Low infection rates could extend data readouts in large-scale COVID-19 trials

• Falling COVID-19 infection rates, especially among adults at least 55 years old, are good news/bad news for public health and vaccine developers. The negative implications of higher infection rates are self-evident for the overall population but are a narrowly defined positive for vaccine makers considering the shorter amount of time to accumulate enough events, i.e., cases of post-dose infections, to gauge efficacy. Lower infection rates are clearly preferred, again for obvious reasons, but extend the timeline in a vaccine trial since event accrual is slower. This is a major reason that late-stage vaccine trials recruit tens of thousands of participants. Event accrual is the tollgate for assessing efficacy, even preliminarily.
• Assuming the FDA's stated 50% minimum efficacy bar, Johnson & Johnson's (NYSE:JNJ) Phase 3 trial, ENSEMBLE, evaluating Ad26.COV2.S in 60K adults will require a minimum of 154 events. Likewise, Pfizer (NYSE:PFE)/BioNTech's (NASDAQ:BNTX) 44K-subject study of BNT162b2 will need 164 events (52% efficacy), AstraZeneca's (NASDAQ:AZN) 30K-subject study of AZD1222 will need 150 and Moderna's (NASDAQ:MRNA) 30K-subject trial of mRNA-1273 will need 151.
• Vaccine efficacy (VE) is calculated as: VE = 100 x (1 - IRR). IRR, the infection rate ratio, is the rate of confirmed COVID-19 infections in the vaccinated group divided by the rate of infections in the placebo group.  For example, if the infection rate in the test group was 0.1% compared to 1.0% in the control group, vaccine efficacy would be 90% [100 x (1 - (0.1%/1.0%)].
• A potential emergency use nod in the U.S. based on an interim analysis of less data will likely need a higher efficacy bar before a developer would proceed with an application considering the current public mood about rushing the vaccines to market without thorough vetting since there is no way to truncate the time required to thoroughly assess medium- and long-term safety and efficacy.
• J&J's study, launched last week, assumes 60% vaccine efficacy and a 1.4% infection rate in the placebo group during the first three months, dropping to 0.7% in month four and ~0.5% thereafter. Monitoring can start after half of the participants have at least two months' follow-up from the second shot, at least six cases of COVID-19 have been reported in subjects at least 60 years old and at least 20 moderate-to-severe/critical COVID-19 cases have occurred (at least five deemed severe/critical) (protocol, page 110). No interim analyses or efficacy monitoring will be done until these conditions are met. Monitoring for efficacy will be done at least once/week by independent committee until 154 events have been reported or the efficacy (statistical) boundaries have been met.
• The minimum number of events in Pfizer/BioNTech's study is also based on an assumed VE of 60%. 164 events will provide 90% power (statistical threshold) to demonstrate efficacy and an early end to the study (protocol, page 92). This would be achieved with 17,600 evaluable subjects in each arm or 21,999 vaccine recipients randomized 1:1 with placebo for a total sample size of 43,998 based on a 1.3% infection rate in the placebo group. As of September 28, enrollment was 35,469 including 24,289 participants who have received the second shot.
• Likewise, Moderna is basing the 151 events in its Phase 3 trial on an assumed VE of 60% (90% power) and a 0.75% six-month infection rate in the control arm (protocol, page 13). There are two interim analyses planned, after 35% and 70%, respectively, of the total targeted cases have been observed. As of September 25, 27,232 people have been enrolled including 15,454 who have received the second jab.
• All things considered, the chance of a U.S. emergency use nod before election day appears remote, but possible before year-end.
• https://seekingalpha.com/news/3618991-low-infection-rates-extend-data-readouts-in-large-scale-covidminus-19-trials