Should we genotype every patient before 5-FU or capecitabine?

CPIC strongly recommends it, and the EMA endorses it across the EU. In the US, the FDA label warns about DPD deficiency but does not mandate testing. Major academic and community oncology centers increasingly implement universal testing because the cost-effectiveness case is strong and the consequences of severe toxicity (including death) are severe.

What is the DPYD activity score?

An additive score across both alleles. Wild-type = 1 per allele (total 2). No-function allele = 0. Decreased-function allele = 0.5. Total scores: 2 = normal; 1-1.5 = intermediate; 0-0.5 = poor.

Does DPYD genotyping cover all DPD deficiency?

No. The standard panel (DPYD *2A, *13, c.2846A>T, HapB3) covers most common actionable variants but not all DPD deficiency. Functional testing (uracil load test, plasma uracil) complements genotyping when clinical suspicion is high.

What if a patient develops severe toxicity despite normal DPYD genotype?

Consider rare variants outside the standard panel, drug interactions, hepatic dysfunction, or other PGx factors. Uridine triacetate is the FDA-approved antidote for life-threatening fluoropyrimidine toxicity within 96 hours.

Pharmacogenomics · CPIC Level A · Fluoropyrimidines

DPYD genotype-guided dosing for 5-fluorouracil and capecitabine

CPIC-aligned reference. Educational use only — not a substitute for clinical pharmacology consultation.

TL;DR

DPYD-deficient patients face severe (sometimes fatal) toxicity on standard 5-FU or capecitabine. CPIC Level A. Pre-treatment genotyping for DPYD *2A (c.1905+1G>A), *13 (c.1679T>G), c.2846A>T, and HapB3 (c.1129-5923C>G / c.1236G>A) is the standard. Intermediate metabolizers (DPYD activity score 1-1.5) get 50% dose reduction; poor metabolizers (score 0-0.5) avoid fluoropyrimidines or use a markedly reduced dose with monitoring.

Variants tested

DPYD *2A (rs3918290, c.1905+1G>A): splice variant, no functional protein. Activity = 0.
DPYD *13 (rs55886062, c.1679T>G / I560S): activity = 0.
c.2846A>T (rs67376798, D949V): activity ~0.5 (intermediate).
HapB3 (rs56038477 / rs75017182): activity ~0.5 (intermediate).
Approximately 3-7% of patients carry at least one actionable DPYD variant.

Phenotype mapping

Normal metabolizer (activity score 2): standard dose.
Intermediate metabolizer (activity score 1 or 1.5): start at 50% standard dose with titration based on toxicity and clinical response. Heterozygous for one no-function or one decreased-function allele.
Poor metabolizer (activity score 0 or 0.5): avoid 5-FU and capecitabine; if no alternative, use markedly reduced dose with intensive monitoring. Homozygous or compound heterozygous for no/decreased-function alleles.

Dosing recommendation

CPIC strongly recommends DPYD pre-treatment genotyping before initiating 5-fluorouracil-based or capecitabine-based therapy. Genotype-informed dose adjustment substantially reduces grade 3-4 toxicity (mucositis, diarrhea, neutropenia, hand-foot syndrome) and treatment-related death.

Intermediate metabolizer: 50% starting dose; titrate by toxicity.
Poor metabolizer: avoid; if no alternative, use ≤25% dose with very close monitoring.
Uridine triacetate (Vistogard) is the antidote for life-threatening 5-FU or capecitabine overdose / severe toxicity within 96 hours of the dose.

Evidence

The European Medicines Agency (EMA) endorsed pre-treatment DPYD testing in 2020. The FDA updated 5-FU and capecitabine labels to include DPYD warnings. Multiple prospective studies (Henricks 2018, Deenen 2016) demonstrated that genotype-guided dosing reduces severe toxicity from ~73% (standard dose) to ~28% in DPYD heterozygotes, without compromising efficacy. The cost-effectiveness case is strong (one prevented severe-toxicity event covers many genotyping tests).

Caveats

DPYD genotyping does not detect every DPD-deficient patient. Rare variants outside the standard panel and post-translational regulation contribute to phenotypic variability. Functional uracil-loading or plasma uracil testing can complement genotyping in high-risk patients. Genotyping is necessary but not sufficient — clinical judgment and toxicity monitoring still drive ongoing dose titration.

Frequently asked questions

Should we genotype every patient before 5-FU or capecitabine?: CPIC strongly recommends it, and the EMA endorses it across the EU. In the US, the FDA label warns about DPD deficiency but does not mandate testing. Major academic and community oncology centers increasingly implement universal testing because the cost-effectiveness case is strong and the consequences of severe toxicity (including death) are severe.
What is the DPYD activity score?: An additive score across both alleles. Wild-type = 1 per allele (total 2). No-function allele = 0. Decreased-function allele = 0.5. Total scores: 2 = normal; 1-1.5 = intermediate; 0-0.5 = poor.
Does DPYD genotyping cover all DPD deficiency?: No. The standard panel (DPYD *2A, *13, c.2846A>T, HapB3) covers most common actionable variants but not all DPD deficiency. Functional testing (uracil load test, plasma uracil) complements genotyping when clinical suspicion is high.
What if a patient develops severe toxicity despite normal DPYD genotype?: Consider rare variants outside the standard panel, drug interactions, hepatic dysfunction, or other PGx factors. Uridine triacetate is the FDA-approved antidote for life-threatening fluoropyrimidine toxicity within 96 hours.