To the Editor.—We read with concern the article by Davis and Williamson1  on the risk of coronavirus disease 2019 (COVID-19) transmission during autopsy, wherein the authors report a low risk (1:675) for autopsy-acquired COVID-19. The analysis is deeply flawed.

The authors report data with an unusual leniency of detail. The authors report “at least 225 autopsies” with “1–6 persons” per autopsy, where “at least 102” had brain removal. The authors do not inform on the duration, response rate, or bias assessments of their survey. There is no information on whether COVID-19–positive status was confirmed at autopsy (plausibly some patients had cleared their COVID-19 infection and died of secondary complications). Most importantly, there is no information on the frequency, duration, or method of monitoring of exposed personnel, or what proportion of personnel underwent laboratory testing.

The authors' calculations and comparisons are creative, conflate risks and rates, and are not corrected for time or geography.2  The risk of COVID-19 acquisition from autopsy cannot be calculated or approximated as 1:675 if the infection is not considered to be autopsy acquired. The tenuously calculated supposed “risk” cannot be compared with the aggregate incidence rate of the US population. The population incidence rate of 1:201 is a measure of the incidence of laboratory-confirmed COVID-19 in all 50 US states for the full 5-month duration of the pandemic. The authors' data do not even overlap with respect to time; also, their data are drawn from only 14 states. Furthermore, the risk of transmission of COVID-19 from autopsy should be compared with the risk of transmission of COVID-19 from not performing COVID-19–positive autopsies. The latter is zero.

The authors' conclusion that a risk of 1:675 is “exceedingly little” is subjective and not factual. Risk is assessed as acceptable or not using a frequency-severity matrix, which balances the probability and severity of harm with potential benefits. “High risk” becomes societally and individually acceptable when something of higher value is at stake; for example, physicians accept the risk for COVID-19 infection and treat/assess COVID-19–infected patients because a patient's life hangs in the balance. In contrast, one would not accept even a much lower risk, if, for example, a mouse's life hung in the balance.

Autopsies are critical to understand the pathobiology of a novel emerging disease, to inform management strategies, and for tissue acquisition for research toward prevention and cure. This report serves to undermine the value of the autopsy by glibly ascribing a “low risk” to autopsy-acquired COVID-19. This report also has the unfortunate potential for harm to trainees, who report being threatened by directors with no/poor recommendations should they refuse to participate in COVID-19–positive autopsies. At the moment, we have no concrete data to inform on the risk of COVID transmission at autopsy.

Risk of COVID-19 transmission during autopsy
[published online
Arch Pathol Lab Med.
Centers for Disease Control and Prevention.
Lesson 3: measures of risk
. Updated May 18, 2012. Accessed July 18, 2020.

The authors have no relevant financial interest in the products or companies described in this article.