Comparison of COVID-19 outcomes in organ transplant recipients (OTr) and non-transplant patients: a study protocol for rapid review

This rapid review will compare clinical outcomes of COVID-19 between OTr and patients without a history of transplant (controls). The review protocol has been registered within the Open Science Framework database (osf.​io/​4n9d7) and is being reported in accordance with the reporting guidance provided in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) statement [7] (Additional file 1). The rapid review will be conducted following the methodological guidance from the Cochrane Handbook [8].

Database searches will be conducted in MEDLINE (via Ovid and PubMed) and EMBASE (via Embase.​com) for articles published between December 1, 2019, and October 31, 2021. The search strategy will be peer-reviewed by a research librarian in accordance with the Peer Review of Electronic Search Strategies (PRESS) 2015 Guideline Statement [9]. The Cochrane Library will not be searched as it focuses on effectiveness of interventions, which is not the focus of this research question.

The search will include a broad range of terms and keywords related to COVID-19 and organ transplant. COVID-19 concept search terms will be based on a previously designed search strategy on COVID-19 and telemedicine designed at Mayo Clinic Libraries on April 6, 2020 [10]. Organ transplant concept terms will be modified from the Ovid MEDLINE search strategy proposed by Raja et al. [11]. A draft search strategy for MEDLINE is provided in Additional File 2.

All articles identified from the literature search will be independently screened by two reviewers. First, titles and abstracts of articles returned from initial searches will be screened based on the eligibility criteria outlined above. Potential conflicts will be resolved by a third reviewer and potential discussion among all investigators. Second, full texts will be examined in detail and screened for eligibility. A flow chart showing the numbers of studies included and excluded at each stage of the study selection process will be provided. Reasons for exclusions will be provided for studies screened at the full-text level.

The risk of bias/quality assessment of primary studies will be evaluated using the Newcastle-Ottawa Scale (NOS) for observational (e.g., cohort and case-control) studies [12]. Using the NOS tool, each study is judged on eight items categorized into three groups: the selection of the study groups, the comparability of the groups, and the ascertainment of either the exposure or outcome of interest for case-control or cohort studies, respectively. Stars are awarded for each quality item, and the highest quality studies are awarded up to nine stars. We will consider studies with 0-3, 4-6, and 7-9 stars to represent low, moderate, and high-quality studies, respectively. The risk of bias for each study will be independently assessed by two reviewers. Discrepant scores will be resolved by a third reviewer and potentially discussion among all investigators.

We will perform narrative synthesis using evidence tables and evidence maps to describe the PICOTS elements, including study characteristics (e.g., author, journal, study design), population characteristics (e.g., eligibility, type of transplant, age distribution, demographics, definitions of COVID positivity), outcome operationalization (e.g., definitions of mortality, measures of effect estimates), covariates for which statistical models are adjusted, and findings. We will also assess whether covariates in adjusted models overlap sufficiently so that the reported conditional effects are estimates of the same estimand. Based on these elements, we will assess clinical and methodological heterogeneity and whether studies are statistically exchangeable to be meta-analyzed. The decision to meta-analyze is based on assumptions that are not testable with observed data as acknowledged by the Cochrane Collaboration [8]. Hence, we will quantitatively synthesize findings across studies through meta-analysis if our evidence tables and maps demonstrate that studies are similar enough in their design, the populations have similar characteristics, the two comparison groups (transplant vs non-transplant patients) are defined in the same way across studies, outcomes are operationalized in a consistent way in the body of evidence, and the reported effects are conditional estimates of the same estimand.

If the assumptions of study exchangeability are reasonably met, we will perform random-effects meta-analysis with the Sidik-Jonkman estimator and the Hartung-Knapp adjustment for confidence intervals to estimate a summary measure of association between history of transplant and each outcome. We are interested in binary outcomes for which we expect that studies report the measures of odds ratio, risk ratio, risk difference, or hazard ratio. To quantify statistical heterogeneity, we will use the I² statistic: the percentage of variance among studies that is attributable to genuine underlying differences in study properties rather than measurement error. We will explore potential sources of heterogeneity (e.g., organ type, number of years post-transplant, number of transplants, transplant center volume, US versus non-US studies, period of data collection) using random effects meta-regression. These parameters are selected as they can influence outcomes through host factors (organ type, number of years post-transplant, number of transplants) or different practices (including available treatments) in the management of COVID-19 (US versus non-US studies, period of data collection). We will assess the presence of small-study effects visually with funnel plots and statistically using Egger’s regression; both methods provide evidence for whether smaller vs larger studies give systematically different results, which may be due to publication bias as well as other meta-biases [13].

Although both adjusted and unadjusted estimates of association will be extracted, we will synthesize separately those for which primary studies have performed some adjustment for confounders, e.g., through regression, matching, or other approach.

The Newcastle-Ottawa Scale (NOS) will be used to evaluate study quality as it is validated, adaptable, and useful as a potential moderator in systematic reviews and meta-analyses.

Given the changing temporal and geographic landscape of the COVID-19 pandemic through the study inclusion period, we will perform subgroup analyses accounting for period of data collection, country and continent of study, and transplant center volume. We will also perform subgroup analyses for type of transplanted organ. To assess the role of methodological quality, we will perform subgroup analyses according to study quality determined by the NOS. We expect that studies will already report associations adjusted for covariates, which we will identify according to the publications. We will not adjust for confounders that are not included in the primary studies. If there is substantial variability in the covariates, we will consider potential categorizations (e.g., demographics, clinical characteristics, COVID-19 prevalence) and perform analyses according to the relevant subgroup, if feasible. Finally, we will perform random-effects meta-regression to explore factors that may drive statistical heterogeneity.

To assess the trustworthiness of the body of evidence and our confidence in the synthesized findings, we will use six quantitative criteria previously proposed [14] including level of statistical significance, sample size, statistical significance for the largest study, 95% prediction intervals, between-study heterogeneity, and the results of tests for small study effects.

All analyses will be conducted in Stata version 17 (StataCorp LP, College Station, TX, USA) and R (R Foundation, Vienna, Austria).