Timely discussion. Do recurrent events always provide greater power than time-to-first events?

In many cases, adapting to recurrent events is favorable as w #SOLOISTWHF #AHA20 @DLBHATTMD

However, despite ↓ events, time-to-first may offer more precise estimates in select cases. https://twitter.com/BrettSperryMD/status/1329043508078972929

Example: #AFFIRMAHF #AHA20

https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)32339-4/fulltext

Primary Endpoint: total HFH+ CVd
# of Events: 665
RR 0·79, 95% CI 0·62–1·01; p=0·059

Time-to-First: first HFH+ CVd
# of Events: 390
HR 0·80, 95% CI 0·66–0·98; p=0·030

Wider CI despite ↑ events. How might this be the case?

In-depth paper @CircAHA featuring #TOPCAT #PARADIGM & #CHARM shows that variables beyond n of events are important in statistical gains, namely heterogeneity of pt risk & drug discontinuation. These parameters will be increasingly important w #COVID19.

https://www.ahajournals.org/doi/full/10.1161/CIRCULATIONAHA.117.033065