T2-weighted and gadolinium enhanced T1-weighted MRI scans measure plaque burden and breakdown of the blood-brain barrier, respectively, in multiple sclerosis (MS) lesions. These have become widely used outcome measures for monitoring disease activity in clinical trials and clinical practice. However, their use as surrogates or biomarkers for disability and relapses, key clinical outcome measures, has remained incompletely validated. In a clinical trial database comprised of 31 relapsing-remitting and secondary progressive MS trial placebo groups, we assessed relationships between 1) T2 lesion load (TLL) change and disability change and 2) gadolinium enhancement of MS lesions and on-study relapses with univariate and multivariate analyses. In relapsing-remitting MS, TLL change (n = 223) made no independent contribution to predicting change in disability from baseline to trials' end. Similarly, inclusion of gadolinium enhancing lesions (n = 170) into multivariate models did not independently contribute to the predictive value for on-trial relapses. In secondary progressive MS, a small effect of TLL was found for disability change (n = 355) but in multivariate analysis this accounted for less than 5% of the variance in end-of-trial disability. Results were replicated in independent datasets, more than doubling effective sample sizes. MRI measures widely used in trials of relapsing-remitting and progressive multiple sclerosis add little if anything independently to the clinically relevant relapse and disability outcomes. These results reemphasize the importance of validating potential surrogate markers against clinical measures and highlight the need for better MRI markers of disease activity and progression.