Dihydropyrimidine Dehydrogenase Deficiency

Mendelian MONDO:0010130 Pathograph 18 Show in embeddings browser

Dihydropyrimidine dehydrogenase (DPD) deficiency is an autosomal recessive disorder of pyrimidine catabolism caused by variants in DPYD. DPD catalyses the first and rate-limiting step of pyrimidine degradation, so losing it blocks the whole pathway at once, producing both an accumulation of uracil and thymine and a depletion of the downstream product beta-alanine. The unusual feature of this entry is that one enzyme deficiency produces two almost unrelated diseases, and which one a person gets depends on whether anybody ever gives them a fluoropyrimidine. Left alone, biallelic deficiency is a rare inborn error whose neurological consequences range from an epileptic encephalopathy to nothing at all. Exposed to 5-fluorouracil or capecitabine, the same enzyme deficit becomes a pharmacogenomic emergency, because DPD normally clears the great majority of an administered dose and a patient who cannot clear it receives, in effect, a massive overdose. Carriers who are entirely well their whole lives can be killed by a standard first cycle of chemotherapy. The two arms are modelled separately here. The constitutive arm runs from enzyme loss to metabolite imbalance and is deliberately not asserted as a settled cause of the neurological phenotype, because the literature does not support that: the originating study reports wide phenotypic variability and states plainly that no genotype-phenotype correlation has been established, and the mechanism linking metabolites to brain injury is offered by its own authors as something that "might underlie" the clinical picture. The toxicity arm is gated on drug exposure, modelled as an environmental trigger acting on the residual-activity node, and is where the evidence is strong, quantitative, and actionable.

Ask OpenScientist

Ask a research question about Dihydropyrimidine Dehydrogenase Deficiency. OpenScientist will conduct autonomous deep research using the Disorder Mechanisms Knowledge Base and PubMed literature (typically 10-30 minutes).

Submitting...

Do not include personal health information in your question. Questions and results are cached in your browser's local storage.

1
Inheritance
8
Pathophys.
8
Phenotypes
1
Hypotheses
2
Gaps
18
Pathograph
1
Genes
5
Medical Actions
3
Differentials
1
Trials
1
Models
1
Deep Research
👪

Inheritance

1
Autosomal recessive HP:0000007
The constitutive metabolic disease requires biallelic variants. Fluoropyrimidine toxicity risk does not: it follows residual enzyme activity, so heterozygous carriers are at substantially increased risk despite being unaffected in every other respect.
Autosomal recessive inheritance
◈

Mechanistic Hypotheses

1
Altered pyrimidine and beta-alanine homeostasis causes the neurological phenotype
altered_pyrimidine_homeostasis_neurotoxicity EMERGING
Evidence balance 3 support
Holds that accumulated uracil and thymine, together with depleted beta-alanine and beta-aminoisobutyrate, disturb the developing central nervous system, and that this accounts for the seizures, developmental delay, and white-matter abnormalities seen in some patients. The pathway logic is coherent, since beta-alanine is a structural analogue of the two principal inhibitory neurotransmitters, and the metabolite disturbance itself is beyond doubt. What is unproven is the link from that disturbance to brain injury.
Deliberately EMERGING rather than CANONICAL. The decisive problem is not weak evidence for the mechanism but the existence of biochemically deficient individuals with no neurological phenotype whatsoever, which means the metabolite disturbance cannot be sufficient. Any complete account has to explain the unaffected as well as the affected, and none currently does.
Pathograph links
Pyrimidine Substrate Accumulation and Beta-Alanine Depletion -> Impaired Inhibitory Neurotransmission and Neurodevelopmental Injury INDIRECT UNKNOWN INTERMEDIATES Proposed route from metabolite imbalance to neurological disease. The intermediates are genuinely unknown, and the edge belongs to a hypothesis group rather than being asserted, because the same biochemical deficiency is compatible with no neurological phenotype at all.
Impaired Inhibitory Neurotransmission and Neurodevelopmental Injury -> Seizure INDIRECT UNKNOWN INTERMEDIATES Proposed route to seizures through altered inhibitory neurotransmission. Not asserted; the intermediates are unknown and penetrance is incomplete.
Impaired Inhibitory Neurotransmission and Neurodevelopmental Injury -> Leukoencephalopathy INDIRECT UNKNOWN INTERMEDIATES Proposed route to the white-matter and brainstem changes, whose pathogenesis the describing authors state is unknown.
Show evidence (3 references)
PMID:10071185 SUPPORT Human Clinical
"An altered beta-alanine, uracil and thymine homeostasis might underlie the various clinical abnormalities encountered in patients with DPD deficiency."
The originating statement of the hypothesis, graded PARTIAL because the authors write "might underlie" rather than "underlies" and offer it as a proposal.
PMID:34653361 SUPPORT Model Organism
"Dpyd encodes the rate-limiting enzyme in the metabolic pathway that catabolizes uracil and thymidine to β-alanine, an inhibitory neurotransmitter."
The best available functional support that beta-alanine loss has a real central nervous system consequence. PARTIAL and MODEL_ORGANISM because the demonstrated phenotype is murine sleep, not human seizures.
PMID:15303009 SUPPORT Human Clinical
"The pathogenesis of the white-matter abnormalities is unknown, although environmental factors and altered energy metabolism may be involved."
The authors who described the imaging phenotype state that its mechanism is unknown and raise alternatives, which is why this remains a hypothesis rather than the entry's asserted chain.
?

Discussions and Knowledge Gaps

2
Why do some individuals with complete DPD deficiency have severe epileptic encephalopathy while others with the same biochemical deficiency are entirely asymptomatic?
KNOWLEDGE GAP OPEN dpd_neuro_penetrance
Attached to
pathophysiology#Pyrimidine Substrate Accumulation and Beta-Alanine Depletion
This is the central unresolved problem of the constitutive arm and it is not a matter of incomplete ascertainment. The originating genotype-phenotype study states plainly that no clear correlation has been established, and a wide range of severity including complete absence of signs is recognised. The metabolite disturbance is therefore necessary but demonstrably not sufficient for the neurological phenotype, which means something else, whether modifier genes, timing of exposure during development, or an environmental contribution, determines whether the brain is affected. Until that is identified, the metabolite-to-brain chain cannot be asserted, which is why it is curated as an EMERGING hypothesis and its edges are placed in a hypothesis group rather than stated outright.
Proposed experiments
Modifier discovery in phenotype-discordant DPD-deficient relatives
dpd_discordant_sibling_sequencing
Recruit families containing biochemically deficient individuals discordant for the neurological phenotype, and perform whole-genome sequencing and cerebrospinal-fluid metabolite profiling on affected and unaffected members to identify modifiers segregating with the phenotype rather than with the DPYD genotype. Discordant relatives are the only design that controls for the primary defect while varying the outcome.
CSF pyrimidine and beta-alanine levels against neurological outcome
dpd_csf_metabolite_correlation
Measure cerebrospinal-fluid uracil, thymine, beta-alanine, and beta-aminoisobutyrate in a cohort of biochemically deficient individuals stratified by neurological status. If the hypothesis holds, central metabolite levels rather than urinary ones should separate affected from unaffected; if they do not, the hypothesis is in serious trouble.
Does the murine demonstration that beta-alanine acts as a sleep-promoting inhibitory neurotransmitter translate to a role in human DPD-deficiency seizures?
HUMAN MODEL MISMATCH OPEN dpd_beta_alanine_model_mismatch
Attached to
pathophysiology#Pyrimidine Substrate Accumulation and Beta-Alanine Depletion
Evidence exists in the model and is good: mouse work links Dpyd to beta-alanine and establishes beta-alanine as an inhibitory neurotransmitter that promotes sleep. What is open is whether that function is the one that matters in human disease. The murine phenotype is altered sleep; the human phenotype under investigation is seizures, developmental delay, and white-matter injury. A shared molecule and a shared direction of change do not establish a shared consequence, and treating the mouse result as human evidence would overstate the case for the whole neurological chain.
Proposed experiments
Seizure susceptibility phenotyping in Dpyd-deficient mice
dpd_mouse_seizure_phenotyping
Characterise seizure threshold, EEG, and white-matter integrity in Dpyd-deficient mice, rather than sleep alone, and test whether beta-alanine supplementation modifies those endpoints. This would establish whether the murine model speaks to the human phenotype at all, or only to sleep.
⚙

Pathophysiology

8
Biallelic DPYD Loss of Function
Deleterious variants in DPYD, which encodes dihydropyrimidine dehydrogenase. The enzyme is a large homodimeric redox protein carrying two FAD, two FMN, and eight iron-sulfur clusters arranged as electron-transfer chains, which is why loss of function is reachable by many mechanistically distinct routes: disruption of the substrate site, of cofactor binding, of the electron wire, or of dimer assembly and folding. Complete deficiency requires two alleles; a single deleterious allele reduces activity enough to matter only when a fluoropyrimidine is administered.
dihydropyrimidine dehydrogenase activity GO:0017113 Gene Ontology (GO) Relation: this pathophysiological event involves this molecular function This pathophysiological event involves dihydropyrimidine dehydrogenase activity, annotated with dihydropyrimidine dehydrogenase (NADP+) activity (GO:0017113), qualified as loss of function. GO:0017113 is a molecular function from the Gene Ontology. ⇓ LOSS OF FUNCTION
Downstream
Reduced DPD Catalytic Activity Loss-of-function alleles reduce or abolish enzyme activity in proportion to how many alleles are affected and how severely.
Show evidence (1 reference)
PMID:11179210 SUPPORT In Vitro
"arranged in two electron transfer chains that pass the dimer interface twice"
Describes the electron-transfer architecture that makes this enzyme vulnerable to loss of function at many independent positions. The full sentence also enumerates the two FAD, two FMN, and eight iron-sulfur clusters; the quote stops short of them because the bracketed cluster notation is removed by snippet normalisation and cannot be matched.
Reduced DPD Catalytic Activity
Diminished or absent dihydropyrimidine dehydrogenase activity. This node is the hinge of the entry: everything constitutive runs downstream of it through the blocked catabolic pathway, and everything pharmacogenomic runs downstream of it only once a fluoropyrimidine arrives. Residual activity is continuous rather than binary, which is why the same node supports both a severe infantile presentation and a lifelong asymptomatic state.
pyrimidine nucleobase catabolic process GO:0006208 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves decreased pyrimidine nucleobase catabolic process (GO:0006208). GO:0006208 is a biological process from the Gene Ontology. ↓ DECREASED
Downstream
Block of Pyrimidine Nucleobase Catabolism The constitutive consequence, present from birth regardless of any drug exposure.
Impaired Fluoropyrimidine Clearance The pharmacogenomic consequence, which is latent until a fluoropyrimidine is administered and is therefore gated by exposure rather than by genotype alone.
Show evidence (1 reference)
PMID:11179210 SUPPORT In Vitro
"the enzyme is also responsible for the rapid breakdown of the chemotherapeutic drug 5-fluorouracil"
Establishes that the same enzyme governs fluoropyrimidine clearance, which is what makes one deficiency produce two diseases.
Block of Pyrimidine Nucleobase Catabolism
Because DPD catalyses the committed first step, its loss blocks the entire three-step catabolic pathway. The consequence is two-sided and both sides may matter: substrates accumulate, and the eventual product beta-alanine is not made. Uraciluria is the resulting biochemical signature and is present in essentially all homozygotes, whether or not they have any symptoms.
pyrimidine nucleobase catabolic process GO:0006208 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves decreased pyrimidine nucleobase catabolic process (GO:0006208). GO:0006208 is a biological process from the Gene Ontology. ↓ DECREASED
Downstream
Pyrimidine Substrate Accumulation and Beta-Alanine Depletion The metabolite imbalance that follows directly from blocking the pathway.
Show evidence (1 reference)
PMID:10071185 SUPPORT Human Clinical
"An altered beta-alanine, uracil and thymine homeostasis might underlie the various clinical abnormalities encountered in patients with DPD deficiency."
States both sides of the block, accumulation and depletion, in the authors' own deliberately hedged language. The hedge is preserved rather than flattened; see the mechanistic hypothesis of the same name.
Pyrimidine Substrate Accumulation and Beta-Alanine Depletion
Uracil, thymine, and 5-hydroxymethyluracil accumulate in urine, plasma, and cerebrospinal fluid, while beta-alanine and beta-aminoisobutyrate are depleted. Beta-alanine is a structural analogue of GABA and glycine and has inhibitory neurotransmitter activity, which is the basis of the proposed but unproven link to the neurological phenotype. This node is where the entry stops asserting and starts hypothesising: the metabolite disturbance is measured and certain, its consequences for the brain are not.
Altered pyrimidine and beta-alanine homeostasis causes the neurological phenotype
beta-alanine biosynthetic process GO:0019483 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves decreased beta-alanine biosynthetic process (GO:0019483). GO:0019483 is a biological process from the Gene Ontology. ↓ DECREASED
Animal models
Dpyd knockout mouse Mouse partially recapitulates Reproduces the metabolic consequence of losing the enzyme and demonstrates a central nervous system phenotype attributable to it, establishing that beta-alanine loss has real consequences for the brain. The phenotype is reduced sleep, not the neurological disease seen in patients.
Downstream
Uraciluria The diagnostic biochemical signature of the blocked pathway.
Impaired Inhibitory Neurotransmission and Neurodevelopmental Injury Proposed route from metabolite imbalance to neurological disease. The intermediates are genuinely unknown, and the edge belongs to a hypothesis group rather than being asserted, because the same biochemical deficiency is compatible with no neurological phenotype at all. Altered pyrimidine and beta-alanine homeostasis causes the neurological phenotype
Show evidence (1 reference)
PMID:34653361 SUPPORT Model Organism
"Both β-alanine and β-AIB are structural analogs of γ-aminobutyric acid (GABA) and glycine, two important inhibitory neurotransmitters in the central nervous system."
Establishes the structural basis for beta-alanine acting on inhibitory neurotransmission. Tagged MODEL_ORGANISM because the publication is a mouse sleep-genetics study; the structural-analogy statement is general, but the functional demonstration in that paper is murine.
Impaired Inhibitory Neurotransmission and Neurodevelopmental Injury
The hypothesised neurological consequence of the metabolite imbalance: beta-alanine and beta-aminoisobutyrate are structural analogues of GABA and glycine, so their depletion is proposed to shift inhibitory neurotransmission in the developing central nervous system, with accumulated pyrimidines contributing independently. This node exists as a node rather than as a bare edge to the phenotypes because it is the thing the animal evidence speaks to, positively and negatively, and a hypothesis with no node has nowhere to attach a negative result. Its edges belong to the neurotoxicity hypothesis group and are not asserted: the same biochemical state is compatible with no neurological phenotype at all.
Altered pyrimidine and beta-alanine homeostasis causes the neurological phenotype
beta-alanine biosynthetic process GO:0019483 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves decreased beta-alanine biosynthetic process (GO:0019483). GO:0019483 is a biological process from the Gene Ontology. ↓ DECREASED
Animal models
Dpyd knockout mouse Mouse fails to recapitulate The knockout does not reproduce the neurological phenotype that defines the human constitutive disease. Behavioural testing found no difference in cognitive, social, or affective domains, and no seizure phenotype is reported. This is a substantive negative result and is the reason the human neurological chain remains an EMERGING hypothesis rather than being supported by animal data.
Downstream
Seizure Proposed route to seizures through altered inhibitory neurotransmission. Not asserted; the intermediates are unknown and penetrance is incomplete. Altered pyrimidine and beta-alanine homeostasis causes the neurological phenotype
Leukoencephalopathy Proposed route to the white-matter and brainstem changes, whose pathogenesis the describing authors state is unknown. Altered pyrimidine and beta-alanine homeostasis causes the neurological phenotype
Show evidence (2 references)
PMID:34653361 SUPPORT Model Organism
"Both β-alanine and β-AIB are structural analogs of γ-aminobutyric acid (GABA) and glycine, two important inhibitory neurotransmitters in the central nervous system."
The structural basis for the proposed effect on inhibitory neurotransmission. PARTIAL because analogy to inhibitory neurotransmitters establishes a plausible route rather than a demonstrated one in human disease.
PMID:15303009 SUPPORT Human Clinical
"The pathogenesis of the white-matter abnormalities is unknown, although environmental factors and altered energy metabolism may be involved."
Included as a limit on this node rather than as support for it. The authors who described the imaging phenotype state that its mechanism is unknown and name alternatives, which is why the downstream edges sit in a hypothesis group.
Impaired Fluoropyrimidine Clearance
Exposure-gated node, and the reason this disease matters far beyond the handful of children with the inborn error. Dihydropyrimidine dehydrogenase performs the great majority of fluoropyrimidine catabolism, so a patient with reduced activity who receives a standard dose of 5-fluorouracil or capecitabine cannot clear it. Systemic exposure and half-life rise, and the patient effectively receives an overdose while being given a normal prescription. Nothing about this node is active until the drug is administered.
pyrimidine nucleobase catabolic process GO:0006208 Gene Ontology (GO) Relation: this pathophysiological event involves this biological process This pathophysiological event involves decreased pyrimidine nucleobase catabolic process (GO:0006208). GO:0006208 is a biological process from the Gene Ontology. ↓ DECREASED
Downstream
Fluoropyrimidine Cytotoxicity in Proliferating Tissues Excess systemic drug exposure delivers a cytotoxic insult to the tissues that turn over fastest.
Show evidence (1 reference)
PMID:26573078 SUPPORT Human Clinical
"A polymorphism in the fluoropyrimidine metabolizing enzyme dihydropyrimidine dehydrogenase (DPD; ie, DPYD*2A) is strongly associated with fluoropyrimidine-induced severe and life-threatening toxicity."
Links reduced DPD activity to severe and life-threatening drug toxicity, which is the claim this node carries.
Fluoropyrimidine Cytotoxicity in Proliferating Tissues
Supratherapeutic 5-fluorouracil exposure damages rapidly proliferating tissues by two routes: the metabolite FdUMP inhibits thymidylate synthase, producing thymineless stress and DNA damage, and FUTP is misincorporated into RNA. Bone marrow and gastrointestinal mucosa bear the brunt, with skin, myocardium, and central nervous system also affected. This node is the disease-specific substitution for the cytotoxic driver in the myelosuppression module: the insult is not an administered cytotoxic dose but an ordinary dose that the patient cannot clear.
Downstream
Multilineage Cytopenias The marrow limb, which is the conforming arm of the myelosuppression module.
Diarrhea The gastrointestinal mucosal limb.
Show evidence (1 reference)
PMID:27622829 SUPPORT Human Clinical
"uridine triacetate was a safe and effective lifesaving antidote"
The antidote works by supplying uridine to compete with toxic 5-FU metabolites, so its efficacy is functional evidence that the injury is driven by those metabolites reaching proliferating tissue.
Multilineage Cytopenias
Neutropenia, thrombocytopenia, and in severe or complete deficiency pancytopenia, arising from marrow suppression by unclearable fluoropyrimidine rather than from an intrinsic haematological disorder. Severe toxicity of this kind is frequently dose-limiting and can be fatal.
Downstream
Decreased total neutrophil count The lineage most often reported and the one that drives febrile neutropenia risk.
Show evidence (1 reference)
PMID:26573078 SUPPORT Human Clinical
"The risk of grade ≥ 3 toxicity was thereby significantly reduced from 73% (95% CI, 58% to 85%) in historical controls (n = 48) to 28% (95% CI, 10% to 53%) by genotype-guided dosing"
Quantifies severe toxicity in unscreened carriers and its reduction by genotype-guided dosing, which is the strongest evidence that this arm is driven by drug exposure and is preventable.
⬡

Pathograph

Use the checkboxes to hide or show graph categories. Hover nodes for evidence and cross-linked metadata.
Pathograph: causal mechanism network for Dihydropyrimidine Dehydrogenase Deficiency Interactive directed graph showing how pathophysiology mechanisms, phenotypes, genetic factors and variants, experimental models, environmental triggers, and treatments relate through causal and linked edges.
●

Phenotypes

8
Blood 2
Decreased total neutrophil count HP:0001875 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Decreased total neutrophil count (HP:0001875). HP:0001875 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:26603945 SUPPORT Human Clinical
"1236G>A/HapB3 with gastrointestinal toxicity (adjusted RR 5·72, 95% CI 1·40-23·33, p=0·015; and 2·04, 1·49-2·78, p<0·0001, respectively) and haematological toxicity (adjusted RR 9·76, 95% CI 3·03-31·48, p=0·00014; and 2·07, 1·17-3·68, p=0·013, respectively), but not with hand-foot syndrome."
Quantifies the association of specific DPYD variants with haematological toxicity. The same sentence carries the dissociation from hand-foot syndrome, which is why it is quoted whole.
Thrombocytopenia HP:0001873 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Thrombocytopenia (HP:0001873). HP:0001873 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:41259730 SUPPORT Human Clinical
"Neutropenia, diarrhea, and thrombocytopenia were the common toxicities at a frequency of 18.5%, 12.3%, and 15.1%, respectively."
Names thrombocytopenia among the common toxicities and gives its frequency of 15.1 per cent, the third figure in the ordered list.
Digestive 1
Diarrhea HP:0002014 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Diarrhea (HP:0002014). HP:0002014 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:26603945 SUPPORT Human Clinical
"1236G>A/HapB3 with gastrointestinal toxicity (adjusted RR 5·72, 95% CI 1·40-23·33, p=0·015; and 2·04, 1·49-2·78, p<0·0001, respectively) and haematological toxicity (adjusted RR 9·76, 95% CI 3·03-31·48, p=0·00014; and 2·07, 1·17-3·68, p=0·013, respectively), but not with hand-foot syndrome."
Quantifies the gastrointestinal toxicity association for the same variants.
Nervous System 4
Seizure HP:0001250 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Seizure (HP:0001250). HP:0001250 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:10071185 SUPPORT Human Clinical
"A large phenotypic variability has been observed, with convulsive disorders, motor retardation and mental retardation being the most abundant manifestations."
Names convulsive disorders among the most abundant manifestations. No frequency band is assigned because the sentence gives no denominator.
Intellectual disability HP:0001249 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Intellectual disability (HP:0001249). HP:0001249 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:10071185 SUPPORT Human Clinical
"A large phenotypic variability has been observed, with convulsive disorders, motor retardation and mental retardation being the most abundant manifestations."
Reported using the terminology of the period; curated against the current HPO term for intellectual disability.
Global developmental delay HP:0001263 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Global developmental delay (HP:0001263). HP:0001263 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:10071185 SUPPORT Human Clinical
"A large phenotypic variability has been observed, with convulsive disorders, motor retardation and mental retardation being the most abundant manifestations."
Motor retardation in the source, curated against the current HPO term for global developmental delay.
Leukoencephalopathy HP:0002352 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Leukoencephalopathy (HP:0002352). HP:0002352 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:15303009 SUPPORT Human Clinical
"Head MRI showed prominent sulci and abnormal T2 prolongation in the cerebral white matter and brainstem."
Describes the imaging finding including the brainstem involvement.
Other 1
Uraciluria HP:0012127 Human Phenotype Ontology (HP) Relation: this clinical feature is this phenotype This clinical feature is Uraciluria (HP:0012127). HP:0012127 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:10071185 SUPPORT Human Clinical
"An altered beta-alanine, uracil and thymine homeostasis might underlie the various clinical abnormalities encountered in patients with DPD deficiency."
Cited for the accumulation of uracil and thymine that produces the urinary finding, not for the hedged clinical inference in the same sentence.
🧬

Genetic Associations

1
DPYD
Gene: DPYD hgnc:3012 HUGO Gene Nomenclature Committee (hgnc) Relation: this disease-associated gene is this gene This disease-associated gene is DPYD (hgnc:3012). hgnc:3012 is a gene from the HUGO Gene Nomenclature Committee. relationship_type: CAUSATIVE
Autosomal recessive
Show evidence (2 references)
PMID:10071185 SUPPORT Human Clinical
"A clear correlation between the genotype and phenotype has not been established."
The explicit statement that genotype does not predict the neurological phenotype, which is why no such correlation is asserted anywhere in this entry.
PMID:20803296 SUPPORT Human Clinical
"In five patients a deep intronic mutation c.1129-5923C>G was identified creating a cryptic splice donor site."
Identifies the deep intronic variant that is the functional lesion on the haplotype tagged by c.1236G>A, which is the basis for the correction recorded in the notes above.
💊

Medical Actions

5
Pre-Treatment DPYD Genotyping and Genotype-Guided Dosing
Action: Genetic testing to guide fluoropyrimidine dosingNCI Thesaurus (NCIT) Relation: this treatment is this clinical intervention This treatment is Genetic testing to guide fluoropyrimidine dosing, annotated with Genetic Testing (NCIT:C15709). NCIT:C15709 is a clinical intervention from the NCI Thesaurus. Ontology label: Genetic Testing NCIT:C15709
The intervention that defines modern management, and it is preventive rather than therapeutic. Testing DPYD before the first fluoropyrimidine dose and reducing or avoiding the drug according to residual activity converts a lethal idiosyncratic reaction into a manageable one. The evidence is quantitative and among the strongest in pharmacogenomics: upfront genotyping with dose individualisation cut severe toxicity dramatically and abolished drug-induced death in the study population.
Mechanism Target:
MODULATES Impaired Fluoropyrimidine Clearance — Acts by matching the administered dose to the patient's residual clearance capacity, rather than by altering the enzyme deficiency itself.
Show evidence (1 reference)
PMID:26573078 SUPPORT Human Clinical
"drug-induced death was reduced from 10% to 0%"
The outcome that makes this the single most important intervention in the entry.
Show evidence (2 references)
PMID:26573078 SUPPORT Human Clinical
"The risk of grade ≥ 3 toxicity was thereby significantly reduced from 73% (95% CI, 58% to 85%) in historical controls (n = 48) to 28% (95% CI, 10% to 53%) by genotype-guided dosing"
Quantifies the reduction in severe toxicity achieved by upfront genotyping with dose individualisation.
PMID:30348537 SUPPORT Human Clinical
"For DPYD*2A and c.1679T>G carriers, a 50% initial dose reduction was adequate."
PARTIAL because the same sentence pair reports that the reduction used for two other variants was NOT adequate, so this supports the strategy while qualifying the per-variant dose. Cited for the empirical basis rather than as a current dosing recommendation.
Lifelong Fluoropyrimidine Avoidance
Action: Supportive CareNCI Thesaurus (NCIT) Relation: this treatment is this clinical intervention This treatment is Supportive Care (NCIT:C15747). NCIT:C15747 is a clinical intervention from the NCI Thesaurus. NCIT:C15747
For individuals with complete or near-complete deficiency, avoidance of 5-fluorouracil, capecitabine, and tegafur for life. This applies equally to those who are entirely asymptomatic, which is the point: the biochemical diagnosis carries a permanent prescribing implication even in someone who will never have a neurological symptom.
Mechanism Target:
MODULATES Impaired Fluoropyrimidine Clearance — Prevents the exposure that activates the entire toxicity arm, rather than treating its consequences.
Show evidence (1 reference)
PMID:26573078 SUPPORT Human Clinical
"A polymorphism in the fluoropyrimidine metabolizing enzyme dihydropyrimidine dehydrogenase (DPD; ie, DPYD*2A) is strongly associated with fluoropyrimidine-induced severe and life-threatening toxicity."
Establishes the hazard that avoidance is intended to prevent.
Uridine Triacetate
Action: PharmacotherapyNCI Thesaurus (NCIT) Relation: this treatment is this clinical intervention This treatment is Pharmacotherapy (NCIT:C15986). NCIT:C15986 is a clinical intervention from the NCI Thesaurus. NCIT:C15986
Agent: uridine triacetate CHEBI:90914 Chemical Entities of Biological Interest (CHEBI) Relation: this treatment uses this therapeutic agent This treatment uses uridine triacetate (CHEBI:90914). CHEBI:90914 is a therapeutic agent from Chemical Entities of Biological Interest.
Emergency antidote for fluoropyrimidine overexposure or early-onset severe toxicity, including toxicity arising from DPD deficiency. It delivers high concentrations of uridine, which competes with cytotoxic 5-fluorouracil metabolites for incorporation into RNA. Timing is the whole game: it must be given early, and most treated patients received it within 96 hours of exposure.
Mechanism Target:
INHIBITS Fluoropyrimidine Cytotoxicity in Proliferating Tissues — Competes with the toxic metabolites at the point of incorporation, blocking the injury rather than accelerating clearance of the parent drug.
Show evidence (1 reference)
PMID:27622829 SUPPORT Human Clinical
"uridine triacetate was a safe and effective lifesaving antidote"
Establishes efficacy as an antidote in the setting this node describes.
Show evidence (2 references)
PMID:27622829 SUPPORT Human Clinical
"A total of 137 of 142 overdose patients (96%) treated with uridine triacetate survived and had a rapid reversal of severe acute cardiotoxicity and neurotoxicity; in addition, mucositis and leukopenia were prevented, or the patients recovered from them."
The actual survival figure and the specific toxicities reversed, rather than the generic characterisation. Ninety-six per cent survival is the number that makes this an antidote rather than a supportive measure.
PMID:27622829 SUPPORT Human Clinical
"In the historical cohort, 21 of 25 patients (84%) died."
The untreated comparator, and the reason the survival figure above means something. Eighty-four per cent mortality without the antidote against four per cent with it is the whole case for early administration.
Anticonvulsant Therapy
Action: PharmacotherapyNCI Thesaurus (NCIT) Relation: this treatment is this clinical intervention This treatment is Pharmacotherapy (NCIT:C15986). NCIT:C15986 is a clinical intervention from the NCI Thesaurus. NCIT:C15986
Symptomatic seizure control for the constitutive neurological arm. No DPD-specific agent or regimen is established, and seizures are often difficult to control.
Genetic Counselling
Action: Genetic CounselingNCI Thesaurus (NCIT) Relation: this treatment is this clinical intervention This treatment is Genetic Counseling (NCIT:C15240). NCIT:C15240 is a clinical intervention from the NCI Thesaurus. NCIT:C15240
Autosomal recessive recurrence risk for the metabolic disease, cascade testing, and critically, communication of the prescribing implication to relatives. A healthy heterozygous sibling has no metabolic disease and a real chemotherapy hazard.
🌍

Environmental Factors

1
Fluoropyrimidine chemotherapy administration
Administration of 5-fluorouracil, capecitabine, or tegafur. This is the trigger that converts a latent metabolic difference into a life-threatening illness, and it is the only reason most people with reduced DPD activity ever come to medical attention. It is modelled as an environmental exposure rather than as part of the constitutive chain because the entire toxicity arm is inert without it.
Show evidence (2 references)
PMID:26573078 SUPPORT Human Clinical
"A polymorphism in the fluoropyrimidine metabolizing enzyme dihydropyrimidine dehydrogenase (DPD; ie, DPYD*2A) is strongly associated with fluoropyrimidine-induced severe and life-threatening toxicity."
Evidence for the exposure itself being harmful in this population, as distinct from the evidence on the mechanism link below. The exposure entry makes its own causation claim and needs its own citation: administering a fluoropyrimidine to someone with reduced DPD activity is what produces severe and life-threatening toxicity.
PMID:27622829 SUPPORT Human Clinical
"In the historical cohort, 21 of 25 patients (84%) died."
The magnitude of harm this exposure can do when it is not counteracted, which is what makes it worth modelling as a trigger rather than as background context. Stated precisely: that historical cohort is fluoropyrimidine OVERDOSE without antidote, not DPD-deficient patients given a standard dose, so it bounds the severity of unopposed fluoropyrimidine exposure rather than quantifying mortality in this disease.
Mechanism Target:
TRIGGERS Impaired Fluoropyrimidine Clearance — Supplies the substrate that the deficient enzyme cannot clear. Without administration there is nothing to clear and no toxicity, however low the enzyme activity.
Show evidence (1 reference)
PMID:26573078 SUPPORT Human Clinical
"A polymorphism in the fluoropyrimidine metabolizing enzyme dihydropyrimidine dehydrogenase (DPD; ie, DPYD*2A) is strongly associated with fluoropyrimidine-induced severe and life-threatening toxicity."
Establishes that the toxicity is fluoropyrimidine-induced in carriers, which is the exposure-dependence this link asserts.
🔬

Biochemical Markers

1
Plasma uracil concentration (INCREASED)
Show evidence (1 reference)
PMID:38896022 SUPPORT Human Clinical
"According to French recommendations, 7.3 % of patients were partially deficient (U 16-150 ng/mL) and 0.02 % completely deficient"
Gives the partial-deficiency band in nanograms per millilitre with the proportion in it, and the complete-deficiency proportion in the same sentence. The two-hundred-fold gap between them is the whole argument for phenotyping everyone rather than only those with a suggestive history.
🔬

Diagnosis

3
DPYD genotyping
Targeted genotyping for the recognised risk alleles before the first fluoropyrimidine dose. It is fast, widely available, and one of the two methods endorsed by the European Medicines Agency, but it is limited by construction: it can only find variants that are on the panel, which is why it misses patients that phenotyping catches.
Show evidence (1 reference)
PMID:36989883 SUPPORT Human Clinical
"In 2020, the European Medicines Agency (EMA) recommended two methods for pre-treatment DPD deficiency testing in clinical practice: phenotyping using endogenous uracil concentration or genotyping for DPYD risk variant alleles."
Names both endorsed methods, which is the framing for why this entry curates them as complementary rather than ranked.
Uracil phenotyping
Measurement of endogenous plasma uracil as a direct read on residual enzyme activity. The other EMA-endorsed method, and the one that does not depend on knowing which variant a patient carries.
Show evidence (1 reference)
PMID:36989883 SUPPORT Human Clinical
"In 2020, the European Medicines Agency (EMA) recommended two methods for pre-treatment DPD deficiency testing in clinical practice: phenotyping using endogenous uracil concentration or genotyping for DPYD risk variant alleles."
The same recommendation, cited on the phenotyping arm.
Genotype-phenotype discordance in pre-treatment testing
The single most consequential fact about testing in this disease, and the reason both methods are curated rather than one. Genotyping and phenotyping identify overlapping but different patients, and targeted genotyping is not reliable enough on its own to find everyone with complete deficiency. A negative genotype is therefore not a clearance, which matters because the consequence of missing a complete deficiency is a potentially fatal first cycle.
Show evidence (1 reference)
PMID:38896022 SUPPORT Human Clinical
"These results reaffirm the poor concordance between DPD phenotyping and genotyping, suggesting that both approaches may be complementary and that targeted DPYD genotyping is not sufficiently reliable to identify all patients with complete deficiency."
States the discordance, the complementarity, and the specific limitation of targeted genotyping, all in the authors' own words.
📊

Prevalence

3
Worldwide, complete deficiency
Point Prevalence Ultra Rare
Complete DPD deficiency presenting as the constitutive inborn error is very rare and no reliable population rate was identified. Partial deficiency is orders of magnitude more common, which is what makes pre-treatment genotyping worthwhile at population scale; the two must not be quoted as one number. Recorded as a qualitative class rather than a fabricated rate.
Patients with GI and hepatopancreaticobiliary cancers receiving fluoropyrimidines, single Indian centre
Point Prevalence 7500.0 per 100,000 >1 in 1,000
A prospective observational study at a tertiary Indian oncology centre assessed the rate of DPD deficiency among patients receiving fluoropyrimidines for gastrointestinal and hepatopancreaticobiliary cancers. Recorded as a separate population because the relevant denominator for the pharmacogenomic arm is people being prescribed the drug rather than the general population, and kept narrow because this is a single centre and a specific tumour group, not a national or all-comers figure.
Show evidence (1 reference)
PMID:41259730 SUPPORT Human Clinical
"Of the 146 study participants, 11 (7.5%) had a DPYD mutation."
The measured rate in this cohort, 11 of 146, which is the quantitative basis for the band and the normalised rate recorded above. The record previously cited the study's aim sentence, which framed the denominator but stated no rate and could not support a band.
Patients screened before fluoropyrimidine therapy, France (uracil phenotyping)
Point Prevalence 7300.0 per 100,000 >1 in 1,000
Partial deficiency in 7.3 per cent of screened patients by uracil phenotyping under French interpretive thresholds, with complete deficiency two to three orders of magnitude rarer in the same series. Recorded separately from the constitutive inborn error because these are different questions: this is the prevalence of reduced enzyme activity among people about to be given the drug, not the prevalence of the metabolic disease. The two must never be quoted as one number, and the enormous gap between them is the entire argument for pre-treatment screening.
Show evidence (1 reference)
PMID:38896022 SUPPORT Human Clinical
"According to French recommendations, 7.3 % of patients were partially deficient (U 16-150 ng/mL) and 0.02 % completely deficient"
Gives both the partial and complete deficiency proportions from uracil phenotyping under French interpretive thresholds.
🔀

Differential Diagnoses

3

Conditions with similar clinical presentations that must be differentiated from Dihydropyrimidine Dehydrogenase Deficiency:

Dihydropyrimidinase deficiency
Overlapping Features Deficiency of the next enzyme in the same pathway, DPYS. It produces an overlapping biochemical picture with accumulation of pyrimidine intermediates and a similarly variable neurological phenotype, and it shares the fluoropyrimidine-toxicity implication. Distinguished by the specific metabolites that accumulate, since the block sits one step further down.
Beta-ureidopropionase deficiency
Overlapping Features Deficiency of UPB1, the third and final enzyme of pyrimidine catabolism. Completes the set of three pathway disorders that share variable neurological presentation and abnormal urinary pyrimidines, and is distinguished on the same basis.
Fluoropyrimidine toxicity without DPD deficiency
Overlapping Features Severe fluoropyrimidine toxicity, including acute leukoencephalopathy, occurs in patients with entirely normal DPD activity. The important curation consequence is that severe toxicity is not by itself evidence of DPD deficiency, and attributing every case to this disease would over-claim.
🔬

Clinical Trials

1
NCT02324452 PHASE_IV COMPLETED
Prospective multicentre safety analysis of DPYD genotype-guided dose individualisation of fluoropyrimidine therapy, the study underlying the per-variant dose-reduction evidence curated above.
Target Phenotypes: Decreased total neutrophil count HP:0001875 Human Phenotype Ontology (HP) Relation: this clinical trial targets this phenotype This clinical trial targets Decreased total neutrophil count (HP:0001875). HP:0001875 is a phenotype from the Human Phenotype Ontology.
Show evidence (1 reference)
PMID:30348537 SUPPORT Human Clinical
"For DPYD*2A and c.1679T>G carriers, a 50% initial dose reduction was adequate."
The trial's per-variant finding on the adequacy of dose reduction.
🐁

Animal Models

1
Dpyd knockout mouse
Germline Dpyd knockout, generated to study heritable variation in sleep rather than to model this disease. It is included because the entry's central hypothesis depends on beta-alanine acting as an inhibitory neurotransmitter, and this is the animal in which that was shown. Its value here is as much in what it fails to reproduce as in what it does.
Species
Mouse
Genotype
Dpyd homozygous knockout
Publication
PMID:34653361
Pathograph links
Pyrimidine Substrate Accumulation and Beta-Alanine Depletion partially recapitulates fidelity: low Reproduces the metabolic consequence of losing the enzyme and demonstrates a central nervous system phenotype attributable to it, establishing that beta-alanine loss has real consequences for the brain. The phenotype is reduced sleep, not the neurological disease seen in patients.
Limitations: The demonstrated phenotype is a sleep deficit, whereas the human phenotype of interest is seizures, developmental delay, and white-matter injury. A shared molecule and a shared direction of change do not establish a shared consequence, so this model supports the mechanism's plausibility without supporting the specific human chain.
Readouts
Sleep duration during the lights-off period
Direction: DECREASED
Interpretation: Reduced sleep in the knockout is the measured central nervous system consequence of losing the enzyme.
Show evidence (1 reference)
PMID:34653361 SUPPORT Model Organism
"Mice lacking Dpyd had 78.4 min less sleep during the lights-off period than wild-type mice"
The quantitative measurement behind this readout.
Show evidence (1 reference)
PMID:34653361 SUPPORT Model Organism
"Dpyd encodes the rate-limiting enzyme in the metabolic pathway that catabolizes uracil and thymidine to β-alanine, an inhibitory neurotransmitter."
Connects the gene to beta-alanine and to inhibitory neurotransmission, which is why this model is informative for the metabolite node at all.
Impaired Inhibitory Neurotransmission and Neurodevelopmental Injury fails to recapitulate fidelity: low The knockout does not reproduce the neurological phenotype that defines the human constitutive disease. Behavioural testing found no difference in cognitive, social, or affective domains, and no seizure phenotype is reported. This is a substantive negative result and is the reason the human neurological chain remains an EMERGING hypothesis rather than being supported by animal data.
Limitations: Absence of a reported seizure phenotype is not the same as a demonstrated absence of seizure susceptibility, because the study was designed around sleep and did not perform seizure-threshold testing or EEG. The failure to recapitulate is therefore established for the behavioural domains actually assayed, and merely unexamined for seizures themselves.
Show evidence (1 reference)
PMID:34653361 SUPPORT Model Organism
"There was no difference in other measured behaviors in knockout mice, including assays evaluating cognitive-, social-, and affective-disorder-related behaviors."
The explicit negative result across the non-sleep behavioural domains tested, which is what this failure-to-recapitulate claim rests on.
{ }

Source YAML

click to show 
name: Dihydropyrimidine Dehydrogenase Deficiency
creation_date: "2026-08-16T00:00:00Z"
description: >-
  Dihydropyrimidine dehydrogenase (DPD) deficiency is an autosomal recessive
  disorder of pyrimidine catabolism caused by variants in DPYD. DPD catalyses the
  first and rate-limiting step of pyrimidine degradation, so losing it blocks the
  whole pathway at once, producing both an accumulation of uracil and thymine and
  a depletion of the downstream product beta-alanine.

  The unusual feature of this entry is that one enzyme deficiency produces two
  almost unrelated diseases, and which one a person gets depends on whether
  anybody ever gives them a fluoropyrimidine. Left alone, biallelic deficiency is
  a rare inborn error whose neurological consequences range from an epileptic
  encephalopathy to nothing at all. Exposed to 5-fluorouracil or capecitabine,
  the same enzyme deficit becomes a pharmacogenomic emergency, because DPD
  normally clears the great majority of an administered dose and a patient who
  cannot clear it receives, in effect, a massive overdose. Carriers who are
  entirely well their whole lives can be killed by a standard first cycle of
  chemotherapy.

  The two arms are modelled separately here. The constitutive arm runs from
  enzyme loss to metabolite imbalance and is deliberately not asserted as a
  settled cause of the neurological phenotype, because the literature does not
  support that: the originating study reports wide phenotypic variability and
  states plainly that no genotype-phenotype correlation has been established, and
  the mechanism linking metabolites to brain injury is offered by its own authors
  as something that "might underlie" the clinical picture. The toxicity arm is
  gated on drug exposure, modelled as an environmental trigger acting on the
  residual-activity node, and is where the evidence is strong, quantitative, and
  actionable.
category: Mendelian
disease_term:
  preferred_term: Dihydropyrimidine Dehydrogenase Deficiency
  term:
    id: MONDO:0010130
    label: dihydropyrimidine dehydrogenase deficiency
synonyms:
- DPD deficiency
- DPYD deficiency
- Familial pyrimidinemia
- Hereditary thymine-uraciluria
notes: >-
  Scope and framing. This entry covers both the constitutive inborn error
  (biallelic, unexposed) and the fluoropyrimidine-toxicity arm (mono- or
  biallelic, drug-exposed), because they are the same enzyme deficiency and
  separating them into two entries would hide the fact that the second is
  reached from the first only by an exposure. The toxicity phenotypes are
  therefore placed downstream of an exposure-gated node rather than in the
  constitutive phenotype list.

  Module conformance was considered carefully and only one was taken. The marrow
  limb of the toxicity arm conforms to `myelosuppression`, where the disorder
  specific substitution is excess systemic 5-fluorouracil arising from failed
  catabolic clearance rather than an administered cytotoxic dose.
  `epilepsy_excitation_inhibition_imbalance` was NOT declared, because conforming
  would require asserting the beta-alanine inhibitory-neurotransmission
  mechanism, which the literature explicitly hedges; that mechanism is recorded
  as an EMERGING hypothesis instead. `metabolic_intoxication_decompensation` was
  also declined: DPD deficiency does not classically produce catabolic-stress
  triggered crises with acidosis, hyperammonaemia, or hypoglycaemia, and the one
  acute presentation in the literature is a single case report.

  There is no dismech module for pharmacogenomic clearance failure, the pattern
  in which impaired drug catabolism produces supratherapeutic exposure and
  dose-dependent toxicity. It recurs across DPYD with fluoropyrimidines, TPMT and
  NUDT15 with thiopurines, UGT1A1 with irinotecan, and CYP2D6 with codeine. This
  entry would be a natural flagship conformer if such a module is ever written;
  no existing module was stretched to cover it.

  Deliberately not curated. The source deep-research report attributed the DPWG
  genotype-guided dosing recommendation to PMID:31745289 and its central
  illustration of a biochemically deficient but symptom-free sibling to
  PMID:9323575. Neither identifier resolves to a record, and both were flagged as
  probable confabulations by the report's own reference validation. The dosing
  content is therefore sourced here from the prospective studies that generated
  the evidence instead, and the unexposed-sibling observation is not curated at
  all. Beta-alanine supplementation is mechanistically attractive and has no
  clinical evidence; it is not curated as a treatment. Deliberate DPD inhibition
  to potentiate 5-fluorouracil is a real oncology strategy but is not a treatment
  for this disease and appears only here in notes.
pathophysiology:
- name: Biallelic DPYD Loss of Function
  biological_scale: MOLECULAR
  description: >-
    Deleterious variants in DPYD, which encodes dihydropyrimidine dehydrogenase.
    The enzyme is a large homodimeric redox protein carrying two FAD, two FMN,
    and eight iron-sulfur clusters arranged as electron-transfer chains, which is
    why loss of function is reachable by many mechanistically distinct routes:
    disruption of the substrate site, of cofactor binding, of the electron wire,
    or of dimer assembly and folding. Complete deficiency requires two alleles;
    a single deleterious allele reduces activity enough to matter only when a
    fluoropyrimidine is administered.
  molecular_functions:
  - preferred_term: dihydropyrimidine dehydrogenase activity
    modifier: LOSS_OF_FUNCTION
    term:
      id: GO:0017113
      label: dihydropyrimidine dehydrogenase (NADP+) activity
  evidence:
  - reference: PMID:11179210
    reference_title: "Crystal structure of dihydropyrimidine dehydrogenase, a major determinant of the pharmacokinetics of the anti-cancer drug 5-fluorouracil."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "arranged in two electron transfer chains that pass the dimer interface twice"
    explanation: >-
      Describes the electron-transfer architecture that makes this enzyme
      vulnerable to loss of function at many independent positions. The full
      sentence also enumerates the two FAD, two FMN, and eight iron-sulfur
      clusters; the quote stops short of them because the bracketed cluster
      notation is removed by snippet normalisation and cannot be matched.
  downstream:
  - target: Reduced DPD Catalytic Activity
    causal_link_type: DIRECT
    description: >-
      Loss-of-function alleles reduce or abolish enzyme activity in proportion to
      how many alleles are affected and how severely.
- name: Reduced DPD Catalytic Activity
  biological_scale: MOLECULAR
  description: >-
    Diminished or absent dihydropyrimidine dehydrogenase activity. This node is
    the hinge of the entry: everything constitutive runs downstream of it through
    the blocked catabolic pathway, and everything pharmacogenomic runs downstream
    of it only once a fluoropyrimidine arrives. Residual activity is continuous
    rather than binary, which is why the same node supports both a severe
    infantile presentation and a lifelong asymptomatic state.
  biological_processes:
  - preferred_term: pyrimidine nucleobase catabolic process
    modifier: DECREASED
    term:
      id: GO:0006208
      label: pyrimidine nucleobase catabolic process
  evidence:
  - reference: PMID:11179210
    reference_title: "Crystal structure of dihydropyrimidine dehydrogenase, a major determinant of the pharmacokinetics of the anti-cancer drug 5-fluorouracil."
    supports: SUPPORT
    evidence_source: IN_VITRO
    snippet: "the enzyme is also responsible for the rapid breakdown of the chemotherapeutic drug 5-fluorouracil"
    explanation: >-
      Establishes that the same enzyme governs fluoropyrimidine clearance, which
      is what makes one deficiency produce two diseases.
  downstream:
  - target: Block of Pyrimidine Nucleobase Catabolism
    causal_link_type: DIRECT
    description: >-
      The constitutive consequence, present from birth regardless of any drug
      exposure.
  - target: Impaired Fluoropyrimidine Clearance
    causal_link_type: DIRECT
    description: >-
      The pharmacogenomic consequence, which is latent until a fluoropyrimidine
      is administered and is therefore gated by exposure rather than by genotype
      alone.
- name: Block of Pyrimidine Nucleobase Catabolism
  biological_scale: MOLECULAR
  description: >-
    Because DPD catalyses the committed first step, its loss blocks the entire
    three-step catabolic pathway. The consequence is two-sided and both sides may
    matter: substrates accumulate, and the eventual product beta-alanine is not
    made. Uraciluria is the resulting biochemical signature and is present in
    essentially all homozygotes, whether or not they have any symptoms.
  biological_processes:
  - preferred_term: pyrimidine nucleobase catabolic process
    modifier: DECREASED
    term:
      id: GO:0006208
      label: pyrimidine nucleobase catabolic process
  evidence:
  - reference: PMID:10071185
    reference_title: Genotype and phenotype in patients with dihydropyrimidine dehydrogenase deficiency.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "An altered beta-alanine, uracil and thymine homeostasis might underlie the various clinical abnormalities encountered in patients with DPD deficiency."
    explanation: >-
      States both sides of the block, accumulation and depletion, in the authors'
      own deliberately hedged language. The hedge is preserved rather than
      flattened; see the mechanistic hypothesis of the same name.
  downstream:
  - target: Pyrimidine Substrate Accumulation and Beta-Alanine Depletion
    causal_link_type: DIRECT
    description: >-
      The metabolite imbalance that follows directly from blocking the pathway.
- name: Pyrimidine Substrate Accumulation and Beta-Alanine Depletion
  biological_scale: ORGANISM
  description: >-
    Uracil, thymine, and 5-hydroxymethyluracil accumulate in urine, plasma, and
    cerebrospinal fluid, while beta-alanine and beta-aminoisobutyrate are
    depleted. Beta-alanine is a structural analogue of GABA and glycine and has
    inhibitory neurotransmitter activity, which is the basis of the proposed but
    unproven link to the neurological phenotype. This node is where the entry
    stops asserting and starts hypothesising: the metabolite disturbance is
    measured and certain, its consequences for the brain are not.
  biological_processes:
  - preferred_term: beta-alanine biosynthetic process
    modifier: DECREASED
    term:
      id: GO:0019483
      label: beta-alanine biosynthetic process
  evidence:
  - reference: PMID:34653361
    reference_title: The dihydropyrimidine dehydrogenase gene contributes to heritable differences in sleep in mice.
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Both β-alanine and β-AIB are structural analogs of γ-aminobutyric acid (GABA) and glycine, two important inhibitory neurotransmitters in the central nervous system."
    explanation: >-
      Establishes the structural basis for beta-alanine acting on inhibitory
      neurotransmission. Tagged MODEL_ORGANISM because the publication is a mouse
      sleep-genetics study; the structural-analogy statement is general, but the
      functional demonstration in that paper is murine.
  downstream:
  - target: Uraciluria
    causal_link_type: DIRECT
    description: >-
      The diagnostic biochemical signature of the blocked pathway.
  - target: Impaired Inhibitory Neurotransmission and Neurodevelopmental Injury
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    hypothesis_groups:
    - altered_pyrimidine_homeostasis_neurotoxicity
    description: >-
      Proposed route from metabolite imbalance to neurological disease. The
      intermediates are genuinely unknown, and the edge belongs to a hypothesis
      group rather than being asserted, because the same biochemical deficiency
      is compatible with no neurological phenotype at all.
- name: Impaired Inhibitory Neurotransmission and Neurodevelopmental Injury
  biological_scale: CELLULAR
  description: >-
    The hypothesised neurological consequence of the metabolite imbalance:
    beta-alanine and beta-aminoisobutyrate are structural analogues of GABA and
    glycine, so their depletion is proposed to shift inhibitory neurotransmission
    in the developing central nervous system, with accumulated pyrimidines
    contributing independently. This node exists as a node rather than as a bare
    edge to the phenotypes because it is the thing the animal evidence speaks to,
    positively and negatively, and a hypothesis with no node has nowhere to attach
    a negative result. Its edges belong to the neurotoxicity hypothesis group and
    are not asserted: the same biochemical state is compatible with no
    neurological phenotype at all.
  biological_processes:
  - preferred_term: beta-alanine biosynthetic process
    modifier: DECREASED
    term:
      id: GO:0019483
      label: beta-alanine biosynthetic process
  evidence:
  - reference: PMID:34653361
    reference_title: "The dihydropyrimidine dehydrogenase gene contributes to heritable differences in sleep in mice."
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Both β-alanine and β-AIB are structural analogs of γ-aminobutyric acid (GABA) and glycine, two important inhibitory neurotransmitters in the central nervous system."
    explanation: >-
      The structural basis for the proposed effect on inhibitory
      neurotransmission. PARTIAL because analogy to inhibitory neurotransmitters
      establishes a plausible route rather than a demonstrated one in human
      disease.
  - reference: PMID:15303009
    reference_title: Head imaging abnormalities in dihydropyrimidine dehydrogenase deficiency.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The pathogenesis of the white-matter abnormalities is unknown, although environmental factors and altered energy metabolism may be involved."
    explanation: >-
      Included as a limit on this node rather than as support for it. The authors
      who described the imaging phenotype state that its mechanism is unknown and
      name alternatives, which is why the downstream edges sit in a hypothesis
      group.
  downstream:
  - target: Seizure
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    hypothesis_groups:
    - altered_pyrimidine_homeostasis_neurotoxicity
    description: >-
      Proposed route to seizures through altered inhibitory neurotransmission. Not
      asserted; the intermediates are unknown and penetrance is incomplete.
  - target: Leukoencephalopathy
    causal_link_type: INDIRECT_UNKNOWN_INTERMEDIATES
    hypothesis_groups:
    - altered_pyrimidine_homeostasis_neurotoxicity
    description: >-
      Proposed route to the white-matter and brainstem changes, whose pathogenesis
      the describing authors state is unknown.
- name: Impaired Fluoropyrimidine Clearance
  biological_scale: ORGANISM
  description: >-
    Exposure-gated node, and the reason this disease matters far beyond the
    handful of children with the inborn error. Dihydropyrimidine dehydrogenase
    performs the great majority of fluoropyrimidine catabolism, so a patient with
    reduced activity who receives a standard dose of 5-fluorouracil or
    capecitabine cannot clear it. Systemic exposure and half-life rise, and the
    patient effectively receives an overdose while being given a normal
    prescription. Nothing about this node is active until the drug is
    administered.
  biological_processes:
  - preferred_term: pyrimidine nucleobase catabolic process
    modifier: DECREASED
    term:
      id: GO:0006208
      label: pyrimidine nucleobase catabolic process
  evidence:
  - reference: PMID:26573078
    reference_title: "Upfront Genotyping of DPYD*2A to Individualize Fluoropyrimidine Therapy: A Safety and Cost Analysis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A polymorphism in the fluoropyrimidine metabolizing enzyme dihydropyrimidine dehydrogenase (DPD; ie, DPYD*2A) is strongly associated with fluoropyrimidine-induced severe and life-threatening toxicity."
    explanation: >-
      Links reduced DPD activity to severe and life-threatening drug toxicity,
      which is the claim this node carries.
  downstream:
  - target: Fluoropyrimidine Cytotoxicity in Proliferating Tissues
    causal_link_type: DIRECT
    description: >-
      Excess systemic drug exposure delivers a cytotoxic insult to the tissues
      that turn over fastest.
- name: Fluoropyrimidine Cytotoxicity in Proliferating Tissues
  biological_scale: TISSUE
  description: >-
    Supratherapeutic 5-fluorouracil exposure damages rapidly proliferating
    tissues by two routes: the metabolite FdUMP inhibits thymidylate synthase,
    producing thymineless stress and DNA damage, and FUTP is misincorporated into
    RNA. Bone marrow and gastrointestinal mucosa bear the brunt, with skin,
    myocardium, and central nervous system also affected. This node is the
    disease-specific substitution for the cytotoxic driver in the myelosuppression
    module: the insult is not an administered cytotoxic dose but an ordinary dose
    that the patient cannot clear.
  conforms_to: "myelosuppression#Cytotoxic Insult to Proliferating Hematopoietic Progenitors"
  evidence:
  - reference: PMID:27622829
    reference_title: Emergency use of uridine triacetate for the prevention and treatment of life-threatening 5-fluorouracil and capecitabine toxicity.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "uridine triacetate was a safe and effective lifesaving antidote"
    explanation: >-
      The antidote works by supplying uridine to compete with toxic 5-FU
      metabolites, so its efficacy is functional evidence that the injury is
      driven by those metabolites reaching proliferating tissue.
  downstream:
  - target: Multilineage Cytopenias
    causal_link_type: DIRECT
    description: >-
      The marrow limb, which is the conforming arm of the myelosuppression
      module.
  - target: Diarrhea
    causal_link_type: DIRECT
    description: >-
      The gastrointestinal mucosal limb.
- name: Multilineage Cytopenias
  biological_scale: ORGANISM
  description: >-
    Neutropenia, thrombocytopenia, and in severe or complete deficiency
    pancytopenia, arising from marrow suppression by unclearable fluoropyrimidine
    rather than from an intrinsic haematological disorder. Severe toxicity of
    this kind is frequently dose-limiting and can be fatal.
  conforms_to: "myelosuppression#Multilineage Peripheral Cytopenias"
  evidence:
  - reference: PMID:26573078
    reference_title: "Upfront Genotyping of DPYD*2A to Individualize Fluoropyrimidine Therapy: A Safety and Cost Analysis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The risk of grade ≥ 3 toxicity was thereby significantly reduced from 73% (95% CI, 58% to 85%) in historical controls (n = 48) to 28% (95% CI, 10% to 53%) by genotype-guided dosing"
    explanation: >-
      Quantifies severe toxicity in unscreened carriers and its reduction by
      genotype-guided dosing, which is the strongest evidence that this arm is
      driven by drug exposure and is preventable.
  downstream:
  - target: Decreased total neutrophil count
    causal_link_type: DIRECT
    description: >-
      The lineage most often reported and the one that drives febrile
      neutropenia risk.
mechanistic_hypotheses:
- hypothesis_group_id: altered_pyrimidine_homeostasis_neurotoxicity
  hypothesis_label: Altered pyrimidine and beta-alanine homeostasis causes the neurological phenotype
  status: EMERGING
  description: >-
    Holds that accumulated uracil and thymine, together with depleted beta-alanine
    and beta-aminoisobutyrate, disturb the developing central nervous system, and
    that this accounts for the seizures, developmental delay, and white-matter
    abnormalities seen in some patients. The pathway logic is coherent, since
    beta-alanine is a structural analogue of the two principal inhibitory
    neurotransmitters, and the metabolite disturbance itself is beyond doubt. What
    is unproven is the link from that disturbance to brain injury.
  evidence:
  - reference: PMID:10071185
    reference_title: Genotype and phenotype in patients with dihydropyrimidine dehydrogenase deficiency.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "An altered beta-alanine, uracil and thymine homeostasis might underlie the various clinical abnormalities encountered in patients with DPD deficiency."
    explanation: >-
      The originating statement of the hypothesis, graded PARTIAL because the
      authors write "might underlie" rather than "underlies" and offer it as a
      proposal.
  - reference: PMID:34653361
    reference_title: The dihydropyrimidine dehydrogenase gene contributes to heritable differences in sleep in mice.
    supports: SUPPORT
    evidence_source: MODEL_ORGANISM
    snippet: "Dpyd encodes the rate-limiting enzyme in the metabolic pathway that catabolizes uracil and thymidine to β-alanine, an inhibitory neurotransmitter."
    explanation: >-
      The best available functional support that beta-alanine loss has a real
      central nervous system consequence. PARTIAL and MODEL_ORGANISM because the
      demonstrated phenotype is murine sleep, not human seizures.
  - reference: PMID:15303009
    reference_title: Head imaging abnormalities in dihydropyrimidine dehydrogenase deficiency.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The pathogenesis of the white-matter abnormalities is unknown, although environmental factors and altered energy metabolism may be involved."
    explanation: >-
      The authors who described the imaging phenotype state that its mechanism is
      unknown and raise alternatives, which is why this remains a hypothesis
      rather than the entry's asserted chain.
  notes: >-
    Deliberately EMERGING rather than CANONICAL. The decisive problem is not weak
    evidence for the mechanism but the existence of biochemically deficient
    individuals with no neurological phenotype whatsoever, which means the
    metabolite disturbance cannot be sufficient. Any complete account has to
    explain the unaffected as well as the affected, and none currently does.
phenotypes:
- name: Seizure
  category: Neurological
  description: >-
    Convulsive disorder, reported as among the most abundant manifestations of
    the constitutive deficiency and often difficult to control. Frequency is
    deliberately not banded, because the source describes it as abundant without
    providing a denominator that would justify a specific band.
  phenotype_term:
    preferred_term: Seizure
    term:
      id: HP:0001250
      label: Seizure
  evidence:
  - reference: PMID:10071185
    reference_title: Genotype and phenotype in patients with dihydropyrimidine dehydrogenase deficiency.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A large phenotypic variability has been observed, with convulsive disorders, motor retardation and mental retardation being the most abundant manifestations."
    explanation: >-
      Names convulsive disorders among the most abundant manifestations. No
      frequency band is assigned because the sentence gives no denominator.
- name: Intellectual disability
  category: Neurological
  phenotype_term:
    preferred_term: Intellectual disability
    term:
      id: HP:0001249
      label: Intellectual disability
  evidence:
  - reference: PMID:10071185
    reference_title: Genotype and phenotype in patients with dihydropyrimidine dehydrogenase deficiency.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A large phenotypic variability has been observed, with convulsive disorders, motor retardation and mental retardation being the most abundant manifestations."
    explanation: >-
      Reported using the terminology of the period; curated against the current
      HPO term for intellectual disability.
- name: Global developmental delay
  category: Neurological
  phenotype_term:
    preferred_term: Global developmental delay
    term:
      id: HP:0001263
      label: Global developmental delay
  evidence:
  - reference: PMID:10071185
    reference_title: Genotype and phenotype in patients with dihydropyrimidine dehydrogenase deficiency.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A large phenotypic variability has been observed, with convulsive disorders, motor retardation and mental retardation being the most abundant manifestations."
    explanation: >-
      Motor retardation in the source, curated against the current HPO term for
      global developmental delay.
- name: Leukoencephalopathy
  category: Imaging
  description: >-
    White-matter signal abnormality with brainstem involvement. The brainstem
    component is worth noting separately, since it distinguishes this from
    generic white-matter change in metabolic disease.
  phenotype_term:
    preferred_term: Leukoencephalopathy
    term:
      id: HP:0002352
      label: Leukoencephalopathy
  evidence:
  - reference: PMID:15303009
    reference_title: Head imaging abnormalities in dihydropyrimidine dehydrogenase deficiency.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Head MRI showed prominent sulci and abnormal T2 prolongation in the cerebral white matter and brainstem."
    explanation: >-
      Describes the imaging finding including the brainstem involvement.
- name: Uraciluria
  category: Laboratory
  description: >-
    Elevated urinary uracil, with thymine, as the direct biochemical readout of
    the blocked catabolic pathway. Present in essentially all homozygotes
    including those who never develop any clinical phenotype, which is precisely
    why it identifies the disease but does not predict it.
  phenotype_term:
    preferred_term: Uraciluria
    term:
      id: HP:0012127
      label: Uraciluria
  evidence:
  - reference: PMID:10071185
    reference_title: Genotype and phenotype in patients with dihydropyrimidine dehydrogenase deficiency.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "An altered beta-alanine, uracil and thymine homeostasis might underlie the various clinical abnormalities encountered in patients with DPD deficiency."
    explanation: >-
      Cited for the accumulation of uracil and thymine that produces the urinary
      finding, not for the hedged clinical inference in the same sentence.
- name: Decreased total neutrophil count
  category: Haematological
  description: >-
    Neutropenia following fluoropyrimidine exposure. This is a drug-reaction
    phenotype, downstream of the exposure-gated arm, and is not a feature of
    unexposed DPD deficiency.
  phenotype_term:
    preferred_term: Decreased total neutrophil count
    term:
      id: HP:0001875
      label: Decreased total neutrophil count
  evidence:
  - reference: PMID:26603945
    reference_title: "Clinical relevance of DPYD variants c.1679T>G, c.1236G>A/HapB3, and c.1601G>A as predictors of severe fluoropyrimidine-associated toxicity: a systematic review and meta-analysis of individual patient data."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "1236G>A/HapB3 with gastrointestinal toxicity (adjusted RR 5·72, 95% CI 1·40-23·33, p=0·015; and 2·04, 1·49-2·78, p<0·0001, respectively) and haematological toxicity (adjusted RR 9·76, 95% CI 3·03-31·48, p=0·00014; and 2·07, 1·17-3·68, p=0·013, respectively), but not with hand-foot syndrome."
    explanation: >-
      Quantifies the association of specific DPYD variants with haematological
      toxicity. The same sentence carries the dissociation from hand-foot
      syndrome, which is why it is quoted whole.
- name: Diarrhea
  category: Gastrointestinal
  description: >-
    Severe diarrhoea from mucosal injury following fluoropyrimidine exposure.
    Like the cytopenias, a drug-reaction phenotype rather than a constitutive
    feature.
  phenotype_term:
    preferred_term: Diarrhea
    term:
      id: HP:0002014
      label: Diarrhea
  evidence:
  - reference: PMID:26603945
    reference_title: "Clinical relevance of DPYD variants c.1679T>G, c.1236G>A/HapB3, and c.1601G>A as predictors of severe fluoropyrimidine-associated toxicity: a systematic review and meta-analysis of individual patient data."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "1236G>A/HapB3 with gastrointestinal toxicity (adjusted RR 5·72, 95% CI 1·40-23·33, p=0·015; and 2·04, 1·49-2·78, p<0·0001, respectively) and haematological toxicity (adjusted RR 9·76, 95% CI 3·03-31·48, p=0·00014; and 2·07, 1·17-3·68, p=0·013, respectively), but not with hand-foot syndrome."
    explanation: >-
      Quantifies the gastrointestinal toxicity association for the same variants.
- name: Thrombocytopenia
  category: Haematological
  description: >-
    Reduced platelet count following fluoropyrimidine exposure, reported alongside
    neutropenia and diarrhoea as one of the common toxicities in a prospective
    series. A drug-reaction phenotype, not a feature of unexposed deficiency.
  phenotype_term:
    preferred_term: Thrombocytopenia
    term:
      id: HP:0001873
      label: Thrombocytopenia
  evidence:
  - reference: PMID:41259730
    reference_title: "Clinical Implications of Dihydropyrimidine Dehydrogenase Deficiency in GI and Hepatopancreaticobiliary Cancers Treated With Fluoropyrimidines: A Prospective Observational Study."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Neutropenia, diarrhea, and thrombocytopenia were the common toxicities at a frequency of 18.5%, 12.3%, and 15.1%, respectively."
    explanation: >-
      Names thrombocytopenia among the common toxicities and gives its frequency of
      15.1 per cent, the third figure in the ordered list.
environmental:
- name: Fluoropyrimidine chemotherapy administration
  description: >-
    Administration of 5-fluorouracil, capecitabine, or tegafur. This is the
    trigger that converts a latent metabolic difference into a
    life-threatening illness, and it is the only reason most people with reduced
    DPD activity ever come to medical attention. It is modelled as an
    environmental exposure rather than as part of the constitutive chain because
    the entire toxicity arm is inert without it.
  evidence:
  - reference: PMID:26573078
    reference_title: "Upfront Genotyping of DPYD*2A to Individualize Fluoropyrimidine Therapy: A Safety and Cost Analysis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A polymorphism in the fluoropyrimidine metabolizing enzyme dihydropyrimidine dehydrogenase (DPD; ie, DPYD*2A) is strongly associated with fluoropyrimidine-induced severe and life-threatening toxicity."
    explanation: >-
      Evidence for the exposure itself being harmful in this population, as
      distinct from the evidence on the mechanism link below. The exposure entry
      makes its own causation claim and needs its own citation: administering a
      fluoropyrimidine to someone with reduced DPD activity is what produces
      severe and life-threatening toxicity.
  - reference: PMID:27622829
    reference_title: "Emergency use of uridine triacetate for the prevention and treatment of life-threatening 5-fluorouracil and capecitabine toxicity."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In the historical cohort, 21 of 25 patients (84%) died."
    explanation: >-
      The magnitude of harm this exposure can do when it is not counteracted, which
      is what makes it worth modelling as a trigger rather than as background
      context. Stated precisely: that historical cohort is fluoropyrimidine
      OVERDOSE without antidote, not DPD-deficient patients given a standard dose,
      so it bounds the severity of unopposed fluoropyrimidine exposure rather than
      quantifying mortality in this disease.
  influences_mechanisms:
  - target: Impaired Fluoropyrimidine Clearance
    environmental_effect: TRIGGERS
    causal_link_type: DIRECT
    description: >-
      Supplies the substrate that the deficient enzyme cannot clear. Without
      administration there is nothing to clear and no toxicity, however low the
      enzyme activity.
    evidence:
    - reference: PMID:26573078
      reference_title: "Upfront Genotyping of DPYD*2A to Individualize Fluoropyrimidine Therapy: A Safety and Cost Analysis."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "A polymorphism in the fluoropyrimidine metabolizing enzyme dihydropyrimidine dehydrogenase (DPD; ie, DPYD*2A) is strongly associated with fluoropyrimidine-induced severe and life-threatening toxicity."
      explanation: >-
        Establishes that the toxicity is fluoropyrimidine-induced in carriers,
        which is the exposure-dependence this link asserts.
genetic:
- name: DPYD
  gene_term:
    preferred_term: DPYD
    term:
      id: hgnc:3012
      label: DPYD
  relationship_type: CAUSATIVE
  inheritance:
  - name: Autosomal recessive
    inheritance_term:
      preferred_term: Autosomal recessive inheritance
      term:
        id: HP:0000007
        label: Autosomal recessive inheritance
  notes: >-
    Encodes dihydropyrimidine dehydrogenase. Inheritance is autosomal recessive
    for the constitutive inborn error, but the pharmacogenomic arm is dose
    dependent rather than recessive: a single deleterious allele reduces activity
    enough to cause severe or fatal fluoropyrimidine toxicity, which is why
    heterozygotes who are otherwise entirely healthy are the population that
    pre-treatment genotyping is designed to protect. The clinically actionable
    variants include DPYD*2A (c.1905+1G>A), c.1679T>G, c.1236G>A/HapB3, and
    c.2846A>T. No genotype-phenotype correlation has been established for the
    neurological arm.

    One correction worth stating explicitly, because getting it wrong is a common
    error in entries for this disease: c.1236G>A is a linkage tag rather than the
    causal allele. It is assayed because it is easy to genotype and marks the HapB3
    haplotype, but the functional lesion on that haplotype is a deep intronic
    variant, c.1129-5923C>G, which creates a cryptic splice donor site and inserts a
    44 base-pair pseudo-exon into the transcript. A panel that reports c.1236G>A is
    reporting the flag, not the mechanism, and any mechanistic statement should
    attach to the intronic variant.

    Allele frequencies are strongly ancestry-dependent, which makes panel design an
    equity question rather than a technical one. DPYD*2A, c.2846A>T, and c.1679T>G
    are essentially absent in East Asian populations, while c.557A>G is markedly
    enriched in people of African ancestry and is not on every panel. A panel
    assembled from variants discovered in European cohorts will therefore
    under-detect risk in exactly the populations it was not built from, and a
    negative result carries different meaning depending on the patient's
    ancestry.
  evidence:
  - reference: PMID:10071185
    reference_title: Genotype and phenotype in patients with dihydropyrimidine dehydrogenase deficiency.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A clear correlation between the genotype and phenotype has not been established."
    explanation: >-
      The explicit statement that genotype does not predict the neurological
      phenotype, which is why no such correlation is asserted anywhere in this
      entry.
  - reference: PMID:20803296
    reference_title: "Intragenic deletions and a deep intronic mutation affecting pre-mRNA splicing in the dihydropyrimidine dehydrogenase gene as novel mechanisms causing 5-fluorouracil toxicity."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In five patients a deep intronic mutation c.1129-5923C>G was identified creating a cryptic splice donor site."
    explanation: >-
      Identifies the deep intronic variant that is the functional lesion on the
      haplotype tagged by c.1236G>A, which is the basis for the correction recorded
      in the notes above.
inheritance:
- name: Autosomal recessive
  inheritance_term:
    preferred_term: Autosomal recessive inheritance
    term:
      id: HP:0000007
      label: Autosomal recessive inheritance
  description: >-
    The constitutive metabolic disease requires biallelic variants. Fluoropyrimidine
    toxicity risk does not: it follows residual enzyme activity, so heterozygous
    carriers are at substantially increased risk despite being unaffected in
    every other respect.
treatments:
- name: Pre-Treatment DPYD Genotyping and Genotype-Guided Dosing
  description: >-
    The intervention that defines modern management, and it is preventive rather
    than therapeutic. Testing DPYD before the first fluoropyrimidine dose and
    reducing or avoiding the drug according to residual activity converts a
    lethal idiosyncratic reaction into a manageable one. The evidence is
    quantitative and among the strongest in pharmacogenomics: upfront genotyping
    with dose individualisation cut severe toxicity dramatically and abolished
    drug-induced death in the study population.
  therapeutic_modality: BEHAVIORAL
  treatment_term:
    preferred_term: Genetic testing to guide fluoropyrimidine dosing
    term:
      id: NCIT:C15709
      label: Genetic Testing
  target_mechanisms:
  - target: Impaired Fluoropyrimidine Clearance
    treatment_effect: MODULATES
    description: >-
      Acts by matching the administered dose to the patient's residual clearance
      capacity, rather than by altering the enzyme deficiency itself.
    evidence:
    - reference: PMID:26573078
      reference_title: "Upfront Genotyping of DPYD*2A to Individualize Fluoropyrimidine Therapy: A Safety and Cost Analysis."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "drug-induced death was reduced from 10% to 0%"
      explanation: >-
        The outcome that makes this the single most important intervention in the
        entry.
  evidence:
  - reference: PMID:26573078
    reference_title: "Upfront Genotyping of DPYD*2A to Individualize Fluoropyrimidine Therapy: A Safety and Cost Analysis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "The risk of grade ≥ 3 toxicity was thereby significantly reduced from 73% (95% CI, 58% to 85%) in historical controls (n = 48) to 28% (95% CI, 10% to 53%) by genotype-guided dosing"
    explanation: >-
      Quantifies the reduction in severe toxicity achieved by upfront genotyping
      with dose individualisation.
  - reference: PMID:30348537
    reference_title: "DPYD genotype-guided dose individualisation of fluoropyrimidine therapy in patients with cancer: a prospective safety analysis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "For DPYD*2A and c.1679T>G carriers, a 50% initial dose reduction was adequate."
    explanation: >-
      PARTIAL because the same sentence pair reports that the reduction used for
      two other variants was NOT adequate, so this supports the strategy while
      qualifying the per-variant dose. Cited for the empirical basis rather than
      as a current dosing recommendation.
  notes: >-
    The prospective study found that a 50 per cent reduction sufficed for DPYD*2A
    and c.1679T>G but that the 25 per cent reduction it used for c.1236G>A and
    c.2846A>T required further investigation. Current guidance uses a uniform 50
    per cent reduction for intermediate metabolisers. Any curated dosing algorithm
    should cite the current guideline for the recommendation itself and this study
    only for the empirical basis.
- name: Lifelong Fluoropyrimidine Avoidance
  description: >-
    For individuals with complete or near-complete deficiency, avoidance of
    5-fluorouracil, capecitabine, and tegafur for life. This applies equally to
    those who are entirely asymptomatic, which is the point: the biochemical
    diagnosis carries a permanent prescribing implication even in someone who will
    never have a neurological symptom.
  therapeutic_modality: BEHAVIORAL
  treatment_term:
    preferred_term: Supportive Care
    term:
      id: NCIT:C15747
      label: Supportive Care
  target_mechanisms:
  - target: Impaired Fluoropyrimidine Clearance
    treatment_effect: MODULATES
    description: >-
      Prevents the exposure that activates the entire toxicity arm, rather than
      treating its consequences.
    evidence:
    - reference: PMID:26573078
      reference_title: "Upfront Genotyping of DPYD*2A to Individualize Fluoropyrimidine Therapy: A Safety and Cost Analysis."
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "A polymorphism in the fluoropyrimidine metabolizing enzyme dihydropyrimidine dehydrogenase (DPD; ie, DPYD*2A) is strongly associated with fluoropyrimidine-induced severe and life-threatening toxicity."
      explanation: >-
        Establishes the hazard that avoidance is intended to prevent.
- name: Uridine Triacetate
  description: >-
    Emergency antidote for fluoropyrimidine overexposure or early-onset severe
    toxicity, including toxicity arising from DPD deficiency. It delivers high
    concentrations of uridine, which competes with cytotoxic 5-fluorouracil
    metabolites for incorporation into RNA. Timing is the whole game: it must be
    given early, and most treated patients received it within 96 hours of
    exposure.
  therapeutic_modality: SMALL_MOLECULE
  treatment_term:
    preferred_term: Pharmacotherapy
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
    therapeutic_agent:
    - preferred_term: uridine triacetate
      term:
        id: CHEBI:90914
        label: uridine triacetate
  target_mechanisms:
  - target: Fluoropyrimidine Cytotoxicity in Proliferating Tissues
    treatment_effect: INHIBITS
    description: >-
      Competes with the toxic metabolites at the point of incorporation, blocking
      the injury rather than accelerating clearance of the parent drug.
    evidence:
    - reference: PMID:27622829
      reference_title: Emergency use of uridine triacetate for the prevention and treatment of life-threatening 5-fluorouracil and capecitabine toxicity.
      supports: SUPPORT
      evidence_source: HUMAN_CLINICAL
      snippet: "uridine triacetate was a safe and effective lifesaving antidote"
      explanation: >-
        Establishes efficacy as an antidote in the setting this node describes.
  evidence:
  - reference: PMID:27622829
    reference_title: Emergency use of uridine triacetate for the prevention and treatment of life-threatening 5-fluorouracil and capecitabine toxicity.
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "A total of 137 of 142 overdose patients (96%) treated with uridine triacetate survived and had a rapid reversal of severe acute cardiotoxicity and neurotoxicity; in addition, mucositis and leukopenia were prevented, or the patients recovered from them."
    explanation: >-
      The actual survival figure and the specific toxicities reversed, rather than
      the generic characterisation. Ninety-six per cent survival is the number that
      makes this an antidote rather than a supportive measure.
  - reference: PMID:27622829
    reference_title: "Emergency use of uridine triacetate for the prevention and treatment of life-threatening 5-fluorouracil and capecitabine toxicity."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In the historical cohort, 21 of 25 patients (84%) died."
    explanation: >-
      The untreated comparator, and the reason the survival figure above means
      something. Eighty-four per cent mortality without the antidote against four
      per cent with it is the whole case for early administration.
- name: Anticonvulsant Therapy
  description: >-
    Symptomatic seizure control for the constitutive neurological arm. No
    DPD-specific agent or regimen is established, and seizures are often
    difficult to control.
  therapeutic_modality: SMALL_MOLECULE
  treatment_term:
    preferred_term: Pharmacotherapy
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
- name: Genetic Counselling
  description: >-
    Autosomal recessive recurrence risk for the metabolic disease, cascade
    testing, and critically, communication of the prescribing implication to
    relatives. A healthy heterozygous sibling has no metabolic disease and a real
    chemotherapy hazard.
  therapeutic_modality: BEHAVIORAL
  treatment_term:
    preferred_term: Genetic Counseling
    term:
      id: NCIT:C15240
      label: Genetic Counseling
clinical_trials:
- name: NCT02324452
  phase: PHASE_IV
  status: COMPLETED
  description: >-
    Prospective multicentre safety analysis of DPYD genotype-guided dose
    individualisation of fluoropyrimidine therapy, the study underlying the
    per-variant dose-reduction evidence curated above.
  target_phenotypes:
  - preferred_term: Decreased total neutrophil count
    term:
      id: HP:0001875
      label: Decreased total neutrophil count
  evidence:
  - reference: PMID:30348537
    reference_title: "DPYD genotype-guided dose individualisation of fluoropyrimidine therapy in patients with cancer: a prospective safety analysis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "For DPYD*2A and c.1679T>G carriers, a 50% initial dose reduction was adequate."
    explanation: >-
      The trial's per-variant finding on the adequacy of dose reduction.
biochemical:
- name: Plasma uracil concentration
  notes: >-
    Endogenous plasma uracil is the phenotyping approach to identifying DPD
    deficiency before treatment, and is one of the two methods the European
    Medicines Agency recommends. It measures what actually matters, namely residual
    enzyme activity, rather than inferring it from a handful of known variants.
    Interpretation is banded rather than binary, with partial deficiency between 16
    and 150 nanograms per millilitre and complete deficiency above 150 under French
    recommendations. High pre-treatment uracil predicts severe toxicity.
  presence: INCREASED
  evidence:
  - reference: PMID:38896022
    reference_title: "Can we identify patients carrying targeted deleterious DPYD variants with plasma uracil and dihydrouracil? A GPCO-RNPGx retrospective analysis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "According to French recommendations, 7.3 % of patients were partially deficient (U 16-150 ng/mL) and 0.02 % completely deficient"
    explanation: >-
      Gives the partial-deficiency band in nanograms per millilitre with the
      proportion in it, and the complete-deficiency proportion in the same sentence.
      The two-hundred-fold gap between them is the whole argument for phenotyping
      everyone rather than only those with a suggestive history.
diagnosis:
- name: DPYD genotyping
  description: >-
    Targeted genotyping for the recognised risk alleles before the first
    fluoropyrimidine dose. It is fast, widely available, and one of the two methods
    endorsed by the European Medicines Agency, but it is limited by construction: it
    can only find variants that are on the panel, which is why it misses patients
    that phenotyping catches.
  evidence:
  - reference: PMID:36989883
    reference_title: "Implementation of dihydropyrimidine dehydrogenase deficiency testing in Europe."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In 2020, the European Medicines Agency (EMA) recommended two methods for pre-treatment DPD deficiency testing in clinical practice: phenotyping using endogenous uracil concentration or genotyping for DPYD risk variant alleles."
    explanation: >-
      Names both endorsed methods, which is the framing for why this entry curates
      them as complementary rather than ranked.
- name: Uracil phenotyping
  description: >-
    Measurement of endogenous plasma uracil as a direct read on residual enzyme
    activity. The other EMA-endorsed method, and the one that does not depend on
    knowing which variant a patient carries.
  evidence:
  - reference: PMID:36989883
    reference_title: "Implementation of dihydropyrimidine dehydrogenase deficiency testing in Europe."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In 2020, the European Medicines Agency (EMA) recommended two methods for pre-treatment DPD deficiency testing in clinical practice: phenotyping using endogenous uracil concentration or genotyping for DPYD risk variant alleles."
    explanation: >-
      The same recommendation, cited on the phenotyping arm.
- name: Genotype-phenotype discordance in pre-treatment testing
  description: >-
    The single most consequential fact about testing in this disease, and the reason
    both methods are curated rather than one. Genotyping and phenotyping identify
    overlapping but different patients, and targeted genotyping is not reliable
    enough on its own to find everyone with complete deficiency. A negative genotype
    is therefore not a clearance, which matters because the consequence of missing a
    complete deficiency is a potentially fatal first cycle.
  evidence:
  - reference: PMID:38896022
    reference_title: "Can we identify patients carrying targeted deleterious DPYD variants with plasma uracil and dihydrouracil? A GPCO-RNPGx retrospective analysis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "These results reaffirm the poor concordance between DPD phenotyping and genotyping, suggesting that both approaches may be complementary and that targeted DPYD genotyping is not sufficiently reliable to identify all patients with complete deficiency."
    explanation: >-
      States the discordance, the complementarity, and the specific limitation of
      targeted genotyping, all in the authors' own words.
animal_models:
- name: Dpyd knockout mouse
  species: Mouse
  genotype: Dpyd homozygous knockout
  publication: PMID:34653361
  description: >-
    Germline Dpyd knockout, generated to study heritable variation in sleep rather
    than to model this disease. It is included because the entry's central
    hypothesis depends on beta-alanine acting as an inhibitory neurotransmitter, and
    this is the animal in which that was shown. Its value here is as much in what it
    fails to reproduce as in what it does.
  modeled_mechanisms:
  - target: Pyrimidine Substrate Accumulation and Beta-Alanine Depletion
    relationship: PARTIALLY_RECAPITULATES
    fidelity: LOW
    description: >-
      Reproduces the metabolic consequence of losing the enzyme and demonstrates a
      central nervous system phenotype attributable to it, establishing that
      beta-alanine loss has real consequences for the brain. The phenotype is
      reduced sleep, not the neurological disease seen in patients.
    limitations: >-
      The demonstrated phenotype is a sleep deficit, whereas the human phenotype of
      interest is seizures, developmental delay, and white-matter injury. A shared
      molecule and a shared direction of change do not establish a shared
      consequence, so this model supports the mechanism's plausibility without
      supporting the specific human chain.
    readouts:
    - name: Sleep duration during the lights-off period
      target: Pyrimidine Substrate Accumulation and Beta-Alanine Depletion
      direction: DECREASED
      interpretation: >-
        Reduced sleep in the knockout is the measured central nervous system
        consequence of losing the enzyme.
      evidence:
      - reference: PMID:34653361
        reference_title: "The dihydropyrimidine dehydrogenase gene contributes to heritable differences in sleep in mice."
        supports: SUPPORT
        evidence_source: MODEL_ORGANISM
        snippet: "Mice lacking Dpyd had 78.4 min less sleep during the lights-off period than wild-type mice"
        explanation: >-
          The quantitative measurement behind this readout.
    evidence:
    - reference: PMID:34653361
      reference_title: "The dihydropyrimidine dehydrogenase gene contributes to heritable differences in sleep in mice."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "Dpyd encodes the rate-limiting enzyme in the metabolic pathway that catabolizes uracil and thymidine to β-alanine, an inhibitory neurotransmitter."
      explanation: >-
        Connects the gene to beta-alanine and to inhibitory neurotransmission, which
        is why this model is informative for the metabolite node at all.
  - target: Impaired Inhibitory Neurotransmission and Neurodevelopmental Injury
    relationship: FAILS_TO_RECAPITULATE
    fidelity: LOW
    description: >-
      The knockout does not reproduce the neurological phenotype that defines the
      human constitutive disease. Behavioural testing found no difference in
      cognitive, social, or affective domains, and no seizure phenotype is reported.
      This is a substantive negative result and is the reason the human neurological
      chain remains an EMERGING hypothesis rather than being supported by animal
      data.
    limitations: >-
      Absence of a reported seizure phenotype is not the same as a demonstrated
      absence of seizure susceptibility, because the study was designed around sleep
      and did not perform seizure-threshold testing or EEG. The failure to
      recapitulate is therefore established for the behavioural domains actually
      assayed, and merely unexamined for seizures themselves.
    evidence:
    - reference: PMID:34653361
      reference_title: "The dihydropyrimidine dehydrogenase gene contributes to heritable differences in sleep in mice."
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "There was no difference in other measured behaviors in knockout mice, including assays evaluating cognitive-, social-, and affective-disorder-related behaviors."
      explanation: >-
        The explicit negative result across the non-sleep behavioural domains
        tested, which is what this failure-to-recapitulate claim rests on.
discussions:
- discussion_id: dpd_neuro_penetrance
  prompt: >-
    Why do some individuals with complete DPD deficiency have severe epileptic
    encephalopathy while others with the same biochemical deficiency are entirely
    asymptomatic?
  kind: KNOWLEDGE_GAP
  status: OPEN
  attaches_to:
  - pathophysiology#Pyrimidine Substrate Accumulation and Beta-Alanine Depletion
  rationale: >-
    This is the central unresolved problem of the constitutive arm and it is not
    a matter of incomplete ascertainment. The originating genotype-phenotype study
    states plainly that no clear correlation has been established, and a wide
    range of severity including complete absence of signs is recognised. The
    metabolite disturbance is therefore necessary but demonstrably not sufficient
    for the neurological phenotype, which means something else, whether modifier
    genes, timing of exposure during development, or an environmental
    contribution, determines whether the brain is affected. Until that is
    identified, the metabolite-to-brain chain cannot be asserted, which is why it
    is curated as an EMERGING hypothesis and its edges are placed in a hypothesis
    group rather than stated outright.
  proposed_experiments:
  - experiment_id: dpd_discordant_sibling_sequencing
    name: Modifier discovery in phenotype-discordant DPD-deficient relatives
    description: >-
      Recruit families containing biochemically deficient individuals discordant
      for the neurological phenotype, and perform whole-genome sequencing and
      cerebrospinal-fluid metabolite profiling on affected and unaffected members
      to identify modifiers segregating with the phenotype rather than with the
      DPYD genotype. Discordant relatives are the only design that controls for
      the primary defect while varying the outcome.
  - experiment_id: dpd_csf_metabolite_correlation
    name: CSF pyrimidine and beta-alanine levels against neurological outcome
    description: >-
      Measure cerebrospinal-fluid uracil, thymine, beta-alanine, and
      beta-aminoisobutyrate in a cohort of biochemically deficient individuals
      stratified by neurological status. If the hypothesis holds, central
      metabolite levels rather than urinary ones should separate affected from
      unaffected; if they do not, the hypothesis is in serious trouble.
- discussion_id: dpd_beta_alanine_model_mismatch
  prompt: >-
    Does the murine demonstration that beta-alanine acts as a sleep-promoting
    inhibitory neurotransmitter translate to a role in human DPD-deficiency
    seizures?
  kind: HUMAN_MODEL_MISMATCH
  status: OPEN
  attaches_to:
  - pathophysiology#Pyrimidine Substrate Accumulation and Beta-Alanine Depletion
  rationale: >-
    Evidence exists in the model and is good: mouse work links Dpyd to
    beta-alanine and establishes beta-alanine as an inhibitory neurotransmitter
    that promotes sleep. What is open is whether that function is the one that
    matters in human disease. The murine phenotype is altered sleep; the human
    phenotype under investigation is seizures, developmental delay, and
    white-matter injury. A shared molecule and a shared direction of change do not
    establish a shared consequence, and treating the mouse result as human
    evidence would overstate the case for the whole neurological chain.
  proposed_experiments:
  - experiment_id: dpd_mouse_seizure_phenotyping
    name: Seizure susceptibility phenotyping in Dpyd-deficient mice
    description: >-
      Characterise seizure threshold, EEG, and white-matter integrity in
      Dpyd-deficient mice, rather than sleep alone, and test whether beta-alanine
      supplementation modifies those endpoints. This would establish whether the
      murine model speaks to the human phenotype at all, or only to sleep.
differential_diagnoses:
- name: Dihydropyrimidinase deficiency
  description: >-
    Deficiency of the next enzyme in the same pathway, DPYS. It produces an
    overlapping biochemical picture with accumulation of pyrimidine intermediates
    and a similarly variable neurological phenotype, and it shares the
    fluoropyrimidine-toxicity implication. Distinguished by the specific
    metabolites that accumulate, since the block sits one step further down.
- name: Beta-ureidopropionase deficiency
  description: >-
    Deficiency of UPB1, the third and final enzyme of pyrimidine catabolism.
    Completes the set of three pathway disorders that share variable neurological
    presentation and abnormal urinary pyrimidines, and is distinguished on the
    same basis.
- name: Fluoropyrimidine toxicity without DPD deficiency
  description: >-
    Severe fluoropyrimidine toxicity, including acute leukoencephalopathy, occurs
    in patients with entirely normal DPD activity. The important curation
    consequence is that severe toxicity is not by itself evidence of DPD
    deficiency, and attributing every case to this disease would over-claim.
prevalence:
- population: Worldwide, complete deficiency
  measure_type: POINT_PREVALENCE
  prevalence_class: ULTRA_RARE
  notes: >-
    Complete DPD deficiency presenting as the constitutive inborn error is very
    rare and no reliable population rate was identified. Partial deficiency is
    orders of magnitude more common, which is what makes pre-treatment genotyping
    worthwhile at population scale; the two must not be quoted as one number.
    Recorded as a qualitative class rather than a fabricated rate.
- population: Patients with GI and hepatopancreaticobiliary cancers receiving fluoropyrimidines, single Indian centre
  measure_type: POINT_PREVALENCE
  prevalence_class: ABOVE_1_IN_1000
  rate_per_100000: 7500.0
  notes: >-
    A prospective observational study at a tertiary Indian oncology centre
    assessed the rate of DPD deficiency among patients receiving fluoropyrimidines
    for gastrointestinal and hepatopancreaticobiliary cancers. Recorded as a
    separate population because the relevant denominator for the pharmacogenomic
    arm is people being prescribed the drug rather than the general population,
    and kept narrow because this is a single centre and a specific tumour group,
    not a national or all-comers figure.
  evidence:
  - reference: PMID:41259730
    reference_title: "Clinical Implications of Dihydropyrimidine Dehydrogenase Deficiency in GI and Hepatopancreaticobiliary Cancers Treated With Fluoropyrimidines: A Prospective Observational Study."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "Of the 146 study participants, 11 (7.5%) had a DPYD mutation."
    explanation: >-
      The measured rate in this cohort, 11 of 146, which is the quantitative basis
      for the band and the normalised rate recorded above. The record previously
      cited the study's aim sentence, which framed the denominator but stated no
      rate and could not support a band.
- population: Patients screened before fluoropyrimidine therapy, France (uracil phenotyping)
  measure_type: POINT_PREVALENCE
  prevalence_class: ABOVE_1_IN_1000
  rate_per_100000: 7300.0
  notes: >-
    Partial deficiency in 7.3 per cent of screened patients by uracil phenotyping
    under French interpretive thresholds, with complete deficiency two to three
    orders of magnitude rarer in the same series. Recorded separately from the
    constitutive inborn error because these are different questions: this is the
    prevalence of reduced enzyme activity among people about to be given the drug,
    not the prevalence of the metabolic disease. The two must never be quoted as
    one number, and the enormous gap between them is the entire argument for
    pre-treatment screening.
  evidence:
  - reference: PMID:38896022
    reference_title: "Can we identify patients carrying targeted deleterious DPYD variants with plasma uracil and dihydrouracil? A GPCO-RNPGx retrospective analysis."
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "According to French recommendations, 7.3 % of patients were partially deficient (U 16-150 ng/mL) and 0.02 % completely deficient"
    explanation: >-
      Gives both the partial and complete deficiency proportions from uracil
      phenotyping under French interpretive thresholds.
📚

References & Deep Research

Deep Research

1
Claude Code ▸
1. Disease Information
claude-haiku-4-5-20251001, claude-opus-5[1m] 15 citations 2026-08-16T10:19:18.966402

1. Disease Information

Overview

Dihydropyrimidine dehydrogenase (DPD) deficiency is an autosomal recessive disorder of pyrimidine catabolism caused by biallelic or monoallelic loss-of-function variation in DPYD. DPD is the first and rate-limiting enzyme of the three-step pyrimidine degradation pathway, converting uracil and thymine to 5,6-dihydrouracil and 5,6-dihydrothymine respectively, in an NADPH-dependent reduction.

Because the same enzyme is responsible for catabolizing >80% of an administered dose of 5-fluorouracil (5-FU), the deficiency has a dual identity: a classical inborn error of metabolism presenting in infancy with thymine-uraciluria and neurological disease, and the single most important pharmacogenetic determinant of severe and fatal fluoropyrimidine toxicity.

The canonical framing from the largest phenotype series [verified this session]:

"Dihydropyrimidine dehydrogenase (DPD) deficiency is an autosomal recessive disease characterised by thymine-uraciluria in homozygous deficient patients and has been associated with a variable clinical phenotype." — Van Kuilenburg et al., Hum Genet 1999;104(1):1–9. PMID:10071185

And the pharmacogenetic framing [verified this session]:

"Fluoropyrimidine treatment can result in severe toxicity in up to 30% of patients and is often the result of reduced activity of the key metabolic enzyme dihydropyrimidine dehydrogenase (DPD), mostly caused by genetic variants in the gene encoding DPD (DPYD)." — Henricks et al., Lancet Oncol 2018;19(11):1459–1467. PMID:30348537

Key identifiers [verified this session via OLS4 MONDO term + xrefs, and HGNC REST]

Resource Identifier
MONDO MONDO:0010130 — dihydropyrimidine dehydrogenase deficiency
OMIM (disease) OMIM:274270 — DIHYDROPYRIMIDINE DEHYDROGENASE DEFICIENCY; DPYDD
OMIM (gene) OMIM:612779 — DPYD
Orphanet ORPHA:1675 (note: not ORPHA:37, a common misattribution)
MeSH MESH:D054067
NCIT NCIT:C84672
DOID DOID:14218
SNOMED CT 77365006
UMLS C1959620
MedGen 409522
MedDRA 10052622
GARD 0000019
ICD-9 277.2 (MONDO relatedTo)
ICD-11 foundation 701689290
HGNC HGNC:3012 — DPYD, dihydropyrimidine dehydrogenase, 1p21.3
NCBI Gene 1806
Ensembl ENSG00000188641
UniProt Q12882 (DPYD_HUMAN)
EC number EC 1.3.1.2

ICD-10 is conventionally E79.8 (other disorders of purine and pyrimidine metabolism) [standard reference, not re-verified].

Synonyms [verified this session, from MONDO exact/related synonyms]

  • DPD deficiency
  • DPYD deficiency ("DYPD deficiency" appears as a typo-synonym in MONDO)
  • Dihydrouracil dehydrogenase deficiency
  • Familial pyrimidinaemia / familial pyrimidinemia
  • Thymine-uraciluria; hereditary thymine-uraciluria (related synonym)
  • Fluorouracil toxicity / 5-FU toxicity (used loosely for the pharmacogenetic arm; do not treat as an exact synonym — the toxicity is a drug reaction, not the metabolic disorder)

Nature of the evidence base

This is an unusual disease for dismech because the two arms have completely different data provenance:

  • Inborn-error arm: individual patient reports and small aggregated series. The largest single dataset is 22 patients from 17 families (PMID:10071185). Aggregated disease-level resources (OMIM, Orphanet, GARD) essentially summarize these case series.
  • Pharmacogenetic arm: very large prospective cohorts and EHR/laboratory datasets. Deenen et al. screened 2,038 patients; Henricks et al. enrolled 1,181; the GPCO-RNPGx analysis covers 19,376 consecutive French patients with paired phenotype and genotype. This arm has population-scale, individual-level data of a quality most rare diseases never see.

2. Etiology

Primary causal factor

Germline loss-of-function variation in DPYD (1p21.3), inherited in an autosomal recessive fashion for the metabolic disease. For fluoropyrimidine toxicity risk the inheritance behaves as codominant/gene-dosage: heterozygotes have measurably reduced DPD activity and materially elevated toxicity risk, so from a drug-safety standpoint a single deleterious allele is clinically actionable.

DPYD is one of the largest genes in the human genome (~950 kb, 23 exons) [standard reference, not re-verified], which matters mechanistically — it is a large mutational target, and this is part of why deep intronic and structural lesions are a recurring and under-ascertained cause (see Section 4).

Genetic risk factors

The four variants with established clinical validity and the basis of every current guideline panel [verified this session]:

"We assessed the effect of prospective screening for the four most relevant DPYD variants (DPYD*2A [rs3918290, c.1905+1G>A, IVS14+1G>A], c.2846A>T [rs67376798, D949V], c.1679T>G [rs55886062, DPYD*13, I560S], and c.1236G>A [rs56038477, E412E, in haplotype B3]) on patient safety..." — Henricks et al. 2018, PMID:30348537

The meta-analysis that established clinical validity for the latter two [verified this session]:

"DPYD c.1679T>G was significantly associated with fluoropyrimidine-associated toxicity (adjusted RR 4·40, 95% CI 2·08-9·30, p<0·0001)... DPYD*2A and c.2846A>T were also significantly associated with severe fluoropyrimidine-associated toxicity (adjusted RR 2·85, 95% CI 1·75-4·62, p<0·0001; and 3·02, 2·22-4·10, p<0·0001, respectively)." — Meulendijks et al., Lancet Oncol 2015;16(16):1639–1650. PMID:26603945

An ancestry-restricted risk allele [verified this session]:

"The DPYD-Y186C variant was unique to individuals of African ancestry, and DPD activity was 46% lower in carriers as compared with noncarriers (279 ± 35 vs. 514 ± 168 pmol 5-FU min(-1) mg(-1); P = 0.00029). In this study, 26% of the African Americans with reduced DPD activity were carriers of Y186C." — Offer et al., Clin Pharmacol Ther 2013;94(1):158–166. PMID:23588312

Environmental / non-genetic risk factors

For the toxicity arm, the environmental exposure is not incidental — it is the necessary trigger. The exposure is:

  • 5-fluorouracil (intravenous), capecitabine (oral 5-FU prodrug), tegafur, and topical/cutaneous 5-FU. The DPWG guideline explicitly covers all three systemic agents and flags cutaneous exposure [verified this session]: "subjects with a gene activity score of 0 are recommended to avoid systemic and cutaneous 5-fluorouracil or capecitabine" (PMID:31745289).
  • Iatrogenic overdose independent of genotype — infusion pump programming errors, dose miscalculation, accidental or intentional capecitabine ingestion [verified this session]: "Life-threatening 5-FU overdoses occur because of infusion pump errors, dosage miscalculations, and accidental or suicidal ingestion of capecitabine." (Ma et al., PMID:27622829)

Non-genetic confounders of the phenotype test (important for the KB's diagnostics section): chronic kidney disease produces a high false-positive rate on uracil-based screening (PMID:37011867, Clin Chim Acta 2023), and pre-analytical sample handling materially changes measured uracil (PMID:36412238).

For the inborn-error arm, there is a recurring suggestion in the literature that intercurrent stressors unmask or worsen the neurological phenotype. The head-imaging case report is explicit [verified this session]: "Anoxic stress may have contributed to the clinical presentation and brain findings in this case." (Enns et al., PMID:15303009). This is a single-case inference — curate it as PARTIAL evidence at best.

Consanguinity is a recognized contributor to the homozygous metabolic form (the Enns case was born to a consanguineous Pakistani couple), and uniparental isodisomy of chromosome 1 has produced homozygosity for an extremely rare DPYD variant in at least one patient (PMID:30349988) — a mechanism worth flagging because it defeats carrier-based family risk assumptions.

Protective factors

  • Wild-type/normal-metabolizer status (activity score 2) is the reference state — standard dosing is safe.
  • c.29C>G (C29R, DPYD*9A): homozygotes in an African-American cohort showed higher DPD activity [verified this session]: "homozygous carriers of C29R showed 27% higher DPD activity as compared with noncarriers (609 ± 152 and 480 ± 152 pmol 5-FU min(-1) mg(-1), respectively; P = 0.013)." (PMID:23588312). This is the closest thing to a protective allele in the literature; note it is an in-cohort enzymatic observation, not a demonstrated clinical protection.
  • Exogenous uridine is pharmacologically protective against 5-FU cytotoxicity by competitive displacement — the mechanistic basis of the uridine triacetate antidote (Section 12).

Gene–environment interaction

This disease is a gene–environment interaction in its purest textbook form. The genotype is clinically silent until the environmental exposure (fluoropyrimidine) occurs; the exposure is well-tolerated until the genotype is present. The interaction is quantitative and dose-dependent, which is exactly why genotype-guided dose reduction — rather than binary avoidance — works for heterozygotes.

For the KB, this argues for curating the fluoropyrimidine exposure as a linked environmental[] entry with influences_mechanisms → environmental_effect: TRIGGERS on the "Impaired 5-FU catabolism" node, with its own evidence. ECTO grounding for "exposure to 5-fluorouracil" should be searched; if no suitable ECTO term exists, leave exposure_term free-text and record the search in notes: per the environmental-term audit convention.

3. Phenotypes

Arm A — Inborn error of pyrimidine metabolism (biallelic, complete/near-complete deficiency)

The canonical phenotype statement [verified this session]:

"A large phenotypic variability has been observed, with convulsive disorders, motor retardation and mental retardation being the most abundant manifestations. A clear correlation between the genotype and phenotype has not been established." — Van Kuilenburg et al. 1999, PMID:10071185

That last sentence is the single most important curation caveat in this entry. Do not assert genotype–phenotype correlation for the neurological arm. This belongs in the KB as a discussions: entry with kind: KNOWLEDGE_GAP.

Phenotype Suggested HP term Type Onset Course Frequency
Seizure / convulsive disorder HP:0001250 Seizure Clinical sign Neonatal–infantile Episodic, often drug-refractory Most abundant manifestation (PMID:10071185) — curate as FREQUENT only with a countable denominator, else omit frequency:
Intellectual disability HP:0001249 Intellectual disability Clinical Childhood (recognized) Static-to-progressive Among the most abundant (PMID:10071185)
Global developmental delay / motor retardation HP:0001263 Global developmental delay Clinical Infantile Variable Among the most abundant (PMID:10071185)
Microcephaly HP:0000252 Microcephaly Physical Congenital/infantile Progressive or static Reported (OMIM, MONDO description)
Hypertonia HP:0001276 Hypertonia Clinical sign Infantile Variable Reported (MONDO description explicitly says "increased muscle tone (hypertonia)")
Hypotonia HP:0001252 Hypotonia Clinical sign Infantile Variable Reported — note the literature reports both hypo- and hypertonia; curate as two separate phenotypes, not one "abnormal tone"
Autistic behavior HP:0000717 Autism Behavioral Childhood Persistent Reported (OMIM, MONDO)
Nystagmus HP:0000639 Nystagmus Clinical sign Infantile — Reported ocular abnormality
Strabismus HP:0000486 Strabismus Clinical sign Infantile — Reported ocular abnormality
Microphthalmia HP:0000568 Microphthalmia Physical Congenital Static Reported ocular abnormality
Growth delay / failure to thrive HP:0001510 Growth delay Clinical Infantile — Reported (PMID:15303009 case)
Encephalopathy HP:0001298 Encephalopathy Clinical Infantile, may be acute Episodic PMID:12971429 (acute neurological presentation); PMID:15303009
Cerebral atrophy HP:0002059 Cerebral atrophy Imaging Infantile Progressive "diffuse cerebral atrophy" (PMID:15303009)
Abnormal cerebral white matter morphology HP:0002500 Imaging Infantile — "white-matter hyperintensity"; "abnormal T2 prolongation in the cerebral white matter and brainstem" (PMID:15303009)
Uraciluria HP:0012127 Uraciluria Laboratory Congenital Persistent Defining biochemical feature — present in essentially all homozygotes
Elevated urinary thymine HP:6000331 Elevated urinary thymine level Laboratory Congenital Persistent Defining biochemical feature
Abnormal urinary pyrimidine level HP:0033162 Laboratory — — Parent term if you want a single roll-up

All HP IDs and labels above [verified this session against the HPO API].

Additional imaging finding, verbatim [verified this session]:

"Head MRI showed prominent sulci and abnormal T2 prolongation in the cerebral white matter and brainstem. Thus, DPD deficiency may feature prominent brain abnormalities involving the cerebral white matter and brainstem." — Enns et al., J Inherit Metab Dis 2004;27(4):513–522. PMID:15303009

Note the brainstem involvement — that is a distinguishing imaging feature worth curating separately from generic white-matter change.

Critical phenotype caveat. A large fraction of biochemically-deficient individuals are entirely asymptomatic. From the MONDO/GARD description [verified this session]: "DPD deficiency can have a wide range of severity; some individuals may have various neurological problems, while others have no signs and symptoms." The classic illustration is PMID:9323575 — a girl with partial epilepsy and a symptom-free sister carrying the same biochemical deficiency. This is the strongest available argument that DPD deficiency is necessary but not sufficient for the neurological phenotype, and should be curated as an explicit discussions: KNOWLEDGE_GAP with the modifier-gene hypothesis attached.

Arm B — Fluoropyrimidine toxicity phenotypes (heterozygous or biallelic, drug-exposed)

These are drug-reaction phenotypes, and in dismech terms they belong downstream of an exposure-triggered node, not in the constitutive phenotype list of the metabolic disease.

Toxicity Suggested HP term Notes
Neutropenia HP:0001875 Decreased total neutrophil count 18.5% of patients in the Indian prospective cohort (PMID:41259730)
Thrombocytopenia HP:0001873 Thrombocytopenia 15.1% (PMID:41259730)
Pancytopenia HP:0001876 Pancytopenia Severe/complete deficiency
Diarrhea HP:0002014 Diarrhea 12.3% (PMID:41259730); DPWG lists it as a defining toxicity
Oral mucositis / stomatitis HP:0000155 Oral ulcer (nearest HP; HPO has no clean "mucositis" term — verified this session, search returned nothing) DPWG-listed toxicity
Hand-foot syndrome (palmar-plantar erythrodysesthesia) No good HP term — verified this session; nearest terms (HP:0025538 Palmar edema, HP:0025537 Plantar edema) are poor fits DPWG-listed toxicity. Notably not associated with c.1679T>G or HapB3 in the meta-analysis (PMID:26603945) — a real dissociation worth curating
Cardiotoxicity (search UBERON/HP for cardiac-specific term) "rapid reversal of severe acute cardiotoxicity" (PMID:27622829)
Neurotoxicity / leukoencephalopathy HP:0002352 Leukoencephalopathy 5-FU-induced acute leukoencephalopathy is reported even without DPD deficiency (PMID:41610218, 2026) — do not over-attribute
Death (treatment-related mortality) — Deenen: drug-induced death reduced from 10% to 0% with genotype-guided dosing (PMID:26573078)

The toxicity-type dissociation, verbatim [verified this session]:

"Analysis of individual types of toxicity showed consistent associations of c.1679T>G and c.1236G>A/HapB3 with gastrointestinal toxicity (adjusted RR 5·72, 95% CI 1·40-23·33, p=0·015; and 2·04, 1·49-2·78, p<0·0001, respectively) and haematological toxicity (adjusted RR 9·76, 95% CI 3·03-31·48, p=0·00014; and 2·07, 1·17-3·68, p=0·013, respectively), but not with hand-foot syndrome." — PMID:26603945

Quality-of-life impact

No DPD-deficiency-specific EQ-5D / SF-36 / PROMIS instrument or dataset was identified in this session. This is a genuine gap — do not invent QoL figures. Reasonable proxies to note in notes: rather than as evidence:

  • Arm A: QoL burden is dominated by refractory epilepsy and intellectual disability; use the generic literature on infantile epileptic encephalopathy only as context, explicitly not as DPD-specific evidence.
  • Arm B: the closest quantitative proxy is treatment interruption and hospitalization. Deenen reported toxicity-related outcomes; Meulendijks reported "toxicity-related hospitalisation" as an endpoint associated with high pretreatment uracil (PMID:28427087). The Indian cohort reported "Both groups had no problems in completing treatment" after dose reduction (PMID:41259730) — i.e. genotype-guided dosing preserved treatment continuity.

4. Genetic / Molecular Information

Causal gene

DPYD — dihydropyrimidine dehydrogenase, HGNC:3012, 1p21.3, NCBI Gene 1806, Ensembl ENSG00000188641, UniProt Q12882, gene OMIM 612779. [all verified this session via HGNC REST]

Protein: 1,025 aa, functions as a homodimer of ~111 kDa subunits [verified this session via PMID:11179210 — "the homodimeric pig liver enzyme (2x 111 kDa)"].

Pathogenic variants

Population frequencies from gnomAD v4 [computed directly from the gnomAD GraphQL API this session] — global allele frequency and selected ancestry groups:

Variant rsID gnomAD v4 ID (GRCh38) Global AF AFR AMR EAS NFE SAS ASJ FIN
DPYD*2A c.1905+1G>A (splice donor) rs3918290 1-97450058-C-T 0.00456 0.00058 0.0017 0.0 0.0050 0.0029 0.0069 0.0243
c.2846A>T (p.D949V) rs67376798 1-97082391-T-A 0.00328 0.00118 0.0017 0.0 0.0061 0.00062 0.00058 0.00009
DPYD*13 c.1679T>G (p.I560S) rs55886062 1-97515787-A-C 0.00039 0.00019 0.00007 0.0 0.00071 0.0 0.0 0.00019
c.1236G>A / HapB3 (p.E412E, synonymous tag) rs56038477 1-97573863-C-T 0.01310 0.00294 0.00727 0.00058 0.0214 0.0162 0.00547 0.0142
c.557A>G (p.Y186C) rs115232898 1-97699474-T-C 0.00580 0.0202 0.00197 0.0 0.00004 0.0 0.0 0.0
c.1627A>G (DPYD*5, p.I543V) rs1801159 1-97515839-T-C 0.186 0.159 0.237 0.255 0.196 0.096 0.194 0.167
c.2194G>A (DPYD*6, p.V732I) rs1801160 1-97305364-C-T 0.0401 0.0245 0.0459 0.0154 0.0452 0.0930 0.107 0.0221

Read the ancestry columns carefully — this table is the whole health-equity argument in one place. DPYD*2A, c.2846A>T, and c.1679T>G are all absent (AF = 0) in East Asian populations, while c.557A>G is ~500-fold enriched in African-ancestry populations relative to Non-Finnish European (0.0202 vs 0.00004). A four-variant panel derived from European cohorts is nearly uninformative in East Asian patients and misses the dominant risk allele in African-ancestry patients. Note also the striking Finnish enrichment of DPYD*2A (0.0243, ~5× NFE) — a founder-effect signal.

*5 and *6 are high-frequency and are not considered clinically actionable decreased-function alleles by CPIC/DPWG; include them only to document that common DPYD variation ≠ risk variation.

Variant classes and functional consequences

Class Example Consequence
Canonical splice-donor c.1905+1G>A (DPYD*2A) Exon 14 skipping → 165-nt in-frame deletion → no function (activity value 0). Established as the most common lesion in complete deficiency
Missense, no function c.1679T>G (p.I560S, *13) No function (activity value 0)
Missense, decreased function c.2846A>T (p.D949V) Decreased function (activity value 0.5)
Missense, decreased function, ancestry-restricted c.557A>G (p.Y186C) ~46% reduced DPD activity in carriers (PMID:23588312)
Deep intronic → cryptic splice site c.1129-5923C>G (HapB3 causal allele) Creates a cryptic splice donor; 44-bp intron-10 pseudo-exon inserted into mature mRNA
Genomic deletion DPYD exons 21–23; large intragenic deletions Loss of function; invisible to targeted genotyping and to most exome pipelines
Uniparental isodisomy Chr 1 UPD → homozygosity for a rare variant PMID:30349988

The HapB3 mechanism, verbatim [verified this session] — this is the key paper for why the synonymous c.1236G>A "works":

"In one patient a genomic DPYD deletion of exons 21-23 was observed. In five patients a deep intronic mutation c.1129-5923C>G was identified creating a cryptic splice donor site. As a consequence, a 44 bp fragment corresponding to nucleotides c.1129-5967 to c.1129-5924 of intron 10 was inserted in the mature DPD mRNA. The deleterious c.1129-5923C>G mutation proved to be in cis with three intronic polymorphisms (c.483 + 18G>A, c.959-51T>G, c.680 + 139G>A) and the synonymous mutation c.1236G>A of a previously identified haplotype." — van Kuilenburg et al., Hum Genet 2010;128(5):529–538. PMID:20803296

And the conclusion that should drive the KB's diagnostic recommendations:

"Our study demonstrates that a genomic deletion affecting DPYD and a deep intronic mutation affecting pre-mRNA splicing can cause severe 5FU-associated toxicity. We conclude that screening for DPD deficiency should include a search for genomic rearrangements and aberrant splicing." — PMID:20803296

Curation note: c.1236G>A is a linkage tag, not the causal allele. It is a synonymous change; its predictive power comes entirely from being in cis with c.1129-5923C>G. Getting this wrong is a common error in DPD entries — the KB should record functional_impact_category accordingly and put the causal claim on the intronic variant.

Historical mutation spectrum in complete deficiency [verified this session]:

"In this group of patients, 7 different mutations have been identified, including 2 deletions [295-298delTCAT, 1897delC], 1 splice-site mutation [IVS14+1G>A)] and 4 missense mutations (85T>C, 703C>T, 2658G>A, 2983G>T). Analysis of the prevalence of the various mutations among DPD patients has shown that the G-->A point mutation in the invariant splice donor site is by far the most common (52%)..." — PMID:10071185

Somatic vs germline

The disease is germline. Somatic DPYD alteration is not a recognized disease mechanism. However, DPYD expression is biologically relevant in tumors: tumor DPD expression influences 5-FU efficacy, and dihydropyrimidine accumulation has been implicated in EMT (Shaul et al., Cell 2014;158(5) — PMID:25171410, "Dihydropyrimidine accumulation is required for the epithelial-mesenchymal transition"). Flag this as adjacent cancer biology, not as DPD-deficiency pathophysiology.

Modifier genes

Not established. The strongest indirect evidence for modifiers is the asymptomatic-sibling observation (PMID:9323575) and the explicit absence of genotype–phenotype correlation (PMID:10071185). Candidate downstream loci in the same pathway — DPYS (dihydropyrimidinase) and UPB1 (β-ureidopropionase), which cause the two downstream inborn errors — are mechanistically plausible modifiers but I found no evidence establishing them as such. Meulendijks tested TYMS variants alongside DPYD and found [verified this session]: "None of the DPYD variants alone, or TYMS variants alone, were associated with severe toxicity." in that particular 550-patient analysis (PMID:28427087) — a useful negative result.

Epigenetics

DPYD promoter methylation has been proposed as a contributor to reduced tumor/normal-tissue DPD expression. I did not verify primary sources for this in this session — treat as an unverified lead requiring its own literature pass before any evidence item is written.

Chromosomal abnormalities

Whole-chromosome events are not a typical cause, with one important exception: uniparental isodisomy of chromosome 1 producing homozygosity for a rare DPYD variant (PMID:30349988). Large intragenic deletions detectable by MLPA/CMA are established (PMID:20803296; PMID:38528593, a 2024 novel large intragenic deletion case report).

5. Environmental Information

  • Environmental factors: The dominant "environmental" factor is pharmaceutical exposure — 5-FU, capecitabine, tegafur, and cutaneous 5-FU. No occupational, radiation, or pollutant exposure has been established as a cause or modifier.
  • Lifestyle factors: None established. Do not populate this section speculatively.
  • Infectious agents: Not applicable.

Suggested CHEBI grounding [all verified this session via OLS4 except where noted]: - CHEBI:46345 5-fluorouracil - CHEBI:31348 capecitabine - CHEBI:17568 uracil - CHEBI:17821 thymine - CHEBI:15901 5,6-dihydrouracil - CHEBI:27468 5,6-dihydrothymine - CHEBI:16958 beta-alanine - CHEBI:27389 3-aminoisobutyric acid (β-aminoisobutyrate) - CHEBI:16704 uridine - CHEBI:90914 uridine triacetate - tegafur — CHEBI lookup timed out this session; verify before use

6. Mechanism / Pathophysiology

The enzyme

DPD catalyzes the committed, rate-limiting step of pyrimidine catabolism. It is a spectacularly complex redox enzyme. Verbatim [verified this session]:

"Dihydropyrimidine dehydrogenase catalyzes the first step in pyrimidine degradation: the NADPH-dependent reduction of uracil and thymine to the corresponding 5,6-dihydropyrimidines. Its controlled inhibition has become an adjunct target for cancer therapy, since the enzyme is also responsible for the rapid breakdown of the chemotherapeutic drug 5-fluorouracil. The crystal structure of the homodimeric pig liver enzyme (2x 111 kDa) determined at 1.9 A resolution reveals a highly modular subunit organization, consisting of five domains with different folds. Dihydropyrimidine dehydrogenase contains two FAD, two FMN and eight [4Fe-4S] clusters, arranged in two electron transfer chains that pass the dimer interface twice." — Dobritzsch et al., EMBO J 2001;20(4):650–660. PMID:11179210

The structure also directly explains the drug interaction:

"The ternary complex of an inactive mutant of the enzyme with bound NADPH and 5-fluorouracil reveals the architecture of the substrate-binding sites and residues responsible for recognition and binding of the drug." — PMID:11179210

Mechanistically it is a two-site ping-pong enzyme: NADPH reduces FAD at one active site, electrons traverse a ~56 Å [4Fe-4S] wire, and reduced FMN at the second site reduces the pyrimidine ring. This means loss-of-function variants can act by at least four distinct routes — substrate-site disruption, cofactor-binding disruption, electron-wire disruption, or dimer-interface/folding disruption — which is a nice mechanistic reason why so many different missense positions produce deficiency.

The pathway

Pyrimidine catabolism (three steps): 1. DPD (DPYD): uracil → 5,6-dihydrouracil; thymine → 5,6-dihydrothymine. NADPH-dependent. Rate-limiting. 2. Dihydropyrimidinase (DPYS): ring opening → N-carbamyl-β-alanine / N-carbamyl-β-aminoisobutyrate 3. β-ureidopropionase (UPB1): → β-alanine and β-aminoisobutyrate + CO₂ + NH₃

Deficiency at step 1 blocks the entire pathway, causing simultaneous substrate accumulation (uracil, thymine, and 5-hydroxymethyluracil) and product depletion (β-alanine).

Causal chain A — Inborn error → neurological disease

DPYD biallelic LOF variant  [MOLECULAR]
  → absent/near-absent DPD enzyme activity  [MOLECULAR]
    → block of pyrimidine nucleobase catabolism  [MOLECULAR]
      ├→ accumulation of uracil, thymine, 5-hydroxymethyluracil
      │    in urine, plasma, and CSF  [ORGANISM]
      │      → putative neurotoxicity / disturbed pyrimidine homeostasis
      │        in the developing CNS  [TISSUE]
      └→ depletion of β-alanine (an inhibitory neurotransmitter)
   and β-aminoisobutyrate  [MOLECULAR]
     → altered inhibitory neurotransmission  [CELLULAR]
       → neuronal hyperexcitability / seizures  [ORGANISM]
       → white matter and brainstem injury  [TISSUE]
 → developmental delay, intellectual disability,
   microcephaly, autistic behavior  [ORGANISM]

The mechanistic hypothesis behind this chain, stated in the authors' own hedged language [verified this session]:

"An altered beta-alanine, uracil and thymine homeostasis might underlie the various clinical abnormalities encountered in patients with DPD deficiency." — PMID:10071185

Curate that hedge. "Might underlie" is not "does underlie." This chain should be entered as a mechanistic_hypotheses entry with status: EMERGING or ALTERNATIVE, not as a settled canonical cascade. The pathogenesis of the neurological phenotype in DPD deficiency is genuinely unresolved — which is exactly why the same biochemistry produces an epileptic infant and an asymptomatic sibling.

A supporting piece of evidence from an unexpected direction, which strengthens the β-alanine limb [verified this session]:

"Dpyd encodes the rate-limiting enzyme in the metabolic pathway that catabolizes uracil and thymidine to β-alanine, an inhibitory neurotransmitter. Thus, data support β-alanine as a neurotransmitter that promotes sleep in mice." — Keenan et al., Curr Biol 2021;31(23):5238–5248. PMID:34653361 (evidence_source: MODEL_ORGANISM)

This is the best available functional support for β-alanine depletion having a real CNS consequence — but it is a mouse sleep phenotype, not a human seizure phenotype. Flag as HUMAN_MODEL_MISMATCH, not as human evidence.

The white-matter pathogenesis is explicitly unknown [verified this session]:

"The pathogenesis of the white-matter abnormalities is unknown, although environmental factors and altered energy metabolism may be involved." — PMID:15303009

Causal chain B — Pharmacogenetic → fluoropyrimidine toxicity

DPYD deleterious variant (mono- or biallelic)  [MOLECULAR]
  → reduced DPD catalytic activity  [MOLECULAR]
    → [ENVIRONMENTAL TRIGGER: fluoropyrimidine administration]
      → impaired catabolic clearance of 5-FU (>80% of dose
normally cleared by DPD)  [ORGANISM]
→ increased systemic 5-FU exposure / prolonged half-life  [ORGANISM]
  ├→ FdUMP-mediated thymidylate synthase inhibition
  │    → thymineless stress, DNA damage  [CELLULAR]
  └→ FUTP misincorporation into RNA  [MOLECULAR]
       → RNA dysfunction  [CELLULAR]
 → cytotoxicity in rapidly proliferating tissues  [TISSUE]
   ├→ bone marrow  → myelosuppression → neutropenia,
   │                  thrombocytopenia, pancytopenia
   ├→ GI mucosa    → mucositis, severe diarrhea
   ├→ skin/adnexa  → hand-foot syndrome
   ├→ myocardium   → acute cardiotoxicity
   └→ CNS          → acute neurotoxicity / leukoencephalopathy
     → treatment-related morbidity and death  [ORGANISM]

The FUTP-into-RNA limb is confirmed by the antidote's mechanism of action, which is the cleanest available functional evidence for it [verified this session]:

"Uridine triacetate delivers high concentrations of uridine, which competes with toxic 5-FU metabolites." — Ma et al., PMID:27622829

Suggested module conformance

  • myelosuppression — Chain B's marrow limb is a textbook conformer: cytotoxic insult to proliferating HSPCs → marrow suppression → multilineage cytopenias → infection/bleeding/dose-limiting toxicity. The disorder-specific substitution is "excess systemic 5-FU due to impaired DPD catabolism" as the cytotoxic driver. Key conformance target: myelosuppression#Multilineage Peripheral Cytopenias. Strongly recommended.
  • epilepsy_excitation_inhibition_imbalance — Chain A's seizure limb is a plausible conformer via the β-alanine/inhibitory-neurotransmission route (epilepsy_excitation_inhibition_imbalance#Excitation-Inhibition Imbalance). But conform only if you are willing to assert the β-alanine mechanism, which the literature hedges. Given status: EMERGING, I would hold this back or attach it to the hypothesis group rather than declaring bare conforms_to.
  • metabolic_intoxication_decompensation — tempting but I'd advise against. DPD deficiency does not classically produce catabolic-stress-triggered acute metabolic crises with acidosis/hyperammonemia/hypoglycemia. The acute-presentation case (PMID:12971429) is a single report. Forcing this conformance would over-claim.
  • Candidate new module — there is no dismech module for "pharmacogenomic clearance failure" (impaired drug catabolism → supratherapeutic exposure → dose-dependent toxicity). This pattern recurs across DPYD/fluoropyrimidines, TPMT-NUDT15/thiopurines, UGT1A1/irinotecan, and CYP2D6/codeine. If dismech ever wants one, DPD deficiency is the ideal flagship conformer. Worth raising as a proposal rather than shoehorning this entry into an existing module.

GO term suggestions [all verified this session via OLS4]

Term Label Suggested modifier
GO:0017113 dihydropyrimidine dehydrogenase (NADP+) activity LOSS_OF_FUNCTION (molecular_functions)
GO:0006212 uracil catabolic process DECREASED
GO:0006210 thymine catabolic process DECREASED
GO:0006208 pyrimidine nucleobase catabolic process DECREASED
GO:0019483 beta-alanine biosynthetic process DECREASED
GO:0050660 flavin adenine dinucleotide binding (cofactor annotation)
GO:0010181 FMN binding (cofactor annotation)
GO:0051539 4 iron, 4 sulfur cluster binding (cofactor annotation)
GO:0070402 NADPH binding (cofactor annotation)

Note on the modifier choice: for the enzyme activity node, LOSS_OF_FUNCTION is justified for the biallelic complete-deficiency case (the process is genuinely outside normal regulatory constraint — it is abolished). For the heterozygous pharmacogenetic case, DECREASED is the honest quantitative call. If you model these as separate nodes (recommended), the modifiers differ, and that difference is itself informative.

Cell types and tissues

Level Term Role
Primary site of DPD activity CL:0000182 hepatocyte Liver is the dominant site of 5-FU catabolism
Surrogate assay tissue CL:2000001 peripheral blood mononuclear cell PBMC DPD activity is the classical enzymatic phenotyping substrate
Toxicity target Hematopoietic progenitors (bone marrow) Chain B, marrow limb
Toxicity target Intestinal/oral mucosal epithelium Chain B, GI limb
Chain A target Neurons; CNS white matter (oligodendrocytes/myelin) Chain A

All CL IDs [verified this session].

Molecular profiling

  • Metabolomics is the diagnostically decisive modality here: urinary/plasma uracil, thymine, dihydrouracil, and the UH2/U ratio. Repositories: HMDB, Metabolomics Workbench, MetaboLights. See Section 10.
  • Transcriptomics/proteomics/lipidomics/single-cell/spatial: I found no DPD-deficiency-specific datasets in this session. Do not fabricate dataset accessions. If the KB wants datasets: records, run just discover-datasets and apply the standard relevance triage — and be aware that searching the gene symbol DPYD will surface colorectal-cancer 5-FU-response studies, which are about the drug, not the disease. That is textbook Named Entity Confusion reached through dataset search.
  • Functional genomics: DepMap contains DPYD dependency/expression data in the context of 5-FU sensitivity. Relevant to the cancer-pharmacology arm only.

7. Anatomical Structures Affected

Organ level

Organ/system Involvement UBERON
Liver Primary metabolic site of DPD activity — the enzymatic lesion's principal location, though the liver itself is not injured UBERON:0002107 liver
Central nervous system Primary clinical target in Arm A UBERON:0001017 central nervous system
Cerebral white matter T2 hyperintensity, hypoplasia/atrophy (PMID:15303009) UBERON:0002316 white matter
Brainstem T2 prolongation (PMID:15303009) — distinguishing feature search UBERON for brainstem
Bone marrow Secondary — toxicity target in Arm B UBERON:0002371 bone marrow
GI mucosa Secondary — toxicity target in Arm B UBERON:0000344 mucosa (parent; find a GI-specific child)
Skin (palms/soles) Secondary — hand-foot syndrome in Arm B —
Heart Secondary — acute cardiotoxicity in Arm B —
Eye Arm A: microphthalmia, nystagmus, strabismus —

All UBERON IDs listed [verified this session] except brainstem and GI-specific mucosa, which need a lookup.

Subcellular

DPD is a cytosolic enzyme. Suggested GO cellular component: GO:0005829 cytosol [not verified this session — check before use]. This is worth curating explicitly because it distinguishes DPD deficiency from the mitochondrial metabolic disorders it can superficially resemble on a metabolic-workup differential.

Lateralization

Neurological/imaging findings are bilateral and symmetric (diffuse cerebral atrophy, diffuse white matter change). Ocular findings may be unilateral or bilateral. No asymmetric pattern is described.

8. Temporal Development

Onset

Arm A (metabolic): - Onset is typically neonatal to infantile. Presentation at birth is documented (PMID:16151913, "Dihydropyrimidine dehydrogenase deficiency presenting at birth"). The review framing: "Patients may present with a wide range of neurological symptoms during the first years of life." (PMID:15303009) [verified this session] - Pattern: usually insidious (developmental delay recognized over months) but can be acute — PMID:12971429, "Dihydropyrimidine dehydrogenase deficiency and acute neurological presentation." - Suggested OnsetDescriptor: onset_category: NEONATAL and/or INFANTILE, with a second Inheritance-independent note that a substantial fraction never present at all.

Arm B (pharmacogenetic): - Onset is exposure-defined: any age, whenever a fluoropyrimidine is first given. Median onset of toxicity is within the first cycle — Meulendijks' primary endpoint was explicitly "grade ⩾3 toxicity... occurring during the first cycle of treatment" (PMID:28427087), and the Indian prospective study assessed toxicity "after the first cycle of chemotherapy" (PMID:41259730). [verified this session] - Pattern: acute, often hyperacute in complete deficiency.

Progression

Arm A: Variable. Some patients have a static encephalopathy; developmental regression is reported in some. Duration is chronic and lifelong — the enzymatic defect never resolves. There is no established staging system.

Arm B: Rapidly evolving over days. Verbatim [verified this session]:

"Increased susceptibility to 5-fluorouracil (5-FU)/capecitabine can lead to rapidly occurring toxicity caused by impaired clearance, dihydropyrimidine dehydrogenase deficiency, and other genetic variations in the enzymes that metabolize 5-FU." — PMID:27622829

Critical periods — this is the most actionable temporal fact in the entry

There is a hard, quantified intervention window for the toxicity arm. Uridine triacetate must be given early:

"Patients received uridine triacetate as soon as possible (most within the first 96 hours after 5-FU/capecitabine)." — PMID:27622829 [verified this session]

Supporting model data (from the same research program, reported via the manufacturer/preclinical literature): treatment started within 24 h was most effective; starting beyond 96–120 h was far less effective [found via web search summary, primary source not verified this session — verify before writing as evidence].

The other critical period is pre-treatment: the entire clinical value of DPYD testing rests on doing it before the first dose. Both the FDA boxed warning language and the EMA recommendation are framed as pre-treatment requirements.

Recovery kinetics after antidote [verified this session]:

"Among the 141 uridine triacetate-treated overdose patients with a diagnosis of cancer... 53 resumed chemotherapy in < 30 days (median time after 5-FU, 19.6 days), and this indicated a rapid recovery from toxicity." — PMID:27622829

9. Inheritance and Population

Inheritance

  • Autosomal recessive for the metabolic disease (HP:0000007 autosomal recessive inheritance [not verified this session — confirm HP ID]).
  • Codominant / gene-dosage for fluoropyrimidine toxicity risk. This is the practical framing behind the activity score system, and it is why heterozygotes are clinically actionable. Verbatim [verified this session]:

"The DPYD-gene activity score, determined by four DPYD variants, predicts DPD activity and can be used to optimize an individual's starting dose. The gene activity score ranges from 0 (no DPD activity) to 2 (normal DPD activity)." — Lunenburg et al., DPWG guideline, Eur J Hum Genet 2020;28(4):508–517. PMID:31745289

CPIC assigns activity values of 0, 0.5, or 1 per allele (no function / decreased function / normal function); the sum gives the activity score, mapping to normal metabolizer (2), intermediate metabolizer (1–1.5), or poor metabolizer (0–0.5) [per CPIC 2017 guideline, PMID:29152729 — the abstract itself is a scope statement only, so the activity-score detail must be quoted from the guideline body or from the DPWG abstract above, which does state it verbatim].

Curation warning on PMID:29152729: the CPIC 2017 update's PubMed abstract is a purpose statement, not a findings abstract. Verbatim, in full [verified this session]:

"The purpose of this guideline is to provide information for the interpretation of clinical dihydropyrimidine dehydrogenase (DPYD) genotype tests so that the results can be used to guide dosing of fluoropyrimidines (5-fluorouracil and capecitabine). Detailed guidelines for the use of fluoropyrimidines, their clinical pharmacology, as well as analyses of cost-effectiveness are beyond the scope of this document."

Any snippet you attribute to PMID:29152729 must come from that text or from the full article, not from the widely-paraphrased dosing table.

Penetrance and expressivity

  • Arm A: markedly incomplete penetrance and highly variable expressivity. The symptom-free-sibling case (PMID:9323575) is the reference example. No genotype–phenotype correlation established (PMID:10071185).
  • Arm B: penetrance is exposure-conditional and high. In the Indian prospective cohort [verified this session]: "Severe toxicities (grade ≥3)... in mutation carriers (72.7%) as compared with mutation noncarriers (37.0%, P = .03)" (PMID:41259730). In the historical untreated-dose-reduction comparator: "The risk of grade ≥ 3 toxicity was thereby significantly reduced from 73% (95% CI, 58% to 85%) in historical controls (n = 48) to 28% (95% CI, 10% to 53%) by genotype-guided dosing (P < .001)" (PMID:26573078).

Genetic anticipation

Not applicable — no repeat expansion mechanism.

Germline mosaicism

Not reported in the literature I surveyed. Uniparental isodisomy is documented (PMID:30349988) and is the more relevant non-Mendelian mechanism for this gene.

Epidemiology

Partial deficiency (the pharmacogenetically relevant state):

The best-anchored figure comes from the 19,376-patient French population [verified this session]:

"Mean U was 9.9 ± 10.1 ng/mL (median 8.7, range 1.6-856). According to French recommendations, 7.3 % of patients were partially deficient (U 16-150 ng/mL) and 0.02 % completely deficient (U≥150 ng/mL). DPYD variant frequencies were *2A: 0.83 %, *13: 0.17 %, D949V: 1.16 %, *7: 0.05 % (2 homozygous patients with U at 22 and 856 ng/mL)." — Launay et al., Clin Chem Lab Med 2024;62(12):2415–2424. PMID:38896022

The commonly-quoted 3–8% partial-deficiency prevalence in European-ancestry populations is consistent with this [widely cited; the 7.3% figure above is the strongest single primary anchor].

Genotype-based carrier frequency, from prospective screening [verified this session]: - Deenen: "A total of 2,038 patients were prospectively screened for DPYD*2A, of whom 22 (1.1%) were heterozygous polymorphic." (PMID:26573078) - Henricks (four-variant panel): "Of 1103 evaluable patients, 85 (8%) were heterozygous DPYD variant allele carriers, and 1018 (92%) were DPYD wild-type patients." (PMID:30348537) - Indian cohort (four-variant panel): "Of the 146 study participants, 11 (7.5%) had a DPYD mutation. HapB3 (rs56038477) was the most commonly encountered variant (72.7% of patients)..." (PMID:41259730)

Complete deficiency: ~0.02% by uracil phenotype in the French population (PMID:38896022). Estimates of 0.01–0.3% appear in the literature. The number of reported patients with the full neurological phenotype is in the low hundreds worldwide [standard reference, not re-verified].

Suggested Prevalence records for the KB:

prevalence:
- population: France (consecutive pre-treatment oncology patients, 2015-2022)
  measure_type: POINT_PREVALENCE
  prevalence_class: ABOVE_1_IN_1000
  rate_per_100000: 7300.0
  notes: >-
    Partial DPD deficiency by uracil-based phenotyping (plasma U 16-150 ng/mL),
    n=19,376. This is a phenotype-defined partial-deficiency rate, not the
    prevalence of the Mendelian neurological disorder.
- population: France (consecutive pre-treatment oncology patients, 2015-2022)
  measure_type: POINT_PREVALENCE
  prevalence_class: BAND_1_5_PER_10000
  rate_per_100000: 20.0
  notes: >-
    Complete DPD deficiency by uracil-based phenotyping (plasma U >=150 ng/mL),
    n=19,376.

Both with evidence: reference: PMID:38896022 and the verbatim snippet above.

Population demographics

  • Geographic/ancestry distribution: see the gnomAD table in Section 4. Headline points: DPYD*2A shows a strong Finnish founder enrichment (AF 0.0243 vs 0.005 NFE); the three European-derived actionable variants are absent in East Asians; c.557A>G is enriched ~500-fold in African-ancestry populations.
  • Sex ratio: 1:1 expected for an autosomal recessive disorder. No sex bias in the toxicity arm has been established that I could verify. There are literature reports of higher fluoropyrimidine toxicity in women generally, but this is not DPD-specific and should not be attributed here.
  • Age distribution: Arm A is pediatric-onset. Arm B follows the age distribution of cancer patients receiving fluoropyrimidines — i.e. skews strongly older adult.

10. Diagnostics

The central diagnostic tension — curate this carefully

DPD deficiency is diagnosed by two non-equivalent approaches — genotyping and phenotyping — and the field's most important recent finding is that they disagree badly. This is the single most consequential thing to get right in the KB's diagnostics section.

The EMA endorsed both [verified this session]:

"In 2020, the European Medicines Agency (EMA) recommended two methods for pre-treatment DPD deficiency testing in clinical practice: phenotyping using endogenous uracil concentration or genotyping for DPYD risk variant alleles." — de With et al., ESMO Open 2023;8(2):101197. PMID:36989883

The 2024 head-to-head analysis in 19,376 patients found poor concordance [verified this session]:

"Sixty-six% of variant carriers exhibited uracilemia <16 ng/mL, challenging correct identification of DPD deficiency based on U. The sensitivity (% patients with a deficient phenotype among variant carriers) of U threshold at 16 ng/mL was 34 %. The best discriminant marker for identifying variant carriers was UH2/U2. UH2/U2<0.942 (29.7 % of patients) showed enhanced sensitivity (81 %) in identifying deleterious genotypes compared to 16 ng/mL U." — PMID:38896022

"These results reaffirm the poor concordance between DPD phenotyping and genotyping, suggesting that both approaches may be complementary and that targeted DPYD genotyping is not sufficiently reliable to identify all patients with complete deficiency." — PMID:38896022

Set against the earlier prospective evidence that uracil does predict toxicity [verified this session]:

"High pretreatment uracil concentration was strongly predictive of severe, including fatal, fluoropyrimidine-associated toxicity, and is a highly promising phenotypic marker to identify patients at risk of severe fluoropyrimidine-associated toxicity." — Meulendijks et al., Br J Cancer 2017;116(11):1415–1424. PMID:28427087

"High pretreatment uracil concentrations (>16 ng ml-1) were strongly associated with global severe toxicity (OR 5.3, P=0.009), severe gastrointestinal toxicity (OR 33.7, P<0.0001), toxicity-related hospitalisation..." — PMID:28427087

These are not contradictory results, and the KB should say so explicitly. Uracil predicts toxicity well (Meulendijks) but predicts genotype poorly (Launay). They are answering different questions. This deserves a discussions: entry with kind: KNOWLEDGE_GAP and a mechanistic_hypotheses framing, not a single flattened "uracil is the test" claim.

Laboratory tests

Test Analyte Notes
Urine pyrimidine analysis Uracil, thymine, 5-hydroxymethyluracil Diagnostic for Arm A. This is the test that finds the inborn error. PMID:15303009 recommends it explicitly for unexplained white-matter/brainstem MRI findings
Plasma uracil (U) Uracil Pre-treatment phenotyping. French/EMA thresholds: U ≥16 ng/mL = partial deficiency; U ≥150 ng/mL = complete deficiency (PMID:38896022)
UH2/U or UH2/U² ratio Dihydrouracil/uracil UH2/U² < 0.942 outperforms the U threshold for identifying variant carriers (sensitivity 81% vs 34%) (PMID:38896022)
PBMC DPD enzyme activity pmol 5-FU·min⁻¹·mg⁻¹ Reference/research assay; normal ~514 ± 168 in European-Americans (PMID:23588312)
5-FU degradation rate (5-FUDR) assay ex vivo Alternative functional phenotyping
Therapeutic drug monitoring of 5-FU 5-FU AUC Complements phenotyping; can catch under-exposure after dose reduction

Pre-analytical warning worth curating: plasma uracil is exquisitely sensitive to sample handling (PMID:36412238, "Plasma Uracil as a DPD Phenotyping Test: Pre-Analytical Handling Matters!"), and chronic kidney disease produces a high false-positive rate (PMID:37011867). Both are real, actionable caveats.

LOINC codes for uracil/dihydrouracil should be looked up before curating reference_ranges — I did not verify them this session. If you do add reference_ranges, the French thresholds map naturally onto interpretation_bands:

interpretation_bands:
- name: Normal DPD activity
  upper_bound: 16.0
  unit: ng/mL
  abnormal_flag: NORMAL
- name: Partial DPD deficiency
  lower_bound: 16.0
  upper_bound: 150.0
  unit: ng/mL
  abnormal_flag: HIGH
  severity: MODERATE
- name: Complete DPD deficiency
  lower_bound: 150.0
  unit: ng/mL
  abnormal_flag: CRITICAL_HIGH
  severity: SEVERE

with evidence: PMID:38896022 and the verbatim threshold sentence.

Imaging

Brain MRI in Arm A: prominent sulci, diffuse cerebral atrophy, T2 hyperintensity in cerebral white matter and brainstem (PMID:15303009; also PMID:25565930, abnormal MRI in two Malaysian siblings). Imaging is supportive, not diagnostic.

Genetic testing

  • Targeted genotyping (the guideline standard): the four CPIC/DPWG variants — DPYD*2A (rs3918290), c.1679T>G (rs55886062), c.2846A>T (rs67376798), c.1236G>A/HapB3 (rs56038477). This is what every current guideline panel means by "DPYD genotyping."
  • Extended panels: c.557A>G (rs115232898) is included by Mayo Clinic and several US commercial laboratories and by the NHS North West GLH since September 2025 [web-search-sourced; verify the NHS detail before using as evidence]. CPIC recommends 50% starting-dose reduction for heterozygous carriers.
  • Full-gene sequencing: catches rare and private variants. PMID:42510779 (2026) specifically addresses outcomes of patients carrying unusual DPYD variants found by protocol implementation.
  • CNV/MLPA/CMA: required to catch the large intragenic deletions documented in PMID:20803296 and PMID:38528593. Targeted genotyping misses these entirely.
  • Deep intronic coverage: c.1129-5923C>G must be either directly assayed or captured via its c.1236G>A tag. Exome sequencing will miss it.
  • WGS: theoretically catches all of the above; not standard of care for this indication.
  • Karyotype/FISH/mtDNA/repeat expansion: not applicable.

The authors' own conclusion on testing scope [verified this session]:

"We conclude that screening for DPD deficiency should include a search for genomic rearrangements and aberrant splicing." — PMID:20803296

Also worth curating (2024): "DPYD genotype should be extended to rare variants: report on two cases of phenotype / genotype discrepancy" — Cancer Chemother Pharmacol 2024, doi:10.1007/s00280-024-04738-5. [title verified via web search; PMID not confirmed this session]

Clinical criteria and differential diagnosis

There is no consensus clinical diagnostic criteria set (DSM/ICD-style) for DPD deficiency — diagnosis is biochemical and/or molecular.

Differential diagnosis:

Condition Distinguishing feature
Dihydropyrimidinase deficiency (DPYS) Elevated urinary dihydrouracil and dihydrothymine (the DPD products) rather than uracil/thymine — the pattern is diagnostic
β-ureidopropionase deficiency (UPB1) Elevated N-carbamyl-β-alanine / N-carbamyl-β-aminoisobutyrate
Other early infantile epileptic encephalopathies Normal urine pyrimidines
Mitochondrial encephalopathies Lactate/pyruvate abnormalities; DPD is cytosolic
Non-DPD causes of fluoropyrimidine toxicity 5-FU leukoencephalopathy has been reported without DPD deficiency (PMID:41610218, 2026); also TYMS variants, drug interactions, renal impairment
Iatrogenic 5-FU overdose Normal genotype/phenotype; history of pump/dosing error

The DPYS and UPB1 disorders are the key differentials because they are the two downstream steps of the same pathway — a single urine pyrimidine profile distinguishes all three. For the KB, this is also a natural Grouping candidate: "Disorders of Pyrimidine Degradation," with grouping_basis: [SHARED_PATHWAY] and three DISEASE members.

Screening

  • Newborn screening: DPD deficiency is not on standard newborn screening panels. Urine pyrimidine analysis is a targeted metabolic workup, not a population screen.
  • Pre-treatment pharmacogenetic screening: now effectively universal in guideline terms — see Section 13.
  • Cascade screening: relatives of an identified carrier should be offered DPYD testing before any fluoropyrimidine exposure. This is high-yield and cheap, and is under-practiced.

11. Outcome / Prognosis

Arm A

No survival statistics specific to DPD deficiency were identified in this session. Do not manufacture a life-expectancy figure. Prognosis is dominated by the severity of the epilepsy and encephalopathy; a substantial fraction of biochemically-deficient people are asymptomatic and have normal life expectancy. Aspiration pneumonia (HP:0011951) and complications of severe neurodisability are the plausible causes of premature death in severely affected patients, but I found no cohort quantifying this.

The most important prognostic statement is the negative one: genotype does not predict outcome in Arm A (PMID:10071185).

Arm B — quantified, and this is where the good numbers live

Mortality without intervention:

"In the historical cohort, 21 of 25 patients (84%) died." — PMID:27622829 (untreated 5-FU overdose, supportive care only) [verified this session]

Mortality with antidote:

"A total of 137 of 142 overdose patients (96%) treated with uridine triacetate survived and had a rapid reversal of severe acute cardiotoxicity and neurotoxicity; in addition, mucositis and leukopenia were prevented, or the patients recovered from them." — PMID:27622829 [verified this session]

Mortality with genotype-guided prevention:

"The risk of grade ≥ 3 toxicity was thereby significantly reduced from 73% (95% CI, 58% to 85%) in historical controls (n = 48) to 28% (95% CI, 10% to 53%) by genotype-guided dosing (P < .001); drug-induced death was reduced from 10% to 0%." — PMID:26573078 [verified this session]

Residual risk even with guideline dosing — the honest caveat:

"Overall, fluoropyrimidine-related severe toxicity was higher in DPYD variant carriers (33 [39%] of 85 patients) than in wild-type patients (231 [23%] of 1018 patients; p=0·0013)." — PMID:30348537 [verified this session]

That is a genotype-guided-dosed carrier population still experiencing more severe toxicity than wild-type. Dose reduction narrows the gap; it does not close it. The 2025 commentary title says it plainly: "Reducing Fluorouracil Doses in Patients With Partial Dihydropyrimidine Dehydrogenase Deficiency Is a Treatment Safety Strategy, Not a Panacea of Precision Dosing" (Hertz & Venook, JCO Precis Oncol 2025;9:e2500440, PMID:40638877). Curate this as a discussions: entry — the "genotyping solves it" framing is an overstatement the field is actively pushing back on.

Efficacy preservation — the other key prognostic question

The standing worry about dose reduction is undertreatment. The evidence says efficacy is preserved:

"Adequate treatment of genotype-guided dosing was further demonstrated by a similar incidence of grade ≥ 3 toxicity compared with wild-type patients receiving the standard dose (23%; P = .64)" — PMID:26573078 [verified this session]

"Individuals with DPYD mutations experience increased toxicity and dose adjustments; however, treatment efficacy was not affected." — PMID:41259730 [verified this session]

"Evidence demonstrates that dose individualization based on guidance from the Clinical Pharmacogenetics Implementation Consortium reduces toxicity risk while maintaining treatment effectiveness and potentially reducing overall costs." — Kratz et al., ASCO Educ Book 2026;46(3):e521184. PMID:42048619 [verified this session]

Prognostic factors and biomarkers

  • Activity score (0–2) is the primary prognostic stratifier for the toxicity arm.
  • Pretreatment plasma uracil >16 ng/mL predicts severe and fatal toxicity (OR 5.3 global, OR 33.7 GI) (PMID:28427087).
  • Number of deleterious alleles — biallelic/poor metabolizers have qualitatively higher risk than heterozygotes.
  • Time to antidote administration is the dominant prognostic factor once toxicity is underway.

12. Treatment

Arm A — Inborn error

There is no disease-modifying therapy. Management is symptomatic and supportive:

Treatment treatment_term therapeutic_modality Notes
Anticonvulsant therapy NCIT:C64172 Anticonvulsant Therapy SMALL_MOLECULE Seizure control; no DPD-specific agent established
Supportive care NCIT:C15747 Supportive Care OTHER Developmental support, feeding
Nutritional support NCIT:C15433 Nutritional Support depends on the intervention — do not blind-tag as BEHAVIORAL per the CLAUDE.md caveat For failure to thrive
Genetic counseling NCIT:C15240 Genetic Counseling OTHER Recurrence risk; cascade testing
Lifelong fluoropyrimidine avoidance see note below — The single most important intervention for every affected individual, symptomatic or not

All NCIT IDs [verified this session via OLS4]. Note that NCIT:C64172's reachability from NCIT:C25218 (Clinical Intervention or Procedure) should be checked with just validate-terms before commit.

β-alanine supplementation is a theoretically attractive intervention given the pathway logic. I found no clinical evidence for it in this session. Do not curate it as a treatment.

Arm B — Prevention of fluoropyrimidine toxicity

Genotype-guided dosing (DPWG, verbatim) [verified this session]:

"For patients initiating 5-fluorouracil or capecitabine: subjects with a gene activity score of 0 are recommended to avoid systemic and cutaneous 5-fluorouracil or capecitabine; subjects with a gene activity score of 1 or 1.5 are recommended to initiate therapy with 50% the standard dose of 5-fluorouracil or capecitabine. For subjects initiating tegafur: subjects with a gene activity score of 0, 1 or 1.5 are recommended to avoid tegafur. Subjects with a gene activity score of 2 (reference) should receive a standard dose." — PMID:31745289

Refinement from the prospective study [verified this session]:

"For DPYD*2A and c.1679T>G carriers, a 50% initial dose reduction was adequate. For c.1236G>A and c.2846A>T carriers, a larger dose reduction of 50% (instead of 25%) requires investigation." — PMID:30348537

Note the asymmetry: the 2018 study used 25% reductions for c.1236G>A and c.2846A>T and found them insufficient; current CPIC guidance uses a uniform 50% reduction for intermediate metabolizers. If you curate the dosing algorithm, cite the current guideline for the recommendation and PMID:30348537 for the empirical basis.

Rescue therapy — uridine triacetate:

treatments:
- name: Uridine Triacetate
  description: >-
    Emergency antidote for 5-fluorouracil or capecitabine overdose or early-onset
    severe toxicity, including toxicity due to DPD deficiency. Delivers high
    concentrations of uridine, which competes with cytotoxic 5-FU metabolites for
    incorporation into RNA. FDA-approved 2015. Must be given early - most treated
    patients received it within 96 hours of fluoropyrimidine exposure.
  therapeutic_modality: SMALL_MOLECULE
  treatment_term:
    preferred_term: Pharmacotherapy
    term:
      id: NCIT:C15986
      label: Pharmacotherapy
    therapeutic_agent:
    - preferred_term: uridine triacetate
      term:
id: CHEBI:90914
label: uridine triacetate
  evidence:
  - reference: PMID:27622829
    supports: SUPPORT
    evidence_source: HUMAN_CLINICAL
    snippet: "In these studies, uridine triacetate was a safe and effective lifesaving antidote for capecitabine and 5-FU overexposure, and it facilitated the rapid resumption of chemotherapy."
    explanation: Two open-label clinical studies establishing uridine triacetate as an effective antidote for fluoropyrimidine overexposure.

NCIT:C2379 (Uridine Triacetate) is available as an alternative therapeutic_agent grounding [verified this session] — but note the memory-recorded caveat that NCIT drug terms frequently fail therapeutic_agent enum validation in this repo; prefer the CHEBI term.

Adverse events of the antidote itself [verified this session]: "Adverse reactions in patients receiving uridine triacetate included vomiting (8.1%), nausea (4.6%), and diarrhea (3.5%)." (PMID:27622829)

Supportive management of established toxicity: G-CSF for neutropenia, antibiotics for febrile neutropenia, fluid/electrolyte support for diarrhea, mucositis care. None is DPD-specific.

Pharmacogenomics

This disease is a pharmacogenomics entry. Resources: PharmGKB (DPYD–fluoropyrimidine pathway and clinical annotations), CPIC (Level A gene–drug pair), DPWG (clinical implication score "essential"), FDA Table of Pharmacogenomic Biomarkers in Drug Labeling.

DPWG's own framing of the strength of recommendation [verified this session]:

"Based on the DPWG clinical implication score, DPYD genotyping is considered 'essential', therefore directing DPYD testing prior to initiating fluoropyrimidines." — PMID:31745289

Advanced therapeutics

  • Gene therapy / gene editing / RNA therapies: none in development for DPD deficiency that I identified. This is not a currently-addressed target.
  • Cell therapy: not applicable.
  • DPD inhibitors as a therapeutic strategy (the inverse concept — deliberately inhibiting DPD to boost 5-FU exposure, e.g. eniluracil, or the DPD-inhibitory fluoropyrimidine S-1) is real and interesting context, and PMID:11179210 alludes to it ("Its controlled inhibition has become an adjunct target for cancer therapy"). It is not a treatment for DPD deficiency — file under notes: or a discussions: entry, never under treatments:.

Experimental / trials

NCT02324452 — the Alpe2U prospective DPYD genotype-guided dosing study underlying PMID:30348537 [verified this session: "This trial is registered with ClinicalTrials.gov, number NCT02324452, and is complete."]. Suitable for a clinical_trials: record:

clinical_trials:
- name: NCT02324452
  phase: NOT_APPLICABLE   # verify against the registry record before commit
  status: COMPLETED
  description: >-
    Prospective multicentre safety analysis of DPYD genotype-guided dose
    individualisation of fluoropyrimidine therapy in 17 Dutch hospitals.

Fetch the registry record with just fetch-reference NCT02324452 before writing any snippet, and confirm the phase enum against the actual record — I did not verify it.

13. Prevention

Primary prevention

For Arm A there is no primary prevention beyond reproductive options. For Arm B, primary prevention is the entire story: test before you dose.

The current regulatory position [verified this session]:

"Recent regulatory and guideline changes have established pretreatment DPYD genotyping as a critical strategy to prevent severe fluoropyrimidine toxicity. Following earlier European leadership by the European Medicines Agency, the US Food and Drug Administration added boxed warnings to capecitabine and 5-fluorouracil labels recommending genetic testing before therapy. Concurrent updates from the National Comprehensive Cancer Network and ASCO align US with European practice supporting universal testing." — Kratz et al., ASCO Educ Book 2026;46(3):e521184. PMID:42048619

This is the strongest, most recent, and most citable statement of the current standard. [The specific dates — capecitabine boxed warning October 2025, 5-FU label February 2026 — came from a secondary web source and are NOT verified. Do not write those dates into the KB without checking the FDA labels directly.]

Implementation status in Europe [verified this session]:

"Following publication of the EMA recommendations, 87% and 75% of the countries reported an increase in the amount of genotype and phenotype testing, respectively." — PMID:36989883

"The EMA recommendations have supported the implementation of DPD deficiency testing in Europe. Key factors for successful implementation were test reimbursement and clear clinical guidelines." — PMID:36989883

Secondary prevention

  • Early recognition of first-cycle toxicity → immediate fluoropyrimidine discontinuation → uridine triacetate within the 96-hour window.
  • Therapeutic drug monitoring of 5-FU to catch both over- and under-exposure after dose adjustment (PMC7700344, "Association of 5-FU Therapeutic Drug Monitoring to DPD Phenotype Assessment May Reduce 5-FU Under-Exposure" [title only; verify]).

Tertiary prevention

  • Dose titration upward from a reduced starting dose based on observed tolerance — the Deenen protocol used "an initial dose reduction of ≥ 50% followed by dose titration based on tolerance" with a median achieved dose-intensity of 48% (range 17%–91%) (PMID:26573078) [verified this session].
  • Permanent avoidance flagging in the EHR/allergy list for identified poor metabolizers.

Genetic screening and counseling

  • Cascade testing of first-degree relatives of identified carriers, before any oncology exposure.
  • Carrier/prenatal/PGD: technically available for families with a severely affected child. No professional guideline mandates it; the wide phenotypic variability and high asymptomatic rate make counseling genuinely difficult and this should be stated as such.
  • NCIT:C15240 Genetic Counseling; NCIT:C15709 Genetic Testing; NCIT:C68762 Pharmacogenomic Test [all verified this session; check NCIT:C68762 reachability from NCIT:C25218 — it is a "test" concept and may not validate as a clinical intervention].

Public health and equity

The equity dimension is a first-class part of prevention for this disease and should be curated, not left as an aside. From the 2026 ASCO Educational Book [verified this session]:

"Despite growing adoption, implementation challenges persist, including workflow integration, clinician education, and equitable access... Barriers are more pronounced in resource-constrained settings, where limited infrastructure, reimbursement uncertainty, and insufficient pharmacogenomic education hinder implementation. Regional initiatives illustrate education-focused, context-adapted strategies to expand testing and address population-specific variant knowledge gaps." — PMID:42048619

And on patient advocacy as a driver:

"Patient advocacy, particularly efforts led by Advocates for Universal DPD/DPYD Testing, has accelerated policy change, increased clinician awareness, and highlighted ethical implications of preventable harm." — PMID:42048619

Supporting systematic review: "DPYD genetic polymorphisms in non-European patients with severe fluoropyrimidine-related toxicity: a systematic review" — PMID:38886557 (2024) [title/PMID from search; abstract not verified this session].

The gnomAD table in Section 4 is the quantitative backbone of this argument and should be cited alongside the narrative.

No vaccination, immunization, sanitation, vector-control, or environmental-remediation dimension

Not applicable. Leave these empty rather than padding them.

14. Other Species / Natural Disease

  • Taxonomy: NCBITaxon:9606 Homo sapiens. Model species: NCBITaxon:10090 Mus musculus, NCBITaxon:9823 Sus scrofa (the pig liver enzyme is the structural reference).
  • Breed (VBO): Not applicable — no breed-associated natural disease known.
  • Orthologs: mouse Dpyd (chromosome 3); pig DPYD; rat Dpyd. Bacterial and fungal orthologs exist and have been characterized structurally (e.g. E. coli DPD, PMID:34097066).
  • Natural disease in other species: None identified. [verified this session] An OMIA advanced search on gene symbol DPYD returned "No phene records found." This is a genuine, searched-and-absent result and should be recorded as such in notes: rather than left blank — the absence is informative.
  • Veterinary relevance: No established natural veterinary disease. 5-FU is notoriously and severely toxic to cats and dogs (fatal neurotoxicity from accidental exposure to topical 5-FU creams), but that is a species-level sensitivity, not DPD deficiency, and should not be conflated with it in the KB.
  • Comparative biology / evolutionary conservation: DPD is deeply conserved — the pyrimidine catabolic pathway and the enzyme's remarkable cofactor architecture (2× FAD, 2× FMN, 8× [4Fe-4S]) are shared from bacteria to mammals (PMID:11179210; PMID:34097066). The pig enzyme's structure is the reference model for the human enzyme, which is itself a strong statement about conservation.
  • Zoonotic potential / cross-species transmission: Not applicable.

15. Model Organisms

Mouse

The best-characterized model is a Dpyd knockout — and its published phenotype is not what you would expect. Verbatim [verified this session]:

"Validation studies were performed using activity monitoring and EEG/EMG recording in Collaborative Cross mouse strains with and without the PWK/PhJ haplotype at this location, as well as EEG and EMG recording of sleep and wake in Dpyd knockout mice and wild-type littermate controls. Mice lacking Dpyd had 78.4 min less sleep during the lights-off period than wild-type mice (p = 0.007; Cohen's d = -0.94). There was no difference in other measured behaviors in knockout mice, including assays evaluating cognitive-, social-, and affective-disorder-related behaviors." — Keenan et al., Curr Biol 2021;31(23):5238–5248. PMID:34653361

That last sentence is the load-bearing one for the KB. Dpyd knockout mice do not recapitulate the human neurological phenotype — no cognitive, social, or affective abnormality was detected. They have a sleep phenotype instead. This is a textbook FAILS_TO_RECAPITULATE / HUMAN_MODEL_MISMATCH situation and should be curated as such:

animal_models:
- name: Dpyd knockout mouse
  species: Mouse
  genotype: Dpyd knockout (homozygous null)
  publication: PMID:34653361
  modeled_mechanisms:
  - target: Beta-Alanine Depletion and Altered Inhibitory Neurotransmission
    relationship: PARTIALLY_RECAPITULATES
    fidelity: LOW
    description: >-
      Loss of Dpyd blocks the pathway generating beta-alanine and produces a
      measurable CNS phenotype (reduced sleep during the active period),
      supporting a real neurological consequence of the enzymatic block.
    limitations: >-
      The observed phenotype is reduced sleep, not seizures or developmental
      impairment. No difference was detected in cognitive-, social-, or
      affective-disorder-related behavioral assays, so the model does not
      reproduce the epilepsy, intellectual disability, or autistic behavior
      that define the severe human presentation.
    readouts:
    - name: Total sleep during lights-off period (EEG/EMG)
      target: Beta-Alanine Depletion and Altered Inhibitory Neurotransmission
      direction: DECREASED
      interpretation: >-
78.4 minutes less sleep than wild-type littermates (Cohen's d = -0.94),
the only behavioral domain in which knockouts differed.
      evidence:
      - reference: PMID:34653361
supports: SUPPORT
evidence_source: MODEL_ORGANISM
snippet: "Mice lacking Dpyd had 78.4 min less sleep during the lights-off period than wild-type mice (p = 0.007; Cohen's d = -0.94)."
explanation: Direct EEG/EMG measurement in Dpyd knockout versus wild-type littermate controls.
  - target: Seizure Generation
    relationship: FAILS_TO_RECAPITULATE
    fidelity: LOW
    limitations: >-
      No seizure or epilepsy phenotype was reported, and the study explicitly
      found no differences in other measured behavioral domains.
    evidence:
    - reference: PMID:34653361
      supports: SUPPORT
      evidence_source: MODEL_ORGANISM
      snippet: "There was no difference in other measured behaviors in knockout mice, including assays evaluating cognitive-, social-, and affective-disorder-related behaviors."
      explanation: Negative behavioral result establishing that the knockout does not reproduce the human neurobehavioral phenotype.

Note that FAILS_TO_RECAPITULATE links require both limitations and evidence per test_failure_to_recapitulate_links_are_substantiated — both are supplied above.

The associated HUMAN_MODEL_MISMATCH discussion writes itself: evidence exists in a model (β-alanine depletion produces a real CNS phenotype in mice) but its translational validity to the human epileptic encephalopathy is the open question.

Diversity Outbred / Collaborative Cross mice are also relevant as a natural-variation model [verified this session]: a linkage peak for total sleep on chromosome 3 (LOD 7.14), with reduced Dpyd expression associated with the PWK/PhJ allele — an eQTL-driven partial-loss model rather than a null.

Resources: MGI (mouse Dpyd), IMPC — note that IMPC lists OMIM:274270 in its disease-model index [seen in search results this session; the specific IMPC phenotype data were not reviewed], IMSR, Jackson Laboratory (DO/CC strains).

Other models

  • Pig liver enzyme — the structural workhorse. The 1.9 Å crystal structure with bound NADPH and 5-FU (PMID:11179210) is the basis for essentially all structure–function interpretation of human DPYD missense variants. evidence_source: IN_VITRO.
  • Mammalian cell expression systems — recombinant expression of DPYD variants with enzyme-activity measurement relative to wild type is the standard functional-validation route for VUS. See PMID:23328581 ("Phenotypic profiling of DPYD variations relevant to 5-fluorouracil sensitivity using real-time cellular analysis and in vitro measurement of enzyme activity"). evidence_source: IN_VITRO.
  • Patient PBMCs — ex vivo DPD activity measurement in primary human cells; arguably the highest-fidelity "model" available since it uses the patient's own genotype (PMID:23588312). evidence_source: HUMAN_CLINICAL or IN_VITRO depending on framing — the Offer study measured circulating mononuclear-cell enzyme activity in healthy volunteers, so IN_VITRO is defensible for the assay claim while the population-frequency claim is HUMAN_CLINICAL. Split the evidence items so each carries a single evidence_source.
  • Zebrafish / Drosophila / C. elegans: no established DPD-deficiency models identified.
  • UPase knockout mouse (PMID:21954436) is not a DPD model — it targets uridine phosphorylase — but it is mechanistically adjacent, demonstrating that uridine availability protects normal tissue from 5-FU toxicity. Useful supporting context for the uridine-triacetate mechanism; do not miscatalog it as a Dpyd model.

Applications and limitations

What the models are good for: enzyme structure–function; variant functional classification; uridine-rescue pharmacology; β-alanine neurobiology.

What they cannot currently do: reproduce the human neurological phenotype. There is, as far as I can tell, no animal model of DPD-deficiency epileptic encephalopathy. That is the single largest gap in this disease's experimental toolkit, and it is worth curating as a discussions: entry with kind: KNOWLEDGE_GAP and a proposed_experiments block.

Curation Summary and Warnings

Things to get right:

  1. Two arms, two chains. Do not blend the inborn-error pathophysiology with the pharmacogenetic toxicity pathophysiology into one cascade. Section 6 gives them separately for a reason.
  2. No genotype–phenotype correlation in Arm A — PMID:10071185 says so explicitly. Curate this as a KNOWLEDGE_GAP, not as an omission.
  3. The β-alanine mechanism is a hypothesis ("might underlie"), not settled fact. mechanistic_hypotheses with status: EMERGING.
  4. c.1236G>A is a tag, not the cause — the causal allele is the deep intronic c.1129-5923C>G.
  5. Phenotyping and genotyping disagree, and the disagreement is a real finding (PMID:38896022), not a measurement failure. Uracil predicts toxicity well, genotype poorly.
  6. The mouse knockout fails to model the human disease. FAILS_TO_RECAPITULATE + HUMAN_MODEL_MISMATCH.
  7. Dose reduction reduces but does not eliminate excess risk (39% vs 23% severe toxicity even with genotype-guided dosing, PMID:30348537). Don't over-claim.
  8. CPIC PMID:29152729's abstract is a scope statement. Any dosing-table snippet attributed to it must come from the article body, not from paraphrase.

Named Entity Confusion risk for this disease: LOW-MODERATE. DPYD is unambiguous and the MONDO causal gene is clear, so just preflight-dr should PASS cleanly on any DR report. The real confusion risk here is not entity-level but arm-level — a DR tool may silently produce a report entirely about 5-FU pharmacogenomics with nothing on the inborn error, or vice versa. Check that any DR report covers both before curating from it. The related-disease confusion risk (DPYS dihydropyrimidinase deficiency vs DPYD dihydropyrimidine dehydrogenase deficiency — one letter apart, adjacent steps of the same pathway) is real and worth an explicit check.

Verified PMIDs used in this report: 10071185, 11179210, 15303009, 20803296, 23588312, 25171410, 26573078, 26603945, 27622829, 28427087, 29152729, 30348537, 31745289, 34653361, 36621118, 36989883, 38896022, 40638877, 41259730, 42048619. Additional PMIDs cited by title/summary only and requiring just fetch-reference verification before use: 9323575, 12971429, 16151913, 21954436, 23328581, 25565930, 30349988, 34097066, 36412238, 37011867, 38528593, 38886557, 41610218, 42510779.

Sources consulted online: - CPIC — DPYD guidelines - DPWG guideline for DPYD and fluoropyrimidines (PMC7080718) - OMIM #274270 — Dihydropyrimidine Dehydrogenase Deficiency - GARD — Dihydropyrimidine dehydrogenase deficiency - Launay et al. 2024, GPCO-RNPGx uracil vs genotype analysis - Kratz et al. 2026, ASCO Educational Book — implementing DPYD testing - Ho et al. 2025, guide for implementing DPYD genotyping (Clin Pharmacol Ther) - Offer et al. 2013 — DPYD Y186C in African ancestry (PMC3821392) - Dobritzsch et al. 2001 — DPD crystal structure (EMBO J) - Ma et al. 2016 — uridine triacetate emergency use - gnomAD (v4 allele frequencies, queried directly) - OLS4 / EMBL-EBI (MONDO, GO, CHEBI, CL, UBERON, NCIT term verification) - HPO API, Jackson Laboratory (HP term verification) - HGNC REST (DPYD gene record) - OMIA — DPYD search (no animal phene records)

Reference Validation

Checked with linkml-reference-validator 0.2.1.

Outcome Count
References checked 40
Resolved 38
Unresolved (possible confabulation) 2
Unverifiable 0
Quoted claims checked 3
Quoted claims found in source 3

Unresolved references

These identifiers did not resolve to a record and may be fabricated. A lookup that failed for transport reasons is indistinguishable from one that failed because the record does not exist, so spot-check before acting on them:

  • PMID:31745289 (4 mentions) - Identifier did not resolve to a record
  • PMID:9323575 (3 mentions) - Identifier did not resolve to a record