Covid-19 vaccine candidate is unimpressive: NNTV is around 256
Important editorial notice for readers: This is a rapid response (online comment by a third party) and not an article in The BMJ. It is attributed in a misleading way on certain websites and social media. The Editor, 14/01/2021.
Pfizer’s vaccine “may be more than 90% effective.” (Mahase, BMJ 2020;371:m4347, November 9) Specific data are not given but it is easy enough to approximate the numbers involved, based on the 94 cases in a trial that has enrolled about 40,000 subjects: 8 cases in a vaccine group of 20,000 and 86 cases in a placebo group of 20,000. This yields a Covid-19 attack rate of 0.0004 in the vaccine group and 0.0043 in the placebo group. Relative risk (RR) for vaccination = 0.093, which translates into a “vaccine effectiveness” of 90.7% [100(1-0.093)]. This sounds impressive, but the absolute risk reduction for an individual is only about 0.4% (0.0043-0.0004=0.0039). The Number Needed To Vaccinate (NNTV) = 256 (1/0.0039), which means that to prevent just 1 Covid-19 case 256 individuals must get the vaccine; the other 255 individuals derive no benefit, but are subject to vaccine adverse effects, whatever they may be and whenever we learn about them……We’ve already heard that an early effect of the vaccine is “like a hangover or the flu.” Will vaccinees who are later exposed to coronaviruses have more severe illness as a result of antibody-dependent enhancement of infection (ADEI), a known hazard of coronavirus vaccines? Is there squalene in the Pfizer vaccine? If so, will vaccinees be subject to autoimmune diseases, like Gulf War Syndrome and narcolepsy that have been associated with the adjuvant?
We already know that current Covid-19 vaccine trials are unlikely to show a reduction in severe illness or deaths. (Doshi, BMJ 2020;371:m4037, October 21) Will they be like seasonal influenza vaccines, which have not proved to be lifesavers, and may even have increased overall mortality in the elderly? (Anderson et al, Ann Intern Med 2020;172:445) We need a lot more time and a lot more data, especially in view of massive uncertainties about Covid-19 case definitions and statistics.
ALLAN S. CUNNINGHAM 13 November 2020
Competing interests: No competing interests