Skip to content

Technology Evaluation
Center (TEC)

Pharmacogenetic Testing to Predict Serious Toxicity from 5-Fluorouracil (5-FU) for Patients Administered 5-FU-Based Chemotherapy for Cancer


Background:  Severe toxicity occurs in about 30% of patients treated with 5-fluorouracil (FU) -based chemotherapy regimens.  5-Fluorouracil has a narrow therapeutic window and the drug concentration required for tumor response is within the general range where toxicity may occur.  Inherited genetic variability in key enzymes involved in the 5-FU metabolic pathway may be related to the variable patient experience of toxicity.  Dihydropyrimidine dehydrogenase (DPD) is a saturable and rate-limiting enzyme in the 5-FU catabolic pathway.  In the anabolic pathway, 5-FU is converted into several active metabolites, one of which inhibits the action of thymidylate synthase (TS), a key enzyme in normal and tumor DNA synthesis.  Genetic polymorphisms in the genes coding for DPD and TS may result in enzyme products with different activity levels, resulting in 5-FU excess, the accumulation of 5-FU anabolic products, and severe toxicity.

Objective:  This Assessment will evaluate the evidence for using genetic polymorphisms in the genes coding for DPD and TS as predictors of toxicity at standard doses, and as indicators for reducing 5-FU dose to avoid toxicity without reducing efficacy.

Search strategy:  MEDLINE® was searched (via PubMed) using the following two strategies:

·          "Fluorouracil"[MeSH] AND ("Thymidylate Synthase"[MeSH] OR ("Dihydrouracil Dehydrogenase (NAD+)"[MeSH] OR "Dihydrouracil Dehydrogenase (NADP)"[MeSH])), limited to human subjects, clinical trials, and the English language. ·          "Fluorouracil"[MeSH] AND ("Thymidylate Synthase"[MeSH] OR ("Dihydrouracil Dehydrogenase (NAD+)"[MeSH] OR "Dihydrouracil Dehydrogenase (NADP)"[MeSH])) AND gene* AND tox*, limited to human subjects and the English language.

In addition, bibliographies from recent review articles and clinical studies were hand-searched for relevant studies.  For information on assay technical performance, the grey literature was also searched, in particular websites of commercial laboratories offering DPD and/or TS genotyping.

Selection criteria:  For clinical utility (effect of using DPD and TS pharmacogenetic tests to determine 5-FU dosing on patient toxicity outcomes), controlled trials that compared standard fixed-dose 5-FU regimens versus 5-FU dose adjusted, if needed, based on the results of pharmacogenetic testing and predicted 5-FU toxicity results.

For clinical validity (i.e., ability of DPD and TS pharmacogenetic tests to predict 5-FU patient toxicity), prospective cohort studies that reported or contained the data to calculate sensitivity, specificity, and predictive values of the genetic variants were preferred.  Retrospective cohort studies that provided relevant information were also considered.  Studies that evaluated highly selected populations (e.g., healthy controls versus selected patients with 5-FU grade 3 to 4 toxicity reactions or case series of patients with 5-FU severe toxicity) were not included.

For analytic validity (technical performance of DPD and TS pharmacogenetic tests), any information from the published or grey literature that provided information on assay methods and technical performance was included.

Main results:  Ten cohort studies, mostly prospective, that evaluated the clinical validity of DPD pharmacogenetic testing were included in this Assessment.  Sample sizes ranged from 21 to 683.  Patients had primarily colorectal cancer, other gastrointestinal cancer, head and neck, or breast cancer and were treated with 5-FU monotherapy or combination chemotherapy.  Most studies tested for only the most common variant of DPYD, the gene that codes for DPD (DPYD *2A) or a small number of known genetic variants, or scanned DPYD exon 14 for variants (includes the *2A variant).  Two studies scanned the entire DPYD coding region and flanking intronic regions.  Because of the variety of approaches, studies report finding different variants; in addition, variants found may also be influenced by geography and ethnic mix.  These differences affect the results of the sensitivity, specificity, and predictive value results, where data reported allowed these calculations to be made.  Nevertheless, some general observations can be made.  First, sensitivity is very low in all cases, because the majority of patients with severe toxicity have no detectable DPYD genetic variants.  Specificity is quite high in most studies, because DPYD genetic variants are rare and patients without variants or severe toxicity are numerous in comparison to patients with DPYD variants but no severe toxicity (i.e., false positives).

Seven cohort studies examined the clinical validity of one or more of the 3 types of genetic variants found in the gene that codes for TS (TYMS); 3 of these studies examined 2 or 3 of these variants in combination.  None of these three studies reported significant results for TYMS genetic variants, alone or in combination, as predictors of 5-FU toxicity.  Results for each of the types of TYMS genetic variants considered individually are generally poor in terms of sensitivity, specificity, and predictive values.

One study, considered one of the best studies to address pharmacogenetic testing for 5-FU toxicity due primarily to its large patient enrollment and to its focus on 5-FU monotherapy, thus avoiding confounding toxicity from other chemotherapy drugs, addressed not only DPYD and TYMS genetic variants, but several other possible factors influencing toxicity in a multivariable regression analysis.  Results indicated the strongest influences were from an interaction between DPYD variants and sex (such that DPYD*2A variants were significant only in male patients), and from leucovorin administration (yes/no); there were also significant influences from method of 5-FU administration, and from variants of the TYMS and MTHFR genes (MTHFR codes for 5,10-methylene-tetrahydrofolate reductase; another enzyme in the 5-FU metabolic pathway).  From these results the authors constructed a nomogram for estimating 5-FU toxicity risk; however, use of this nomogram was not tested prospectively in the reported study.

No published studies addressed clinical utility, that is, whether reducing the starting dose of 5-FU when serious toxicity is predicted by pretreatment pharmacogenetic testing reduces episodes of serious toxicity without reducing treatment response compared to standard dosing and dose adjustment according to symptoms.

No published studies addressed analytic validity, or the technical performance of commercially available DPD and TS pharmacogenetic assays.  One commercial laboratory provides a limited summary of technical specifications for their assays on their website.  False negatives and false positives are estimated to be less than 1%.

Author's conclusions and comments:  It has been tempting to postulate alterations in activity of key enzymes such as DPD and TS in the 5-FU metabolic pathway as the causal basis for 5-FU toxicity, and specific genetic variants of the genes coding for those enzymes as the starting points in the causal chain.  Indeed, patients with homozygous, DPD-inactivating mutations in the DPD gene uniformly experience severe, potentially fatal toxicity reactions when administered standard 5-FU doses.  However, homozygosity for DPD-inactivating mutations was rarely reported in the studies included in this Assessment and heterozygous DPYD variants were observed in relatively small proportions of patients with grade 3 to 4 toxicity.  Moreover, not all patients with DPYD variants experience toxicity (even when variants assessed are limited to those with prior associations with toxicity), making DPYD variants poor predictors.  The clinical validity evidence for each of the 3 types of TS gene variants is similarly poor in terms of the ability of TYMS variants to predict which patients are likely to experience severe 5-FU toxicity.

In summary, testing for genetic variants of the genes coding for DPD and TS enzymes has poor predictive value for 5-FU toxicity and no studies have shown that it is useful in directing 5-FU dose alterations to reduce toxicity without adversely impacting tumor response. 

Based on the available evidence, the Blue Cross and Blue Shield Association Medical Advisory Panel made the following judgments about whether the use of pharmacogenetic testing to predict serious toxicity from 5-FU for patients administered 5-FU-based chemotherapy for cancer meets the Blue Cross and Blue Shield Association's Technology Evaluation Center (TEC) criteria.

1.  The technology must have final approval from the appropriate government regulatory bodies.

There are no assay kits approved by the U.S. Food and Drug Administration (FDA) for genetic testing for DPYD or TYMS genotypes, nor are any kits being actively manufactured and marketed for distribution.  Clinical laboratories may develop and validate tests in-house ("home-brew") and market them as a laboratory service; such tests must meet the general regulatory standards of the Clinical Laboratory Improvement Act (CLIA).  The laboratory offering the service must be licensed by CLIA for high-complexity testing.  While the FDA has technical authority to regulate home-brew tests, there is currently no active oversight.

2.  The scientific evidence must permit conclusions concerning the effect of the technology on health outcomes.

The evidence for this Assessment consists of cohort studies that address the clinical validity of DPD and TS pharmacogenetic testing.  In general, both assays have poor ability to identify patients likely to experience severe 5-FU toxicity.  No studies address clinical utility.  The evidence is insufficient to permit conclusions regarding the effect of DPD and TS pharmacogenetic testing on benefits (reduced toxicity) and harms (poorer response to treatment).

3.  The technology must improve the net health outcome; and

4.  The technology must be as beneficial as any established alternatives.

There is insufficient evidence to permit conclusions regarding the use of pharmacogenetic testing to predict serious toxicity from 5-FU for patients administered 5-FU-based chemotherapy for cancer.

5.  The improvement must be attainable outside the investigational settings.

Whether or not the use of use of pharmacogenetic testing to predict serious toxicity from 5-FU for patients administered 5-FU-based chemotherapy for cancer improves health outcomes has not been demonstrated in the investigational setting.

For the above reasons, the use of pharmacogenetic testing to predict serious toxicity from 5-FU for patients administered 5-FU-based chemotherapy for cancer does not meet the TEC criteria.


NOTICE OF PURPOSE:  TEC Assessments are scientific opinions, provided solely for informational purposes. TEC Assessments should not be construed to suggest that the Blue Cross Blue Shield Association, Kaiser Permanente Medical Care Program or the TEC Program recommends, advocates, requires, encourages, or discourages any particular treatment, procedure, or service; any particular course of treatment, procedure, or service; or the payment or non-payment of the technology or technologies evaluated.