the signal peptide or the pre-pro-peptide. The offset between the two numberings is specific to each protein — 19 amino acids in fibrinogen Aα, 24 in apolipoprotein A-I, 23 in apolipoprotein A-II, 18 in lysozyme, 27 in gelsolin and 20 in β2-microglobulin — and cannot be calculated by analogy. The following table brings together the validated correspondences.

Four amyloidogenic variants of these entities have, in ClinVar, a classification that does not match the clinical evidence, and the divergence should be known by anyone interpreting a report. FGA R554L, p.(Arg573Leu), has conflicting classifications, although its amyloidogenicity was proven by amino acid sequencing of amyloid isolated from the kidney and it accounts for 10 of the 81 renal cases typed by mass spectrometry in a reference series [1, 36, 47]. E540V in the same gene, p.(Glu559Val), is classified as of uncertain significance, although it accounts for three patients in the largest published cohort [3, 47]. APOA1 L90P, p.(Leu114Pro), is classified as of uncertain significance, although the substitution was confirmed by structural analysis of amyloid isolated from cardiac deposits [6, 47]. And B2M P32L, p.(Pro52Leu), is classified as of uncertain significance, despite family segregation, biophysical characterisation and the identification of a second patient of the same ancestry [23, 25, 47]. In all four cases, the classification of an aggregated database should not prevail over the demonstration that the corresponding protein makes up the deposit.

Three pitfalls in this table deserve to be made explicit. The first is that the OMIM entry for Finnish-type amyloidosis mixes, in the same list, one variant in mature numbering, D187N, and two in precursor numbering, W493R and E580K, so applying the 27-residue offset to these two produces non-existent designations [20]. The second is that the new variant described in 2026 in the gelsolin gene is given as N160K according to the numbering of the shorter cytoplasmic isoform and described, in the same paragraph, by its structural interaction with residue D187, which is mature plasma protein numbering; the two positions correspond to precursor residues 211 and 214, almost contiguous, which makes the interaction described coherent, but requires the numbering to be harmonised when citing [48]. The third is that two independent nucleotide changes at the same nucleotide of the lysozyme gene produce the same protein substitution, which is why OMIM assigns them distinct allelic variant numbers and why they should not be merged [34, 49].

2.4 Fibrinogen A alpha chain (FGA)

Fibrinogen is synthesised in the liver and is the substrate of thrombin in the coagulation cascade. It is composed of three polypeptide chains, encoded by a cluster of three genes on chromosome 4 — FGA, FGB and FGG —, two copies of each of which assemble into the 340 kDa plasma hexamer that thrombin cleaves into fibrin, the structural backbone of the clot; fibrinogen also guides cell migration in wound repair and is required for placental integrity [39, 45]. FGA is located at 4q31.3, has MIM number 134820, reference transcript NM_000508 and corresponds to UniProt protein P02671 [45].

Of the three chains, only the alpha chain produces hereditary renal amyloidosis [45]. All described amyloidogenic variants are concentrated in the carboxy-terminal portion of the chain, encoded by exon 5, which is why targeted genetic screening amplifies and sequences that exon [3]. The protease sensitivity of the carboxyl terminus of the Aα chain is suspected to facilitate deposition, and mutated Aα chains have not been detected in plasma, presumably because the mutated protein is metabolised faster than the wild-type protein [39]. Circulating variant fibrinogen is entirely of hepatic origin and the wild-type protein does not form amyloid fibrils in vivo, two premises that underpin the rationale for liver transplantation [3].

The number of published amyloidogenic FGA variants is 16, of which at least 6 have been described since 2009 — four single-base substitutions and two frameshifts —, and two were first identified in patients assessed at the UK national centre [3]. The count differs slightly between sources: the Mayo Clinic series refers to 15 pathogenic variants [1] and the nephrology review to “more than 15” [39], reflecting the date of consultation of each. One variant dominates overwhelmingly: the substitution of glutamic acid by valine at codon 526 of the mature protein, NM_000508.5:c.1634A>T, NP_000499.1:p.(Glu545Val) in the numbering that includes the 19-amino-acid signal peptide, designated E526V in almost all of the clinical literature [1, 50]. This variant accounts for 86 to 90% of published cases, has been described mainly in Northern Europeans but identified in families worldwide, and accounted for 187 of the 194 patients (96.4%) in the largest published cohort [1, 3]. In that cohort, the remaining seven patients had p.Glu559Val (three patients, 1.5%, formerly E540V), p.Arg573Leu (two patients, 1.0%, formerly R554L, NM_000508.5:c.1718G>T) and one patient each with p.Pro571His (formerly P552H) and p.Gly574Phe (formerly G555F) [3, 50]. In the North American Mayo Clinic series, mass spectrometry detected a pathogenic variant in 81 of 115 renal cases (70%), with E545V in 71 (88%) and R573L in 10 (12%) [1]. Frameshift variants have been found mainly in isolated families [1]: the first, c.4904delG at the third nucleotide of codon 524, with frameshift and premature termination at codon 548, was described by Uemichi in 1996 and was also the first amyloidosis attributed to a frameshift; the resulting protein loses 14% of the normal carboxy-terminal sequence, residues 525 to 610, and gains an abnormal 23-residue peptide, which is speculated to be degraded much faster than the normal protein [36]. The second, c.4897delT, with frameshift at codon 522 and termination at codon 548, was identified by Hamidi Asl in 1997 in a French family; amyloid isolated from the proband's transplanted kidney contained a 49-residue hybrid peptide whose 23 amino-terminal amino acids were identical to residues 499 to 521 of normal fibrinogen and whose remaining 26 residues represented a sequence entirely new to mammalian proteins [36].

The genotype–age at onset correlation is consistent across series and has practical value. Ages at onset of renal disease were, in the French series of 32 patients, 57 years on average (range 36 to 77) in E526V carriers, 45 years (12 to 59) in R554L carriers and 24.5 years (12 to 31) in carriers of frameshift variants, with statistical significance in both comparisons (p<0.001) [4]. OMIM describes the same ordering: onset in the third decade with R554L, at the end of the fourth and beginning of the fifth decade with the 1996 frameshift variant, and in the fifth to seventh decade with E526V [36]. Haplotype analysis of four families with E526V, two Irish-American, one Polish-Canadian and an additional Irish-American one, suggested origin from a single founder [36]. Among the 18 patients with E526V identified in the British study reclassifying presumed AL amyloidoses, all of Northern European ancestry, none was aware of a relevant family history, but genealogical studies revealed that two were cousins and that ancestors of two others had lived in neighbouring villages [29].

The allelic disorders of FGA are the quantitative and qualitative fibrinogen disorders, and they are clinically distinct from amyloidosis. Congenital afibrinogenaemia, autosomal recessive, with undetectable fibrinogen, and hypofibrinogenaemia manifest as bleeding disorders: umbilical cord bleeding is the classical neonatal presentation, followed by muscle haematoma, oral cavity and post-surgical bleeding, cutaneous haematoma and menorrhagia, with central nervous system haemorrhage the main cause of morbidity and mortality; afibrinogenaemia also carries a paradoxical thrombotic risk, because thrombin generation persists in the absence of fibrin [45]. Dysfibrinogenaemia, qualitative and generally dominant, with normal antigen and low activity, may cause bleeding, thrombosis or neither, and the FGA variants that cause it cluster at the thrombin cleavage site in exon 2, Arg35Cys, Arg35His and Arg35Ser, corresponding to Arg16 of the mature chain [45]. The MIM numbers of these entities are 202400 (afibrinogenaemia and hypofibrinogenaemia) and 616004 (dysfibrinogenaemia and hypodysfibrinogenaemia), and the Orphanet code for congenital fibrinogen deficiency is ORPHA:335, with a point prevalence of 1.5 per 1,000,000 in Europe, not validated and referring to the three genes together [45, 51]. In carriers of amyloidogenic variants, haemostasis is typically normal, although one family with the 1996 frameshift variant had a low plasma fibrinogen concentration, and two of the proband's four children, who were carriers, had lower fibrinogen than their siblings without symptoms of amyloidosis in the second decade of life [36, 45].

2.5 Apolipoprotein A-I (APOA1)

Apolipoprotein A-I is the main protein of high-density lipoprotein in plasma and plays a central role in cholesterol efflux from peripheral tissues to the liver, where it is excreted; it is also the activating cofactor of lecithin–cholesterol acyltransferase [52, 53]. The gene is located on the long arm of chromosome 11, at 11q23.3, adjacent to the apolipoprotein C-III and A-IV genes, has MIM number 107680 and transcript NM_000039 [52, 53]. Of the approximately 70 distinct variants described, most in the heterozygous state, some are functionally relevant, being associated with reduced concentrations of apoA-I and HDL cholesterol, and some structural substitutions in the amino-terminal region have been detected in patients with familial amyloidosis [53].

The amyloidogenic APOA1 variants catalogued by OMIM, in mature protein numbering, are G26R, L60R, W50R, an exon 4 deletion and insertion with loss of residues 60 to 71 and insertion of valine and threonine, L90P, R173P, L174S and A175P [6]. The fibrils are composed of amino-terminal fragments of the variant, ending mainly at residues 83, 92 or 94 [6]. The hypothesis that an extra positive charge in this region would explain amyloidogenicity, supported by all variants known until 1996, was contradicted by L90P, which does not alter charge, leading the authors to consider it unlikely that fibril formation relates to the change in charge of a specific residue, in parallel with what had been observed in transthyretin amyloidosis [6]. L174S was the first variant located outside the deposited fragment: in the three-dimensional structure of lipid-free apoA-I, composed of four identical polypeptide chains, position 174 of one chain lies close to position 93 of the adjacent chain, which suggested that the substitution at position 174 allows proteolytic cleavage at the carboxyl terminus of valine 93 [6]. In R173P carriers, plasma HDL cholesterol concentrations were decreased, suggesting increased catabolism, as had been shown for G26R, and indicating that the altered metabolism caused by the variant may be a significant factor in fibrillogenesis [6]. A variant described as new in 2026, NM_000039.3:c.605T>C, NP_000030.1:p.(Leu202Pro), lies in exon 4 and in the amino-terminal region, described by the authors as a hotspot region for amyloidogenic variants; the Human Genome Variation database had no entry for the same change, but had another at the same locus (c.605T>G), and in ClinVar the corresponding record has two reports classified as pathogenic, without specification of the associated disease [8, 54]. In mature protein numbering, this variant corresponds to the L178P substitution, which is the same change previously described as a cause of low HDL cholesterol associated with vascular dysfunction and accelerated carotid arterial wall thickening [53, 54]. The reconciliation of the two descriptions is that a single apolipoprotein A-I variant can simultaneously produce reduced HDL cholesterol, through accelerated catabolism of the protein, and amyloidosis, through amyloidogenicity of the variant protein, a mechanism already documented for the G26R and R173P variants, in whose carriers HDL cholesterol concentrations were decreased [6]. The L178P variant should not be confused with L174S, which is a distinct change, c.593T>C, p.(Leu198Ser) [54]. The work of Rowczenio and colleagues added five new variants, four amyloidogenic and one incidental in a patient with systemic AL amyloidosis, and established by proteomics that only the mutated protein is deposited [7].

The allelic disorders of APOA1 are apolipoprotein A-I deficiency and hypoalphalipoproteinaemia. ApoA-I deficiency is a rare autosomal recessive disorder characterised by decreased HDL cholesterol concentrations: heterozygous carriers of a functionally relevant variant usually have half the normal values of apoA-I and HDL cholesterol, some variants cause more marked reductions, and in most cases heterozygotes have no specific symptoms [53]. Biallelic loss of function gives undetectable plasma apoA-I with very reduced HDL cholesterol, xanthomas, corneal opacities and, in most patients, premature atherosclerotic cardiovascular disease; it is undetectability, and not just a very low value, that distinguishes apoA-I deficiency from Tangier disease in the differential diagnosis of severe HDL deficiency [52]. Susceptibility to premature coronary disease differs markedly between apoA-I variants: low HDL cholesterol concentrations due to heterozygosity for a specific variant, L178P, were associated with vascular dysfunction and accelerated carotid arterial wall thickening [53] — the same variant that the 2026 case described above associates with cardiac amyloidosis. The Orphanet code for apolipoprotein A-I deficiency is ORPHA:425, with a point prevalence below 1 per 1,000,000 worldwide and in Europe and about 30 families described, a validated record [55]. Reading a low HDL cholesterol value in a carrier of an APOA1 variant is diagnostic and not therapeutic: genetically elevated HDL cholesterol did not track myocardial infarction risk in Mendelian randomisation, so it is not in itself a modifiable therapeutic target [52].

2.6 Apolipoprotein A-II (APOA2)

APOA2 is located at 1q23.3, with coordinates 1:161,222,292-161,223,628 in the GRCh38 reference genome, has MIM number 107670 and transcript NM_001643 [9, 10]. Apolipoprotein A-II is the second most abundant protein of high-density lipoprotein particles, present in plasma as a monomer, a homodimer or a heterodimer with apolipoprotein D, and is thought to stabilise HDL structure and influence its metabolism [9, 10].

The mechanism of apoA-II amyloidosis is unique among the hereditary amyloidoses: all variants described affect the stop codon, and none is a substitution within the coding sequence. The known variants are Stop78Gly, Stop78Ser and Stop78Arg; each replaces the normal stop codon with an amino acid, extending translation to the next in-frame stop codon, 60 nucleotides downstream, and producing an apolipoprotein A-II with a 21-residue carboxy-terminal extension, identical to that found in the amyloid [10, 11, 37]. The variant protein with the extension circulates in plasma alongside the normal one and is deposited as fibrils, mainly in the kidney [10]. The variant described by Benson in 2001 creates a new BstNI restriction site, usable for carrier identification by restriction fragment length polymorphism analysis [37]; the one described by Yazaki in 2003 is a thymine to cytosine transition at the first position of the stop codon, replacing it with arginine at residue 78 [11]. Fibril formation from apoA-II, whose predicted structure is amphipathic helical, suggests that this lipoprotein undergoes a transition to beta-pleated sheet by a mechanism shared with other amyloidogenic lipoproteins [37].

OMIM assigns no phenotype number to this amyloidosis, and the two entries associated with APOA2 concern entities that are not the amyloid disease. Apolipoprotein A-II deficiency (MIM 621417, autosomal recessive, recorded as provisional) is known from a Japanese family in which two sisters were homozygous for a guanine to adenine transition at the first position of intron 3, designated apoA-II(Hiroshima); absence of plasma apoA-II had little effect on lipids and coronary disease [9, 10]. The modifier role in familial hypercholesterolaemia (MIM 143890) is based on the promoter polymorphism -265T>C (rs5082): in an American kindred of 1135 members, carriers of an LDLR variant who were also homozygous for the -265C allele had markedly lower total cholesterol and LDL cholesterol values (155 versus 237 mg/dl), and transfection assays showed 30% reduced promoter transcriptional activity with the -265C allele [9, 10]. None of these entries is a diagnosis of amyloidosis, and no amyloidogenic variant is catalogued among OMIM's selected allelic variants for this gene [9].

2.7 Lysozyme (LYZ)

LYZ is located at 12q15, has MIM number 153450 and transcript NM_000239, and encodes human lysozyme, a secreted bacteriolytic enzyme that cleaves the beta-1,4 bond between N-acetylmuramic acid and N-acetylglucosamine of bacterial peptidoglycan [44]. The protein is produced by hepatocytes, neutrophils and macrophages and is present in milk, saliva, tears, plasma and most secretions [16, 44]. The amyloidogenic missense variants do not cause immunodeficiency: they lower the conformational stability of the protein, so that full-length variant lysozyme unfolds and is deposited as amyloid fibrils in the viscera [44]. Pepys and colleagues showed that the fibrils of an individual from the first family described were composed of the full-length Thr-56 variant lysozyme molecule, and noted that, since the structures of human and hen egg-white lysozyme are known at atomic resolution and their folding and structure–function relationships exhaustively analysed, the disease would provide a powerful model for understanding amyloidogenesis [35].

The catalogued amyloidogenic variants are I56T and D67H, in the two original English families; W64R, in a French family; F57I, in an Italian-Canadian family; and a second W64R, resulting from a different nucleotide change, in a woman from Piedmont, later found in another woman from the same region and in an English man [34]. The two nucleotide changes that produce p.(Trp82Arg) in precursor numbering, corresponding to historical W64R, are NM_000239.3:c.244T>C and c.244T>A, both classified as pathogenic in ClinVar [49]. The family with hepatic rupture described by Harrison in 1996 also had the D67H variant, but was apparently unrelated to the original English family [34]. Altogether, nine amyloidogenic lysozyme variants and about thirty affected families have been described worldwide [14]. Structural analysis of lysozyme with comparative computational predictions of aggregation propensity and conformational stability, carried out in the German cohort, supports decreased conformational stability as the key factor that makes the variants aggregation-prone [14]. An amyloid fibril structure derived from hereditary amyloidosis lysozyme has been determined by cryo-electron microscopy [44].

There is no genotype–phenotype correlation in this entity: the phenotype is similar within each family but does not correlate with genotype, and marked differences have been observed between families carrying the same variant [14, 16]. Clinical penetrance is, by contrast, high: in the British cohort, a family history of amyloidosis was present in every affected individual [16].

2.8 Gelsolin (GSN)

GSN is located at 9q33.2, has MIM number 137350 and transcript NM_000177, and encodes gelsolin, a calcium-regulated actin-modulating protein present in a cytoplasmic form and a secreted plasma form [21]. Gelsolin caps the plus, or barbed, ends of actin monomers and filaments, blocking monomer exchange, nucleates the assembly of monomers into filaments and severs existing filaments; it contributes to ciliogenesis and, in the nervous system, is involved in axonal transport, myelination, neurite outgrowth and neuroprotection [21].

The mechanism is now known in detail. The codon 187 substitutions, the G654A and G654T changes, which produce p.Asp187Asn and p.Asp187Tyr in the mature protein, destabilise the second domain of gelsolin and impair calcium ion binding, so that the secreted plasma isoform is aberrantly processed by furin in the Golgi apparatus; a 68 kDa carboxy-terminal fragment, designated C68, is generated, which is then processed extracellularly into amyloidogenic 8 and 5 kDa fragments, which are deposited as AGel amyloid in the cornea, peripheral and cranial nerves, skin, heart and kidney [21, 48]. The classical D187N variant corresponds to NM_000177.5:c.640G>A, NP_000168.1:p.(Asp214Asn), and the same change in NM_198252.3, a shorter isoform, is c.487G>A, p.(Asp163Asn); it is also referred to in the literature as D214N [20, 54]. The variant was found in all Finnish families studied and also in the affected son of the proband of the Scottish-American family described by Sack in 1981 [20]. Two later variants lay outside the canonical domain, both designated in precursor and not mature protein numbering: E580K, identified by exome sequencing in six members of a family, segregated with the disease, was absent from gnomAD and lies in domain G5, homologous to the second domain that contains D187N, corresponding to E553K in mature numbering; and W493R, the first variant located in domain G4, corresponding to W466R in mature numbering, with immunohistochemistry of corneal tissue identifying gelsolin in the histologically defined amyloid deposits [20, 54]. In 2026 the variant given as c.480C>A, p.Asn160Lys in cytoplasmic isoform numbering was described, corresponding to NM_000177.5:c.633C>A, p.(Asn211Lys) in the reference transcript, heterozygous and likely pathogenic, absent in the patient's son: analysis of the gelsolin crystal structure showed that in the wild type the N160 side chain interacts closely with D187 of domain 2, stabilising the conformation of the calcium-binding pocket, and that with the lysine substitution this interaction is weakened, with the distance increasing from 2.9 to 4.1 ångström, although the distance between D187 and the calcium ion remains at about 4.0 ångström [48].

Inheritance is autosomal dominant, but homozygotes have been described. Two patients from the 1973 Finnish national survey, whose parents were both affected and whose disease was more severe than that of the others, were considered homozygous [20]. In 1993, Maury described two sisters homozygous for D187N, with unusually severe disease, manifesting with nephrotic syndrome and end-stage renal failure; immunohistochemical studies of the kidneys showed abundant glomerular deposits of gelsolin-derived amyloid, and immunostaining also revealed gelsolin in the tubular epithelium, where it was Congo red negative [20]. Phenotypic variation between variants is marked: in the 2026 case with the N160K variant, the interaction between domain 2 and the calcium ion was preserved, which the authors propose as the explanation for the absence of typical neurological symptoms over a three-decade disease course, and an Asp163Asn variant had been described with overt renal injury and no involvement of other organs [48].

2.9 β2-microglobulin (B2M)

B2M is located at 15q21.1, has MIM number 109700 and transcript NM_004048, and encodes β2-microglobulin, the invariant light chain that stabilises the major histocompatibility complex class I and the related molecules CD1a, CD1b, CD1c and the neonatal Fc receptor at the cell surface; its loss abolishes class I antigen presentation and shortens the catabolic half-life of immunoglobulin G and albumin [38]. The same small beta-sheet protein is intrinsically amyloidogenic, and destabilising germline variants form fibrils at normal plasma concentrations [38].

Three distinct entities derive from this protein, and confusing them has clinical consequences. The first, autosomal recessive, is MHC class I deficiency, catalogued as immunodeficiency 43 (MIM 241600) and historically as familial hypercatabolic hypoproteinaemia, caused by biallelic variants: in the two siblings first sequenced, rapid degradation of albumin and immunoglobulin G produced marked deficiency of both, and eight relatives had moderate IgG deficiency; a later pair of Turkish siblings had recurrent respiratory infections, with severe skin disease in the sister and bronchiectasis with few symptoms in the brother, showing that severity varies widely within the family [38]. The second, autosomal dominant, is hereditary systemic amyloidosis 6 (MIM 620659), caused by destabilising heterozygous variants [23, 38]. The third, entirely distinct and without germline genetic significance, is dialysis-related amyloidosis, acquired, caused by high and sustained plasma concentrations of the normal wild-type protein in long-term haemodialysis, in which amyloid accumulates around bones and joints [24, 38].

Two amyloidogenic variants have been described. D96N, c.286G>A, NP_004039.1:p.(Asp96Asn), was published as D76N because the numbering used did not include the 20-amino-acid signal peptide; the recombinant mutant protein showed reduced stability in the fully folded form and greatly increased conversion into fibrils with amyloid properties under physiological conditions, whereas the normal protein did not aggregate; the high-resolution crystal structure, determined at 1.40 ångström, is similar to that of the normal protein, but the substitution establishes new hydrogen bonds and raises the isoelectric point [23, 24, 54]. The second, P52L in precursor numbering and P32L in mature protein numbering, results from a single heterozygous dinucleotide mutation, c.154_155delinsTT — it is therefore not a single-base substitution, and a notation such as c.155C>T would be incorrect —, and was described in a Portuguese family; in vitro studies demonstrated thermodynamic and chemical instability and greater susceptibility to proteolysis, with rapid formation of prefibrillar oligomeric structures by amino- and carboxy-terminally truncated species under physiological conditions [23, 25]. Haslett and colleagues identified the same variant in another patient of Portuguese ancestry and considered a common ancestor possible [23].

3. Clinical Spectrum and Manifestations by System

3.1 Manifestations common to the systemic amyloidoses

Regardless of the precursor protein, amyloid deposition produces a set of manifestations that the clinician recognises as belonging to this family of diseases. In the heart, accumulation in the myocardial extracellular matrix increases wall thickness and causes diastolic dysfunction, progressing to restrictive cardiomyopathy; the most common picture is congestive heart failure, with oedema, dyspnoea and ascites, and when systolic function is also compromised low output leads to forward failure [32]. Hypotension or orthostasis may result from systolic heart failure and also from autonomic dysfunction [32]. The cardiac hallmarks described across the amyloidoses are atrial fibrillation, concentric biventricular hypertrophy, conduction delay on the electrocardiogram, coronary arterial obstruction, disproportionately low voltage on the electrocardiogram, diastolic or combined heart failure, increased echodensity and thickness of the atria and cardiac valves, and pericardial effusion; the extracardiac hallmarks are ascites, autonomic neuropathy, carpal tunnel syndrome, constipation alternating with diarrhoea and malabsorption, tongue enlargement with dental indentations, haematuria due to bladder infiltration, periorbital purpura, peripheral neuropathy, peripheral oedema, pleural effusion and biceps tendon rupture [32].

The pathophysiology of restrictive heart failure explains the drug intolerance that marks treatment. With progressive worsening of left ventricular hypertrophy, the cavity typically becomes smaller; despite preserved ejection fraction, stroke volume decreases and cardiac output becomes directly proportional to heart rate within the physiological range, so that fast rates are poorly tolerated because of ventricular stiffness and slow rates worsen heart failure [32]. Angina pectoris in the absence of obstructive epicardial coronary disease is recognised in cardiac amyloidosis and probably originates from microvascular dysfunction: autopsy studies showed deposits not only in the interstitium between cardiomyocytes but also in perivascular regions and in the intima and media of coronary vessels, and in a prospective study of 21 patients with cardiac amyloidosis without obstructive coronary disease, compared with 10 patients with hypertrophy due to hypertensive cardiomyopathy, vasodilator positron emission tomography showed varying degrees of regional ischaemia in the former and no regional perfusion defect in the latter, with rest and stress myocardial blood flow and coronary flow reserve significantly lower in the amyloidosis group even after correction for differences in wall thickness [32]. Conduction disorders are frequent, with sinoatrial node dysfunction, chronotropic incompetence, bundle branch blocks and atrioventricular blocks of various degrees; ventricular ectopy is frequent but rarely progresses to ventricular tachycardia, and the most common tachyarrhythmia is atrial fibrillation, whose control is hampered by poor tolerance of drugs that slow atrioventricular conduction [32]. Cardiac thromboembolism has been described even without documented atrial fibrillation: in an autopsy study of hearts with amyloidosis, intracardiac thrombus was found in up to 33% of the hearts examined, most in the atria, with a prevalence no different between patients with and without documented atrial fibrillation; the mechanism is unknown, with endothelial dysfunction, hypercoagulability or reduced atrial mechanical activity despite sinus rhythm proposed, and routine assessments for thrombus detection are not performed in daily practice, the risks and benefits of prophylactic anticoagulation in sinus rhythm being unknown [32].

3.2 Fibrinogen A alpha-chain amyloidosis

AFib is, in its clinical expression, a kidney disease. The largest published series, with 194 patients followed at a national centre between 1993 and November 2025, describes a median age at diagnosis of 60 years, with a range of 39 to 84 years, without variation between the two periods studied, and a slight male predominance (55.6%) [3]. The presentation is proteinuric chronic kidney disease: mean urinary protein quantification at diagnosis was 4.6 g per day (range 0.1 to 19.0) and mean urinary albumin/creatinine ratio 384 mg/mmol (8 to 1577), most patients being in albuminuria category A3; 103 of the 194 patients (53.1%) presented with arterial hypertension [3]. Diagnosis is often late: almost one third of the cohort (30.4%) was already in chronic kidney disease stage IV, and 15 patients (7.7%) were dialysis-dependent at diagnosis, the mean estimated glomerular filtration rate in the remainder being 37 ml/min per 1.73 m² [3]. In the French series of 32 patients, the median age at diagnosis was 51.5 years (12 to 77), with proteinuria in 93%, hypertension in 83% and renal failure in 68% at presentation [4]. In the North American Mayo Clinic series of 13 patients, the second largest in the United States, with a median age of 61 years (47 to 74), all had proteinuria (median 4.5 g; 1.6 to 14 g), 12 (92%) hypertension, 11 (85%) raised serum creatinine (median 4 mg/dl; 1 to 15), 6 (50%) hypoalbuminaemia (median albumin 3.4 g/dl; 2 to 4.3), 4 (36%) peripheral oedema, 2 (17%) haematuria and only 1 (8%) full nephrotic syndrome [1].

The renal natural history is one of progressive and relatively rapid loss of function. In the British cohort, the mean rate of loss of estimated glomerular filtration in patients with a value above 20 ml/min per 1.73 m² at diagnosis was 5.3 ml/min per 1.73 m² per year overall, with a difference between periods (6.7 before 2009 and 4.8 after), and 122 of the 194 patients (62.8%) progressed to end-stage kidney disease requiring replacement therapy, at a median age of 61 years (36 to 84) [3]. The historical series of 71 patients from the same centre, published in 2009, with a median age at presentation of 58 years (33 to 83) and primary renal involvement in all, described a median time from presentation with proteinuria to end-stage kidney disease of 4.6 years (0 to 10.2) and of only 2 years from the diagnosis of amyloid, with 44 of the 71 patients (62%) reaching end-stage kidney disease at a median age of 60 years (36 to 82); the mean rate of glomerular filtration loss, in the 23 patients with a baseline value above 20 ml/min assessed for this purpose, was 11.5 ml/min per year (0.5 to 27.7), more than double that observed in the contemporary cohort from the same centre, and renal progression was slower than in AL amyloidosis, in which the median time from diagnosis to end-stage kidney disease is 7.5 to 14 months, but substantially faster than in AApoAI [56]. Estimated survival in that cohort was 15.2 years (0 to 25.2) from clinical presentation and 10.9 years from diagnosis, with 17 of the 71 patients (24%) dead at a median age of 67 years [56]. In the North American series, renal survival at 1, 2 and 5 years from the diagnostic biopsy was 44%, 33% and 11%, and patient survival at the same time points 92%, 82% and 64% [1]. The difference in renal survival between the series should be read in light of the chronic kidney disease stage at the time of biopsy, more advanced in the American series, in which four of the 13 patients were already in stage 5.

Extrarenal amyloid is frequent and, in most cases, clinically silent. In the British cohort, in which iodine-123-labelled serum amyloid P component scintigraphy was performed at least once in 192 of the 194 patients (99.0%), splenic amyloid was identified in 155 (80.7%), adrenal in 33 (17.2%) and hepatic in 12 (6.3%); three of the 12 with hepatic amyloid had a progressive rise in alkaline phosphatase and gamma-glutamyltransferase attributable to amyloid, but, despite a median follow-up of 9.8 years, none developed failure of hepatic synthetic function [3]. Spontaneous splenic rupture has been described: it occurred in two of the 18 patients with E526V identified in the British reclassification study and in three of the nine patients who underwent combined liver–kidney transplantation, in the perioperative period [29].

Cardiac involvement in AFib is a subject of divergence between series, and its resolution is clinically important. The King's College series, with 22 patients assessed for combined transplantation, described the disease as systemic amyloidosis with visceral, vascular, cardiac and neurological involvement: coronary atherosclerosis in 68%, severe systemic atheromatous disease in 55%, autonomic neuropathy in half the patients and autonomic gastrointestinal involvement with constipation or diarrhoea, early satiety and nausea or slow gastric emptying in 15 patients (68%); vascular atheroma excised at endarterectomy and endomyocardial biopsies contained amyloid composed purely of variant fibrinogen [45, 57]. The nephrology review adds that cardiac involvement may occur in the form of atheromatous plaque disease composed of mutant Aα-chain material, which may even precede proteinuria and renal dysfunction [39]. The contemporary cohort of 194 patients reaches a different conclusion: echocardiography did not show cardiac amyloidosis in most; 12 of the 194 patients (6.2%) had been considered, at the referring centres, to have compatible features, such as thickened walls and abnormal strain, but detailed assessment at the reference centre revealed that these were not typical and were attributable to other causes — all 12 had a long history of hypertension and ten were on haemodialysis with volume overload —, and in 12 patients who underwent cardiac magnetic resonance imaging, a modality that distinguishes uraemic cardiomyopathy from cardiac amyloidosis, there was no evidence of cardiac amyloidosis in any case [3]. The authors conclude that, despite the presence of extrarenal deposits, mainly splenic, AFib is an amyloidosis syndrome clinically limited to the kidney [3]. In the historical series of 71 patients from the same centre, there were no typical cardiac findings on non-invasive tests, two patients underwent endomyocardial biopsy and only one had any deposit [39]. The reasonable interpretation is that myocardial and vascular deposition may occur and be histologically demonstrable, but does not produce, in the great majority of patients, a clinically significant amyloid cardiomyopathy, and that the echocardiographic changes in these patients should first and foremost be attributed to hypertension, coronary disease and the volume overload of chronic kidney disease.

Cardiovascular morbidity, on the other hand, is high, as is the rule in chronic kidney disease: in the British cohort, 39 of the 194 patients (20.1%) had angiographically confirmed ischaemic heart disease during their lifetime, including 16 who underwent percutaneous coronary intervention and 6 coronary artery bypass surgery; at least 9 strokes or transient ischaemic attacks, 2 carotid endarterectomies and documented peripheral vascular disease in 3 patients were recorded [3]. Overall survival is long compared with other systemic amyloidoses: in the British cohort, median survival from diagnosis was 12.7 years (95% confidence interval 10.0 to 19.1) in patients diagnosed before 2009 and 12.2 years (9.85 to not reached) in those diagnosed afterwards, with no significant difference (p=0.69); 50 of the 75 patients in the earlier cohort died, at a median age of 72 years (56 to 88), and 32 of the 119 in the recent cohort, at a median age of 69 years (48 to 85) [3]. In the historical series, patients with end-stage kidney disease had a median survival of 8.2 years (0.2 to 24.8) from the start of replacement therapy and of 9.3 years when the analysis was censored at transplantation, substantially higher than that of patients with treated or untreated AL amyloidosis and no different from that of British patients with non-diabetic nephropathy under 55 years of age, although the median age at starting dialysis was 60 years [56].

A family history is notably often absent, and its absence does not exclude the diagnosis: among the 18 patients with E526V identified among 350 patients with presumed AL amyloidosis, none was aware of a relevant family history, and one patient only retrospectively found out that her dizygotic twin had died of renal failure at 76 years [29]. In the North American series, a family history of amyloidosis or chronic kidney disease was present in 5 of the 13 patients (42%) [1].

3.3 Apolipoprotein A-I amyloidosis

AApoAI is the most heterogeneous of the six entities, because the phenotype depends on the variant. The spectrum includes amyloid neuropathy, nephropathy, liver disease and cardiomyopathy, and deposition may occur in the skin, in the larynx, causing hoarseness, and in the testis, causing infertility [6]. Some variants give extensive visceral deposits and end-stage renal failure in young adults, and others only laryngeal or cutaneous amyloid of little clinical consequence [7, 52].

The series from the UK national centre, which brought together all patients presenting between 1986 and 2019, describes 57 patients with 14 different APOA1 variants: the median age at presentation was 43 years and the median time from presentation to referral 3 years (0 to 31 years), a direct measure of diagnostic difficulty; amyloid involvement of the kidney, liver and heart was detected in 81%, 67% and 28% of patients [5]. Renal amyloidosis was universal in carriers of the most frequently identified variant, p.Gly26Arg (28 patients) [5]. Across all variants, patients with renal amyloidosis had, at diagnosis, a median creatinine of 159 µmol/L, versus 103 µmol/L in patients without renal involvement (p=0.0004), a median estimated glomerular filtration rate of 37 versus 67 ml/min (p=0.0001) and median proteinuria of 0.3 g per 24 hours, versus 0.15 g, a difference that did not reach statistical significance (p=0.06) [5]. This proteinuria value is much lower than in AFib and has a direct practical consequence: chronic kidney disease without significant proteinuria is compatible with AApoAI, unlike AL, AA and AFib, in which the presentation is typically nephrotic, and in the Italian series of the Leu75Pro variant only 12% of affected patients had more than 0.5 g per day of proteinuria [5]. Of the 46 patients who presented with renal involvement, 21 reached end-stage kidney disease, and the median time to that point was 19.0 years (95% confidence interval 7.5 to 30.6) from presentation and 15.0 years (10.0 to 20.0) from diagnosis, which defines the disease as slowly progressive [5]. At diagnosis, the median estimated glomerular filtration rate for the cohort as a whole was 41 ml/min, median alkaline phosphatase 145.5 IU/L, median N-terminal pro-B-type natriuretic peptide 355 ng/L and median troponin T 10 pg/mL; the visceral amyloid load on iodine-123-labelled serum amyloid P component scintigraphy was graded as absent in 12 patients (23%), small in 12 (23%), moderate in 13 (25%) and large in 16 (30%), not having been performed in 4 [5]. Median overall survival from diagnosis was 27.0 years (21.3 to 32.7), the longest of all the entities covered in this article [5].

The phenotypes described in the founder families illustrate the breadth of the spectrum. In the Iowa kindred, with five generations of English, Scottish and Irish ancestry, eight members of the third generation had clinical and pathological evidence of amyloidosis and four others were probably affected: neuropathy dominated the clinical picture in the early phase and nephropathy in the late phase, the mean age at onset was about 35 years and mean survival after onset about 12 years, with death attributable in most cases to renal amyloidosis; there was severe peptic ulcer disease in some, frequent hearing loss and cataracts in several, without vitreous opacities [6]. In a four-generation English family with the L60R variant, the proband was completely asymptomatic at 24 years, when extensive splenic and hepatic amyloidosis was discovered on radiolabelled serum amyloid P component scintigraphy, and later developed progressive hypertension, thrombocytopenia and easy bruising, with normalisation of platelets, cessation of bruising and better control of hypertension after splenectomy performed for left upper quadrant pain; another member presented with renal amyloidosis [6]. In a man who was the only child of Ashkenazi Jewish parents, with the W50R variant, presentation was at 34 years with microscopic haematuria detected on routine testing, with mild hepatomegaly, moderately raised alkaline phosphatase, trace proteinuria and moderately impaired renal function; renal and hepatic function deteriorated slowly, intermittent explosive diarrhoea appeared, renal and intestinal biopsies two years later revealed extensive amyloidosis, scintigraphy at eight years indicated massive splenic and hepatic amyloidosis, he started peritoneal dialysis ten years after presentation and had a successful kidney transplant at eleven years, dying two months later with cytomegalovirus infection and liver failure; his father had died at 45 years with histologically confirmed massive hepatic and renal amyloidosis [6]. A large four-generation Spanish family with the exon 4 deletion and insertion had a unique hepatic presentation with death from liver failure generally by the sixth decade: the deposits were extremely widespread at autopsy, but the presentation, subsequent disease and death related exclusively to massive hepatic involvement, and, as in most systemic amyloidoses, there was an asymptomatic prodromal phase lasting up to 15 to 20 years, with only biochemical abnormalities [6].

Two predominantly cutaneous and cardiac phenotypes were described by Hamidi Asl in 1999. In the family with the L90P variant, the proband was 54 years old and presented with recent exertional dyspnoea and long-standing yellow maculopapular skin lesions, which first appeared on the forehead and spread rapidly to the face, neck, shoulders and axillary and antecubital areas; she had cardiomegaly, right bundle branch block, concentric left ventricular wall thickening with a small cavity, a typical restrictive haemodynamic pattern at cardiac catheterisation and amyloid deposits on endomyocardial biopsy, and died six years later with progressive heart failure; a 57-year-old second cousin had three years of extensive maculopapular cutaneous amyloidosis, with petechial purpura on the skin, ocular conjunctiva, tonsillar pillars, buccal mucosa and lips, and four years later developed exertional dyspnoea and hoarseness, with positive laryngeal, oesophageal and rectal biopsies, persistently normal clinical and electrical neurological examination, cardiomegaly, signs of left ventricular hypertrophy on the electrocardiogram, thickening of the tricuspid and mitral valves and pericardial effusion on the echocardiogram, dying 14 years after the initial presentation [6]. In the American family with the R173P variant, expression was mainly cutaneous and cardiac: the proband, aged 33, was referred to dermatology for a diffuse rash resembling acanthosis nigricans in the axillae, whose Congo red-stained biopsy revealed amyloid deposits; her father had cerebral aneurysms at 37 years, was later diagnosed with multiorgan systemic amyloidosis and died at 63 years with cardiomyopathy and hepatic and renal failure; two sisters had a brown axillary rash since the age of 20, with petechiae and skin thickening of the hands; a third sister, not assessed, had a hoarse voice, a symptom shared by several members; a 71-year-old paternal aunt had voice changes due to biopsy-proven vocal cord deposition, cutaneous amyloid and hypertrophic cardiomyopathy on echocardiography; and the paternal grandmother had a diagnosis of cardiac and vocal cord amyloidosis [6].

A distinct renal phenotype, of tubulointerstitial nephritis, was described in five Italian families carrying the Leu75Pro variant in the heterozygous state: the presentation was defective urinary concentrating ability, moderate polyuria, negative urinalysis and mild tubular proteinuria, and the amyloid was restricted to non-glomerular regions and limited to the renal medulla, unlike other systemic amyloidoses; renal involvement was almost universal, with highly variable severity [58]. The practical implication is direct: renal biopsy may be falsely negative if it does not include medulla [58]. The medullary and tubulointerstitial pattern was also observed in the three Irish families that account for all known cases of chronic kidney disease due to this entity in that country, all carrying the Gly26Arg variant in the heterozygous state, with demonstrable kidney disease in the great majority of those affected, variable liver disease, usual presentation as slowly progressive renal failure with mild proteinuria, histology with tubulointerstitial fibrosis and amyloid deposition in the renal medulla, and very high penetrance; in one of these families, severe and disabling peripheral polyneuropathy was common [59]. The same pure interstitial pattern, without glomerular, vascular or tubular deposits, was described in the only report of recurrence of apolipoprotein A-I amyloidosis in a renal allograft, in a 63-year-old woman with no family history, with slowly progressive graft dysfunction and no overt proteinuria, in whom proteomics identified apolipoprotein A-I and sequencing revealed two new variants, Leu202Arg and Lys262Asn; the authors of that report note that tubulointerstitial injury was diagnosed in half of the carriers of a variant in the gene and that only 13% of patients progressed to renal failure requiring haemodialysis or transplantation, recurrence after kidney transplantation being very rare [60].

The pure cardiac phenotype is documented with the L174S variant: a man from a family in central Italy presented at 48 years with asthenia and exertional dyspnoea, a testicular biopsy for infertility six years earlier having already shown amyloid deposits; lipid studies revealed a significant reduction in serum apoA-I and HDL cholesterol in the patient but not in unaffected relatives; abdominal fat biopsy was Congo red positive; echocardiography at presentation revealed hypertrophic cardiomyopathy with a thickened interventricular septum and an ejection fraction of 40%; progression to heart failure refractory to medical treatment led to heart transplantation at 56 years, after which the patient was described as in excellent condition and without myocardial deposits at nine months, with mild renal impairment; a maternal uncle, similarly affected, died at 52 years of heart failure [6]. In 2026 a case with the p.L202P variant was described that illustrates the same cardiac phenotype and is detailed in section 10.

3.4 Apolipoprotein A-II amyloidosis

The dominant manifestation is renal amyloidosis, with proteinuria or renal dysfunction noted between the end of the third and the beginning of the fifth decade in the men described, progressing to renal failure [10]. Cardiac deposition is documented: in the patient with the Stop78Arg variant, a Caucasian man with progressive renal dysfunction from the age of 34 and amyloidosis on renal biopsy at 56 years, echocardiography showed mild thickening of the interventricular septum and technetium-99m-labelled pyrophosphate scintigraphy demonstrated uptake by cardiac muscle, compatible with myocardial deposition; his father died of renal failure and a paternal half-brother had renal dysfunction [11]. In the patient with the Stop78Gly variant, a Caucasian man developed proteinuria at the end of the third decade, was diagnosed with renal amyloidosis on biopsy at 33 years, started continuous ambulatory peritoneal dialysis at 45 years and received a deceased-donor kidney transplant at 47 years, maintaining stable renal function for nine years without other evidence of organ dysfunction attributable to amyloidosis [12]. Death in this form is due to renal failure [12].

3.5 Lysozyme amyloidosis

ALys is a disease of the gastrointestinal tract, liver and kidneys, with a slow natural history [16]. The largest published cohort brings together all patients diagnosed at a UK national centre between 1988 and 2025: of the 38 identified, 3 did not return for assessment after the genetic diagnosis, so the detailed analysis covers 35, with a median age at diagnosis of 45.2 years (15.5 to 75.7) and 14 men (40%) [15]. The cohort is exclusively of white European ancestry — 26 British, seven Australian, four American and one Romanian —, 30 patients (86%) had a family history and 27 (77%) belonged to eight distinct families, a limitation on generalisation that the authors acknowledge [15]. The central finding is that symptom onset is much earlier than the age at diagnosis suggests: the median age at first reported symptom was 23.7 years (8.0 to 72.5) and 15 patients (43%) recalled symptoms in childhood or adolescence [15]. The earliest manifestation was sicca syndrome, in 16 patients (46%), at a median age of 15 years, followed by gastrointestinal symptoms, in eight (23%), at a median of 30 years, and easy bruising, in three (9%), at a median of 10 years; over the course of the disease, 26 patients (74%) reported sicca symptoms, with ocular dryness typically preceding oral dryness, and 11 (31%) easy bruising or a petechial rash [15]. Hepatic involvement occurred in 28 patients (80%) and renal involvement in 20 (57%); overt gastrointestinal bleeding in 11 (31%); eight (23%) developed renal failure requiring replacement therapy, and six received a kidney transplant, without amyloid recurrence in any [15]. In that cohort, 25 of the 35 patients (71%) carried the Asp67His variant, p.(Asp85His), four Ile56Thr, p.(Ile74Thr), two each of the variants Asp87Gly, p.(Asp105Gly), Leu84Ser, p.(Leu102Ser), and Trp64Arg, p.(Trp82Arg), and one Asp67Gly, p.(Asp85Gly) [15]. The amyloid load documented by scintigraphy is high and stable: of the 28 patients (80%) who underwent iodine-123-labelled serum amyloid P component scintigraphy at the first visit, 15 had a large and 12 a moderate total body load, with amyloid identified in the liver in 93%, in the spleen in 89% and in the kidneys in 68%, and a mean liver length of 16.4 cm; in the 22 patients with serial imaging, total body load remained stable in 15 (68%), progressed in five (23%) and partially regressed in two (9%) over a median imaging interval of 7.6 years (1.0 to 27.9) [15]. Cardiac and neurological involvement are rare, unlike in transthyretin and light-chain amyloidosis: cardiac amyloid was considered only possible in a single patient, a carrier of the Leu84Ser variant, with equivocal echocardiographic findings, a negative bone scintigram and underlying obstructive cardiomyopathy, and three patients (9%) reported length-dependent sensory neuropathy and three (9%) carpal tunnel syndrome, with lymphadenopathy in five (14%) and respiratory symptoms in eight (23%), two of whom had established smoking-related chronic obstructive pulmonary disease [15]. Reproductive outcomes were preserved: 15 women and four men in the cohort reported having had children, a total of 32 children, one woman with the Asp67His variant completed two successful pregnancies after an emergency liver transplant, one had six children in six pregnancies, there was no obstetric haemorrhage or uterine rupture, and the only complication recorded was one case of intrahepatic cholestasis of pregnancy — preservation that the authors relate to the absence of lysozyme expression in the reproductive organs [15]. The authors conclude that the disease combines early symptom onset, indolent systemic progression and a risk of catastrophic hepatic or splenic rupture, with implications for surveillance, transplant planning and the design of future therapeutic studies [15]. The most common manifestations, in the German cohort of ten patients from six families, were renal dysfunction, gastrointestinal symptoms and sicca syndrome [14]. In the earlier series from the same British centre, with 16 patients, symptomatic gastrointestinal amyloid was prevalent and macroscopically visible amyloidotic lesions were present in nine of ten patients who underwent gastrointestinal endoscopy; all symptomatic individuals had hepatic amyloid [16].

The most severe and most characteristic manifestation is spontaneous hepatic haemorrhage, with liver rupture. Harrison and colleagues described a family in which it occurred in three generations: the proband, a 15-year-old boy, underwent emergency laparotomy for massive intra-abdominal haemorrhage related to a hepatic haemorrhage and a large capsular haematoma; at the time of transplantation, the liver was described as fragile and friable, with loss of the reticulin framework, and histological examination revealed a surprisingly modest amount of amyloid, shown immunohistochemically to be lysozyme-derived; the father had undergone surgery for massive hepatic haemorrhages and died of uncontrollable intra-abdominal haemorrhage, and the paternal grandfather also died shortly afterwards of uncontrollable hepatic haemorrhage [34]. In the British cohort, three of the four patients who reached liver transplantation were transplanted for spontaneous hepatic rupture [16].

Renal involvement is common but slowly progressive, and its presentation is that of subnephrotic-range proteinuric chronic kidney disease: of the 20 patients with renal involvement in the 2026 cohort, only one was frankly nephrotic, the median estimated glomerular filtration rate at the first visit was 58 ml/min per 1.73 m² (20 to more than 90) and the median urinary protein/creatinine ratio 55 mg/mmol (0 to 5000); in patients who had not reached end-stage kidney disease by the census, the median rate of decline in estimated glomerular filtration rate was only −0.7 ml/min per 1.73 m² per year over a median follow-up of 8.8 years, an order of magnitude below that observed in AFib [15]. Eight of the 35 patients (23%) progressed to end-stage kidney disease, at a median of 10.1 years from amyloid presentation, and received replacement therapy — three haemodialysis, three peritoneal dialysis and two pre-emptive living related donor kidney transplantation; five of the 16 patients in the earlier series from the same centre had progressive amyloidotic renal dysfunction to that point [15, 16]. The renal histological pattern is distinctive: in Lanham's original family, with six affected members in two generations and onset of kidney disease between 23 and 45 years, deposition was characteristically interstitial and not glomerular, unlike other forms of amyloidosis [34]. In other families, deposition was glomerular, arteriolar and in the tubular basement membranes: in the Italian-Canadian family with the F57I variant, the proband developed renal failure at 42 years and had a kidney transplant at 44, developed ischaemic heart disease at the end of the fourth decade and died at 52 years of postoperative complications of triple coronary bypass; her daughter had hypertension at 17 years, with severe amyloid deposition in the glomeruli and arterioles, had a kidney transplant at 23 years and remained well over the following 13 years; and her sister had proteinuria at 35 years and kidney transplantation at 43 [34].

Sicca syndrome may be the earliest manifestation and precede kidney disease by decades. In the French family with the W64R variant, the 66-year-old proband had ocular and oral sicca syndrome since the age of 14; at 47 years he underwent emergency laparotomy for suspected small bowel obstruction, and histology revealed amyloid deposits in the lymph nodes and in the muscular layer of the small bowel, as well as in the vascular wall of the rectum, along the hepatic sinusoids and in the bone marrow; renal failure was detected at 54 years and was progressive, without signs of peripheral neuropathy or cardiac involvement, and he developed keratitis secondary to sicca syndrome; three first cousins were affected, all with kidney disease and two with ocular sicca, and two individuals of the younger generation were affected, one with oral sicca and amyloid deposition in the labial salivary gland [34]. Petechial rash is another finding described: in Zalin's English family, it was marked and petechial haemorrhages were induced by minimal abrasion, with extensive amyloid deposition in the lungs; in that family, all members with amyloidosis had had a petechial rash since childhood, as had some from earlier generations, and the proband's mother died of renal failure and had petechiae [34].

Gastrointestinal involvement may be the isolated presentation and lead to diagnostic error. Granel and colleagues described a 59-year-old Italian woman diagnosed with systemic gastrointestinal and bone marrow amyloidosis, who had undergone partial gastrectomy for presumed peptic ulcer; immunohistochemistry identified lysozyme as the fibril protein, there was no evidence of autonomic neuropathy and ophthalmological assessment was normal; her father had gastrointestinal amyloidosis diagnosed at 67 years, with marked weight loss, abdominal pain and malnutrition; the authors noted that the patient could have been misdiagnosed as having acquired AL amyloidosis and received dangerous and inappropriate chemotherapy had the fibril type not been characterised and had the family history not been obtained [34]. Peripheral polyneuropathy is not typical of this entity, but was observed in a mother and son with a new variant described in the German cohort [14]. The disease may manifest at an early age [14].

3.6 Gelsolin amyloidosis

AGel differs from the others in a phenotype that begins in the eye and in survival that does not differ from that of the general population until advanced age. The classical triad is ophthalmological, neurological and dermatological [20, 21]. In the Finnish national registry of 227 patients, the first symptom was ophthalmological in 167 (73.6%), at a mean age of 39 years; corneal lattice dystrophy was reported at a mean of 43 years, and decreased vision, polyneuropathy, facial nerve palsy and cutis laxa appeared at a mean of between 52 and 57 years; carpal tunnel syndrome was reported by 86 patients (37.9%), nine (4.0%) had a pacemaker and 12 (6.1%) cardiomyopathy; apart from corneal dystrophy, no salient difference in age at onset was found between the main symptoms, and carpal tunnel, pacemakers and cardiomyopathy were more common than in the general population [19].

In the update of the same registry, with 261 patients and survival analysis in 478, symptom prevalences were: dry eye in 93% and corneal lattice amyloidosis in 89%; numbness, tingling and other paraesthesiae in 75% and facial palsy in 67%; drooping eyelids in 86% and cutis laxa in 84% [18]. Systemic manifestations were less frequent: cardiac arrhythmias reported by 15% of patients, with a pacemaker in 5%, proteinuria in 13% and renal failure in 5% [18]. The surgical burden is considerable: 65% of patients had undergone skin or soft-tissue surgery, 26% carpal tunnel surgery and 24% at least unilateral cataract surgery [18]. Relative survival exceeded 1 for men and women up to the 70 to 74 years age group, in which it was 0.96, that is, the disease does not shorten life expectancy at least in the first 75 years, which the authors attribute in part to the rarity of severe renal and cardiac manifestations compared with other systemic amyloidoses [18]. Cranial neuropathy, especially progressive bilateral facial palsy, is the characteristic neurological manifestation: in Kiuru's clinical series of 30 patients, there were signs of cranial neuropathy, mainly affecting the facial nerve, in all, and peripheral polyneuropathy mainly impairing vibration and touch in 26; autonomic dysfunction was minor in most, and clinically significant autonomic dysfunction or heart disease were not characteristic [20].

The facial presentation is recognisable. In the patient described by Sack, facial paralysis began at about 56 years as an inability to control the drooping lower lip, which became markedly protruding and everted with exposure of the lower gingival mucosa; five years after onset he could not wrinkle his forehead and had intermittent fasciculation on the right side of the lower lip, with minimally affected extraocular muscles and no ptosis; skin laxity was marked, to the point of raising the question of cutis laxa, and the slit lamp showed bilateral lattice corneal opacity; his mother had an identical picture with onset at the same stage of life, and the proband had bulbar manifestations [20]. Two families described in 2021 add details: in a family of six members, there was corneal dystrophy in five, lax skin in five and cardiac arrhythmia requiring an implantable cardioverter-defibrillator in three, and one patient had bilateral optic neuropathy, possibly related to the disease and critical to recognise before a corneal transplant, with eyelid skin histopathology showing atrophic epidermis and dermis and amyloid deposits around the hair follicles [20]; in two unrelated families, one with an affected woman with no family history and the other with an affected father, son and daughter, all four patients had polymorphic corneal stromal opacities, dermatochalasis requiring upper blepharoplasty and dry eyes, the father and son had open-angle glaucoma, and systemic manifestations included cutis laxa in all four, proteinuria in the daughter, sinus bradycardia in the father, supraventricular tachycardia and idiopathic cardiomyopathy in the daughter and hypothyroidism in the son and daughter [20].

Clinically significant renal involvement is rare in this entity, with renal failure observed in only 3% of the Finnish cohort of 227 patients and in 5% of the updated registry [18, 48]. When it occurs, it typically manifests in older patients, who frequently have comorbidities such as hypertension and diabetes, and the disease has a more indolent course and a more favourable prognosis than other systemic amyloidoses [48]. Homozygosity radically alters the picture: the two sisters homozygous for D187N described by Maury had unusually severe disease, manifesting with nephrotic syndrome and end-stage renal failure [20]. Of the 207 patients identified in the 1973 Finnish national survey, some developed nephrotic syndrome and renal failure and some had cardiac involvement, and amyloid involvement was quite widespread at autopsy [20].

3.7 Variant β2-microglobulin amyloidosis

In the three-generation French index family, the picture was one of slowly progressive gastrointestinal symptoms with autonomic neuropathy, with onset around 50 to 60 years: patients developed slowly progressive chronic diarrhoea with weight loss and sicca syndrome, one had sensorimotor axonal polyneuropathy and orthostatic hypotension and two had severe autonomic neuropathy; the authors observed that the gastrointestinal problems resulted both from direct deposition in the gastrointestinal tract and from autonomic neuropathy [23]. Post-mortem examination of a patient who died at 70 years showed extensive deposits containing β2-microglobulin in the spleen, liver, heart, salivary glands and nerves, and a colonic biopsy from another affected individual also contained deposits; scintigraphy in two patients showed a high visceral load in the spleen and adrenals, but not in the heart [23].

The Portuguese family with the P32L variant had a dominant cardiac phenotype. The proband, a 59-year-old woman, had progressive exertional dyspnoea and early satiety since the fifth decade; bilateral carpal tunnel syndrome developed at 45 years and was only decompressed 14 years later; at 57 years paroxysmal atrial fibrillation was diagnosed and at 59 years coronary angiography showed a 40 to 50% left main stem lesion; she had interstitial lung disease, normal renal function, low serum β2-microglobulin concentration and mild length-dependent sensory peripheral neuropathy, with endomyocardial biopsy and abdominal fat aspirate positive for amyloid [23, 25]. Her brother was diagnosed with cardiac amyloidosis at 68 years, one year after the proband, and had bilateral carpal tunnel and chronic obstructive pulmonary disease; a cousin was diagnosed at 60 years, three years before the proband, was transplanted, and also had carpal tunnel and nodular lung disease; two other relatives had a history of amyloidosis [23]. Haslett and colleagues described a 63-year-old man of Portuguese ancestry, with a family history of amyloidosis, who presented with exertional dyspnoea and lower limb oedema and had unoperated bilateral carpal tunnel; cardiac magnetic resonance suggested transthyretin amyloidosis, but endomyocardial biopsy showed β2-microglobulin amyloidosis, and the heart explanted at transplantation confirmed the deposition [23, 61].

The laboratory finding that separates this entity from the dialysis form is the serum concentration of the protein: in the dialysis-related form, β2-microglobulin is greatly increased, whereas in the hereditary form it is slightly low or normal, and patients have normal renal function [23, 24]. In the patient with the P32L variant, almost equal amounts of variant and wild-type protein circulated, and the deposits were composed exclusively of the variant [25].

4. Clinical Recognition and Red Flags

4.1 When to suspect systemic amyloidosis

Suspicion of amyloidosis generally arises from one of three contexts. The first is proteinuric nephropathy of unexplained cause, especially in a middle-aged adult with hypertension, in whom renal biopsy reveals congophilic deposits: systemic amyloidosis is the diagnosis in 2.5% of all renal biopsies, according to the British glomerulonephritis registry cited in OMIM [29]. The second is infiltrative cardiomyopathy, suspected from clinical symptoms, left ventricular hypertrophy on echocardiography or magnetic resonance imaging and, in particular, from the discordance between hypertrophy on imaging and the absence of hypertrophy on the electrocardiogram [32]. The third is the combination of apparently unrelated multisystem manifestations — bilateral carpal tunnel syndrome, dysautonomia, chronic diarrhoea, macroglossia, periorbital purpura, biceps tendon rupture — in a patient with some degree of organ dysfunction [32].

Suspicion of a hereditary form, and in particular of a hereditary form not mediated by transthyretin, requires an additional step of reasoning, because in most cases these entities have no exclusive clinical sign. The European Society of Cardiology position statement organises red flags by amyloid type, and those relevant to the forms covered here are renal failure and proteinuria, which point to AApoAI, AApoAII, AApoAIV, Aβ2M and AFib; skin discolouration, which points to AApoAI; cutis laxa and corneal lattice dystrophy, which point to AGel; and family history, which points to ATTRv, AApoAI and AApoAII [28]. The cardiac red flags — low QRS voltage relative to wall thickness, atrioventricular conduction disease and pseudoinfarct pattern — apply to all types and do not discriminate between them [28]. In 2026, the authors of a case of cardiac AApoAI pointed out that, in the European statement, the advice on when to test for the rare subtypes remains vague, being limited to mentioning concomitant kidney disease, and that the 2023 North American consensus document refers to dialysis status and family history, and argued for the need for more specific guidance, namely in algorithms, on when to look for rare amyloidoses [8, 62].

4.2 Red flags by entity

In AFib, the clinical profile of suspicion is relatively homogeneous: an adult between the fifth and seventh decade with nephrotic or subnephrotic-range proteinuria, arterial hypertension and progressive loss of glomerular filtration, often without any other manifestation of amyloid disease; the finding of splenic amyloid on scintigraphy or of splenomegaly is an additional clue, and spontaneous splenic rupture, although rare, is a specific red flag [3, 29]. The absence of a family history should not rule out the diagnosis, given that none of the 18 patients with E526V identified in a cohort of presumed AL amyloidoses was aware of one [29]. With frameshift variants, the disease may begin in adolescence: in the French series, the mean age at onset of kidney disease with these variants was 24.5 years, with cases from 12 years of age, so a familial proteinuric nephropathy of juvenile onset requires this entity to be considered [4].

In AApoAI, the extrarenal signs are what give the clue: hoarseness due to laryngeal amyloid, yellow maculopapular skin lesions or lesions resembling acanthosis nigricans in the axillae and neck, petechial purpura of the mucosae, male infertility with testicular deposition, hepatomegaly with cholestasis and adrenal insufficiency [6, 28]. In a patient with infiltrative cardiomyopathy, the combination of bilateral carpal tunnel, dysphonia and chronic kidney disease of undetermined cause, with a family history of amyloidosis, is the set that led to the diagnosis in the case described in 2026 [8]. A family history is often absent: only one of four patients in one series had it, so the authors recommend sequencing the apolipoprotein A-I gene in patients with apparently localised amyloidosis in whom immunohistochemistry is not diagnostic of the fibril protein, even in the absence of a family history [7].

In ALys, the red flags are early-onset sicca syndrome, which may precede kidney disease by decades, gastrointestinal complaints with weight loss and abdominal pain, petechial rash with capillary fragility since childhood and, above all, any episode of spontaneous intra-abdominal or hepatic haemorrhage in a young adult or adolescent, which constitutes an emergency sign [16, 34]. A family history is the rule in this entity, unlike in AFib and AApoAI [16].

In AGel, suspicion is typically ophthalmological and early. Corneal lattice dystrophy is the first ophthalmological sign to appear, is practically a prerequisite for the diagnosis and, accompanied by a positive family history, is sufficient to establish it [20]. The classical sequence is dry eye and lattice corneal opacities in the fourth decade, followed by progressive facial palsy, drooping eyelids, cutis laxa and polyneuropathy in the sixth decade [18, 19]. A red flag of operative urgency is bilateral optic neuropathy, whose identification before a corneal transplant is critical [20]. In atypical cases, the disease may present with isolated nephropathy, without ocular, neurological or cutaneous manifestations and without a family history, as in the case described in section 10 [48].

In hereditary Aβ2M, the decisive red flag is the dissociation between the clinical picture of amyloidosis and the renal context: the patient has normal renal function and normal or low serum β2-microglobulin, with no history of dialysis, which excludes the acquired form [23, 24]. The phenotype combines chronic diarrhoea with weight loss and dysautonomia, sicca syndrome, bilateral carpal tunnel syndrome and cardiomyopathy, and Portuguese ancestry is, in light of the families described, a contextual element to be taken into account [23, 25].

In AApoAII, the picture is one of proteinuric nephropathy with onset between the third and fifth decade and a family history of renal failure, which may be associated with septal thickening and cardiac uptake on pyrophosphate scintigraphy [11, 12].

4.3 Newborn screening and population screening

None of these entities is the subject of newborn screening in any programme, and none would be indicated for it: all have adult onset, none has effective presymptomatic therapy, and the targeted therapy available — organ transplantation — only applies at the stage of organ failure. The screening that has a place in these diseases is cascade family screening after identification of an index case, covered in section 8.

4.4 The attribution error and its two directions

Clinical recognition of these entities is inseparable from recognising that clinical attribution of the amyloid type is unreliable. The error runs in two directions, and both have been documented. In the first direction, a hereditary amyloidosis is taken for AL: in the study of 350 patients with a presumed clinical diagnosis of AL, 34 (9.7%) had amyloidogenic mutations, and in 8 of these 34 (24%) there was a low-grade monoclonal gammopathy that falsely supported the diagnosis [40]. The consequence is the administration of chemotherapy to people without a plasma cell dyscrasia and the loss of the opportunity for potentially curative transplantation [6]. In the second direction, AL amyloidosis is taken for something else, including hereditary amyloidosis, even when there is genetic confirmation. In the Portuguese series of five diagnostically difficult patients, a 61-year-old man with peripheral neuropathy, cardiac involvement, a positive family history and a pathogenic genetic test for transthyretin amyloidosis, whose salivary gland biopsy showed amyloid deposits, had mass spectrometry identifying AL lambda amyloidosis, which raised the possibility of concomitant transthyretin and AL lambda amyloidosis and prompted haematological reassessment [63]. In the same series, the presence of a monoclonal gammopathy did not reliably predict the amyloid subtype in three patients [63]. The practical rule that follows from these two sets of observations is that typing must be done, and that a monoclonal protein does not prove AL amyloidosis nor does a positive genetic result exclude a concomitant acquired amyloidosis.

5. Diagnosis: Criteria, Tests and Algorithms

5.1 Criteria and logical sequence

The diagnosis of a hereditary systemic amyloidosis requires three linked and non-substitutable steps: demonstrating amyloid in tissue, identifying the protein that makes up the fibrils and identifying the genetic variant that produces it. None of the three is dispensable, and each answers a different question — whether amyloid is present, what protein it is made of and which gene and variant are responsible.

Amyloid is demonstrated by biopsy with Congo red staining and examination under polarised light, looking for the characteristic birefringence of the deposits; the ISA Nomenclature Committee recommends abandoning the expression “apple-green birefringence” in favour of a more neutral description, because in practice the colour varies and is usually a mixture of green, red, orange and yellow [31, 32]. For cardiac amyloidosis, the European Society of Cardiology position statement establishes that an endomyocardial biopsy with Congo red-positive amyloid confirms the diagnosis regardless of wall thickness, and that amyloid deposits on an extracardiac biopsy also confirm it if echocardiography or magnetic resonance imaging show typical features and there is no other cause of wall thickening; the non-invasive criteria, based on bone scintigraphy with negative monoclonal protein, are accepted only for transthyretin amyloidosis [28]. This asymmetry has a direct practical consequence: in the rare forms, including the six covered here, biopsy is mandatory. If bone scintigraphy and monoclonal protein tests are both negative, an alternative diagnosis should be considered, but, if suspicion persists, cardiac magnetic resonance imaging and then biopsy are recommended, because some transthyretin variants and the rare subtypes may not take up the tracer [28].

5.2 Typing of the fibril protein

Typing is the step that changed practice in this class of diseases. The methods in routine use until the end of the 2000s, including immunohistochemistry, have low specificity and sensitivity [64]. The combination of laser microdissection with tandem mass spectrometry, validated on routine clinical biopsy specimens, identified the amyloid type with 100% specificity and sensitivity in a training set of 50 cases well characterised by reference clinicopathological criteria and with 98% in an independent validation set of 41 cases of cardiac amyloidosis [64]. The advent of mass spectrometry-based proteomics revolutionised typing and led to the discovery of new amyloid types, and exact typing of the precursor protein is of paramount importance because the type dictates a specific therapeutic approach [30]. The European statement considers mass spectrometry the reference standard for typing, admitting immunohistochemistry and immuno-electron microscopy as methods in routine use in specialised centres [28]. When endomyocardial biopsy shows amyloid, immunostaining for light chains and for transthyretin may be positive for both, and definitive attribution of AL or ATTR on that basis is limited, whereas liquid chromatography with mass spectrometry determines the subtype more definitively and sometimes identifies mutant peptides [32].

The proteomic protocol of a national reference centre, described for AFib, gives a measure of the rigour required: amyloid deposits from the renal biopsy are laser-microdissected and digested with trypsin, the analysis is run on a hybrid ion trap–Orbitrap spectrometer, and the raw spectral data are searched against the Swiss-Prot human protein database; amyloid is identified if two of three amyloid signature proteins are present — apolipoprotein E, apolipoprotein A-IV and serum amyloid P component; typing is assigned to AFib if the fibrinogen score is above 80 with a minimum of two significant unique peptides matching the human fibrinogen sequence; and, after the initial search, a library with the peptide sequences of the known amyloidogenic FGA variants is added to the database, which allows the specific variant to be identified from the deposit itself [3]. A panel of experts in amyloid diagnosis then analyses the proteomic data together with genetic sequencing, histology and clinical features, to establish the definitive diagnosis [3]. Conventional histology proceeds in parallel: formalin-fixed, paraffin-embedded biopsies, stained with Congo red and examined under cross-polarised light by two independent operators, with immunohistochemistry using a panel of antibodies against serum amyloid A protein, kappa and lambda light chains, transthyretin and the fibrinogen A alpha chain [3].

The yield of immunohistochemistry varies widely between entities, and AFib is a favourable exception. In the historical British cohort, immunohistochemical staining of amyloid with an antibody against the fibrinogen Aα chain was definitive in 93% of patients, with staining absent in all of them for serum amyloid A protein, kappa and lambda light chains and apolipoprotein A-I; in the 7% in whom it was not definitive, all had a previously described amyloidogenic fibrinogen variant, the same characteristic renal morphology and a typical clinical course, so non-definitive immunohistochemistry does not exclude the diagnosis [56]. This yield was not reproduced, however, in the French series of 32 patients, where the sensitivity of routine immunohistochemistry for AFib was only 8% on paraffin-embedded sections and 35% on frozen sections, values that led the authors also to question its specificity [4]. The discrepancy is one of method and centre, not of disease: staining depends on the antibody, the fixation and the laboratory's experience, and the practical consequence is that negative immunohistochemistry for fibrinogen in a compatible renal pattern requires proceeding to mass spectrometry and genetic testing, not abandoning the hypothesis. In the contemporary cohort from the same centre, immunohistochemistry was diagnostic of AFib-type amyloid in 35 of 75 patients (46.7%) from the period before 2009, and proteomic analysis became the predominant method after that date, performed in 72 of 119 patients (60.5%); available specimens from the earlier cohort were processed retrospectively for proteomics, which confirmed the AFib type with the associated variant in 24 of 24 cases (100%) [3]. In other entities, immunohistochemistry is frequently non-diagnostic: in AApoAI, it is reported as often non-diagnostic of the fibril protein, and that is precisely the situation in which sequencing the gene is recommended even without a family history [7, 52]. ALECT2 may be suggested by immunohistochemistry with an anti-LECT2 antibody, but the diagnosis should be confirmed by liquid chromatography with mass spectrometry whenever possible [39].

The size of the diagnostic gain has been quantified in two series. In a blinded comparison between the immunohistochemistry of a UK national centre and the microdissection with mass spectrometry of a North American centre, performed on 142 consecutive biopsies from 38 different tissue types, immunohistochemistry failed to demonstrate the amyloid type in about 20 to 25% of cases, requiring additional tests; concordance between positive immunohistochemistry and mass spectrometry was 100%, and the addition of mass spectrometry raised diagnostic accuracy from 76% to 94% [65]. In the French national cohort of 833 specimens analysed between January 2010 and September 2021, with a two-set design, mass spectrometry identified the type in all 92 specimens of the training set, with 100% concordance with immunofluorescence and immunohistochemistry, and in 95.0% of the 741 specimens of the validation set, sent because of absent or inadequate frozen sample (70.5%), equivocal immunohistochemistry (15.7%), negative immunohistochemistry (9.9%) or a result inconsistent with the clinical picture (3.9%), identifying 19 different types [66]. The most consequential finding of this series is the quantification of immunostaining error in the subgroup in which it was equivocal or negative: in cases of light-chain amyloidosis (101 patients), anti-light-chain antibodies gave false negatives in 48.5% and false positives in 41.5%; in cases of transthyretin amyloidosis (58 patients), the anti-transthyretin antibody gave false negatives in 12.0% and false positives in 56.8% [66]. The authors attribute the phenomenon to commercial antibodies not being optimised for mutated and truncated proteins and to the beta-sheet conformation generating non-specific positivity [66]. In that cohort, the universal amyloid signature — serum amyloid P component, apolipoprotein E and apolipoprotein A-IV — was present in 81.6% of cases, and there was over-representation of apolipoprotein A-I in light-chain amyloidosis compared with transthyretin and AA amyloidosis (64.3% versus 27.0% and 29.3%; p<0.001) and under-representation of apolipoprotein A-IV in AA amyloidosis (56.1% versus 91.6% and 90.4%; p<0.001), differences that persisted after adjustment for age, sex and organ [66].

The experience of a Portuguese reference laboratory shows the added value of mass spectrometry when conventional methods fail or mislead. In a series of five diagnostically difficult patients, in which an approach without laser microdissection followed by proteomic analysis with ultra-high-performance liquid chromatography coupled to high-resolution mass spectrometry was used, mass spectrometry established or revised the subtype in all cases: in a 78-year-old woman with nephrotic syndrome diagnosed on renal biopsy as AA amyloidosis, the absence of underlying inflammatory disease prompted the analysis, which identified AL lambda and led to a change of treatment; in an 86-year-old woman with nephrotic syndrome and IgM lambda monoclonal gammopathy, with deposits without light-chain restriction on conventional tests, mass spectrometry confirmed AL lambda and supported the start of treatment; in a 68-year-old man with chronic kidney disease, proteinuria and IgG lambda monoclonal gammopathy, initially under watchful waiting because bone marrow biopsy showed 11% plasma cells without myeloma-defining events, the renal biopsy with congophilic deposits and negative immunofluorescence and immunohistochemistry, even after pronase digestion, was clarified by mass spectrometry as AL lambda; and in a 78-year-old man with IgA kappa multiple myeloma and severe acute kidney injury requiring dialysis, mass spectrometry of the abdominal fat deposit demonstrated AL lambda amyloid, revealing a discordance between the plasma cell clone detected and the amyloid subtype, a discordance the authors attribute to the possibility of fibril formation by a quantitatively minor but structurally unstable light chain, to the lower specificity of conventional techniques in detecting truncated immunoglobulins, or to biclonal expression [63].

5.3 Genetic testing

Identification of the variant is indispensable and cannot be replaced by proteomics. Microdissection with mass spectrometry identifies the protein present in the deposit, not the gene or the variant: it can say that the deposits are lysozyme, but LYZ sequencing is still needed to identify the causal variant and allow family testing [44]. Clinical guidelines require DNA sequencing of all amyloidogenic pathogenic variants, usually on a peripheral blood sample [1]. In AFib, targeted screening amplifies exon 5 of FGA by polymerase chain reaction and sequences it by an automated method [3].

Access to testing is, in practice, the weak point of the process. Genetic testing for transthyretin amyloidosis is widely available, but testing for the other forms of hereditary amyloidosis — AFib, ALys, AGel, AApoAI, AApoAII, AApoCII and AApoCIII — exists only in a few laboratories [1]. The tissue proteomics laboratory of a North American national reference centre found that genetic testing is not performed in most cases of non-transthyretin hereditary amyloidosis that it diagnoses there by mass spectrometry, despite recommending it in its reports: in the cohort of 13 patients with AFib, peripheral blood testing had been performed in only 3 (23%) [1]. This gap is relevant because the genetic and phenotypic diversity inherent to hereditary amyloidosis makes genetic testing necessary for an accurate diagnosis and for genetic counselling of patients and their relatives [1].

When no blood sample is available, archival tissue can be used. A proof-of-concept study demonstrated for the first time detection of the E545V variant by analysis of DNA extracted from residual formalin-fixed, paraffin-embedded renal biopsy tissue, allowing rapid confirmation of the diagnosis: of 13 cases tested, 7 (54%) gave a positive result with identification of the variant, and the remaining 6 gave no result because of insufficient input DNA (four cases), degraded DNA (three cases) or low coverage of the sequence of interest (four cases), categories that are not mutually exclusive [1]. The limitations of the method are known: DNA from paraffin-embedded tissue is fragmented and partially degraded, its quality, quantity and fragment size depend on the interval between tissue collection and fixation, on the nuclease content of the tissue and on the storage time of the block, and formalin induces nucleic acid cross-links that may give rise to artefactual nucleotide substitutions; in the study, the residual tissue was very small, because the block had already provided 16 or more 3 µm levels for light microscopy, one 10 µm section for Congo red and another 10 µm section for mass spectrometry [1]. The authors speculate that variants in other genes responsible for other forms of hereditary amyloidosis could be detected by this method, but this extension was not tested [1].

5.4 Content of the genetic report and variant interpretation

The report should identify the gene, the transcript with its version, the DNA- and protein-level change in HGVS nomenclature, zygosity and the pathogenicity classification. Dual protein numbering is the main pitfall: the most common AFib variant is p.Glu545Val in precursor protein numbering, which includes the 19-amino-acid signal peptide, and E526V in the mature protein numbering used in almost all of the clinical literature; in the contemporary British cohort, variants were reported as p.Glu545Val (formerly E526V), p.Glu559Val (formerly E540V), p.Arg573Leu (formerly R554L), p.Pro571His (formerly P552H) and p.Gly574Phe (formerly G555F) [1, 3]. The correspondences confirmed in a normalisation tool are NM_000508.5:c.1634A>T for p.(Glu545Val) and NM_000508.5:c.1718G>T for p.(Arg573Leu) in FGA; NM_000039.3:c.605T>C for p.(Leu202Pro) in APOA1; NM_004048.4:c.286G>A for p.(Asp96Asn) in B2M; and NM_000177.5:c.640G>A for p.(Asp214Asn) in GSN, corresponding to historical D187N of the mature protein and also referred to as D214N [20, 50, 54]. In LYZ, the two nucleotide changes that produce p.(Trp82Arg), corresponding to historical W64R, are NM_000239.3:c.244T>C and c.244T>A, both classified as pathogenic in ClinVar [49]. The existence of two distinct nucleotide changes producing the same substitution, documented in LYZ by two independent families, shows that the variant must be identified at the nucleotide level and not only at the amino acid level [34].

In AApoAII, interpretation requires particular attention, because the pathogenic variants lie in the stop codon and may be classified by automated systems as variants of uncertain consequence: the effect is extension of translation by 60 nucleotides up to the next in-frame stop codon, with a 21-residue carboxy-terminal extension that is precisely the sequence found in the amyloid [11, 37]. The possible functional confirmation is plasma western blot under reducing conditions with an anti-apoA-II antibody, which reveals an extra band at about 10 kDa in addition to the normal band at 8 kDa, and western blot of the solubilised fibrils, which contain only the variant apoA-II [10, 11].

A variant of uncertain significance in one of these genes only acquires clinical significance when accompanied by demonstration that the corresponding protein makes up the deposits. The case of the GSN p.Asn160Lys variant, classified as likely pathogenic, was corroborated by mass spectrometric detection of peptide fragments specific to the mutant allele in tissue, without corresponding wild-type peptides, and by structural analysis of the predicted disturbance in the calcium-binding pocket [48]. The absence of the variant in the only offspring studied and the impossibility of assessing the parents, both deceased, left open the hypothesis of a de novo event or incomplete penetrance [48]. For the APOA1 p.L202P variant, the classification relied on two pathogenic reports in ClinVar, on a change at the same locus recorded in a human genome variation database and on location in an amino-terminal region described as a hotspot for amyloidogenic variants, with confirmation by immunoreactivity of the gastric deposits with anti-apoA-I antibodies [8].

5.5 Limits of the tests

A negative result in any of the three steps does not exclude the diagnosis. A biopsy negative for amyloid does not exclude amyloidosis, because the distribution of deposits is uneven, and the sampling site may determine the result. In apolipoprotein A-I amyloidosis, the deposit may be limited to the renal medulla and non-glomerular regions, which makes the biopsy falsely negative if it does not include medulla [58, 59]. In other entities, the problem is the reverse: in AFib, the pattern is one of complete histological involvement of the glomeruli, often obliterating the architecture, with sparing of the vessels, which makes renal biopsy very high-yield in this entity [39], but in the North American series the abdominal fat biopsies performed in four patients and the bone marrow biopsies in three were all negative, although the renal diagnosis was established [1]. In ALys, gastrointestinal endoscopy is particularly informative, because macroscopically visible amyloid lesions were found in nine of ten patients who underwent endoscopy [16]. In AGel, the diagnosis may require repetition: in the case described in 2026, the first renal biopsy had equivocal Congo red and did not yield a diagnosis, and only the second biopsy, with weakly positive Congo red and equivocal birefringence, combined with mass spectrometry, established the typing [48].

Proteomics left 5.0% of the validation set of the French cohort without a protein identified, and its availability is limited to a few expert centres [66]. Mass spectrometry does not identify the variant in every case. In the review of 115 renal AFib cases typed in a reference laboratory, liquid chromatography with tandem mass spectrometry detected a pathogenic variant in 81 (70%), not detecting it in the remaining 34 [1]. Scintigraphy with radiopharmaceuticals can demonstrate amyloid load but is not specific for the protein type [39], and bone scintigraphy with labelled diphosphonates, which has high sensitivity and specificity for ATTR in the absence of monoclonal gammopathy or an abnormal free light-chain ratio, has documented false positives in the rare forms: in AApoAI and AApoAII with concomitant kidney disease, in which case genetic testing is suggested; in AApoAIV with kidney disease; and in Aβ2M after more than nine years of dialysis, histology being required in these last two cases [28, 32]. Grade 2 or 3 uptake on scintigraphy, or a heart-to-contralateral-rib uptake ratio above 2.0, is considered diagnostic of transthyretin cardiac amyloidosis in the absence of monoclonal gammopathy, but this inference does not transfer to the rare forms [32].

As for genetic testing, the panel used determines the result: a panel restricted to transthyretin, which is the widely available test, necessarily gives a negative result in the six entities covered here, and the absence of a variant on such a panel does not allow the conclusion that the amyloidosis is not hereditary [1]. In the case described in 2026, sequencing of TTR exons 1 to 4 was normal and only subsequent APOA1 sequencing, requested after the patient's brother had been diagnosed, revealed the variant [8].

5.6 Biomarkers and assessment of disease extent

Initial assessment and follow-up use biomarkers that are not specific to the amyloid type but quantify organ involvement. In the kidney, estimated glomerular filtration rate in millilitres per minute per 1.73 m², 24-hour proteinuria quantification in grams per day and urinary albumin/creatinine ratio in milligrams per millimole or per gram define the chronic kidney disease stage and albuminuria category, and are the independent predictors of progression to end-stage kidney disease in AFib, with hazard ratios of 0.97 per unit of glomerular filtration and 1.08 per gram of proteinuria [3]. In the heart, N-terminal pro-B-type natriuretic peptide and troponin are the markers used in follow-up, with values of 1210 and 1580 pg/ml in the AApoAI case before treatment and 804 ng/L after the clinical response [8, 28]. Visceral amyloid load can be quantified by iodine-123-labelled serum amyloid P component scintigraphy, with anterior and posterior whole-body imaging obtained 6 or 24 hours after administration, and allows serial assessment of the evolution of the deposits over time [3].

Cardiac assessment follows the same instruments as for all amyloidoses. On the electrocardiogram, the most common finding is low voltage in the limb leads, present in about 50 to 60% of patients with AL, less frequent in the other forms, with about 30% in familial amyloidoses and 40% in senile ones; even when formal low-voltage criteria are not met, there is usually a discordance between the presence of hypertrophy on imaging and its absence on the electrocardiogram [32]. On echocardiography, the diffuse and generally symmetrical increase in wall thickness results from accumulation in the extracellular matrix and not from myocyte hypertrophy, and may also thicken the right ventricle, the atrial walls and the valves; the bright appearance of the myocardium is described as typical but has contested diagnostic value, with low sensitivity and specificity in several studies, serving mainly to raise the hypothesis [32]. Diastolic dysfunction is frequent and may be the first and only sign in the early stages, progressing in advanced stages to restrictive cardiomyopathy with biatrial dilatation; tissue Doppler imaging and speckle tracking show early impairment of longitudinal strain, with impairment of the basal and mid-ventricular segments more pronounced than at the apex, a pattern associated with high sensitivity and specificity [32]. On cardiac magnetic resonance imaging, late gadolinium enhancement is diffuse, myocardial nulling on late enhancement sequences is difficult, native T1 relaxation time is greatly increased by expansion of the extracellular matrix, and an estimated extracellular volume above 40% is highly predictive of cardiac amyloidosis [32]. It is worth noting that none of these findings is type-specific: the apical sparing pattern of longitudinal strain, for example, was present in the case of apolipoprotein A-IV amyloidosis and absent in the case of apolipoprotein A-I amyloidosis, in both unrelated to the protein type [8, 46].

5.7 Prenatal and preimplantation diagnosis

As in any genetic disease, preimplantation genetic diagnosis is an option to avoid transmission to subsequent generations, and with variants of relatively high penetrance the challenges of this technique are more relevant than with those of late onset and probably low penetrance [32]. Its application to these entities is documented, albeit sporadically: in the British cohort of 35 patients with lysozyme amyloidosis, one couple used preimplantation genetic diagnosis [15]. The technical possibility is general, because all variants are known at the nucleotide level and can therefore be analysed in a single cell or a trophectoderm, but there are no series describing its use in the other entities. The decision in these entities is conditioned by two factors that counselling should set out: the age at onset is adult, generally from the fifth or sixth decade, except with FGA frameshift variants, in which it may be juvenile; and there is effective treatment of organ failure, with long survival, notably in AFib and AApoAI, although there is no treatment that prevents deposition [3–5].

6. Differential Diagnosis

6.1 Comparative table

6.2 Separating hereditary from acquired forms

The most consequential differential diagnosis is not between the six hereditary entities, but between them and the acquired forms, because treatment depends on this distinction. AL amyloidosis is treated with chemotherapy directed at the plasma cell clone, and chemotherapy has absolutely no role in hereditary apolipoprotein A-I amyloidosis or in the other hereditary forms [7, 52]. Conversely, treatment with transthyretin stabilisers or silencers has no effect on the six entities covered in this article, because none of them is mediated by that protein.

Prior probability always favours the acquired forms: AL and ATTR together account for more than 98% of diagnosed cardiac amyloidoses, and hereditary amyloidoses account for less than 5% of renal amyloidoses [1, 28]. But the frequency of error in the opposite direction is high enough to justify systematic testing: 9.7% of patients with a clinical diagnosis of AL based on clinical and laboratory findings and the absence of a family history in fact had amyloidogenic mutations, 18 of them with the FGA E526V variant and 13 with transthyretin variants [40]. The presence of a monoclonal gammopathy does not resolve the doubt, in either direction: a low-grade monoclonal gammopathy was present in 24% of patients with hereditary amyloidosis wrongly classified as AL [40], and in the Portuguese proteomic typing series the presence of a monoclonal gammopathy did not reliably predict the subtype in three of five patients [63]. The coexistence of the two situations is possible and has been documented: a patient with genetically confirmed transthyretin amyloidosis and a compatible clinical picture had AL lambda amyloid on mass spectrometry [63]. In the historical AFib cohort, seven patients had incidental plasma cell clones [39], and in the North American series two of the 13 patients (15%) had a history of monoclonal gammopathy [1].

The error also has a purely technical component, which molecular typing corrects. In the French national typing cohort, in the subgroup of specimens with equivocal or negative immunostaining, anti-light-chain antibodies gave false positives in 41.5% of the 101 cases of light-chain amyloidosis and the anti-transthyretin antibody gave false positives in 56.8% of the 58 cases of transthyretin amyloidosis, which quantifies the risk of assigning either of these two types by immunohistochemistry alone in a patient whose deposit is of another protein [66]. In a blinded comparison of 142 consecutive biopsies, immunohistochemistry was inconclusive in 20 to 25% of cases, and the addition of mass spectrometry raised diagnostic accuracy from 76% to 94% [65].

The distinction between hereditary and dialysis-related β2-microglobulin amyloidosis rests on a simple and reliable criterion. The dialysis-related form is caused by high and sustained plasma concentrations of the normal protein in renal failure treated by long-term haemodialysis, deposits preferentially around bones and joints and has an estimated point prevalence of 4.5 per 100,000 in Europe [24, 38, 67]; the hereditary form occurs with normal renal function and normal or low serum β2-microglobulin, and deposits in the gastrointestinal tract, heart, soft tissues and nerves [23–25]. Bone scintigraphy may give a false positive in the dialysis form after more than nine years of treatment, requiring histology [28].

6.3 The rare forms that escape the non-invasive algorithm

Two entities deserve emphasis because they illustrate the limit of the diagnostic algorithm based on scintigraphy and monoclonal protein testing. The 2023 North American consensus document is explicit about this limit in both directions: there was positive uptake on technetium-99m-labelled 3,3-diphosphono-1,2-propanodicarboxylic acid scintigraphy in apolipoprotein A-I, apolipoprotein A-IV and β2-microglobulin amyloidoses, and there are false negatives with the Phe64Leu transthyretin variant, so bone uptake with negative monoclonal protein screening is not a sufficient criterion for transthyretin amyloidosis when these entities are in question, and histological assessment by biopsy is required whenever clinical suspicion of a rare form arises [62]. The first is apolipoprotein A-IV amyloidosis, which the current classification considers non-hereditary and of unknown frequency, with mainly renal involvement and a median survival of 79 months [28, 68]. The case published in 2026 of a 71-year-old man with an infiltrative cardiomyopathy phenotype — maximum wall thickness of 24 mm, interatrial septal hypertrophy, mild pericardial effusion, diffuse and partly transmural late enhancement, increased native T1 and extracellular volume of 40% in the mid-ventricular septum on magnetic resonance imaging — shows the problem clearly: bone scintigraphy and monoclonal protein screening were negative, immunohistochemistry of the endomyocardial biopsy was likewise negative, and only mass spectrometry-based proteomics identified apolipoprotein A-IV [46]. The authors conclude that a pathognomonic phenotype on magnetic resonance, despite negative scintigraphy and monoclonal protein screening, should lead to invasive investigation, and that endomyocardial biopsy with mass spectrometry remains the definitive typing technique in these cases [46]. The second is ALECT2, which typically involves the kidney and liver with clear sparing of the heart, disproportionately affects certain populations — 88 to 92% of patients are Hispanic —, and whose global prevalence has not yet been accurately reported [39]. Despite the evident ethnic variation, no hereditary cause has been identified and a sporadic disorder is presumed: no mutations in the LECT2 gene were detected in patients, although a common polymorphism was detected, homozygosity for the G allele at the single nucleotide polymorphism rs31517, of unknown mechanistic significance, and LECT2 concentrations were not increased in patients, which excludes sustained overexpression as the mechanism [39]. Histologically, deposition occurs in the glomerular interstitium and mesangium with relative sparing of the medullary interstitium, clinical findings range from chronic kidney disease with minimal proteinuria to nephrotic syndrome, median renal survival was estimated at 62 to 98.3 months, and kidney transplantation has been performed successfully with no reports of recurrence [39].

6.4 Renal differential: what mimics amyloidosis on biopsy

On renal biopsy, the organised deposition diseases enter the same differential. Fibrillary glomerulonephritis is characterised by DNAJB9 deposits, with thicker fibrils and IgG-predominant immunofluorescence, and fibronectin glomerulopathy by mesangial and subendothelial fibronectin deposits; both are Congo red negative and are distinguished by targeted immunohistochemistry [48]. The case of renal AGel described in 2026 illustrates the difficulty: immunofluorescence of the first biopsy showed segmental glomerular deposits of all immunoglobulins, complement and light chains, with IgA predominance; Congo red was equivocal; immunohistochemistry for DNAJB9 and fibronectin was negative, excluding fibrillary glomerulonephritis and fibronectin glomerulopathy; and electron microscopy showed mesangial expansion with abundant short, rigid fibrils and parallel 10 nm filaments, also present in the subendothelial and subepithelial spaces, with amorphous electron-dense deposits in the mesangium and diffuse foot process effacement, findings that, combined, did not yield a definitive pathological diagnosis [48]. Only laser microdissection with mass spectrometry, performed after the second biopsy, identified high spectral counts of amyloid-associated proteins — apolipoprotein E and serum amyloid P component — and of gelsolin, establishing the diagnosis of hereditary gelsolin amyloidosis with suspected concomitant IgA nephropathy [48]. The authors note that in a subset of cases of renal AGel, three of twelve, there are concomitant immunoglobulin deposits, which may lead to a misdiagnosis of fibrillary glomerulonephritis, and that the distinction is made by the predominance of IgA and C3 instead of IgG, by fibril diameter, 10 nm and therefore thinner than in fibrillary glomerulonephritis, and by mandatory confirmation by mass spectrometry [48].

6.5 Cardiac differential

In a restrictive cardiomyopathy phenotype with preserved ejection fraction, severe concentric thickening and an apical sparing pattern of longitudinal strain, the differential considered includes transthyretin amyloidosis, hereditary or wild-type, as the primary suspicion when there are suggestive extracardiac signs; AL amyloidosis, to be rigorously excluded, especially when there is unexplained proteinuria and renal dysfunction; the rare forms of cardiac amyloidosis, including the hereditary ones due to APOA1 variants; hypertrophic cardiomyopathy of concentric phenotype, which may show similar degrees of wall thickening and severe diastolic impairment; hypertensive heart disease, unlikely when the severity of thickening is disproportionate to the blood pressure history; and Fabry disease, unlikely with late onset and in the absence of typical extracardiac features [8, 46]. Fabry disease was indeed excluded by enzyme testing in the patient with cardiac AApoAI before the diagnosis was reached [8]. In histological diagnosis, the presence of atrial amyloid derived from atrial natriuretic peptide is frequent and requires conclusive evidence of ventricular amyloid, by imaging or ventricular biopsy, before cardiac systemic amyloidosis is diagnosed [32].

7. Treatments and Risk Management Strategies

7.1 Principles

None of the six entities covered in this article has drug therapy that acts on the amyloidogenic process. Treatment is organised on three levels: support of the affected organ, with the usual nephroprotective, cardiological and surgical tools; removal of the hepatic source of the variant protein, when it is hepatic and transplantation is feasible; and replacement of the failing organ. The therapeutic decision depends entirely on correct typing of the deposit, because chemotherapy, the treatment of AL amyloidosis, has absolutely no role in hereditary apolipoprotein A-I amyloidosis or in the other hereditary forms, and its inappropriate administration exposes the patient to toxicity without benefit [7, 52].

It is equally necessary to state what does not apply. The drugs approved for amyloidosis do not treat these diseases and should not be offered for them. Tafamidis (Vyndaqel®, Vyndamax®), patisiran (Onpattro®), vutrisiran (Amvuttra®), eplontersen (Wainua®) and acoramidis (Attruby®) are indicated, in their respective US agency labels, exclusively for transthyretin-mediated amyloidosis, with label effective dates of 12 May 2026, 8 September 2025, 6 November 2025, 15 April 2026 and 29 June 2026; inotersen (Tegsedi®), likewise restricted to transthyretin amyloidosis, with a label in effect since 25 January 2024, is recorded as discontinued in the US agency's drug registry, under application number NDA 211172; and daratumumab with hyaluronidase (Darzalex Faspro®) is indicated for light-chain amyloidosis [44]. None of these drugs targets fibrinogen, apolipoproteins A-I and A-II, lysozyme, gelsolin or β2-microglobulin.

7.2 Nephroprotection: the baseline therapy of amyloid nephropathy

In the four predominantly renal entities — AFib, AApoAI, AApoAII and ALys —, and in AGel when there is nephropathy, the effective treatment available is the treatment of proteinuric chronic kidney disease. There are no trials or guidelines specific to these diseases; the recommendations applied are those for chronic kidney disease of any cause, and their application to these entities constitutes an extrapolation, which is declared here.

The 2024 Kidney Disease: Improving Global Outcomes guideline recommends starting renin–angiotensin system inhibitors — angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker — in people with chronic kidney disease and severely increased albuminuria, categories G1 to G4 with A3, without diabetes, with recommendation grade 1B, and suggests starting them in moderately increased albuminuria, A2, without diabetes, grade 2C; it recommends avoiding any combination of converting enzyme inhibitor, receptor blocker and direct renin inhibitor, grade 1B [69]. The practice points accompanying these recommendations state that the drug should be given at the highest approved dose that is tolerated, because the proven benefits were obtained in trials with those doses; that blood pressure, serum creatinine and serum potassium should be checked 2 to 4 weeks after starting or increasing the dose, depending on glomerular filtration and potassium; that associated hyperkalaemia can often be managed with measures to lower potassium rather than by reducing or stopping the drug; that treatment should be continued unless serum creatinine rises by more than 30% within 4 weeks of starting or increasing the dose; that dose reduction or discontinuation should be considered in the event of symptomatic hypotension or uncontrolled hyperkalaemia despite treatment, or to reduce uraemic symptoms while treating kidney failure with estimated glomerular filtration rate below 15 ml/min per 1.73 m²; and that the drug should be maintained even when estimated glomerular filtration rate falls below 30 ml/min per 1.73 m² [69].

The same guideline recommends treating adults with chronic kidney disease with a sodium–glucose cotransporter 2 inhibitor when estimated glomerular filtration rate is 20 ml/min per 1.73 m² or above and urinary albumin/creatinine ratio is 200 mg/g (20 mg/mmol) or above, or when there is heart failure, irrespective of albuminuria level, with recommendation grade 1A, and suggests treating adults with estimated glomerular filtration rate between 20 and 45 ml/min per 1.73 m² and albumin/creatinine ratio below 200 mg/g, grade 2B [69]. Once the drug has been started, it is reasonable to continue it even if estimated glomerular filtration rate falls below 20 ml/min per 1.73 m², unless it is not tolerated or kidney replacement therapy is started; it is reasonable to withhold it during periods of prolonged fasting, surgery or critical illness, because of the greater risk of ketosis; and starting it does not require a change in monitoring frequency, the reversible reduction in glomerular filtration at initiation not being an indication to stop [69]. The guideline further suggests a systolic blood pressure target below 120 mmHg, when tolerated, using standardised office measurement, grade 2B, considering less intensive therapy in people with frailty, high risk of falls and fractures, very limited life expectancy or symptomatic postural hypotension; a sodium intake below 2 g per day, equivalent to less than 90 mmol of sodium or less than 5 g of sodium chloride per day, grade 2C; and, in adults aged 50 years or over with estimated glomerular filtration rate below 60 ml/min per 1.73 m² not treated with chronic dialysis or transplantation, a statin or statin–ezetimibe combination, grade 1A, extending the statin recommendation, grade 1B, to patients with estimated glomerular filtration rate of 60 ml/min per 1.73 m² or above [69].

The dose of the sodium–glucose cotransporter 2 inhibitor most used in this indication is given by the label, because there is no dose specific to the hereditary amyloidoses. The European summary of product characteristics of dapagliflozin (Forxiga®) indicates it in adults for the treatment of chronic kidney disease, with a recommended dose of 10 mg once daily; no dose adjustment is needed based on renal function; because of limited experience, starting treatment is not recommended in patients with glomerular filtration below 25 ml/min; in severe hepatic impairment a starting dose of 5 mg is recommended, which may be increased to 10 mg if well tolerated; there is no adjustment for age; and safety and efficacy in chronic kidney disease in those under 18 years have not been established [70]. The US label for the same active substance (Farxiga®), effective from 10 July 2026, frames the indication as reducing the risk of sustained decline in estimated glomerular filtration rate, end-stage kidney disease, cardiovascular death and hospitalisation for heart failure in adults with chronic kidney disease at risk of progression, with a recommended dose of 10 mg orally once daily for all non-glycaemic indications, and states that renal function should be assessed before initiation and then as clinically indicated, that volume depletion should be assessed and corrected before starting, and that the drug should be withheld at least 3 days before surgery or procedures associated with prolonged fasting, where possible [71]. The two jurisdictions differ in the wording of the indication — treatment of chronic kidney disease in the European Union and risk reduction in the United States — and agree on the dose.

The evidence that these measures alter the course of hereditary amyloid nephropathy is indirect and comes from comparison of periods. In the British AFib cohort, the proportion of patients receiving a converting enzyme inhibitor or angiotensin receptor blocker rose from 25 of 75 (33.3%) before 2009 to 66 of 119 (55.5%) afterwards, although with more missing data in the earlier cohort; and renal survival improved over the same period, especially in patients diagnosed in chronic kidney disease stage IIIA or IIIB, in whom the median time to end-stage kidney disease rose from 2.1 years to 6.8 years, with a hazard ratio of 3.4 (95% confidence interval 1.5 to 7.5; p=0.0007) [3]. The period of diagnosis was the strongest predictor of renal survival in the multivariable analysis, with a hazard ratio of 0.44 (0.28 to 0.70) for the period after 2009 [3]. The authors acknowledge not knowing what explains this effect: in the absence of new anti-amyloid therapy for AFib, they consider it possible that advances in general nephroprotective care, including optimisation of renin–angiotensin–aldosterone system blockade, better access to nephrology and better blood pressure control, explain it, but Cox regression revealed no significant effect of converting enzyme inhibitor or receptor blocker prescription on the hazard ratio for end-stage kidney disease when adjusted for other potential confounders, possibly because of the high proportion of missing prescription data in the earlier cohort [3]. Uptake of sodium–glucose cotransporter inhibitors in this population is still low: only 1 patient was considered eligible in the pre-2009 cohort, and was not prescribed one, and in the later cohort 9 of 38 eligible patients (23.7%) received them, patients being considered eligible with estimated glomerular filtration rate above 25 ml/min per 1.73 m², no obvious contraindication and no kidney transplant at any time after 2022 [3]. It remains to be seen whether widespread adoption of this class will bring benefit in patients with AFib [3].

7.3 Organ transplantation

Transplantation is the only targeted therapy available in these diseases, and its rationale differs between entities according to the origin of the variant protein.

In AFib, the variant protein is entirely of hepatic origin and the wild-type protein does not form fibrils in vivo, which underpins the possibility of halting production by liver transplantation [3]. The King's College series, published in 2010, assessed 22 patients for combined liver–kidney transplantation: 21 had the E526V variant and one R554L, the median age at presentation was 55 years (range 33 to 63), proteinuria was the most common presenting manifestation, in 55%, and renal biopsy showed amyloidosis in all cases; 14 were accepted for listing and 9 were transplanted, transplantation having been declined for 8 patients because of poor ventricular function or uncorrectable cardiovascular or systemic atheromatous disease; the procedure was complicated by perioperative spontaneous splenic rupture in 3 patients and postoperative deep vein thrombosis in 2; six of the nine transplanted were alive (67%), with good allograft function and without amyloidosis, at a median follow-up of 67 months (33 to 155); four explanted livers were used successfully for domino transplantation; and serial renal scintigraphy with technetium-99m-labelled dimercaptosuccinic acid in two cases of pre-emptive combined transplantation demonstrated preserved residual function of the native kidneys at 5 years [29, 57].

The contemporary cohort from the UK national centre, with 194 patients, reversed this recommendation. Between 1996 and 2023, 13 patients underwent combined liver–kidney transplantation, 9 in the cohort diagnosed before 2009 and 4 in the later one, indicating that the procedure became less favoured over time; three of the nine in the earlier cohort (33.3%) died of complications of the procedure in the first year, one of necrotising pancreatitis and the others of perioperative complications, and one died with deteriorating liver function in the context of biliary strictures of the liver allograft 16 years later; all four transplanted after 2009 were alive at census; and at census, at a median of 13.6 years after the procedure, 6 patients had died, including 2 with both allografts functioning at the date of death, and 7 were alive, 5 with both original allografts functioning, 2 with a second kidney transplant for transplant glomerulopathy and for calcineurin inhibitor toxicity, and 1 back on dialysis [3]. Over the same period, 48 patients underwent isolated kidney transplantation, 11 before 2009 and 37 after; amyloid recurrence with graft failure occurred in 8 patients, 4 of 11 in the earlier period and 4 of 37 in the recent one, one patient in the earlier cohort having lost two allografts at 5.8 and 7.0 years from recurrence in each; there were 5 of 37 patients in the recent cohort considered to have recurrence but with a functioning graft at census, so the proportion of transplants with recurrence was 5 of 12 (41.6%) in the earlier period and 9 of 37 (24.3%) in the recent one; and median renal allograft survival, censored for death, rose from 7.7 years (5.8 to undefined) to 15.5 years (10.0 to undefined), with p=0.006 and a hazard ratio for failure of 4.5 (95% confidence interval 1.2 to 22.6) in the earlier period [3]. Overall, 48 of the 194 patients (24.7%) received an isolated kidney transplant and 13 (6.7%) a combined transplant, corresponding to exactly half of the 122 patients who reached end-stage kidney disease receiving a renal allograft [3]. The authors' reading is that combined transplantation effectively eliminates the risk of recurrent amyloid in the renal allograft, but remains associated with significant procedural risk, which, in the face of the increasingly favourable results of isolated kidney transplantation, justifies the declining trend in its use; the data suggest that isolated kidney transplantation is probably appropriate for most patients with end-stage kidney disease due to AFib, combined transplantation remaining a valid option for younger patients in better condition [3]. The presence of a kidney transplant was, in the multivariable analysis, the strongest modifiable protective factor against death, with a hazard ratio of 0.54 (0.29 to 0.99; p=0.05) [3].

The choice between isolated and combined transplantation in AFib may be informed by the variant. In the French series of 32 patients, 18 were transplanted, 15 with isolated kidney transplantation and 4 with combined liver–kidney transplantation, one of them after failure of the first kidney transplant; in the 15 kidney transplant recipients, amyloidosis recurrence in the graft occurred in 7 (47%), at a median of 96 months (12 to 120), and graft loss in 8 of the 18 patients (44%), attributable to recurrence in 6 of these 8 losses (75%) [4]. Recurrence was very unevenly distributed across variants: 2 of 9 E526V carriers (22%), at a median of 119.5 months, 2 of 3 R554L carriers (67%), at a median of 105 months, and 3 of 3 carriers of frameshift variants (100%), at a median of only 24 months (12 to 96); the statistically significant comparison is that of the E526V group with the frameshift group (22% versus 100%; p=0.045), while graft loss due to recurrence was 11% in the E526V group versus 100% in the non-E526V group as a whole, without statistical significance (p=0.1) because of the size of the subgroups [4]. The article's abstract groups them differently — E526V and R554L versus frameshift, with 22% versus 83% and p=0.03 — and the figures in the body of the article, followed here, are those that allow the denominators to be reconstructed [4]. Renal allograft survival in the series was 88%, 80% and 72% at 5, 7 and 9 years, without significant difference between variant groups, and patient survival from diagnosis 84% at 10 years and 70% at 30 years [4]. No amyloid recurrence in the renal graft was observed after combined transplantation in any of the 4 patients, and the authors concluded that isolated kidney transplantation is a valid option in carriers of the E526V variant without systemic manifestations, whereas combined transplantation may be preferable in carriers of frameshift or R554L variants, because they are younger and at greater risk of early allograft loss from recurrence, and should be considered regardless of variant when there are severe biopsy-proven systemic manifestations [4]. This correlation was not reproduced in the British cohort, where only one carrier of a non-E526V variant was transplanted, with robust graft survival, so the difference in recurrence between periods in that cohort is not explained by the distribution of variants [3]. The historical series from the same centre, with 71 patients, described kidney transplantation in 10 patients, two of whom received two grafts, a total of 12 allografts: median overall graft survival was 5.9 years (range 0 to 12.2) by Kaplan-Meier analysis and 6.7 years (0.9 to 12.2) after excluding three grafts lost immediately for technical reasons; of the 12 grafts, 5 were functioning at census and 7 had failed, three from recurrent amyloidosis at 5.8, 6.0 and 7.4 years — two of them sequentially in the same patient — and one from transplant glomerulopathy at 5.8 years, without amyloid on histology [56]. The longest-surviving graft was functioning at census 12.2 years after transplantation, despite recurrence demonstrated by serum amyloid P component scintigraphy associated with increased proteinuria in the first 7.6 years; this patient is both an individual case described and the upper limit of the graft survival range in that series [56]. The series also included the first patient to receive a combined liver–kidney transplant for AFib, well and completely free of amyloid 11.5 years later, and six more patients transplanted in combined fashion at King's College Hospital, three of them pre-emptively, with one perioperative death from acute necrotising pancreatitis and the rest well, without recurrence, at a median follow-up of 24 months [56]. Ten-year allograft survival in kidney-only transplants is 5.5%, and the nature of the figure conditions its reading: the King's College series presents it as an aggregate of the 18 isolated kidney transplants published up to 2010, and not as a result of its own cohort, accompanied by the observation that amyloid recurrence in the graft and subsequent loss are almost universal and occur between one and seven years [57]. The comparator given in the same text is 10-year survival of 64% for the graft and 68% for the patient in deceased-donor kidney transplantation for all causes of chronic kidney disease, with a graft half-life of 10 to 12 years [57]. Understood in this way, the percentage is compatible with the median graft survivals observed — 5.9 and 6.7 years in the historical series and 7.7 years before 2009 versus 15.5 years after in the contemporary cohort —, because a median below seven years implies that most grafts are lost before ten years [3, 56]. In the same series, of the nine patients who underwent combined liver–kidney transplantation between January 1996 and September 2009, six were alive and well at a median follow-up of 67 months, with a range of 33 to 155 months, five with good renal graft function, median creatinine of 120 µmol/L and median glomerular filtration rate of 54 mL/min, and the sixth with chronic allograft nephropathy at 68 months [57]. The timing of the intervention separated the results markedly: all patients transplanted before starting renal replacement therapy or within six months of starting it survived, compared with half of those on prolonged haemodialysis at the time of transplantation, the group in which the two procedure-related deaths occurred [57]. The practical implication is that assessment for combined transplantation is done on recognising amyloid nephropathy, and not when dialysis is already established. Pre-emptive isolated liver transplantation, performed at the onset of amyloid nephropathy to avoid progression to haemodialysis and kidney transplantation, has been proposed on the basis of serial renal scintigraphy in two cases, but is not established and remains an investigational strategy [45].

In AApoAI, management is organ support and transplantation, and the results are favourable. In the cohort of 57 patients from the UK national centre, 18 underwent kidney transplantation, including 11 isolated kidney transplants (3 living donor and 8 deceased donor), 5 combined liver–kidney and 2 combined heart–kidney transplants; median renal allograft survival overall was 22.0 years (95% confidence interval 13.0 to 31.0), higher than the 13.1 years the same centre had previously published, there was one early death from cytomegalovirus infection within 2 months of a kidney transplant and no episode of early rejection with graft failure [5]. In the subgroup of isolated kidney transplant recipients, including the two patients who received a second allograft, median graft survival censored for deaths with a functioning graft was 16.0 years (3.1 to 28.9), median survival from diagnosis 32 years (20.9 to 43.1) and from transplantation 19.7 years (15.2 to 24.1); two patients died with documented amyloid recurrence in the allograft 13 years after transplantation, but with functioning grafts at the date of death, and two lost the graft to recurrence and received a second allograft 6 and 16 years after the first [5]. None of the patients who underwent isolated kidney transplantation had clinically significant cardiac amyloidosis, which, in this entity, is not an absolute contraindication, but requires careful assessment of cardiac function and reserve before surgery [5]. Liver transplantation was followed by amyloid regression on serial iodine-123-labelled serum amyloid P component scintigraphy in the four patients so assessed, with marked splenic regression in all and evident at 6 months in one of them, which is explained by about 50% of circulating apolipoprotein A-I being of hepatic synthesis [5, 52]. The authors' operational conclusion is restrictive and worth retaining: given the excellent outcome of isolated kidney and heart transplantation, liver transplantation is not justified routinely in AApoAI, being reserved for the failing liver, possibly due to extensive hepatic amyloidosis, or for progressive amyloidotic dysfunction of another organ that threatens life or causes significant functional impairment [5]. Recurrence in the renal allograft is rare in this entity, with a single published report, and only 13% of variant carriers progressed to renal failure requiring haemodialysis or transplantation [60]. In the Irish experience, of the five patients transplanted one lost the graft at 5 years from recurrent disease and one died of sepsis shortly after transplantation, but, overall, kidney transplantation conferred a survival advantage [59]. In the pure cardiac phenotype, heart transplantation has been performed successfully: in the Italian patient with the L174S variant, transplanted at 56 years after progression to refractory heart failure, there were no myocardial deposits at nine months [6]. A relevant operative difference separates this entity from AFib: histological assessment of two livers explanted from patients with the Arg26 variant revealed extensive parenchymal involvement, with large areas of amyloid separating the hepatic cords, central perivenular accentuation, all portal tracts markedly involved with compression of bile ducts and vascular structures and irregular deposits in the hilar vessels, leading the authors to conclude that the livers of these patients are not suitable for domino donation, unlike those from familial amyloidotic polyneuropathy [72] — a conclusion that contrasts with the four livers from patients with AFib used successfully in domino transplantation [57].

In ALys, the 2026 cohort describes, in the body of the article, two patients with spontaneous hepatic haemorrhage — one contained, subcapsular, and the other catastrophic, the latter requiring emergency liver transplantation — and two patients with hepatic rupture, both also with emergency transplantation, in addition to a pre-emptive liver transplant in a patient with progressive hepatomegaly and a family history of four deaths from hepatic rupture or haemorrhage; the article's abstract totals 5 patients and 4 emergency transplants [15]. There were also two spontaneous splenic ruptures, one after kidney transplantation and the other after streptococcal infection, which indicates that intercurrent illness or major surgery can precipitate rupture of amyloid-laden organs, and that the mechanical risk is independent of conventional markers of hepatic decompensation: none of these patients had hepatic synthetic failure or portal hypertension [15]. Of the 8 patients who developed renal failure, 6 received a kidney transplant, with a median graft survival of 6.6 years (0.8 to 15.0) and a single failure, from chronic allograft rejection, without any evidence of amyloid recurrence on serum amyloid P component scintigraphy over a median follow-up of 10.4 years after transplantation [15]. In the earlier series from the same centre, four patients received orthotopic liver transplantation, three for spontaneous hepatic rupture and one, with extensive hepatic amyloid and a strong family history of rupture, pre-emptively, and all four liver grafts were functioning at 1.7, 5.8, 9.0 and 11.0 years; three of the five patients with end-stage renal failure had a kidney transplant, without allograft dysfunction at 6.6, 1.8 and 0.8 years [16].

In AGel, there is no established indication for organ transplantation, because severe renal and cardiac manifestations are rare and life expectancy is near normal; the disease burden is ophthalmological, neurological, cutaneous and surgical [18, 21]. In AApoAII, dialysis and kidney transplantation are the two therapeutic options, and a recipient heterozygous for the Stop78Gly variant maintained stable graft function for nine years after transplantation, without involvement of other organs, which led the authors to consider transplantation effective because recurrence in the graft and extrarenal progression appear slow; the evidence is, however, from a single patient [10, 12]. In Aβ2M, organ transplantation has been performed in organ failure: a member of the Portuguese family was transplanted for cardiomyopathy, and the patient described in 2023 underwent heart transplantation, with the explanted heart confirming β2-microglobulin deposition [23].

In the amyloidoses in general, transplantation is accepted as an established therapy with reasonable long-term results, but selection of eligible patients is strict and requires multidisciplinary assessment of extracardiac and extrarenal involvement; because of organ scarcity, it should not be considered routine therapy, and even patients placed on the waiting list may deteriorate beyond eligibility or die before receiving the organ [32]. Involvement of other systems, such as gastrointestinal, hepatic or renal, and severe neuropathy, which limits mobility and postoperative rehabilitation, may constitute a contraindication to heart transplantation [32]. In AFib, combined transplantation was declined in 8 of 22 patients assessed because of poor ventricular function or uncorrectable atheromatous disease [57].

7.4 Treatment of amyloid heart failure

In the entities with cardiac involvement — AApoAI, AApoAII, Aβ2M and, in a minority, AGel —, the treatment of heart failure differs from conventional treatment, and there is little evidence of the safety and efficacy of the usual drugs in these patients. Loop diuretics are the most important therapy, often needed to maintain euvolaemia, because raised intracardiac pressures and diastolic dysfunction produce volume overload; treatment aims to maintain the euvolaemic state, and loop diuretics with sodium and fluid restriction are paramount [32]. Angiotensin-converting enzyme inhibitors and angiotensin receptor blockers have never been proven effective in cardiac amyloidosis and may be harmful, because they worsen hypotension, including orthostatic hypotension, since the amyloid-infiltrated heart cannot increase stroke volume in response to vasodilation; beta-blockers may be harmful because stroke volume is relatively fixed in restrictive heart failure and cardiac output depends largely on heart rate; and digoxin and calcium channel blockers may bind to cardiac amyloid and cause toxicity even at normal circulating concentrations, although one report suggests that some patients tolerate digoxin if renal function and serum concentrations are closely monitored [32]. There is an explicit tension here between the nephrological recommendation, which calls for renin–angiotensin system blockade at the maximum tolerated dose in proteinuric kidney disease [69], and the cardiological recommendation, which considers it potentially harmful in cardiac amyloidosis [32]: in the predominantly renal entities without amyloid cardiomyopathy, such as AFib, the former prevails; in a patient with established infiltrative amyloid cardiomyopathy, haemodynamic tolerance determines the decision.

A case published in 2026 documents the use of a sodium–glucose cotransporter 2 inhibitor in a non-transthyretin hereditary amyloidosis. In a 61-year-old man with AApoAI and infiltrative cardiomyopathy in New York Heart Association class III, dapagliflozin was started, with substantial improvement in symptoms to class I, a fall in N-terminal natriuretic peptide from 1580 to 804 ng/L and clinical stability at 9 months, without heart failure hospitalisations [8]. The authors state explicitly that the use of this class in AApoAI is extrapolated from observational studies in transthyretin cardiomyopathy, possibly improving symptomatic status, and that liver transplantation remains the definitive therapy to reduce amyloid production [8]. This is a single case, and the reasonable conclusion is that the drug is an acceptable option, based on the indication already approved for heart failure and for chronic kidney disease, and not a therapy demonstrated in this disease.

As for devices, in the late stages cardiac amyloid manifests as refractory heart failure; ventricular ectopy is common but sudden death from tachyarrhythmia is uncommon, possibly because of the diffuse conduction delay mediated by the deposits. In a retrospective study of 45 people with cardiac amyloid who received an implantable cardioverter-defibrillator, 38 (84%) had no previous syncope, in a mixed cohort with 27% AL amyloidosis and the remainder transthyretin amyloidosis (60% hereditary and 13% wild type); with a mean follow-up of 17 months, 27% had at least one appropriate device therapy, although transplant-free survival remained low, at 60% over that interval; use for primary prevention is reasonable to consider, but data to support it more broadly are lacking [32]. In AGel, cardiac arrhythmia is reported by 15% of patients in the Finnish national registry and 5% have a pacemaker, and cardiac pacing is part of symptomatic treatment when the arrhythmia requires it [18, 21]. Anticoagulation is another area without an answer: intracardiac thrombus was found in up to 33% of hearts with amyloidosis in an autopsy study, with a prevalence no different between patients with and without documented atrial fibrillation, routine assessments to detect it are not performed in daily practice, and the risks and benefits of prophylactic anticoagulation in sinus rhythm are unknown, a question the European statement lists among the open questions [28, 32].

7.5 Entity-specific symptomatic and surgical therapy

In AGel, there is no specific or disease-modifying therapy, and management is symptomatic and organ-directed: ophthalmological care for corneal lattice dystrophy and dry eye, oculoplastic and skin or soft-tissue surgery for cutis laxa, blepharochalasis and facial palsy, carpal tunnel release, cataract surgery and cardiac and renal surveillance, with cardiac pacing when the arrhythmia requires it; correct diagnosis matters because appropriate symptomatic treatment substantially improves quality of life [21]. The surgical burden is considerable: in the updated Finnish registry, 65% of patients had undergone skin or soft-tissue surgery, 26% carpal tunnel surgery and 24% at least unilateral cataract surgery [18]. Before a corneal transplant, optic neuropathy should be excluded, a finding that has been described and is critical to recognise [20].

In ALys, the disease is treated symptomatically, and surgery has a particular role: spontaneous hepatic rupture is a surgical emergency, liver transplantation may be performed in an emergency setting or pre-emptively in patients with extensive hepatic amyloid and a strong family history of rupture, and the fragility of the liver, with loss of the reticulin framework, conditions the operative technique [16, 34]. In AApoAI, splenectomy was performed for left upper quadrant pain in a patient with massive splenic amyloidosis and thrombocytopenia, with normalisation of platelets, cessation of bruising and better control of hypertension [6]. In AFib, spontaneous splenic rupture occurred in two patients of one cohort and in the perioperative period in three of nine transplant recipients, which requires the spleen to be borne in mind in surgical planning for these patients [29, 57].

7.6 What not to do: immunosuppression in hereditary amyloid nephropathy

Immunosuppression has no role in these diseases and may cause harm. The case of gelsolin amyloidosis with nephrotic syndrome published in 2026 documents this: the patient received prednisolone 20 mg per day, hydroxychloroquine, dapagliflozin and irbesartan, without improvement, with slow progression of renal function; immunosuppression was switched to a Tripterygium wilfordii preparation, the nephrotic syndrome persisting; and, since overlapping IgA nephropathy was considered, prednisolone 40 mg per day with cyclophosphamide was started, with a cumulative dose of 2.8 g over four months, the treatment being stopped because of a lung abscess and, once the infection was controlled, reduced to prednisolone 15 mg per day; there was no renal response, although renal function remained stable [48]. The authors conclude that the lack of response to initial and intensified immunosuppression does not support an overlapping immune-complex glomerulonephritis, and that the strategy should shift to supportive care aimed at delaying progression — strict blood pressure control, avoidance of nephrotoxins, regular monitoring of nephrological and cardiovascular parameters and patient education —, clinical management focusing on control of the complications of amyloidosis and on delaying the progression of renal failure [48].

7.7 Regulatory table

7.8 Trials and development

Reading these data requires caution. None of the programmes listed has any of the six entities covered in this article as its target population: the AT-02 programme is described as an investigational medicinal product in development to treat systemic amyloidosis, but the inclusion criteria of the only study with active status require a confirmed diagnosis of AL amyloidosis, with histological confirmation of congophilic deposits with apple-green birefringence or other amyloid stain and confirmatory immunohistochemistry or mass spectrometry [76]. The anti-amyloid P component strategy, which targets a component common to all subtypes and consists of first removing the serum protein with the bis-D-proline compound known as miridesap, or CPHPC, and then using a humanised IgG1 monoclonal antibody to activate macrophage destruction of the deposits, is theoretically applicable to any amyloidosis and was actually tested in patients with systemic amyloidosis without type restriction, with dose-dependent hepatic amyloid clearance, but the programme was discontinued and none of these drugs is approved for human use [32, 77–79]. The inclusion criteria of two of these studies required a medical diagnosis of systemic amyloidosis with amyloid load in the spleen and, in some patients, in the liver, without restricting the amyloid type, so patients with AFib, which has splenic amyloid in 80.7%, and with ALys, which has hepatic involvement in 80%, were eligible in principle; the publications describe the populations as systemic amyloidosis without breaking down the subtype [3, 15, 77, 78, 80].

The search of the international trials registry, carried out on 1 November 2026 by term and by condition, identified no clinical trial whose condition studied was fibrinogen Aα, apolipoprotein A-I, apolipoprotein A-II, lysozyme or gelsolin amyloidosis, in any phase and by any sponsor. One search pitfall deserves to be flagged: the records returned by the expression “gelsolin amyloidosis” all refer to recombinant human plasma gelsolin as a therapeutic agent in sepsis, community-acquired pneumonia, severe covid-19 and acute respiratory distress syndrome, a use opposite to that of the disease, in which gelsolin is the amyloidogenic protein, and should not be read as therapy for gelsolin amyloidosis. A negative search does not demonstrate the absence of development, but describes the state of the public registry on that date.

The therapeutic prospects set out by the authors of the largest AFib cohort are of three kinds: agents that promote clearance of existing deposits; agents that bind circulating variant fibrinogen and facilitate its removal; and suppression of production of the pathogenic protein, by allele-specific gene silencing or by gene editing strategies directed at the mutant FGA allele [3]. Such approaches, analogous to the therapeutic advances already achieved in other forms of systemic amyloidosis, would have the potential to delay progression or prevent amyloidosis, prolong renal survival in patients with established disease and eliminate the risk of recurrence after kidney transplantation [3]. Their current classification is that of a therapeutic concept, with no programme in clinical development identified. The exception is gelsolin amyloidosis, the only one of these entities with a published targeted preclinical programme: there is no therapeutic agent that specifically targets and neutralises the amyloidogenic gelsolin fragment, and the antibody fragments described in 2024 are, to date, exclusively in vitro data, with no animal model or human study [85]. A 2026 molecular review proposes classifying gelsolin gene variants by their molecular and clinical features, noting that for more than 40 years the disease was linked exclusively to the D187N and D187Y variants of the second domain, the Finnish and Danish variants, but that since 2013 numerous amyloidogenic variants have been identified in other regions of the protein, with distinct phenotypes and mechanisms, and concludes that the lack of information, biological models and drug development efforts calls for further therapeutic exploration [86].

In β2-microglobulin amyloidosis, the distinction between the hereditary and acquired forms has a direct therapeutic consequence. In the hereditary form, dialysis has no mechanistic basis, because there is no retention of wild-type β2-microglobulin: those affected have normal renal function and a normal or low circulating concentration of the protein [24, 25]. In the dialysis-related form, by contrast, reducing the plasma concentration is the therapeutic goal: most β2-microglobulin is eliminated by glomerular filtration with subsequent proximal tubular reabsorption and catabolism, so serum concentrations are inversely proportional to glomerular filtration rate and rise up to 60-fold in end-stage kidney disease, and residual renal function, probably the best determinant of these concentrations, should be preserved as far as possible [87]. The incidence of this acquired form has fallen with improvements in dialysis methods: in two Japanese national surveys that used first carpal tunnel surgery in the following year as a proxy for disease onset, in registries covering 202,726 patients, incidence was 1.77% in the 1998 cohort versus 1.30% in the 2010 cohort (p<0.05), a reduction the authors attribute to improved dialysis methods, especially dialysate quality control [88]. A β2-microglobulin-specific adsorption column, used in series with the high-flux dialyser three times a week for one year in 17 patients, reduced serum concentration from 34.5 ± 8.4 to 28.8 ± 7.3 mg/L, with improvement in pinch strength, terminal motor latency and activities of daily living, after a previous year of high-flux dialysis alone had not changed the concentration [89]; the post-approval study of this device in the United States is recorded as suspended [90]. Two randomised trials of medium cut-off dialysers, compared with high-flux dialysers, have been completed with dialysis amyloidosis among the conditions registered [91, 92]. Doxycycline has a European orphan designation for systemic amyloidosis caused by β2-microglobulin, which covers both the hereditary and the wild-type form, but the published clinical evidence is limited to a report of an analgesic effect in the dialysis-related form with 100 mg per day, with pain falling from 7 to 8 to 4 to 5 on a 0 to 10 scale over 6 months and unchanged or increased uptake on fluorodeoxyglucose positron emission tomography at 6 months, which dissociates the analgesic effect from the inflammatory burden of the lesions; there is no controlled trial in this indication [74, 93]. None of these measures applies to the hereditary form. The absence of amyloid-directed therapy remains the main limitation of care in these diseases, and is what explains why, in the British AFib cohort, renal and allograft survival improved significantly between periods while overall survival remained unchanged, with a median of about 12 years in both [3].

Among clearance therapies, the only one with documented clinical results in these entities is indirect and surgical: liver transplantation was followed by amyloid regression on serial iodine-123-labelled serum amyloid P component scintigraphy in the four patients with AApoAI so assessed [5]. This result shows that, once the source of the variant protein is removed, amyloid load can decrease in at least some of these diseases, which contrasts with the progression through wild-type protein deposition observed after liver transplantation in transthyretin amyloidosis [32] and is consistent with the observation that, in AFib, wild-type fibrinogen does not form fibrils in vivo [3].

7.9 Agents to avoid, pregnancy, perioperative care and transition

Agents to avoid. Plasma cell-directed chemotherapy has no role in these diseases and constitutes inappropriate treatment when given because of a misdiagnosis of AL amyloidosis [7, 34]. Immunosuppression with corticosteroids and cyclophosphamide in hereditary amyloid nephropathy produced no renal response and was associated with a lung abscess in one case [48]. In amyloid cardiomyopathy, beta-blockers, converting enzyme inhibitors, angiotensin receptor blockers, digoxin and calcium channel blockers should be used with caution or avoided, for the haemodynamic and toxicity reasons described [32]. Non-steroidal anti-inflammatory drugs should be avoided in nephropathy because of the risk of deterioration of renal function, as should nephrotoxins in general, in line with the avoidance recommendation set out in the management of a case of amyloid nephropathy [48]. The sodium–glucose cotransporter 2 inhibitor should be withheld during periods of prolonged fasting, surgery or critical illness, because of the risk of ketosis, and, according to the US label, at least 3 days before surgery or procedures with prolonged fasting [69, 71].

Pregnancy and lactation. The age at onset of most systemic amyloidoses is postmenopausal in women, who tend to present later than men, so the question rarely arises; with early-onset genotypes, it becomes relevant for women carriers considering pregnancy, and, since early-onset forms of transthyretin amyloidosis affect the peripheral nerves more than the heart, pregnancy does not impose an increased risk of heart failure in that rare circumstance; there are no trials or published case reports of women with heart failure due to cardiac amyloid during pregnancy [32]. The entities covered here pose different problems: with FGA frameshift variants, with onset of nephropathy at a mean of 24.5 years and cases from 12 years of age, a pregnancy may occur in a patient with established chronic kidney disease [4], a situation requiring joint planning with nephrology and obstetrics; and in a transplanted patient, pregnancy follows the rules of pregnancy after organ transplantation. The sources consulted do not describe pregnancy series in any of the six entities, and the nephroprotective drugs in question — renin–angiotensin system inhibitors and sodium–glucose cotransporter inhibitors — are not used in pregnancy, so pharmacological nephroprotection must be reviewed before conception.

Perioperative care. Three risks are specific to these diseases. The first is spontaneous visceral haemorrhage: splenic rupture occurred in the perioperative period in three of nine patients with AFib undergoing combined transplantation [57], and hepatic rupture is the characteristic complication of ALys, with the liver described as fragile and friable, with loss of the reticulin framework [34]. The second is the haemodynamic intolerance of restrictive cardiomyopathy, in which cardiac output depends on heart rate and vasodilation is poorly tolerated, which conditions anaesthesia [32]. The third is dysautonomia, present in half the patients in the King's College series with AFib and in some patients with Aβ2M, with orthostatic hypotension and delayed gastric emptying [23, 57]. Postoperative deep vein thrombosis occurred in two of the nine transplanted patients [57].

Transition of care. Transition from paediatric to adult medicine is relevant in one specific circumstance: with FGA frameshift variants, nephrotic syndrome may begin in the second decade — the son of the proband of the French family presented at 12 years, started peritoneal dialysis one year after the onset of nephrotic syndrome and had a kidney transplant at 15 years [36] —, and in the German ALys cohort the disease manifested at an early age, with a mother and son followed in paediatrics [14]. In these situations, transition should ensure continuity of nephrological follow-up, recording of the familial variant and reassessment of the indication for transplantation.

7.10 Guideline recommendations and divergences

There is no guideline or consensus dedicated to the non-transthyretin hereditary amyloidoses. The applicable recommendations come from three documents with different scopes, none of which has these diseases as its main subject. The ISA provides the nomenclature and classification, without therapeutic recommendations [31]. The European Society of Cardiology position statement establishes the diagnostic criteria, the red flags by amyloid type, the requirement for typing by mass spectrometry as the reference standard, the rules for genetic testing and family screening and the follow-up intervals, and advocates a network model with referral to reference centres [28]. The 2024 nephrology guideline provides the nephroprotection recommendations, with explicit grade and level, but for chronic kidney disease in general [69]. The 2023 North American consensus document on cardiac amyloidosis devotes a short section to the diagnosis of the rare forms, and its literal content matters because it is often misquoted: it lists as rare causes of cardiac amyloidosis AA amyloidosis and the hereditary apolipoprotein A-I and apolipoprotein A-IV amyloidoses, and attributes a history of chronic inflammatory disease — rheumatoid arthritis, spondyloarthritis, inflammatory bowel disease, chronic infections and familial periodic fever syndromes — specifically to AA amyloidosis, and not to apolipoprotein A-I amyloidosis [62]. It recognises that cardiac amyloidosis has been described from nine precursor proteins, that more than 90% to 95% of cases result from two of them, and that the useful clinical clues to identify the rare forms are family history and dialysis status, requiring in all cases histological assessment by biopsy when clinical suspicion exists [62]. For the practice of this article, the most consequential statement in that document is another: there was positive uptake on technetium-99m-labelled 3,3-diphosphono-1,2-propanodicarboxylic acid scintigraphy in apolipoprotein A-I, apolipoprotein A-IV and β2-microglobulin amyloidoses, while the Phe64Leu transthyretin variant gives false negatives, so a positive bone scintigraphy does not prove transthyretin amyloidosis in any of these entities [62]. The same document sets liquid chromatography with tandem mass spectrometry as the reference standard for tissue diagnosis, with a sensitivity of 88% and specificity of 96% [62].

The divergences between documents are three and have consequences. The first is the classification of β2-microglobulin amyloidosis: the European statement considers it non-hereditary and dialysis-related, whereas the hereditary form has been established since 2012, has two variants described in independent families and has had its own OMIM number since December 2023 [23–25, 28]. The second is the classification of apolipoprotein A-IV amyloidosis, which the European statement considers non-hereditary [28]. The third is the indication for transplantation in AFib: the 2010 series proposed combined liver–kidney transplantation as a potentially curative strategy, based on the failure of isolated renal allografts between one and seven years from recurrent amyloidosis [45, 57], and the 2026 cohort from the same centre, with a median isolated renal allograft survival of 15.5 years in transplants performed after 2009, concludes that isolated kidney transplantation is probably appropriate for most patients, reserving combined transplantation for younger patients in better condition [3]. The French series adds an intermediate criterion, favouring isolated transplantation in carriers of the E526V variant and combined transplantation in carriers of frameshift variants [4]. The reference for the current decision is the most recent and largest cohort, read together with the variant criterion.

As for control of blood pressure and proteinuria, the reference is the nephrology guideline, with the grade of recommendation indicated for each measure, and the dapagliflozin labels serve as a comparison for the dose, which coincides between jurisdictions at 10 mg once daily [69–71]. None of these recommendations was obtained in patients with hereditary amyloidosis, and their application to these entities is an extrapolation that should be made explicit to the patient.

8. Reproductive Options and Family Counselling

8.1 Recurrence risk

The six entities are transmitted in an autosomal dominant manner, so each child of a heterozygous parent has a 50% probability of inheriting the variant, regardless of sex [8, 41]. The siblings of an index case with a heterozygous parent have the same 50% risk. Gelsolin amyloidosis is the only one of the six in which homozygotes have been documented: two patients from the 1973 Finnish national survey, whose parents were both affected, had more severe disease than the others and were considered homozygous, and two sisters homozygous for the D187N variant described in 1993 had unusually severe disease, with nephrotic syndrome and end-stage renal failure, with abundant glomerular deposits [20]. In couples in which both members carry a GSN variant, the risk of homozygous offspring is 25%, with the expectation of a more severe, predominantly renal phenotype, which is relevant information for counselling in populations with a high allele frequency, such as the Finnish population.

The proportion of variants arising de novo has not been established in any of these entities. In the case of gelsolin amyloidosis with the p.Asn160Lys variant described in 2026, the variant was absent in the patient's son and assessment of the parents was not possible because both had died, so the authors left open the possibility of a de novo event or incomplete penetrance [48]. The sources consulted do not describe germline mosaicism in any of these diseases; in the absence of data, the prudent approach is the one used in other adult-onset dominant diseases, which is not to assume zero recurrence risk for siblings when the variant is not detected in either parent.

8.2 Penetrance, expressivity and the weight of family history

Information on penetrance and expressivity determines the quality of counselling and differs considerably between the six entities. In ALys, clinical penetrance is high: a family history of amyloidosis was present in each of the 16 affected individuals in the British cohort, indicating high clinical penetrance in the presence of an amyloidogenic lysozyme variant; but the phenotype, although similar within each family, does not correlate with genotype, and marked differences have been observed between families carrying the same variant [14, 16]. In AApoAI, the situation is the opposite: a family history is often absent — only one of four patients in one series had it —, and the spectrum ranges from extensive visceral deposits with end-stage renal failure in young adults to laryngeal or cutaneous amyloid of little clinical consequence, depending on the variant [7, 52]. In AFib, none of the 18 patients with the E526V variant identified among 350 patients with presumed AL amyloidosis knew of a relevant family history, although genealogical studies revealed kinship in four of them and geographical proximity of ancestors in another two, and one patient only retrospectively found out that her dizygotic twin had died of renal failure at 76 years [29]. In the North American series of 13 patients with AFib, a family history of amyloidosis or chronic kidney disease was present in five (42%) [1].

The practical consequence is that the absence of a family history does not allow hereditary amyloidosis to be excluded, either in the diagnostic assessment or in counselling. The authors who described five new APOA1 variants expressly recommend sequencing the gene in patients with apparently localised amyloidosis in whom immunohistochemistry is not diagnostic of the fibril protein, even in the absence of a family history [7]. The explanation for apparently negative family histories includes reduced penetrance, early death of the parent from another cause, late onset and misdiagnosis in earlier generations — end-stage nephropathy in a generation without access to renal biopsy and typing was not recorded as amyloidosis.

Intrafamilial expressivity is variable even when the overall phenotype is similar. In AFib, the age at onset of kidney disease varied within the same variants: 36 to 77 years in E526V carriers and 12 to 59 years in R554L carriers, in the French series [4]. In AApoAI, the American family with the R173P variant had members with purely cutaneous expression, members with voice changes due to vocal cord deposition, members with hypertrophic cardiomyopathy and one member with multiorgan amyloidosis and hepatic and renal failure [6]. In AGel, five of six members of one family had corneal dystrophy and lax skin but only three had arrhythmia requiring an implantable device [20].

8.3 Cascade testing

Genetic testing is recommended for relatives of people with hereditary forms of amyloidosis, accompanied by genetic counselling [28]. Testing confirms the diagnosis, informs prognosis and allows cascade screening of first-degree relatives, and the autosomal dominant pattern makes such screening important [8]. In the AApoAI case published in 2026, the diagnosis in the index patient was possible because his brother had been diagnosed days earlier, with identification of the APOA1 variant, which led to sequencing of that gene being requested after TTR sequencing had been normal; first-degree relatives then underwent cascade genetic screening [8]. This sequence illustrates the value of family screening as a diagnostic tool, and not only a predictive one: in a patient with amyloidosis that is difficult to type, the genotype of an affected relative can direct the investigation.

Genetic testing of first-degree relatives should be offered at the ages at which the familial variant tends to manifest, or earlier if the person concerned wishes it for family planning [32]. Testing of minors is discouraged, but may be offered in early adulthood if the results help with career or reproductive planning [28]. There is, however, a clear exception in these entities: with FGA frameshift variants, in which nephropathy began at 12 years in two patients described and at a mean of 24.5 years in the French series [4, 36], and in families with ALys, in which the median age at symptom onset is 23.7 years and 43% of patients report symptoms in childhood [14, 15], bringing testing forward is justified by the possibility of symptomatic disease in adolescence, in which case the test ceases to be predictive and becomes diagnostic.

Monitoring of asymptomatic carriers should, in general, begin about ten years before the typical age at onset in affected relatives, or earlier if symptoms appear [28]. This rule, formulated for transthyretin amyloidosis, is the only one available and transfers to the other entities with the caveat that age at onset varies within the same family and with the same variant. In the predominantly renal entities, the minimum content of such monitoring is simple and accessible: blood pressure, serum creatinine with estimated glomerular filtration rate and quantification of albuminuria, which are also the parameters that define disease stage and eligibility for nephroprotection [3, 69]. In AGel, slit-lamp examination is the earliest detection tool, given that corneal lattice dystrophy is practically a prerequisite for the diagnosis and the first sign to appear, at a mean age of 43 years [19, 20].

8.4 Prenatal and preimplantation diagnosis

Preimplantation genetic diagnosis is technically possible in all these entities, because all variants are known at the nucleotide level, and is an option to avoid transmission to subsequent generations; with variants of relatively high penetrance, the challenges of this technique are more relevant than with those of late onset and probably low penetrance [32]. The sources consulted document its application to transthyretin amyloidosis, and do not describe cases in any of the six entities covered here.

The considerations that counselling should set out differ from those of a disease of paediatric onset or one that is invariably fatal. In the predominantly renal forms, the disease generally manifests between the fifth and seventh decade, median survival from diagnosis is about 12 years in AFib and the time from diagnosis to end-stage kidney disease is 15 years in AApoAI, with effective replacement therapy and transplantation available, with a median allograft survival of 15.5 years in AFib and 22 years in AApoAI [3, 5]. In AGel, relative survival does not differ from that of the general population up to 75 years, although the disease burden is significant, with ophthalmological, neurological and cutaneous morbidity and repeated surgery [18]. With FGA frameshift variants, by contrast, end-stage kidney disease may occur in the second or third decade, with dialysis and transplantation in adolescence [4, 36], and ALys carries the added risk of potentially fatal spontaneous hepatic haemorrhage, with three consecutive deaths in three generations in one family [34]. The reproductive decision should be informed by these data, by the family's specific variant and by its expression in affected members, and not by the generic category of hereditary amyloidosis.

8.5 Information to give the family

Three points deserve explicit communication. The first is that identification of the fibril protein and the variant has a direct therapeutic consequence, because it determines whether or not chemotherapy is indicated and whether there is a rationale for liver transplantation; wrong attribution to AL amyloidosis exposes the patient to useless chemotherapy and deprives them of potentially curative transplantation [6, 40]. The second is that screening of relatives has a dual value, predictive for them and diagnostic for the index case when typing is ambiguous [8]. The third is that there is, to date, no drug therapy that prevents or reverses deposition in any of these entities, and that follow-up of the asymptomatic carrier aims to detect organ involvement early in order to start nephroprotection and plan replacement of function, and not to administer preventive treatment [3, 21].

9. Special Considerations: Population Diversity and Epidemiology

9.1 What is known and what cannot be measured

None of the six entities has a validated population prevalence estimate. Orphanet has recorded a point prevalence below 1 per 1,000,000 for lysozyme amyloidosis, gelsolin amyloidosis, variant β2-microglobulin amyloidosis and apolipoprotein A-IV amyloidosis, all as a class and not as a value, and has no numerical estimate for fibrinogen Aα amyloidosis, apolipoprotein A-I amyloidosis or apolipoprotein A-II amyloidosis; the parent entity “hereditary amyloidosis with primary renal involvement” is recorded with unknown and non-validated prevalence [2, 13, 17, 22, 42, 43, 68, 94]. The figures available are counts of cases or families — 475 cases of gelsolin amyloidosis, 7 families with lysozyme amyloidosis, 5 cases of variant β2-microglobulin amyloidosis and 2 cases of apolipoprotein A-IV amyloidosis, all validated records [13, 17, 22, 68] —, and counts of referrals to national centres. A referral count is not a population denominator: the series of 194 patients with fibrinogen amyloidosis from a UK national centre, accumulated between 1993 and 2025, that of 57 patients with apolipoprotein A-I amyloidosis from the same centre, accumulated between 1986 and 2019, and that of 35 patients with lysozyme amyloidosis, accumulated between 1988 and 2025 and corresponding to eight distinct families, describe the activity of a reference centre and not the frequency of the disease in the population [3, 5, 15].

The best available approximation to the relative frequency of these entities comes from the case mix of reference typing laboratories. In a North American central laboratory, 16,175 specimens were typed by mass spectrometry-based proteomics between 1 January 2008 and 31 December 2018, identifying 21 amyloid types: light-chain amyloidosis in 9542 cases (59.0%), transthyretin amyloidosis in 4600 (28.4%), ALECT2 in 511 (3.2%), AA amyloidosis in 463 (2.9%), heavy-chain amyloidosis in 367 (2.3%), insulin amyloidosis in 182 (1.2%), keratin 5 and 14 amyloidosis in 94, fibrinogen Aα amyloidosis in 71, apolipoprotein A-IV amyloidosis in 57, apolipoprotein A-I amyloidosis in 56, atrial natriuretic factor amyloidosis in 47, β2-microglobulin amyloidosis in 38, semenogelin 1 amyloidosis in 34, gelsolin amyloidosis in 29, transforming growth factor beta 1 amyloidosis in 29, lysozyme amyloidosis in 15, islet amyloid polypeptide amyloidosis in 13, apolipoprotein C-II amyloidosis in 11, prolactin amyloidosis in 8, enfuvirtide amyloidosis in 6 and calcitonin amyloidosis in 2 [95]. With a modified bioinformatic pipeline, amino acid substitutions in hereditary amyloidosis cases were detected with 100% specificity, and the work produced the first organ-by-type map of the 21 entities in 31 organs [95]. Converted into proportions of that case mix, the values for the entities covered here are 0.44% for fibrinogen Aα amyloidosis, 0.35% for apolipoprotein A-I amyloidosis, 0.23% for β2-microglobulin amyloidosis, 0.18% for gelsolin amyloidosis and 0.09% for lysozyme amyloidosis; these are not population prevalences, but the composition of a reference case mix sent for typing, and are therefore subject to referral patterns and the geographical availability of the technique. In a series of 127 cases of renal amyloidosis typed at the same centre between 2008 and 2010, the proteins identified in the deposits included immunoglobulin light and heavy chains, serum amyloid A, LECT2, fibrinogen alpha chain, transthyretin, apolipoproteins A-I and A-IV, gelsolin and β2-microglobulin, that is, the six entities of this article all appeared in the renal case mix of a single centre over three years [96].

Two contextual measures are available and useful. Systemic amyloidosis is the diagnosis in 2.5% of all renal biopsies, according to the British glomerulonephritis registry, and is the cause of death in more than 1 in every 1500 people per year in the United Kingdom, data cited in the amyloidosis reclassification study [29]. Hereditary amyloidoses account for less than 5% of cases of renal amyloidosis, and fibrinogen Aα amyloidosis is the most common hereditary form affecting the kidney, more frequent in Europe than in the United States [1]. In cardiac amyloidosis, the proportion is even smaller: light-chain amyloidosis and transthyretin amyloidosis together account for more than 98% of diagnosed cases, and of the more than 30 proteins capable of forming amyloid only nine accumulate in the myocardium in amounts that cause significant disease [28].

9.2 Geographical distribution and founder populations

Fibrinogen Aα amyloidosis is the most common hereditary form of renal amyloidosis in Europe, and possibly also in the United States [57]. The dominant variant, E526V, has been described mainly in Northern Europeans and identified in families worldwide [1]. Haplotype analysis of four families with this variant, two Irish-American, one Polish-Canadian and another Irish-American, suggested derivation from a single founder [36], and the 18 patients with the variant identified in the United Kingdom were all of Northern European ancestry, two of them cousins and the ancestors of another two coming from neighbouring villages [29]. Outside Europe, the entity has been described in people of Chinese, Korean, Russian, North African, Japanese and Peruvian ancestry [39]; Benson's original family was Peruvian and carried the R554L variant [29]. In the North American Mayo Clinic series of 13 patients, all were white [1], and in Uemichi's two families the ancestry was Irish [29].

Gelsolin amyloidosis is the entity with the best-characterised geography: it is one of the most common diseases of the Finnish disease heritage, with a national survey that identified 207 patients in 1973 and 307 in 1978, and a national registry that included 227 patients in 2015 and 261 in its updated phase, covering about one third of Finnish patients, with the survival analysis covering 478 patients [18–20]. The D187N variant was found in all Finnish families studied [20]. Outside Finland, cases have been described in the United States, the Netherlands and Denmark, three Czechoslovak sisters with bulbar palsy, hyperelastic cutis and corneal lattice dystrophy described in 1959 may have had the disease, and two Japanese siblings, aged 70 and 68, were described with corneal lattice dystrophy in the context of this amyloidosis; the same D187N variant was found in the affected son of the proband of a Scottish-American family [20]. Two families with variants outside the canonical domain were described in 2021, one with the E580K variant and the other, with three affected first-degree relatives, with the W493R variant [20]. The case described in 2026, of a 58-year-old Chinese man with the p.Asn160Lys variant and isolated renal presentation, shows that the disease is recognised outside Europe with atypical phenotypes [48].

Lysozyme amyloidosis has been described in about 30 families worldwide, with nine amyloidogenic variants [14]. The published families are English (the two original families and a third with hepatic rupture), French, Italian-Canadian and Italian from the Piedmont region, with two independent women carrying the same W64R variant resulting from a nucleotide change distinct from the French one, and an English man with the same variant; the German cohort added ten patients from six families, seven of them with the W64R variant [14, 34]. The aggregation of three W64R carriers in the same Italian region suggests a local founder effect, but the sources consulted do not quantify it.

Apolipoprotein A-I amyloidosis has 14 different variants documented in a single British national series of 57 patients, and the most frequently identified variant, p.Gly26Arg, was present in 28 of these patients [5]. The founder families described are of English, Scottish and Irish ancestry (the Iowa kindred), English (L60R variant), Ashkenazi Jewish (W50R variant), Spanish (exon 4 deletion and insertion), French (L90P variant), American (R173P variant) and Italian (L174S variant), in addition to the German family described in 2026 with the p.L202P variant [6, 8].

Variant β2-microglobulin amyloidosis has only two variants described, and the second defines a focus: the P32L variant was described in a Portuguese family with three affected members, and later identified in another patient of Portuguese ancestry, the authors considering a common ancestor possible [23, 25]. The index family, with the D96N variant, is French [24]. This is, therefore, the only one of the six entities with an identified Portuguese population focus, albeit built on four patients from two families, a number that does not allow any frequency estimate.

Apolipoprotein A-II amyloidosis is known from isolated families and the three variants described were identified in Caucasian patients, without a defined geographical pattern [10–12, 37].

9.3 Data from Portugal and Brazil

In the sources consulted, Portuguese data on these entities are of two kinds. The first is the familial focus of variant β2-microglobulin amyloidosis described above [23, 25]. The second is the diagnostic experience of a reference laboratory: a series of five diagnostically difficult patients followed at a hospital unit in Lisbon, in collaboration with a structural chemistry institute, in which mass spectrometry established or revised the amyloid subtype in all cases, despite misleading or inconclusive conventional studies [63]. That series did not include any of the six non-transthyretin hereditary entities, but documents the availability of the proteomic typing technique in Portugal and illustrates the national context: in one of the patients, with a pathogenic genetic test for transthyretin amyloidosis and a compatible clinical picture, mass spectrometry identified lambda light-chain amyloidosis, and the clinical decision took into account the significant prevalence of hereditary transthyretin amyloidosis in Portugal, with tafamidis maintained under close monitoring of the response and of the plasma cell dyscrasia [63]. The point prevalence of hereditary transthyretin amyloidosis in Portugal, recorded and validated by Orphanet, is 19.91 per 100,000, about 130 times higher than the estimate recorded for the United Kingdom, of 0.149 per 100,000, which constitutes the epidemiological context in which any amyloidosis is diagnosed in the country [26].

The sources consulted contain no Brazilian data on these six entities, nor series or cases published in Brazil. This is a real gap and should not be interpreted as an absence of patients: fibrinogen Aα amyloidosis has been described in populations of European ancestry, Iberian included, and in patients of Peruvian ancestry, and variant β2-microglobulin amyloidosis is associated with a Portuguese focus, two circumstances that make its occurrence in Brazilian populations plausible. The search was not directed by country, in line with the methodology adopted, so the absence also reflects the design of the search.

9.4 Special populations

Sex differences are not described as relevant in any of the six entities, with two isolated observations. In the British fibrinogen amyloidosis cohort, there was a slight male predominance, with 55.6% men [3], and in the North American series 10 of the 13 patients were men [1]. In the Finnish gelsolin amyloidosis registry, relative survival exceeded 1 for men and women up to the 70 to 74 years group, without difference between sexes [18]. In the amyloidoses in general, age at onset is later in women, who tend to present after men [32].

Paediatric age is relevant in two situations. With FGA frameshift variants, kidney disease appeared at a mean of 24.5 years, with cases from 12 years of age, and the son of the proband of the French family presented with nephrotic syndrome at 12 years, started peritoneal dialysis one year later and had a kidney transplant at 15 years [4, 36]. In lysozyme amyloidosis, the disease may manifest at an early age, and the German cohort included paediatric follow-up [14]; the proband of the family with hepatic rupture was 15 years old when he underwent emergency laparotomy [34]. The French fibrinogen amyloidosis series has a minimum age at diagnosis of 12 years [4], and the nephrology review notes that the presentation is heterogeneous and may occur from childhood to old age [39].

Patients on dialysis and transplant recipients are the predominant population in the predominantly renal entities: in the British fibrinogen amyloidosis cohort, 122 of the 194 patients (62.8%) reached end-stage kidney disease, 71 started haemodialysis, 37 peritoneal dialysis and 14 had a pre-emptive kidney transplant, and 16 of those who started peritoneal dialysis (43.3%) converted to haemodialysis after a median of 1.4 years (0.5 to 3.4) [3]. Frequent conversion from peritoneal dialysis to haemodialysis is a practical consideration in planning the replacement modality.

9.5 Reference networks and patient associations

There is no European Reference Network dedicated to the hereditary amyloidoses. The thematic groups of the European network for rare kidney diseases, ERKNet, are hereditary glomerulopathies, immune glomerulopathies, tubulopathies, metabolic diseases and stones, thrombotic microangiopathies, autosomal dominant structural kidney diseases, congenital malformations and ciliopathies, obstructive uropathies, paediatric chronic kidney disease, paediatric dialysis and paediatric kidney transplantation, and amyloidosis is not included in any of them [27]. In European practice, amyloid polyneuropathy falls within the network for neuromuscular diseases, ERN EURO-NMD, and amyloid cardiomyopathy within the network for rare, low-prevalence and complex diseases of the heart, ERN GUARD-Heart. The European Society of Cardiology position statement advocates, for amyloidosis, a network model in which local centres refer patients to regional or national reference centres for selected procedures or complex decisions, and recommends that patients with light-chain amyloidosis be referred to specialised centres whenever possible, with multidisciplinary haemato-oncology and cardiology teams [28].

In the absence of a formal network, the effective organisation of care in these diseases rests on national amyloidosis centres. The UK national centre, based at the Royal Free Hospital in London, brought together the reference cohorts of four of the six entities — 194 patients with fibrinogen amyloidosis, 71 in the historical series of the same entity, 57 with apolipoprotein A-I amyloidosis and 35 with lysozyme amyloidosis —, with an initial diagnostic assessment followed by follow-up at regular intervals, usually 12 months, including iodine-123-labelled serum amyloid P component scintigraphy [3, 5, 15, 56]. The tissue proteomics laboratory of a North American centre functions as a national reference laboratory for amyloid typing [1]. The Finnish gelsolin amyloidosis registry, FIN-GAR, founded in 2013, is maintained in conjunction with the Finnish patient association, to which the questionnaires that feed the registry were sent [18]. These structures are the realistic route of access to proteomic typing, genetic testing and transplant assessment in these diseases.

Among patient organisations, the Amyloidosis Alliance presents itself as a worldwide alliance of patient organisations, with the aim of improving awareness, diagnosis, care, research and representation in amyloidosis, based in Paris; the institutional page consulted does not list the member organisations [97]. The Finnish amyloidosis association collaborates in maintaining the national gelsolin amyloidosis registry [18]. The sources consulted do not identify patient associations dedicated specifically to any of the six entities, either in Portugal or in Brazil. These structures do not produce epidemiological data, but are relevant for counselling, referral and access to research.

9.6 Declared gaps

Besides the absence of prevalence estimates, three gaps deserve recording. The first is the discrepant count of FGA variants: 16 published according to the 2026 cohort, 15 according to the 2025 series and more than 15 according to the nephrology review of the same year, a divergence that reflects the date of consultation of each source and the hereditary amyloidosis variant database that serves as their reference, not consulted directly [1, 3, 39, 98]. The second is the absence of survival data for apolipoprotein A-II amyloidosis in the European statement, which recorded median survival values for apolipoprotein A-IV amyloidosis (79 months) and fibrinogen amyloidosis (180 months) but left the others blank [28]. The third is that the complete list of the 42 human fibril proteins in the 2024 nomenclature, with the systemic or localised classification of each, is not accessible in an open source; the aggregate figures used in this article come from the abstract of the document, and the list of entities in section 2 was built from the European statement, the nephrology review, the OMIM phenotypic series and the Orphanet records [28, 31, 39, 41].

10. Illustrative Clinical Case

10.1 Apolipoprotein A-I cardiac amyloidosis diagnosed after the brother

A 61-year-old man, followed at a German university centre, presented with progressive exertional dyspnoea, fatigue in New York Heart Association functional class III and mild dysphonia, with a blood pressure of 160/100 mmHg and a heart rate of 68 per minute [8]. Before this presentation he had been assessed for several years by his family doctor and by local nephrology and cardiology. His history included bilateral carpal tunnel syndrome, not operated on, and chronic kidney disease category G3bA1, of undetermined aetiology for three years. N-terminal pro-B-type natriuretic peptide had risen over the years. Transthoracic echocardiography showed increased myocardial wall thickness with preserved ejection fraction and diastolic dysfunction, and exercise testing revealed reduced functional capacity without signs of ischaemia on the electrocardiogram; stress echocardiography showed no wall motion abnormalities. Carotid artery ultrasound revealed atherosclerosis without significant stenosis. Because his older brother had severe coronary disease that had required revascularisation, computed tomography coronary angiography was performed, which revealed non-obstructive coronary disease, with a calcium score of 79. Years earlier, Fabry disease had been excluded by enzyme testing. The family history was significant: his brother had amyloidosis, with the diagnosis confirmed a few days before this presentation and identification of an apolipoprotein A-I gene variant, and his mother had died of presumed amyloid-related disease and heart failure. Medication at presentation was acetylsalicylic acid 100 mg and atorvastatin 20 mg, both once daily [8].

The 12-lead electrocardiogram showed normal sinus rhythm with pre-terminal negative T waves, without a pseudoinfarct pattern or formal low-voltage criteria, but with QRS voltage disproportionately low relative to the increased left ventricular wall thickness [8]. Echocardiography revealed increased wall thickness of the left and right ventricles, with an end-diastolic interventricular septum of 17 mm and an end-diastolic right ventricular free wall of 10 mm, preserved ejection fraction (55% by the biplane method and 52% in three dimensions), diastolic dysfunction with a mean E/E' ratio of 18, biatrial dilatation, minimal pericardial effusion, valvular thickening and moderate tricuspid regurgitation; left ventricular global longitudinal strain did not show the characteristic apical sparing pattern [8]. Cardiac magnetic resonance imaging demonstrated increased wall thickness in both ventricles, interatrial septal hypertrophy, minimal pericardial effusion, inadequate myocardial nulling and diffuse late gadolinium enhancement, findings compatible with cardiac amyloidosis [8]. Tests showed raised N-terminal natriuretic peptide, of 1210 pg/ml, which rose to 1580 pg/ml before the start of treatment; serum and urine immunofixation was negative; serum free light chains were slightly raised, with lambda of 32 mg/L and kappa of 33 mg/L, and a normal kappa/lambda ratio of 1.04; and high-density lipoprotein cholesterol was slightly decreased, at 0.68 mmol/L [8].

Sequencing of exons 1 to 4 of the transthyretin gene revealed no changes from the reference sequence. With the brother's result then known, sequencing of the apolipoprotein A-I gene was requested, which identified the change NM_000039.3:c.605T>C, corresponding to a pathogenic variant in exon 4, p.L202P, establishing the diagnosis of apolipoprotein A-I amyloidosis [8]. For liver transplant assessment and follow-up of a previous gastric Helicobacter pylori infection, upper gastrointestinal endoscopy was performed, with biopsies of a small ulcer in the gastric antrum: amyloid deposits were identified in the lamina propria by polarisation microscopy, and subsequent histochemical analysis showed immunoreactivity for complement C9 and for anti-apolipoprotein A-I antibodies [8].

Dapagliflozin was started and the patient was referred to hepatology for liver transplant assessment; first-degree relatives underwent cascade genetic screening, the results of which were not disclosed [8]. Symptoms improved substantially, to functional class I, with N-terminal natriuretic peptide falling to 804 ng/L, and at nine months of follow-up the patient remained clinically stable, without heart failure hospitalisations [8].

The management in this report does not constitute established routine. The authors themselves state that the efficacy of sodium–glucose cotransporter 2 inhibitors has been described in transthyretin cardiomyopathy and that their use in apolipoprotein A-I amyloidosis is extrapolated, possibly improving symptomatic status, and that liver transplantation remains the definitive therapy to reduce amyloid production [8]. The case has three transferable lessons that the authors highlight: consider a hereditary amyloidosis such as apolipoprotein A-I amyloidosis in patients with a familial pattern and unusual presentations, including systemic features and right heart disease; genetic testing confirms the diagnosis, informs prognosis and allows cascade screening of first-degree relatives; and sodium–glucose cotransporter 2 inhibitors may improve symptoms even in rare hereditary cardiac amyloidosis [8]. The prognosis of this entity is, compared with light-chain amyloidosis, favourable, with a median survival of more than 172 months according to the data cited by the authors [8].

10.2 Gelsolin amyloidosis with isolated renal presentation

A 58-year-old Chinese man was admitted in January 2025 for assessment of persistent urinary abnormalities documented for more than 30 years and raised serum creatinine for more than 2 years [48]. The initial urinalysis, three decades earlier, showed proteinuria fluctuating between trace and one plus, without microscopic haematuria. Two years before admission, urinalysis showed two-plus proteinuria and serum creatinine of 120 µmol/L. His history included arterial hypertension for 12 years, with a maximum blood pressure above 160/100 mmHg, chronic non-atrophic gastritis and reflux oesophagitis for 3 years, and 35 years of smoking with an average of 20 cigarettes per day. There was no known family history of kidney disease. Medication included sacubitril with valsartan 100 mg once daily and levamlodipine besilate 2.5 mg once daily for blood pressure control, and omeprazole for gastroprotection [48].

The first renal biopsy, performed at another centre in August 2023, gave an ambiguous result. Immunofluorescence showed segmental glomerular deposits of all immunoglobulins, complement and light chains, with IgA predominance. Light microscopy of 27 glomeruli identified global glomerulosclerosis in 8 of 13 and segmental in 5 of 13, with mesangial hypercellularity, matrix expansion containing periodic acid–Schiff-positive amorphous material, moderate interstitial fibrosis and tubular atrophy and arteriolar sclerosis. Congo red staining was equivocal, and immunohistochemistry for DNAJB9 and fibronectin was negative, excluding fibrillary glomerulonephritis and fibronectin glomerulopathy. Electron microscopy revealed mesangial expansion with abundant short, rigid, rod-shaped fibrils and parallel 10 nm filaments, also present in the subendothelial and subepithelial spaces, with amorphous electron-dense deposits in the mesangium and diffuse foot process effacement. Taken together, these morphological features did not yield a definitive pathological diagnosis [48].

Investigation of secondary causes — monoclonal immunoglobulin, cryoglobulins, neoplasia and autoimmune diseases — was negative. The patient started prednisolone 20 mg per day, hydroxychloroquine, dapagliflozin and irbesartan, without improvement, with slow progression of renal function: 24-hour proteinuria of 6.6 g per day, albumin of 29.9 g/L and creatinine of 145 µmol/L. Immunosuppression was switched to a Tripterygium wilfordii preparation, the nephrotic syndrome persisting with a slight further decline in renal function, which prompted readmission in January 2025 [48].

The second renal biopsy showed, on immunofluorescence, diffuse mesangial IgA, C3 and light chains at one to two plus, with the other immunoglobulins and complement fractions minimal; on light microscopy, severe glomerulosclerosis, global in 8 of 10 glomeruli and segmental in 1 of 10, with marked mesangial expansion containing weakly periodic acid–Schiff-positive amorphous nodular deposits, severe interstitial fibrosis and tubular atrophy and arteriosclerosis; and, on electron microscopy, a thickened and laminated glomerular basement membrane with amorphous electron-dense deposits and random 10 nm fibrils, without the bundles of parallel filaments seen in the first biopsy. Congo red was weakly positive, mainly in the mesangial area, with equivocal birefringence, and stains for DNAJB9 and fibronectin were again negative [48].

For a definitive diagnosis, laser microdissection with mass spectrometry identified high spectral counts of amyloid-associated proteins — apolipoprotein E and serum amyloid P component — and of gelsolin, alongside immunoglobulins and complement, indicating hereditary gelsolin amyloidosis with suspected concomitant IgA nephropathy; immunohistochemistry confirmed intense gelsolin deposition in the glomeruli, interstitium and vessels, in both biopsies [48].

Considering the overlapping IgA nephropathy, prednisolone 40 mg per day with cyclophosphamide was started in January 2025, reaching a cumulative cyclophosphamide dose of 2.8 g by the end of April. Treatment was stopped in May because of the development of a lung abscess and, once the infection was controlled, the regimen was adjusted to low-dose prednisolone, 15 mg per day. No renal response was observed, although renal function remained stable. Follow-up examinations were negative for neurological, corneal or cutaneous involvement, but Congo red-positive amyloid deposits were identified in the gastric arterioles, in 2023, and intestinal arterioles, in 2025, on endoscopic biopsies [48].

Genetic testing revealed a heterozygous likely pathogenic variant in the gelsolin gene, c.480C>A, p.Asn160Lys, absent in the patient's son; genetic analysis of the parents was not possible, as both had died. Mass spectrometry corroborated the findings, detecting peptide fragments specific to the mutant allele in the patient, without corresponding wild-type peptides being identified. Molecular structure analysis, based on the gelsolin crystal structure, showed that in the wild type the N160 side chain forms a close interaction with D187 in domain 2, stabilising the conformation of the calcium-binding pocket, and that, with the lysine substitution, this interaction is weakened, with the distance increasing from 2.9 to 4.1 ångström, although the distance between D187 and the calcium ion remains at about 4.0 ångström [48].

The lack of clinical response to initial and intensified immunosuppression does not support an overlapping immune-complex-mediated glomerulonephritis, and the strategy shifted to supportive care aimed at delaying progression, with strict blood pressure control, avoidance of nephrotoxic agents, regular monitoring of nephrological and cardiovascular parameters and patient education [48]. The immunosuppression given in this case should not be read as recommended management: it was tried because of suspicion of a second glomerulopathy, produced no renal response and was associated with a serious infectious complication, and the authors themselves conclude that clinical management should focus on controlling the complications of amyloidosis and delaying the progression of renal failure [48]. The case illustrates three points: gelsolin amyloidosis may present with predominant renal and systemic arteriolar involvement, without the typical family history and without cutaneous, neurological and ocular manifestations; the pathological findings may mimic fibrillary glomerulonephritis or other hereditary nephropathies; and mass spectrometry and genetic testing are decisive in avoiding misdiagnosis [48].

11. Prognosis and Long-Term Follow-up

11.1 Prognosis by entity

Fibrinogen Aα amyloidosis has an unfavourable renal prognosis and a relatively favourable vital prognosis. In the largest published cohort, 62.8% of patients progressed to end-stage kidney disease, at a median age of 61 years, and the mean rate of glomerular filtration loss was 5.3 ml/min per 1.73 m² per year; median survival from diagnosis was 12.7 years in patients diagnosed before 2009 and 12.2 years in those diagnosed afterwards, with no significant difference [3]. The historical series from the same centre, with 71 patients, described a median time from presentation to end-stage kidney disease of 4.6 years, with 62% of patients reaching that point, and estimated survival of 15.2 years from clinical presentation and 10.9 years from diagnosis [56], and the European position statement recorded a median survival of 180 months for this entity [28]. Patients with end-stage kidney disease had a median survival of 8.2 years from the start of replacement therapy and of 9.3 years censored at transplantation, substantially higher than that of patients with light-chain amyloidosis, treated or untreated, and not lower than that of younger patients with non-diabetic kidney disease [56]. The independent predictors of progression to end-stage kidney disease, in the multivariable analysis, are period of diagnosis (hazard ratio 0.44 for the period after 2009; 95% confidence interval 0.28 to 0.70; p=0.018), age at diagnosis (0.95 per year; 0.92 to 0.98; p=0.0003), baseline estimated glomerular filtration rate (0.97 per ml/min per 1.73 m²; 0.96 to 0.99; p<0.0001) and baseline 24-hour proteinuria (1.08 per gram; 1.00 to 1.16; p=0.04) [3]. The only independent predictor of death was age at diagnosis (hazard ratio 1.04 per year; 1.00 to 1.07; p=0.03), and the protective factor was the presence of a kidney transplant (0.54; 0.29 to 0.99; p=0.05) [3]. Renal survival in the North American series is lower, at 44%, 33% and 11% at 1, 2 and 5 years from the diagnostic biopsy, and patient survival 92%, 82% and 64% at the same time points, in a cohort in which four of the 13 patients were already in chronic kidney disease stage 5 at diagnosis [1].

Apolipoprotein A-I amyloidosis is described as a slowly progressive disease, difficult to diagnose, and transplantation outcomes are favourable: the median time from diagnosis to end-stage kidney disease was 15.0 years (95% confidence interval 10.0 to 20.0) and 19.0 years (7.5 to 30.6) from presentation, median overall survival from diagnosis 27.0 years (21.3 to 32.7) and median renal allograft survival 22.0 years (13.0 to 31.0) across the 18 transplant recipients and 16.0 years (3.1 to 28.9) in the 11 who received an isolated kidney transplant [5]. The European position statement recorded no survival data for this entity, estimating it as probably above 120 months [28], and the authors of a 2026 case cite a median survival of more than 172 months, a prognosis they describe as quite favourable compared with light-chain amyloidosis [8].

Lysozyme amyloidosis has indolent systemic progression with early symptom onset and an added risk, catastrophic when it materialises, of hepatic or splenic rupture: in the largest published cohort, with 35 patients, the median age at symptom onset was 23.7 years, 43% reported symptoms in childhood, hepatic involvement reached 80% and renal 57%, five patients had spontaneous hepatic rupture or haemorrhage, four of them requiring emergency liver transplantation, eight developed renal failure and six received a kidney transplant, without amyloid recurrence [15]. Transplantation is successful both for hepatic rupture and for end-stage kidney disease: in the 2026 cohort, the six renal allografts had a median survival of 6.6 years (0.8 to 15.0), with a single failure from chronic rejection and no amyloid recurrence over a median follow-up of 10.4 years, and in the earlier series from the same centre the four liver grafts were functioning at 1.7, 5.8, 9.0 and 11.0 years, and the three renal allografts had no dysfunction at 6.6, 1.8 and 0.8 years [15, 16]. Mortality is now quantified in a series: in the cohort of 35 patients, 10 (29%) had died by the census date of November 2025, with a median time from diagnosis to death of 18.0 years (95% confidence interval 8.6 to not reached) and a median survival from birth of 71.5 years (62.0 to not reached) [15]. Uncontrollable hepatic haemorrhage caused the death of two consecutive members of one family, and in one of the families in the British cohort four deaths from hepatic rupture or haemorrhage were recorded [15, 34]. Reproductive outcomes in this cohort were preserved, with 32 children of 15 affected women and four affected men, without obstetric haemorrhage or uterine rupture [15]. In the historical description of the Ostertag type, visceral deposition was usually fatal by the fifth decade [35], a characterisation that predates the availability of dialysis, transplantation and typing and should not be transposed to current practice.

Gelsolin amyloidosis is the entity with the best vital prognosis. In the Finnish national registry, relative survival exceeded 1 for men and women up to the 70 to 74 years age group, in which it was 0.96, that is, the disease does not shorten life expectancy at least in the first 75 years, which the authors attribute in part to the rarity of severe renal and cardiac manifestations compared with other systemic amyloidoses [18]. The functional prognosis is, by contrast, significantly affected: the disease causes a broad spectrum of ophthalmological, neurological, cutaneous and oral symptoms which, with repeated surgery, create a clinically significant disease burden [18]. The European position statement recorded cardiac involvement in about 5% of patients, mainly conduction disease, with near-normal life expectancy [28]. Homozygotes are an exception, with early-onset and much more severe, predominantly renal disease [20, 21].

In apolipoprotein A-II amyloidosis, the cause of death is renal failure, and dialysis and kidney transplantation are the two therapeutic options; transplantation was considered effective in a patient who maintained stable graft function for nine years, because recurrence in the graft and progression in other organs appear very slow [12]. The European statement recorded no survival data for this entity [28].

Variant β2-microglobulin amyloidosis has a slowly progressive course in the index family, with one patient dying at 70 years with extensive deposits at autopsy, and a dominant cardiac course in the Portuguese family, with two of the five affected members undergoing organ transplantation [23]. There are no published survival data.

The most important cross-cutting observation is that, in the only entity with comparable cohorts in two periods, renal and allograft survival improved significantly but overall survival did not change. The authors attribute this dissociation to the competing risks of kidney disease itself and of cardiovascular disease and to the absence of therapies that modify the amyloidogenic process, stressing that interpretation of the mortality data is limited by the fact that the database does not record cause of death [3].

11.2 Surveillance table

The intervals indicated come from two sources with different scopes: 12-monthly follow-up is the practice at the national centre that brought together the reference cohorts of these entities [3], and the 6-monthly schedule for the visit with electrocardiogram and tests and 12-monthly schedule for echocardiography and 24-hour recording, as well as the carrier surveillance scheme, come from the European position statement, formulated for cardiac amyloidosis and transthyretin amyloidosis and applied here by extension [28]. The frequency of monitoring of renal function and albuminuria follows, in the nephrology guideline, the risk grid defined by glomerular filtration and albuminuria category, so that, in patients with A3 albuminuria and reduced glomerular filtration, monitoring more frequent than annual is justified [69].

11.3 Targets and action thresholds

The thresholds that trigger therapeutic decisions in these diseases are almost entirely generic nephrological and cardiological thresholds. In the kidney, a urinary albumin/creatinine ratio of 200 mg/g or above with estimated glomerular filtration rate of 20 ml/min per 1.73 m² or above indicates a sodium–glucose cotransporter 2 inhibitor with recommendation grade 1A, and A2 or A3 albuminuria indicates a renin–angiotensin system inhibitor, grade 2C and 1B respectively; the systolic blood pressure target is below 120 mmHg when tolerated, grade 2B; and an estimated glomerular filtration rate below 60 ml/min per 1.73 m² in patients aged 50 years or over indicates a statin, grade 1A [69]. The deterioration that prompts reassessment is a creatinine rise above 30% in the four weeks after starting or increasing the dose of a renin–angiotensin system inhibitor, and a fall in glomerular filtration above 30% after starting haemodynamically active therapy exceeds the expected variability and warrants investigation [69]. In the heart, an estimated extracellular volume above 40% on magnetic resonance imaging is highly predictive of cardiac amyloidosis and constitutes the most specific imaging threshold available [32].

The threshold for considering transplantation is the approach of end-stage kidney disease, and the decision between isolated kidney transplantation and combined liver–kidney transplantation should weigh the variant, age and general condition: in fibrinogen amyloidosis, isolated kidney transplantation is probably appropriate for most patients, with a median allograft survival of 15.5 years in transplants performed after 2009, combined transplantation remaining a valid option for younger patients in better condition and, in light of the French series, for carriers of variants other than E526V [3, 4]. Transplant assessment is also a moment of exclusion: in the King's College series, eight of 22 patients assessed were declined because of poor ventricular function or uncorrectable cardiovascular or systemic atheromatous disease [57]. In lysozyme amyloidosis, the threshold for pre-emptive liver transplantation is the presence of extensive hepatic amyloid together with a strong family history of hepatic rupture [16].

11.4 Surveillance of asymptomatic carriers

Carrier surveillance aims to detect the onset of organ involvement, in order to start nephroprotection and plan replacement of function, and not to administer preventive therapy, which does not exist. The starting rule is to begin about ten years before the typical age at onset in affected relatives, or earlier if symptoms appear [28]. The scheme applicable in the predominantly renal entities is simple: blood pressure, serum creatinine with estimated glomerular filtration rate and quantification of albuminuria, annually [3, 69]. In the entities with cardiac involvement, annual electrocardiogram, tests and echocardiogram are added, 24-hour electrocardiographic recording every two years and scintigraphy or cardiac magnetic resonance imaging every three years, or earlier if any test is abnormal [28]. In gelsolin amyloidosis, annual ophthalmological assessment with a slit lamp is the earliest detection tool, given that corneal lattice dystrophy appears at a mean age of 43 years and is practically a prerequisite for the diagnosis [19, 20]. In lysozyme amyloidosis, surveillance should include assessment of gastrointestinal symptoms and liver function, given the risk of spontaneous hepatic rupture, and endoscopy has a high yield when there are symptoms [16]. The safety of endoscopy in this entity is documented and matters, because these patients will undergo repeated examinations throughout life: of the 21 patients (60%) in the 2026 cohort who underwent endoscopy, three had contact bleeding, but no serious iatrogenic complication was recorded in experienced hands [15]. Patients with significant hepatosplenomegaly require enhanced monitoring and careful procedural planning, because rupture has occurred in amyloid-laden organs in the absence of portal hypertension and hepatic synthetic failure, and has been precipitated by infection and major surgery [15]. In this entity, the rule of starting ten years before the familial age at onset is insufficient: median symptom onset is at 23.7 years and 43% of patients report symptoms in childhood, so surveillance of at-risk relatives should begin in adolescence, looking for sicca syndrome, gastrointestinal complaints and proteinuria [15].

The limitation of this surveillance is its evidence base: none of the schemes has been validated in these entities, the 12-month interval is the one practised at the centres that brought together the cohorts, and starting ten years before the familial age at onset is a rule formulated for transthyretin amyloidosis. Intrafamilial variability in age at onset, which in fibrinogen amyloidosis reaches four decades with the same variant, limits the precision of this rule [4].

12. Summary and Final Considerations

The hereditary systemic amyloidoses constitute a class defined by a common mechanism — extracellular deposition of fibrils formed by a plasma protein that a germline variant makes conformationally unstable — and differentiated by the identity of that protein, which determines the organs affected, the natural history and the treatment. The current nomenclature, established by the Nomenclature Committee of the International Society of Amyloidosis and updated in 2024, counts 42 human amyloid fibril proteins, 14 only systemic, 23 only localised and 5 with both forms of deposition, giving a total of 19 associated with systemic deposition, and designates each type by the letter A followed by the abbreviation of the precursor protein. The OMIM phenotypic series recognises six hereditary systemic amyloidoses, all autosomal dominant: the one mediated by transthyretin and the five mediated by the fibrinogen A alpha chain, apolipoprotein A-I, gelsolin, lysozyme and β2-microglobulin. Apolipoprotein A-II amyloidosis completes the group of non-transthyretin hereditary forms, documented in independent families and with its own Orphanet code, but without a phenotype MIM number.

The most frequent entity in this group is fibrinogen Aα amyloidosis, the most common hereditary form of renal amyloidosis in Europe, caused by variants in the carboxy-terminal portion of the alpha chain encoded by FGA exon 5, with one dominant variant — p.Glu545Val, the historical E526V — responsible for 86 to 90% of cases. The variant protein is entirely of hepatic origin, the wild-type protein does not form fibrils, and haemostasis remains normal, so the disease presents as proteinuric nephropathy with hypertension and not as a coagulopathy. The contemporary cohort of 194 patients from a UK national centre established the current profile: diagnosis at a median of 60 years, often at an advanced stage of chronic kidney disease, progression to end-stage kidney disease in 62.8% of patients, a mean loss of 5.3 ml/min per 1.73 m² per year of glomerular filtration, splenic amyloid in 80.7% without clinical expression and no amyloid cardiomyopathy demonstrable by magnetic resonance imaging. Median survival from diagnosis is about 12 years, and the improvement in renal and allograft survival between periods did not translate into improved overall survival, a dissociation the authors attribute to competing risks and to the absence of therapy that modifies the amyloidogenic process.

The other entities fall into distinct phenotypes. Apolipoprotein A-I amyloidosis is the most heterogeneous, with 14 variants documented in a single series of 57 patients and a spectrum ranging from slowly progressive nephropathy, 15 years to end-stage kidney disease, to fatal liver disease, restrictive cardiomyopathy, cutaneous and laryngeal amyloidosis with hoarseness and infertility due to testicular deposition, with a family history often absent. Lysozyme amyloidosis is a disease of the gastrointestinal tract, liver and kidneys, with high clinical penetrance and symptom onset much earlier than the diagnosis suggests — median 23.7 years, with childhood symptoms in 43% of patients in the largest published cohort —, presenting most often with sicca syndrome, with hepatic involvement in 80% and renal in 57%, amyloid lesions visible at endoscopy and the characteristic risk of spontaneous hepatic rupture, with the liver friable through loss of the reticulin framework. Gelsolin amyloidosis, one of the diseases of the Finnish disease heritage, is defined by the triad of corneal lattice dystrophy at a mean age of 39 years, progressive facial palsy and cutis laxa, with renal involvement in 5% and cardiac in 15%, and is the only one in which relative survival does not differ from the general population up to 75 years, at the cost of significant morbidity and repeated surgery. Apolipoprotein A-II amyloidosis results exclusively from stop codon variants, which extend translation by 21 residues, and manifests as nephropathy in the third to fifth decade. Variant β2-microglobulin amyloidosis is distinguished from the acquired dialysis form by occurring with normal renal function and normal or low serum concentration of the protein, and has two phenotypes described, one gastrointestinal and dysautonomic in the French index family and another cardiac in a Portuguese family.

Diagnostic reasoning has three steps that no shortcut replaces: demonstrating amyloid in tissue with Congo red, identifying the fibril protein and identifying the genetic variant. The non-invasive criteria, based on bone scintigraphy with negative monoclonal protein screening, are valid only for transthyretin amyloidosis, so in these entities biopsy is mandatory, and scintigraphy has documented false positives in several of them. Typing by laser microdissection and tandem mass spectrometry, with 100% specificity and sensitivity in the original training set and 98% in validation, is the reference standard, and its addition to immunohistochemistry raised diagnostic accuracy from 76% to 94% in a blinded comparison of 142 biopsies; in the subgroup of samples with equivocal or negative immunostaining, false positives of anti-light-chain and anti-transthyretin antibodies reach 41.5% and 56.8%. Proteomics identifies the protein, not the gene, so sequencing remains indispensable, and genetic testing for these forms is available in few laboratories, and is not performed in most cases diagnosed by mass spectrometry at a reference centre. The consequence of not typing is quantified: 9.7% of 350 patients with a clinical diagnosis of light-chain amyloidosis had amyloidogenic variants, 18 of them in FGA, and a quarter of these hereditary patients had a low-grade monoclonal gammopathy that reinforced the wrong label.

Treatment is organised on three levels, none of which acts on amyloidogenesis. Support of the affected organ rests, in the renal forms, on the recommendations of the 2024 nephrology guideline for proteinuric chronic kidney disease — renin–angiotensin system inhibitor at the maximum tolerated dose in A2 and A3 albuminuria, sodium–glucose cotransporter 2 inhibitor with an albumin/creatinine ratio of 200 mg/g or above and estimated glomerular filtration rate of 20 ml/min per 1.73 m² or above, systolic pressure target below 120 mmHg when tolerated, sodium restriction and statin —, whose application to these entities is a declared extrapolation. Removal of the hepatic source of the variant protein is the only targeted therapy for which there is proof of effect: liver transplantation was followed by regression of amyloid load on serial scintigraphy in the four patients with apolipoprotein A-I amyloidosis so assessed. Replacement of the failing organ now has much better results than the reference literature of the 2000s suggested: in fibrinogen amyloidosis, median isolated renal allograft survival rose from 7.7 to 15.5 years between periods, which led the centre that proposed combined liver–kidney transplantation as a curative strategy to reconsider the indication, favouring isolated kidney transplantation for most patients and reserving combined transplantation for the youngest and those in better condition, with the French series additionally suggesting combined transplantation in carriers of variants other than E526V, because of more frequent recurrence in the graft. It should be stated clearly what does not apply: none of the drugs approved for amyloidosis — the transthyretin stabilisers and silencers, and the anti-CD38 antibody indicated in light-chain amyloidosis — has an indication in any of these entities, and chemotherapy has no role at all in the hereditary forms.

The landscape of therapeutic development is, to date, empty for these diseases. The search of the public trials registry identified no clinical trial whose condition studied is one of these entities; the only pan-amyloid clearance programme with an active study has its phase 2 extension restricted to light-chain amyloidosis; and the strategy of amyloid P component depletion with a monoclonal antibody, which showed dose-dependent hepatic amyloid clearance in patients with systemic amyloidosis without type restriction, has been discontinued. There is a single orphan designation naming one of these entities, that of doxycycline for systemic amyloidosis caused by β2-microglobulin, granted in 2012 and without a controlled trial to support it, and a single targeted preclinical programme, the antibody fragments that block fibrillogenesis of variant gelsolin, with data obtained exclusively in vitro. The prospects set out by the authors of the largest fibrinogen amyloidosis cohort — clearance of deposits, binding to circulating variant fibrinogen and suppression of production by allele-specific silencing or gene editing directed at the mutant allele — are therapeutic concepts without an identified clinical programme. As long as this remains the case, the clinical gain possible in these diseases depends on three things within reach of practice: correct typing of the deposit, which avoids useless treatments and opens the door to transplantation; early access to nephrology, with optimised nephroprotection, whose association with improved renal survival between periods is the most encouraging observation in the recent literature; and cascade family screening, which identifies carriers before organ involvement and allows targeted surveillance from about ten years before the age at onset in the family.

13. References

1. Buglioni A, Dasari S, Bu L, Sethi S, Moyer AM, Baudhuin LM, Hasadsri L, Leung N, McPhail ED, Nasr SH. Feasibility of DNA-based diagnosis of fibrinogen A alpha-chain amyloidosis from paraffin-embedded kidney tissue. Kidney Int Rep. 2025;10(10):3670-3. doi:10.1016/j.ekir.2025.08.006
2. Orphanet. AFib amyloidosis, ORPHA:93562 (clinical subtype). https://www.orpha.net/en/disease/detail/93562
3. Carruthers JE, Rauf MU, Goodman H, Petrie A, Rowczenio D, Taylor G, Canetti D, Fernandes R, Gilbertson JA, Martinez Naharro A, Lachmann HJ, Wechalekar AD, Fontana M, Hawkins PN, Gillmore JD. Renal outcomes in hereditary fibrinogen A alpha-chain amyloidosis. Kidney Int Rep. 2026;11(11):107076. doi:10.1016/j.ekir.2026.107076
4. Meyer L, Ulrich M, Ducloux D, Garrigue V, Vigneau C, Nochy D, et al. Organ transplantation in hereditary fibrinogen A α-chain amyloidosis: a case series of French patients. Am J Kidney Dis. 2020;76(3):384–91. doi:10.1053/j.ajkd.2020.02.445
5. Cohen OC, Blakeney IJ, Law S, Ravichandran S, Gilbertson J, Rowczenio D, et al. The experience of hereditary apolipoprotein A-I amyloidosis at the UK National Amyloidosis Centre. Amyloid. 2022;29(4):237–44. doi:10.1080/13506129.2022.2070741
6. Online Mendelian Inheritance in Man. #620657 Amyloidosis, hereditary systemic 3; AMYLD3. https://www.omim.org/entry/620657
7. Rowczenio D, Dogan A, Theis JD, Vrana JA, Lachmann HJ, Wechalekar AD, et al. Amyloidogenicity and clinical phenotype associated with five novel mutations in apolipoprotein A-I. Am J Pathol. 2011;179(4):1978–87. doi:10.1016/j.ajpath.2011.06.024
8. Voran JC, Radić Ž, Kilian LS, Braun F, Röcken C, Frank D, Seoudy H. Rare hereditary cardiac amyloidosis treated with SGLT2 inhibitor: beyond ATTR/AL. JACC Case Rep. 2026;31(38):109227. doi:10.1016/j.jaccas.2026.109227
9. Online Mendelian Inheritance in Man. *107670 Apolipoprotein A-II; APOA2. https://www.omim.org/entry/107670
10. GeneGuidance, gene record for APOA2.
11. Yazaki M, Liepnieks JJ, Barats MS, Cohen AH, Benson MD. Hereditary systemic amyloidosis associated with a new apolipoprotein AII stop codon mutation Stop78Arg. Kidney Int. 2003;64(1):11–6. doi:10.1046/j.1523-1755.2003.00047.x
12. Magy N, Liepnieks JJ, Yazaki M, Kluve-Beckerman B, Benson MD. Renal transplantation for apolipoprotein AII amyloidosis. Amyloid. 2003;10(4):224–8. doi:10.3109/13506120309041739
13. Orphanet. ALys amyloidosis, ORPHA:93561 (clinical subtype). https://www.orpha.net/en/disease/detail/93561
14. Anker S, Hinderhofer K, Baur J, Haupt C, Röcken C, Beimler J, Zeier M, Weiler M, Wühl E, Kimmich C, Schönland S, Hegenbart U. Lysozyme amyloidosis — a report on a large German cohort and the characterisation of a novel amyloidogenic lysozyme gene variant. Amyloid. 2022;29(4):245–54. doi:10.1080/13506129.2022.2072198
15. Carruthers JE, Gilbertson JA, Rowczenio D, Hutt DF, Cohen OC, Mahmood S, Martinez-Naharro A, Ravichandran S, Whelan C, Wechalekar AD, Hawkins PN, Fontana M, Gillmore JD. Childhood onset, indolent progression, and risk of hepatic rupture: the natural history of hereditary lysozyme amyloidosis. Amyloid. 2026:1-9. doi:10.1080/13506129.2026.2735580
16. Sattianayagam PT, Gibbs SD, Rowczenio D, Pinney JH, Wechalekar AD, Gilbertson JA, Hawkins PN, Lachmann HJ, Gillmore JD. Hereditary lysozyme amyloidosis — phenotypic heterogeneity and the role of solid organ transplantation. J Intern Med. 2012;272(1):36–44. doi:10.1111/j.1365-2796.2011.02470.x
17. Orphanet. AGel amyloidosis, ORPHA:85448. https://www.orpha.net/en/disease/detail/85448
18. Schmidt EK, Mustonen T, Kiuru-Enari S, Kivelä TT, Atula S. Finnish gelsolin amyloidosis causes significant disease burden but does not affect survival: FIN-GAR phase II study. Orphanet J Rare Dis. 2020;15(1):19. doi:10.1186/s13023-020-1300-5
19. Nikoskinen T, Schmidt EK, Strbian D, Kiuru-Enari S, Atula S. Natural course of Finnish gelsolin amyloidosis. Ann Med. 2015;47(6):506–11. doi:10.3109/07853890.2015.1075063
20. Online Mendelian Inheritance in Man. #105120 Amyloidosis, Finnish type (AMYLD4). https://www.omim.org/entry/105120
21. GeneGuidance, gene record for GSN.
22. Orphanet. Variant ABeta2M amyloidosis, ORPHA:314652. https://www.orpha.net/en/disease/detail/314652
23. Online Mendelian Inheritance in Man. #620659 Amyloidosis, hereditary systemic 6; AMYLD6. https://www.omim.org/entry/620659
24. Valleix S, Gillmore JD, Bridoux F, Mangione PP, Dogan A, Nedelec B, et al. Hereditary systemic amyloidosis due to Asp76Asn variant β2-microglobulin. N Engl J Med. 2012;366(24):2276–83. doi:10.1056/NEJMoa1201356
25. Prokaeva T, Joshi T, Klimtchuk ES, Gibson VM, Spencer B, Siddiqi O, et al. A novel substitution of proline (P32L) destabilises β2-microglobulin inducing hereditary systemic amyloidosis. Amyloid. 2022;29(4):255–62. doi:10.1080/13506129.2022.2072199
26. Orphanet. Hereditary ATTR amyloidosis, ORPHA:271861. https://www.orpha.net/en/disease/detail/271861
27. European Reference Network for Rare Kidney Diseases (ERKNet). https://www.erknet.org/
28. Garcia-Pavia P, Rapezzi C, Adler Y, Arad M, Basso C, Brucato A, et al. Diagnosis and treatment of cardiac amyloidosis: a position statement of the ESC Working Group on Myocardial and Pericardial Diseases. Eur Heart J. 2021;42(16):1554-68. doi:10.1093/eurheartj/ehab072
29. Online Mendelian Inheritance in Man. #105200 Amyloidosis, hereditary systemic 2; AMYLD2. Johns Hopkins University. Available from: https://www.omim.org/entry/105200
30. Muchtar E, Dispenzieri A, Magen H, Grogan M, Mauermann M, McPhail ED, et al. Systemic amyloidosis from A (AA) to T (ATTR): a review. J Intern Med. 2021;289(3):268–92. doi:10.1111/joim.13169
31. Buxbaum JN, Eisenberg DS, Fändrich M, McPhail ED, Merlini G, Saraiva MJM, Sekijima Y, Westermark P. Amyloid nomenclature 2024: update, novel proteins, and recommendations by the International Society of Amyloidosis (ISA) Nomenclature Committee. Amyloid. 2024;31(4):249-56. doi:10.1080/13506129.2024.2405948
32. Wong LSM, Judge DP. Cardiac Amyloidosis. In: Baars HF, van der Smagt JJ, Doevendans PAFM, van Tintelen JP, editors. Clinical Cardiogenetics. 3rd ed. Cham: Springer; 2020. p. 167–177.
33. Buxbaum JN, Dispenzieri A, Eisenberg DS, Fändrich M, Merlini G, Saraiva MJM, Sekijima Y, Westermark P. Amyloid nomenclature 2022: update, novel proteins, and recommendations by the International Society of Amyloidosis (ISA) Nomenclature Committee. Amyloid. 2022;29(4):213–9. doi:10.1080/13506129.2022.2147636
34. Online Mendelian Inheritance in Man. #620658 Amyloidosis, hereditary systemic 5; AMYLD5. https://www.omim.org/entry/620658
35. Pepys MB, Hawkins PN, Booth DR, Vigushin DM, Tennent GA, Soutar AK, et al. Human lysozyme gene mutations cause hereditary systemic amyloidosis. Nature. 1993;362(6420):553-7. doi:10.1038/362553a0
36. Online Mendelian Inheritance in Man. *134820 Fibrinogen alpha chain; FGA. Johns Hopkins University. https://www.omim.org/entry/134820
37. Benson MD, Liepnieks JJ, Yazaki M, Yamashita T, Hamidi Asl K, Guenther B, Kluve-Beckerman B. A new human hereditary amyloidosis: the result of a stop-codon mutation in the apolipoprotein AII gene. Genomics. 2001;72(3):272–7. doi:10.1006/geno.2000.6499
38. GeneGuidance, gene record for B2M.
39. Hu A, Chowdhury RB, Bianchi G. Pathogenesis and kidney prognosis of renal amyloidosis. Cell Mol Life Sci. 2026;83(1):213. doi:10.1007/s00018-025-06030-y
40. Lachmann HJ, Booth DR, Booth SE, Bybee A, Gilbertson JA, Gillmore JD, Pepys MB, Hawkins PN. Misdiagnosis of hereditary amyloidosis as AL (primary) amyloidosis. N Engl J Med. 2002;346(23):1786–91. doi:10.1056/NEJMoa013354
41. Online Mendelian Inheritance in Man. Phenotypic Series PS105210, Amyloidosis, hereditary systemic. https://www.omim.org/phenotypicSeries/PS105210
42. Orphanet. Hereditary amyloidosis with primary renal involvement, ORPHA:85450. https://www.orpha.net/en/disease/detail/85450
43. Orphanet. AApoAI amyloidosis, ORPHA:93560 (clinical subtype). https://www.orpha.net/en/disease/detail/93560
44. GeneGuidance, gene record for LYZ.
45. GeneGuidance, gene record for FGA.
46. Koppelaar AJ, Zwetsloot PP, von der Thüsen JH, Lam KH, Gilbertson JA, Gillmore JD, Michels M, Hirsch A. Apolipoprotein A-IV cardiac amyloidosis as a diagnostic pitfall: beyond the noninvasive algorithm. JACC Case Rep. 2026;31(38):108669. doi:10.1016/j.jaccas.2026.108669
47. ClinVar (National Center for Biotechnology Information). Records of the amyloidogenic variants of FGA, APOA1, APOA2, LYZ, GSN and B2M. https://www.ncbi.nlm.nih.gov/clinvar/
48. Xiang S, Bi P, Chen D, Tang W, Li Y, Zhou Y, Kang D, Pan C, Wan F, Yu J, Tang X. Gelsolin amyloidosis presenting with nephrotic syndrome: a case report and molecular insights. Front Med (Lausanne). 2026;13:1798985. doi:10.3389/fmed.2026.1798985
49. ClinVar (NCBI), records of NM_000239.3:c.244T>C and c.244T>A, p.(Trp82Arg) in the LYZ gene. https://www.ncbi.nlm.nih.gov/clinvar/
50. VariantValidator, validation of NM_000508.5:c.1634A>T and NM_000508.5:c.1718G>T on GRCh38.
51. Orphanet. Congenital fibrinogen deficiency, ORPHA:335. https://www.orpha.net/en/disease/detail/335
52. GeneGuidance, gene record for APOA1.
53. Hovingh GK, et al. Genetic Disorders of Lipoprotein Metabolism: Diagnosis and Management. In: Baars HF, van der Smagt JJ, Doevendans PAFM, van Tintelen JP, editors. Clinical Cardiogenetics. 3rd ed. Cham: Springer; 2020. p. 398 (section 23.14, Apolipoprotein AI Deficiency).
54. VariantValidator, validation of NM_000039.3:c.605T>C (APOA1), NM_004048.4:c.286G>A (B2M) and NM_000177.5:c.640G>A (GSN) on GRCh38.
55. Orphanet. Apolipoprotein A-I deficiency, ORPHA:425. https://www.orpha.net/en/disease/detail/425
56. Gillmore JD, Lachmann HJ, Rowczenio D, Gilbertson JA, Zeng CH, Liu ZH, Li LS, Wechalekar A, Hawkins PN. Diagnosis, pathogenesis, treatment, and prognosis of hereditary fibrinogen A alpha-chain amyloidosis. J Am Soc Nephrol. 2009;20(2):444-51. doi:10.1681/ASN.2008060614
57. Stangou AJ, Banner NR, Hendry BM, Rela M, Portmann B, Wendon J, et al. Hereditary fibrinogen A alpha-chain amyloidosis: phenotypic characterization of a systemic disease and the role of liver transplantation. Blood. 2010;115(15):2998–3007. doi:10.1182/blood-2009-06-223792
58. Gregorini G, Izzi C, Obici L, Tardanico R, Röcken C, Viola BF, Capistrano M, Donadei S, Biasi L, Scalvini T, Merlini G, et al. Renal apolipoprotein A-I amyloidosis: a rare and usually ignored cause of hereditary tubulointerstitial nephritis. J Am Soc Nephrol. 2005;16(12):3680-6. doi:10.1681/ASN.2005040382
59. Traynor CA, Tighe D, O'Brien FJ, Leavey SF, Dorman AM, Denton MD, Magee C, Conlon PJ. Clinical and pathologic characteristics of hereditary apolipoprotein A-I amyloidosis in Ireland. Nephrology (Carlton). 2013;18(8):549-54. doi:10.1111/nep.12108
60. Horike K, Takeda A, Tsujita M, Goto N, Watarai Y, Uchida K, Katayama A, Nishihira M, Shimizu A, Nozu K, Morozumi K. Two novel APOA1 gene mutations in a Japanese renal transplant recipient with recurrent apolipoprotein A-I related amyloidosis. Nephrology (Carlton). 2018;23 Suppl 2:17-21. doi:10.1111/nep.13278
61. Haslett JJ, Patel JK, Kittleson MM. Beta 2-microglobulin: case report of a rare cause of cardiac amyloidosis. Eur Heart J Case Rep. 2023;7(5):ytad239. doi:10.1093/ehjcr/ytad239
62. Kittleson MM, Ruberg FL, Ambardekar AV, Brannagan TH, Cheng RK, Clarke JO, et al. 2023 ACC Expert Consensus Decision Pathway on Comprehensive Multidisciplinary Care for the Patient With Cardiac Amyloidosis. J Am Coll Cardiol. 2023;81(11):1076–126. doi:10.1016/j.jacc.2022.11.022
63. Lopes da Silva M, Pais T, Boavida J, Branco C, Pinheiro P, Justino G, Lopes JA, Rodrigues N. Diagnostic discordance in amyloidosis: lessons from proteomic amyloid typing. Clin Kidney J. 2026;19(9):sfag297. doi:10.1093/ckj/sfag297
64. Vrana JA, Gamez JD, Madden BJ, Theis JD, Bergen HR 3rd, Dogan A. Classification of amyloidosis by laser microdissection and mass spectrometry-based proteomic analysis in clinical biopsy specimens. Blood. 2009;114(24):4957–9. doi:10.1182/blood-2009-07-230722
65. Gilbertson JA, Theis JD, Vrana JA, Lachmann H, Wechalekar A, Whelan C, Hawkins PN, Dogan A, Gillmore JD. A comparison of immunohistochemistry and mass spectrometry for determining the amyloid fibril protein from formalin-fixed biopsy tissue. J Clin Pathol. 2015;68(4):314-7. doi:10.1136/jclinpath-2014-202722
66. Colombat M, Gaspard M, Camus M, Dalloux-Chioccioli J, Delas A, Poullot E, Moktefi A, François A, Moreau A, Gibier JB, Raynaud P, Huart A, Piedrafita A, et al. Mass spectrometry-based proteomics in clinical practice amyloid typing: state-of-the-art from a French nationwide cohort. Haematologica. 2022;107(12):2983-7. doi:10.3324/haematol.2022.281431
67. Orphanet. Wild type ABeta2M amyloidosis, ORPHA:85446. https://www.orpha.net/en/disease/detail/85446
68. Orphanet. AApoAIV amyloidosis, ORPHA:439232. https://www.orpha.net/en/disease/detail/439232
69. Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease. Kidney Int. 2024;105(4S):S117–314.
70. European Medicines Agency. Forxiga (dapagliflozin) — Summary of Product Characteristics (EPAR product information). https://www.ema.europa.eu/en/documents/product-information/forxiga-epar-product-information_pt.pdf
71. U.S. Food and Drug Administration. FARXIGA (dapagliflozin) tablets — prescribing information. NDA 202293.
72. Shaz BH, Lewis WD, Skinner M, Khettry U. Livers from patients with apolipoprotein A-I amyloidosis are not suitable as "domino" donors. Mod Pathol. 2001;14(6):577-80. doi:10.1038/modpathol.3880353
73. Orphanet. Orphan drugs: search by disease (ORPHAcode). Version 1.6.11, updated 2026-10-21. https://www.orpha.net/en/drug
74. European Medicines Agency. EU/3/12/961 — orphan designation for the treatment of systemic amyloidosis caused by beta-2 microglobulin (doxycycline hyclate). Designation of 5 March 2012. https://www.ema.europa.eu/en/medicines/human/orphan-designations/eu-3-12-961
75. ClinicalTrials.gov. NCT05521022 — Study of AT-02 in Healthy Volunteers and Subjects With Systemic Amyloidosis.
76. ClinicalTrials.gov. NCT05951049 — A Phase 2, Open-label Extension Study to Evaluate the Long-term Safety and Tolerability of AT-02 (Attralus).
77. Richards DB, Cookson LM, Berges AC, Barton SV, Lane T, Ritter JM, Fontana M, Moon JC, Pinzani M, Gillmore JD, Hawkins PN, Pepys MB. Therapeutic clearance of amyloid by antibodies to serum amyloid P component. N Engl J Med. 2015;373(12):1106-14. doi:10.1056/NEJMoa1504942
78. Richards DB, Cookson LM, Barton SV, Liefaard L, Lane T, Hutt DF, Ritter JM, Fontana M, Moon JC, Gillmore JD, Wechalekar A, Hawkins PN, Pepys MB. Repeat doses of antibody to serum amyloid P component clear amyloid deposits in patients with systemic amyloidosis. Sci Transl Med. 2018;10(422):eaan3128. doi:10.1126/scitranslmed.aan3128
79. Multiple treatment session study to assess GSK2398852 administered following and along with GSK2315698. ClinicalTrials.gov NCT03044353. GlaxoSmithKline. https://clinicaltrials.gov/study/NCT03044353
80. A study to evaluate the safety of GSK2398852 when co-administered with GSK2315698 in patients with systemic amyloidosis. ClinicalTrials.gov NCT01777243. GlaxoSmithKline. https://clinicaltrials.gov/study/NCT01777243
81. Bodin K, Ellmerich S, Kahan MC, Tennent GA, Loesch A, Gilbertson JA, et al. Antibodies to human serum amyloid P component eliminate visceral amyloid deposits. Nature. 2010;468(7320):93-7. doi:10.1038/nature09494
82. Safety and effect of doxycycline in patients with amyloidosis. ClinicalTrials.gov NCT01677286. Boston University. https://clinicaltrials.gov/study/NCT01677286
83. A study to characterize the biodistribution of 124I-labeled AT-03 in patients with systemic amyloidosis. ClinicalTrials.gov NCT05201911. Attralus, Inc. https://clinicaltrials.gov/study/NCT05201911
84. 124I-p5+14 injection safety in subjects with systemic amyloidosis. ClinicalTrials.gov NCT03678259. University of Tennessee Graduate School of Medicine. https://clinicaltrials.gov/study/NCT03678259
85. Leimu L, Holm P, Gąciarz A, Haavisto O, Prince S, Pesonen U, Huovinen T, Lamminmäki U. Epitope-specific antibody fragments block aggregation of AGelD187N, an aberrant peptide in gelsolin amyloidosis. J Biol Chem. 2024;300(8):107507. doi:10.1016/j.jbc.2024.107507
86. Bollati M, Natale C, Girois L, Conz A, Peqini K, Pieraccini S, Pellegrino S, Diomede L, et al. A molecular perspective of gelsolin amyloidosis: an old foe with new faces. Cell Mol Life Sci. 2026;83:177. doi:10.1007/s00018-026-06172-7
87. Scarpioni R, Ricardi M, Albertazzi V, De Amicis S, Rastelli F, Zerbini L. Dialysis-related amyloidosis: challenges and solutions. Int J Nephrol Renovasc Dis. 2016;9:319-28. doi:10.2147/IJNRD.S84784
88. Hoshino J, Yamagata K, Nishi S, Nakai S, Masakane I, Iseki K, Tsubakihara Y. Significance of the decreased risk of dialysis-related amyloidosis now proven by results from Japanese nationwide surveys in 1998 and 2010. Nephrol Dial Transplant. 2016;31(4):595-602. doi:10.1093/ndt/gfv276
89. Abe T, Uchita K, Orita H, Kamimura M, Oda M, Hasegawa H, Kobata H, Fukunishi M, Shimazaki M, Abe T, Akizawa T, Ahmad S. Effect of beta(2)-microglobulin adsorption column on dialysis-related amyloidosis. Kidney Int. 2003;64(4):1522-8. doi:10.1046/j.1523-1755.2003.00235.x
90. Post approval study of Lixelle for the treatment of dialysis-related amyloidosis. ClinicalTrials.gov NCT02952144. Kaneka Medical America LLC. https://clinicaltrials.gov/study/NCT02952144
91. Comparison of hemodialysis with medium cut-off dialyzer (Theranova) and high flux dialyzer. ClinicalTrials.gov NCT03211676. Poitiers University Hospital. https://clinicaltrials.gov/study/NCT03211676
92. A randomised, open-label, cross over group, single-centre controlled study (Theranova 400 vs Elisio-19H). ClinicalTrials.gov NCT04160351. Selayang Hospital. https://clinicaltrials.gov/study/NCT04160351
93. Piccoli GB, Hachemi M, Molfino I, Coindre JP, Boursot C. Doxycycline treatment in dialysis related amyloidosis: discrepancy between antalgic effect and inflammation, studied with FDG-positron emission tomography: a case report. BMC Nephrol. 2017;18(1):285. doi:10.1186/s12882-017-0698-z
94. Orphanet. AApoAII amyloidosis, ORPHA:238269 (clinical subtype). https://www.orpha.net/en/disease/detail/238269
95. Dasari S, Theis JD, Vrana JA, Rech KL, Dao LN, Howard MT, Dispenzieri A, Gertz MA, Hasadsri L, Highsmith WE, Kurtin PJ, McPhail ED. Amyloid typing by mass spectrometry in clinical practice: a comprehensive review of 16,175 samples. Mayo Clin Proc. 2020;95(9):1852-64. doi:10.1016/j.mayocp.2020.06.029
96. Sethi S, Vrana JA, Theis JD, Leung N, Sethi A, Nasr SH, Fervenza FC, Cornell LD, Fidler ME, Dogan A. Laser microdissection and mass spectrometry-based proteomics aids the diagnosis and typing of renal amyloidosis. Kidney Int. 2012;82(2):226-34. doi:10.1038/ki.2012.108
97. Amyloidosis Alliance. https://www.amyloidosisalliance.org/
98. Mutations in Hereditary Amyloidosis database. http://www.amyloidosismutations.com/