Social media channels and prominent anti-vaxxers like Alex Berenson have been on fire with talk about an unreviewed Israeli study suggesting that infection with COVID-19 provides better immunity than the COVID-19 vaccines. It takes some effort find this study, since it is not published in any reputable journal and is only posted on the medRxiv website. Even a cursory review reveals several serious problems that challenge its validity.
First, this was a retrospective, observational study comparing three groups. There was no blinding, and there was no prospective study to validate the retrospective results. Thus, the authors could pick the outcome they wished and arrange the data for their desired outcome. This reminds me of the observational studies showing that hydroxychloroquine effectively treated COVID that were then refuted by prospective, blinded studies. Every study like this is clearly open to bias.
Second, this study used three, multivariate, logistic regression models to compare their groups. This means they tried to simultaneously estimate the effect of multiple variables (such as vaccination, prior infection or a combination of factors) on an outcome; in this case the likelihood of COVID-19 infection (or reinfection). While there is nothing inherently wrong with this, how the modeling was developed and the presumptions given to each of the different variables can entirely skew the results. The conclusions are not based on actual data, but on the projections derived from their models. It is also possible that the result is secondary to another factor (hospitalization or monoclonal antibody therapy for example) that was not evaluated but closely associated with infection.
In this regard, several variables the authors either analyzed or omitted are interesting. No pre-existing medical conditions were associated with re-infection, which is surprising since this can alter immunity. They also do not analyze timing of prior immunization (but do evaluate timing from prior infection) since all immunity wanes over time. This is important since Israel was the first nation to immunize their population. Given the reinfection period of evaluation was at a time when all infections were with the Delta variant and before Israel began booster shots, this is an issue.
Most importantly, the population they investigated was significantly skewed. Over 65% of individuals were under 40 years old, and only 5% were over 60 years of age! This totally diminishes the benefit of vaccination in the elderly. One might also suggest that there was a survivor effect here since many of the older folks (especially with co-morbidities) who were not immunized died, and only the healthiest survived the infection (to then be analyzed in the study)!
No one denies that COVID-19 infection provides immunity to re-infection. The real questions are at what cost to the individual and is natural immunity adequate to protect against future infections? Anyone who has looked at the ICU’s in Florida knows the cost of natural infection, since the vast majority of seriously ill individuals are unvaccinated. Even without worrying about long-term COVID the concept of people welcoming COVID infection as an alternative to vaccination is not rational.
But it is interesting that this study showed that individuals with “natural immunity” from infection had better long-term protection from reinfection if they received a dose of RNA vaccine after the infection.
So those of you that favor immunity from infection, if you survive COVID to prime your immune system then remember to get a dose of the vaccine!