Small Studies in Systematic Reviews: to Include or Not to Include?

Background: COVID-19 provided a real challenge for evidence synthesis due to the rapid growth of evidence. We aim to assess the impact of including all studies versus including larger studies only in systematic reviews when there is plethora of evidence. We use a case study of COVID-19 and chronic kidney disease (CKD). Methods: The review team conducted a systematic review of multiple databases. The review assessed the effect of CKD on mortality in patients with COVID-19. We performed a sensitivity analysis to assess the effect of study size on the robustness of the results based on cutoffs of 500, 1000 and 2000 patients. Results: We included 75 studies. Out of which there were 40 studies with a sample size of >2,000 patients, seven studies with 1,000-2,000 patients, 11 studies with 500-1,000 patients, and 17 studies with <500 patients. CKD increased the risk of mortality with a pooled hazard ratio (HR) 1.57 (95% confidence interval (CI) 1.42 - 1.73), odds ratio (OR) 1.86 (95%CI 1.64 - 2.11), and risk ratio (RR) 1.74 (95%CI 1.13 - 2.69). Across the three cutoffs, excluding the smaller studies resulted in no statistical significance difference in the results with an overlapping confidence interval. Conclusions: These findings suggested that, in prognosis reviews, it could be acceptable to limit meta-analyses to larger studies when there is abundance of evidence. Specific thresholds to determine which studies are considered large will depend on the context, clinical setting and number of studies and participants included in the review and meta-analysis.