Background
In 2020, the European Medicines Agency recommended testing patients for dihydropyrimidine dehydrogenase (DPD) deficiency before systemic treatment with fluoropyrimidines (FP). DPD activity testing identifies patients at elevated risk of severe FP-related toxicity (FP-TOX). The two most used methods for DPD testing are DPYD genotyping and DPD phenotyping (plasma uracil concentration). The primary objective of this study was to compare the overall frequency of overall grade ≥3 FP-TOX before and after the implementation of DPYD genotyping.

Patients and methods
Two hundred thirty Danish, primarily gastrointestinal cancer patients, were DPYD-genotyped before their first dose of FP, and blood was sampled for post hoc assessment of P-uracil. The initial dose was reduced for variant carriers. Grade ≥3 FP-TOX was registered after the first three treatment cycles of FP. The frequency of toxicity was compared to a historical cohort of 492 patients with post hoc determined DPYD genotype from a biobank.

Results
The frequency of overall grade ≥3 FP-TOX was 27% in the DPYD genotype-guided group compared to 24% in the historical cohort. In DPYD variant carriers, DPYD genotyping reduced the frequency of FP-related hospitalization from 19% to 0%. In the control group, 4.8% of DPYD variant carriers died due to FP-TOX compared to 0% in the group receiving DPYD genotype-guided dosing of FP. In the intervention group, wild-type patients with uracil ≥16 ng/ml had a higher frequency of FP-TOX than wild-type patients with uracil <16 ng/ml (55% versus 28%).

Conclusions
We found no population-level benefit of DPYD genotyping when comparing the risk of grade ≥3 FP-TOX before and after clinical implementation. We observed no deaths or FP-related hospitalizations in patients whose FP treatment was guided by a variant DPYD genotype. The use of DPD phenotyping may add valuable information in DPYD wild-type patients.