Background: Consensus is lacking regarding an effective surveillance strategy for the detection of colorectal cancer (CRC) recurrence at a curative stage. This study aims to evaluate the effect of carcinoembryonic antigen (CEA) measurement alone vs. CEA as part of an intensive surveillance strategy for the detection of curative CRC recurrence. Methods: A systematic review was performed using MEDLINE and EMBASE from January 2000 to April 2016 to identify randomized controlled trials (RCTs), cohort studies, quasi-experimental design studies, and case control studies including stage II/III CRC patients with curative resection and surveillance with repeated CEA measurements. Data screening, abstraction, and risk of bias assessment was independently performed by two reviewers. Risks of recurrence amenable to surgical resection, local and distant recurrence, disease-specific mortality and overall mortality were pooled using random effects meta-analysis. Results: Two RCTs and 9 cohort studies were included. A meta-analysis of 861 patients in RCTs demonstrated a non-significant reduction in the detection of curative recurrence using CEA measurement alone vs. intensive surveillance strategies with CEA (RR 0.76; 95% CI 0.41-1.41). There were no significant differences observed in the risk of locoregional or distant recurrence. Patients with CEA alone had a non-significant higher risk of disease specific and overall mortality ([RR 1.19; 95% CI 0.89-1.60], [RR 1.21; 95% CI 0.84-1.76], respectively). The heterogeneity of the cohort studies inhibited a quantitative meta-analysis. Among 2783 patients in 9 cohort studies using intensive surveillance strategies with CEA measurement, the rate of curative recurrence ranged from 21.3% to 56.8%. Conclusions: Limited data exists to guide the optimal surveillance strategy post-curative CRC resection. This study highlights the potential benefit of intensive surveillance strategies with CEA measurement as compared to CEA measurement alone in detecting curative CRC recurrence. However, further research to determine optimal CRC surveillance strategies is warranted.