Covid-19 in end-stage renal disease patients with renal replacement therapies: A systematic review and meta-analysis

Background The novel coronavirus (COVID-19), caused by SARS-CoV-2, showed various prevalence and case-fatality rates (CFR) among patients with different pre-existing chronic conditions. End-stage renal disease (ESRD) patients with renal replacement therapy (RRT) might have a higher prevalence and CFR due to reduced immune function from uremia and kidney tropism of SARS-CoV-2, but there was a lack of systematic study on the infection and mortality of the SARS-CoV-2 infection in ESRD patients with various RRT. Methodology/Principal findings We searched five electronic databases and performed a systematic review and meta-analysis up to June 30, 2020, to evaluate the prevalence and case fatality rate (CFR) of the COVID-19 infection among ESRD patients with RRT. The global COVID-19 data were retrieved from the international database on June 30, 2020, for estimating the prevalence and CFR of the general population as referencing points. Of 3,272 potential studies, 34 were eligible studies consisted of 1,944 COVID-19 confirmed cases in 21,873 ESRD patients with RRT from 12 countries in four WHO regions. The overall pooled prevalence in ESRD patients with RRT was 3.10% [95% confidence interval (CI) 1.25–5.72] which was higher than referencing 0.14% global average prevalence. The overall estimated CFR of COVID-19 in ESRD patients with RRT was 18.06% (95% CI 14.09–22.32) which was higher than the global average at 4.98%. Conclusions This meta-analysis suggested high COVID-19 prevalence and CFR in ESRD patients with RRT. ESRD patients with RRT should have their specific protocol of COVID-19 prevention and treatment to mitigate excess cases and deaths.


Statistical analysis
The primary outcomes were COVID-19 prevalence and COVID-19 deaths. The prevalence outcomes were 1 4 1 measured with the percentage of COVID-19 confirmed cases and total ESRD with RRT patients with an associated 1 4 2 95% confidence interval (CI). The death outcomes were measured as CFR with the percentage of COVID-19 deaths 1 4 3 and total COVID-19 confirmed cases in ESRD with RRT patients with an associated 95% CI. The additional 1 4 4 outcomes were the need for mechanical ventilation and ICU admission among hospitalized COVID-19 confirmed 1 4 5 cases in ESRD with RRT patients with an associated 95% CI. The results of the studies were included in the meta-1 4 6 analysis and presented in a forest plot, which also showed statistical powers, confidence intervals, and heterogeneity.

4 7
This meta-analysis was performed to find pooled estimated value with a 95% confidence interval and did not have a 1 4 8 comparator group, thus no null hypothesis was tested. Instead, we compared the estimated values of primary 1 4 9 outcomes with global averages.

5 0
We assessed clinical and methodological heterogeneity by examining participant characteristics, follow-up 1 5 1 period, outcomes, and study designs. We then assessed statistical heterogeneity using the I 2 statistic. We regarded 1 5 2 level of heterogeneity for I 2 statistic as defined in chapter 9 of the Cochrane Handbook for Systematic Reviews of 1 5 3 Interventions: 0-40% might not be important; 30-60% may represent moderate heterogeneity; 50-90% may 1 5 4 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted January 26, 2021. ; https://doi.org/10.1101/2021.01.25.21250454 doi: medRxiv preprint represent substantial heterogeneity; 75-100% considerable heterogeneity. The Freeman-Tukey double arcsine 1 5 5 transformation was used to ensure admissible confidence intervals. The random-effects meta-analysis by 1 5 6 DerSimonian and Laird method was used as clinical, methodological, and statistical heterogeneity encountered. The 1 5 7 exact method was used for confidence interval computation. Prespecified subgroup analyses, including country 1 5 8 income level, WHO country region, and RRT modality, were performed. We assessed publication bias in two ways.

5 9
First, we computed each study effect size against standard error and plotted it as a funnel plot to assess asymmetry 1 6 0 visually. Second, we used Egger's test to statistically test for asymmetry. The significant asymmetry indicated the 1 6 1 possibility of publication bias or heterogeneity. The meta-analysis was performed using STATA 16.1 (StataCorp,  Characteristics of the studies 1 6 5 The database search identified 3,272 potential records. After duplicate removal, 1,885 titles passed the 1 6 6 initial screening, and 212 theme-related abstracts were selected for further full-text articles assessment for eligibility 1 6 7 (Fig 1). A total of 178 articles were excluded as the following: 44 letters to the editor, 36 case reports, 26 wrong 1 6 8 populations, 19 wrong outcomes, 16 editorials, 12 review articles, nine duplicates, nine recommendation, three 1 6 9 commentaries, two non-English, and two protocols. Only 34 studies were eligible for data synthesis and meta-1 7 0 analysis. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 26, 2021. ; .

CC-BY-NC-ND 4.0 International license
It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted January 26, 2021.  is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted January 26, 2021.

1
In 27 studies, the diagnostic criteria to confirm COVID-19 case were polymerase chain reaction (PCR) 1 9 0 assay only. While five studies used a combination of PCR and additional criteria to diagnose COVID-19: imaging 1 9 1 (three studies), suggestive symptoms (one study), and a combination of imaging, suggestive symptoms, and 1 9 2 laboratory tests (one study). One study did not explicitly state the diagnostic criteria. For COVID-19 treatment, 1 9 3 Hydroxychloroquine was the most common treatment used in 20 studies, followed by Lopinavir/Ritonavir used in 1 9 4 12 studies. Some studies used other medications including Darunavir/Ritonavir, Oseltamivir, Umifenovir, Ribavirin, 1 9 5 and Ganciclovir. Three studies reported no specific COVID-19 treatment, and eight studies did not report 1 9 6 information on COVID-19 treatment.  prevalence (Fig 2). . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted January 26, 2021.   0.95% (95%CI 0.28-1.97) which was lower than the pooled prevalence in high-income countries at 5.11% (95% 2 2 6 CI1.37-10.97) (Fig 3)       Thirty-one studies with death outcomes among 1,774 COVID-19 confirmed cases in ESRD with RRT from 2 6 9 1 5 RRT for each study (black boxes), with 95% confidence intervals (95% CI; horizontal black lines). The estimated 2 9 4 pooled case fatality rate for each subgroup was presented with a red diamond. The overall estimated pooled case 2 9 5 fatality rate (last red diamond) was 18.06% (95%CI 14.09-22.32). The global case fatality rate (vertical red line) 2 9 6 was 4.98%. The meta-analysis used a random-effects model with the exact method for confidence interval 2 9 7 estimation. ES, effect size. I 2 , test for heterogeneity.    . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted January 26, 2021. ;  To our knowledge, this is the first systematic review and meta-analysis of studies focusing on the 3 3 4 prevalence and CFR of the novel coronavirus in ESRD patients with various renal replacement therapy modalities.

5
The pooled data showed that the prevalence and CFR of COVID-19 in ESRD patients with RRT were significantly 3 3 6 higher than the global average in all populations [3]. The need for mechanical ventilation and ICU admission rate in 3 3 7 hospitalized ESRD patients with RRT was also high. The outcomes were varied among country income level, WHO 3 3 8 region, and type of RRT modality.

9
The overall pooled prevalence of COVID-19 in ESRD patients with RRT was higher than the global

4 4
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted January 26, 2021. ; https://doi.org/10.1101/2021.01.25.21250454 doi: medRxiv preprint While no evidence of publication bias, there was substantial heterogeneity between studies on prevalence 3 4 5 outcomes which might associate with country income level and RRT modality. Significant heterogeneity between 3 4 6 upper-middle-income and high-income countries was suggested from the lower estimated pooled prevalence of 3 4 7 upper-middle-income countries than high-income countries (0.95% vs. 5.11%). This might be partially explained by 3 4 8 the fact that China, which had a relatively lower overall national COVID-19 prevalence than the high-income 3 4 9 countries[3], was regarded as an upper-middle-income country. There was also a significant heterogeneity across 3 5 0 RRT modalities which the HD patients had a considerably higher prevalence than the KT patients (4.26% vs.

5 5
The overall CFR of COVID-19 in ESRD patients with RRT was higher than the global average (18.06% vs.

6 5
Several limitations of this systematic review and meta-analysis should be noted. First, the included studies 3 6 6 had high heterogeneities for both prevalence and CFR outcomes. The subgroup analysis could not fully explore all 3 6 7 sources of heterogeneity, given the limited availability of the information presented in the included studies. Second, 3 6 8 there was no included study focused on ESRD patients with PD only. Thus, the prevalence and CFR of COVID-19 3 6 9 in PD patients might not be available. Third, we did not explore deeper into specific details of each RRT modality 3 7 0 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted January 26, 2021. ; https://doi.org/10.1101/2021.01.25.21250454 doi: medRxiv preprint 1 9