Of 45778 studies retrieved from the searchers, 66 trials with 32 treatments (19, 095 patients) met inclusion criteria (Fig. 1), including 8 different antiviral agents, 6 antibiotics, high and low dose CQ (CQ_HD, CQ_LD), 5 add-on treatments (LPV_RIT_IFN, LPV_RIT_RIB_IFN, RIB_IFN, NOV_LPV_RIT, CQ_LD_AZI), COR, TCM, CON_PLA, ALA, RUX, NOV, COL, IFN, IVIG, SOC and PLA. Among these treatments, 6 antibiotics were separately compared in 1 trial for SARS-COV  and 1 trial for SARS-COV-2 , which could not be connected with other treatments in the trial network. Excluding the six antibiotics, 26 treatments within 64 trials were finally analyzed in the network.
90.6% (58/64) of trials were two-arm studies and only 6 were multiple-arm studies (Appendix 2). Overall, 18,881 patients were involved in the meta-analysis, of whom, 17,416, 14,708, 1,197 and 11,698 patients contributed to four outcomes of mortality, cure rate, VNC and OAE, respectively (Fig. 2). Appendix 2 summarized the characteristics of included trials. Publication year were focused from 2003 to 2005 for SARS-COV (19 trials) and 2020 for SARS-COV-2 (45 trials). Trial duration ranged from 5 to 75 days with a median duration of 15 days [interquartile range (IQR): 13–28 days]. The mean age of included patients was 48.9 years [standard deviation (SD) 9.8] and mean proportion of female was 45.7% (IQR:40%-53.5%). Disease severity was mild/moderate in 23 trials and moderate/severe in the others. SOC, TCM, antiviral agents and CQ were the most common studied drugs, within51, 31, 20 and 14 trials, respectively, followed by PLA (6 trials) and COR (4 trials).
Results of pairwise meta-analysis
The effects of different drugs on mortality, cure rate, VNC and OAE from pairwise meta-analyses are shown in Appendix 4. TCM and COR were associated with a significant reduction in mortality (OR = 0.33, 95%CI: 0.20–0.55 and OR = 0.85, 0.75–0.96 respectively) and a significant increase in cure rate (OR = 2.17, 1.61–2.92 and OR = 1.17, 1.05–1.30 respectively) compared with SOC. Compared with CQ_LD, CQ_HD showed an evident increase in mortality (OR = 3.33, 1.19–9.35)). Compared with SOC, TCM showed a significant decrease in the risk of OAE(OR = 0.52, 0.38–0.71)), whereas CQ_HD, IFN, COL and CQ_LD_AZI were all associated with increased risk, with the ORs ranging from 1.65(1.11–2.47,CQ_LD_AZI) to 3.83 (1.68–8.70, CQ_HD). CQ_LD was significantly associated with increased risk of OAE versus placebo (OR = 2.75, 1.80–4.20).
Results of NMA
Results of NMA are reported in Fig. 3–4. Significant decrease in mortality was observed for TCM (OR = 0.34, 0.20–0.56) and COR (OR = 0.84, 0.75–0.96) versus SOC, while CQ_HD could significantly increase the risk (OR = 3.20, 1.18–8.73). Besides, a significant reduction in mortality was detected for TCM versus COR, CQ_HD and CQ_LD with the ORs.
Regarding cure rate, the pooled results favor TCM (OR = 2.16, 1.60–2.91) and COR (OR = 1.17, 1.05–1.30) in comparison with SOC. When compared with COR and CQ_LD, TCM could significantly improve cure rate with an OR of 1.85(1.35–2.56) and 2.38(1.72–3.33), respectively. No significant results were observed on VNC for any comparison.
In terms of OAE, both TCM (OR = 0.52, 0.38–0.70) and REM_10 (OR = 0.31, 0.19–0.52) were associated with decreased risk, whereas COL, CQ_HD and IFN were associated with increased risk versus SOC, with the ORs ranging from 2.51 (1.20–5.24, CQ_HD) to 3.81 (1.55–9.29, COL). In comparison with COL, COR, CQ_HD, CQ_LD, IFN and LPV_RIT, a significant reduction of OAE was found for TCM).
Results of NMA for secondary outcomes are listed in Appendix 5. Among separate AEs, a significant increased risk of diarrhea was detected for COL, CQ_HD, CQ_LD, LPV_RIT and LPV_RIT_RIB_IFN versus SOC (OR ranging from 3.80(1.55–9.26, COL) to 9.62(1.70-56.11, LPV_RIT)) and TCM (OR ranging from 8.77(2.99–25.74, COL) to 22.27(3.53-142.16, LPV_RIT)). Meanwhile, TCM showed significantly decrease in the risk of diarrhea (OR = 0.43, 0.24–0.79) and SEI (OR = 0.33, 0.17–0.64) when compared with SOC. Additionally, an evident reduction of ARDS was found for LPV_RIT versus SOC (OR = 0.37, 0.18–0.80). No other significant results were detected for other separate AEs (AKI, SAE, TE and HF).
TCM, RIB_IFN, CON_PLA, LPV_RIT_RIB_IFN and LPV_RIT were all associated with shorter hospitalization length ranging from − 17.80 (-29.92, -5.74, RIB_IFN) to -2.53 (-3.43, -1.63, TCM) days when compared with SOC. Furthermore, TCM could reduce hospitalization length for 2.63(CQ_LD) to 5.09(IFN) days versus CQ_LD, CQ_LD_AZI and IFN. Besides, a significant reduction on TFR was found for TCM versus SOC (WMD=-1.03 days, 95%CI: -1.09, -0.97).
According to contribution tables of the network (Appendix 6), comparison of SOC versus antiviral agents, COR and CQ had the largest contribution in all 4 entire networks for primary outcomes with 51.8%, 56.5%, 56.9% and 50.0% for mortality, cure rate, VNC and OAE, respectively.
Transitivity, inconsistency and heterogeneity
Assessment of transitivity by box plots indicated mean age and proportion of female across treatment comparisons were relatively similar (Appendix 7). The test for global inconsistency did not detect any significant difference between consistency and inconsistency models for all the 4 primary outcomes (p = 0.994 for mortality, p = 0.763 for cure, p = 0.952 for VNC and p = 0.773 for OAE, respectively) and 3 secondary outcomes (p = 0.984 for DIA, p = 0.845 for SAE and p = 0.614 for TE, respectively), except for other 6 secondary outcomes(AKI, ARDS, HD, HF, SEI and TFR) could not conduct consistency test due to no loop in the whole network. The test for inconsistency from node-splitting model showed no significant difference in all comparisons across all outcomes (Appendix 8). Most comparisons in all 13 outcomes were with low heterogeneity as indicated in the predictive interval. At visual inspection, funnel plots for all 7 outcomes with number of studies greater than 10 (Appendix 9) were symmetric and did not suggest any significant risk of publication bias.
Rank heat plot and SUCRA of all treatments
Figure 5 and Appendix 10 showed the mean values of SUCRA for providing the hierarchy ranking of different treatments on all 13 outcomes. Due to sparse network data and non-definitive results in most comparisons, the ranking might be biased and interpretation should be made with caution.
GRADE evaluation on quality of evidence
According to GRADE, the quality of evidence ranged between very low and moderate (Appendix 11). In terms of TCM versus SOC, the quality was moderate for mortality, low for cure rate, and very low for VNC and OAE. As for COR versus SOC, the quality was low for mortality, cure rate and OAE while very low for VNC. Regarding to CQ_HD versus SOC, the quality was low for mortality and very low for OAE.
In addition, subgroup pairwise meta-analysis by virus type confirmed the beneficial effect on mortality and cure rate of TCM and COR versus SOC, reduction effect on OAE of TCM versus SOC, increased risk of mortality for CQ_HD versus CQ_LD, and increased risk of OAE for CQ_HD, IFN and CQ_LD_AZI versus SOC in SARS-COV-2, which were in agreement with those previous produced (Appendix 12).