L. Duchateau et al., Estimating number of events from the Kaplan-Meier curve for incorporation in a literature-based meta-analysis: What you don't see you can't get!, BIOMETRICS, 56(3), 2000, pp. 886-892
In literature-based meta-analyses of time-to-event data, the number of even
ts in the treated and control groups together with the total number of pati
ents randomized to the two treatment arms are often used as summary statist
ics. If interest is in mortality at a specified moment in time, the number
of events can, in most cases, only be obtained from the Kaplan-Meier curve.
The estimated number of events, however, is typically larger than the true
number of events. The effect of this overestimation on the Mantel-Haenszel
test and the odds ratio is studied in this paper. From these results, it c
an be concluded that the number of events should not be estimated from the
Kaplan-Meier curves for meta-analytic purposes unless virtually no patients
are lost to follow-up or censored and there are still many patients at ris
k in the two groups at the time at which the number of events is to be dete
rmined.