Thirteen studies were included (n=2,263). All included studies were retrospective. Seven studies had a quality score of 5 or less.
Early detection of the recurrence was associated with significantly better survival than late detection (HR 1.68, 95% CI 1.48 to 1.91; 13 studies). This was not associated with significant heterogeneity. The chance of benefit of early detection was significantly higher in studies that presented data regarding loco-regional recurrence and contra-lateral recurrence separately from distant metastases (HR 2.55, 95% CI 1.76 to 3.70). Survival was better in patients without symptoms compared with those with symptoms (HR 1.56, 95% CI 1.36 to 1.79) and in studies where mammography was used compared with those in which recurrences were assessed clinically (HR 2.44, 95% CI 1.78 to 3.35).
Hazard ratios were not significantly different for early and late detection of starting point for computation of follow-up time, type of follow-up scheme, quality scores and type of outcome.
The authors stated that if all breast cancer recurrences were detected earlier there would be an absolute reduction in mortality of 17% 28%.
There was no evidence of publication bias.