Shwachman–Diamond Syndrome

A multisystem ribosomopathy with predominantly autosomal recessive inheritance, exocrine pancreatic insufficiency, cytopenias predominantly affecting neutrophils, metaphyseal skeletal dysplasia and lifelong predisposition to myelodysplastic syndrome and acute myeloid leukaemia; SBDS accounts for the vast majority of cases and DNAJC21, EFL1 and SRP54 for rare forms.

Literature and regulatory information updated: 10 October 2026.

0. Initial Reference Table

1. Introduction

1.1. Original description and clinical consolidation

In 1964, Shwachman, Diamond, Oski and Khaw published in the Journal of Pediatrics the seminal description of a condition they called “the syndrome of pancreatic insufficiency and bone marrow dysfunction” (The syndrome of pancreatic insufficiency and bone marrow dysfunction) [13]. In subsequent years, Pringle and colleagues (1968) drew attention to the association between pancreatic insufficiency, blood dyscrasia and metaphyseal dysplasia, and the series by Aggett and colleagues (1980) consolidated the pancreatic, haematological and skeletal triad that came to define the entity [2]. Bodian's eponym survives in the synonym Shwachman–Bodian–Diamond syndrome, and the radiological literature has retained descriptive terms such as Shwachman-type metaphyseal chondrodysplasia or congenital pancreatic lipomatosis [1, 2]. SDS is now recognised as the second most frequent cause of hereditary exocrine pancreatic insufficiency, after cystic fibrosis [2].

1.2. From genetic linkage to ribosomopathies

Linkage mapping, published by Goobie and colleagues in 2001, placed the locus in the centromeric region of chromosome 7 [1]. In 2003, Boocock and colleagues identified mutations in a gene of previously unknown function, located within a 1.9 cM interval at 7q11, which they named SBDS: the transcript is 1.6 kb long, encodes a predicted protein of 250 amino acids and coexists with a pseudogene (SBDSP1) with 97 % nucleotide identity, contained within a local genomic duplication of 305 kb [14]. In that study, 89 % of unrelated individuals (141/158) carried recurrent alterations resulting from gene conversion from the pseudogene, and 60 % (95/158) had both alleles converted [14]. The subsequent demonstration that the SBDS protein, in cooperation with the GTPase EFL1, removes the anti-association factor eIF6 from the 60S ribosomal subunit, allowing formation of the active 80S ribosome, placed SDS among the ribosomopathies, alongside Diamond–Blackfan anaemia and Treacher Collins syndrome [1, 15].

1.3. New genes and reclassification

Between 2016 and 2018, three genes responsible for clinically overlapping conditions were described. Tummala and colleagues (2016) associated biallelic DNAJC21 variants with a bone marrow failure syndrome with cancer predisposition and disturbed 60S subunit maturation; Dhanraj and colleagues (2017) documented that these variants cause an SDS phenotype [16, 17]. Stepensky and colleagues (2017) described biallelic variants in EFL1, the functional partner of SBDS, in children with pancytopenia, exocrine pancreatic insufficiency and skeletal anomalies [18]. Carapito and colleagues (2017) and Bellanné-Chantelot and colleagues (2018) identified heterozygous SRP54 variants, mostly de novo, in severe congenital neutropenia with features of SDS [4, 19]. The 2024 version of GeneReviews defines the molecular diagnosis of SDS by the presence of biallelic variants in SBDS, DNAJC21 or EFL1, or a heterozygous variant in SRP54 [1]. Other classifications designate non-SBDS forms as “SDS-like”: the 2024 update from the American Association for Cancer Research (AACR) groups SBDS and EFL1 as SDS and DNAJC21 and SRP54 as SDS-like [20]. A table in a textbook of metabolic diseases includes EIF6 among the SDS genes [15]; this inclusion does not correspond to the germline genetics of the disease, because the EIF6 alterations described in SDS are somatic, acquired by haematopoiesis and compensatory in significance, and another section of the same book correctly lists SBDS, DNAJC21, EFL1 and SRP54 [21, 22].

1.4. Definition of the entities covered

This article addresses classical SBDS-related SDS (SDS1) and the EFL1-related (SDS2), DNAJC21-related and SRP54-related forms, with dedicated subsections where the phenotype diverges. It also includes severe neonatal SBDS-related spondylometaphyseal dysplasia, which Orphanet distinguishes as an entity (ORPHA:622934) and which represents the severe end of the skeletal spectrum [7]. The other ribosomopathies, notably Diamond–Blackfan anaemia and Treacher Collins syndrome, Fanconi anaemia and cystic fibrosis appear here only in the differential diagnosis. The 2011 international consensus (Dror and colleagues), developed by experts from Canada, Europe and the United States, was the first diagnostic and treatment document dedicated to SDS, and was supplemented by a European Society for Blood and Marrow Transplantation (EBMT) consensus on transplantation (2022), by the AACR haematological surveillance recommendations (2024) and, in France, by the French National Diagnosis and Care Protocol (Protocole National de Diagnostic et de Soins, PNDS) of the Haute Autorité de Santé (HAS) dedicated to SDS (March 2023) and by the PNDS on chronic neutropenias (May 2024) [9–11, 20, 23]. Both PNDS were developed by the French reference centre for chronic neutropenias according to the 2010 HAS method, the first with literature up to the end of 2022 and data from the national registry; the authors stress that no therapeutic trial exists in SDS and that most recommendations are grade C (low level of evidence), rarely B (scientific presumption, mainly from cohort studies) [9, 23].

2. Molecular Genetics and Pathophysiology

2.1. Genetic architecture

SDS is genetically heterogeneous, but strongly concentrated in a single gene. GeneReviews attributes approximately 92 % of cases to SBDS, less than 1 % each to DNAJC21, EFL1 and SRP54, and leaves fewer than 10 % of patients with a clinical diagnosis without an identified cause [1]. The proportion without a molecular cause depends on how the cohort is defined: in the National Cancer Institute (NCI) study of inherited bone marrow failure syndromes, only 32 of 47 patients (68.1 %) with neutropenia and pancreatic insufficiency had SBDS variants, and 15 patients with SDS and 7 with an SDS-like phenotype remained without a cause despite extensive assessment [24]; in the Boston Children's Hospital registry (n=419, including clinical phenotypes), SBDS accounted for 62 % of cases, rare genes for 4 %, and 34 % remained without a cause, according to the review by Yang and colleagues [25]. In cohorts selected for molecular confirmation, the proportion of SBDS approaches 99 % [26].

Sources: [1, 4, 5, 16, 18]. AR, autosomal recessive; AD, autosomal dominant.

The relationship between SBDS and SDS has been classified as definitive by ClinGen (Clinical Genome Resource) since 30 July 2020, with autosomal recessive inheritance; ClinGen has issued no actionability assertion for SBDS or DNAJC21 [5].

2.2. The SBDS protein and 80S ribosome assembly

The 60S ribosomal subunit is assembled in the nucleus with the factor eIF6 (eukaryotic initiation factor 6) bound to its interface, where it prevents premature association with the 40S subunit. In the cytoplasm, the SBDS protein and the GTPase EFL1 (elongation factor-like 1) cooperate to displace eIF6, a step that licenses joining of the two subunits into the active 80S ribosome and recycles eIF6 to the nucleus [5, 18]. SDS is therefore a disease of the final stage of ribosome assembly, distinct from ribosomopathies affecting transcription of the rRNA precursor (pre-rRNA; Treacher Collins), pre-rRNA processing (dyskeratosis congenita, cartilage–hair hypoplasia) or subunit maturation (Diamond–Blackfan) [15]. Finch and colleagues (2011) showed that, in SDS, GTP hydrolysis by EFL1 becomes uncoupled from eIF6 release, and the EFL1 mutants described by Stepensky and colleagues retain GTPase activity but do not promote release of the eIF6 homologue in yeast [1, 18]. SBDS also associates with multiple ribosomal proteins and has been assigned roles in mitochondrial function (its reduction decreases mitochondrial membrane potential and increases reactive oxygen species), the DNA damage response, mitotic spindle stability and the stromal microenvironment [1, 5, 24]. There is no clinical enzymatic assay for SBDS function [5].

Complete loss of SBDS appears incompatible with life. The Sbds−/− mouse dies in early embryonic life, zebrafish larvae lacking Sbds have limited survival, and no patient homozygous for the null allele c.183_184delinsCT has ever been described, despite its frequency [1, 27]. By contrast, individuals homozygous for c.258+2T>C exist; this is a hypomorphic splicing variant allowing production of a small amount of normal transcript [24, 28]. In the North American Registry, 158 of 159 patients with biallelic SBDS (99.4 %) had at least one copy of c.258+2T>C; the sole patient without it had an extremely severe phenotype, and mouse models lacking the hypomorphic allele have fulminant phenotypes [28].

2.3. Links between the ribosomal defect and tissues

Insufficient protein synthesis preferentially affects tissues with a high translational demand. In the pancreas, the disease manifests as depletion of acinar cells, with extensive fatty replacement and preservation of the ducts and islets; ductal secretion of water and electrolytes is normal, whereas secretion of proteolytic enzymes is markedly reduced [1, 10]. Acinar dysfunction is generalised and also affects the parotid gland [1]. In a murine model with conditional inactivation of Sbds in the pancreas, acinar atrophy resulted from a TP53-dependent decrease in digestive enzyme synthesis and from early cellular senescence; as pancreatic protein synthesis peaks in the neonatal period, this finding is consistent with the peak of digestive symptoms at that stage [9]. In the marrow, accumulation of eIF6 on the 60S subunit activates the TP53 pathway, and reducing the eIF6 dose attenuates this activation in zebrafish models; in this preclinical model, reduction of Eif6 partially rescues survival but does not correct the neutropenia [27]. In addition to the quantitative deficiency, neutrophils have impaired chemotaxis and motility, which contributes to infections [1, 2]. In the growth cartilage, the defect results in metaphyseal chondrodysplasia, and in the liver in early hepatocellular injury and steatosis [1, 29].

2.4. Clonal haematopoiesis: compensatory and premalignant pathways

The selective pressure imposed by the ribosomal defect favours haematopoietic clones that attenuate it. Kennedy and colleagues (2021) studied 110 patients with a clinical diagnosis of SDS and used deep sequencing with double-stranded unique molecular identifiers, with a sensitivity of 0.1 to 1 % variant allele fraction [22]. They found 327 somatic mutations in 74 of 98 patients (76 %) with germline SBDS, and none in SDS-like patients without variants in SBDS, EFL1, DNAJC21 or SRP54; the most frequently affected genes were EIF6 (61 %), TP53 (45 %), PRPF8 (12 %) and CSNK1A1 (6 %), and age was independently associated with the presence of a mutation (odds ratio of 1.1 per year) [22]. Detectable clonal haematopoiesis increased from 59 % up to 10 years of age to 89 % between 11 and 20 years and 100 % from 21 years onwards [22].

The two pathways have opposite significance. EIF6 mutations (recurrent p.Asn106Ser in 20 % and p.Arg96Trp in 13 % of patients) disrupt eIF6 binding to the 60S subunit or destabilise the protein, alleviating the ribosomal defect; the interstitial deletion del(20q), which includes EIF6, acts by the same mechanism, and neither alteration was associated with myeloid neoplasia [22]. TP53 mutations were more frequent in patients with myeloid neoplasia (73.3 % versus 39.8 %), but the presence, number, persistence and variant allele fraction of monoallelic TP53 clones did not predict imminent leukaemia; progression was associated with biallelic inactivation of TP53, through a second mutation, 17p deletion or copy-number-neutral loss of heterozygosity [22]. Single-cell analysis of 33,426 cells from 6 patients showed that EIF6 and TP53 mutations almost always occur in independent clones [22]. The AACR update estimates clonal haematopoiesis in 70 % of patients at 20 years and considers the TP53 pathway, particularly biallelic alterations, to be the progression pathway [20]. Reilly and Shimamura (2023) emphasise that most myeloid neoplasms in SDS have biallelic loss of TP53 [30].

Other studies, summarised by the French PNDS (French National Diagnosis and Care Protocol), complete this picture. In a cross-sectional study from St. Louis of 27 patients with SDS without MDS or AML, 13 (48 %) had clonal haematopoiesis with one or more TP53 variants, almost always at very low allele fractions (0.1 to 7 %), and none of the patients with ELANE neutropenia in the same study had them; in a Boston series of 1514 patients transplanted for MDS without a prior diagnosis of SDS, the 7 youngest (under 35 years) among the 289 with a TP53 mutation had SBDS variants and an unfavourable post-transplant course [9]. In the French study published as an abstract (80 patients), about 50 % had TP53 variants, with allele fractions above 20 % both in malignant severe cytopenia and in severe cytopenia without a cytogenetic clone, and intermediate fractions (0.1 to 20 %) in the stable phase; serial follow-up showed TP53 clones that disappear and others that emerge, and the clones at transformation differed from those of the chronic phase [9]. In 12 patients, clone detection and allele fraction were similar in blood and marrow, which supports less invasive molecular surveillance in peripheral blood [9]. The PNDS also interprets isochromosome i(7)(q10) as somatic compensation: loss of 7p and duplication of 7q, where SBDS is located, duplicate the more productive SBDS allele, with a protective effect in the cells that carry it [9].

2.5. SBDS variant spectrum and nomenclature

The MANE Select (Matched Annotation from NCBI and EMBL-EBI) reference transcript for SBDS is NM_016038.4; GeneReviews uses version NM_016038.2, whose c. coordinates match those of version .4 for the variants listed here [1, 31]. More than 90 % of patients have at least one pathogenic variant in exon 2 arising through gene conversion, that is, replacement of an SBDS segment by sequence copied from SBDSP1, introducing variants typical of the pseudogene [1]. The two recurrent variants are pan-ethnic and result from this mechanism rather than a classical founder effect; c.183_184delinsCT frequently occurs in cis with c.201A>G, within the same converted segment [31, 32].

Sources: [1, 31, 33].

The frequency of the compound genotype c.258+2T>C/c.183_184delinsCT varies between cohorts: 65 % in the European severe chronic neutropenia registry, 78.1 % in the NCI cohort and 61.4 % in the North American Registry growth cohort; in the European registry, 96 % of patients with mutations had at least one copy of c.258+2T>C [24, 26, 34]. As c.183_184delinsCT and c.258+2T>C can coexist on the same converted allele, the complex allele c.[183_184delinsCT;258+2T>C] requires parental testing to establish whether the two variants are in trans (affected patient) or in cis (carrier) [1]. Rare variants such as c.297_300del also appear to result from gene conversion [1]. According to the French PNDS, more than 90 % of patients have at least one of the three recurrent forms (c.183_184delinsCT, c.258+2T>C or the complex allele) and about two thirds have two; the two recurrent variants account for about 90 % of identified alleles, and the PNDS assigns them population allele frequencies of the order of 1/125 and 1/3125 [9]. These values are not consistent with the carrier frequency of 1 in 205 estimated in Italy nor with gnomAD v4.1, in which c.258+2T>C (NC_000007.14:g.66994210A>G) has an allele frequency of about 1 in 270 in exomes, in a region where the pseudogene hampers alignment of short reads [9, 35].

Deletions and rearrangements form a minority but are clinically relevant. GeneReviews estimates that fewer than 2 % of SBDS variants are whole-exon deletions, extended conversions of exon 2 and the neighbouring introns or rearrangements involving exon 2, and that approximately 8 % of pathogenic variants are not identified by either sequencing or targeted deletion and duplication testing [1]. Costa and colleagues (2007) described an AluSx-mediated exon 3 deletion in a Portuguese patient [1]. In the North American Registry, one patient with a single variant detected by sequencing had, on microarray, a heterozygous 4 kb deletion encompassing the entire exon 5, of paternal origin and confirmed by quantitative polymerase chain reaction (PCR) [28]; in the NCI cohort, long-read sequencing revealed a 19 kb deletion (part of intron 4, exon 5 and the 3' untranslated region) and an 872 bp deletion including exon 3 [24]. Steele and colleagues (2014) described an SBDS founder variant in the French-Canadian population, whose identity is not stated in the sources consulted [1].

2.6. SRP54, EFL1 and DNAJC21 variants

The most frequent SRP54 variant is the in-frame deletion p.Thr117del, present in 14 of the 23 cases in the French congenital neutropenia registry; normalised to the MANE transcript, it corresponds to NM_003136.4:c.349_351del p.(Thr117del), NC_000014.9:g.35007376_35007378del, recorded in ClinVar (VCV000430852, where it appears as c.343ACA[2]) as pathogenic/likely pathogenic [4, 31, 33]. The remaining described SRP54 variants are missense variants affecting conserved residues in the GTPase domain, with reduced GTPase activity; the accepted disease mechanism for all four genes is loss of function [1, 19]. For EFL1, the patients reported by Stepensky and colleagues were homozygous for p.Arg1095Gln (4 patients) or p.Met882Lys (2 patients), in families from Israel, Palestine and Chicago [18]. For DNAJC21 and EFL1, the sources consulted present variants only in protein nomenclature, without a reference transcript.

2.7. Pathophysiology of non-SBDS forms

DNAJC21 is a DNAJ/Hsp40 family co-chaperone, with two N-terminal TPR repeats and a C-terminal DNAJ domain, which binds to the 45S rRNA precursor and participates in early nuclear rRNA biogenesis and late cytoplasmic 60S maturation with PA2G4, HSPA8 and ZNF622 [5]. The mutant proteins interact less with these cofactors; patients' lymphoblastoid cell lines have reduced rRNA, aberrant ribosomal profiles, cytoplasmic PA2G4 accumulation, increased cell death and increased sensitivity to actinomycin D [16]. SRP54 is the GTPase of the signal recognition particle, a complex of one RNA and six polypeptides (SRP9, SRP14, SRP19, SRP54, SRP68 and SRP72) that directs proteins with a signal peptide to the endoplasmic reticulum [1]. Its expression rises during granulocytic differentiation; mutation or silencing reduces granulocytic proliferation through p53-dependent apoptosis, and patients' marrow shows endoplasmic reticulum stress and autophagy [4, 5]. In the marrow of patients with an SRP54 variant, SRP54 mRNA was 3.6-fold below normal, and zebrafish with gene silencing reproduce the neutropenia, reduced chemotaxis and small exocrine pancreas [19].

2.8. Genotype–phenotype correlation, penetrance and modifiers

There is no established genotype–phenotype correlation for any of the four genes, nor systematic phenotypic differences between genes [1]. The French cohort of 102 patients described 15 new SBDS variants without prognostic value of the genotype, and the European registry and NCI cohort reached the same conclusion, the latter with a narrow genotypic spectrum [24, 26, 36]. Expressivity is highly variable, even between affected siblings, and SDS has been diagnosed in asymptomatic siblings of probands [1, 37]. A relative exception is the absence of the hypomorphic c.258+2T>C allele, observed in the most severely affected patient in the North American Registry [28].

2.9. Allelic diseases and significance of carrier status

GeneReviews recognises no other phenotypes associated with germline variants in DNAJC21, EFL1 or SBDS; heterozygous SRP54 variants are also associated with isolated congenital neutropenia, without other SDS features [1, 4]. Severe neonatal SBDS-related spondylometaphyseal dysplasia is treated by Orphanet as a distinct entity, but results from the same biallelic variants [7]. Orphanet links SBDS to idiopathic aplastic anaemia (ORPHA:88) as a major susceptibility factor, following the work by Calado and colleagues (2007), which suggested increased risk in heterozygous carriers [1, 6]. This finding has not been reproduced: the North American Registry observed no aplastic anaemia in carriers from more than 200 families (unpublished data cited by GeneReviews), marrow sequencing in 77 patients with AML revealed no SBDS variants, and in a cohort of 160 children with sporadic AML the frequency of heterozygous variants was similar to that in healthy blood donors, with no biallelic genotype [1].

3. Clinical Spectrum and Manifestations by System

3.1. Overview and data sources

The current clinical description rests on more than 500 individuals with a molecular diagnosis, gathered in the systematic review by Han and colleagues (2023) used by GeneReviews, and on national and international registries with long follow-up: the North American SDS Registry (Boston and Cincinnati), the French Severe Chronic Neutropenia Registry, the Italian Registry (RI-SDS, Verona), the European Severe Chronic Neutropenia International Registry (SCNIR), the Canadian Inherited Bone Marrow Failure Study and the NCI cohort [1, 24, 26, 28, 36]. The registries differ in inclusion criterion (clinical diagnosis or molecular confirmation), patient age and period, and the frequencies that follow should be read with their respective denominator. The French PNDS (French National Diagnosis and Care Protocol) summarises the estimated overall frequencies: neutropenia 90 %, exocrine pancreatic insufficiency 90 %, developmental delay or neuropsychological difficulties 70 %, metaphyseal dysplasia 30 %, anaemia 20 %, hepatic cytolysis 20 %, heart disease 10 % and type 1 diabetes 3 %; these are expert estimates based on the literature and the national registry, without a single denominator [9]. The same document estimates that haematological complications affect nearly 40 % of patients and cause most early deaths [9].

Sources: [1, 24, 26]. ALL, acute lymphoblastic leukaemia.

3.2. Presentation and variability

Most newborns have no manifestations, and the disease usually presents in infancy with poor weight gain and growth impairment secondary to pancreatic insufficiency, recurrent infections (respiratory tract, middle ear, skin) or detection of neutropenia [1, 2]. A minority present in the neonatal period with severe infections, severe bone marrow failure or aplastic anaemia, asphyxiating thoracic dystrophy due to a restrictive rib cage or severe spondylometaphyseal dysplasia [1]. In the North American Registry, among 37 patients with biallelic SBDS, neutropenia was the most common haematological abnormality at presentation (30/37, 81 %), but only 19 (51 %) had the classical combination of neutropenia and steatorrhoea; 24 (65 %) had congenital anomalies, and two asymptomatic siblings of probands were diagnosed through family screening [37]. Median age at diagnosis was 1.78 years in the NCI cohort (0.12 to 41.82 years) and 1.75 years in the European registry (0 to 23.62 years) [24, 26]. Sex ratios are discordant: the 2011 consensus reports 1.7 males per female in the literature, whereas the NCI cohort found 0.74:1 [10, 24].

3.3. Exocrine pancreas and gastrointestinal tract

Pancreatic dysfunction results from severe depletion of acinar cells and is most evident in the first year of life, frequently within the first six months; the 2011 consensus states that it is diagnosed within the first six months and in 90 % of patients in the first year [1, 10]. Severity ranges from an asymptomatic form to steatorrhoea with malabsorption and fat-soluble vitamin deficiency. Loss of approximately 98 % of pancreatic reserve is required for maldigestion to occur, so faecal fat may be normal [10]. Histology shows few acinar cells and extensive fatty infiltration; ultrasound and computed tomography (CT) show a pancreas small for age, and magnetic resonance imaging (MRI) shows fatty infiltration with preserved ductal and islet components, but imaging may initially be normal [1, 37].

Pancreatic function improves with age. Mack and colleagues (1996) observed that up to 50 % of patients can stop enzymes by 4 years of age with normal fat absorption, although enzyme secretion remains below normal; the 2011 consensus stresses that the quantitative secretory deficiency persists in all [1, 10]. More recent reviews estimate that 40 to 60 % of initially insufficient patients become pancreatic sufficient [34, 38]. In the North American Registry growth cohort (127 patients), 94 % had exocrine pancreatic insufficiency and 94 % were receiving pancreatic enzyme replacement therapy [34]. The French PNDS considers that practically all patients have an exocrine pancreatic deficiency, but that about 30 % are clinically sufficient, and that residual function is expressed mainly between 3 months and 5 years; beyond early childhood, symptoms are usually subtle (persistent anorexia, chronic diarrhoea with bulky, frothy and fatty stools, abdominal pain, slowing of the growth curve) [9].

Gastrointestinal involvement extends beyond the pancreas. Shah and colleagues (2010) found duodenal inflammation in more than half of symptomatic patients undergoing biopsy, suggesting an enteropathic component capable of contributing to persistent vitamin deficiencies despite enzymes [1]. In the NCI cohort, 85.7 % had gastrointestinal manifestations, with malabsorption in 86.2 %, gastro-oesophageal reflux in 78.6 % and steatorrhoea in 70.4 %; 12 patients required tube feeding and 3 parenteral nutrition [24]. An endoscopic series from the North American Registry (45 patients, 102 procedures) found abnormal biopsies in 44.4 % of patients with symptom data, with higher yield in the presence of dysphagia or dyspepsia (80 %), gastrointestinal bleeding, elevated C-reactive protein, elevated faecal calprotectin or elevated liver enzymes; before HSCT, biopsies were generally normal or showed mild inflammation, and duodenal inflammation was less frequent than in Shah's study [39].

3.4. Haematology: neutropenia, anaemia, thrombocytopenia and bone marrow failure

Persistent or intermittent neutropenia is the first recognised abnormality in almost all patients (88–100 %), often before diagnosis, and can be neonatal [1, 10]. Neutropenia and impaired neutrophil chemotaxis are the main contributors to infections in early childhood; patients nevertheless retain the capacity to form abscesses and empyema, and deep infections can be fatal in young children [1]. Anaemia is hyporegenerative in up to 80 % of patients, normocytic or macrocytic, with elevated fetal haemoglobin (HbF) in 80 %; thrombocytopenia (<150 × 10⁹/L) is variable, and a minority develop severe aplasia [10]. The marrow is generally hypocellular with increased fat, left-shifted granulopoiesis and frequent, fluctuating mild trilineage dysplasia; prominent multilineage dysplasia may indicate transformation [10].

Longitudinal data show that counts do not deteriorate linearly. In the North American Registry (153 patients with biallelic SBDS, 143 families, median age at last follow-up 10.4 years), analysis of 2146 full blood counts from 119 patients, excluding values during granulocyte colony-stimulating factor (G-CSF) treatment or transfusions, showed severe trilineage cytopenia in 9 patients (7.6 %), at a median age of 5.4 years, and a normal full blood count on at least one occasion in 56 (47.1 %) [28]. The absolute neutrophil count (ANC) increased with age, haemoglobin rose up to 18 years and fell thereafter, and platelets decreased, more rapidly before 7 years [28]. In 547 marrow samples from 119 patients, median cellularity was 40 %; only 35.3 % of marrow samples before 1 year were normocellular, all marrow samples after 16 years were hypocellular for age, and cellularity correlated weakly with haemoglobin and platelets and not with ANC [28]. In the European registry, the initial full blood count (108 patients, median age 1.45 years) showed neutropenia in 81 % (severe, below 500/µL, in 35 % of these), bicytopenia in 25 % and near-normal counts in 17 %; with age, leucocytes and platelets decreased and haemoglobin and mean corpuscular volume (MCV) increased, with a stable median ANC above 800/µL and above 1000/µL in adults; the fall in leucocytes was mainly due to lymphocytes [26].

Severe bone marrow failure is concentrated in early childhood. In the North American Registry, 8 patients (5 %) underwent transplantation for severe bone marrow failure, at a median age of 1.7 years and 7 of 8 before 8 years, and none had clonal haematopoiesis [28]. In the French national cohort (102 patients, median follow-up 11.6 years), definitive severe cytopenia (haemoglobin <7 g/dL or platelets <20 × 10⁹/L) occurred in 21 patients (9 malignant, 9 non-malignant and 3 malignant preceded by a non-malignant phase), with a 20-year cumulative risk of 24.3 % (95 % confidence interval, 95 % CI, 15.3–38.5 %); 41 patients (40 %) had significant haematological manifestations, including intermittent severe cytopenias [1, 36]. Symptom onset before 3 months and low haematological parameters at diagnosis and during follow-up were associated with severe complications, and transient severe cytopenias or indolent cytogenetic clones had no deleterious significance [36]. In the Italian Registry (121 patients with biallelic SBDS), cumulative incidence at 30 years was 59.9 % for severe neutropenia, 66.8 % for thrombocytopenia and 20.2 % for anaemia, according to Furutani's review [28].

The French series details the stable phase [9]. At diagnosis, median ANC was 0.82 × 10⁹/L (interquartile range 0.5–1.7), haemoglobin 11.4 g/dL (9.9–12.2) and platelets 217 × 10⁹/L (142–319); 38 of the 102 patients had cytopenia, 23 with ANC <0.5 × 10⁹/L, 11 with platelets <100 × 10⁹/L and 13 with haemoglobin <9 g/dL [9]. During chronic-phase follow-up (median of 17 blood counts per patient), values fluctuated without a regular pattern, with median ANC of 0.75 × 10⁹/L, haemoglobin of 11.7 g/dL and platelets of 185 × 10⁹/L, and 28 patients had persistent cytopenia (19 with median ANC <0.5 × 10⁹/L, 11 with platelets <100 × 10⁹/L and 7 with haemoglobin <9 g/dL) [9]. Of 81 marrow aspirates at diagnosis, 20 (24.7 %) showed granulocytic maturation arrest, and marrow smears usually had neutrophils with condensed chromatin and hyposegmentation in more than 10 % of cells, a proportion that exceeded 20 % when severe cytopenia was present [9]. Two patients with the same compound genotype, whose neutrophil counts the PNDS reproduces over 35 years, had medians of 0.51 and 2 × 10⁹/L, illustrating variability with an identical genotype [9].

The French classification of haematological complications, later adopted by the North American and Italian registries, distinguishes the stable phase, with mildly symptomatic moderate cytopenias, from complications with severe cytopenia (SC), defined as non-regenerative anaemia <7 g/dL or thrombocytopenia <20 × 10⁹/L in addition to neutropenia, with transfusion requirement [9]. In the presence of SC, marrow examination with cytology, cytogenetics and molecular studies is mandatory, and spontaneous recovery within 3 months, with hypoplastic marrow and no malignant clone, defines the transient form, typical after infection [9].

Source: [9].

In the French cohort, malignant SC arose at a median age of 11.1 years and non-malignant SC at 0.13 years; non-malignant SC never arose after 15 years of age, whereas the risk of malignant transformation was high before 1 year, decreased up to 15 years and remained approximately constant thereafter, including after 30 years [9]. Three patients with non-malignant SC treated with transfusions and G-CSF, all diagnosed in the neonatal period and none transplanted, later progressed to MDS [9]. Definitive SC accounted for 15 of the 17 deaths in the cohort, and the 6 survivors with definitive SC had all been transplanted [9]. Twelve patients had transient SC (median age 0.83 years), with normalisation within 3 months and subsequent stability, except for one patient with a new definitive non-malignant SC 8 years later; the picture was compatible with viral infection, although parvovirus, Epstein–Barr virus and cytomegalovirus had been excluded [9]. In the French and Italian registries, the probability of SC at 20 years, with or without a clone, lies between 20 and 35 % [9].

3.5. Immunological abnormalities beyond neutrophils

The immunodeficiency of SDS is not limited to neutropenia. In the study by Gloude and colleagues (98 patients with biallelic SBDS, 409 time points before any HSCT), B lymphocytes (CD19) were below the reference range in 31 % of measurements and in 41 % of patients at least once; IgM was low in about 30 % of measurements and in 50 % of patients at least once, and IgG in 16 % of patients at least once [40]. B-cell immunophenotyping was abnormal in 8 of 11 patients studied, and lymphocyte proliferation was abnormal or mixed in 6 of 18 [40]. In the 18 patients with low B lymphocytes, 129 sinopulmonary or skin and soft tissue infections were recorded, and only 18 % of infections occurred with an ANC below 500/µL, indicating that neutropenia alone does not explain the infectious risk [40]. In the NCI cohort, 92 % of patients had recurrent infections (otitis media in 94 %), 20.5 % required admission with intravenous antibiotics and 45.8 % received intravenous immunoglobulin [24]. In the French cohort, 52 of 81 patients with evaluable immunoglobulins (65 %) had polyclonal hypergammaglobulinaemia [9]. Forty-three of the 102 patients (42 %) had at least one severe infection, the first at a median age of 9.6 years (interquartile range 2.4–16.9), more frequent in childhood but possible up to the fourth decade; of the 72 severe infections, 30 were pneumonias, 26 cellulitis, 19 septicaemias (8 associated with another infection), 3 arthritis, 1 meningitis and 1 colitis, and 13 patients had acute stomatitis [9]. Three severe viral infections were recorded (severe disseminated varicella, fatal measles and influenza with cardiomyopathy), and the PNDS on chronic neutropenias stresses that influenza, varicella and respiratory syncytial virus can be very severe in SDS, unlike what is observed in ELANE neutropenia [9, 23]. Periodontal disease appears less frequent than in ELANE neutropenias, but more frequent than in the population without neutropenia [9].

3.6. Myelodysplastic syndrome and acute myeloid leukaemia

The risk of myeloid transformation is the main determinant of prognosis, but estimates vary with the definition of MDS, cohort age and period. Published series report transformation rates of 5 to 36 %, and the AACR update estimates an MDS/AML incidence of 10 to 30 %, chiefly in patients with SBDS, from early childhood to adulthood [20, 41].

Sources: [1, 26, 28].

The absence of cases in the Israeli and 2010 NCI cohorts reflects small and young samples and does not demonstrate zero risk [1]. In the North American Registry, MDS arose at a median age of 12.3 years (0.5 to 45.0) and AML at 28.4 years (14.4 to 47.3), which shifts the concern, after early childhood, to late childhood and young adulthood [28]. The 2011 consensus, based on a literature review, reports that about half of patients with a clonal cytogenetic abnormality or MDS in childhood progressed to leukaemia within 1 to 37 years, and that males accounted for 68 % of cases of clonal abnormality or MDS and 92 % of leukaemias; the atlas of skeletal dysplasias describes haematological malignancies in more than 5 % of cases, mainly in males with myelodysplasia and clonal abnormalities [2, 10]. The AML cases described include the M0, M2, M4, M5 and M6 subtypes of the French–American–British (FAB) classification, with M6 erythroleukaemia in about 30 % of classifiable leukaemias; acute lymphoblastic leukaemia and juvenile myelomonocytic leukaemia are rare [10]. In the French cohort, the 12 malignant transformations comprised 8 AML (M2 in 2, M4 in 2, M6 in 3 and M0 in 1) and 4 MDS (3 refractory cytopenias with multilineage dysplasia and 1 refractory anaemia with excess blasts, MDS-EB1 in the 2016 classification), and monosomy 7 was the most frequent abnormality (6 of the 12), with isolated i(7)(q10) in 2 [9]. The table of constitutional bone marrow failure syndromes in the PNDS on bone marrow failure summarises SDS with a cumulative incidence of MDS/AML between 18 and 36 % at 30 years, male predominance and a rule for marrow interpretation: as granulocytic dysplasia is frequent without MDS, a diagnosis of MDS is accepted only with more than 5 % blasts; it adds that small TP53 clones are frequent and without significance, unlike TP53 variants with a high allele fraction [3].

The North American multicentre cohort of Myers and colleagues (2020; 17 centres, 36 patients with SDS and MDS or AML between 2001 and 2017, median age 18 years) describes the characteristics of transformation [41]. Baseline marrow in SDS is hypocellular, with mild dyspoiesis, predominantly myeloid (neutrophils with incomplete nuclear segmentation and hypogranulation); in MDS, prominent nuclear hypolobulation (pseudo-Pelger–Huët forms), marked hypogranulation, increased myeloid precursors with basophilic cytoplasm and significant dyserythropoiesis emerge [41]. Erythroid dysplasia is unusual in baseline SDS, so dyserythropoiesis with progressively increasing cellularity suggests MDS even without cytogenetic changes, and leukaemia can arise without preceding dysplasia or abnormal cytogenetics [41]. Agreement between local diagnosis and central review was only 56 %, chiefly because of the difficulty of counting blasts in hypocellular marrow with baseline dysmorphia, and complex karyotype was more frequent in AML (80 %) than MDS (36 %) [41]. Thirty-three of the 36 patients had already been diagnosed with SDS before the neoplasm; 3-year survival was 11 % in AML (median survival 0.99 years) and 51 % in MDS (median 7.7 years) [41].

3.7. Clonal cytogenetics

The characteristic cytogenetic abnormalities are deletion del(20q11) and isochromosome 7, i(7)(q10); they may persist, fluctuate or disappear without leading to MDS/AML [1, 10]. In the North American Registry, 41 of 111 patients with cytogenetics (36.9 %) had clonal abnormalities on karyotype or fluorescence in situ hybridisation (FISH): del(20q) in 19.8 % and i(7q) in 9.9 %; the clones disappeared in 15 patients, including 3 of the 11 with i(7q) and 4 of the 22 with del(20q) [28]. Progression to MDS/AML was observed in persistent i(7q) clones and in del(20q) clones with additional evolution (del(12p), monosomy 5 or monosomy 7), as well as with other rearrangements [28]. In the European registry, initial cytogenetics (97 patients, median age 4.73 years) was abnormal in 11 % [26]. The Italian cohort of 91 patients described new abnormalities, including unbalanced structural abnormalities of chromosome 7, complex del(20q) rearrangements and an unbalanced translocation with partial trisomy 3q and partial monosomy 6q, and the same cohort published in 2026 cytogenetic monitoring of up to 27 years [1, 42]. In the French cohort, 9 patients without SC had clones detected in routine marrow examinations, i(7)(q10) in 3 and del(20q) in 5, with cytology comparable to that of routine marrow examinations and no consequence for survival; the PNDS considers these two abnormalities, when isolated, to have no adverse prognostic value and not to be an indication for HSCT [9]. Chromosome 7 abnormalities other than i(7q), especially monosomy 7, carry a poor prognosis [10].

3.8. Lymphoid neoplasms and solid tumours

The North American Registry estimated a risk of lymphoid neoplasia 38 times higher than expected in the population, with increased toxicity of conventional therapies; the sources consulted do not give the denominators [43]. The first case with confirmed biallelic SBDS was a boy with c.258+2T>C/c.653G>A who developed primary mediastinal B-cell lymphoma at 16.9 years; treatment with R-CHOP was complicated by a pressure ulcer, febrile neutropenia, deep vein thrombosis and a decrease in ejection fraction, which required doxorubicin to be omitted, and the patient was in remission at 15 months [28]. Solid tumours are rare and of late onset: ovarian cancer at 41.8 years and breast cancer at 69 years in the NCI cohort, pancreaticoduodenal carcinoma at 24 years in another report, and three solid tumours (breast, ovary and oesophagus) among 155 patients in the French Registry, which GeneReviews places after 50 years of age [1, 24]. GeneReviews also reports dermatofibrosarcoma, pancreatic adenocarcinoma and central nervous system lymphoma [1]. The series published by the French team itself, reproduced in both PNDS, describes these three cases at younger ages: squamous cell carcinoma of the oesophagus in a 48-year-old man and papillary serous carcinoma of the peritoneum in a 47-year-old woman, both treated with neoadjuvant chemotherapy and surgery and who died of septic shock 13 and 6 months after diagnosis, and breast adenocarcinoma in a 43-year-old woman; the literature review adds a dermatofibrosarcoma at 17 years, a breast adenocarcinoma at 31, a pancreatic adenocarcinoma at 38 and an ovarian cancer at 43 [9, 23]. The age discrepancy with GeneReviews is resolved in favour of the primary French series, and the authors consider that the excess risk and the tumour profile are not yet confirmed [9]. According to earlier analyses of the NCI cohort cited by Thompson and colleagues, the overall cancer risk in SDS is 8.5 times that of the US population, and the ratio of observed to expected cases reaches 202 for haematological malignancies; osteosarcoma has not been described, unlike in Diamond–Blackfan anaemia [24].

3.9. Liver

Liver involvement is more frequent and more persistent than the classical description suggested. GeneReviews, based on Finnish studies, describes hepatomegaly and elevated aminotransferases in young children with a tendency to resolve by 5 years of age, elevated bile acids in 7 of 12 patients (3 with persistent or intermittent elevation) and mild histological changes that did not appear progressive, with liver complications in older patients after bone marrow transplantation, notably liver failure due to steatohepatitis [1]. The 2011 consensus reported elevated liver enzymes in up to 75 % of patients, mainly infants, with little long-term consequence [10]. Mitochondrial changes mimicking inborn errors of metabolism and early-onset cirrhosis have been described [1]. The French PNDS describes frequent hepatomegaly in young children and elevated enzymes in up to 75 % of patients, with a tendency to normalise after 5 years, but also severe forms in adolescence and young adulthood, including cirrhosis; liver failure revealed during conditioning for HSCT and a liver transplant attributed to chronic vitamin A intoxication have been published, an attribution that the PNDS authors consider poorly compatible with the exposure described [9]. The French registry includes 2 late-onset liver diseases without vitamin A excess, and histology, including that of newborns, shows chronic inflammation of the bile ducts and fatty infiltration of the liver [9].

The study by Koo and colleagues (2026), in the North American Registry (171 evaluable patients with biallelic SBDS, 2009–2024), changed this interpretation [29]. Defining chronic hepatitis as aspartate aminotransferase (AST) and alanine aminotransferase (ALT) above the upper limit of normal (ULN) in at least three samples separated by one month or more, it found chronic hepatitis in 94 patients (55 %), with onset at a median age of 1.0 year; in children, initial ALT was at least twice the ULN in 71.3 %, with a median value of 183.5 U/L and maximum of 1630 U/L, whereas the 5 adults with adult onset had only elevations of 1 to 2 times the ULN [29]. Median duration was 6.0 years and biochemical resolution occurred in 36.2 %, at a median age of 8.9 years; cholestasis was rare (normal bilirubin in 91.7 %) and albumin was normal in 89.5 % [29]. In approximately 17 % of patients, enzyme elevation preceded the diagnosis of SDS, making unexplained hepatocellular injury in infancy a diagnostic entry point [29]. In 30 biopsies, children had portal or lobular inflammation in 69.5 % and fibrosis in 65.2 % (METAVIR F1 in 53.3 %, F2 in 20 % and F3 in 13.3 %, these two under 2 years), and adults had universal steatosis, fibrosis and, in one case, cirrhosis [29]. Enzyme normalisation does not equate to histological resolution, and there were only two deaths from liver disease, both in transplant recipients [29]. In a series of 8 patients undergoing transient elastography, liver stiffness was above stage F0 in 3, two of whom had normal transaminases, suggesting subclinical liver disease [38].

3.10. Growth and adult height

Mean birth weight lies at the 25th centile, and by the first birthday more than half of children are below the 3rd centile for weight and height [10]. With adequate nutrition and enzymes, growth velocity normalises and weight for height is appropriate, but about 50 % of children remain below the 3rd centile [1]. The French PNDS describes intrauterine growth restriction in about a quarter of patients in the national series (30 % in its table of warning signs) and more than half of children below the 3rd centile for weight and height at one year, and reproduces the growth curves of the Italian registry from 0 to 8 years [9]. The specific curves of the North American Registry (127 patients, 2446 measurements from 0 to 20 years, before HSCT or growth hormone) place the 50th centile for height in SDS close to the 5th centile of the Centers for Disease Control and Prevention (CDC) charts, with comparable body mass index; median adult height is about 7 cm below the reference in women and about 11 cm in men (the same article gives 12 cm in the discussion) [34]. Italian curves from 0 to 18 years exist [1]. The atlas of skeletal dysplasias reports adult heights of 140–150 cm in women and 145–170 cm in men [2]. After HSCT, the slope of the height z-score decreased (−0.26; p=0.0004) [34].

3.11. Skeleton

Metaphyseal chondrodysplasia is the characteristic skeletal component. Mäkitie and colleagues (2004), with cross-sectional and longitudinal data from patients with molecular confirmation, described delayed appearance of the secondary ossification centres (delayed bone age), variable widening and irregularity of the metaphyses in early childhood, followed by progressive thickening and irregularity of the growth plates, and generalised osteopenia; the epiphyseal maturation defects tended to normalise with age, whereas the metaphyseal changes tended to worsen [1]. Rib and joint abnormalities are added, the latter due to asymmetric growth, sometimes with a surgical indication [1]. The 2011 consensus organises the course by age: in infancy, delayed secondary ossification, wide and irregular metaphyses (ribs, wrist), osteopenia and Wormian bones; in mid-childhood, metaphyseal irregularity and sclerosis of the femur; in late childhood and puberty, asymmetric metaphyseal growth, stress fractures and femoral deformity, and vertebral compression fractures, which persist as a risk into adulthood [10]. The French PNDS, which reproduces this table of the course, attributes the bone disease to a chondrocyte abnormality, mainly of protein synthesis in the endoplasmic reticulum and the Golgi apparatus, and describes the functional consequences: a narrow thorax with possible respiratory repercussions from the neonatal period; after walking age, cartilage fragility with slipped epiphyses and malposition of the hips, lower-limb malalignment due to metaphyseal dysplasia of the distal femur (torsional changes, genu valgum or genu varum, with limited walking distance and chronic pain) and severe scoliosis during growth, especially at the onset of puberty; in adulthood, fractures also favoured by demineralisation due to vitamin D deficiency [9]. Adult height is sometimes well below 150 cm, and supernumerary digits and syndactyly have been described [9].

Sources differ regarding frequency and evolution. The atlas of skeletal dysplasias states that skeletal findings are present in only approximately half of patients and that metaphyseal lesions are often mild, disappear with age or never occur; GeneReviews, based on Mäkitie, considers that characteristic abnormalities exist in all patients with molecular confirmation, potentially subclinically, and that metaphyseal abnormalities progress; the GeneReviews frequency table states 60 %, and the NCI cohort 80.6 % [1, 2, 24]. The difference is largely explained by method: Mäkitie's study applied systematic, serial radiological assessment to patients with a molecular diagnosis, whereas clinical frequencies depend on radiographs requested for symptoms. The atlas figures themselves document progression in the same patient, with minimal proximal femoral irregularities at 8 months, marked femoral neck irregularities at 3 years and very short femoral necks in varus at 9 years [2]. In the NCI cohort, abnormalities affected the extremities (69 %), thorax and ribs (54.8 %), metaphyses (45.8 %) and spine in the form of scoliosis (33.3 %) [24].

Radiologically, the atlas describes minimal to moderate metaphyseal irregularity of the tubular bones, especially of the proximal femora, occasionally with coxa vara, a narrow thorax due to short ribs with flared ends in young children, delayed bone maturation, minor anomalies of the digits, progressive kyphoscoliosis in some patients and, rarely, severe changes in infancy resembling those of Sedaghatian-type spondylometaphyseal dysplasia or even neonatal achondroplasia [2]. Low-turnover osteoporosis is another component: in a Finnish study, 10 of 11 patients with molecular confirmation had bone abnormalities, including reduced mineral density, 3 had vertebral compression fractures and 6 had vitamin D and vitamin K deficiency [1]. In adults of short stature, bone mineral density may be underestimated if not adjusted for height [1]. Across congenital neutropenias in the French registry, 40 % of patients have early osteopenia or osteoporosis, apparently independent of genotype, and an aggravating role of G-CSF has not been demonstrated [23].

3.12. Neurodevelopment, behaviour and brain

Cognitive and behavioural changes are frequent and heterogeneous, with reduced brain volume and a smaller posterior fossa in imaging studies and increased striatal dopamine transporters [1]. Kerr and colleagues (2010) compared 32 children with SDS with children with cystic fibrosis and with siblings: the range of abilities was much wider in SDS, from severe impairment to superior level; about 20 % had intellectual disability in at least one domain, mainly perceptual reasoning, pervasive developmental disorder occurred in 6 % (0.6 % in the general population) and attention deficits were more frequent in children with SDS and their unaffected siblings than in children with cystic fibrosis [1]. The 2011 consensus reports language delay in more than 50 %, intellectual disability (IQ below the 2nd centile) in about 1 in 5 and, in a sample of 32 children and adolescents, attention deficit hyperactivity disorder (ADHD), pervasive developmental disorder or oppositional disorder in 19 %, regardless of pancreatic involvement, otitis, chronic disease, family and age [10]. In the French series, about 60 % of children aged 6 to 11 years were more than 2 years behind at school or attended an inclusive education class (ULIS); only a minority had marked cognitive impairment (IQ <60), and most had satisfactory cognitive abilities with limitation of executive functions and social skills [9]. The PNDS also reports early psychoaffective difficulties with autistic traits in some patients and the psychological burden of chronic disease and haematological risk on parents and siblings, which justifies access to psychological support throughout follow-up [9].

In the North American Registry (240 individuals with biallelic SBDS), the 14 patients undergoing formal neuropsychological assessment (mean age 10.66 years) had a median full-scale IQ z-score of −0.47; 41.7 % had intellectual weakness or impairment and all had impaired adaptive functioning, with frequent involvement of executive functions, visuomotor integration, spelling and mathematics and preserved immediate verbal learning [44]. In the registry's educational data, 73.9 % attended specialised programmes and 51.9 % had an individualised education plan; 49 % had neurodevelopmental disorders (chiefly ADHD) and 23.5 % had mental health diagnoses, chiefly anxiety [44].

3.13. Endocrinology and metabolism

Congenital hypopituitarism, growth hormone deficiency, delayed puberty and diabetes have been described [1]. In the French registry, insulin-dependent diabetes is exceptional (2 patients; 3 % in the PNDS frequency table), and true growth hormone deficiency may contribute to short stature [9]. The Italian Registry confirms an increased risk of diabetes, mainly type 1, although the published abstract does not give the frequency [45]. A North American endocrine surveillance study with an oral glucose tolerance test and a mixed-meal test (NCT04275479) was completed with 11 participants [46]. In the growth cohort, 10 % of patients received growth hormone [34].

3.14. Skin, teeth, sensory organs and malformations

Ichthyosis and eczematous lesions are frequent; in the NCI cohort, 29 of 30 patients had skin changes, with eczema in 62.5 %, café-au-lait macules in 55 %, hypopigmentation in 46.7 % and hyperpigmentation in 21.4 % [1, 24]. Oral disease includes delayed dental development, enamel defects, increased caries in both dentitions, periodontal disease and recurrent oral ulcers associated with neutropenia [1, 10]. Congenital cardiac (septal defects, patent ductus arteriosus), gastrointestinal, urinary and renal, ocular and auditory anomalies are documented, and the heart defects are generally minor [1, 24]. The French series describes more diverse cardiac involvement, in about 15 % of patients: 6 malformations (coarctation of the aorta operated on at 10 years, double aortic arch operated on in the first month, two ostium secundum atrial septal defects, one ventricular septal defect and one tetralogy of Fallot operated on at 6 months) and 6 dilated cardiomyopathies, diagnosed between 6 months and 19 years of age, in the context of H1N1 influenza, measles or sepsis, or 5 years after HSCT with total body irradiation and cyclophosphamide; two patients with cardiomyopathy died in the context of infection and shock, and the others recovered or stabilised with symptomatic treatment [9]. Coarctation of the aorta appears particularly frequent among the malformations [9]. A myopathic syndrome with palpebral ptosis, cleft lip and palate, thumb anomalies, a chronic inflammatory syndrome and, occasionally, inflammatory bowel disease have also been described [9, 23]. Rod–cone dystrophy, juvenile idiopathic arthritis, chronic recurrent multifocal osteomyelitis, scleroderma, ocular inflammation, occasional renal tubular dysfunction with glycosuria, aminoaciduria or tubular acidosis, and hepatic microcysts on MRI have also been described [1, 2].

3.15. Severe neonatal SBDS-related form

The severe end of the skeletal spectrum is severe neonatal SBDS-related spondylometaphyseal dysplasia (ORPHA:622934), described by Nishimura and colleagues (2007): narrow thorax, rhizomelic shortening, mild to severe platyspondyly, lace-like iliac crests, delayed ossification of the caudal ilium, irregular and cupped metaphyses and hypoplastic pubic rami, with severe respiratory failure (some patients requiring lifelong ventilation), severe short stature and developmental delay [1, 7]. Atrial septal defect, seizures, pulmonary hypoplasia, gallbladder hypoplasia and liver fibrosis may coexist, and neutropenia and low amylase may be present [7]. The atlas figures show newborns with confirmed SBDS, short ribs cupped anteriorly, vertically short iliac bones and a wide pubic symphysis, and a patient with horizontal acetabula, underossified pubic rami and short long bones that could be mistaken for achondroplasia [2]. Prenatal cases have been published: two fetuses in Brussels with severe skeletal anomalies suggestive of short-rib thoracic dysplasia, whose molecular diagnosis was delayed by pseudogene interference, and one fetus in China with only slightly straightened ribs at 22 weeks [47, 48]. Asphyxiating thoracic dystrophy caused 2 of the 20 deaths in the North American Registry [28].

3.16. DNAJC21-, EFL1- and SRP54-related forms

The DNAJC21-related forms are characterised by global bone marrow failure or pancytopenia in early childhood, growth and developmental delay, recurrent infections and changes of the skin, teeth or hair; some patients have retinal abnormalities, exocrine pancreatic insufficiency, liver cirrhosis, skeletal anomalies, developmental dysplasia of the hip, joint hypermobility, cryptorchidism or microcephaly [5, 49]. The French PNDS stresses that these patients have mainly anaemia and thrombocytopenia, rather than neutropenia, associated with exocrine pancreatic insufficiency [9]. A child in the original cohort developed AML in childhood, and D'Amours and colleagues propose an entity overlapping SDS and dyskeratosis congenita, with a possible risk of leukaemia [16, 49]. A series of two patients from Brussels (2026) describes phenotypic heterogeneity, clonal evolution with TP53 and allogeneic HSCT, and a series from Birmingham (2026) broadens the phenotypic spectrum; the sources consulted allowed only the metadata of these series to be read [50, 51]. Several patients with DNAJC21 were diagnosed only by exome sequencing [5].

In the EFL1-related forms, the six patients from three families described by Stepensky and colleagues had infantile pancytopenia, exocrine pancreatic insufficiency and skeletal anomalies, in a picture overlapping that of classical SDS [18]. Functional validation of the EFL1 variants, absent from the first report, was obtained in cellular and animal models from patients in the French registry, by teams in Cambridge and at the Imagine Institute; according to the PNDS, the phenotype has the same components as classical SDS, but anaemia and thrombocytopenia are more frequent than neutropenia, especially in the neonatal period, the bone disease is very severe, with extreme short stature (below −5 standard deviations), and a liver disease progressing to cirrhosis has been described in one patient [9]. The French registry included 4 patients with EFL1 in June 2023 [23].

SRP54-related disease is mainly a severe congenital neutropenia. In the French congenital neutropenia registry, SRP54 was the second most frequent cause of congenital neutropenia with maturation arrest: 23 cases (16 sporadic and 7 familial), with profound chronic neutropenia from the first months, promyelocytic maturation arrest and marked dysgranulopoiesis, severe neurodevelopmental delay in 5 and exocrine pancreatic insufficiency requiring enzymes in 3 [4]. The spectrum ranges from isolated neutropenia to a picture overlapping SDS, and episodically normal counts have been described, so a normal count does not exclude the disease [4, 5]. Almost all patients required prolonged G-CSF, with a poor response [4]. The PNDS on chronic neutropenias distinguishes two presentations: in about two thirds of patients, profound chronic neutropenia from the first months with promyelocytic arrest, severe or recurrent bacterial infections and gingivitis, with no other organ involved, with a response to G-CSF lower than in ELANE neutropenias and more refractory forms; in about one third, the same neutropenia with components of SDS (exocrine pancreatic insufficiency, neurocognitive, socialisation and learning difficulties) [23]. Despite high doses of G-CSF, malignant transformation is much less frequent than in ELANE neutropenia and in SBDS-related SDS (estimated risk below 5 %), and the myeloid maturation block with promyelocyte abnormalities distinguishes it from classical SDS; the French registry included 27 patients with SRP54, with 1 death, in June 2023 [9, 23]. By 2023, 38 patients had been published [52]. In the largest series (23 patients, median follow-up of 15 years) there were no neoplasms, but a 15-year-old boy with p.Thr117del was described who developed AML with myelodysplasia-related changes, detected in a screening marrow, with somatic CSF3R and RUNX1 mutations [53].

The PNDS points out that, in about 5 % of presentations of neutropenia with exocrine pancreatic insufficiency, the cause does not lie in SBDS, and that variants of JAGN1, SRP54, SRPRA, G6PC3, EFL1, COX4I2, DNAJC21 and NDUFS2 (the latter usually associated with Leigh syndrome) have been described in this context, each with differences from the classical picture; given this diversity, it recommends high-throughput sequencing of all known genes and, if negative, exome or genome sequencing [9].

4. Clinical Recognition and Warning Signs

4.1. When to suspect the diagnosis

SDS should be considered in any combination of exocrine pancreatic dysfunction, marrow dysfunction and ancillary manifestations, without requiring the complete triad [1]. The North American registry showed that only half of patients had neutropenia and steatorrhoea at presentation, and that absence of pancreatic lipomatosis, normal faecal elastase or a normal skeletal survey do not exclude the diagnosis; classical criteria would miss or delay diagnosis in a significant proportion of patients [37]. Absence of one of the three major components (malabsorption, metaphyseal abnormalities, neutropenia) does not exclude SDS [2].

Sources: [1, 24, 25, 29, 37, 54].

The French PNDS (French National Diagnosis and Care Protocol) organises suspicion by sign and requires, for the less specific signs, the association with exocrine pancreatic insufficiency and neutropenia [9]. The typical neutropenia is moderate, rarely with maturation arrest, and associated with involvement of other organs; anaemia is moderate (usually above 9 g/dL) and thrombocytopenia lies between 20 and 150 × 10⁹/L; the diarrhoea due to pancreatic insufficiency is steatorrhoea (faecal fat above 5 g/24 hours on a controlled diet) with fat-soluble vitamin deficiency; complete neonatal thoracic dystrophy with respiratory distress affects about 5 % of patients, but moderate forms of pectus carinatum are frequent, with Jeune syndrome as a differential diagnosis; and MDS before 40 years of age, especially with a complex karyotype and a TP53 variant, should prompt consideration of SDS alongside other predispositions, such as GATA2 deficiency [9]. Intrauterine growth restriction (30 %), early diffuse dermatosis (ichthyosis), short stature with metaphyseal dysplasia, vitamin deficiency and developmental delay or autism spectrum disorder are only meaningful when associated with pancreatic insufficiency and neutropenia [9]. In symptomatic relatives of an index case, any minor sign justifies genetic testing [9].

For the child with chronic neutropenia of unexplained cause, the PNDS on chronic neutropenias proposes a score, validated in children, that estimates the probability of congenital neutropenia and helps to distinguish it from primary autoimmune neutropenia, which is profound but mildly symptomatic [23].

Source: [23]. Sum of points: −2 to 0, no risk; 1 to 5, probability of congenital neutropenia of 21 %; 6 to 9, of 62 %; 10 or more, of 100 %.

Suspicion in adulthood warrants specific mention. In a study from the Center for International Blood and Marrow Transplant Research (CIBMTR), 4 % of young people transplanted for MDS had SDS, and the authors of the North American transplantation series recommend considering cryptic SDS in patients with MDS before 40 years and testing family donors before selection [41, 54]. The AACR update, citing the National Comprehensive Cancer Network (NCCN), reports the recommendation to test for hereditary predisposition to haematopoietic neoplasia in all patients with MDS and adults under 50 years with AML [20]. The case of Yang and colleagues, with molecular confirmation 29 years after clinical diagnosis, illustrates the similarity to cystic fibrosis in the respiratory clinic [25].

4.2. Warning signs and emergencies

Fever in a neutropenic patient, or any acute symptom, requires urgent assessment; severe neutropenia or severe infection warrants hospital admission and broad-spectrum intravenous antibiotics [10]. Haemoglobin below 7 g/dL or symptomatic anaemia and platelets below 20 × 10⁹/L or symptomatic thrombocytopenia require prompt haematological assessment [10]. During follow-up, warning signs include worsening cytopenias, falling ANC despite G-CSF, progressively increasing cellularity with dyserythropoiesis, emergence of cytogenetic abnormalities other than isolated i(7q) and del(20q) (particularly monosomy 7 or a complex karyotype) and detection of biallelic TP53 alterations [10, 22, 41]. Mean corpuscular volume is not a consistent indicator of transformation [41]. In young children, respiratory distress due to a narrow thorax requires early referral to paediatric respiratory medicine and orthopaedics [1].

The emergency management annex of the French PNDS (French National Diagnosis and Care Protocol) considers pain, respiratory distress, fever, haemorrhagic syndrome and acute anaemic syndrome as emergencies: in the outpatient setting, it recommends contacting the local hospital team; in hospital, immediate symptomatic care has no particular features, and contact with the usual team or the reference centre should be made within 2 days [9]. The PNDS on chronic neutropenias defines as an emergency neutropenia associated with anaemia, thrombocytopenia or a tumour syndrome, or severe infection (pneumonia, septicaemia, deep infection, cellulitis, including perineal cellulitis, adenitis or peritonsillar abscess), with assessment within 24 hours and within 6 hours in the acute infectious setting; the depth of neutropenia is not, by itself, a criterion for emergency [23]. The management of fever is described in 7.4.

4.3. Newborn screening

There is no newborn screening directed at SDS in the sources consulted. Immunoreactive trypsin, used in cystic fibrosis screening, may be abnormal in SDS, as in the case of Yang and colleagues, but has not been evaluated as a screening test for this disease [25]. A Belgian genomic newborn screening study (Baby Detect, n=6824) appears in the trial search for SDS, but the sources consulted do not confirm whether the panel includes SDS genes [46]. Faecal elastase is physiologically low in meconium and reaches normal values only on the third day in term newborns and at two weeks in those born before 28 weeks, limiting its interpretation in the first days [55].

5. Diagnosis: Criteria, Tests and Algorithms

5.1. Diagnostic criteria

GeneReviews considers that no published consensus clinical criteria exist and proposes suggestive findings and an operational definition: clinical diagnosis requires exocrine pancreatic dysfunction (low serum trypsinogen below 3 years, low isoamylase above 3 years and/or low faecal elastase) and marrow dysfunction (cytopenia, macrocytosis and/or characteristic marrow findings); molecular diagnosis requires biallelic pathogenic or likely pathogenic variants in SBDS, DNAJC21 or EFL1, or a heterozygous variant in SRP54 [1]. Neither clinical branch is sufficient alone, and no biochemical assay is independently confirmatory [5]. The 2011 consensus formalised more detailed criteria, which remain the clinical reference [10].

Source: [10].

The French PNDS (French National Diagnosis and Care Protocols) define neutropenia as ANC <1.5 × 10⁹/L, profound below 0.5 × 10⁹/L and chronic when it persists beyond 3 months, and reserve the term “severe” for neutropenia with recurrent or life-threatening infections [23]. Diagnostic confirmation of SDS requires ACMG class 4 or 5 variants (likely pathogenic or pathogenic), in the homozygous or compound heterozygous state in SBDS or another associated gene; variants of uncertain significance appear in the report only when an interest has been identified with the requesting clinician [9, 23].

Suggestive findings listed by GeneReviews are, in the marrow domain, neutropenia (ANC <1500/mm³) on at least two occasions, unexplained anaemia or macrocytosis, thrombocytopenia (<150,000/mm³) on at least two occasions, hypocellular marrow for age, myelodysplasia, leukaemia and characteristic cytogenetic abnormalities (del(20q11), monosomy 7, isochromosome 7); in the pancreatic domain, trypsinogen or isoamylase low for age, low faecal elastase, pancreatic lipomatosis or abnormal faecal fat balance after exclusion of mucosal disease and cholestasis; and, as ancillary findings, proportionate postnatal short stature, skeletal anomalies, congenital heart disease, ear malformations or deafness, eczematous rash or ichthyosis and hepatomegaly with or without elevated aminotransferases [1].

5.2. Pancreatic assessment

The choice of marker depends on age. Serum cationic trypsinogen is below 6 µg/L in patients with pancreatic insufficiency and is generally 6 µg/L or higher in pancreatic-sufficient patients, being normal in about one fifth of these [10]. Pancreatic isoamylase is uniformly low in SDS at all ages, but is physiologically low in infants up to about 3 years, so it should not be used alone in this age group; a classification tree rule combining trypsinogen, isoamylase and age distinguishes patients from controls except below 3 years [10]. Hence the practical rule of using trypsinogen below 3 years and isoamylase above 3 years [1]. Faecal elastase below 200 µg/g indicates severe dysfunction and below 100 µg/g suggests maldigestion due to insufficiency; it is widely used in Europe, but has not been validated in large SDS series [10]. The French PNDS considers faecal elastase the most reliable test and the determinant of sufficient or insufficient status, with values below 200 µg/g significant and below 50 µg/g suggestive of severe exocrine pancreatic insufficiency; it states that serum immunoreactive trypsinogen is discriminating but rarely available, that serum amylase and lipase are low at all ages but also physiologically low in infants, and that imaging (ultrasound, CT or MRI) can demonstrate fatty replacement [9]. In the newborn, elastase reaches normal values only on the third day (term) or at two weeks (less than 28 weeks) [55]. When trypsinogen is low, the 72-hour faecal fat balance confirms insufficiency; when it is 6 µg/L or higher, it confirms sufficiency, with enzymes stopped at least 24 hours beforehand [1, 10]. The secretin–cholecystokinin stimulation test is a research tool, and duodenal aspiration is not recommended, as it misclassifies about 25 % of pancreatic-sufficient patients [10]. A normal sweat test excludes cystic fibrosis, and immunoglobulins are classically normal, although recent studies show low IgM in a subgroup [2, 40]. Abdominal ultrasound forms part of the initial assessment [10].

5.3. Haematological assessment

A full blood count with differential, MCV, blood film, reticulocytes and HbF, iron, folate and vitamin B12, and immunoglobulins IgG, IgA and IgM form part of the initial assessment [10]. Bone marrow aspiration and biopsy, with cellularity, differential, iron staining and cytogenetics, are essential at diagnosis, because i(7q) and del(20q) are strongly associated with SDS and because some patients have MDS or AML at their first biopsy [1, 10]. The AACR update recommends that baseline marrow assessment include aspiration, biopsy, flow cytometry and cytogenetics, ideally reviewed at a centre experienced in SDS given the baseline dysmorphia, and that a somatic gene panel including the genes involved in SDS clonal evolution be added [20]. The criteria for clonal cytogenetic abnormality adopted by the 2011 consensus are at least 2 of 20 cells with the same gain or at least 3 with the same loss on G-banding, or FISH above reference values; paediatric MDS is defined by 2 of 3 of chronic trilineage or prominent bilineage cytopenia, a clonal abnormality and 5 to 29 % blasts, and AML by at least 20 % blasts (World Health Organization) [10]. Severe aplastic anaemia is defined by 2 of 3 criteria (reticulocytes <40,000/µL, platelets <20,000/µL, neutrophils <500/µL) with a hypocellular biopsy (<25 %, or <50 % with less than 30 % haematopoietic cells) without significant fibrosis; the moderate form, by at least two low counts (reticulocytes <40,000, platelets <40,000, neutrophils <1500) with hypocellular marrow [10]. Systematic assessment of chemotaxis is not necessary in routine follow-up [10]. The French PNDS adds, at diagnosis, a marrow aspirate with marrow cytogenetics (karyotype and FISH), a next-generation sequencing (NGS) myeloid panel and cryopreservation of marrow cells, immunoglobulins G, A and M, vaccine serologies, T, B and NK lymphocyte immunophenotyping, and coagulation tests in the presence of bleeding [9].

The PNDS on bone marrow failure uses a modified version of the Camitta criteria for severity: severe aplasia with haematopoietic cellularity below 30 % on biopsy and at least 2 of 3 peripheral criteria (platelets <20 × 10⁹/L, neutrophils <0.5 × 10⁹/L, reticulocytes <60 × 10⁹/L on automated count, the threshold adopted in the RACE trial instead of the 20 × 10⁹/L of manual counting), and very severe aplasia when neutrophils are below 0.2 × 10⁹/L, according to the EBMT [3]. These thresholds do not coincide with those of the 2011 consensus (reticulocytes <40,000/µL, cellularity <25 %), and the difference in reticulocytes results mainly from the counting method (see 7.15) [3, 10]. The same PNDS recalls three points for reading the marrow karyotype: del(20q), trisomy 8 and loss of Y remain compatible with aplasia; the poverty of the marrow causes the karyotype to fail fairly often; and interphase FISH for chromosomes 7 and 8 complements it in this situation [3]. Bone marrow biopsy is necessary to measure cellularity and exclude myelofibrosis or infiltration, but is not systematic in the constitutional forms of childhood [3].

5.4. Molecular strategy

GeneReviews proposes two approaches [1]. In the first, a multigene panel with DNAJC21, EFL1, SBDS, SRP54 and genes of the differential diagnosis; if negative, methods for complex SBDS/SBDSP1 rearrangements are used (targeted sequencing, single nucleotide polymorphism arrays, targeted deletion and duplication analysis, microarray). In the second, used when the phenotype is atypical, exome or genome sequencing [1]. In both cases, the complex allele c.[183_184delinsCT;258+2T>C] requires parental testing to define phase [1]. When biallelic variants are not found, proteogenomic analysis can identify reduced SBDS expression [1]. The 2011 consensus advised SBDS testing in almost all suspected cases, including those without clinical pancreatic insufficiency [10]. The NCI cohort recommends looking for structural variants whenever only one variant is identified [24]. The French PNDS recommends sampling the suspected case and both parents, to confirm that the two variants are on different alleles and to define the risk in the case of a de novo variant [9]. Both PNDS designate the reference laboratory of the chronic neutropenia centre (Pitié-Salpêtrière Hospital) for constitutional diagnosis and a reference laboratory (Saint-Antoine Hospital) for somatic molecular follow-up, recommend that results from laboratories outside the network be validated at a national RCP and, if the work-up is inconclusive, extension by genome-wide analysis within the France Genomic Medicine Plan 2025 [9, 23].

5.5. Limitations of genetic tests

The SBDSP1 pseudogene is the main source of error. A panel not designed for SDS or an exome may fail to detect common variants because short reads are mismapped [1]. Yamada and colleagues described two Japanese compound heterozygous patients in whom exome sequencing detected c.258+2T>C at an allele fraction of approximately 0.85, mimicking homozygosity, and did not detect the c.183_184delinsCT allele (with c.201A>G), which was assigned to the pseudogene; PCR with SBDS-specific primers or RNA sequencing resolved the diagnosis [32]. In the two prenatal cases from Brussels, initial tests were inconclusive owing to the same interference [47]. Designing multiplex ligation-dependent probe amplification (MLPA) probes, quantitative PCR or targeted microarray is complicated by homology [1]. Large SBDS deletions escape sequencing and may require microarray, quantitative PCR or long-read sequencing [24, 28].

The GeneReviews detection table (sequencing >90 % and deletion/duplication <2 % for SBDS; 100 % by sequencing for EFL1 and SRP54; <90 % for DNAJC21) expresses the proportion of known variants detectable by each method, not the clinical sensitivity of the test in a patient [1]. Approximately 8 % of pathogenic SBDS variants are not detected by either method, and fewer than 10 % of patients with a clinical diagnosis have no identified cause, so a negative result does not exclude SDS when the phenotype is compatible [1]. When interpreting apparent homozygosity, an undetected deletion on the other allele (false homozygosity) or uniparental isodisomy must be considered [1]. Variants of uncertain significance neither confirm nor exclude the diagnosis [1].

5.6. Report content and reclassification

The report should identify the gene, versioned transcript (for SBDS, NM_016038.4; for SRP54, NM_003136.4), variant in HGVS nomenclature with its historical designation where available, zygosity and phase (cis or trans, established by parental testing), classification according to the criteria of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP), and the method used to circumvent the pseudogene [1, 31]. The variant c.523C>T p.(Arg175Trp), present in the most severely affected patient in the North American Registry, has conflicting classifications in ClinVar, illustrating the need for periodic reassessment [28, 33]. For a variant of uncertain significance, family segregation and phenotypic assessment of siblings (pancreatic function and full blood count) aid interpretation [1].

5.7. Prenatal and preimplantation diagnosis

Once familial variants have been identified, prenatal diagnosis and preimplantation genetic testing are possible [1]. In fetuses without a family history, ultrasound signs are skeletal and non-specific: a small thorax with short ribs suggestive of thoracic dysplasia, or merely slightly straightened ribs [47, 48]. In this setting, fetal exome sequencing may miss the converted allele, and phenotypic reassessment with molecular analysis targeted to the pseudogene enabled the diagnosis in the Brussels cases [47].

5.8. Diagnostic algorithm

Sources: [1, 10, 20, 24].

6. Differential Diagnosis

6.1. Comparative table

Sources: [1–3, 9, 10, 15, 21, 23, 55–59].

6.2. Pancreatic insufficiency: cystic fibrosis, Pearson, Johanson–Blizzard and isolated forms

Cystic fibrosis is the most common cause of hereditary pancreatic insufficiency and should be excluded in all patients by sweat testing [2, 10]. Confusion does occur: Lozada-Muñoz and Aliaga described SDS as “the clinical mimic of cystic fibrosis” in relation to an autopsy in Puerto Rico, and the patient of Yang and colleagues had a borderline sweat test (sodium 43 and 48 mEq/L) and received genetic confirmation only decades later [25, 60]. In SDS, ductular electrolyte secretion is normal, so the sweat test is normal, and chronic suppurative lung disease is not characteristic [10, 25]. Pearson syndrome combines pancreatic insufficiency and cytopenia, but the marrow shows ring sideroblasts and vacuolation of erythroid and myeloid precursors, with generally normal cellularity; both the metabolic diseases textbook and GeneReviews place it alongside SDS in the differential diagnosis of infants with pancreatic insufficiency and neutropenia or pancytopenia [1, 10, 21]. Johanson–Blizzard syndrome has pancreatic insufficiency without haematological abnormalities, with hypoplastic nasal alae, scalp defects, spiky hair and frequent learning difficulties [1, 55]. Isolated exocrine pancreatic insufficiency due to biallelic SPINK1 variants does not feature neutropenia [1].

6.3. Bone marrow failure and congenital neutropenias

Other inherited bone marrow failure syndromes share hypocellularity and MDS/AML risk, but do not have primary exocrine pancreatic dysfunction. The 2011 consensus recommends chromosome breakage testing with diepoxybutane or mitomycin C to exclude Fanconi anaemia unless the diagnosis is obvious, and exclusion of dyskeratosis congenita and other marrow failure disorders; virology (Epstein–Barr virus, cytomegalovirus, parvovirus B19) is undertaken when indicated [10]. Diamond–Blackfan anaemia, the most frequent ribosomopathy, is dominated by macrocytic erythroid aplasia, with congenital anomalies in 30 to 50 % and a risk of osteosarcoma, which has not been described in SDS [15, 24]. Congenital neutropenias due to ELANE and HAX1 have no pancreatic component, and SRP54-related neutropenia is discussed as a form of SDS itself, with a spectrum ranging from isolated neutropenia to the syndromic presentation [1, 4]. In glycogen storage disease type Ib, neutrophils show apoptosis (caspase activation, condensed nuclei, perinuclear mitochondria), a finding not observed in SDS [59]. Drug- or infection-associated neutropenia is transient [1]. The PNDS (French National Diagnosis and Care Protocol) for chronic neutropenias notes that transient neutropenia (lasting less than 3 months) is mainly post-viral or drug-induced and that, in isolated neutropenia without severe infection, the course should be monitored for at least 3 months before invasive investigations; benign chronic neutropenia associated with the Duffy-null red cell phenotype, common in people of African or Middle Eastern origin, is free of infections and oral involvement and does not warrant further investigation [23]. In children, the initial work-up includes immunoglobulins, lymphocyte immunophenotyping and antineutrophil antibodies, to identify immune deficiencies and autoimmune neutropenia, and in adults large granular lymphocyte proliferation and autoimmunity are also sought [23].

SDS may also come to diagnosis through presentation with aplastic anaemia. The French PNDS for aplastic anaemia estimates that 15 to 20 % of cases are constitutional, a higher proportion in children and adolescents, against an overall incidence of about 2 to 3 cases per million inhabitants per year in Europe and the USA, and places SDS among the constitutional disorders that only occasionally present as aplastic anaemia, alongside GATA2 deficiency and Diamond–Blackfan anaemia [3]. Clues pointing to SDS are intrauterine and postnatal growth restriction, metaphyseal changes, fatty diarrhoea and abdominal pain; a blood count already abnormal for more than 6 months, macrocytosis, mild dysplasia and HbF above 5 % before any transfusion favour a constitutional form, whereas a recent normal blood count, very rapid deterioration, recent hepatitis or a paroxysmal nocturnal haemoglobinuria (PNH) clone of 1 % or more favour the acquired form [3]. Below 18 years of age, chromosome breakage testing and lymphocyte telomere length measurement by flow-FISH are systematic; telomeres below the 1st centile suggest a telomere biology disorder, but are also seen in SDS, Fanconi anaemia, GATA2 deficiency and idiopathic aplastic anaemia, so a short telomere does not exclude SDS [3]. The constitutional bone marrow failure gene panel, which includes SBDS, SRP54 and DNAJC21, is recommended systematically before 10 years of age and case by case thereafter, preferably on fibroblasts obtained by skin biopsy, because reversion or clonal evolution in the marrow may produce false-negative results in blood; for SDS, fat-soluble vitamins, lipase, faecal elastase and, if in doubt, pancreatic magnetic resonance imaging are added [3]. In the absence of evidence for a constitutional form, specific treatment of aplastic anaemia in a child or young adult should not be delayed more than 6 weeks pending these results, and, if conventional genetic testing is negative despite strong clinical suspicion or failure of immunosuppression in a child, genome sequencing of fibroblast DNA within the France Genomic Medicine Plan is discussed at a national multidisciplinary team meeting (RCP) [3]. This distinction has direct therapeutic consequences, because constitutional aplastic anaemias are not an indication for immunosuppression (see 7.4).

6.4. Skeletal dysplasias

In skeletal dysplasia tables, SDS appears as metaphyseal chondrodysplasia with delayed bone maturation, pancreatic insufficiency and cytopenias, associated with SBDS [61]. In cartilage–hair hypoplasia (described chiefly in the Amish population), metaphyseal lesions are generally more severe, with fine hair, lymphopenia and cellular immunodeficiency, and no pancreatic insufficiency [10, 57]. Schmid metaphyseal dysplasia has more severe metaphyseal abnormalities and characteristic coxa vara, without neutropenia or malabsorption, although coxa vara can occasionally occur in SDS [2, 56]. In infants, Sedaghatian-type spondylometaphyseal dysplasia (GPX4, lethal within the first days) may have similar abnormalities and require molecular testing [58]. Spahr-type metaphyseal dysplasia has minimal abnormalities without the clinical signs of SDS; EXTL3 deficiency and Omenn syndrome have combined immunodeficiency without pancreatic insufficiency; asphyxiating thoracic dysplasia and achondroplasia may cause confusion because of the narrow thorax, but both have a trident acetabulum [2]. The atlas observes that classical SDS does not present with platyspondyly, unlike axial spondylometaphyseal dysplasia, but the severe neonatal SBDS-related form can have it [7, 62].

7. Treatments and Risk Management Strategies

7.1. General principles and organisation of care

There is no authorised targeted therapy for SDS, and treatment is supportive, preventive and, for severe haematological complications, curative through HSCT [1, 63]. GeneReviews recommends a multidisciplinary team including haematology, gastroenterology, clinical genetics, orthopaedics, endocrinology, immunology, dentistry, child development, psychology and social work [1]. Recommendations are based on the 2011 consensus, which the authors themselves acknowledge relies on frequently anecdotal data, supplemented by the 2022 EBMT consensus for transplantation and the 2024 AACR update for haematological surveillance [10, 11, 20]. The French PNDS defines treatment as day-to-day multidisciplinary care of the patient and family, with six objectives: compensating for pancreatic insufficiency with pancreatic enzymes and vitamins (mainly A, E and D), treating infectious episodes promptly, preventing infections (prolonged antibiotics or G-CSF, when indicated, and extended vaccination), ensuring educational and occupational integration, detecting haematological complications early by blood count and NGS, and treating these complications, which are indications for HSCT, with a collegial decision at a multidisciplinary team meeting (RCP) [9]. The care network includes general practice, paediatrics, haematology and oncology, gastroenterology, hepatology, endocrinology, orthopaedics, physical and rehabilitation medicine, respiratory medicine, neurology, child psychiatry, dermatology, otorhinolaryngology, stomatology and dentistry, radiology, nursing, physiotherapy, psychology, social work and school support assistants, with recourse to the national RCP for difficult decisions [9].

Sources: [1, 10].

7.2. Pancreatic enzyme replacement therapy

Pancrelipase or pancreatin preparations (for example, Creon® in the USA and Kreon® in Europe) contain porcine-derived lipase, protease and amylase in gastro-resistant microspheres and replace the missing acinar secretion [64]. The 2011 consensus acknowledges that no SDS-specific dosing guidelines exist and recommends following those for cystic fibrosis: an initial dose of 2000 lipase units/kg body weight per day, with a maximum of 10,000 lipase units/kg/day, administered with all meals and snacks containing protein, fat or complex carbohydrates [10]. The US Creon® label (generic indication for exocrine pancreatic insufficiency, without SDS studies) defines dosing by meal and age [64]. The French PNDS gives a dosage of 5000 to 10,000 lipase units/kg/day with a gastro-resistant preparation, taken with food, and lists faltering weight gain, steatorrhoea or frequent stools, abdominal pain and persistent anorexia as indications for replacement [9]. This range starts above the initial dose of the 2011 consensus (2000 units/kg/day) and is capped by the daily maximum of the consensus and the label (10,000 units/kg/day); the guideline is the reference and the label serves as a comparison (see 7.15) [9, 10, 64].

Source: [64].

Capsules should be swallowed whole; when this is not possible, they are opened and the contents sprinkled over acidic soft food (pH ≤4.5, such as apple sauce, banana or plain yoghurt) and consumed immediately, without crushing or chewing, to avoid destroying the enteric coating and irritating the oral mucosa; the 2011 consensus permits mixing the granules with milk, juice or puréed fruit and recommends avoiding heat [10, 64]. The main adverse effects are fibrosing colonopathy, described with prolonged high doses chiefly in children with cystic fibrosis (colonic stricture with more than 6000 lipase units/kg/meal before 12 years, requiring surveillance of those receiving doses above this level and reduction if possible), hyperuricaemia at high doses, hypersensitivity to porcine proteins and the theoretical risk of porcine viral transmission [64]. Doses exceeding the maxima are justified only if efficacy is documented by faecal fat or improvement in malabsorption [64]. In the European Union, pancreatins are authorised through national procedures, without centralised assessment by the European Medicines Agency (EMA); the sources consulted do not include a national summary of product characteristics, so European label doses are not included in this article.

If malabsorption persists despite an optimised dose, an H2 antagonist or proton pump inhibitor may be added, and high requirements should prompt investigation for associated enteropathy [10]. Clinical response is generally excellent, but growth may remain limited by skeletal causes [10]. As many patients regain sufficiency with age, pancreatic status should be reassessed periodically to decide on enzyme discontinuation [10]. The PNDS places this possibility mainly from 10 years of age onwards, with faecal elastase as the key test for reassessment, and stresses that the response to enzymes is generally excellent, unlike in cystic fibrosis, where intestinal factors may coexist [9]. Bezzerri and colleagues point out that 10 to 20 % of functioning pancreatic tissue is sufficient for adequate fat absorption and suggest that the dose might be reduced after assessing steatorrhoea, a hypothesis without formal demonstration [63].

7.3. Nutrition and fat-soluble vitamins

Malnutrition present at diagnosis should be corrected by the first year of age with appropriate therapy [10]. Weight and height are measured at every visit, with dietitian assessment; behavioural feeding difficulties warrant psychological support, and reflux, food allergy and enteropathy should be excluded; oral supplements are used if intake is insufficient, and gastrostomy is the last resort, with overnight feeding and weaning when stable [10]. Vitamins A, D and E and prothrombin time (an indicator of vitamin K deficiency) are measured initially, repeated approximately one month after starting enzymes and then at least once to twice yearly (every 6–12 months in young children), with supplementation if they remain low [10]. The French PNDS adds that about 15 % of patients, particularly in the first months, depend on nutritional support, especially continuous enteral nutrition, because of persistent anorexia, and that 10 % of patients in the French series had a gastrostomy [9]. In a French study published as an abstract (124 patients with SDS compared with 171 with cystic fibrosis; 807 samples taken in a stable phase, all on enzymes and vitamins A, D and E), median values were 250 µg/L for vitamin A, 26.7 ng/mL for 25-hydroxyvitamin D and 8.15 mg/L for vitamin E; with thresholds of 400 µg/L, 10 ng/mL and 8 mg/L, vitamin A deficiency was present in 95 %, vitamin E deficiency in 49 % and vitamin D deficiency in 24 % of patients, and vitamins A and E were lower than in cystic fibrosis, despite a median vitamin A intake of 259 IU/kg/day [9]. Vitamin A correlated with neutrophils and vitamin E with haemoglobin and platelets; retinol-binding protein (RBP), measured in a few patients, had a median of 21 mg/L and a vitamin A/RBP molar ratio of 0.98, which did not allow a transporter deficiency to be distinguished from insufficient hepatic stores [9]. The PNDS therefore recommends always measuring vitamin A together with RBP, considers that cystic fibrosis vitamin A regimens do not apply to SDS and refers to the general intake recommendations, from about 350 µg retinol equivalents per day in infants (about 1100 IU, or 120 IU/kg/day) to 700 µg/day in adults (about 2500 IU, or 40 IU/kg/day) [9]. In 21 French patients whose supplementation was analysed, the median was 459 IU/kg/day (137.8 µg/kg/day), which for a 60 kg adult corresponds to about 8270 µg/day, far above recommended intakes, without normalising values; the risk of hypervitaminosis A, with one case of cirrhosis attributed to this mechanism, requires careful consideration of the dose [9]. For vitamins D, E and K, supplementation raises no particular difficulty and follows the general recommendations, started if values are low and accompanied by checking adherence to enzymes, because deficiencies are an indirect marker of malabsorption [9]. 25-hydroxyvitamin D and parathyroid hormone should remain normal, with calcium and vitamin D if intake is insufficient [10].

7.4. Non-malignant haematological complications

Granulocyte colony-stimulating factor. Filgrastim (recombinant G-CSF; Neupogen® in the USA and biosimilars such as Zarzio® in the EU) stimulates proliferation and differentiation of neutrophil precursors. Most patients with SDS do not need G-CSF because infection incidence is low; the 2011 consensus reserves chronic use for patients with recurrent invasive bacterial or fungal infections and severe neutropenia, allows intermittent low-dose regimens (“for example 2–3 µg/kg every 3 days”) or higher continuous doses, and defines the goal as preventing infections rather than normalising counts; G-CSF may be considered during infections [10]. GeneReviews considers it for recurrent severe infections with persistently ≤500/mm³ ANC, and before complex dental procedures or orthopaedic surgery [1]. Labels define doses for severe congenital neutropenia in general, which greatly exceed those used in SDS. The French PNDS accepts prolonged G-CSF only for recurrent infections or severe chronic stomatitis with profound neutropenia (ANC <0.5 × 10⁹/L) and adopts the intermittent low-dose regimen of the consensus (for example, 2–3 µg/kg every 3 days), with the aim of preventing infections rather than normalising the blood count [9]. The PNDS for chronic neutropenias describes the general regimen for congenital neutropenias, which serves as a comparison: an initial dose of 5 µg/kg/day for 14 days, adjustment in steps of 3–5 µg/kg according to clinical response, assessed after 2–3 weeks, spacing of injections (for example, on alternate days) if there are no infections or aphthous ulcers, and use of the lowest effective dose in a continuous, on-demand (less than 10 days) or intermittent regimen; once the dose is stable, a blood count every 6 months is sufficient [23]. It sets no maximum dose, but defines as a non-responder a patient who does not reach 1 × 10⁹/L with more than 50 µg/kg/day and as a high-threshold responder one who needs 10 to 50 µg/kg/day; prolonged exposure to more than 10 µg/kg/day increases leukaemic risk, and prolonged use above 15 µg/kg/day has become an indication for HSCT in ELANE-related neutropenia [23]. Filgrastim and lenograstim (Granocyte®) have equivalent clinical efficacy; pegfilgrastim (Neulasta®) is not recommended, because of the instability of its effect and grade 3–4 adverse effects; with bone pain, myalgia or headache, the dose should be reduced [23].

Sources: [10, 26, 65, 66].

The initial total daily dose is identical in the two labels (12 µg/kg/day), with different fractionation: two administrations in the USA, a single or divided dose in the EU [65, 66]. Neither label sets a ceiling: the US label mentions patients with congenital neutropenia who rarely required ≥100 µg/kg/day, and the European label states that 97 % of responders had a complete response at ≤24 µg/kg/day and that long-term safety above this value has not been established [65, 66]. The practical consequence is that, in SDS, dosing should be titrated from low levels towards the clinical goal, as shown by the 25 patients in the European registry on prolonged therapy (started at a median age of 2.48 years, median duration 2.87 years), all responders with a median ANC of 1394/µL [26]. The most frequent adverse effects are fever, musculoskeletal pain (bone pain, back pain, arthralgia, myalgia), anaemia, vomiting and nausea; serious effects include anaphylaxis, acute respiratory distress syndrome, capillary leak syndrome, splenomegaly and splenic rupture (pain in the left upper quadrant or shoulder should prompt assessment; in severe chronic neutropenia, 3 % of patients required splenectomy), glomerulonephritis, thrombocytopenia and large-vessel arteritis [65, 66]. In continuous therapy for more than 6 months in patients with underlying osteoporosis, the European label suggests monitoring bone density, which is relevant to SDS [66]. G-CSF should not be administered in the presence of leukaemia or leukaemic evolution [66].

Leukaemic risk and G-CSF. GeneReviews recommends caution with prolonged use because of a possible contribution to leukaemic transformation, extrapolated from data on severe congenital neutropenia under prolonged G-CSF [1]. Labels state that MDS and AML occur in the natural history of congenital neutropenia and in treated patients, with approximately 3 % MDS/leukaemia in trials and 12 % new cytogenetic abnormalities in patients with a normal baseline karyotype, and that with abnormal cytogenetics or myelodysplasia the risks and benefits of continuation should be weighed [65, 66]. SDS-specific data show no association: in the European registry, only 2 of the 10 patients with MDS, AML or ALL were on prolonged G-CSF at diagnosis and 7 had never received it, none of those treated developed somatic CSF3R mutations, and the authors conclude that G-CSF does not increase risk, in a retrospective observational analysis [26]. In the North American Registry, 45 % of patients with biallelic SBDS used G-CSF, and 35 % of MDS and 10 % of AML arose during G-CSF, in a design that cannot assess the effect on risk [41]. The 2011 consensus records that transformation during G-CSF has been described without demonstrated causality [10]. The PNDS for chronic neutropenias confirms that a leukaemogenic role of G-CSF has only been demonstrated in patients dependent on high continuous doses, mainly in ELANE-related neutropenia, and not in SDS, in which G-CSF is rarely needed [23]. In SRP54-related disease, response to G-CSF is poor, and one infant with p.Thr117del remained without transformation during continuous G-CSF [4, 52].

Infections. Fever or acute symptoms require urgent assessment; some patients can be treated with oral antibiotics, but severe neutropenia or severe infection warrants hospital admission and broad-spectrum intravenous antibiotics [10]. Immunoglobulins and post-vaccination antibodies are assessed for recurrent infections, and baseline immunological assessment (lymphocyte subsets and immunoglobulins) is recommended by Gloude and colleagues in all patients [10, 40]. Five patients in Gloude's study received immunoglobulin replacement, three of whom had low IgG, and in the NCI cohort 45.8 % received intermittent or chronic intravenous immunoglobulin [24, 40].

The French PNDS documents set out the management in detail. Antibiotic therapy for fever is not systematic and depends on the clinical picture, but signs of severity (altered consciousness or haemodynamic instability) or an infectious focus (cellulitis, pneumonia, colitis) require admission and empirical broad-spectrum intravenous antibiotic therapy, as in post-chemotherapy febrile neutropenia, covering Gram-positive cocci (Staphylococcus aureus, oral streptococci) and Gram-negative bacilli including Pseudomonas aeruginosa: an antipseudomonal β-lactam (cefepime or piperacillin + tazobactam, the latter preferred if there is an oral or gastrointestinal focus), an aminoglycoside added in septic shock, and adaptation if there is colonisation by meticillin-resistant S. aureus or multidrug-resistant bacteria [23]. During the episode, G-CSF is prescribed at the patient's usual dose or, failing that, at 5 µg/kg/day; granulocyte transfusion is discussed in septicaemia or uncontrolled aggressive local infection; protective isolation is of no benefit, and normalisation of neutrophils should not be awaited before discharge or stopping antibiotics [23]. Intramuscular injections and rectal temperature measurement are avoided, and there are no restrictions on diet or communal life [23].

Antibiotic prophylaxis and vaccination. In the absence of efficacy studies, the PNDS for chronic neutropenias allows, in patients at highest risk of bacterial infection, prophylaxis with oral sulfamethoxazole + trimethoprim (co-trimoxazole; Bactrim® in France), 10–30 mg/kg/day of sulfamethoxazole in 1 or 2 doses in children and 1 tablet of 800 mg sulfamethoxazole daily in adults (grade C), with no effect on oral and dental infections; the dosage is extrapolated by the authors from chronic granulomatous disease, and antifungal or antiviral prophylaxis is not recommended outside WHIM syndrome and GATA2 deficiency [23]. In SDS, most patients need neither antibiotic prophylaxis nor G-CSF [9]. Vaccination should be extended, because influenza, varicella and measles can be very severe: a complete vaccination schedule with boosters, influenza and COVID-19 vaccines, and live attenuated vaccines are not contraindicated, except in severe haematological complications [9]. The PNDS for chronic neutropenias, following the French High Council for Public Health, adds an extra dose of pneumococcal conjugate vaccine in patients at risk of severe bacterial infections, vaccines against SARS-CoV-2, respiratory syncytial virus and varicella, and does not contraindicate BCG (bacillus Calmette–Guérin) in congenital neutropenias, all grade C [23]. The PNDS for aplastic anaemia, for the aplastic phase, defines fever as a temperature above 38.5 °C once or above 38 °C twice 4 hours apart and requires, with neutrophils <0.5 × 10⁹/L, urgent admission and intravenous antibiotic therapy, with intensive care assessment in haemodynamic instability; without fever, admission is not necessary [3]. In constitutional forms, in which fungal risk appears lower than in acquired aplastic anaemia, antifungal prophylaxis (posaconazole, voriconazole or itraconazole) is decided case by case according to the severity of neutropenia; it recommends vaccines against pneumococcus, meningococcus, Haemophilus influenzae and influenza if not already given; and it considers G-CSF useful only in short courses in severe infection or to facilitate discharge, and erythropoietin of no use outside chronic kidney disease [3].

Transfusions and bleeding. Chronic transfusions use irradiated products, with iron chelation when indicated [10]. The PNDS for aplastic anaemia, which covers constitutional forms, adapts the red cell threshold (7 to 9 g/dL) to age, tolerance, comorbidities and quality of life, with systematic transfusion below 7 g/dL; it transfuses platelets for bleeding or, prophylactically, below 20 × 10⁹/L in adults and 10 to 20 × 10⁹/L in children, a threshold that may be lowered to 10 × 10⁹/L in regularly transfused patients with few symptoms [3]. Red cell concentrates should be phenotyped (Rh and Kell) to prevent alloimmunisation and irradiated only after antilymphocyte globulin or from the start of HSCT conditioning; platelet concentrates are no longer irradiated nor necessarily from apheresis; and the number of platelet transfusions should be limited when HSCT is anticipated, with anti-HLA antibody testing in the event of a nil transfusion yield or in high-risk patients, such as multiparous women [3]. This practice differs from the systematic irradiation of products proposed by the 2011 consensus (see 7.15) [3, 10]. Mild to moderate bleeding is managed with local measures (for example, 0.05 % nasal xylometazoline), tranexamic acid or aminocaproic acid, vitamin K if deficient, platelets for severe bleeding and prophylactic platelet transfusion below 10 × 10⁹/L or with a bleeding tendency; plasma or prothrombin complex concentrate for coagulopathy, and hormonal suppression of heavy menstrual bleeding [10]. The sources consulted do not state doses of these drugs for SDS.

Iron overload and chelation. SDS-specific sources only mention chelation when indicated; the PNDS for aplastic anaemia, applicable to transfused constitutional forms, defines screening and doses, which are applied here by extrapolation [3, 10]. Iron overload impairs haematopoiesis and, potentially, morbidity and mortality after HSCT; it is monitored with ferritin and transferrin saturation every 2 to 3 months depending on the transfusion rate (every 3 months in the appendix of the same PNDS), sampled immediately before a transfusion and preferably fasting, bearing in mind that saturation above 60 % indicates circulating free iron, which is more toxic to organs [3]. Chelation is indicated with ferritin above 500 µg/L or after more than 20 red cell transfusions, and hepatic and cardiac magnetic resonance imaging quantifies tissue overload when ferritin exceeds 1000 µg/L [3]. No chelator is optimal in aplastic anaemia, and there is no validated recommendation on the choice: subcutaneous desferrioxamine is the first option; deferasirox carries greater renal toxicity in combination with ciclosporin, rare cases of marrow toxicity and hepatic toxicity; deferiprone, because of the risk of agranulocytosis and neutropenia, should be avoided, which weighs particularly in a disease with underlying neutropenia such as SDS [3]. Monitoring of renal function and blood count is required, and, in massive overload before HSCT, combination chelation is acceptable [3]. After remission or HSCT, venesection replaces chelators, because of its efficacy and absence of adverse effects [3].

Sources: [3, 67, 68]. The PNDS starting threshold (ferritin >500 µg/L or more than 20 transfusions) is earlier than that of the US deferasirox label (at least 100 mL/kg of red cell concentrate and ferritin persistently above 1000 µg/L); the guideline is the reference, and the label serves as a comparison [3, 67].

Immunosuppression. The combination of antilymphocyte globulin and ciclosporin, with or without eltrombopag, is the treatment of immune-mediated acquired aplastic anaemia; in constitutional aplastic anaemias, including SDS, the PNDS for aplastic anaemia considers that there is no indication for immunosuppressive treatment, and severe bone marrow failure follows the HSCT pathway (see 7.5) [3].

Androgens. Data are scarce; they are not recommended first line because of potentially greater hepatotoxicity, and are reserved for bone marrow failure without a donor [10]. The PNDS for aplastic anaemia reserves danazol for refractory or relapsing acquired aplastic anaemia, Fanconi anaemia without a donor and telomere biology disorders, and does not extend it to SDS; when used, it requires annual hepatic imaging after more than one year of treatment, because of the risk of liver tumours, which persists after discontinuation [3].

7.5. Haematopoietic stem cell transplantation: indications

HSCT is the only curative therapy for bone marrow failure and malignant transformation, and the EBMT Severe Aplastic Anaemia Working Party (SAAWP) consensus estimates that 10 to 20 % of patients with SDS will require it [11]. The 2011 consensus defines three indications: severe cytopenia (haemoglobin <7 g/dL, neutrophils <0.5 × 10⁹/L with recurrent infections, platelets <20 × 10⁹/L), MDS with excess blasts and overt leukaemia [10]. As aplasia may be transient (less than 3 months), an observation period of up to 3 months is allowed while human leucocyte antigen (HLA) typing and donor search proceed; in leukaemia, donor search begins at diagnosis [10]. GeneReviews adds that marrow abnormalities are not treated with HSCT unless there is severe aplasia, myelodysplastic changes or leukaemic transformation, and the North American Registry shows that cellularity alone is not an indication for transplantation [1, 28]. Preventive HSCT before any complication is not recommended because only a subgroup progresses [54]. Any family donor must be assessed for SDS to avoid selecting an asymptomatic affected sibling [1, 54].

The French PNDS, which reproduces the recommendations of the EBMT consensus (all grade B, except one expert opinion), estimates that HSCT is needed in at most 30 to 40 % of patients and that the aim of surveillance is to bring them to transplantation in the pre-leukaemic phase, before cytogenetic aberrations of poor prognosis [9]. According to that consensus, the donor search should be activated in the presence of progressive erythroid or platelet cytopenia, transfusion dependence, recurrent infections or unfavourable clonal evolution; severe cytopenia without a clone is a formal indication for HSCT after a few weeks of observation during which the cytopenia may recover spontaneously; HSCT should precede leukaemia or overt MDS; and TP53 and EIF6 testing should be performed, although its use in decision-making has not yet been validated [9]. Regarding the donor, it recommends genetic and HLA testing of parents and siblings from diagnosis; accepts a healthy relative as donor, whether or not a heterozygous carrier; prefers bone marrow to peripheral blood; chooses first an HLA-identical related donor, then a 10/10 unrelated donor (HLA-A, -B, -C, -DRB1, -DQB1) and, second-line, 9/10; accepts unrelated cord blood with more than 3.7 × 10⁷ nucleated cells/kg and 8/8 or 7/8 matching; and reserves haploidentical donors or donors with two or more mismatches mainly for clinical emergencies [9].

7.6. Transplantation: outcomes, conditioning and divergences

Sources: [26, 41, 54, 69, 70]. GVHD, graft-versus-host disease.

In the CIBMTR series, patients transplanted for bone marrow failure (median age 7 years) achieved neutrophil recovery at a median of 15 days and platelet recovery at 26 days; there was one primary and two secondary graft failures, all fatal after a second transplant, grade II–IV acute GVHD in 26 % and chronic GVHD in 13 % at 5 years [54]. Five-year chronic GVHD was 14 % with marrow versus 27 % with peripheral blood, and the authors recommend avoiding peripheral blood as the source [54]. In the meta-analysis, for myeloid neoplasia, pooled relapse was 25.8 % and 3-year non-relapse mortality 52.6 % [70].

Conditioning toxicity is the central problem. GeneReviews states that cyclophosphamide and busulfan may be unsuitable because of possible cardiac toxicity, and that substitution with fludarabine, treosulfan and melphalan, with cautious myeloablation or reduced intensity, yielded better results in small cohorts [1]. The 2011 consensus, reviewing 36 patients treated with chemotherapy or HSCT, found 83 % deaths from complications (prolonged aplasia, infections, cardiotoxicity, neurological and renal complications, veno-occlusive disease, lung disease, graft failure, GVHD), with survival of approximately 80 % in aplastic anaemia versus 30–40 % in MDS/AML [10]. Regimens used in the North American Registry for severe bone marrow failure were alemtuzumab, fludarabine and melphalan (4 patients), fludarabine and alemtuzumab (1) and treosulfan, fludarabine and antithymocyte globulin (1), with 2 transplant-related deaths [28]. Under the CIBMTR definition, a regimen is myeloablative with busulfan >8 mg/kg, melphalan >140 mg/m², thiotepa >10 mg/kg or fractionated total body irradiation ≥1000 cGy [54]. A phase 2 trial of treosulfan and fludarabine before HSCT in bone marrow failure syndromes (NCT04965597, 40 patients) is completed, without published results in the sources consulted [46].

The 2022 EBMT consensus recommends regular, structured haematological follow-up, reduced-intensity conditioning to decrease transplant-related mortality, limitation of total body irradiation, particularly in severe cytopenia and non-malignant bone marrow failure, early diagnosis of malignant clonal evolution and early recognition of the indication for HSCT, and acknowledges poor outcomes in AML regardless of previous cytoreductive chemotherapy [11]. The detailed recommendations, reproduced in the appendix of the French PNDS, distinguish the indications: reduced intensity in non-clonal severe cytopenia (aplastic anaemia); chemotherapy-based myeloablative conditioning in MDS/AML; no myeloablative regimen based on total body irradiation; cytoreduction before HSCT as an option of debated efficacy; and, in advanced MDS/AML with reduced stem cell reserve, sequential approaches with moderately intensive antileukaemic chemotherapy followed by reduced-intensity HSCT [9]. For GVHD prevention, calcineurin inhibitors (ciclosporin or tacrolimus), alone or with short-course methotrexate or mycophenolate mofetil, are the reference regimen for unmanipulated grafts, and in vivo T-cell depletion with antithymocyte globulin or alemtuzumab is the reference for reducing graft failure and GVHD; with haploidentical or highly mismatched donors, T-cell depletion and post-transplant cyclophosphamide are considered experimental (expert opinion) [9]. The EBMT position in transformed disease thus comes close to that of Myers and colleagues, who in the CIBMTR series observed that all patients with MDS/AML transplanted with reduced intensity died and suggest preferring myeloablation for better leukaemic control [9, 54]. The remaining divergence is one of degree: in MDS/AML, the choice of regimen still lacks comparative evidence, with small series and an internal discrepancy in the number of myeloablative regimens in the CIBMTR series itself (9 in the text, 8 in the abstract) [54].

The French PNDS adds a review of 229 published HSCTs up to June 2022: 77 patients (33.6 %) died of infection, organ failure or graft failure; survival was 79 % (100/126) when the indication was severe cytopenia without a clone and about 21 % in MDS or acute leukaemia, a difference explained partly by age (aplastic anaemia in the first decade, MDS/AML in the second and third) and by donor type (better outcomes with an HLA-identical sibling) [9]. Toxicity-related deaths reached 21 of 74 patients (28 %) in 1988–2016 and 12 of 54 (23 %) in 2000–2017, values far higher than in paediatric HSCT in general (cardiac toxicity, pulmonary haemorrhage, severe veno-occlusive disease), and no survival difference was shown between myeloablative and reduced-intensity regimens [9]. In the EBMT series, although 82 % of indications were non-clonal severe cytopenia, 54 % of transplants used a myeloablative regimen; in the CIBMTR series, 13 of 39 patients with non-clonal severe cytopenia had a myeloablative regimen; and, of the 13 EBMT patients with MDS (7) or AML (6), 4 were alive, and the 4 conditioned with treosulfan and cyclophosphamide died [9]. The PNDS recommends validating conditioning at a national multidisciplinary team meeting (RCP) and proposes regimens by indication:

Source: [9]. The text of the same PNDS reports good outcomes, in non-clonal severe cytopenia, with reduced-toxicity conditioning based on treosulfan, fludarabine or melphalan, avoiding cyclophosphamide, which contrasts with the presence of cyclophosphamide in the aplastic anaemia row of the table; the specific choice, including doses, is validated at the RCP [9].

The PNDS for aplastic anaemia adds two points applicable to SDS. Among constitutional aplastic anaemias, only Fanconi anaemia and telomere biology disorders require a genetic diagnosis to choose the conditioning; in the others, including SDS, the genetic diagnosis mainly serves counselling, the choice of a family donor and surveillance of extra-haematological manifestations and clonal evolution [3]. In follow-up after HSCT for aplastic anaemia, ciclosporin trough concentrations of 200 to 250 ng/mL are maintained for about 9 to 12 months, with progressive tapering over 6 to 12 months if chimerism is predominantly donor and stable; mixed chimerism (more than 5 % recipient cells), common with alemtuzumab, allows immunosuppression to be reduced every 3 months if it is stable and the blood count satisfactory (haemoglobin >10 g/dL, neutrophils >1 × 10⁹/L and platelets >100 × 10⁹/L), and requires ciclosporin to be increased again if chimerism or counts fall [3]. Primary graft failure in the first 30 days and rejection in the first 6 months are not rare in aplastic anaemia, partly because of multiple previous transfusions, and may justify a second transplant [3].

7.7. Treatment of MDS and AML

In SDS-associated MDS, conventional chemotherapy is not indicated and HSCT should be urgent; high-dose chemotherapy serves essentially as conditioning [10]. In AML, conventional chemotherapy controls the disease only temporarily; a donor is sought urgently and minimal chemotherapy is used for interim control [10]. In the North American MDS/AML cohort, of 10 patients with AML treated before HSCT (2 with hypomethylating agents and 8 with cytoreductive chemotherapy), only 1 achieved complete remission, and the other 9 died, with a median survival of 9 months; pretransplant cytoreduction did not prevent relapse and had unacceptable toxicity [41]. Of 4 patients with MDS initially treated with chemotherapy, only 1 achieved complete remission, and two patients with MDS with excess blasts progressed to AML at 2.5 and 3.8 months before any therapy [41]. The authors draw an analogy with Fanconi anaemia and TP53-mutated neoplasms, which respond poorly to cytotoxic chemotherapy and to some extent to hypomethylating agents, with or without venetoclax [41]. GeneReviews reports use of azacitidine with venetoclax as a bridge to HSCT in one TP53-mutated case, without this constituting established practice [1]. The French PNDS considers that chemotherapy has never achieved prolonged complete remission in this disease (grade B), that it is conceivable only as a bridge to HSCT and that classical AML induction with an anthracycline and cytarabine is highly toxic and of little efficacy; by analogy with frail older patients, it proposes considering hypomethylating agents (azacitidine or decitabine) and BCL-2 inhibitors, or a sequential approach with the FLAMSA regimen followed by reduced-intensity HSCT [9]. It also invokes the experience of Fanconi anaemia, in which HSCT in complete remission was associated with better survival than HSCT with active disease (71 % versus 37 %), with cytoreduction mainly with fludarabine and cytarabine or with azacitidine [9]. For congenital neutropenias in general, the 2024 PNDS does not consider it essential to achieve complete remission before HSCT, accepts demethylating agents to reduce the blast count without demonstrated benefit and notes the high risk of fungal infection with chemotherapy [23]. The Revised International Prognostic Scoring System (IPSS-R) and the World Health Organization Prognostic Scoring System (WPSS) are not validated in MDS associated with inherited bone marrow failure syndromes [41].

7.8. Skeleton and bone

Asphyxiating thoracic dystrophy requires early referral to paediatric respiratory medicine and orthopaedics [1]. Deformities of metaphyseal chondrodysplasia, particularly of the hips and knees, may require orthopaedic management or surgery, such as hemiepiphysiodesis for genu valgum, and radiographs assess deformities and stress fractures [10]. Low-turnover osteoporosis is treated with nutrition, fat-soluble vitamins, weight-bearing exercise, and calcium and vitamin D if intake is insufficient; the safety and efficacy of bisphosphonates in SDS are unknown [10]. The French PNDS maintains these positions, attributes the osteoporosis to a primary defect of bone metabolism related to marrow dysfunction and neutropenia, and considers that the optimal treatment of osteoporosis in SDS remains to be established [9]. Densitometry should be interpreted in light of short stature and delayed puberty [1, 10].

7.9. Liver, gastrointestinal tract, teeth, neurodevelopment and endocrinology

Chronic hepatitis in infancy has no specific treatment. The study by Koo and colleagues recommends continued liver surveillance even after enzyme normalisation, reserves biopsy for portal hypertension, declining synthetic function or pretransplant assessment, and concludes that prior chronic hepatitis should not preclude indicated HSCT, given the low incidence of liver failure [29]. Routine endoscopy in asymptomatic patients has limited value and should be guided by symptoms or laboratory abnormalities [39]. The French PNDS confirms that there is no specific treatment for hepatic cytolysis and adds that SDS does not contraindicate liver transplantation in progression to cirrhosis [9]. Dental care starts early, with intensive oral hygiene and topical fluoride and at least annual review, ideally by professionals experienced in orthodontics and periodontology [1, 10]. The French PNDS includes dental care in the 100 % coverage of the disease and recommends a recent blood count before repeated treatments: with platelets above 100 × 10⁹/L and neutrophils above 1 × 10⁹/L no precautions are needed, and thrombocytopenia or profound neutropenia warrant a prophylactic approach decided with the haematologist [9]. Enamel mineralisation defects can lead to caries and tooth loss, and acid from gastro-oesophageal reflux can cause erosion [9]. The PNDS for chronic neutropenias and its dental appendix propose follow-up from the first teeth or, at the latest, from 3 years of age, 2 to 4 visits per year depending on oral status, scaling 1 to 2 times per year, brushing twice daily with a soft toothbrush and fluoride toothpaste, floss or interdental brushes from adolescence and antiseptic mouthwashes for at most one week per month; they advise against implants, because of the risk of maxillary bone infection, except by collegial decision, and note that orthodontics requires a periodontal health that is often difficult to achieve [23]. For dental surgical procedures with bleeding, they recommend single-dose antibiotic prophylaxis in the preceding hour: amoxicillin 50 mg/kg in children and 2 g in adults, or, in penicillin allergy, clindamycin 20 mg/kg in children and 600 mg in adults, with the paediatric dose not exceeding the adult dose [23]. The clindamycin alternative corresponds to earlier French regimens: the 2023 European guideline on endocarditis no longer recommends clindamycin for dental prophylaxis, because of the risk of Clostridioides difficile infection, and proposes, in penicillin allergy, cefalexin 2 g (children 50 mg/kg, maximum 2 g), azithromycin or clarithromycin 500 mg (children 15 mg/kg, maximum 500 mg) or doxycycline 100 mg (2.2 mg/kg below 45 kg), as a single dose 30–60 minutes beforehand, avoiding cephalosporins after anaphylaxis, angioedema or urticaria with penicillin [71]. That guideline was written for the prevention of endocarditis, not for neutropenia; it serves here as the reference for the choice of antibiotic in allergy, and the PNDS clindamycin regimen remains a dated French position. Aphthous ulcers are treated with local antiseptic or analgesic measures (chlorhexidine, povidone-iodine, lidocaine gel), of modest efficacy, and deep ulcerations with systemic repercussions may justify antibiotics and resumption or intensification of G-CSF [23]. For neurodevelopment, the 2011 consensus recommends developmental checklists from diagnosis to 4 years, with early referral for speech and language therapy, occupational therapy, developmental paediatrics and psychology, and concurrent parental counselling; the 2026 data reinforce routine neuropsychological monitoring and early intervention [10, 44]. The French PNDS proposes neuropsychological assessments at around 6, 12 and 20–24 years of age, adjusted to educational and occupational demands, covering intellectual abilities, attention (span, selective, sustained and divided), higher-order language, visual perception, visuomotor functioning, executive functions, school readiness and attainment, behaviour and functional independence, with speech and language therapy, occupational therapy and psychomotor therapy as needed, counselling for parents and psychological support for the patient, parents and siblings [9]. Growth hormone is indicated for proven deficiency [1]; in the North American cohort, 12 treated patients (5 with confirmed deficiency and 5 for short stature, started at a median age of 9.3 years for a median duration of 4 years) showed no change in height z-score slope, and no MDS/AML case was attributed to treatment, whereas a previous study of 6 patients described improvement in the first two years [34]. The French PNDS, based on earlier studies, considers that growth hormone can improve height without adverse effects and without apparent leukaemic promotion and that SDS is not a contraindication on account of the risk of malignant transformation; it recommends endocrinological assessment at around 5–6 years and before puberty, with the indication discussed with the endocrinologist according to the predicted final height and, if necessary, at an RCP, and stresses that the pubertal growth spurt is critical for final height and that bone age improves the prediction [9]. The more recent North American data, showing no gain in height slope, diverge from this favourable reading (see 7.15). Delayed puberty and other endocrine manifestations are followed by endocrinology [1].

7.10. Agents to avoid and precautions

Prolonged G-CSF use requires caution and annual marrow surveillance (see 7.4) [1, 10]. Cyclophosphamide and busulfan at conventional conditioning doses may be unsuitable because of cardiac toxicity, and total body irradiation should be limited [1, 11]. Conventional cytoreductive chemotherapy before HSCT in AML had unacceptable toxicity without demonstrated benefit [41]. Patients with lymphoid neoplasms have increased toxicity with conventional regimens, as in the lymphoma case treated with R-CHOP in which doxorubicin had to be omitted because of ventricular dysfunction [28, 43]. Androgens are not first line because of hepatotoxicity [10]. The French PNDS documents advise against pegfilgrastim, intramuscular injections, rectal temperature measurement and, except by collegial decision, dental implants [23]. The sources consulted identify no other drugs specifically contraindicated in SDS.

7.11. Pregnancy, lactation, perioperative care and transition

Pregnancy in a woman with SDS should be managed as high risk, with haematology consultation [1]. GeneReviews reports an increased risk of recurrent miscarriages, with rates of live births, spontaneous miscarriages and elective terminations comparable to those in the general population, citing a study by Giri and colleagues (2018) whose title refers to mothers of children with bone marrow failure syndromes, limiting interpretation of this finding [1]. The European registry documents two successful pregnancies in 25-year-old mothers with SDS, and the patient of Yang and colleagues had a miscarriage at 25 weeks and pre-eclampsia in two subsequent pregnancies [25, 26]. The French PNDS considers that fertility is not reduced, records in the national registry 8 children of parents with SDS (7 born to mothers and 1 to a father), without vertical transmission, and identifies no specific pregnancy complications, which diverges from the GeneReviews recommendation of management as a high-risk pregnancy [1, 9]. Filgrastim is not recommended during pregnancy according to the European label, because of scarce human data, reproductive toxicity in animals and reported transplacental passage; its excretion in milk is unknown, and the decision on breastfeeding weighs benefits for child and mother [66]. The PNDS for chronic neutropenias considers, on the contrary, that G-CSF may be used throughout pregnancy if necessary, adapting anti-infective prophylaxis in the neutropenic pregnant woman; the guideline is the reference, and the more restrictive European label serves as a comparison [23, 66]. Before complex dental procedures or orthopaedic surgery, G-CSF and prophylactic antibiotics may be useful, and cytopenias should be corrected according to the thresholds in 7.4 [1, 10]. Transition to adult care should include continued haematological and liver surveillance, given that AML arises at a median age of 28 years and histological liver disease can emerge in adolescence or adulthood, and genetic counselling regarding family planning [20, 28, 29]. When HSCT is anticipated, the PNDS for aplastic anaemia recommends a reproductive medicine consultation before treatment, even with non-myeloablative conditioning of low gonadal risk, weighing the haemorrhagic and infectious risk of the procedures and the harm of delaying transplantation: before puberty only gonadal tissue can be cryopreserved (ovarian cortex in girls; testicular tissue in boys, an experimental technique), and after puberty sperm or oocytes after stimulation are cryopreserved, procedures that profound thrombocytopenia or neutropenia may contraindicate; preservation may also be offered after HSCT [3]. SDS-specific sources describe no structured transition models; the PNDS for aplastic anaemia proposes a three-phase programme (preparation in paediatrics, with empowerment of the adolescent, part of the visit without parents and early designation of the adult physician; flexible transfer between 16 and 20 years of age, preferably in a stable phase, with a joint consultation and alternating visits in the first year; and reception in the adult service with an extended consultation), applicable to SDS [3]. In France, SDS has 100 % coverage as a long-term condition (ALD) and access to disability schemes (the education allowance for a disabled child, AEEH, and the disabled adult allowance, AAH) and, at school, to a school life assistant and to the assessment guide for compensation needs in schooling (GEVA-Sco); communal life is permitted without restrictions [9, 23]. Quality of life is affected by psychocognitive difficulties (school progression, occupational integration), motor limitations and pancreatic insufficiency [9].

7.12. Regulatory status

Sources: [63–66, 72].

In France, according to the PNDS for chronic neutropenias (situation as of April 2024), sulfamethoxazole + trimethoprim and filgrastim hold a marketing authorisation (AMM) for these indications, lenograstim is an acceptable off-label use with the same level of evidence as filgrastim, and pegfilgrastim is not recommended; multi-year dental care and dental prostheses are recommended, although outside the basket of reimbursed care [23].

No SDS-targeted therapy is authorised by the Food and Drug Administration (FDA) or EMA according to the sources consulted; a negative search does not demonstrate the absence of ongoing processes [46, 72].

7.13. Intervention and therapeutic outcome studies

Sources: [26, 34, 54, 63, 69].

There are no randomised trials in SDS. The data consist of retrospective registry series with different inclusion criteria, periods and definitions, which do not allow indirect comparisons between studies; the ataluren series has 3 patients without a control group and with laboratory outcomes, placing it at case-series level; and post-transplant survival reflects selection bias among patients who reach transplantation [63, 70]. ClinicalTrials.gov (search for the condition “Shwachman-Diamond” on 10 October 2026) lists observational studies (Boston Children's Hospital registry, NCT06056908, recruiting since 2016 with a target of 5000 participants; NCI cancer study in bone marrow failure syndromes, NCT00027274; SDS Alliance global survey, NCT06999954), the completed endocrine surveillance study (NCT04275479), completed treosulfan conditioning trials (NCT04965597; NCT00919503, 98 patients with non-malignant diseases), an old G-CSF pilot in bone marrow failure syndromes (NCT00004787, n=20) and a phase 3 mavorixafor trial in chronic neutropenias (NCT06056297, n=176), whose eligibility for SDS the sources do not clarify; there is no registered trial of ataluren or SDS-targeted therapy, and the European trial registry was not searched [46].

7.14. Developments in the last 6–12 months

The ataluren series warrants careful reading: the rationale is readthrough of the premature termination codon p.Lys62Ter, present in more than half of registry patients, and the regimen used was that on the Duchenne dystrophy label (10 mg/kg in the morning, 10 mg/kg at midday and 20 mg/kg in the evening, as oral granules) [63]. In addition to the improvements described, marrow cellularity rose in all three patients, marrow neutrophil maturation increased, SBDS protein normalised in marrow mononuclear cells from two patients, and marrow TP53 status remained unchanged at 6 months [63]. With non-renewal of European authorisation in 2025, ataluren ceased to be authorised in the EU for any indication, affecting access; this is case-series evidence rather than a treatment recommendation [63, 72]. Mammalian target of rapamycin inhibitors (mTOR; everolimus, dactolisib) did not improve granulopoiesis in vitro in the same work [63].

7.15. Guideline recommendations and divergences

Sources: [1, 3, 9–11, 20, 23, 26, 29, 34, 54, 64, 66, 67].

The main divergence is the intensity of marrow surveillance. The 2011 consensus allows marrow examinations every 1 to 3 years in stable patients, whereas the AACR update recommends annual marrow examination and somatic panel from diagnosis, supported by subsequent data on clonal haematopoiesis and poor survival after AML [10, 20]. In the North American MDS/AML cohort, 3-year survival was 62 % in patients with serial marrow surveillance versus 28 % without surveillance, a difference without statistical significance (p=0.13) and subject to adherence bias, and surveillance by full blood count alone was less sensitive than the combination with marrow examination [41]. This article adopts the AACR recommendations as current for haematological surveillance and notes that evidence of survival benefit is observational.

The French PNDS (French National Diagnosis and Care Protocol) introduce three divergences of wider scope. The first concerns marrow surveillance: instead of an annual marrow examination, they propose a marrow examination at diagnosis and thereafter only on clinical or molecular indication, with an annual NGS panel in blood, supported by the concordance between blood and marrow observed in 12 patients and by its lower invasiveness [9, 23]. The AACR recommendation remains the reference for marrow examination, as it is more recent and issued by an international predisposition panel, but the French option is a reasoned alternative when the blood panel is performed regularly; neither strategy has been compared in outcome studies [9, 20]. The second is the pancreatic enzyme dose: the PNDS indicates 5000–10,000 units of lipase/kg/day, a range that starts above the initial dose of the 2011 consensus and respects the same daily maximum; the choice within the range depends on growth, steatorrhoea and elastase, without exceeding 10,000 units/kg/day without documented efficacy [9, 10, 64]. The third is testing of asymptomatic minor siblings, which the PNDS does not propose outside compatible signs or donation, whereas GeneReviews and the AACR recommend it for the benefit of early haematological surveillance [1, 9, 20]. The remaining divergences (ages for neuropsychological assessment and radiographs, growth hormone, pregnancy and G-CSF in pregnancy, elastase threshold) reflect expert consensus with a low grade of evidence (C) and have no comparative data [9, 23].

The PNDS for bone marrow aplasias adds four divergences of practical scope. The systematic irradiation of blood products proposed in 2011 gives way to irradiation of red cell concentrates only after antithymocyte globulin or from conditioning onwards, without irradiation of platelets; since SDS does not receive immunosuppression, the French rule limits irradiation to the HSCT period, and the local choice depends on each country's transfusion standards [3, 10]. The severity criteria differ in the thresholds for reticulocytes (40 versus 60 × 10⁹/L, the latter for automated counting) and cellularity (25 versus 30 %) [3, 10]. Chelation now has explicit thresholds and doses, earlier than those of the US deferasirox label [3, 67]. As for antifungal prophylaxis, the PNDS for chronic neutropenias does not recommend it, and the PNDS for aplasias allows it case by case in constitutional forms with profound neutropenia; the two positions are reconciled by the phase of the disease, the second applying only to the aplastic phase [3, 23]. The appendix table of the same PNDS assigns autosomal recessive inheritance to SRP54, which contradicts the dominant, almost always de novo, inheritance described in GeneReviews and in the original series, which is the one supported by family data [1, 3, 4].

8. Reproductive Options and Family Counselling

8.1. Autosomal recessive forms (SBDS, DNAJC21, EFL1)

The parents of an affected child are usually asymptomatic heterozygotes. Occasionally, only one parent is a carrier and the child has one inherited and one de novo variant: approximately 10 % of pathogenic SBDS variants arise de novo, and the equivalent frequency in EFL1 and DNAJC21 is unknown [1]. Testing both parents is therefore recommended; if only one is a carrier and biological maternity and paternity are confirmed, one of the proband's variants occurred de novo in the proband or in a parent with postzygotic mosaicism [1]. If the proband appears homozygous, an undetected deletion on the other allele or uniparental isodisomy should be considered [1]. Parental testing is also essential to establish the phase of the complex allele c.[183_184delinsCT;258+2T>C] [1].

Source: [1].

The clinical spectrum varies between affected siblings, and absence of symptoms does not exclude disease in a sibling [1, 37]. In light of available data, heterozygous SBDS carriers have no demonstrated risk of aplastic anaemia or AML (see 2.9) [1]. Carrier testing in relatives requires that the proband's variants be known [1].

8.2. Autosomal dominant form (SRP54)

Most patients with an SRP54 variant have a de novo variant [1, 19]. If the variant is not detected in parental leucocytes, the risk to siblings is approximately 1 %, owing to possible germline mosaicism; if a parent carries it, the risk is 50 %; if the parents have not been tested but are clinically unaffected, the risk is low but above population risk because of possible reduced penetrance and mosaicism [1]. Each child of an affected patient has a 50 % probability of inheriting the variant [1]. In the French registry, 7 of 23 cases were familial, showing that dominant transmission from a parent with a mild phenotype or isolated neutropenia is possible [4].

The dominant mode of transmission rests on the two founding series and on variant databases. Carapito and colleagues (2017) identified, by trio exome sequencing in a family with SDS without an SBDS variant and by targeted sequencing in other patients, three unrelated patients, each with a de novo missense variant in the GTPase domain of SRP54, and concluded that these were autosomal dominant mutations [19]. Bellanné-Chantelot and colleagues (2018) found heterozygous variants in three sporadic cases and in one family with autosomal dominant transmission and, across the whole French registry, in 23 cases (16 sporadic and 7 familial), with 7 distinct germline variants [4]. OMIM designates entry 618752 as severe congenital neutropenia 8, autosomal dominant, and the Human Phenotype Ontology inheritance annotation for that entry is autosomal dominant, based on Carapito 2017 [76]. In ClinVar, accessed on 6 November 2026, the point variants classified as pathogenic or likely pathogenic (among them p.(Thr117del), p.(Thr115Ala), p.(Gly226Glu), p.(Ala223Asp) and p.(Gly274Asp)) are recorded as heterozygous de novo, with autosomal dominant inheritance when the mode is stated, and no submission associates SRP54 with a recessive condition [77]. The two French PNDS that address the disease in detail, those for SDS (2023) and for chronic neutropenias (2024), also describe it as autosomal dominant [9, 23]. The assignment of AR inheritance to SRP54 in the table of appendix 2 of the PNDS for bone marrow aplasias (2023), which groups SBDS, SRP54 and EFL1 as recessive, has no counterpart in these data and is retained here as a divergence of the source, not as a second pattern [3].

In a PubMed search (6 November 2026) for SRP54 with biallelic, homozygous or compound heterozygous variants, the only human report was that of Rahmani and colleagues (2021), in an Iranian cohort of 25 children with infantile-onset inflammatory bowel disease, in which 11 of the 14 patients with a monogenic defect were born to consanguineous unions: a novel SRP54 variant in the homozygous state in a patient who died and the variant in the heterozygous state in a sibling with a milder phenotype [78]. The abstract does not state the variant, its classification or the presence of neutropenia or pancreatic insufficiency, and the full text was not obtained; an isolated case, in a context of consanguinity and with an affected heterozygote as well, does not establish a recessive form and is compatible with a dosage effect of a variant with dominant effect. When faced with a homozygous SRP54 variant, counselling should start from the dominant model (50 % risk for the offspring of each carrier) and the heterozygous parents should be assessed clinically.

8.3. Assessment of at-risk relatives

Older and younger siblings of a proband should be assessed as early as possible: by molecular testing if familial variants are known, or, if unavailable, by assessment of exocrine pancreatic function and full blood count [1]. This screening has two aims: identifying asymptomatic affected siblings, who benefit from haematological surveillance, and excluding disease in potential family HSCT donors [1, 54]. The AACR update recommends cascade testing of relatives at the age when risk increases and germline testing of family donors, and includes natural history, inheritance, family planning, organ surveillance and transition to adult care in counselling [20]. In minors, sibling testing has direct clinical benefit because SDS can be asymptomatic and haematological surveillance should begin at diagnosis [1, 20]; the sources consulted do not specifically address carrier status testing in healthy minors.

The French PNDS takes a more restrictive position: testing of minor siblings is justified in the presence of symptoms, even subtle ones, compatible with the disease, or with a view to intrafamilial marrow donation, and presymptomatic diagnosis is not offered [9]. For the partner of a patient planning a family, it proposes, in a genetic counselling consultation, targeted testing for the two recurrent variants, which account for about 90 % of alleles; with a carrier frequency estimated at 1 in 205 in the Italian population, the probability that the partner is a carrier is low, which explains the absence of observed vertical transmission [9]. The PNDS for chronic neutropenias allows, for the same purpose, a genetic study targeted at the gene concerned [23]. The diagnosis is disclosed in an individual consultation by an experienced physician, with both parents in the case of a child, and covers the mode of transmission, the course, the risk of recurrence, the associated organ involvement, the organisation of follow-up and of the care network, the gene-dependent surveillance strategy and the value of family investigation [9].

8.4. Prenatal diagnosis and preimplantation genetic testing

Counselling should take place before pregnancy and be offered to affected young adults, carriers or those at risk [1]. When there is no molecular diagnosis, DNA banking from the proband allows future studies [1]. Once the familial variants have been identified, prenatal diagnosis and preimplantation genetic testing are possible, and counselling should acknowledge that professionals and families may hold different views on their use [1]. In prenatal diagnosis, the laboratory method must circumvent the SBDSP1 pseudogene, as shown by cases in which fetal exome sequencing was initially inconclusive [47]. The French PNDS considers a medical genetics clinic consultation indispensable in the initial period after diagnosis and for young adults planning conception, and considers preimplantation diagnosis and prenatal diagnosis justified, given the 1 in 4 risk and the severity of the disease; double preimplantation diagnosis, with simultaneous selection of an embryo HLA-matched to the affected sibling, is theoretically justified but hardly feasible in France [9].

9. Special Considerations: Population Diversity and Epidemiology

9.1. Prevalence and what each number measures

The most commonly cited estimate, 1 in 77,000 births, comes from the mapping work of Goobie and colleagues (2001), who inferred it from the observation that SDS was approximately 1/20 as frequent as cystic fibrosis in North America; this is not an epidemiological study, and should be treated as an order of magnitude [1, 5]. This estimate does not apply to the rare forms, whose frequency has never been measured [5]. Other figures circulate through indirect attribution: 1 in 76,000 in the 2011 consensus, 1 in 160,000 and 1 in 168,000 in 2025 articles, and 1 in 75,000 to 100,000 in another review; none constitutes a single validated estimate [10, 25, 34, 63].

Sources: [1, 6, 9, 36, 60, 79].

The point prevalence of 0.28 per 100,000 that Orphanet presents as worldwide rests on the French national cohort of 102 patients, and should be read as a French estimate rather than validated worldwide prevalence [6, 36]. The Puerto Rican value refers to an autopsy report focused on the differential diagnosis with cystic fibrosis, whose abstract contains no prevalence data, weakening this estimate [60]. Orphanet counts 26 cases or families for SRP54-related disease in the two founding series, with prevalence below 1 per million; by 2023, 38 patients had been published [8, 52]. GeneReviews counts more than 500 individuals with a molecular diagnosis across the four genes [1]. The French PNDS, drawing on the two population-based registries with published estimates, gives for Italy an annual incidence of 5.9 per million births (1996–2000; about 3 births per year out of 504,000) and a heterozygous carrier frequency of 1 in 205, and for France an incidence of 5.1 per million (95 % CI 4.1–6.2; 1995–2017; about 4.4 cases per year out of 800,000 births) and a prevalence of 4.2 patients per million inhabitants (grade B) [9]. These registry values are lower than the indirect North American estimate of 1 in 77,000; the difference may reflect the method and the underdiagnosis of mild forms, which no registry quantifies [1, 9]. For congenital neutropenias as a whole, the 2024 PNDS estimates a prevalence of about 1 in 100,000 and 20 to 30 cases out of 850,000 births; in June 2023, the French registry included 170 patients with SDS, the most numerous genetic entity, ahead of ELANE neutropenias (155) and GATA2 deficiency (125) [23].

9.2. Ancestry, recurrent variants and founder populations

SDS has been described in populations of European, Indian, Native American, Chinese, Japanese and African ancestry [1]. Unlike many recessive diseases, the predominant variants reflect recurrent gene conversions from SBDSP1 rather than regional founder effects, explaining the similarity of spectra between North America, Europe and Japan [14, 26, 32]. The frequency of the compound genotype c.258+2T>C/c.183_184delinsCT was 65 % in the European registry (which includes 55 patients from the Russian registry), 78.1 % in the US NCI cohort and 61.4 % in the North American Registry growth cohort [24, 26, 34]. A founder variant has been described in the French-Canadian population, whose identity the sources consulted do not state [1]. EFL1 variants were described in families from Israel and Palestine, and SRP54 variants chiefly in the French cohort [4, 18].

9.3. Portugal and Brazil

The sources consulted contain no prevalence estimates for Portugal or Brazil. The sole Portuguese finding identified is molecular: Costa and colleagues (2007) described an AluSx-mediated deletion of SBDS exon 3 in a Portuguese patient, one example of whole-exon deletions that escape sequencing [1]. Absence of data in the sources consulted does not demonstrate absence of published studies.

9.4. Special populations

Adults with SDS are a growing population: 20 % of patients in the European registry and 25.5 % in the North American Registry were adults at last follow-up, and AML risk is concentrated in young adulthood [26, 28]. Suspected SDS in young adults with MDS or AML has implications for conditioning and donor selection [20, 54]. Women of childbearing age require counselling regarding high-risk pregnancy (see 7.11), and transplant recipients require endocrine, liver and growth surveillance after HSCT [1, 29, 34].

9.5. Reference networks, registries and patient associations

In Europe, the relevant network is ERN-EuroBloodNet, dedicated to rare haematological diseases, including bone marrow failure syndromes [12]. In France, SDS is followed by the rare disease reference centre (CRMR) Neutropénies Chroniques, coordinated at Trousseau Hospital (Assistance Publique–Hôpitaux de Paris), with a constituent site in Toulouse (dedicated to GATA2 deficiency) and competence centres at Saint-Louis Hospital (adults and HSCT), in Rennes, in the dentistry department of Bretonneau Hospital, at Necker Hospital and in Angers; the national multidisciplinary team meeting (RCP) meets at least once a month and can be contacted through the MaRIH health network (rare immuno-haematological diseases) [9, 23]. In the bone marrow failure phase, the French point of contact is the CRMR Aplasies médullaires acquises et constitutionnelles, of the same MaRIH filière, coordinated at Saint-Louis Hospital (adults), with a paediatric site at Robert-Debré Hospital, and with a national multidisciplinary team meeting usually held on the first and third Wednesday of each month; the national RIME observatory, created in early 2017, records acquired and constitutional bone marrow failure in anonymised form and is linked to a biobank of blood, marrow and fibroblasts, and the reference medical biology laboratory for aplasias (Saint-Louis Hospital) receives samples for the genetic panel as a priority [3]. The centre distributes a bone marrow aplasia emergency card, to be given to all patients, information booklets for the patient and family and, for transplant recipients, the national follow-up booklet of the Société Francophone de Greffe de Moelle et de Thérapie Cellulaire (SFGM-TC); therapeutic education includes nursing and clinical pharmacy consultations and patient days organised by associations, among them HPN France–Aplasie médullaire [3]. The French Chronic Neutropenia Registry, with a favourable opinion from the Commission Nationale de l'Informatique et des Libertés since 1997, included 1307 patients (994 with congenital neutropenia) in the latest analysis, enrols patients with written consent and transmits anonymised data to the European neutropenia registry (Hannover) and to the registry of the European Society for Immunodeficiencies (ESID, Freiburg); annual reports are available on the reference centre's website (www.neutropenie.fr) [9, 23]. The PNDS list IRIS, the Association Césaire and the Association syndrome de Shwachman-Diamond as French patient associations and, for dental care, the centres of the O-Rares network [9, 23]. Active registries include the Boston Children's Hospital SDS registry (NCT06056908), the NCI cancer study in inherited bone marrow failure syndromes (NCT00027274), the Italian Registry (RI-SDS, Verona), the French Severe Chronic Neutropenia Registry and the European SCNIR [26, 36, 42, 46]. Patient associations include the Shwachman-Diamond Syndrome Foundation and the Shwachman-Diamond Syndrome Alliance (USA), Associazione Italiana Sindrome di Shwachman (Italy), Stichting Shwachman Syndroom Support (the Netherlands), which makes the 2011 consensus available, and SDS Deutschland (Germany) [1, 80–82]. The SBDSbase variant database records described mutations [1].

10. Illustrative Clinical Case

10.1. Severe neonatal form without the hypomorphic allele (North American Registry)

Furutani and colleagues described, in the North American Registry analysis, the only female patient in the cohort of 159 individuals with biallelic SBDS who did not have the hypomorphic c.258+2T>C variant [28]. Her genotype comprised a null allele and a missense variant, NM_016038.4:c.183_184delinsCT p.(Lys62Ter) and c.523C>T p.(Arg175Trp), the latter with conflicting classifications in ClinVar [28, 33]. At birth, she had intrauterine growth restriction and severe neonatal aplastic anaemia, with platelet and red cell transfusion dependence and severe neutropenia unresponsive to G-CSF [28]. She had clinodactyly, severe thoracic dystrophy and exocrine pancreatic insufficiency requiring enzyme replacement [28]. The marrow biopsy showed irregular cartilage islands surrounded by osteoid, with an increased ratio of bone to marrow space and a loose fibroblastic network around the bony islands (Figure 1A, original magnification ×40, and Figure 1B, ×200, haematoxylin and eosin stain; https://pmc.ncbi.nlm.nih.gov/articles/PMC8753194/figure/F1/) [28]. The figure description is based on the caption published in the article. The patient underwent unrelated-donor HSCT with reduced-intensity conditioning and died from overwhelming bacterial infection before engraftment [28].

The case has three lessons. First, it reinforces the role of the c.258+2T>C allele, which allows residual SBDS production and is present in 99.4 % of patients in the cohort; its absence was associated with the most severe phenotype observed, in agreement with mouse models lacking a hypomorphic allele [28]. Second, it shows that SDS can present in the newborn as complete bone marrow failure and thoracic dysplasia, overlapping the severe neonatal form distinguished by Orphanet [7]. Third, it illustrates the severity of peritransplant infectious complications in patients with severe bone marrow failure, consistent with series in which infection, graft failure and toxicity are the main causes of early death [54]. The management adopted (HSCT for refractory severe bone marrow failure) corresponds to the indications in the 2011 consensus [10].

10.2. Cystic fibrosis-like presentation with 42 years of follow-up

Yang and colleagues reported a woman followed for 42 years in a UK respiratory clinic [25]. At 6 years, having been born at 41 weeks and always “smaller than her peers”, she presented with a productive cough for one week and chronic steatorrhoea, a history of sinusitis, eczema and recurrent respiratory, urinary and skin infections, and bone abnormalities on examination [25]. Assessment showed elevated liver enzymes (gamma-glutamyltransferase, ALT, AST) and immunoreactive trypsin, Haemophilus influenzae in sputum, impaired neutrophil locomotion and chemotaxis and normal ciliary function; the sweat test was borderline, with sodium of 43 and 48 mEq/L [25]. A presumptive clinical diagnosis of SDS was made; approximately 12 years later, CFTR testing was negative, and Sanger sequencing of the five SBDS exons identified c.258+2T>C and, in trans, the novel missense variant c.284T>A p.(Val95Glu), at the same codon as a previously described variant (p.Val95Gly), molecularly confirming SDS 29 years after clinical diagnosis [25]. In adulthood, she had a miscarriage at 25 weeks and pre-eclampsia in her two subsequent pregnancies; at 48 years, she still had intermittent neutropenia, normal transaminases and pancreatic insufficiency, with frequent lower respiratory infections, normal spirometry and bronchial wall thickening with mucus plugs without bronchiectasis on CT [25].

The case shows that mild phenotypes survive into adulthood and that SDS should be considered when a cystic fibrosis-like presentation has equivocal or negative tests [25]. The report does not describe marrow surveillance; current recommendations provide for a full blood count every 3–4 months and annual marrow examination with a somatic panel for any patient with SDS [20]. Normalisation of transaminases in adulthood is consistent with the biochemical evolution described in the North American Registry, but does not exclude histological liver disease [29].

11. Prognosis and Long-Term Follow-up

11.1. Prognosis

Prognosis is dominated by haematological complications. In the North American Registry (153 patients with biallelic SBDS, 2020 data cut-off), estimated median survival was 38.2 years (95 % CI 29.5–48.5); of the 20 deaths (12.6 %, median age 21.2 years), 15 (75 %) were due to neoplasia, 2 to bone marrow failure, 2 to asphyxiating thoracic dystrophy and 1 to liver failure, and none to infection outside the context of neoplasia or transplantation [28]. Other registries describe deaths from sepsis, meningitis and other infections, particularly in early childhood [1, 28]. In the European registry, the 11 deaths (median age 14.45 years) were due to leukaemia (2), transplant-related mortality (5), post-transplant relapse (2) and infections (2) [26]. In the North American MDS/AML cohort, 3-year survival was 11 % in AML and 51 % in MDS, lower than that of young adults with AML without SDS (about 50 % at 5 years), and a complex karyotype was associated with poorer survival in MDS (15 % versus 64 % at 3 years, p=0.15) [41]. The factors associated with severe haematological complications in the French cohort were symptom onset before 3 months and low haematological parameters [36]. The French PNDS, based on the French and Italian registries, estimates a probability of death of about 10 % at 20 years, approximately stable between 20 and 30 years and rising markedly after 30, with survival of 85 % (France) and 87 % (Italy) at 20 years (grade B); about 80 % of deaths before 40 years are due to leukaemic transformation, and the remainder to infections, mainly viral (influenza), and to solid tumours [9]. Elsewhere, the same document mentions about 30 % mortality before 30 years from haematological and infectious complications, a figure that does not agree with the survival curves it presents; the data from the curves prevail [9]. In June 2023, the French registry recorded 26 deaths among 170 patients with SDS (15 %) [23]. The review by Yang and colleagues cites a life expectancy of 38 to 41 years, by indirect attribution [25]. These values come from cohorts with still limited adult follow-up and different transplant eras, and should not be read as individual life expectancy.

Outside haematological complications, treated natural history is favourable in several domains: pancreatic function improves in a large proportion of patients, growth velocity normalises with enzymes and nutrition, and neutrophil counts tend to rise with age, but short stature, chronic liver disease in a subgroup, progressive metaphyseal abnormalities and special educational needs persist [1, 28, 29, 44].

11.2. Assessment after diagnosis

Sources: [1, 9, 10, 20, 40].

11.3. Surveillance

Haematological surveillance follows the AACR update, which is more intensive than the 2011 consensus and GeneReviews (full blood count every 3–6 months and marrow every 1–3 years); the reasons for this choice and the observational nature of the evidence are discussed in 7.15 [1, 10, 20]. Somatic panel results should be interpreted with caution, because the presence, number and allele fraction of monoallelic TP53 clones did not predict imminent leukaemia, and conventional sequencing does not distinguish the monoallelic from the biallelic state; single-cell sequencing remains a research tool [20, 22]. Case SDS-048 from the Kennedy study illustrates the problem: AML arose despite stable blood counts and surveillance marrows without neoplasia, with a TP53 p.Cys242Phe clone detectable at 0.17 % allele fraction 4.5 years before transformation and expansion of the biallelic subclone over the following four years, while other TP53 clones remained monoallelic for 6.5 years [22]. For solid tumours, there is no evidence-based surveillance guideline [24]. In SRP54- and DNAJC21-related forms, the AACR applies the same haematological surveillance schedule [20]. The French PNDS prefers the annual NGS panel in blood and reserves marrow examination for clinical or molecular indications, a divergence discussed in 7.15; C-reactive protein and renal function come from the chronic neutropenia PNDS and apply to congenital neutropenias as a whole [9, 23].

12. Summary and Final Considerations

Shwachman–Diamond syndrome is a ribosomopathy in which failure to release eIF6 from the 60S subunit, due to SBDS or EFL1 deficiency, or neighbouring disturbances of ribosomal biogenesis and protein synthesis due to DNAJC21 and SRP54 variants, compromises the acinar pancreas, haematopoiesis, growth cartilage, liver and neurodevelopment [1, 18]. Approximately 92 % of cases are due to biallelic SBDS variants, almost always arising through gene conversion from the SBDSP1 pseudogene, with pan-ethnic predominance of c.258+2T>C and c.183_184delinsCT (NM_016038.4); the hypomorphic c.258+2T>C allele is present in practically all patients [1, 14, 28].

Diagnosis rests on the combination of cytopenia, chiefly neutropenia, with exocrine pancreatic dysfunction assessed using age-adjusted markers, and is confirmed by molecular testing capable of circumventing the pseudogene; half of patients lack the classical presentation, and an exome or panel not designed for SDS may miss the diagnosis [10, 32, 37]. Pancreatic insufficiency tends to improve with age, but short stature, progressive metaphyseal chondrodysplasia, chronic hepatitis in infancy with fibrosis in a subgroup, and neurocognitive and adaptive difficulties accompany many patients into adulthood [1, 29, 34, 44].

The lifelong risk of MDS and AML, estimated at 10 to 30 % and concentrated in late childhood and young adulthood, is the main cause of death, and survival after AML is very low [20, 28, 41]. Clonal haematopoiesis is almost universal with age and follows two pathways: the compensatory pathway, through EIF6 or del(20q), and the premalignant pathway, through biallelic TP53 inactivation [22]. Current surveillance therefore combines a full blood count every 3–4 months with annual marrow examination and somatic panel, to identify clonal evolution and indicate HSCT before overt leukaemia, although survival benefit is supported only by observational data [20, 41].

Treatment remains supportive: pancreatic enzyme replacement according to cystic fibrosis rules, fat-soluble vitamins, low-dose filgrastim only for severe neutropenia with infections, irradiated transfusions and HSCT with reduced-intensity conditioning for severe bone marrow failure, with far poorer outcomes in transformed disease [10, 11, 26, 70]. Targeted therapies are in early stages, from the compassionate use series of three patients treated with ataluren, which lost European authorisation for its original indication, to ex vivo gene editing at the preclinical stage [63, 72, 74]. Genetic counselling should include parental testing for variant phase, testing of siblings and family donors, and discussion of prenatal and preimplantation diagnosis using methods appropriate to the pseudogene [1, 20]. In France, the HAS PNDS of 2023 (SDS) and 2024 (chronic neutropenias) organise this care around the reference centre and the national multidisciplinary team meeting, adopt the EBMT consensus for transplantation and propose, as an alternative to annual marrow examination, annual molecular surveillance in blood [9, 23].

13. Bibliographical References

1. Nelson A, Myers K. Shwachman-Diamond Syndrome. 2008 Jul 17 [updated 2024 Sep 19]. In: Adam MP, Bick S, Mirzaa GM, Wallace SE, Amemiya A, editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. PMID: 20301722. Available at: https://www.ncbi.nlm.nih.gov/books/NBK1756/
2. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford University Press; 2018. p. 50–53 (3.5 Shwachman Syndrome).
3. Haute Autorité de Santé. Protocole National de Diagnostic et de Soins (PNDS): Aplasies médullaires acquises et constitutionnelles. Saint-Denis La Plaine: HAS; 2023 (version of 1 March 2023, published 2023-05-04; reference centre for acquired and constitutional bone marrow aplasias, MaRIH filière). Available from: https://www.has-sante.fr/jcms/p_3431517
4. Bellanné-Chantelot C, Schmaltz-Panneau B, Marty C, Fenneteau O, Callebaut I, Clauin S, et al. Mutations in the SRP54 gene cause severe congenital neutropenia as well as Shwachman-Diamond-like syndrome. Blood. 2018;132(12):1318-31. doi:10.1182/blood-2017-12-820308
5. GeneGuidance [local database]. Gene records for SBDS, DNAJC21 and SRP54 (EFL1 absent). Extracted on 2026-10-10.
6. Orphanet. Shwachman-Diamond syndrome (ORPHA:811). Available from: https://www.orpha.net/en/disease/detail/811
7. Orphanet. SBDS-related severe neonatal spondylometaphyseal dysplasia (ORPHA:622934). Available from: https://www.orpha.net/en/disease/detail/622934
8. Orphanet. Severe congenital neutropenia-developmental delay syndrome due to SRP54 deficiency (ORPHA:675767). Available from: https://www.orpha.net/en/disease/detail/675767
9. Haute Autorité de Santé. Protocole National de Diagnostic et de Soins (PNDS): Maladie de Shwachman Diamond. Saint-Denis La Plaine: HAS; 2023 Apr 5 [document dated March 2023; coordination: Centre de référence Maladies Rares Neutropénies Chroniques, Alimi A, Donadieu J]. Available from: https://www.has-sante.fr/jcms/p_3425534
10. Dror Y, Donadieu J, Koglmeier J, Dodge J, Toiviainen-Salo S, Makitie O, et al. Draft consensus guidelines for diagnosis and treatment of Shwachman-Diamond syndrome. Ann N Y Acad Sci. 2011;1242:40-55. doi:10.1111/j.1749-6632.2011.06349.x
11. Cesaro S, Donadieu J, Cipolli M, Dalle JH, Styczynski J, Masetti R, et al. Stem cell transplantation in patients affected by Shwachman-Diamond syndrome: expert consensus and recommendations from the EBMT Severe Aplastic Anaemia Working Party. Transplant Cell Ther. 2022;28(10):637-49. doi:10.1016/j.jtct.2022.07.010
12. European Reference Network on Rare Hematological Diseases (ERN-EuroBloodNet) [Internet]. [accessed 2026 Oct 10]. Available from: https://www.eurobloodnet.eu
13. Shwachman H, Diamond LK, Oski FA, Khaw KT. The syndrome of pancreatic insufficiency and bone marrow dysfunction. J Pediatr. 1964;65:645-63. doi:10.1016/s0022-3476(64)80150-5
14. Boocock GR, Morrison JA, Popovic M, Richards N, Ellis L, Durie PR, et al. Mutations in SBDS are associated with Shwachman-Diamond syndrome. Nat Genet. 2003;33(1):97-101. doi:10.1038/ng1062
15. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 39, section 39.3 Ribosomal Biogenesis (39.3.1, 39.3.4, 39.3.5; Table 39.1).
16. Tummala H, Walne AJ, Williams M, Bockett N, Collopy L, Cardoso S, et al. DNAJC21 mutations link a cancer-prone bone marrow failure syndrome to corruption in 60S ribosome subunit maturation. Am J Hum Genet. 2016;99(1):115-24. doi:10.1016/j.ajhg.2016.05.002
17. Dhanraj S, Matveev A, Li H, Lauhasurayotin S, Jardine L, Cada M, et al. Biallelic mutations in DNAJC21 cause Shwachman-Diamond syndrome. Blood. 2017;129(11):1557-62. doi:10.1182/blood-2016-08-735431
18. Stepensky P, Chacón-Flores M, Kim KH, Abuzaitoun O, Bautista-Santos A, Simanovsky N, et al. Mutations in EFL1, an SBDS partner, are associated with infantile pancytopenia, exocrine pancreatic insufficiency and skeletal anomalies in a Shwachman-Diamond like syndrome. J Med Genet. 2017;54(8):558-66. doi:10.1136/jmedgenet-2016-104366
19. Carapito R, Konantz M, Paillard C, Miao Z, Pichot A, Leduc MS, et al. Mutations in signal recognition particle SRP54 cause syndromic neutropenia with Shwachman-Diamond-like features. J Clin Invest. 2017;127(11):4090-103. doi:10.1172/JCI92876
20. Maese LD, Wlodarski MW, Kim SY, Bertuch AA, Bougeard G, Chang VY, et al. Update on recommendations for surveillance for children with predisposition to hematopoietic malignancy. Clin Cancer Res. 2024;30(19):4286-95. doi:10.1158/1078-0432.CCR-24-0685
21. Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1, Clinical Approach to Inborn Errors of Metabolism in Paediatrics (1.6.5 Haematology; Tables 1.36 and 1.38).
22. Kennedy AL, Myers KC, Bowman J, Gibson CJ, Camarda ND, Furutani E, et al. Distinct genetic pathways define pre-malignant versus compensatory clonal hematopoiesis in Shwachman-Diamond syndrome. Nat Commun. 2021;12(1):1334. doi:10.1038/s41467-021-21588-4
23. Haute Autorité de Santé. Protocole National de Diagnostic et de Soins (PNDS): Neutropénies chroniques. Saint-Denis La Plaine: HAS; 2024 Aug 6 [document dated May 2024; Centre de Référence Maladies Rares Neutropénies Chroniques]. Available from: https://www.has-sante.fr/jcms/p_3522406
24. Thompson AS, Giri N, Gianferante DM, Jones K, Savage SA, Alter BP, et al. Shwachman Diamond syndrome: narrow genotypic spectrum and variable clinical features. Pediatr Res. 2022;92(6):1671-80. doi:10.1038/s41390-022-02009-8
25. Yang F, Weitnauer L, Felton I, Simmonds NJ. Beyond cystic fibrosis: recognising Shwachman-Diamond syndrome in the respiratory clinic. Respirol Case Rep. 2025;13(12):e70440. doi:10.1002/rcr2.70440
26. Mellor-Heineke S, Skokowa J, Gerschmann N, Deordieva E, Tesakov I, Kinsey S, et al. Genetic and clinical characteristics of patients with Shwachman Diamond syndrome with special consideration of treatment with granulocyte-colony stimulating factor. Haematologica. 2025;110(9):2171-5. doi:10.3324/haematol.2024.286119
27. Oyarbide U, Bezzerri V, Staton M, Boni C, Shah A, Cipolli M, et al. Reduced EIF6 dosage attenuates TP53 activation in models of Shwachman-Diamond syndrome. J Clin Invest. 2025;135(8):e187778. doi:10.1172/JCI187778
28. Furutani E, Liu S, Galvin A, Steltz S, Malsch MM, Loveless SK, et al. Hematologic complications with age in Shwachman-Diamond syndrome. Blood Adv. 2022;6(1):297-306. doi:10.1182/bloodadvances.2021005539
29. Koo J, Schwarz D, Haney M, Putra J, Kim J, Cheng L, et al. Liver disease in Shwachman-Diamond syndrome: a comprehensive characterization across the age spectrum. J Pediatr Gastroenterol Nutr. 2026;82(6):1357-65. doi:10.1002/jpn3.70430
30. Reilly CR, Shimamura A. Predisposition to myeloid malignancies in Shwachman-Diamond syndrome: biological insights and clinical advances. Blood. 2023;141(13):1513-23. doi:10.1182/blood.2022017739
31. Freeman PJ, Hart RK, Gretton LJ, Brookes AJ, Dalgleish R. VariantValidator: accurate validation, mapping, and formatting of sequence variation descriptions. Hum Mutat. 2018;39(1):61-8. doi:10.1002/humu.23348. Available from: https://variantvalidator.org
32. Yamada M, Uehara T, Suzuki H, Takenouchi T, Inui A, Ikemiyagi M, et al. Shortfall of exome analysis for diagnosis of Shwachman-Diamond syndrome: mismapping due to the pseudogene SBDSP1. Am J Med Genet A. 2020;182(7):1631-6. doi:10.1002/ajmg.a.61598
33. National Center for Biotechnology Information. ClinVar [Internet]. VCV000003196, VCV000003195, VCV000449095, VCV000418467, VCV001328553, VCV002498198, VCV002498197 (SBDS); VCV000430852 (SRP54). Available from: https://www.ncbi.nlm.nih.gov/clinvar/
34. Schwarz L, Cheng L, Steltz S, Schwarz D, Korn E, Loveless S, et al. Growth patterns in Shwachman-Diamond syndrome: findings from the North American Shwachman-Diamond Syndrome Registry. J Pediatr. 2025;287:114780. doi:10.1016/j.jpeds.2025.114780
35. Genome Aggregation Database (gnomAD) v4.1. Variant 7-66994210-A-G (SBDS c.258+2T>C): 5343/1 459,192 alleles in exomes and 510/152 106 in genomes [Internet]. Cambridge (MA): Broad Institute; [accessed 2026 Nov 6]. Available from: https://gnomad.broadinstitute.org/variant/7-66994210-A-G?dataset=gnomad_r4
36. Donadieu J, Fenneteau O, Beaupain B, Beaufils S, Bellanger F, Mahlaoui N, et al. Classification of and risk factors for hematologic complications in a French national cohort of 102 patients with Shwachman-Diamond syndrome. Haematologica. 2012;97(9):1312-9. doi:10.3324/haematol.2011.057489
37. Myers KC, Bolyard AA, Otto B, Wong TE, Jones AT, Harris RE, et al. Variable clinical presentation of Shwachman-Diamond syndrome: update from the North American Shwachman-Diamond Syndrome Registry. J Pediatr. 2014;164(4):866-70. doi:10.1016/j.jpeds.2013.11.039
38. Sabharwal S, Grover A, Mitchell P, Serino-Cipoletta J, Shimamura A, Fawaz R. Liver stiffness in patients with Shwachman-Diamond syndrome. JPGN Rep. 2026;7(2):200-6. doi:10.1002/jpr3.70138
39. Korn E, Koo J, Schwarz D, Haney M, Putra J, Kim J, et al. Endoscopic findings in patients with Shwachman-Diamond syndrome: a report from the North American Shwachman-Diamond syndrome registry. JPGN Rep. 2026. doi:10.1002/jpr3.70218
40. Gloude NJ, Brundige K, Loveless S, Atkinson I, Chandra S, Cooper R, et al. Beyond neutropenia: immunological evaluation of patients with Shwachman-Diamond syndrome. Br J Haematol. 2026;209(1):180-8. doi:10.1111/bjh.70553
41. Myers KC, Furutani E, Weller E, Siegele B, Galvin A, Arsenault V, et al. Clinical features and outcomes of patients with Shwachman-Diamond syndrome and myelodysplastic syndrome or acute myeloid leukaemia: a multicentre, retrospective, cohort study. Lancet Haematol. 2020;7(3):e238-46. doi:10.1016/S2352-3026(19)30206-6
42. Micheloni G, Frattini A, Montalbano G, Cipolli M, Pintani E, Cesaro S, et al. Cytogenetics of Shwachman Diamond syndrome: 27 years report of the Italian cohort and review of the literature. Eur J Med Genet. 2026;83:105095. doi:10.1016/j.ejmg.2026.105095
43. Reed HD, Do H, Weller E, Harris MH, Reilly CR, Tsai HK, et al. Lymphoid malignancies in patients with Shwachman-Diamond syndrome. Blood. 2025;145(21):2528-32. doi:10.1182/blood.2024026507
44. Koo J, Gerstle M, Cheng L, Cooper R, Coyne K, Dusa C, et al. Neuropsychological and educational outcomes in Shwachman-Diamond syndrome — a report from the North American Shwachman-Diamond Syndrome Registry. Pediatr Blood Cancer. 2026;73(11):e70650. doi:10.1002/1545-5017.70650
45. Minelli A, Pintani E, Valli R, Tridello G, Porta G, Fioredda F, et al. Shwachman-Diamond syndrome and diabetes: an update from the Italian registry and review of the literature. Exp Clin Endocrinol Diabetes. 2025;133(2):78-82. doi:10.1055/a-2460-6977
46. ClinicalTrials.gov. Studies for condition «Shwachman-Diamond» [Internet]. Bethesda (MD): National Library of Medicine; [accessed 2026 Oct 10]. Available from: https://clinicaltrials.gov/api/v2/studies?query.cond=Shwachman-Diamond
47. Vanden Eynde N, Slegers I, Vantroys E, Symoens S, Doné E, Vorsselmans A, et al. Prenatal Shwachman-Diamond syndrome: diagnostic challenges in two unrelated cases with a rare clinical presentation and pseudogene interference, and a review of the literature. Prenat Diagn. 2026;46(2):270-6. doi:10.1002/pd.70073
48. Zhu L, Wang X, Shi Y, Huang X, Ding H. Prenatal diagnosis of Shwachman-Diamond syndrome: fetal compound heterozygous variants in the SBDS gene associated with mildly straight ribs. Prenat Diagn. 2025;45(4):568-71. doi:10.1002/pd.6762
49. D'Amours G, Lopes F, Gauthier J, Saillour V, Nassif C, Wynn R, et al. Refining the phenotype associated with biallelic DNAJC21 mutations. Clin Genet. 2018;94(2):252-8. doi:10.1111/cge.13370
50. Bernat CE, Vilain C, Calo P, Lagrou J, Mazilier P. Biallelic DNAJC21 variants in two children: phenotypic heterogeneity, TP53 clonal evolution and allogeneic HSCT. Pediatr Blood Cancer. 2026;73(9):e70423. doi:10.1002/1545-5017.70423
51. Deshpande A, Mathur R, Velangi M, Norton A, Cherungonath A, Lawson S, et al. Expanding the phenotypic spectrum of DNAJC21-associated bone marrow failure syndrome: a single-centre experience. Br J Haematol. 2026;208(5):1863-6. doi:10.1111/bjh.70372
52. Fan EM, Vagher J, Meznarich JA, Ubico EM, Goteti S, Peterson D, et al. Severe congenital neutropenia, SRP54 pathogenicity, and a framework for surveillance. Am J Med Genet A. 2023;191(5):1434-41. doi:10.1002/ajmg.a.63156
53. Sabulski A, Grier DD, Myers KC, Davies SM, Rubinstein JD. Acute myeloid leukemia in SRP54-mutated congenital neutropenia. EJHaem. 2022;3(2):521-5. doi:10.1002/jha2.413
54. Myers K, Hebert K, Antin J, Boulad F, Burroughs L, Hofmann I, et al. Hematopoietic stem cell transplantation for Shwachman-Diamond syndrome. Biol Blood Marrow Transplant. 2020;26(8):1446-51. doi:10.1016/j.bbmt.2020.04.029
55. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 150–151 (Failure to thrive).
56. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford University Press; 2018. p. 33–36 (3.1 Metaphyseal Chondrodysplasia, Schmid Type).
57. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford University Press; 2018. p. 37–40 (3.2 Cartilage-Hair Hypoplasia).
58. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford University Press; 2018. p. 169–171 (7.5 Spondylometaphyseal Dysplasia, Sedaghatian Type).
59. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. p. 439–448 (Glycogenosis type I — von Gierke disease).
60. Lozada-Muñoz L, Aliaga MD. Shwachman-Diamond syndrome: the clinical imitator of cystic fibrosis. P R Health Sci J. 1995;14(4):275-7.
61. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 853–855 (Appendix: Skeletal dysplasia charts).
62. Spranger JW, Brill PW, Hall C, Nishimura G, Superti-Furga A, Unger S. Bone Dysplasias: An Atlas of Genetic Disorders of Skeletal Development. 4th ed. Oxford University Press; 2018. p. 406–408 (11.8 Axial Spondylometaphyseal Dysplasia).
63. Bezzerri V, Pegoraro A, Hristodor AM, Crane GM, Meneghelli I, Brignole C, et al. Ataluren improves hematopoietic and pancreatic disorders in Shwachman-Diamond syndrome patients: a compassionate program case-series. Nat Commun. 2025;16(1):8189. doi:10.1038/s41467-025-63137-3
64. AbbVie Inc. CREON (pancrelipase) delayed-release capsules, for oral use. Prescribing information. Revised 2024 Feb [effective 2024-02-28]. BLA020725. DailyMed set ID 073201aa-556d-4a70-918e-84e9616fd88d. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=073201aa-556d-4a70-918e-84e9616fd88d
65. Amgen Inc. NEUPOGEN (filgrastim) injection, for subcutaneous or intravenous use. Prescribing information [effective 2026-07-23]. BLA103353. DailyMed set ID 97cc73cc-b5b7-458a-a933-77b00523e193. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=97cc73cc-b5b7-458a-a933-77b00523e193
66. European Medicines Agency. Zarzio (filgrastim): EPAR — product information [Internet]. Amsterdam: EMA; first authorised 2009 Feb 6. Available from: https://www.ema.europa.eu/en/documents/product-information/zarzio-epar-product-information_en.pdf
67. MSN Laboratories Private Limited. Deferasirox tablets, for oral use. Prescribing information [effective 2021-01-08]. ANDA210945. DailyMed set ID 0354cbc7-9b32-4513-976a-2ef2da39b6b3. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0354cbc7-9b32-4513-976a-2ef2da39b6b3
68. AvKARE. Deferiprone tablets, for oral use. Prescribing information [effective 2026-08-12]. ANDA208800. DailyMed set ID 58dd1370-5042-833f-e063-6394a90a0777. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=58dd1370-5042-833f-e063-6394a90a0777
69. Cesaro S, Pillon M, Sauer M, Smiers F, Faraci M, de Heredia CD, et al. Long-term outcome after allogeneic hematopoietic stem cell transplantation for Shwachman-Diamond syndrome: a retrospective analysis and a review of the literature by the Severe Aplastic Anemia Working Party of the European Society for Blood and Marrow Transplantation (SAAWP-EBMT). Bone Marrow Transplant. 2020;55(9):1796-809. doi:10.1038/s41409-020-0863-z. Erratum: Bone Marrow Transplant. 2020;55(9):1884.
70. Ma C, Huang X, Chen Y, Shi F, Li X, Tian S, et al. Outcomes of allogeneic hematopoietic stem cell transplantation in Shwachman-Diamond syndrome: a systematic review and meta-analysis. Cytotherapy. 2025;27(4):465-74. doi:10.1016/j.jcyt.2024.12.004
71. Delgado V, Ajmone Marsan N, de Waha S, Bonaros N, Brida M, Burri H, et al. 2023 ESC Guidelines for the management of endocarditis. Eur Heart J. 2023;44(39):3948–4042. doi:10.1093/eurheartj/ehad193. PMID:37622656.
72. European Medicines Agency. Translarna (ataluren): EPAR [Internet]. Amsterdam: EMA; authorisation expired (EC non-renewal 2025 Mar 28). Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/translarna
73. Myers KC, Donadieu J, Goldberg B, Delafoy M, Zeidler C, Mellor-Heineke S, et al. Cumulative hematologic morbidity and outcomes in Shwachman-Diamond syndrome. N Engl J Med. 2026;395(12):1241-4. doi:10.1056/NEJMc2605001
74. EQT Foundation. Daniel Bauer — Advancing targeted therapy for Shwachman-Diamond syndrome through prime editing [Internet]. 2026. Available from: https://eqtgroup.com/eqt-foundation/grants/daniel-bauer
75. Shwachman-Diamond Syndrome Alliance. RTW Foundation grant: prime editing gene therapy for SDS [Internet]. Available from: https://www.sdsalliance.org/post/rtwf-grant-prime-editing-gene-therapy-sds
76. Human Phenotype Ontology. OMIM:618752 Neutropenia, severe congenital, 8, autosomal dominant: inheritance annotation HP:0000006 (Autosomal dominant inheritance; source PMID:28972538). Bar Harbor (ME): The Jackson Laboratory. Available from: https://hpo.jax.org/browse/disease/OMIM:618752 [accessed 6 November 2026; the OMIM page returned HTTP 403]
77. National Center for Biotechnology Information. ClinVar [Internet]. Gene SRP54 (NM_003136.4): VCV000430852, VCV000430851, VCV000430850, VCV000810688, VCV000810689, VCV000810690, VCV000985721. Bethesda (MD): NCBI. Available from: https://www.ncbi.nlm.nih.gov/clinvar/?term=SRP54%5Bgene%5D [accessed 6 November 2026]
78. Rahmani F, Rayzan E, Rahmani MR, Shahkarami S, Zoghi S, Rezaei A, et al. Clinical and mutation description of the first Iranian cohort of infantile inflammatory bowel disease: the Iranian Primary Immunodeficiency Registry (IPIDR). Immunol Invest. 2021;50(4):445-59. doi:10.1080/08820139.2020.1776725. PMID:32633164
79. Minelli A, Nicolis E, Cannioto Z, Longoni D, Perobelli S, Pasquali F, et al. Incidence of Shwachman-Diamond syndrome. Pediatr Blood Cancer. 2012;59(7):1334-5. doi:10.1002/pbc.24260
80. Shwachman-Diamond Syndrome Alliance. Publications and diagnostic & treatment guidelines [Internet]. [accessed 2026 Oct 10]. Available from: https://sdsalliance.org/diagnostic-and-treatment-guidelines
81. Stichting Shwachman Syndroom Support [Internet]. The Netherlands. [accessed 2026 Oct 10]. Available from: https://www.shwachman.nl
82. SDS Deutschland e.V. [Internet]. Germany. [accessed 2026 Oct 10]. Available from: https://sdsdeutschland.de/en/