Objectives: After removal of dentine caries lesions, cavity lining has been advocated. Non-clinical data support this approach, but clinical data are sparse and ambiguous. We aimed at evaluating the benefits and harms of cavity lining using meta-analysis and Trial Sequential Analysis. Data: We included randomized clinical trials comparing restorations without versus with cavity lining for treating primary caries lesions. Only trials reporting failure (defined as need to re-retreat) after >= 1 year follow-up were included. Trial selection, data extraction, and risk of bias assessment were conducted independently by two reviewers. We conducted random-effects intention-to-treat and per-protocol meta-analyses, and Trial Sequential Analysis to control for random errors. Sources: Electronic databases (PubMed, Embase, CENTRAL) were systematically screened, and hand searches and cross-referencing performed. Study selection: From 128 studies, three randomized trials (89/130 patients or teeth), all treating primary teeth, were included. The trials had high risk of bias. All trials compared no lining versus calcium hydroxide lining after selective caries removal followed by adhesive restoration. Follow-up was 36 to 53 months. Restoring the cavity without lining did not significantly affect the risk of failure (intention-to-treat relative risk (RR) (95% confidence interval) 0.71 (0.49-1.04), per-protocol RR 0.52 (0.24-1.10). According to Trial Sequential Analysis, no firm evidence was reached. The quality of evidence was very low. Conclusions: Strong recommendations for using cavity liners are unsubstantiated, but firm evidence for omitting lining is also unavailable. Our findings apply only to primary teeth and calcium hydroxide liner. Clinical significance: Whilst lining is frequently performed in dental practice, very few randomized clinical trials investigated this issue. The three trials included in this review treated deciduous teeth and did not find lining with calcium hydroxide beneficial. Lining is not supported by sufficient clinical evidence. (C) 2015 Elsevier Ltd. All rights reserved.