Trimethylaminuria (FMO3 Deficiency)

Inborn error of amine metabolism in which gut-derived trimethylamine escapes hepatic N-oxygenation by flavin-containing monooxygenase 3 and is eliminated in sweat, breath, urine, and other secretions, producing a rotting-fish odour with no other physical signs but with psychosocial consequences; it includes the primary autosomal recessive form, the mild and transient forms, and the secondary forms.

Literature and regulatory information updated: 27 October 2026.

0. Initial Reference Table

1. Introduction

1.1. Historical background

Descriptions of people who smell of fish predate modern medicine. GeneReviews and a recent British survey cite as the earliest reference the character Satyavati, a young woman who smelled of rotting fish in the Mahabharata, an Indian epic compiled around 400 AD, and who was exiled to work as a ferrywoman; the second literary reference is Caliban, of whom Trinculo says in Shakespeare's The Tempest that "he smells like a fish" (Act II, Scene ii) [1, 8, 17]. These allusions do not amount to a diagnosis, but they show that the odour and the social exclusion that accompanies it were recognised long before biochemistry explained them.

The first clinical description of trimethylaminuria (TMAU) is that of Humbert, Hammond, and Hathaway, published in the Lancet in October 1970 under the title "Trimethylaminuria: the fish-odour syndrome", which introduced both names still in use [4, 18]. The patient was a 6-year-old child being investigated for immunodeficiency, whose mother reported an intermittent fishy odour; unmetabolised trimethylamine (TMA) was present in her secretions and urine, and urinary TMA rose by 67% after an oral TMA load of 15 mg/kg, which did not occur in three controls; the article has no indexed abstract, and its content is known through the review by Mackay and colleagues [8]. In 1972, Higgins and colleagues demonstrated "trimethylamine oxidase" deficiency on a liver biopsy from the same patient, and during that period Daniel Ziegler characterised the flavin-containing monooxygenases in pig liver microsomes [4]. In 1987, Al-Waiz and colleagues framed the disease as an inborn error of oxidative metabolism [19]. In 1993, Ayesh and colleagues at St Mary's Hospital in London used a newspaper article to recruit 187 people complaining of body odour, 156 of whom underwent biochemical testing; 11 had trimethylaminuria, with a fraction of TMA oxidised to TMAO below 55% on a normal diet and below 25% after an oral load of 600 mg of TMA (normal values above 80%), and the parents of 6 patients showed N-oxidation below 80% after the load, consistent with heterozygosity, which established autosomal recessive inheritance [20]. By 2011 more than 200 cases had been described, and a 2020 review refers to several hundred in the literature [8, 14].

1.2. Molecular discovery

In 1997, Dolphin and colleagues determined the organisation of the FMO3 gene directly from genomic DNA (one noncoding exon followed by eight coding exons), which allowed exon-by-exon amplification, and in the same year identified in a patient the variant c.458C>T p.(Pro153Leu) in exon 4, which segregated with the disease in the pedigree and abolished the catalytic activity of the enzyme [21, 22]. The first molecular series followed: three causal mutations in nine probands born in Australia (1998), a North American cohort with one common mutation and many private ones (1999), and the description, also in 1999, of mild to transient trimethylaminuria caused by common FMO3 variants with markedly reduced in vivo activity [23–25]. A mutation database was established in 2003 and has since been replaced by the shared LOVD database [2, 26].

1.3. Nomenclature and classification

The term "fish odour syndrome" still appears in databases, but in a British survey 71% of patients preferred "TMAU", and 27 of 38 respondents considered the English expression pejorative, not least because the odour is not always described as fishy (sometimes faecal or like rubbish) [17]. Recent reviews distinguish TMAU1, primary and caused by FMO3 variants, from TMAU2, which is secondary [4]. The classification of Mitchell and Smith (2001), cited by GeneReviews and by Bouchemal and colleagues, separates five situations: primary genetic autosomal recessive trimethylaminuria; acquired (for example, after viral hepatitis); transient childhood; transient menstrual; and precursor overload (betaine in homocystinuria, L-carnitine) [1, 5]. Saudubray organises the same reality into four clinical forms: severe, mild/moderate, transient, and secondary to an underlying disease [3]. Orphanet keeps ORPHA:468726 active for the severe primary form and has inactivated ORPHA:35056 because it considers trimethylaminuria as a whole not rare in Europe [9, 10].

1.4. Definition of the entities covered in this article

Severe primary trimethylaminuria is defined by the excretion, on a normal diet, of more than 10% of total TMA as the free amine (generally more than 40%), associated with biallelic pathogenic loss-of-function variants in FMO3 [1]. Mild or moderate trimethylaminuria corresponds to reduced N-oxygenation capacity of multifactorial etiology, combining reduced-activity alleles, physiologically decreased expression of the enzyme, and substrate overload [3]. The transient forms occur in preterm newborns and infants, owing to the immaturity of FMO3, and in the perimenstrual period [1, 27]. The secondary forms result from an excess of precursors or from disease that increases intestinal TMA production or impairs its hepatic clearance [4, 14]. FMO3 is the main flavin-containing monooxygenase of the adult liver, and GeneGuidance describes primary trimethylaminuria as a nonprogressive condition affecting quality of life and social functioning [6].

2. Molecular Genetics and Pathophysiology

2.1. Trimethylamine metabolism

TMA is a tertiary amine and a strong base (pKa 9.8), volatile and malodorous in its free, unprotonated form at room temperature [3, 14]. At physiological pH most of it is protonated, but in equilibrium with a small free-base fraction that the human sense of smell detects in the parts-per-billion range; the odour threshold given by Mackay and colleagues is 0.12 parts per billion (ppb), with men being more sensitive, and about 7% of people have a specific anosmia for TMA [8, 14]. At pH 6.0, less than 0.02% of TMA is in the volatile free-base form, which is the rationale for using acidic soaps [1].

TMA is a bacterial metabolite produced in the colon from free and lecithin-bound choline, from the TMAO of saltwater fish, from L-carnitine, and from betaine (Fig. 31.1 in Saudubray) [3]. Choline is absorbed in the small intestine by carrier-mediated transport, with half-saturation at 200–300 µM; when absorptive capacity is exceeded, choline reaches the colon and is metabolised to TMA [1, 14]. The bacterial enzymatic pathways identified are the cutC/cutD gene cluster (choline TMA-lyase, a glycyl radical enzyme, and its activator, described in Desulfovibrio desulfuricans), the CntA/B system (a Rieske-type oxygenase for carnitine and γ-butyrobetaine), and the promiscuous YeaW/X system (choline, carnitine, γ-butyrobetaine, and betaine) [4]. TMA producers predominate among the Firmicutes and are rare among the Bacteroidetes; oral bacteria, such as Streptococcus, also produce TMA [4, 28]. Marine fish contain up to 3 mg/g of TMAO, the highest level of any food; about half of ingested TMAO is excreted unchanged, and the remainder is reduced to TMA by bacterial TMAO reductase [28]. A typical Western diet generates about 50 mg of TMA per day, about 95% of which is converted to TMAO [28]. Normal urinary excretion is about 1 mg of TMA and 40 mg of TMAO per day [5].

TMA is absorbed by passive diffusion and reaches the liver via the portal circulation, where more than 90% is oxygenated by microsomal FMO3, which uses flavin adenine dinucleotide (FAD) as a prosthetic group, to TMAO, which is water-soluble and odourless and is excreted in the urine [3, 14]. The Michaelis constant (Km) of FMO3 for TMA has been estimated at between about 15 and 310 µmol/L, depending on pH and experimental conditions [8]. The catalytic cycle uses NADPH (reduced nicotinamide adenine dinucleotide phosphate) to reduce FAD to FADH2, which reacts with oxygen to form 4a-hydroperoxyflavin; nucleophilic attack of the substrate on this intermediate is the rate-limiting step, followed by release of water and NADP+ [4].

2.2. The FMO3 gene and the FMO family

The human family has five functional genes (FMO1 to FMO5) and six pseudogenes (FMO6P to FMO11P); FMO1–FMO4 and FMO6P form a cluster at 1q24.3 [4, 8]. Sources disagree on the position of the second cluster: Mackay and colleagues place it about 4 Mb telomeric, whereas Sidoti and colleagues place FMO5 about 26 Mb more centromeric, at 1q21.1, and FMO7P–FMO11P at 1q24.2 [4, 8]. FMO3 spans about 27 kb and has nine exons, of which the first and part of the last are noncoding; it encodes a 532-amino-acid, 60,047-Da transmembrane protein of the endoplasmic reticulum, expressed mainly in the adult liver and also in skin, pancreas, and adrenal cortex and medulla [4, 8]. The promoter contains binding elements for NFY, USF1, YY1, Pbx2/Hox, and NF-E2, and there is an upstream C/EBPβ element [8]. Hepatic FMO3 expression varies up to 20-fold between individuals and with ancestry [3, 6].

FMO1 is the fetal hepatic form and disappears around the third postnatal day, without TMAO being a relevant product; FMO2 is functional in the lung of 26% of African Americans and 4.5% of Hispanics, but not in people of European or Asian ancestry; FMO5 has a hepatic mRNA level similar to that of FMO3 [8]. Unlike humans, male mice switch off hepatic Fmo3 at about 6 weeks of age, a difference to be taken into account when interpreting animal models [14, 28].

2.3. Ontogeny and physiological regulation

FMO3 is present in 30% of first-trimester fetal livers, disappears in the second and third trimesters, and reappears at about 21 days after birth, regardless of gestational age; it reaches about 8% of the adult level at 9 months and 20% at 11 years, and remains below the adult level at 18 years [8]. This slow maturation explains the transient forms of childhood and the susceptibility of infants carrying reduced-activity alleles to choline or TMAO overload [1, 2]. FMO3 expression decreases in the perimenstrual period; in a 30-day study, variation in men was slight, whereas women showed a perimenstrual rise in TMA lasting about 7 days, attributed to downregulation of FMO3 expression by female sex hormones; high testosterone also reduces FMO3 capacity, and testosterone therapy is reported as an aggravating factor [3, 14, 28]. FMO3 activity is inhibited by the indoles and sulfur-containing glucosinolates of brassicas, by tannins, and by nitric oxide–mediated S-nitrosylation, and is induced in pregnancy [3, 14, 28].

2.4. Disease mechanism

Trimethylaminuria occurs when absorbed TMA exceeds the capacity of hepatic FMO3; the amount of TMA depends on the amount and form of the precursors ingested and on the relative abundance of the producing bacterial species [3]. GeneReviews distinguishes two interrelated pathophysiological types: (a) decreased amount or activity of FMO3, owing to genetic (FMO3 variants), physiological (hormones), or environmental (chemical inhibitors) factors, with an elevated urinary TMA/TMAO ratio; and (b) substrate overload, owing to excess precursors or variation in the gut flora, with elevated urinary TMA and a normal TMA/TMAO ratio [1]. A load tolerated by one individual may represent an overload for another with reduced residual activity; in the severe form the genetic deficiency predominates, whereas in the mild/moderate form the etiology is multifactorial and the odour appears intermittently when a risk factor tips the balance of the system [1, 3]. The molecular mechanism of the causal variants is loss of function [1, 9].

Functional characterisation shows two modes of failure: kinetic incompetence and failure of holoenzyme assembly. The p.(Met66Ile) and p.(Arg492Trp) variants do not incorporate or do not retain FAD, whereas p.(Pro153Leu) and p.(Asn61Ser) have a catalytic efficiency below 10% for TMA N-oxygenation, the former through altered Km and catalytic constant and the latter through increased Km; the homology model places residue 61 close to the FAD centre [29]. p.(Asn61Ser) abolishes TMA N-oxygenation but does not affect methimazole S-oxygenation, demonstrating substrate specificity, whereas p.(Met434Ile) and p.(Arg492Trp) impair both [1, 30].

TMAO has been studied as a marker and possible mediator of cardiovascular disease, with conflicting results across studies; there is no evidence that the low TMAO of trimethylaminuria is protective, and associations of common FMO3 variants with other diseases are likewise inconsistent [3, 14, 31]. The hypothesis of a gut–brain axis, in which TMAO at physiological concentrations would strengthen the blood–brain barrier and TMA would compromise it, rests on in vitro and animal studies [4]. Dysfunctional metabolism of endogenous amines that are FMO3 substrates, such as tyramine, has been proposed as a contributor to depression in some patients, again without proof [1].

2.5. Variant spectrum and nomenclature

FMO3 is highly polymorphic. As of August 2020, 231 variants were recorded in the shared LOVD (Leiden Open Variation Database); Mackay and colleagues report more than 300 variants or single-nucleotide polymorphisms (SNPs), more than 40 of them associated with trimethylaminuria, and Sidoti and colleagues more than 90 associated variants [3, 4, 8]. On 20 July 2024, HGMD (Human Gene Mutation Database) listed 94 mutations: 81 missense/nonsense, 1 splicing, 1 regulatory, 7 small deletions, 2 small insertions, 1 small indel, and 1 gross deletion [4]. More than 40 variants cause the severe form, mostly missense, and the rest are nonsense or small frameshift deletions [3]. The only gross deletion described, of 12.2 kb and involving exons 1 and 2, was found in the homozygous state in a Greek patient [1, 2]. Causal variants vary by population: in Europeans c.458C>T p.(Pro153Leu) and c.913G>T p.(Glu305Ter) predominate; in Japanese individuals, c.613C>T p.(Arg205Cys) and c.1498C>T p.(Arg500Ter) [2]. Molecular series have added private variants: in the North American cohort with TMA oxidation below 50% of normal, the missense variants p.(Ala52Thr) and p.(Arg387Leu), the nonsense variant p.(Glu314Ter), and an allele made up of the two common polymorphisms; in Italian families, the deletion G1182del (codon 394) and the missense variant p.(Arg238Pro); in Japan, p.(Ile441Thr), p.(Ser195Leu), p.(Val58Ile), p.(Pro70Leu), p.(Gly421Val), and p.(Thr488Ala), all with an allele frequency below 1% and reduced activity on bacterial expression, plus the nonsense variants p.(Cys197Ter) and p.(Arg500Ter) (notation as in the sources, not normalised) [24, 32, 33].

The MANE Select transcript of FMO3 is NM_001002294.3 (ENST00000367755.9), encoding NP_001002294.1; GeneReviews uses NM_006894.5/NP_008825.4, with the same c. numbering (verified for c.458C>T, NC_000001.11:g.171107811C>T) [1, 7]. Table 2.1 normalises the variants cited in this article; coordinates marked (37) were returned in GRCh37.

Table 2.1. FMO3 variants (NM_001002294.3) normalised in VariantValidator, with ClinVar classification accessed on 27 October 2026 [7, 34].

The variant NM_001002294.3:c.1079T>C p.(Leu360Pro), confirmed in VariantValidator, is described as the only one with increased activity [7, 8]. Some typographical errors in the sources are worth knowing: Mackay and colleagues write "c.472C>T" and "c.23A>G (p.Glu308Gly)" where the correct notation is c.472G>A and c.923A>G; the Irish cohort writes "c.472G>T" for one patient and the Irish psychosocial study "c.932A>G"; some articles write "p.Glu158Gly" for p.Glu158Lys; and the indexed abstract of Dolphin 2000 gives "M4341" for p.Met434Ile [8, 11, 30, 37]. ClinVar classifications do not replace clinical interpretation: the isolated polymorphisms are benign, the cis haplotype has old assertions without criteria, and its penetrance is undetermined [11].

2.6. The common p.[(Glu158Lys;Glu308Gly)] haplotype

Fifteen nonsynonymous SNPs are known; only c.183C>A p.(Asn61Lys), which is rare, severely reduces activity [2]. The SNPs c.472G>A p.(Glu158Lys) and c.923A>G p.(Glu308Gly) in cis moderately reduce activity in vitro and drug metabolism in vivo; in the homozygous state they can cause mild or transient trimethylaminuria, especially in infants and young children, who have low FMO3 expression [1, 2, 31]. p.(Glu158Lys) is the most common variant in African, Asian, and European populations; p.(Glu308Gly) is common in Europeans and Asians and less so in Africans; the two in cis reach about 20% in Europeans and Asians [4]. Bouchemal and colleagues report an allele frequency of about 40% for p.(Glu158Lys), varying with ancestry, and about 10% homozygotes [5]. In Ireland, in an earlier study of 2036 individuals, the allele frequency of the haplotype was about 0.2, corresponding to about 4% homozygotes, most of them asymptomatic, and in the USA it was about 0.15 [11]. Saudubray gives an estimated homozygosity of 2–5% in Germany and 4% in Ireland [3]. In Sicily and Sardinia (528 healthy people), the haplotype was less common, and 158KK/308EG individuals had a considerable reduction in activity measured by mass spectrometry without any odour, showing that biochemical trimethylaminuria is not always accompanied by odour [36]. Molecular docking models suggest that the haplotypes may narrow the enzyme pocket and alter TMA transit [4].

Other common variants have ancestry-dependent frequencies: p.(Val257Met) is more common in Asians and has no catalytic effect; p.(Asp132His) is more common in Africans; p.(Val277Ala) and p.(Glu362Gln) are common only in Africans, with an effect yet to be determined; p.(Gly180Val) occurs only in Europeans, with no effect [4]. Two haplotypes containing promoter SNPs severely reduce expression in vitro, without in vivo validation, and four haplotypes containing g.−5109G>C reduce it moderately without a marked effect on TMA metabolism [2]. In Japan, four combinations of upstream variants that reduce transcriptional activity have been described [40].

2.7. Genotype–phenotype correlation, penetrance, and expressivity

On a normal diet, individuals with biallelic loss-of-function variants excrete more than 40% of total TMA as the free amine and have an odour; the more severe the reduction in activity, the more severe the manifestations and the poorer the response to treatment [1]. The European gene card states that all homozygotes or compound heterozygotes for variants that abolish activity have severe trimethylaminuria [2]. Haplotype effects are clinically relevant: p.(Val187Ala) alone does not alter activity, but in cis with p.(Glu158Lys) it severely reduces it; a young woman with severe trimethylaminuria had p.(Glu158Lys), p.(Val187Ala), and p.(Glu308Gly) on the maternal allele and p.(Glu305Ter) on the paternal allele [2, 35]. In Ireland, homozygosity for the common haplotype was associated with one severe and two mild phenotypes, a difference attributed to diet, inhibitors, and other genes or environmental factors [11]. In Italian families without an identified FMO3 variant, the authors proposed a spectrum of odour phenotypes modulated by dietary and environmental factors [32]. In the Japanese cohort of Shimizu (2014), the severe form was caused by FMO3 loss of function, whereas the moderate and mild forms resulted mainly from factors other than genotype [40].

Heterozygosity calls for an age-related nuance: GeneReviews describes heterozygotes as asymptomatic and not at risk of developing the disease, but in the differential diagnosis it acknowledges mild symptoms in young children heterozygous for a loss-of-function variant or with certain combinations; the case of an infant with a truncating variant and the two maternal polymorphisms illustrates manifestation under overload [1, 39]. The NHGRI (National Human Genome Research Institute) states that carriers may have mild symptoms or temporary episodes [41].

2.8. Genes beyond FMO3, allelic disorders, and pharmacogenetics

Exome sequencing has raised the hypothesis that variants in other genes disturb TMA metabolism: in individuals with biochemical trimethylaminuria and no detectable FMO3 variants, variants of predicted pathogenicity were found in genes of oxidoreductase pathways; of 10 individuals with impaired TMA oxidation, 5 had rare variants in these genes [3, 11]. No other phenotype is known to be associated with germline FMO3 variants [1]. FMO3 oxygenates a wide range of nitrogen- and sulfur-containing drugs and dietary compounds; severe variants impair the metabolism of substrate drugs, and the common haplotype may increase the efficacy or toxicity of some [6, 31]. Severe nonsense and missense variants have been associated with altered metabolism of other nitrogen- or sulfur-containing substrates, including biogenic amines [23]. Therapeutic inhibition of hepatic FMO3 with antisense oligonucleotides has been proposed for obesity on the basis of results in mice, which raises the theoretical possibility of a trimethylaminuria phenocopy, not yet evaluated [42].

3. Clinical Spectrum and Manifestations by System

3.1. The odour and the absence of physical signs

The sole manifestation of trimethylaminuria is a rotten or rotting-fish odour, sometimes described as pungent and ammoniacal, resulting from the excretion of TMA in urine, breath, sweat, and reproductive fluids, including vaginal secretions; TMA is also excreted in saliva and may be produced in the oral cavity [1, 3, 4]. Patients look normal and healthy and have no physical symptoms [1, 9]. Most have a moderate odour of variable intensity, and some have a constant odour; in a British survey, patients sometimes described the odour as faecal or like rubbish rather than fishy [17, 41]. Many learn empirically to control the odour (avoiding marine fish and washing frequently) and may have little odour at the clinic visit; some do not perceive their own odour, and cases have been identified without evident odour [8]. This variability relates to the specific anosmia for TMA in about 7% of people and to the episodic nature of the symptom [8, 17].

3.2. Natural history by clinical form

The severe form is characterised by a persistent odour, usually from early childhood, sometimes with later onset, aggravated by sweating and menstruation [3]. Although the disease is present from birth, it becomes apparent when foods rich in choline or marine fish are introduced; Chalmers and colleagues report that persistent trimethylaminuria in children is rare in their experience [12]. Symptoms may worsen at puberty and, in women, are more intense immediately before and during menstruation, after oral contraceptives, and in the perimenopause [1, 41]. The NHGRI notes that the disease appears to be more common in women [41]. Breastfed infants may develop an odour after the mother eats shellfish [1].

The mild or moderate form is the most common, with unknown incidence; the odour is episodic, may begin at any age, and often appears at puberty or in the perimenstrual period [3]. The transient form has occurred in preterm newborns fed choline-supplemented formulas and resolved when the supplement was withdrawn; the transient forms of infants and young children resulted from physiologically low enzyme expression combined with gene polymorphisms, and a lifelong risk of recurrence may persist [1, 3]. Mayatepek and Kohlmüller described two healthy, unrelated children with transient trimethylaminuria without N-oxidation deficiency, suggesting that the transient form may be common and requires follow-up [43].

The perimenstrual form was studied longitudinally in Japan in five women aged 21 to 37 years over 120 days. A woman homozygous for p.(Arg500Ter) maintained an FMO3 metabolic capacity of 13 ± 10% (mean ± standard deviation, 20 collections); a woman homozygous for the p.[(Glu158Lys;Glu308Gly)] haplotype maintained about 90% except on a few days around menstruation; a woman heterozygous for the haplotype maintained values above 90%, falling to 60–70% around menstruation; two women with the reference genotype maintained values above 90% except around menstruation [27]. Mackay and colleagues cite a study of six healthy women and six men with daily TMA measurement for one month, in which all the women had a perimenstrual increase [8]. At the Monell Center odour clinic, menstruating women with a positive test tended to produce more TMA in the 7 days around menstruation (P = 0.07) [44].

The secondary forms include severe chronic liver disease and, rarely, viral hepatitis; advanced renal failure with small intestinal bacterial overgrowth; urinary tract infection with bacteria that degrade TMAO; intestinal anatomical anomalies with blind loops; vaginitis; and gingivitis [3]. Trimethylaminuria acquired after hepatitis appears in adulthood, with no personal or family history, and the metabolic abnormalities persist after the liver disease resolves [1]. Mackay and colleagues describe three cases of adult onset after possible viral hepatitis and a Japanese cohort with elevated liver enzymes that had more TMA and lower oxidative capacity [8, 45]. Precursor overload has occurred in patients with Huntington disease or Alzheimer disease treated with oral choline at doses of 8–20 g/day; GeneReviews gives doses of 20 g/day or more [1, 8]. High-dose betaine in homocystinuria is associated with a fishy odour, and pharmacological doses of L-carnitine can cause it [8, 46, 47]. A 58-year-old man with Child-Pugh A cirrhosis who took 1000 mg of L-carnitine before each workout developed an acute fishy odour, noticed by colleagues but not by himself [13]. An infant with food protein–induced enterocolitis syndrome (FPIES) had an intense fishy odour only during episodes [39]. A patient with features of both primary and secondary trimethylaminuria has also been described: biochemistry consistent with a secondary form, three FMO3 polymorphisms (two of them activity-reducing), diagnosis without an evident characteristic odour, and a temporary response to metronidazole and neomycin [48].

3.3. Psychosocial impact and mental health

The social and psychological impact is the main source of morbidity. In childhood, being avoided, ridiculed, or bullied at school is associated with aggressive or disruptive behaviour and poor school performance; later come shame, low self-esteem, reluctance to seek medical help, avoidance of contact, isolation, loneliness, frustration, depression, and difficulty starting and maintaining relationships; in extreme cases, paranoid behaviour, despair, and suicidal tendencies have been described [1, 3]. Bouchemal and colleagues add alcoholism and obsessive behaviour, and Mackay and colleagues record rare suicide attempts [3, 5, 8].

Qualitative and survey studies quantify this impact in small or self-selected samples. In Ireland, a phenomenological study from the National Centre for Inherited Metabolic Disorders (NCIMD) included 6 adults with a confirmed diagnosis (3 men, 3 women; median age 36.5 years, range 30 to 75 years in the text and 30 to 72 in the table): the path to diagnosis took years to decades, most had been dismissed by health professionals, the diagnosis brought relief and validation, and participants described daily fear, anxiety, and paranoia, negative automatic thoughts, rituals (menthol lozenges, excessive showering, perfume), social avoidance, deliberate breaking off of intimate relationships, alcohol use in one participant, and suicidal ideation in another; one participant had diagnosed depression and anxiety, one undiagnosed depression, one obsessive-compulsive disorder, and three no psychiatric diagnosis [49]. In the United Kingdom, an anonymous online survey by University College London with the MEBO patient group included 44 adults (34 women, 9 men, 1 no answer) and 3 parents or guardians: symptom onset ranged from 5 to 60 years of age (median 19); 32 (73%) had undergone urine testing, and of those tested 17 (52%) had trimethylaminuria (5 primary, 11 secondary, 1 unclassified) and 16 (48%) had a negative test; 23 reported difficulties or delays in obtaining the test; all 40 who answered the question described bullying, harassment, or ostracism, at work (36; 90%) and in social settings (35; 88%) [17]. The sample includes people without a confirmed diagnosis, so these percentages describe those seeking help for odour and not only patients with trimethylaminuria [17].

A 2026 preprint, not peer reviewed, gathered 401 valid responses from more than 40 countries (61.1% women); 65.8% had neither an official diagnosis nor a clear medical explanation, 78.8% reported physical rejection or social exclusion, 61.1% bullying or demeaning treatment, and 84.3% considered treatments inadequate; in the cognitive follow-up (n = 306), 68.0% reported cognitive difficulties several times a week or more [50]. The lack of diagnostic confirmation in most respondents precludes attributing these complaints to trimethylaminuria [50].

3.4. Other reported associations

Mackay and colleagues cite an unpublished study of 104 African Americans with more uncontrolled hypertension, with no difference in oxidation ratios, and anecdotal reports of an association with hypertension; they consider it uncertain whether the mood disorder results from social isolation or from a biochemical component [8]. Patients with primary trimethylaminuria have abnormal benzydamine metabolism; alterations in other FMO3 substrates would be expected, but no such effects have been reported [1].

4. Clinical Recognition and Red Flags

4.1. When to suspect

Suspicion arises from a complaint of body odour or halitosis without an evident cause. In a US odour clinic (Monell Center), about one-third of 353 patients with idiopathic odour (118) had a positive choline challenge; women, especially African American women, were overrepresented among both positives and negatives, so neither ancestry nor sex helps select whom to test, and the authors propose the test for all patients with idiopathic odour [44]. In an odour clinic in Philadelphia, 35% of those referred had trimethylaminuria [2]. These figures come from referral cohorts, not from population samples.

By age and setting, the scenarios described are: the infant or young child with a fishy odour after the introduction of marine fish, eggs, or other choline-rich foods, or the breastfed infant after the mother's shellfish meals [1, 12, 38]; the preterm newborn fed a choline-supplemented formula [3]; the infant with odour only during FPIES episodes [39]; the adolescent girl or woman with perimenstrual worsening, with oral contraceptives, or in the perimenopause [27, 41]; the adult with cirrhosis, portosystemic shunt, uraemia, or bacterial overgrowth [1, 8]; and the patient treated with betaine, L-carnitine, or choline at pharmacological doses (8–20 g/day) [8, 13, 47]. Fever, infection, exercise, stress, and emotional upset aggravate the odour through sweating and increased TMA production [1, 8].

4.2. Features favouring true trimethylaminuria

In the Hôpital Necker cohort, the features that distinguished children with trimethylaminuria from adults with an unconfirmed complaint were odour confirmed by others (parents or clinicians), onset in the first year of life, and elevated urinary TMA; the adult complaining of an odour not perceived by others, who interprets other people's behaviour as a reaction to their odour and has psychiatric comorbidity, should be evaluated for olfactory reference syndrome and referred for psychological or psychiatric care [5]. Even so, the odour of trimethylaminuria is episodic and may be absent at the visit, and some patients do not detect it; in the British survey, 17 of the 23 patients who had difficulty obtaining the test attributed this to the primary care physician's lack of awareness [17]. Wise and colleagues acknowledge that undiagnosed trimethylaminuria may exist among people diagnosed with olfactory reference syndrome, so biochemical evaluation should precede a psychiatric conclusion [44].

4.3. Red flags and emergency signs

Trimethylaminuria has no acute organic complications; the red flags are psychiatric and social: depression, suicidal ideation (reported in one of six Irish patients), isolation, alcohol use, and dropping out of school or work, all of which warrant prompt mental health evaluation [5, 17, 49]. In the secondary setting, a fishy odour associated with advanced liver disease may accompany fetor hepaticus and hepatic encephalopathy, and elevated blood TMA in uraemia may contribute to neurological manifestations; in these patients the warning sign is the underlying disease [1].

4.4. Newborn screening

There is no newborn screening for trimethylaminuria in the sources consulted [4]. FMO3 genetic testing is not listed in the National Genomic Test Directory for rare diseases in England, and the NHS (National Health Service) has no standardised diagnostic pathway; the authors propose a pathway from primary care to a specialist unit with robust testing, defined criteria, and genetics when needed [17].

5. Diagnosis: Criteria, Testing, and Algorithms

5.1. Principles

The diagnosis cannot rely on the examiner's sense of smell: the odour is often episodic; the human nose is very sensitive (some people detect one part in 10^9), but olfaction is subjective, some people cannot detect TMA, and the odour may include other compounds [1]. The first-line approach is urinary analysis of TMA and TMAO, and biochemistry should remain the first step even with wider access to sequencing [3, 11]. Before testing, poor body and oral hygiene, gingivitis, vaginosis, and urinary tract infection should be excluded, and other inborn errors of metabolism with an odour should be considered [3]. There is no formal society clinical guideline: GeneReviews states that no clinical practice guidelines have been published, and the guidance documents are the EuroGentest clinical utility gene card (2015), the paediatric "good practice" recommendations of Chalmers and colleagues (2006), and laboratory reviews [1, 2, 8, 12].

5.2. Sample collection and loading tests

Urinary TMA excretion should ideally be tested on two separate occasions on a normal diet; a choline-rich meal before the test, for example 2 eggs with 400 g of baked (haricot) beans or soybeans, may help [1, 2, 4]. Women should not be tested at the onset of or during menstruation, because menstruation causes transient trimethylaminuria even in unaffected women [1, 27]. Urine is collected in a container with 2 mL of 6 M HCl (urinary pH 2.0), which converts free TMA into a nonvolatile salt, and frozen until analysis [3]. A spot sample is used when the odour is evident; otherwise, a random collection 2–12 h after 300 g of marine fish or a choline-rich meal, a timed 6–8-h collection (Nijmegen protocol), or a collection after 5 g of choline in orange juice; the test is repeated if negative or borderline and may need to be repeated during periods with odour [3, 8]. The load is essential because many patients already restrict their diet, and a combined excretion below a few hundred µmol/mmol creatinine may give a misleading result [8]. The Irish cohort used first-morning urine on a normal diet, without choline restriction and outside the perimenstrual period [11]. Hôpital Necker used a 3-day dietary load (eggs, cabbage, fish, shellfish) with collection on the fourth day [5].

Chalmers and colleagues compared a choline load with a marine fish meal in six children: the fish meal is a simple, acceptable method for confirming the diagnosis in children, with effects cleared more quickly than with choline, whereas an oral load of choline bitartrate allows estimation of residual oxidative capacity in vivo and complements molecular studies [12]. In that series, homozygotes for p.(Pro153Leu) had virtually no residual N-oxidation capacity, a patient with p.(Met82Thr) had some, a compound heterozygote for two novel variants, p.(Gly193Glu) and p.(Arg483Thr), had considerable residual capacity, and another patient, heterozygous for two novel variants, had an odour and an elevated TMA/TMAO ratio at baseline but a negative response to both loads, showing that a negative loading test does not exclude the diagnosis [12]. Direct TMA loading also has historical value: 600 mg distinguishes obligate carriers from noncarriers and from affected individuals (300 mg does not discriminate and 900 mg saturates the enzyme), and load failures result from intestinal colonisation, transit time, and gastric emptying [8]. According to the NHGRI, after 600 mg of TMA carriers excrete 20–30% of total TMA as the free amine and noncarriers less than 13%; Schmidt and Leroux report about 25% in heterozygotes after a load, compared with about 4% without a load [14, 41]. A TMAO load shifts total TMA towards the free form in affected individuals and carriers, but it is not a recognised diagnostic tool [14]. In the 2025 South African case, the direct load consisted of 1.3 mL of 43–49% TMA solution in 250 mL of orange juice, with baseline and hourly collections over 4 h under medical supervision and outside the menstrual period [37]. The gene card records one individual with fever and vomiting after the choline challenge [2].

5.3. Analytical methods

TMA and TMAO have no chromophore; specific methods require sophisticated equipment and experienced staff [1, 14]. Proton nuclear magnetic resonance (NMR) spectroscopy measures TMA and TMAO in the same spectrum, directly in urine, without extraction; acidification separates TMAO from glycine betaine, aluminum chloride prevents splitting of the TMA peak, and the detection limit is often above 10 µmol/L [1, 8]. In the Necker protocol (500 MHz; 600 µL of urine with 100 µL of D2O, no pretreatment), TMA appears at 2.92 ± 0.02 ppm, TMAO at about 3.27 ± 0.03 ppm (pH-dependent), and creatinine at 3.05 ppm, and samples frozen at −20 °C remain measurable [5]. NMR allows trimethylaminuria and dimethylglycine dehydrogenase (DMGDH) deficiency to be diagnosed in the same run [3, 5]. Alternatives are headspace gas chromatography (with reduction of TMAO to TMA by titanium(III) and nitrogen–phosphorus detection, 7 times more sensitive than flame ionisation), electrospray ionisation tandem mass spectrometry after derivatisation with ethyl bromoacetate, MALDI-TOF with iodomethane, and liquid chromatography–tandem mass spectrometry (LC-MS/MS) on diluted spot urine [1, 8]. A stable isotope dilution method for rapid sequential analysis was described in 1995 [51]. In the United Kingdom, the measurement is not routinely available in the NHS [17]. In the USA, the NHGRI page lists the biochemical genetics laboratory of Children's Hospital Colorado and the Monell Chemical Senses Center, the latter with a long waiting list; the same page, dating from 2018, describes genetic testing as available only in research laboratories, which is outdated information [41].

5.4. Metrics, severity cutoffs, and odour threshold

The result is expressed in µmol/mmol creatinine and as N-oxygenation capacity, calculated as TMAO/(TMA + TMAO) × 100, or, in complementary fashion, as the percentage of total TMA excreted as the free amine, TMA/(TMA + TMAO) × 100 [3, 8]. The two percentages add up to 100%, and some articles write one formula when the values correspond to the other; in the South African case, for example, the published formula is that of free TMA, but the values are oxidation capacity [37]. All the ratios are needed, because the TMAO/creatinine ratio depends heavily on diet [5]. The normal values cited by Bouchemal from Chalmers are TMA/creatinine below 10 mmol/mol, TMAO/creatinine of 50–1000 mmol/mol, TMA/TMAO below 0.1 mol/mol, and TMAO/(TMA + TMAO) above 0.8; a healthy individual converts at least 95% of TMA to TMAO, and no Japanese control had values below 95% [5, 35, 52].

Table 5.1. Published severity classification schemes (TMAO% = TMAO/(TMA + TMAO) × 100; free TMA% = 100 − TMAO%).

The divergence between schemes reflects differing populations, methods, and collection conditions; the free-TMA scheme of 0–9%, 10–39%, and more than 40% is the one adopted by the European gene card and by GeneReviews for deciding on genetic testing, and it is the one this article uses as its reference, mentioning the others when a study applies them [1, 2]. The severity cutoff differs only at the boundary (>40% in GeneReviews, ≥40% in Saudubray), with no relevant practical consequence [1, 3].

As for the odour threshold, GeneReviews, citing Mitchell and Smith (2001), gives a urinary free TMA concentration of 10 µg/mL or more, equivalent to 18–20 µmol/mmol creatinine and about 15–20 mg/day [1]. Saudubray states that the odour may be detectable with TMA above 10 µmol/L and 18–20 µmol/mmol creatinine [3]. Given the molar mass of TMA (59.11 g/mol), 10 µg/mL corresponds to about 169 µmol/L, so the two absolute concentrations differ by more than an order of magnitude and one of them probably results from a unit error; the original article was not consulted. The creatinine-normalised value, 18–20 µmol/mmol, is identical in both sources and in Mackay and colleagues, and it is the threshold used in this article; Saudubray's value of 10 µmol/L coincides with the usual NMR detection limit [1, 3, 8].

5.5. Diagnostic criteria

GeneReviews establishes the diagnosis of primary trimethylaminuria in a proband who, on a normal diet, excretes more than 10% of total TMA as the free amine and has known biallelic pathogenic loss-of-function variants in FMO3 [1]. The European gene card recommends offering genetic testing to anyone who, on a normal diet, excretes more than 10% of total TMA as the free amine, and states that molecular testing identifies affected individuals and carriers and distinguishes the severe primary form (loss of function) from the mild form (compound heterozygosity for mild missense variants and polymorphisms) and from the secondary forms [2]. The Irish cohort applies the equivalent criterion of conversion below 90% [11]. Biochemistry does not distinguish TMAU1 from TMAU2; FMO3 genetics identifies TMAU1 [4].

The added value of genetic testing in severe cases has been questioned: with persistent severe symptoms and elevated TMA even on a low intake, genetics adds little to the diagnosis [3, 8]. Even so, a molecular result provides firm evidence that can speed up school and psychological support decisions and allows early testing of asymptomatic siblings; in the mild or intermittent forms, knowing the genetic contribution may help [3, 8]. The Irish cohort warns of the opposite risk: identifying common variants in the mild forms may lead to excessive choline restriction [11].

5.6. Genetic testing: methods and limitations

For clinical suspicion, GeneReviews proposes single-gene sequencing followed by targeted deletion/duplication analysis if only one or no variant is found, or a multigene panel; when the disorder has not been suspected, the diagnosis comes from exome sequencing (the most commonly used route) or genome sequencing [1]. The gene card describes amplification and bidirectional sequencing of the eight coding exons, with MLPA (multiplex ligation-dependent probe amplification) for the rare gross deletions [2]. The Irish cohort sequenced the entire coding region and exon–intron boundaries, with classification according to ACMG/AMP (American College of Medical Genetics and Genomics/Association for Molecular Pathology) criteria and pretest counselling [11].

Sequencing detects pathogenic variants in about 99% of probands with known variants (HGMD data) [1]. The gene card gives an analytical sensitivity of about 99% for coding variants, with no analytical specificity data, and an unknown clinical sensitivity, because the secondary forms do not give a positive test [2]. These proportions refer to detectable variants and are not equivalent to the clinical sensitivity of the test for trimethylaminuria. The positive predictive value is high (all individuals biallelic for variants that abolish activity have the severe form), and the negative predictive value approaches 100% in relatives when the proband's variants are known [2].

A negative result does not exclude the disease: in a Japanese study cited by the gene card, 10 of 13 individuals with the severe form had a causal biallelic genotype, 3 severe cases had no coding variant, and none of the 65 moderate or mild cases was biallelic [2]. Unresolved cases may be due to deep intronic variants (detectable by RNA sequencing), regulatory variants, untested deletions/duplications or structural rearrangements, or variants in other oxidoreductase genes [3, 11]. Biallelic variants of uncertain significance (VUS), or a pathogenic variant with a VUS, neither establish nor exclude the diagnosis [1]. A novel variant requires functional evidence (heterologous expression with measurement of TMA N-oxygenation) and, when two heterozygous variants are present, parental testing to determine phase [2]. There are no global classification guidelines for FMO3 variants, particularly for the polymorphisms [4]. Common polymorphisms can be misleading: in Japan, only 7 of 640 volunteers with a self-reported complaint had FMO3 variants and only 19 had less than 40% of normal metabolic capacity; in a study of 10 patients whose odour was not detected by trained sensory judges, 7 had FMO3 variants that correlated weakly with the ratios [5].

The report should state the gene, the versioned transcript (NM_001002294.3, MANE Select), the HGVS c. and p. nomenclature, the zygosity and, when determined, the phase (cis or trans) of the variants, and the ACMG/AMP class; for haplotypes, describing the cis configuration is essential, because p.(Val187Ala) or the p.(Glu158Lys)/p.(Glu308Gly) pair are meaningful only in combination [2, 7, 11, 35]. Faecal metagenomics by next-generation sequencing has been proposed as an adjunct, in a research setting [4].

5.7. Yield in cohorts

Table 5.2. Biochemical and genetic diagnostic yield in published cohorts.

5.8. Carrier detection

On a normal diet, carriers and unaffected individuals excrete less than 10% as free TMA and cannot be distinguished biochemically; carriers are identified by molecular testing once the familial variants are known [1]. A 600-mg TMA load distinguishes carriers in research (all obligate carriers increased TMA excretion, whereas healthy volunteers did so only at 900 mg or more), but it is not a routine clinical test [8, 20, 39].

5.9. Practical algorithm

Table 5.3. Diagnostic sequence proposed from the sources.

5.10. Prenatal and preimplantation diagnosis

Prenatal diagnosis is technically possible once the familial variants have been identified, but it is not indicated, given that this is a condition without physical morbidity [2, 3]. GeneReviews states that prenatal diagnosis and preimplantation genetic testing are possible after identification of the variants and that perspectives on their use differ, the decision being a personal one [1].

6. Differential Diagnosis

6.1. Biochemical principle

The differential rests on two questions: whether urinary TMA is increased and, if so, whether the TMA/TMAO ratio is altered [1]. Without increased urinary TMA, the odour has another origin: poor hygiene, gingivitis, blood-borne halitosis from compounds other than TMA, DMGDH deficiency (with low urinary TMA and a normal ratio), or methanethiol oxidase deficiency [1, 46]. With increased TMA but a normal TMA/TMAO ratio, the causes are urinary tract infection, bacterial vaginosis, advanced liver or kidney disease, cervical carcinoma, and precursor overload; primary trimethylaminuria is characterised by a high TMA/TMAO ratio [1, 4]. In practice, the secondary form with coexisting polymorphisms may have an altered ratio, and a history of precursor intake is indispensable [2, 48]. Molecular diagnosis distinguishes the primary form from the others [1].

Table 6.1. Differential diagnosis of fishy body odour or halitosis.

6.2. Dimethylglycine dehydrogenase deficiency

DMGDH and sarcosine dehydrogenase are mitochondrial flavoenzymes with covalently bound FAD that sequentially remove two methyl groups in choline catabolism, transferring them to tetrahydrofolate, which also explains the need for folate when choline is restricted [3]. The first patient described was an adult with a fishy odour since age 5, aggravated by stress and exertion, with major social, psychological, and occupational problems, chronic muscle fatigue, and a persistently elevated creatine kinase (CK) about 4 times normal, with normal intelligence; he was homozygous for the variant historically described as 326A>G (H109R) (historical numbering, = NM_013391.3:c.326A>G p.(His109Arg) in DMGDH, confirmed in VariantValidator) [3, 7]. The second patient was an 11-month-old infant without odour, with CK 10 times the upper limit, diagnosed by exome sequencing [3]. Dimethylglycine accumulates about 100-fold in plasma and 20-fold in urine and is not detected by routine metabolic screening; reference values are 1–5 µmol/L in adult plasma and, in urine, below 550 µmol/mmol creatinine from 0 to 2 months and below 26 µmol/mmol from 2 months onwards [3]. Serum dimethylglycine also rises in folate deficiency (up to 10-fold), cobalamin deficiency (up to 2-fold), and renal failure (up to 2-fold) [3]. Proton NMR diagnoses both deficiencies, and in the Necker spectrum dimethylglycine was detectable but never quantifiable, while dimethylamine appears at 2.73 ppm [3, 5]. The interpretation of the odour in dimethylglycinuria is debated: in the case of homocystinuria treated with betaine, markedly elevated dimethylglycine persisted after the odour resolved, indicating that in that patient TMA, not dimethylglycine, accounted for the odour [47]. Treatment of DMGDH deficiency includes counselling, choline restriction, and avoidance of sweating, as in trimethylaminuria; antibiotics are not indicated, the adult patient did not benefit from riboflavin, and the infant received high-dose riboflavin with folic acid for 6 months without normalisation of CK [3].

6.3. Methanethiol oxidase deficiency (SELENBP1)

Selenium-binding protein 1 (SELENBP1) is a methanethiol oxidase that converts methanethiol into hydrogen peroxide, formaldehyde, and hydrogen sulfide [54]. Pol and colleagues identified SELENBP1 variants in five patients from three sibships with cabbage-like breath, attributed to elevated concentrations of methanethiol and dimethyl sulfide in exhaled air, with increased urinary excretion of dimethyl sulfoxide; fibroblasts had little protein and deficient activity, corrected by lentiviral expression of the gene, and the Selenbp1 knockout mouse reproduced the biochemistry [54]. Inheritance is autosomal recessive; in two consanguineous families there were additional neurological problems of uncertain causal relationship, and the condition is probably underdiagnosed owing to the scarcity of specialised assays [46]. Diagnosis requires demonstration of increased methanethiol and dimethyl sulfide in exhaled air and dimethyl sulfone and dimethyl sulfoxide in urine, or of the enzyme deficiency in fibroblasts, and SELENBP1 analysis may be the initial test; metronidazole reduces the malodorous compounds, and methionine restriction is another option [46]. Saudubray includes methanethiol oxidase deficiency in the list of inborn errors with abnormal odour (extraoral halitosis) [55]. The Irish cohort cites dimethylglycinuria and SELENBP1 deficiency as differentials to consider in the presence of fishy odour or halitosis [11].

6.4. Olfactory reference syndrome

Olfactory reference syndrome is a preoccupation with a body odour that others do not perceive, with distress and impairment, and it overlaps with schizophrenia, social phobia, obsessive-compulsive disorder, and delusional disorder; in Japan it corresponds in part to "taijin-kyofusho"; selective serotonin reuptake inhibitors and antipsychotics reduce symptoms [5]. Most people who complain of body odour end up with this diagnosis [5]. In the Necker cohort, of 11 adults whose complaint was not confirmed by clinicians, onset was after age 16 in 10 (one at age 9), 10 interpreted other people's behaviour as a sign of odour, all had a psychiatric history (depressive symptoms in 6), 8 reported an impact on their sex life, and almost all scored below 70 on the Global Assessment of Functioning (GAF) scale [5]. The distinguishing feature indicated by GeneReviews is late onset [1]. The overlap with undiagnosed true trimethylaminuria makes it mandatory to obtain biochemistry under load before reaching a conclusion [44].

6.5. Treatable secondary forms

Congenital portosystemic shunt deserves emphasis because the trimethylaminuria may resolve with endovascular closure [1]. Cirrhosis, hepatocellular dysfunction, and shunts impair the clearance of intestinal TMA, and the resulting TMAU may contribute to hepatic encephalopathy and fetor hepaticus; in uraemia, TMA release increases through small intestinal bacterial overgrowth and renal clearance decreases [1]. Omega-3 fatty acids may leave a fishy taste, which is a taste complaint and not trimethylaminuria [57]. In the United Kingdom, 48% of people tested for an odour complaint had a negative result, which calls for investigation of other causes of body odour and halitosis [17].

7. Treatment and Risk Management Strategies

7.1. Goals, evidence, and guidelines

There is no cure or disease-modifying treatment, and no medicinal product is approved for trimethylaminuria in the USA or the EU [14, 15, 41]. GeneReviews states that no clinical practice guidelines have been published, GeneGuidance classifies the available recommendations as expert consensus without randomised trials, and the search of 27 October 2026 identified no new guideline from the SSIEM (Society for the Study of Inborn Errors of Metabolism), MetabERN, the ACMG, or any other society; a negative search does not prove absence [1, 6, 16]. The guidance documents are the 2015 European gene card, focused on diagnosis; the paediatric "good practice" recommendations of Chalmers and colleagues (2006), of which only the abstract was read; the NHGRI information page (updated 18 December 2018); and the reviews by Mackay (2011), Fennema (2016), Schmidt and Leroux (2020), and Sidoti (2024) [2, 4, 8, 12, 14, 28, 41]. Strategies fall into four targets: limiting precursors, protonating excreted TMA, acting on intestinal metabolism, and acting on FMO3 [4, 14]. The individual response to restrictions differs and should be assessed by urinary TMA and TMAO; in general, the more severe the reduction in enzyme activity, the poorer the response to treatment [1]. In the Irish cohort, the mild and moderate forms responded to diet and acidic soaps, and the severe forms may require antibiotics and riboflavin [11]. Precursor restriction combined with frequent washing with acidic soaps is the mainstay of treatment; it may suffice in the mild or temporary forms, but offers limited benefit in the primary form [14].

7.2. Precursor-restricted diet

Precursor restriction is effective in some cases, especially in the mild and moderate forms [1, 3]. The primary measure is to exclude marine fish and other seafood (fish, cephalopods, crustaceans), the main source of TMAO; freshwater fish contain less TMAO and are not a problem [1, 8, 41]. Choline-rich foods to limit are eggs (yolk), liver, kidney, peas, beans, peanuts, soy, brassicas (Brussels sprouts, broccoli, cabbage, cauliflower), and rapeseed products (oil and meal); sources of lecithin include brassicas, soy and other legumes, and rapeseed and sunflower oils [1, 3]. Brassicas are doubly inadvisable, because they contain precursors as well as indoles that inhibit FMO3, and tannins and Brussels sprouts inhibited human FMO3 with prolonged administration [1, 3, 14]. Milk from wheat-fed cows may contain significant amounts of TMA, and lecithin and fish oil supplements containing lecithin should be avoided [1, 41]. L-carnitine is abundant in red meat, poultry, and some dairy products, and some centres advise avoiding them [14, 49]. Precursor bioavailability differs: humans make poor use of the choline and carnitine in foods such as red meat and peanuts, but efficiently extract TMAO from marine fish and shellfish, so the priority is marine fish, shellfish, and foods containing free choline, such as soy milk [28]. Dietary L-carnitine did not increase urinary methylamines in one study, nor did 11.65 g of lecithin (13% choline), whereas 2.1 g of choline did [8]. In a Brazilian patient, tuna and eggs did not increase TMA, but cação (dogfish), which is rich in TMAO, did, showing the need to define which fish to avoid [58].

Choline is essential for the synthesis of phospholipids and acetylcholine and cannot be eliminated from the diet: healthy people on a choline-deficient diet develop signs of incipient liver dysfunction within 3 weeks [14]. The diet should be planned and monitored so that choline and folate meet the recommendations for age and sex; the adequate intake of choline in adults is 550 mg/day for men and 425 mg/day for women [1]. Choline restriction increases the requirement for folate, a methyl donor, and Saudubray recommends supplementing it; the sources consulted do not give a folic acid dose specific to trimethylaminuria, and the goal is to meet the recommended intake for age and sex [1, 3, 8]. Moderate restriction is tolerable and often effective in the mild and moderate forms; more restrictive diets should be supervised by a dietitian, because of the risks to the brain and liver, and diets without professional guidance are potentially harmful [3, 4, 8]. Choline- and betaine-defined diets exist, validated by laboratory analysis of aliquots of the total diet and well accepted for up to 2 months, applicable to counselling patients who need a controlled diet [59]. Adherence is often suboptimal, and the restrictions are difficult to maintain [11, 17]. Choline should not be restricted during pregnancy, breastfeeding, or early childhood [1, 3].

7.3. Hygiene and protonation of TMA

Soaps and lotions with a pH close to that of the skin (5.5–6.5) keep TMA in its salt form, which is less volatile and can be washed off [1, 41]. Although mentioned only anecdotally in the literature, frequent washing with low-pH soaps is, together with diet, the standard measure [14]. Additional measures are antiperspirants, light clothing, good ventilation in warm environments, and regular laundering of clothes; patients should not take unprescribed supplements (fish oil capsules, carnitine, "natural" products rich in lecithin and choline) and should promptly report adverse drug reactions [3, 8]. Factors that promote sweating (exercise, stress, emotional upset) should be avoided as far as possible [1, 41]. Clothing with an activated charcoal layer has been patented [4]. Masking the odour with (E,E)-2,4-undecadienal, derived from coriander leaves, deodorised TMA up to 1 mM at concentrations of 10 ppm, but was toxic to human fibroblasts at 25 µM [14].

7.4. Short courses of oral antibiotics

Antibiotics reduce the TMA-producing flora and are useful in short courses when the diet has to be relaxed (social occasions) or when TMA production increases (menstruation, infection, emotional upset, stress, exercise) [1, 3]. They should not be taken continuously: prolonged use promotes resistance, dysbiosis, and increased intestinal permeability, and Schmidt and Leroux consider that the benefit–risk ratio may not be acceptable for chronic use, given that trimethylaminuria does not affect vital functions [3, 4, 14]. Chalmers and colleagues compared ex vivo and in vivo the effect of metronidazole, amoxicillin, and neomycin on bacterial TMA production from choline: all reduced it to a limited extent, and neomycin was the most effective [1, 12]. Reports of metronidazole given for weeks show variable results, with urinary TMA reduced to one-third in some patients and no effect in others, and better results with diet or with a second antibiotic; neomycin alone lowered TMA in mice and in small groups of patients, but values returned to baseline after discontinuation; rifaximin reduced TMA and was well tolerated [4, 14]. Treacy and colleagues (1995) described a clinical and biochemical response to metronidazole, and a patient with primary and secondary features had a temporary response to metronidazole and neomycin [48, 51]. The NHGRI mentions low-dose antibiotics, without specifying doses [41].

No accessible source publishes antibiotic doses specific to trimethylaminuria; the abstract of Chalmers and colleagues does not include them, and the full text was not obtained. In the absence of a condition-specific dose, Table 7.1 presents the dose from the official label (summary of product characteristics [SmPC] from INFARMED, and the US Food and Drug Administration [FDA] label), with the approved indication used as the reference and the extrapolation stated; all uses in trimethylaminuria are off-label. The actual regimen (dose, duration, and repetition) is a decision for the metabolic disease specialist, and the courses described in the literature are short and intermittent [3, 14, 37].

Table 7.1. Antibiotics and an osmotic laxative used to reduce intestinal TMA production: label doses as an extrapolated reference (off-label use).

In Portugal there is no oral neomycin monotherapy suitable for suppressing the intestinal flora: Infomed lists only topical, ophthalmic, and otic combinations and Dimicina® (bacitracin 35.2 mg with neomycin 18.25 mg, tablet), whose dose is not suitable for this purpose [64]. In Brazil, open data from ANVISA (Agência Nacional de Vigilância Sanitária, the Brazilian Health Regulatory Agency) show active registrations of metronidazole (Flagyl® and generics), rifaximin (Xifaxan®), and six registrations of neomycin whose dosage form is not stated in the file; the text of the Brazilian package inserts was not consulted [62]. Metronidazole is used chronically in methylmalonic and propionic acidemias to reduce metabolites produced by anaerobes, and has been tested in the short term in hepatic encephalopathy; neomycin has been FDA-approved for overt hepatic encephalopathy since the late 1960s, with nephrotoxicity in about one-third of patients; rifaximin was approved by the FDA for hepatic encephalopathy in 2010 [14]. Unlike systemically available antibiotics, rifaximin acts locally in the intestinal lumen, which minimises the risk of adverse effects [14, 66].

7.5. Riboflavin

Riboflavin (vitamin B2) is the precursor of FAD, the prosthetic group of FMO3, and may maximise residual activity; it may reduce and even normalise TMA excretion in those with residual function [1, 3]. Mackay and colleagues considered it theoretical in 2011, and Schmidt and Leroux, in 2020, emphasise that the data are limited to isolated cases [8, 14]. The clinical evidence comes from uncontrolled case series. In a 17-year-old girl with homocystinuria treated with betaine (2 × 8 g/day), homozygous for the common haplotype, with urinary TMA of 392 mmol/mol creatinine (normal 2–11) and 93% free TMA, riboflavin improved the odour immediately, without changes to diet or betaine, with TMA of 77 mmol/mol at 30 days (36% free) and 20 at 330 days (7%); the article gives 200 mg/day in the abstract and 100 mg/day in the discussion [47]. At Necker, two biallelic children (aged 3 and 9 years) received 50 mg/day and the odour normalised, with a fall in TMA/creatinine within a few weeks; an adult without biochemical trimethylaminuria who started riboflavin had no effect and discontinued it [5]. At the Irish centre, all patients in the psychosocial study were taking 100 mg once daily with a restricted diet, as centre practice and without analysis of efficacy [49].

Published doses diverge. In the absence of a society guideline, the reference adopted is that of GeneReviews, the expert consensus source closest to a guideline, which matches that of the NHGRI: 30–40 mg 3 to 5 times daily with food (90–200 mg/day), with monitoring of children, because excessive amounts can cause gastrointestinal discomfort [1, 41]. For comparison, the drug chapter of Saudubray gives 100–400 mg/day orally in 2 to 3 doses, and the series used 50 mg/day in children, 100 mg/day in Irish practice, and 100 or 200 mg/day in the British case [5, 47, 49, 72]. There is no drug label for comparison: Infomed returns no authorised medicinal product containing riboflavin alone in Portugal, and the ANVISA open data file has no active registration of riboflavin alone as a medicinal product, so in Brazil and Portugal it is obtained as a food supplement [62, 73].

7.6. Oral sequestrants

In Japan, in 7 patients with a capacity of 70–90%, activated charcoal (1.5 g/day in total for 10 days) reduced urinary free TMA and normalised the ratio during administration, and copper chlorophyllin (180 mg/day in total for 3 weeks) had a longer-lasting effect, maintained for several weeks after discontinuation [52]. GeneReviews and the NHGRI translate these totals into 750 mg of activated charcoal twice daily for 10 days and 60 mg of copper chlorophyllin three times daily after meals for 3 weeks [1, 41]. Mackay and colleagues cite a suggestive reduction in TMA with charcoal in 2 patients and with chlorophyllin in 3; another case confirmed the effect of charcoal, which was rapidly lost after discontinuation [8, 14]. The mechanism of chlorophyllin is complexation of TMA, and ingestion of large amounts of white button mushroom (Agaricus bisporus) and chlorophyllin decreased TMAO-reducing Firmicutes [3, 14]. None of these supplements has been evaluated systematically; Sidoti and colleagues advise against them because they are nonspecific and may bind other metabolites, whereas Schmidt and Leroux list them as supplements recommended to patients, with the same caveat [3, 4, 14]. A patent based on zeolites was abandoned without published data [14].

7.7. Probiotics, microbiota transplantation, and other molecules

Probiotics have no proven benefit [3]. Faecal microbiota transplantation, after pretreatment with metronidazole, in 2 patients attenuated the odour in the first 6 months in only one, with a return to baseline after 1 year [4, 14]. Methanogenic archaea, which reduce TMA to methane, have been patented but not tested in humans; Enterobacter aerogenes ZDY01, Eubacterium limosum, and Clostridium clade XV have been studied only in vitro or in mice; the multistrain probiotic VSL#3 did not prevent the rise in plasma TMAO on a high-fat diet [4, 14, 28]. Desmopressin was associated with decreased odour in one patient, acetylsalicylic acid (81 mg/day for 1 month) lowered plasma TMAO in volunteers without TMA being measured, and meldonium and resveratrol acted in animal models; none has been studied in trimethylaminuria [14]. Allicin and raw garlic juice reduced TMAO production capacity in a one-week human study in people with elevated TMAO [4].

7.8. Psychological and social support

Recognition that this is a medical condition helps patients, and psychosocial support should be part of management [1, 6]. Severely distressed patients may need psychiatric care [3]. The Irish study recommends emphasising psychological intervention; participants who attended the centre's support group rated it positively, and experiences with mental health professionals were mixed [49]. In the United Kingdom, patients are asking for trimethylaminuria to be recognised as an invisible disability under the Equality Act 2010 [17]. Molecular diagnosis can speed up school and psychological support decisions for children [3].

7.9. Drugs and agents to avoid

Drugs metabolised by FMO3 compete with TMA for residual activity; in addition to aggravating trimethylaminuria, reduced drug metabolism may cause adverse effects [1, 17]. Most substrates also have cytochrome P450 or glucuronidation pathways, and FMO3 is not the main route; the exceptions, with metabolism predominantly by FMO3, are benzydamine, itopride, ranitidine, cimetidine, ethionamide, and sulindac sulfide (the active metabolite of sulindac) [3, 31]. FMO3 variants decreased in vivo oxygenation of benzydamine, sulindac sulfide, and methimazole, and patients with primary trimethylaminuria have abnormal benzydamine metabolism [1, 3]. Oral contraceptives aggravate symptoms [1, 4].

Table 7.2. Agents to avoid or to use with caution.

In homocystinuria, betaine has a starting dose of 100 mg/kg/day in children and 6 g/day in adults, in two doses, and doses above 150–200 mg/kg/day bring no additional benefit; high doses are associated with a fishy odour [46]. In the case reported by Manning and colleagues, riboflavin made it possible to continue betaine without odour [47].

7.10. Regulatory status

Table 7.3. Regulatory status of agents used in trimethylaminuria.

7.11. Interventional studies

Table 7.4. Published interventional studies in trimethylaminuria. There are no randomised trials in patients; no study reports confidence intervals for the primary outcome.

These studies have samples of 1 to 16 people, no control group in the patient studies, heterogeneous biochemical or olfactory outcomes, and different analytical methods between measurements (for example, in the Irish study, initial and recent TMA values changed from 20.7 to 0.6, from 12.1 to less than 2.5, from 24.8 to 40.3, from 25.4 to 0.6, from 11.8 to 14.4, and from 35.2 to 30.1 µmol/mmol creatinine, without analysis of efficacy) [49]. They do not allow indirect comparison between interventions or estimation of absolute effect, and the response is poorer in the more severe forms [1, 14]. The polymersome study is a proof of concept in volunteers without the disease, and clinical efficacy remains to be demonstrated [14, 74]. No negative, suspended, or withdrawn trials were found, because there are no registered interventional trials [15].

7.12. New developments and research

A search of ClinicalTrials.gov (27 October 2026) for "trimethylaminuria", "fish odor syndrome", and FMO3 found no interventional trial dedicated to trimethylaminuria; the only result is a generic rare disease registry (NCT01793168) [15]. All therapeutic development is preclinical.

Table 7.5. Lines of research and level of evidence.

In the window from October 2025 to October 2026, two further preprints, not peer reviewed, appeared: the multinational survey described in section 3.3, and a case from Uganda of a 19-year-old woman with a continuous fishy odour for 3 years, beginning shortly after menarche, with elevated urinary TMA and a reduced ratio, managed with precursor restriction, a short course of antibiotics, activated charcoal, and psychological support, with no genotype in the abstract [50, 78]. FCAR, published two months before the start of the window, is the candidate closest to clinical development described by the authors [75].

7.13. Secondary and transient forms

In the secondary forms the cause is treated: withdrawal of L-carnitine and a choline- and carnitine-restricted diet resolved the odour and lowered plasma TMAO from 28.5 to 5.3 µmol/L within 2 months in the cirrhotic patient, and withdrawal of carnitine was considered the most effective measure [13]. A congenital portosystemic shunt can be closed endovascularly, with resolution of the trimethylaminuria [1]. In the transient infantile form, withdrawal of supplemental choline or reduction of fish portions, with gradual reintroduction, was sufficient in the cases described [3, 38]. Lactulose and antibiotics provide relief mainly in the secondary and transient forms [76].

7.14. Pregnancy, lactation, childhood, and transition

Choline should not be restricted during pregnancy or lactation: it is essential for fetal neurodevelopment, large amounts cross the placenta and pass into breast milk, and there is a risk of maternal depletion [1, 4, 8]. In pregnancy, metronidazole is contraindicated in the first trimester and should be weighed carefully thereafter, and it passes into breast milk; aminoglycosides can cause fetal harm; lactulose can be used during pregnancy and breastfeeding [60, 63, 69]. FMO3 activity is induced in pregnancy [28]. In infants and children, restriction should not be excessive, and children taking riboflavin should be monitored for the risk of gastrointestinal discomfort [1]. Rifaximin and lactulose for encephalopathy have no established safety under 18 years of age [65, 69]. FMO3 maturation continues through adolescence, which explains the resolution of the transient forms and justifies reassessing children's diets [8, 38]. The sources consulted do not describe perioperative care specific to trimethylaminuria or protocols for transition to adult care; the issues the sources associate with adolescence and young adulthood are worsening at puberty and with oral contraceptives, and genetic counselling offered to young adults who are affected, carriers, or at risk [1].

7.15. Recommendations of the guidance sources and divergences

Table 7.6. Management recommendations and divergences between sources.

8. Reproductive Options and Family Counselling

8.1. Inheritance and recurrence risks

Primary trimethylaminuria is autosomal recessive [1, 2]. The parents of an affected individual are obligate heterozygotes, and molecular testing of the parents is recommended to confirm heterozygosity and assess the risk reliably, because de novo variants occur in recessive disorders at a low but appreciable rate; in Italy the first de novo causal missense variant of FMO3 was described [1, 32]. The siblings of a proband have a 25% risk of being affected, a 50% risk of being asymptomatic carriers, and a 25% chance of being neither affected nor carriers; the children of a proband are obligate carriers; the siblings of the parents have a 50% risk of being carriers [1, 2]. The risk that a child of an affected individual will also be affected depends on the partner's status, and the Irish patients expressed concern about transmission to their children and about their partner's carrier status [49]. In the mild forms associated with common variants, the recurrence risk is difficult to calculate, because the penetrance of the haplotypes is undetermined [11].

Table 8.1. Risks by family scenario in the primary form.

8.2. Cascade testing and carriers

Carrier detection requires prior identification of the familial variants, because biochemistry does not distinguish carriers from unaffected individuals [1]. Biochemical testing of siblings identifies affected individuals who would benefit from management, and genetic testing allows them to be tested early, while still asymptomatic [1, 3]. Carriers are described as asymptomatic, but a carrier of a severe variant with coexisting polymorphisms may become symptomatic under overload, and the NHGRI mentions mild symptoms or temporary episodes in some carriers [39, 41].

8.3. Prenatal and preimplantation diagnosis

Prenatal diagnosis and preimplantation genetic testing are possible once the familial variants have been identified; the gene card and Saudubray consider prenatal diagnosis possible but not indicated, and GeneReviews notes that perspectives differ and that the decision is a personal one [1–3]. Reproductive counselling should ideally take place before pregnancy and be offered to young adults who are affected, carriers, or at risk; DNA banking may be considered [1].

8.4. Content of counselling

Genetic counselling before and after testing is recommended; patients often ask about the recurrence risk, and identifying common variants in the mild forms may not be beneficial, as it can prompt excessive choline restriction [11]. In the United Kingdom, genetic counselling for primary trimethylaminuria is not offered routinely, and FMO3 testing is not listed in the National Genomic Test Directory [17]. The topics to address are the nonprogressive nature of the condition and its lack of physical morbidity, the variability of expression with diet, hormones, and age, the limited significance of the polymorphisms, the aggravating factors (menstruation, fever, exercise, stress), choline requirements during pregnancy and lactation, and psychosocial support [1, 3, 6]. In children, testing of asymptomatic siblings is justified by the benefit of early management and by school-related decisions [1, 3].

9. Special Considerations: Population Diversity and Epidemiology

9.1. What the available figures measure

There is no direct measurement of the prevalence of affected individuals; Orphanet gives prevalence as unknown, and GeneGuidance declines to derive prevalence from carrier frequency, because that would require a Hardy-Weinberg assumption that the source does not make [6]. The incidence of the severe form of up to 1/40,000 in the United Kingdom is an estimate derived from a heterozygote frequency of 0.5–1.0% in the British white population and is considered probably underestimated owing to underdiagnosis [2, 3]. Carrier frequencies for other populations come from Mitchell and colleagues (1997), a study of biochemical phenotyping of discontinuous TMA N-oxidation rather than genotyping: 1.7% in Jordan, 3.8% in Ecuador, and 11.0% in New Guinea; from these frequencies the gene card derives disease frequencies of up to 1/3000 in Ecuador and 1/400 in New Guinea [1, 2]. A 2025 review cites an estimated incidence of 1/25,000 to 1/40,000, also without direct measurement [76]. The mild/moderate form is more common than the severe form, with unknown incidence [3]. Referral cohorts, such as that of the Monell Center (about one-third of 353 patients with idiopathic odour) or Philadelphia (35%), measure the proportion of positives among those seeking an odour clinic and not the prevalence in the population [2, 44].

Table 9.1. Population and cohort data by region.

9.2. Diversity and variants by ancestry

Primary trimethylaminuria is pan-ethnic: it has been genetically confirmed in the USA (in people of European ancestry and in African American, Hispanic, and Asian individuals), Canada, the United Kingdom, Spain, Italy, Korea, Japan, Australia (a patient of Greek origin), Norway, and Thailand, and has been described in Hong Kong [3, 8, 32]. The most common causal variants differ between Europeans (p.(Pro153Leu), p.(Glu305Ter)) and Japanese individuals (p.(Arg205Cys), p.(Arg500Ter)), and in the Irish cohort the p.(Glu305Ter) and p.(Pro153Leu) variants of the severe patients coincide with those of Australian and North American families of Irish or British ancestry [2, 11]. The polymorphism frequencies vary with ancestry, with variants exclusive to Africans whose effect is yet to be determined, which limits the interpretation of genotypes in understudied populations [4]. The sources consulted do not describe founder populations with a common causal variant. The absence of epidemiological data from Portugal and Brazil in the sources consulted, beyond the INSA cohort and the case reports, does not demonstrate an absence of published studies.

9.3. Special populations

Women have perimenstrual worsening, with oral contraceptives, and in the perimenopause, and should be tested outside menstruation [1, 27]. Infants, especially preterm newborns, have low FMO3 activity, and children carrying reduced-activity alleles may become transiently symptomatic [1, 8]. Patients with homocystinuria treated with betaine, patients with cirrhosis or shunts, and those taking L-carnitine supplements constitute groups at risk of the secondary form [1, 13, 47]. In the pregnant woman, choline should not be restricted [1].

9.4. Reference networks and patient organisations

The relevant European reference network is MetabERN (European Reference Network for Hereditary Metabolic Disorders), with no subnetwork specific to trimethylaminuria identified on the website [16]. The patient resources cited are Metabolic Support UK (United Kingdom), MEBO Research (a North American research and advocacy organisation for patients with metabolic body odour, registered in Florida since 2010), GARD (Genetic and Rare Diseases Information Center), NORD (National Organization for Rare Disorders), MedlinePlus, and the NHGRI page [1, 41, 79, 80]. The Irish centre runs a support group linked to the clinic, rated positively by patients [49].

10. Illustrative Clinical Case

10.1. Severe primary trimethylaminuria with a rare genotype (South Africa)

In 2025, Dercksen and colleagues described an 11-year-old White girl, an only child of nonconsanguineous parents, referred by the dietitian for an intermittent "fishy" odour, more intense around menstruation and after shellfish; she was teased at school, and the school was concerned about her hygiene [37]. The patient noticed no difference in odour before and after the load, although the samples reached the laboratory with an extreme odour, which the authors interpreted as possible specific anosmia for TMA [37]. The direct load test consisted of 1.3 mL of 43–49% TMA solution diluted in 250 mL of orange juice, with baseline and hourly urine over 4 h, under medical supervision and outside the menstrual period, and quantification by proton NMR (TMA at 2.88 ppm; TMAO at 3.278 ppm) [37]. Figure 1 of the article shows the evolution of the TMA and TMAO peaks in the NMR spectra during the load (https://pmc.ncbi.nlm.nih.gov/articles/PMC11897904/figure/jmd270005-fig-0001/; page and legend consulted) [37].

Table 10.1. Results of the load test (mmol/mol creatinine) [37].

A baseline capacity below 43% classified the form as severe; the article writes the formula as TMA/(TMA + TMAO) × 100, but the values correspond to TMAO/(TMA + TMAO) × 100 [37]. Sanger sequencing of exons 2 to 9 (NM_001002294.3; NG_012690.1) showed homozygosity for c.23T>C p.(Ile8Thr) (rs565935391; VUS in ClinVar, with two heterozygotes in 125,748 exomes of the Genome Aggregation Database, gnomAD v2.1.1) and for the polymorphisms c.472G>A p.(Glu158Lys) and c.923A>G p.(Glu308Gly); the parents were not tested, so the homozygosity was not confirmed by segregation and a deletion on one allele was not excluded [34, 37]. Management included a choline-restricted diet (including fish) per the NHGRI, activated charcoal twice daily for 10 days, and copper chlorophyllin three times daily after meals for 3 weeks (no doses in the article, which refers to the NHGRI), an unspecified antibiotic twice a year, soaps of pH 5.5–5.6, and riboflavin [37]. At age 19 she was a functional young adult, with symptoms moderately controlled by elimination of shellfish [37]. This was the first case of primary trimethylaminuria described in sub-Saharan Africa [37].

This case illustrates three points: the direct TMA load reveals an almost absent oxidation capacity that baseline urine underestimates; the genotype rests on a VUS, and the diagnosis of the primary form depends on biochemistry, consistent with the GeneReviews criterion that a VUS neither establishes nor excludes the diagnosis; and the twice-yearly antibiotic is a centre practice, not described in the guidance sources, which recommend short, occasional courses [1, 3].

10.2. Transient childhood form (Portugal)

In 2025, Resende and colleagues described a Portuguese boy followed at the Centro Materno-Infantil do Norte (Porto), exclusively breastfed until 6 months of age with fish introduced at 7 months [38]. At 10 months, after eating swordfish, the mother noticed an intense "rotting-fish" odour, mainly on the head and hands, which recurred with other fish; with 2 months off fish there were no symptoms, and at 12 months the reintroduction of hake and then of forkbeard made the odour reappear [38]. Growth and neurodevelopment were normal, there was no odour at the visit (he had not eaten fish beforehand), there was no family history or consanguinity, and renal, thyroid, and liver function were normal [38]. Molecular study of FMO3 showed c.627+10C>G (intronic) and c.472G>A p.(Glu158Lys), both heterozygous, with phase undetermined, and urinary TMA and TMAO measurement was not available [38]. Management was to reintroduce fish in smaller portions and use a low-pH soap; since 19 months of age he has eaten the recommended fish portions (5–6 meals per week) without recurrence, with normal growth and development [38]. Figure 1 of the article is a diagram of the TMA metabolic pathway (https://pmc.ncbi.nlm.nih.gov/articles/PMC11929150/figure/FIG1/; consulted) [38].

The authors interpreted the picture as a transient form from the combination of the variants with FMO3 immaturity [38]. That interpretation is limited: c.627+10C>G is classified as benign or likely benign in ClinVar, and p.(Glu158Lys) alone is a benign polymorphism; without urine biochemistry, the diagnosis rests on the clinical picture and the response to the dietary change [34, 38]. The case shows a conservative management pathway, with gradual reintroduction of fish, consistent with the described maturation of FMO3, and the importance of not assigning pathogenic significance to benign variants [8, 11].

11. Prognosis and Long-Term Follow-Up

11.1. Prognosis

Primary trimethylaminuria does not reduce life expectancy or cause any known organ morbidity; it is a nonprogressive condition affecting quality of life and social functioning [6, 38]. Symptoms are generally present from birth and may worsen at puberty, and treatment and diet relieve some but not all [1]. All individuals biallelic for loss-of-function variants have the severe form, and the more severe the reduction in activity the poorer the response to treatment; the moderate and mild forms depend on factors beyond genotype [1, 2, 40]. The transient infantile forms resolve with maturation or withdrawal of the precursor, but a lifelong risk of recurrence may persist [3, 38]. The secondary forms may resolve with correction of the cause, although the post-hepatitis form persists after the hepatitis resolves [1, 13]. The dominant morbidity is psychosocial, with isolation, depression, and, rarely, suicidal ideation or attempts, and follow-up at a reference centre with a support group was associated with better understanding of the disease, self-esteem, and sense of support [1, 8, 49].

11.2. Initial evaluation

After diagnosis, GeneReviews recommends determining the TMAO/total TMA ratio on a normal diet to estimate severity, assessing social problems (harassment, bullying, discrimination, negative self-image, isolation, relationship problems), and referring to a genetics clinic [1]. Rutkowski and colleagues proposed a self-assessment instrument for the social and mental health effects and the efficacy of treatment (odour characteristics, social well-being, support from professionals, and mental well-being) [1, 49].

11.3. Surveillance

No source defines formal surveillance intervals for trimethylaminuria, and GeneReviews has no surveillance section [1]. Table 11.1 brings together the recommended assessments and the time points used in the published studies, which are centre practice rather than a validated protocol.

Table 11.1. Surveillance of trimethylaminuria.

12. Summary and Concluding Remarks

Trimethylaminuria results from an imbalance between intestinal TMA production, from dietary choline, carnitine, betaine, and TMAO, and the N-oxygenation capacity of hepatic FMO3. The severe primary form is autosomal recessive, caused by biallelic loss-of-function variants of FMO3 (MANE NM_001002294.3), with more than 40% of total TMA excreted as the free amine; the mild or moderate form is multifactorial and associated with the common haplotype p.[(Glu158Lys;Glu308Gly)], enzyme immaturity in childhood, menstruation, and precursor overload; the secondary forms accompany liver disease, portosystemic shunts, uraemia, bacterial overgrowth, and the use of betaine, carnitine, or choline. The incidence of the severe form is estimated at up to 1/40,000 in the United Kingdom, from a carrier frequency of 0.5–1.0%, without direct measurement of prevalence.

The odour is the only physical manifestation, and the social and psychological impact is the dominant morbidity; odour confirmed by others, onset in the first year of life, and elevated urinary TMA distinguish the disease from olfactory reference syndrome, but only biochemistry under load allows a conclusion. Diagnosis rests on urinary TMA and TMAO on two occasions, outside menstruation, in acidified urine, with genetic testing when free TMA exceeds 10%; the severity cutoffs and the odour threshold diverge between sources, and the value of 18–20 µmol/mmol creatinine is the most concordant. Sequencing detects most known variants, but a negative result does not exclude the disease, VUS and polymorphisms do not establish the diagnosis, and the phase of the variants must be determined.

Treatment is symptomatic and without randomised trials: precursor restriction with adequate choline and folate, with no choline restriction during pregnancy, lactation, or early childhood; soaps of pH 5.5–6.5; short courses of antibiotics, with doses extrapolated from labels for lack of a published condition-specific dose; riboflavin 30–40 mg 3 to 5 times daily per GeneReviews, with divergent doses in other sources; activated charcoal and copper chlorophyllin; avoidance of FMO3 substrate drugs; and psychological support. Therapeutic development is preclinical, with inhibitors of bacterial TMA production, postbiotics, and topical formulations. Genetic counselling follows the autosomal recessive model, with cascade testing of siblings and prenatal diagnosis possible but not indicated.

13. References

1. Phillips IR, Shephard EA. Primary Trimethylaminuria. 2007 Oct 8 [updated 2020 Nov 5]. In: Adam MP, Bick S, Mirzaa GM, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2026. Available from: https://www.ncbi.nlm.nih.gov/books/NBK1103/. PMID: 20301282.
2. Shephard EA, Treacy EP, Phillips IR. Clinical utility gene card for: Trimethylaminuria – update 2014. Eur J Hum Genet. 2015;23(9):1269. doi:10.1038/ejhg.2014.226. PMID: 25335494. PMC4538216.
3. Wevers RA, Mayatepek E, Walker V. Disorders of Peptide and Amine Metabolism. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 31, sections 31.1 (31.1.1 Trimethylaminuria) and 31.2.1. doi:10.1007/978-3-662-63123-2_31
4. Sidoti A, D'Angelo R, Castagnetti A, Viciani E, Scimone C, Alibrandi S, et al. Exploring Trimethylaminuria: Genetics and Molecular Mechanisms, Epidemiology, and Emerging Therapeutic Strategies. Biology (Basel). 2024;13(12). doi:10.3390/biology13120961. PMID: 39765628. PMC11726875.
5. Bouchemal N, Ouss L, Brassier A, Barbier V, Gobin S, Hubert L, et al. Diagnosis and phenotypic assessment of trimethylaminuria, and its treatment with riboflavin: (1)H NMR spectroscopy and genetic testing. Orphanet J Rare Dis. 2019;14(1):222. doi:10.1186/s13023-019-1174-6. PMID: 31533761. PMC6751875.
6. GeneGuidance (version of 2026-09-26), gene FMO3 and condition Trimethylaminuria (OMIM 602079, ORPHA 468726).
7. 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):61-68. doi:10.1002/humu.23348. PMID: 28967166. Available from: https://variantvalidator.org [accessed 2026-10-27 and 2026-11-01].
8. Mackay RJ, McEntyre CJ, Henderson C, Lever M, George PM. Trimethylaminuria: causes and diagnosis of a socially distressing condition. Clin Biochem Rev. 2011;32(1):33-43. PMID: 21451776. PMC3052392.
9. Orphanet. Severe primary trimethylaminuria (ORPHA:468726). Available from: https://www.orpha.net/en/disease/detail/468726
10. Orphanet. NON RARE IN EUROPE: Trimethylaminuria (ORPHA:35056). Available from: https://www.orpha.net/en/disease/detail/35056
11. Doyle S, O'Byrne JJ, Nesbitt M, Murphy DN, Abidin Z, Byrne N, et al. The genetic and biochemical basis of trimethylaminuria in an Irish cohort. JIMD Rep. 2019;47(1):35-40. doi:10.1002/jmd2.12028. PMID: 31240165. PMC6498825.
12. Chalmers RA, Bain MD, Michelakakis H, Zschocke J, Iles RA. Diagnosis and management of trimethylaminuria (FMO3 deficiency) in children. J Inherit Metab Dis. 2006;29(1):162-72. doi:10.1007/s10545-006-0158-6. PMID: 16601883.
13. Knapp M. Trimethylaminuria In A 58-Year-Old Male with Liver Cirrhosis and Hepatitis C. Eur J Case Rep Intern Med. 2025;12(7):005520. doi:10.12890/2025_005520. PMID: 40636240. PMC12236676.
14. Schmidt AC, Leroux JC. Treatments of trimethylaminuria: where we are and where we might be heading. Drug Discov Today. 2020;25(9):1710-1717. doi:10.1016/j.drudis.2020.06.026. PMID: 32615074.
15. U.S. National Library of Medicine. ClinicalTrials.gov API v2 — search "trimethylaminuria", "fish odor syndrome", "FMO3" [Internet]. Available from: https://clinicaltrials.gov/ [cited 2026 Oct 27].
16. European Reference Network for Hereditary Metabolic Disorders (MetabERN) [Internet]. Available from: https://metab.ern-net.eu/ [cited 2026 Oct 27].
17. Flaherty CC, Phillips IR, Janmohamed A, Shephard EA. Living with trimethylaminuria and body and breath malodour: personal perspectives. BMC Public Health. 2024;24(1):222. doi:10.1186/s12889-024-17685-w. PMID: 38238734. PMC10797923.
18. Humbert JA, Hammond KB, Hathaway WE. Trimethylaminuria: the fish-odour syndrome. Lancet. 1970;2(7676):770-1. doi:10.1016/s0140-6736(70)90241-2. PMID: 4195988.
19. Al-Waiz M, Ayesh R, Mitchell SC, Idle JR, Smith RL. Trimethylaminuria (fish-odour syndrome): an inborn error of oxidative metabolism. Lancet. 1987;1(8533):634-5. doi:10.1016/s0140-6736(87)90280-7. PMID: 2881174.
20. Ayesh R, Mitchell SC, Zhang A, Smith RL. The fish odour syndrome: biochemical, familial, and clinical aspects. BMJ. 1993;307(6905):655-7. doi:10.1136/bmj.307.6905.655. PMID: 8401051. PMC1678986.
21. Dolphin CT, Riley JH, Smith RL, Shephard EA, Phillips IR. Structural organization of the human flavin-containing monooxygenase 3 gene (FMO3), the favored candidate for fish-odor syndrome, determined directly from genomic DNA. Genomics. 1997;46(2):260-7. doi:10.1006/geno.1997.5031. PMID: 9417913.
22. Dolphin CT, Janmohamed A, Smith RL, Shephard EA, Phillips IR. Missense mutation in flavin-containing mono-oxygenase 3 gene, FMO3, underlies fish-odour syndrome. Nat Genet. 1997;17(4):491-4. doi:10.1038/ng1297-491. PMID: 9398858.
23. Treacy EP, Akerman BR, Chow LM, Youil R, Bibeau C, Lin J, et al. Mutations of the flavin-containing monooxygenase gene (FMO3) cause trimethylaminuria, a defect in detoxication. Hum Mol Genet. 1998;7(5):839-45. doi:10.1093/hmg/7.5.839. PMID: 9536088.
24. Akerman BR, Lemass H, Chow LM, Lambert DM, Greenberg C, Bibeau C, et al. Trimethylaminuria is caused by mutations of the FMO3 gene in a North American cohort. Mol Genet Metab. 1999;68(1):24-31. doi:10.1006/mgme.1999.2885. PMID: 10479479.
25. Zschocke J, Kohlmueller D, Quak E, Meissner T, Hoffmann GF, Mayatepek E. Mild trimethylaminuria caused by common variants in FMO3 gene. Lancet. 1999;354(9181):834-5. doi:10.1016/s0140-6736(99)80019-1. PMID: 10485731.
26. Hernandez D, Addou S, Lee D, Orengo C, Shephard EA, Phillips IR. Trimethylaminuria and a human FMO3 mutation database. Hum Mutat. 2003;22(3):209-13. doi:10.1002/humu.10252. PMID: 12938085.
27. Shimizu M, Cashman JR, Yamazaki H. Transient trimethylaminuria related to menstruation. BMC Med Genet. 2007;8:2. doi:10.1186/1471-2350-8-2. PMID: 17257434. PMC1790885.
28. Fennema D, Phillips IR, Shephard EA. Trimethylamine and Trimethylamine N-Oxide, a Flavin-Containing Monooxygenase 3 (FMO3)-Mediated Host-Microbiome Metabolic Axis Implicated in Health and Disease. Drug Metab Dispos. 2016;44(11):1839-1850. doi:10.1124/dmd.116.070615. PMID: 27190056. PMC5074467.
29. Yeung CK, Adman ET, Rettie AE. Functional characterization of genetic variants of human FMO3 associated with trimethylaminuria. Arch Biochem Biophys. 2007;464(2):251-9. doi:10.1016/j.abb.2007.04.014. PMID: 17531949. PMC2039921.
30. Dolphin CT, Janmohamed A, Smith RL, Shephard EA, Phillips IR. Compound heterozygosity for missense mutations in the flavin-containing monooxygenase 3 (FM03) gene in patients with fish-odour syndrome. Pharmacogenetics. 2000;10(9):799-807. doi:10.1097/00008571-200012000-00005. PMID: 11191884.
31. Phillips IR, Shephard EA. Flavin-containing monooxygenase 3 (FMO3): genetic variants and their consequences for drug metabolism and disease. Xenobiotica. 2020;50(1):19-33. doi:10.1080/00498254.2019.1643515. PMID: 31317802.
32. Teresa E, Lonardo F, Fiumara A, Lombardi C, Russo P, Zuppi C, et al. A spectrum of molecular variation in a cohort of Italian families with trimethylaminuria: identification of three novel mutations of the FM03 gene. Mol Genet Metab. 2006;88(2):192-5. doi:10.1016/j.ymgme.2006.02.014. PMID: 16600650.
33. Shimizu M, Kobayashi Y, Hayashi S, Aoki Y, Yamazaki H. Variants in the flavin-containing monooxygenase 3 (FMO3) gene responsible for trimethylaminuria in a Japanese population. Mol Genet Metab. 2012;107(3):330-4. doi:10.1016/j.ymgme.2012.06.014. PMID: 22819296.
34. National Center for Biotechnology Information. ClinVar [Internet]. Bethesda (MD): NCBI. FMO3 variants, VCV 16308, 16304, 38394, 38395, 16318, 876845, 2734034, 16313, 16309, 16305, 3892342, 5003929, 260076, 3366396, 4526110, 4750109, 3779673, 2720934, 217371, 242472, 3366476, 2577331, 631576. Available from: https://www.ncbi.nlm.nih.gov/clinvar/ [cited 2026 Oct 27; updated 2026 Nov 1].
35. Motika MS, Zhang J, Zheng X, Riedler K, Cashman JR. Novel variants of the human flavin-containing monooxygenase 3 (FMO3) gene associated with trimethylaminuria. Mol Genet Metab. 2009;97(2):128-35. doi:10.1016/j.ymgme.2009.02.006. PMID: 19321370. PMC2739593.
36. D'Angelo R, Esposito T, Calabrò M, Rinaldi C, Robledo R, Varriale B, et al. FMO3 allelic variants in Sicilian and Sardinian populations: trimethylaminuria and absence of fish-like body odor. Gene. 2013;515(2):410-5. doi:10.1016/j.gene.2012.12.047. PMID: 23266626.
37. Dercksen M, Perumal M, Davoren E, Reed DR, Murry-Maritz C, van der Sluis R, et al. Diagnosis of Primary Trimethylaminuria in an Affected Patient With a Rare Genotype in Sub-Saharan Africa. JIMD Rep. 2025;66(2):e70005. doi:10.1002/jmd2.70005. PMID: 40078825. PMC11897904.
38. Resende MM, Leite-Almeida L, Campos P, Sobreira I, Garcia P. Unpleasant Smell: A Case Report of Trimethylaminuria (Fish Odour Syndrome) in a Child. Cureus. 2025;17(2):e79318. doi:10.7759/cureus.79318. PMID: 40125190. PMC11929150.
39. Miller NB, Beigelman A, Utterson E, Shinawi M. Transient massive trimethylaminuria associated with food protein-induced enterocolitis syndrome. JIMD Rep. 2014;12:11-5. doi:10.1007/8904_2013_238. PMID: 23821320. PMC3897788.
40. Shimizu M, Allerston CK, Shephard EA, Yamazaki H, Phillips IR. Relationships between flavin-containing mono-oxygenase 3 (FMO3) genotype and trimethylaminuria phenotype in a Japanese population. Br J Clin Pharmacol. 2014;77(5):839-51. doi:10.1111/bcp.12240. PMID: 24028545. PMC4004404.
41. National Human Genome Research Institute. About Trimethylaminuria [Internet]. Bethesda (MD): NHGRI; last updated 2018 Dec 18 [cited 2026 Oct 27]. Available from: https://www.genome.gov/Genetic-Disorders/Trimethylaminuria
42. Yan H, Cao C, Tan X, Huang X, Liu Y, Chu Y, et al. Hepatic FMO3 Downregulation Contributes to the Metabolic Benefits of Vertical Sleeve Gastrectomy. Obesity (Silver Spring). 2026;34(10):1880-1891. doi:10.1002/oby.70281. PMID: 42555205.
43. Mayatepek E, Kohlmüller D. Transient trimethylaminuria in childhood. Acta Paediatr. 1998;87(11):1205-7. doi:10.1080/080352598750031257. PMID: 9846928.
44. Wise PM, Eades J, Tjoa S, Fennessey PV, Preti G. Individuals reporting idiopathic malodor production: demographics and incidence of trimethylaminuria. Am J Med. 2011;124(11):1058-63. doi:10.1016/j.amjmed.2011.05.030. PMID: 21851918.
45. Yamazaki H, Fujieda M, Cashman JR, Kamataki T. Mild trimethylaminuria observed in a Japanese cohort with liver damage. Am J Med. 2005;118(7):803-5. doi:10.1016/j.amjmed.2004.08.015. PMID: 15989923.
46. Kožich V, Morris AAM, Blom HJ. Disorders of Sulfur Amino Acid Metabolism. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 20, sections 20.2 (Methanethiol Oxidase Deficiency) and 20.6.5. doi:10.1007/978-3-662-63123-2_20
47. Manning NJ, Allen EK, Kirk RJ, Sharrard MJ, Smith EJ. Riboflavin-responsive trimethylaminuria in a patient with homocystinuria on betaine therapy. JIMD Rep. 2012;5:71-5. doi:10.1007/8904_2011_99. PMID: 23430919. PMC3509925.
48. Fraser-Andrews EA, Manning NJ, Ashton GH, Eldridge P, McGrath J, Menagé Hdu P. Fish odour syndrome with features of both primary and secondary trimethylaminuria. Clin Exp Dermatol. 2003;28(2):203-5. doi:10.1046/j.1365-2230.2003.01230.x. PMID: 12653714.
49. Roddy D, McCarthy P, Nerney D, Mulligan-Rabbitt J, Smith E, Treacy EP. Impact of trimethylaminuria on daily psychosocial functioning. JIMD Rep. 2021;57(1):67-75. doi:10.1002/jmd2.12170. PMID: 33473342. PMC7802621.
50. Gomez Lopez LA, Yoon SS, Espana VM, Walteros CV. Beyond Malodor: A Multinational Patient-Reported Study of the Diagnostic, Psychosocial, and Cognitive Burden of Trimethylaminuria [preprint]. medRxiv. 2026 Sep 11. doi:10.64898/2026.09.08.26362298.
51. Treacy E, Johnson D, Pitt JJ, Danks DM. Trimethylaminuria, fish odour syndrome: a new method of detection and response to treatment with metronidazole. J Inherit Metab Dis. 1995;18(3):306-12. doi:10.1007/BF00710420. PMID: 7474897.
52. Yamazaki H, Fujieda M, Togashi M, Saito T, Preti G, Cashman JR, et al. Effects of the dietary supplements, activated charcoal and copper chlorophyllin, on urinary excretion of trimethylamine in Japanese trimethylaminuria patients. Life Sci. 2004;74(22):2739-47. doi:10.1016/j.lfs.2003.10.022. PMID: 15043988.
53. Ferreira F, Esteves S, Almeida LS, Gaspar A, da Costa CD, Janeiro P, et al. Trimethylaminuria (fish odor syndrome): genotype characterization among Portuguese patients. Gene. 2013;527(1):366-70. doi:10.1016/j.gene.2013.05.025. PMID: 23791655.
54. Pol A, Renkema GH, Tangerman A, Winkel EG, Engelke UF, de Brouwer APM, et al. Mutations in SELENBP1, encoding a novel human methanethiol oxidase, cause extraoral halitosis. Nat Genet. 2018;50(1):120-129. doi:10.1038/s41588-017-0006-7. PMID: 29255262. PMC5742538.
55. Saudubray JM, García-Cazorla Á. Clinical Approach to Inborn Errors of Metabolism in Paediatrics. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 1, section 1.6.9, Table 1.40. doi:10.1007/978-3-662-63123-2_1
56. Nyhan WL, Hoffmann GF, Al-Aqeel AI, Barshop BA. Atlas of Inherited Metabolic Diseases. 4th ed. Boca Raton: CRC Press; 2020. Appendix: Differential diagnosis of clinical phenotypes, p. 824 (Odd or unusual odor).
57. Ramaswami U, Humphries SE. Inborn Errors of Lipoprotein Metabolism Presenting in Childhood. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 36. doi:10.1007/978-3-662-63123-2_36
58. Podadera P, Sipahi AM, Arêas JA, Lanfer-Marquez UM. Diagnosis of suspected trimethylaminuria by NMR spectroscopy. Clin Chim Acta. 2005;351(1-2):149-54. doi:10.1016/j.cccn.2004.09.006. PMID: 15563884.
59. Busby MG, Fischer L, da Costa KA, Thompson D, Mar MH, Zeisel SH. Choline- and betaine-defined diets for use in clinical research and for the management of trimethylaminuria. J Am Diet Assoc. 2004;104(12):1836-45. doi:10.1016/j.jada.2004.09.027. PMID: 15565078.
60. INFARMED. Summary of Product Characteristics: Flagyl 250 mg film-coated tablets (metronidazole) [Resumo das Características do Medicamento: Flagyl 250 mg comprimidos revestidos (metronidazol)]. Faes Farma Portugal, S.A. Approved 2023-04-24. Infomed med_id 3423. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=3423&tipo_doc=rcm
61. Pfizer Laboratories. FLAGYL 375 (metronidazole) capsules. Prescribing information (NDA 020334). DailyMed set id a2883ca1-5a9a-4259-9d80-46ab67274384; effective 2026-05-22. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=a2883ca1-5a9a-4259-9d80-46ab67274384
62. Agência Nacional de Vigilância Sanitária (ANVISA). Open data — registered medicinal products (DADOS_ABERTOS_MEDICAMENTOS.csv) [Dados abertos — medicamentos registados] [Internet]. Available from: https://dados.anvisa.gov.br/dados/DADOS_ABERTOS_MEDICAMENTOS.csv [cited 2026 Oct 27].
63. Chartwell RX. Neomycin sulfate tablets USP, 500 mg. Prescribing information (ANDA 204435). DailyMed set id 0aa6947b-4739-4b00-be93-7303a5fc687d; effective 2026-02-27. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=0aa6947b-4739-4b00-be93-7303a5fc687d
64. INFARMED. Infomed — search "Neomicina" (authorized/marketed medicinal products) [Infomed — pesquisa «Neomicina» (medicamentos autorizados/comercializados)] [Internet]. Available from: https://extranet.infarmed.pt/INFOMED-fo/ [cited 2026 Oct 27].
65. INFARMED. Summary of Product Characteristics: Refero 550 mg film-coated tablets (rifaximin) [Resumo das Características do Medicamento: Refero 550 mg comprimidos revestidos por película (rifaximina)]. Alfasigma Portugal, Lda. First authorization April 22, 2013; last renewal November 29, 2017. Infomed med_id 54942. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=54942&tipo_doc=rcm [cited 2026 Oct 27].
66. Salix Pharmaceuticals (repackager Cardinal Health 107). XIFAXAN (rifaximin) tablets. Prescribing information (NDA 021361). DailyMed set id 7afafa63-3236-4212-9589-045db6b01c0b; effective 2025-08-27. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=7afafa63-3236-4212-9589-045db6b01c0b
67. INFARMED. Summary of Product Characteristics: Clamoxyl 500 mg capsules (amoxicillin) [Resumo das Características do Medicamento: Clamoxyl 500 mg cápsulas (amoxicilina)]. Beecham Portuguesa - Produtos Farmacêuticos e Químicos, Lda. Infomed med_id 1880. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=1880&tipo_doc=rcm [cited 2026 Oct 27].
68. DirectRx (repackager). Amoxil (amoxicillin) for oral suspension. Prescribing information (ANDA 065325). DailyMed set id ec7dd735-dc92-3d29-e053-2a95a90af549; effective 2025-01-22. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=ec7dd735-dc92-3d29-e053-2a95a90af549
69. INFARMED. Summary of Product Characteristics: Duphalac 667 g/l syrup (lactulose) [Resumo das Características do Medicamento: Duphalac 667 g/l xarope (lactulose)]. Cooper Consumer Health B.V. Approved 2023-08-07. Infomed med_id 2877. Available from: https://extranet.infarmed.pt/INFOMED-fo/download-ficheiro.xhtml?med_id=2877&tipo_doc=rcm [cited 2026 Oct 27].
70. Actavis Pharma. Enulose (lactulose) oral solution. Prescribing information (ANDA 090502). DailyMed set id 00c42c5c-d19e-4130-aa08-a8bbd47d3e5b; effective 2022-06-01. Available from: https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=00c42c5c-d19e-4130-aa08-a8bbd47d3e5b
71. Pike MG, King GS, Pettit BR, Leonard JV, Atherton DJ. Lactulose in trimethylaminuria, the fish-odour syndrome. Helv Paediatr Acta. 1989;43(4):345-8. PMID: 2708075.
72. Morris AAM, Jones S. Medications Used in the Treatment of Inborn Errors of Metabolism. In: Saudubray JM, Baumgartner MR, García-Cazorla Á, Walter J, editors. Inborn Metabolic Diseases: Diagnosis and Treatment. 7th ed. Berlin: Springer; 2022. Chapter 45 (Riboflavin). doi:10.1007/978-3-662-63123-2_45
73. INFARMED. Infomed — search "riboflavina" (authorized/marketed medicinal products) [Infomed — pesquisa «riboflavina» (medicamentos autorizados/comercializados)] [Internet]. Available from: https://extranet.infarmed.pt/INFOMED-fo/ [cited 2026 Oct 27].
74. Schmidt AC, Hebels ER, Weitzel C, Kletzmayr A, Bao Y, Steuer C, et al. Engineered Polymersomes for the Treatment of Fish Odor Syndrome: A First Randomized Double Blind Olfactory Study. Adv Sci (Weinh). 2020;7(8):1903697. doi:10.1002/advs.201903697. PMID: 32328434. PMC7175261.
75. Cini E, Vinciarelli G, Siciliano S, Del Castello G, Dreassi E, Poggialini F, et al. Fluoromethylcarnitine, a novel inhibitor of trimethylamine levels in trimethylaminuria and trimethylamine N-oxide related disorders. Eur J Med Chem. 2025;300:118073. doi:10.1016/j.ejmech.2025.118073. PMID: 40907381.
76. Giannini G, Soldi S, Elli M, Sagheddu V, Castagnetti A, Viciani E, et al. A mixture of postbiotics/tyndallized probiotics reduces trimethylamine (TMA) in trimethylaminuria models: Evidence from in vitro and in vivo studies. Front Pharmacol. 2025;16:1591825. doi:10.3389/fphar.2025.1591825. PMID: 41132530. PMC12540306.
77. Gao W, Bai Y, Chang L, Han X, Yu J, Zhu P. Regulating the Flexibility of Two-Dimensional Covalent Organic Frameworks for High-Performance Gas Sensing in Noninvasive Monitoring of Trimethylaminuria. ACS Sens. 2026;11(3):2090-2100. doi:10.1021/acssensors.5c03761. PMID: 41789564.
78. Hussein K, Muyomba I, Simon T, Bwesigye P, William N, Ssebulime I, et al. Trimethylaminuria Presenting as Persistent Malodor in an Adolescent Female: A Case Report [preprint]. Research Square. 2026 May 6. doi:10.21203/rs.3.rs-9614832/v1.
79. Metabolic Support UK. Trimethylaminuria [Internet]. Available from: https://metabolicsupport.org/condition/trimethylaminuria/ [cited 2026 Oct 27].
80. MEBO Research, Inc. Metabolic body odor research and patient advocacy [Internet]. Available from: https://www.meboresearch.org/ [cited 2026 Oct 27].