Results

Review process

After elimination of duplicates, 1591 records were assessed by title and abstract, of which 164 records were further assessed by full-text. Finally, a total of 12 studies were included (Figure 1).

Characteristics of included studies

Of the 12 studies included, eleven 16–26 investigated the role of influenza co-infection with SARS-CoV-2 in the outcomes of interest, and three 22,23,27 investigated the role of RSV co-infection with SARS-CoV-2. The included studies represented 955 laboratory-confirmed COVID-19 patients coinfected with influenza or RSV and 6907 patients infected with SARS-CoV-2 mono-infection. All studies reported deaths as one of the outcomes of interest; four, six and six studies reported the outcomes of need or use of supplemental oxygen, mechanical ventilation and ICU admission, respectively.
All studies were conducted in inpatient settings except one study 19 with mixed settings (i.e., inpatients, emergence department and outpatients). Subject age varied greatly across the included studies and was reported in various statistical forms (e.g., frequency by age group, mean and standard deviation, median, etc.). The total number of SARS-CoV-2 coinfected with influenza/RSV and mono-infected patients per study ranged from 22 to 4501. The included studies were mainly conducted in Asia (9/10), among which six 16,17,21–23,25 were from China. For viral detection method, SARS-CoV-2 was detected using PCR for all studies whereas influenza and / or RSV co-infection was confirmed by PCR, serological testing or antigen assays. The quality assessment of included studies is provided in Table S1. Four studies 16,17,19,25 were assessed as high-quality and they used multivariate statistical methods to account for common confounders such as age, sex, and comorbidities. The basic characteristics of the included studies are available in Table 1.

Co-infection and risk of need or use of supplemental oxygen

Four studies16,22,25,26 reported the need or use of supplemental oxygen as an outcome (four on influenza and one 22 on RSV co-infection).
Our meta-analysis results showed that SARS-CoV-2 co-infection with influenza did not seem to be associated with increased need or use of supplemental oxygen compared with SARS-CoV-2 mono-infections (OR=1.04, 95% CI: 0.37-2.95) (Figure 2, panel A). When excluding studies with small sample size or low-quality studies, the meta-estimates did not differ substantially from the main analyses (Figure S1, panel A; Figure S2, panel A). SARS-CoV-2 co-infection with influenza A virus was also not observed to be associated with increased need or use of supplemental oxygen (OR=1.28, 95% CI: 0.36-4.53) (Figure 2, panel B). Two high-quality studies showed contrasting findings: one study 16 showed SARS-CoV-2 co-infection with influenza A virus was associated with decreased need or use of supplemental oxygen (OR=0.61, 95% CI: 0.48-0.76) whereas no such difference was observed on SARS-CoV-2 co-infection with influenza B virus (OR=0.97, 95% CI: 0.56-1.67); the other study 25 showed that SARS-CoV-2 co-infection with influenza was associated with increased need or use of supplemental oxygen (OR=2.47, 95%CI: 1.04-5.86).
Only one study 22 had available data on SARS-CoV-2 co-infection with RSV; however, none of the included patients in that study required supplemental oxygen.

Co-infection and risk of mechanical ventilation

Six studies17,19,22,24,25,27 reported mechanical ventilation as an outcome (all but one 27 on influenza and two studies 22,27 on RSV co-infection).
Based on the meta-analysis results, SARS-CoV-2 co-infection with influenza was found to be associated with a higher risk of mechanical ventilation as compared to SARS-CoV-2 mono-infection (OR=2.31, 95% CI: 1.10-4.85) (Figure 3, panel A). SARS-CoV-2 co-infection with influenza A virus was also associated with a higher risk mechanical ventilation (OR=5.04, 95% CI: 2.19-11.62) (Figure 3, panel B). After excluding low-quality studies, a similar meta-estimate was observed (Figure S2, panel B). Results from two high-quality studies 19,25using multivariable models were consistent with our meta-estimates although another high-quality study 17 showed no significant difference in receiving mechanical ventilation between the two groups.
Regarding the SARS-CoV-2 co-infection with RSV, one moderate-quality study 27 indicated that the co-infection was not associated with increased risk of mechanical ventilation (OR=5.00, 95% CI: 0.27-93.96) in children under two years old. The other low-quality22 study reported no patients receiving mechanical ventilation in either group.

Co-infection and risk of admission to intensive care unit (ICU)

Six studies18,19,22,24,25,27compared the utilisation of intensive care between mono-infection and co-infection (all but one27 on influenza and two 22,27 on RSV co-infection).
Based on the meta-analysis results, SARS-CoV-2 co-infection with influenza was associated with a higher risk of ICU admission compared with SARS-CoV-2 mono-infection (OR=2.09, 95% CI: 1.64-2.68) (Figure 4, panel A). Sensitivity analysis revealed a similar meta-estimate when excluding low-quality studies (Figure. S2, panel C). SARS-CoV-2 co-infection with influenza A virus was also associated with a higher risk of ICU admission (OR=2.11, 95% CI: 1.61-2.76) (Figure 4, panel B). Results from the two high-quality studies 19,25 were in accordance with the meta-analysis results.
Two studies 22,27 investigated the SARS-CoV-2 co-infection with RSV. One moderate-quality study 27found no significant difference in ICU admission (OR: 2.40, 95% CI: 0.18-31.88) between mono- and co-infection groups. Another one low-quality study 22 reported no patients being admitted to ICU.

Co-infection and risk of death

All studies included in our review reported the proportion of deaths in mono- and co-infection groups (all but one 27 on influenza and three22,23,27 on RSV co-infection).
For SARS-CoV-2 co-infection with influenza, the meta-analysis results showed that the co-infection was not associated with increased risk of death compared with SARS-CoV-2 mono-infection (OR=1.41, 95% CI: 0.65-3.08) (Figure 5, panel A). Sensitivity analyses that excluded studies with small sample size and low-quality studies showed similar meta-estimates (Figure S1, panel B; Figure S2, panel D). Similar results were also found in subgroup analysis by co-infection of influenza A and B virus (Figure 5, panel B, C). Findings from high-quality studies showed contrasting results: two studies 16,17 reported decreased risk of death in co-infection group (OR=0.51, 95% CI: 0.36-0.73; OR=0.26, 95% CI: 0.07-0.95, respectively) whereas another study 19 reported increased risk of death (OR: 2.27,95% CI: 1.23-4.19); in addition, another two studies24,25 (moderate-quality and high-quality, respectively) reported no differences in risk of death between the two groups (OR=4.61,95% CI: 0.98-21.67; OR=21.09, 95% CI: 0.84-527.66, respectively).
With respect to SARS-CoV-2 co-infection with RSV, three lower-quality studies 22,23,27 reported very small number of coinfected patients (range: 1-6). Our meta-analysis results suggested that no significant association was found between the co-infection status and death (OR=5.27, 95% CI: 0.58-47.87) (Figure 5, panel D).