Patient or population: Children or adults with severe COVID-19 infection Intervention: Convalescent or hyperimmune intravenous immunoglobulin Comparison: Usual care + placebo (saline or intravenous immunoglobulin) | |||||
---|---|---|---|---|---|
Outcome | Relative effects, source of evidence | Absolute effects | Certainty/quality of evidence | Plain-language summary | |
Baseline risk for control group (per 1000) | Difference (95% CI) (per 1000) | ||||
Mortality (7–28 d) | RR 0.94 (95% CI 0.49 to 1.80) Based on 572 patients with influenza in 4 RCTs | 104* | −6 (−53 to 84) | Very low ⊕⊖⊖⊖ (Very serious indirectness and serious imprecision)† | Convalescent plasma may have little to no effect on mortality, but the evidence is very uncertain. |
Mortality (22 d) | RR 0.10 (95% CI 0.01 to 1.70) Based on 40 patients with SARS in 1 observational study | 104* | −94 (−103 to 73) | Very low ⊕⊖⊖⊖ (Serious indirectness, very serious risk of bias and serious imprecision)‡ | Convalescent plasma could have an important effect on decreasing or increasing mortality, but the evidence is very uncertain. |
Recovery by 28 days as measured by a 6-point ordinal scale§ | Proportional OR for recovery§ OR 1.05 (95% CI 0.67 to 1.64) Based on 438 patients with influenza from 2 RCTs | 104* | 5 (−30 to 56) | Very low ⊕⊖⊖⊖ (Very serious indirectness and serious imprecision)† | Convalescent plasma may have little to no effect on recovery, but the evidence is very uncertain. |
Length of hospital stay, d | Based on 259 patients with influenza in 3 RCTs | Median 13¶ | MD −1.62 (−3.82 to 0.58) | Very low ⊕⊖⊖⊖ (Very serious indirectness and serious imprecision)† | Convalescent plasma may confer a small reduction in hospital length of stay, but the evidence is very uncertain. |
Length of ICU stay, d | Based on 149 patients with influenza in 2 RCTs | Median 7** | MD −0.32 (CI −3.20 to 2.56) | Very low ⊕⊖⊖⊖ (Very serious indirectness and serious imprecision)† | Convalescent plasma may have little to no effect in reducing duration of ICU stay, but the evidence is very uncertain. |
Time on mechanical ventilation, d | Based on 83 patients with influenza in 2 RCTs | Median 9.25** | MD −3.67 (CI −7.70 to 0.36) | Very low ⊕⊖⊖⊖ (Very serious indirectness and serious imprecision)† | Convalescent plasma may reduce days of mechanical ventilation, but the evidence is very uncertain. |
Serious adverse events | RR 0.85 (95% CI 0.56 to 1.29) Based on 576 patients with influenza in 3 RCTs | 80†† | −12 (−35 to 23) | Low ⊕⊕⊖⊖ (Serious indirectness and imprecision)‡‡ | Convalescent plasma may result in little or no difference in the number of serious adverse events. |
Note: CI = confidence interval; COVID-19 = coronavirus disease 2019; GRADE = Grading of Recommendations Assessment, Development, and Evaluation; ICU = intensive care unit; MD = mean difference; OR = odds ratio; RCT = randomized controlled trial; RR = relative risk; SARS = severe acute respiratory syndrome.
↵* We chose the baseline risk from patients admitted to hospital with COVID-19 who did not receive convalescent plasma and steroids from the article by Guan et al.3 This paper reports 96/173 severely ill patients who did not receive steroids or hyperimmune plasma, of whom 10 patients died (Dr. W. Guan, State Key Laboratory of Respiratory Disease, National Clinical Research Center for Respiratory Disease, Guangzhou Institute of Respiratory Health, First Affiliated Hospital of Guangzhou Medical University, Guangzhou, Guangdong Province, China: personal communication, 2020). Hence, the baseline mortality risk is 10/96 = 10.4%. The median duration of hospital admission was 12.0 days (mean 12.8 d).
↵† We rated down 2 levels for indirectness because clinical and epidemiologic characteristics of patients with influenza vary in COVID-19. We rated down 1 level for imprecision because the confidence interval included both important benefit and important harm.
↵‡ Evidence from observational studies begins as low-quality evidence. We rated down 1 level for indirectness because evidence came from SARS rather than COVID-19. We rated 1 level down for imprecision because the confidence intervals included both important benefit and important harm.
↵§ Recovery defined by an ordinal outcome (6 mutually exclusive categories) at 28 days: death, in ICU, in hospital with oxygen support, in hospital without oxygen support, discharged but not fully recovered, discharged and fully recovered. An OR of > 1 indicates treatment is better than control, interpreted as odds of better recovery is 1.24 times higher among those treated with hyperimmune plasma than control arm. This OR is similar across categories. We also assumed the risk differences between treatment groups is the same across categories of the outcome.
↵¶ We chose the median duration of hospital admission from hospital-admitted COVID-19 patients with severe disease from the article by Guan et al.3
↵** This is the median number of days in ICU obtained from the control arm of RCTs including patients with severe influenza.
↵†† The baseline risk of serious adverse events obtained from the control arm of studies including influenza (3 studies).
↵‡‡ We rated down 1 level for indirectness for this safety outcome, inferring that the adverse effects are likely to be similar across viral illnesses, and 1 level down for imprecision because the confidence intervals included both important benefit and important harm.