Page 94 - Personalised medicine of fluoropyrimidines using DPYD pharmacogenetics Carin Lunenburg

Chapter 4
However, the increase was non-significant in studies including 340% of patients with toxicity.
The results were similar if only prospective studies, only higher quality studies or only studies including 3 200 patients were analysed. In prospective studies, the risk also increased as the incidence of grade III-V toxicity decreased in the study.
The risk was also increased when only studies investigating 5-FU-based therapy or 5-FU monotherapy were analysed.
Increased risk of grade III-V haematological toxicity (OR = 15.77; 95% CI: 6.36-39.06) (S)
Increased risk of grade III-V diarrhoea (OR = 5.54; 95% CI: 2.31-13.29) (S)
Increased risk of grade III-V mucositis (OR = 7.48; 95% CI: 3.03-18.47) (S)
*2A had a sensitivity of 5.2% (95% CI: 3.0-8.9) and a specificity of 99.2% (95% CI: 98.8-99.4) for predicting grade III-V toxicity (S)
The sensitivity was 9.0% for studies that showed less than 40% grade III-V toxicity (95% CI: 5.7-13.9) (S). There was study heterogeneity in the overall group, but not in the group with less than 40% toxicity.
*2A had a sensitivity of 13% (95% CI: 6.6-24.1) for predicting grade III-V haematological toxicity (S) *2A had a sensitivity of 5.6% (95% CI: 3.2-9.7) for predicting grade III-V diarrhoea (S)
*2A had a sensitivity of 11.5% (95% CI: 6.2-20.5) for predicting grade III-V mucositis (S)
c.2846A>T versus (no c.2846A>T):
Increased risk of grade III-V toxicity (OR = 8.18; 95% CI: 2.65-25.25; increase in the percentage of patients with grade III-V toxicity from 34% to 71%) (S) Exclusion of each of the studies from the meta- analysis did not lead to substantially different results (OR = 6.20 - 12.88 (S)).
The risk was increased in studies in which the percentage of patients with grade III-V toxicity was less than 40% (OR = 16.59; 95% CI: 5.06-54.43) (S). However, the increase was non-significant in studies including 340% of patients with toxicity.
The results were similar if higher only quality studies or only studies including 3 200 patients were analysed.
The risk was also increased when only prospective studies were analysed (OR = 18.14; 95% CI: 6.26- 52.58) (S) or only studies investigating 5-FU-based therapy (OR = 21.38; 95% CI: 6.71-68.15) (S).
There was moderate study heterogeneity in the overall group, but not in the low or high toxicity subgroups, among prospective studies or among those investigating 5-fluorouracil-based therapy. There may have been publication bias.
92
table continues