CDC Sponsored Research

An update on a few CDC studies or data sources.  In the case of the studies, as has become the case with CDC, buyer beware, they are publishing stuff that suits the official line in many cases, but still a source of useful information.

If you like looking at big data sets with potentially lots of nuggets, here is the CDC case surveillance file, which contains details on hundreds of thousands of CV-19 cases.  (CDC File)

Antibody response was examined in this CDC published study.    (CDC Study)   The study looked at antibody responses of 58 patients who had mild or asymptomatic infections, for as long as 8 months after disease onset.  They found that almost all had a lasting response, and they were only looking at the actual antibody response, not other aspects of the adaptive immune response.  The authors used multiple assays and noted wide variation in results across those assays, suggesting that it is more likely inadequacy of the assay than actual lack of adaptive immunity that leads to concerns about waning antibody protection.

Asymptomatic transmission was the focus of this article by CDC researchers.  (JAMA Article)   And I am going to spend basically no time on it, because it is solely based on a model.  No contact tracing, no observational work, no nothing but built a model.  And tell that model that there is lots of asymptomatic transmission, so guess what, the model tells you that 59% of transmission is from asymptomatic people, 24% of whom never develop symptoms.  Garbage in, garbage out.

This article looked at the effect of shortening quarantine periods.  (CDC Article)   People with positive test results or with exposure had been advised to isolate for 14 days, which is a long time.  The CDC recently shortened that to ten days, based on lack of likelihood of infectiousness after that period.  This study looked households in Tennessee and Wisconsin to determine if shortening the period might result in more infections.  Household contacts of index patients were tested daily and tracked any symptoms.  Of 185 contacts, 59% had a positive test at some point, and 86% of those positive tests occurred within ten days of the index case developing symptoms.  So there was some positivity after ten days, but when completely analyzed and adjusted, that risk was around 5%.  But these are PCR tests and we got no information on cycle numbers.  I am willing to bet a very high threshold was used and they recorded a lot of non-viable virus positives.

This article examined the performance of an antigen test on a university campus.  (CDC Study)   Antigen tests can be done more rapidly but have accuracy concerns.  Over 1000 people were tested at two universities in Wisconsin and antigen testing compared to PCR testing done at the same time.  Antigen testing had very poor positive predictive value, i.e., lots of false negatives, but this rate was much better when only testing symptomatic people.  But among asymptomatic persons there were a number of false positives.  Interestingly, among symptomatic persons, only around 15% were positive by PCR testing.  Once again, we should be reminded that there are many causes of similar symptoms that are not CV-19.  Another interesting nugget is that among the antigen tests’ false negatives, many had high CT values, indicating that maybe antigen tests actually do a better job of identifying who actually is likely to be infectious.  When cultured, multiple supposedly positive antigen and PCR tests yielded no viable virus.  In fact, more than half of all positive tests did not culture virus.  As these authors note, antigen test-based widespread screening will carry risks of large numbers of inaccurate results, especially in a low prevalence population.

Another CDC published paper covers nursing home resident and staff cases.  (CDC Paper)   The study tracks data provided for the period from May 25 to November 22, following the federal requirement that nursing homes provide this data to the federal government.  Over 15,000 nursing homes were included in the analysis.  About 572,000 cases were reported, closely split between staff and residents.  The spring wave was missed but the trend in cases shows the early summer wave in some states, the lull in late summer and the re-emergence in fall, both in cases among residents and staff.  Trends is nursing home cases tended to mirror community transmission trends.  They also followed the general geographic pattern of the epidemic.