Table 3.

Comparison of treatment effects of pivotal trial using surrogate markers with matched postapproval trials using surrogate markers: primary and sensitivity analyses

	Summary difference between standardized mean differences or relative odds ratios (95% CI)
Method of analyses
At least two matching criteria (n = 43)
Reported non-continuous endpoints	1.50 (1.01 to 2.23)
Reported continuous endpoints	0.01 (−0.15 to 0.16)
All standardized as odds ratios (secondary)	1.12 (0.88 to 1.42)
At least three matching criteria (n = 39)
Reported non-continuous endpoints	1.45 (0.99 to 2.14)a
Reported continuous endpoints	0.05 (−0.08 to 0.19)
All standardized as odds ratios (secondary)	1.17 (0.95 to 1.45)
All four matching criteria (n = 33)
Reported non-continuous endpoints	1.21 (0.89 to 1.64)
Reported continuous endpoints	0.06 (−0.10 to 0.21)
All standardized as odds ratios (secondary)	1.13 (0.90 to 1.42)

A positive difference between standardized mean differences or ratio of odds ratios >1.0 indicates a larger benefit of treatment compared to the comparator in pivotal trials compared to postapproval trials. Pooled using DerSimonian and Laird random-effects meta-analyses

CI confidence interval

^a95% CI of 1.00 to 2.10 with a between-study correlation of 0.5