A meta-analysis in Caucasians was conducted to investigate the possible association of uridine diphosphate glucuronosyltransferase (UGT) 1A1 gene polymorphisms with irinotecan (IRI)-induced neutropenia and diarrhoea in colorectal cancer (CRC). We searched PubMed and Embase until May 2012 to identify eligible studies, extracted data, assessed methodological quality, and performed statistical analysis using REVMAN 5.1 and R software. Subgroups meta-analyses were performed in groups representing different IRI combination regimens and IRI doses. Sixteen trials were included. UGT1A1*28/*28 genotype was associated with more than fourfold (odds ratio (OR)=4.79, 95% confidence intervals (CI): 3.28-7.01; P<0.00001) and threefold (OR=3.44, 95% CI: 2.45-4.82; P<0.00001) increases in the risk of neutropenia when compared with wild type and with at least one UGT1A1*1 allele, respectively. UGT1A1*1/*28 genotype had an OR of 1.90 (95% CI: 1.44-2.51; P<0.00001) for an increased risk of neutropenia. A twofold increase in risk of diarrhoea was associated with UGT1A1*28/*28 genotype (OR=1.84, 95% CI: 1.24-2.72; P=0.002). In subgroup meta-analysis, the higher incidence of diarrhoea in UGT1A1*28/*28 patients was limited to studies where when IRI was given at higher doses (OR=2.37, 95% CI: 1.39-4.04; P=0.002) or combined with 5-fluorouracil (FU or analogue) (OR=1.78, 95% CI: 1.16-2.75; P=0.009). Genotyping of UGT1A1*28 polymorphism before treatment for CRC can tailor IRI therapy and reduce the IRI-related toxicities. IRI-combined 5-FU (or analogue) and a high-dose IRI therapy enhance IRI-induced diarrhoea among patients bearing the UGT1A1*28 allele. Although the toxicity relationships were much stronger with the UGT1A1*28 homozygous variant, associations were also found with the UGT1A1*28 heterozygous variant.