Many studies have estimated the prevalence of chronic kidney disease (CKD) but results have varied due partly to the type of equation used to estimate GFR, type of subjects, and ethnicity. This review aimed to estimate the prevalence of CKD Stage III, accounting for these factors. 403 studies were identified from Medline using the PubMed search engine, of which 34 studies were eligible. Data were independently extracted by two reviewers, and heterogeneity was assessed using metaregression. The pooled prevalence was estimated using a random effects model. In the general population, the prevalences of CKD Stage III using MDRD equation were 3.6% (95% CI: 2.5, 4.8), 10.7% (95% CI: 4.5 – 16.9%), and 16.3% (95% CI: 2.1 – 30.5%) for age groups ≤ 50, 50 – 60 and > 60 years. The prevalence was about double using the Cockcroft-Gault equations, i.e. 7.5% (95% CI: 6.9 – 8.2%) and 34.9 (95% CI: 25.9 – 44.8%) in age ≤ 50 and > 50 years, respectively. The prevalence was similar in Caucasians and Asians aged ≤ 60, i.e. 9.9 versus 9.3%. The prevalence was also higher in the diabetic population than in the general population, i.e. 18.2% versus 10.6%. The pooled prevalence of CKD in the general population varied according to age groups. The prevalence is similar in Caucasians and Asians within age 60 years or younger but other age groups need more studies in order to pool. Individual patient meta-analysis would be appropriate to resolve the causes of heterogeneity.