COVID: seroprevalence 10x higher than Covid cases; prison transmission

recent evidence suggests that the seroprevalence of SARS-CoV-2 is >10 times higher than reported (see the CDC summary at https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/commercial-lab-surveys.html , and the article preprint at https://www.medrxiv.org/content/10.1101/2020.06.25.20140384v1.full.pdf

Details:

--the CDC partnered with 2 commercial labs to assess SARS-CoV-2 antibodies from 6 areas: Connecticut, South Florida, New YorkCity metro area, Missouri, Utah, and Western Washington State. they will expand to 4 other states as well: California, Louisiana, Minnesota, and Pennsylvania

--they tested 11,933 blood samples collected for other reasons (sick or routine visits unrelated to Covid-19, such as lipid screening)

--goal:1,800 samples from each area, approximately every 3-4 weeks

--they balanced the sampling by age (0-18yo, 19-49, 50-64, >65) and sex

--this was compared to the positivity rate of PCR of SARS-CoV-2 in those areas

--sera were tested at the CDC, having "developed and validated an ELISA assay"

    --they assumed that the antibody test had 96.0% sensitivity and 99.3% specificity

Results (from the first six sites only, others are scheduled), will include the most current number of cases for the states, per Johns Hopkins:

--Western Washington state (march 23-april 1): 4,300 cases reported, 48,300 estimated cases by seroprevalence (1.13%); seroprevalence 11x higher

--New York City metro region (march 23-april 1): 53,800 cases, 641,800 estimated cases by seroprevalence (6.93%); seroprevalence 12x higher

--South Florida  (april 6-april 10): 10,500 cases reported, 117,400 estimated cases by seroprevalence (1.85%); seroprevalence 11x higher

--Missouri (april 20-april 26): 6,800 cases, 161,900 estimated cases by seroprevalence (2.65%); seroprevalence 24x higher

    --now 20.7K cases

--Utah (april 20-may 3): 4,500 cases, 47,400 estimated cases by seroprevalence (2.18%); seroprevalence 11x higher

    --now 17.1K cases

--Connecticut (april 26-may 3): 29,300 cases, 176,700 estimated cases by seroprevalence (4.94%); seroprevalence 6x higher

    --now 46.3K cases

--overall, there was no real difference by age groups, though there were some differences in the different sites. and not much difference by sex

Commentary:

--finally, some important prevalence numbers!!!! (caveats below)

--no granular data on the actual symptoms of these patients, though all were presumably in patients not being seen for Covid-19

--they did put in a fudge factor for the antibody tests' imperfections (somewhat decreased sensitivity and specificity). there was  no such fudge factor for the relatively poor performance of the PCR test, but that is appropriate since, for better or worse, that is what we are using to assess SARS-CoV-2

--it is important to remember that seropositivity found by antibody testing, even with the most accurate tests, does not necessarily indicate protection from reinfection (eg see the WHO advisory: http://gmodestmedblogs.blogspot.com/2020/05/covid-who-dont-trust-antibody-testing.html). the major use of antibody testing, as done in this study, is to get a better idea of the epidemiology of the virus (SARS-CoV-2) and how that relates to clinical disease (covid-19). And, prior estimates of the level of past infection to create herd immunity (about 65%) also relies on measuring protective antibodies (which we do not know at this point)

--some strengths of the study: they used a CDC-validated antibody test; and this is an ongoing study to assess changes over time

limitations:

--the above study was done 2-3 months ago, the numbers are much higher now, though there will be ongoing sampling over time,which should give great insight into the epidemiology and spread of the disease, and, theoretically if the political powers so choose, lead to appropriate changes in prevention of disease transmission

    --for example, as noted above from the Johns Hopkins data for Utah is now 17.1 K cases, vs 4.5K above; though not sure that this increase necessarily translates to 200K cases by serology, as below

--the blood samples may not reflect the general population, since these samples were from health care facilities. And, at the time of the above collections, many states did not have adequate testing supplies and targeted only those with overt symptoms of fever and cough (ie, with more lenient testing now, the PCR testing might be much higher than in these studies, leading potentially to a lower seroprevalence to PCR ratio)

--and there was no ability to assure that the same person may have had several samples tested, also potentially biasing the results

--the rates found in some of the larger areas (eg states) may not truly represent the overall state: samples may have come disproportionately from some areas (eg more urban,.or higher infection rates), and generalizing these numbers per the overall state population may be inaccurate

--antibody response is also not uniform and may be lower or undetectable in those with milder disease (eg see http://gmodestmedblogs.blogspot.com/2020/06/covid-asymptomatic-patient.html ) ie, there may be many more infections than antibody testing will pick up

so, i think this data does support the following conclusions:

    --lots of SARS-CoV-2 is asymptomatic, or at least minimally symptomatic

    --the possibility of signficantly decreasing the spread of SRS-CoV-2 will depend on lots of screening of asymptomatic people and contacts if positive

    --even states with low numbers of cases of Covid-19 seem to have pretty high populations who have been (or still are) infected/infectious: eg Utah and Missouri in the above 6 areas, with pretty few cases of Covid-19 but estimated seroprevalence in the 2-3% of their population (!!!)

    --so, the idea of opening up travel and relaxing social distancing/masks just because the number of reported Covid-19 cases has diminished is really, really likely to increase the spread of the virus (ie, that it is okay for people from low prevalence of Covid-19 to travel since the likelihood of transmitting the virus is very low, seems to be folly and likely to propel the continuing spread of the virus)

--------------------------------------------------------------------------

And, another study was just published in MMWR finding that serial testing of 98 incarcerated people in a Louisiana prison found a high % who were asymptomatic or presymptomatic (see https://www.cdc.gov/mmwr/volumes/69/wr/mm6926e2.htm?s_cid=mm6926e2_w ):

--on March 29, a staff member developed Covid-19

--April 2-May 7: 2 more cases in staff, and 36 in incarcerated persons (all confirmed caes were immediatly transferred to another facility for isolation)

--May 7: CDC and Louisiana dept of health began study

--May 7-21: 98 incarcerated people tested: 71 additional cases of infection (32, or 45%, were asymptomatic at time of testing; 3 were presymptomic); 18 of the cases (25%) were in people who were negative in prior rounds of testing

    --"serial testing of contacts from shared living quarters identified persons with SARS-CoV-2 infection who would not have been detected by symptom screening alone or by testing at a single point". so,

        --identification of asymptomatic and presymptomatic people is really important, as they can often be the source of virus spread (and, in this study, there might have been a real difference if aggressive testing began earlier, not 2-4 weeks after identifying the first Covid-positive person)

        --contact tracing is important in focusing on higher risk people

        --and identification of asymptomtatic infected people is really important and should lead to isolation/quarantine

--demographics of the prison: median age 33, 93% male, 66% non-Hispanic Black/32% non-Hispanic Whie; 40% with an underlying health condition, 23% had BMI>30

--sequence of detection over time:

    --of 98 people tested on day 1: 54% were SARS-CoV-2 positive

    --of the remaining 45 who where negative, 16 (36%) were positive on day 4

    --of the 29 negative on both days, 2 (7%) were positive on day 14

--asymptomatics: of the 71 cases, 2 were presymptomatic, and 29 (41%) were in asymptomatic people

--among the 27 people testing negative, 18 (67%) had Covid-like symptoms (8 with loss of smell, 7 with loss of taste), suggesting that many of these other people likely had Covid-19 but had false negative tests (for the high incidene of false negative PCR testing, see http://gmodestmedblogs.blogspot.com/search?q=covid+false+negative )

--attack rates were quite high: from 57% in one dormitory to 82% in another

.

so, this study adds to others suggesting that there is very high transmission rates in confined persons (eg in nursing homes: see http://gmodestmedblogs.blogspot.com/2020/04/covid-presymptomatic-transmission-in.html ) or in homeless persons who often live in close contact with others (see http://gmodestmedblogs.blogspot.com/2020/04/covid-asymptomatic-homeless-patients-in.html ). And it reinforces the conclusions of the above study of really focusing on testing asymptomatic people and aggressively tracking their contacts, and not relying just on the reported Covid-19 cases as currently tracked daily.

geoff

 

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@uphams.org

to get access to all of the blogs:

1. go to http://gmodestmedblogs.blogspot.com/ to see them in reverse chronological order

2. click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

3. or you can just click on the magnifying glass on top right, then  type in a name in the search box and get all the blogs with that name in them

 

please feel free to circulate this to others. also, if you send me their emails, i can add them to the list