Back

Covid-19 positive test cycle threshold trends predict covid-19 mortality in Rhode Island

Bostom, A.; Eaton, C. B.; Kenyon, T.

2021-01-29 infectious diseases
10.1101/2021.01.26.21250557 medRxiv
Show abstract

The cycle thresholds (Cts) at which reverse transcriptase polymerase chain reaction (rtPCR) tests for covid-19 become positive are intimately associated with both viral load, and covid-19 infectiousness (i.e., ability to culture live virus). Clinical data indicate lower Cts--and hence larger viral loads--independently predict greater covid-19 mortality when patients are hospitalized for symptomatic covid-19 pneumonia. We merged public covid-19 mortality data from the Rhode Island Department of Health with a de-identified dataset of n=5036 positive rtPCR test Cts from the Rhode Island Department of Health State Laboratory to explore the potential relationship between positive covid-19 test Ct distribution trends, and covid-19 mortality in the state of Rhode Island, from March through early to mid-June, 2020. Mean daily covid-19 positive test Ct data were compiled, and 7-day rolling average covid-19 mortality was offset by 21-days, given the lag between infection and death. We divided the Ct data into three strata, >32, 28-32, and <28, which were operationally defined as "not infectious," "maybe infectious," and "infectious," respectively. Between late March and June, mean daily Ct values rose linearly (R-squared=0.789) so that by early June, as the covid-19 pandemic ebbed in severity, all means reached the noninfectious (Ct >32) range. Most notably, this May-June trend for Cts was accompanied by a marked, steady decline in Rhode Islands daily covid-19 mortality. Our results suggest that monitoring, and public reporting of mean population covid-19 test Cts over time is warranted to gauge the vacillations of covid-19 outbreak severity, including covid-19 mortality trends.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.