Evidences for the personalized use of nonsteroidal anti-inflammatory drugs (NSAIDs) in colorectal cancer (CRC) prevention and treatment that include consideration of prostaglandin E2 levels are necessary. This study was designed as a case-control study including 60 CRC patients and 120 cancer-free controls. A sensitive empirical method, precolumn derivatization HPLC, was used to determine plasma PGE2 levels. The TaqMan SNP Genotyping Assay was used for the genotyping of prostaglandin-endoperoxide synthase 2 (PTGS2) polymorphisms. Multivariate logistic regression analysis suggested that 1 log10(PGE2) increase would result in a 3.64-fold increase in the risk of CRC. Moreover, subjects with log10(PGE2) level in the 75th percentile had a significantly higher risk of CRC than those with log10(PGE2) levels in the 25th percentile [odds ratio (OR), 3.50; 95% confidence interval (CI), 1.35-9.05]. This association was more evident after adjustment for history of NSAIDs use (OR, 3.85; 95% CI, 1.46-10.16). Preliminarily, 260.02 and 414.95 pg/ml might be proposed as the preventive and warning cutoff values of plasma PGE2 for CRC. The preferred NSAIDs dose for patients with the AG+GG (rs689466) and CC+CT (rs5275) genotypes should be higher than that of patients carrying AA or TT genotypes, despite the presence of equal plasma PGE2 levels. We show for the first time that the plasma PGE2 level is associated with the risk of CRC. We provide a preliminary suggestion for NSAIDs doses adjustment according to PTGS2 genotypes after consideration of plasma PGE2 levels.