To determine whether peak IgG suppression depth constitutes a valid study-level surrogate for MG-ADL treatment effect across FcRn antagonist trials in generalized myasthenia gravis, and to derive a surrogate threshold effect for early-phase dose-selection guidance.

Whether the depth of IgG suppression produced by FcRn antagonists predicts the magnitude of clinical response in generalized myasthenia gravis (gMG) at the trial level is unknown.

A Bayesian inverse-variance-weighted meta-regression of MG-ADL treatment effect on peak IgG suppression was conducted across active arms from randomized placebo-controlled trials of FcRn antagonists in adults with gMG. R²-trial was estimated using the Bayesian R² formulation, and the surrogate threshold effect (STE) was derived from the posterior predictive distribution. Prespecified sensitivity analyses restricted the dataset to a low risk-of-bias subset and excluded one arm with pharmacodynamic-clinical timing mismatch.

Five trials contributed 9 active arms (n = 274) covering efgartigimod, rozanolixizumab, nipocalimab, and batoclimab. Each 1% increase in peak IgG suppression was associated with a −0.035-point change in MG-ADL treatment effect (95% CrI, −0.098 to 0.030; posterior probability of a negative slope, 0.86). R²-trial was 0.24 (95% CrI, 0.00 to 0.55), below the prespecified surrogacy threshold of ≥ 0.50. The STE was 57.5% IgG suppression. A drug-level dummy did not improve fit (LOO ΔELPD, −0.8; SE, 1.2). The low risk-of-bias subset of 6 arms produced a consistent slope of −0.044 (95% CrI, −0.118 to 0.035).

Peak IgG suppression depth was directionally associated with MG-ADL treatment effect but did not meet criteria for a study-level surrogate. The 57.5% STE provides a quantitative anchor for early-phase dose selection; MG-ADL should remain the primary clinical endpoint in confirmatory FcRn-antagonist trials.