SARS-CoV-2-Legionella Co-Infections: A Systematic Review and Meta-Analysis (2020-2021)

Abstract

Legionnaires' Disease (LD) is a severe, sometimes fatal interstitial pneumonia due to Legionella pneumophila. Since the inception of the SARS-CoV-2 pandemic, some contradictory reports about the effects of lockdown measures on its epidemiology have been published, but no summary evidence has been collected to date. Therefore, we searched two different databases (PubMed and EMBASE) focusing on studies that reported the occurrence of LD among SARS-CoV-2 cases. Data were extracted using a standardized assessment form, and the results of such analyses were systematically reported, summarized, and compared. We identified a total of 38 articles, including 27 observational studies (either prospective or retrospective ones), 10 case reports, and 1 case series. Overall, data on 10,936 SARS-CoV-2 cases were included in the analyses. Of them, 5035 (46.0%) were tested for Legionella either through urinary antigen test or PCR, with 18 positive cases (0.4%). A pooled prevalence of 0.288% (95% Confidence Interval (95% CI) 0.129-0.641), was eventually calculated. Moreover, detailed data on 19 co-infections LD + SARS-CoV-2 were obtained (males: 84.2%; mean age: 61.9 years, range 35 to 83; 78.9% with 1 or more underlying comorbidities), including 16 (84.2%) admissions to the ICU, with a Case Fatality Ratio of 26.3%. In summary, our analyses suggest that the occurrence of SARS-CoV-2-Legionella infections may represent a relatively rare but not irrelevant event, and incident cases are characterized by a dismal prognosis.

Keywords: COVID-19; Legionella pneumophila typing; Legionnaires’ disease; SARS-CoV-2; diagnosis; epidemiology.

Figure A1

Forest plot representing the estimated case fatality ratio (CFR) among individuals affected by SARS-CoV-2 in the retrieved studies. Pooled CFR was estimated in 12.820% (95% Confidence Interval (95% CI) 7.199–21.801), with estimates that were considerably greater in PCR-based studies (14.688%, 95% CI 5.765–32.636) compared to those based on LUAT (11.571%, 95% CI 5.573–22.487) (note: Pos. = deaths; Tot. = total cases; Prev. = Estimated CFR).

Figure A2

Border-enhanced funnel plot for studies included in the meta-analysis (a), and the corresponding radial plot (b). Visual inspection of the contour-enhanced funnel plot suggested substantial evidence of publication bias, but this was substantially ruled out by the Egger test (i.e., t = −1.05, df = 12, p-value = 0.3131). On the other hand, in radial plots, the studies were substantially scattered across the regression line, suggesting no significant small study effect.

Figure 1

PRISMA flowchart for retrieved studies.

Figure 2

Forest plot representing the estimated pooled prevalence (Prev.) for Legionella infection (Pos.) among individuals affected by SARS-CoV-2 (Tot.). The pooled prevalence rate was estimated at 0.288% (95% Confidence Interval (95% CI) 0.129–0.641), with estimates that were considerably greater in PCR-based studies (0.377%, 95% CI 0.118–1.194) compared to those based on LUAT (0.247%, 95% CI 0.078–0.777).

Figure 3

Border-enhanced funnel plot for studies included in the meta-analysis (a), and the corresponding radial plot (b). Visual inspection of the contour-enhanced funnel plot suggested substantial evidence of publication bias, and this was substantially excluded by the Egger test (i.e., t = −1.25, df = 25, p-value = 0.2235). In fact, in radial plots, the studies on were substantially scattered across the regression line, suggesting no significant small study effect.