ABSTRACT
Background: The optimal isolation duration for COVID-19 patients remains unclear. To support an update of WHO Living Clinical management guidelines for COVID-19 (https://www.who.int/publications/i/item/WHO-2019-nCoV-clinical-2022.2), this rapid systematic review and modelling study addresses the effects of different isolation periods for preventing onward transmission leading to hospitalization and death among secondary cases. Methods: We searched World Health Organization (WHO) COVID-19 database for clinical studies evaluating the impact of isolation periods for COVID-19 patients up to July 28, 2022. We performed random-effects meta-analyses to summarize testing rates of persistent test positivity rates after COVID-19 infection. We developed a model to compare the effects of the five-day isolation and removal of isolation based on a negative antigen test with ten-day isolation on onward transmission leading to hospitalization and death. We assumed that patients with a positive test are infectious and those with a negative test are not. If the test becomes negative, patients will stay negative. The model included estimates of test positivity rates, effective reproduction number, and hospitalization rate or case fatality rate. Findings: Twelve studies addressing persistent test positivity rates including 2799 patients proved eligible. Asymptomatic patients (27.1%, 95% CI: 15.8% to 40.0%) had a significantly lower rapid antigen test (RAT) positive rate than symptomatic patients (68.1%, 95% CI: 40.6% to 90.3%) on day 5. The RAT positive rate was 21.5% (95% CI: 0 to 64.1%; moderate certainty) on day 10. Our modelling study suggested that the risk difference (RD) for asymptomatic patients between five-day isolation and ten-day isolation in hospitalization (2 more hospitalizations of secondary cases per 1000 patients isolated, 95% uncertainty interval (UI) 2 more to 3 more) and mortality (1 more per 1000 patients, 95% UI 0 to 1 more) of secondary cases proved very small (very low certainty). For symptomatic patients, the potential impact of five- versus ten-day isolation was much greater in hospitalizations (RD 19 more per 1000 patients, 95% UI 14 more to 24 more; very low certainty) and mortality (RD 5 more per 1000 patients, 95% UI 4 more to 6 more; very low certainty). There may be no difference between removing isolation based on a negative antigen test and ten-day isolation in the onward transmission leading to hospitalization or death, but the average isolation period (mean difference -3 days) will be shorter for the removal of isolation based on a negative antigen test (moderate certainty). Interpretation: Five versus 10 days of isolation in asymptomatic patients may result in a small amount of onward transmission and negligible hospitalization and mortality, but in symptomatic patients concerning transmission and resulting hospitalization and mortality. The evidence is, however, very uncertain.