The Food and Drug Administration and Pfizer changed the package insert for irinotecan to include a patient's UGT1A1*28 genotype as a risk factor for severe neutropenia on the basis of the findings of four pharmacogenetic studies, which found that irinotecan-treated patients who were homozygous for the UGT1A1*28 allele had a greater risk of hematologic toxic effects than patients who had one or two copies of the wild-type allele (UGT1A1*1). Findings of subsequent irinotecan pharmacogenetic studies have been inconsistent. In a meta-analysis, we reviewed data presented in nine studies that included a total of 10 sets of patients (for a total of 821 patients) and assessed the association of irinotecan dose with the risk of irinotecan-related hematologic toxicities (grade III-IV) for patients with a UGT1A1*28/*28 genotype. The risk of toxicity was higher among patients with a UGT1A1*28/*28 genotype than among those with a UGT1A1*1/*1 or UGT1A1*1/*28 genotype at both medium (odds ratio [OR] = 3.22, 95% confidence interval [CI] = 1.52 to 6.81; P = .008) and high (OR = 27.8, 95% CI = 4.0 to 195; P = .005) doses of irinotecan. However, risk was similar at lower doses (OR = 1.80, 95% CI = 0.37 to 8.84; P = .41). Low doses of irinotecan (100-125 mg/m2) are in the commonly used therapeutic range. The risk of experiencing irinotecan-induced hematologic toxicity for patients with a UGT1A1*28/*28 genotype thus appears to be a function of the dose of irinotecan administered.