We performed a systematic review and meta-analysis to explore the optimum management strategy for women with atypical squamous cells of undetermined significance (ASCUS/borderline) or low-grade squamous intra-epithelial lesions (LSIL/mild dyskaryosis) cytological abnormalities at primary screening in the absence of HPV DNA test. We searched MEDLINE, EMBASE and CENTRAL and included randomised controlled trials comparing immediate colposcopy to cytological surveillance in women with ASCUS/LSIL. The outcomes of interest were occurrence of different histological grades of cervical intraepithelial neoplasia (CIN) and default rates during follow-up. Pooled risk ratios (RR) and 95% confidence intervals (CI) were calculated using a random-effect model and with inverse variance weighting. Interstudy heterogeneity was assessed using I(2) statistics. Six RCTs were included. Immediate colposcopy significantly increased detection of unimportant abnormalities as opposed to repeat cytology (koilocytosis: 32 vs. 21%, RR: 1.49, 95% CI = 1.17-1.90); CIN1: 21 vs. 8%, RR: 2.58, 95% CI = 1.69-3.94). Although immediate colposcopy detected CIN2, CIN2+, and CIN3+ earlier than cytology, the differences were no longer observed at 24 months (CIN3+: 10.3 vs.11.9%, RR: 1.02, 95% CI = 0.53-1.97), with significant interstudy heterogeneity (p < 0.001, I(2)  = 93%). Default risk was significantly higher for repeat cytology (6 months: 6.3 vs. 13.3%, RR: 3.85, 95% CI = 1.27-11.63; 12 months: 6.3 vs. 14.8%, RR: 6.39, 95% CI = 1.24-32.95; 24 months: 0.9 vs. 16.1%, RR: 19.1, 95% CI = 9.02-40.4). Detection of CIN2+ for cytological surveillance over two years is similar to that of immediate colposcopy, although patients may default. Colposcopy may be first choice when good compliance is not assured, but may increase detection of insignificant lesions. This emphasizes the need for a reflex triage test to distinguish women who need diagnostic work-up from those who can return to routine recall.