Current guidelines recommend pre-therapeutic UGT1A1 genotyping to guide irinotecan dosing, but the usefulness of this approach remains to be clarified. In 247 patients with advanced gastrointestinal cancers undergoing irinotecan-based chemotherapy, we prospectively performed UGT1A1*28 genotyping and we analyzed the incidence of severe neutropenia according to genotype-guided dose reductions. Overall, 28 (11.3%) and 92 (37.2%) patients were homozygous or heterozygous UGT1A1*28 carriers, respectively. Grade ≥ 3 neutropenia was reported in 39% of homozygous patients receiving an upfront dose reduction of irinotecan (median 40%, range 22–58%), in 20% of heterozygous or wild-type patients receiving full dose (OR_{vs*28/*28 genotype} = 0.38; 95% CI: 0.14–1.03; p = 0.058), and in 15.3% of those receiving a reduced dose for clinical reasons (OR _{vs*28/*28 genotype} = 0.28, 95% IC: 0.12–0.67; p = 0.004). Occurrence of severe neutropenia was inversely associated with dose reduction in UGT1A1*28 homozygous carriers (OR_{x10 unit} = 0.62, 95% CI: 0.27–1.40, p = 0.249) and UGT1A1 heterozygous or wild-type patients (OR_{x10 unit} = 0.87, 95% CI: 0.59–1.28, p = 0.478). Incidence of severe neutropenia was related to irinotecan doses and UGT1A1 polymorphisms. Upfront irinotecan dose reductions do not reduce the burden of grade ≥ 3 neutropenia in UGT1A1*28 homozygous carriers.