Context: In adults, permissive hypercapnia reduces mortality and ventilation duration. However, in preterm infants, findings from past research regarding the efficacy and safety of permissive hypercapnia are controversial. Objective: To evaluate the efficacy and safety of permissive hypercapnia versus normocapnia in preterm infants on mechanical ventilation. Data Sources: MEDLINE, EMBASE, CENTRAL, and CINAHL Study Selection: Published randomized controlled trials (RCTs), non-RCTs, interrupted time series, cohort studies, case-control studies, and controlled before-and-after studies were included. Data Extraction: Two reviewers independently screened the title and abstract and full text, extracted data, assessed the risk of bias, and evaluated certainty of evidence (CoE) according to the Grading of Recommendations Assessment, Development and Evaluation approach. A meta-analysis of RCTs was performed using the random-effects model. Results: Four RCTs (693 infants) and one cohort study (371 infants) were included. No significant differences existed between the permissive hypercapnia and normocapnia groups for bronchopulmonary dysplasia (BPD) (risk ratio [RR] 0.94; 95% confidence interval [CI] 0.74-1.18; very low CoE) and a composite outcome of death or BPD (RR 1.05; 95% CI 0.90-1.23; very low CoE). Permissive hypercapnia may increase necrotizing enterocolitis (RR 1.69; 95% CI 0.98-2.91; very low CoE), although the null or trivial effect cannot be excluded. No significant differences existed between the two groups for any other outcome assessed (very low-to-low CoE). Limitations: The sample sizes were less than the optimal sizes for all outcomes assessed, indicating the need for further trials. Conclusions: Permissive hypercapnia did not have any significant benefit or harm in preterm infants.