Evaluating the role of ENOSF1 and TYMS variants as predictors in fluoropyrimidine-related toxicities: An IPD meta-analysis.

To assess the proposed associations of the c.742-227G>A (rs2612091) polymorphism within the Enolase Superfamily Member 1 gene (ENOSF1) and two variants in the adjacent Thymidylate Synthase gene (TYMS): the 5'VNTR 28bp-repeat (rs45445694) and 3'UTR 6bp-indel (rs11280056) with severe toxicity in fluoropyrimidine-treated cancer patients, we performed an individual patient data meta-analysis. Only studies investigating all three-abovementioned variants with fluoropyrimidine-related toxicities were considered for meta-analysis. Associations were tested individually for each study using multivariate regression. Meta-analysis was performed using a random-effects model. One-stage multivariate regressions including tests for independent SNP effects were applied to investigate individual effects of the variants. Multivariate haplotype regression analyses were performed on a pooled dataset to test multi-SNP effects. Of four studies including 2'067 patients, 1'912 were eligible for meta-analysis. All variants were exclusively associated with severe hand-foot-syndrome (HFS) (TYMS 2R: OR = 1.50, p = 0.0002; TYMS 6bp-ins: OR = 1.42 p = 0.0036; ENOSF1 c.742-227G: OR = 1.64 p < 0.0001, per allele). We observed independent effects for ENOSF1 c.742-227G>A and the TYMS 28bp-repeat: each toxicity-associated allele increased the risk for severe HFS (OR = 1.32 per allele, p < 0.0001). Patients homozygous for both variants were at the 3-fold higher risk for severe HFS compared to wild-type patients. Our results confirm an essential role for ENOSF1 c.742-227G and TYMS 2R-alleles in the development of fluoropyrimidine-related HFS. This suggests an important function of these genes in the development of severe HFS. Furthermore, these variants might help stratify patients in studies investigating measures of HFS prevention.