Inferring transmission risk of respiratory viral infection from the viral load kinetics of SARS-CoV-2, England, 2020 to 2021 and influenza A virus, Hong Kong, 2008 to 2012

Abstract

BackgroundInfectiousness of respiratory viral infections is quantified as plaque forming units (PFU), requiring resource-intensive viral culture that is not routinely performed. We hypothesised that RNA viral load (VL) decline time (e-folding time) in people might serve as an alternative marker of infectiousness.AimThis study's objective was to evaluate the association of RNA VL decline time with RNA and PFU VL area under the curve (AUC) and transmission risk for SARS-CoV-2 and influenza A virus.MethodsIn SARS-CoV-2 and influenza A virus community cohorts, viral RNA was quantified by reverse transcription quantitative PCR in serial upper respiratory tract (URT)-samples collected within households after an initial household-member tested positive for one virus. We evaluated correlations between RNA VL decline time and RNA and PFU-VL AUC. Associations between VL decline time and transmission risk in index-contact pairs were assessed.ResultsIn SARS-CoV-2 cases, we observed positive correlations between RNA VL decline time and RNA and PFU VL AUC with posterior probabilities 1 and 0.96 respectively. In influenza A cases a positive correlation between RNA VL decline time and RNA VL AUC was observed, with posterior probability of 0.87. Index case VL decline times one standard deviation above the cohort-mean showed a relative increase in secondary attack rates of 39% (95% credible interval (CrI): -6.9 to 95%) for SARS-CoV-2 and 25% (95% CrI: -11 to 71%) for influenza A virus.ConclusionWe identify VL decline time as a potential marker of infectiousness and transmission risk for SARS-CoV-2 and influenza A virus. Early ascertainment of VL kinetics as part of surveillance of new viruses or variants could inform public health decision making.

Keywords: SARS-CoV-2; clearance time; decline rate; decline time; infectiousness; influenza A; transmission risk; viral load kinetics.

Figure 1

Illustration of the advantage of using two samples to estimate the viral load area under the curve

Figure 2

Flowchart of the recruited SARS-CoV-2 and influenza A virus incident cohorts and paired index-contact cohorts, with SARS-CoV-2 cases recruited in England, United Kingdom, 2020–2021, and influenza A cases in Hong Kong Special Administrative Region, China, 2008–2012 (n = 2,541 participants recruited)

Figure 3

Scatter plots of the VL AUC as a function of the VL decline time for A) SARS-CoV-2 RNA (n = 57), B) SARS-CoV-2 PFU (n = 33), and C) influenza A virus RNA (n = 14), with SARS-CoV-2 cases recruited in England, United Kingdom, 2020–2021, and influenza A cases in Hong Kong Special Administrative Region, China, 2008–2012

Figure 4

Scatter plot of the accuracy of using two samples to estimate the viral load decline time, with SARS-CoV-2 cases recruited in England, United Kingdom, 2020–2021 (n = 57)

Figure 5

(Top) Frequency distributions of the viral RNA decline times in index cases and (bottom) the association between index viral RNA decline time and SAR for (A) SARS-CoV-2 (n = 38) and (B) influenza A virus (n = 153), with SARS-CoV-2 cases recruited in England, United Kingdom, 2020–2021, and influenza A cases in Hong Kong Special Administrative Region, China, 2008–2012