• hemorrhage
• stroke

Success is not final; failure is not final; it’s the courage to continue that counts. Winston Churchill

Beginning over half a century ago with a randomized clinical trial by McKissock et al and continuing through the high-quality international trials STICH I, STICH II, and MISTIE III, most trials evaluating surgical evacuation of ICH have repeatedly been unable to demonstrate improvement in functional outcome.1–4 The only outlier being a smaller study by Pantazis et al that demonstrated a functional benefit from early surgery (<8 hours) in 108 patients with putaminal or lobar ICH. ENRICH robustly confirms the findings of Pantazis et al, demonstrating that effective evacuation of lobar hemorrhages provides patients with functional benefit.5 This is a major finding with immediate and broad-reaching impact.

The ENRICH trial’s primary efficacy endpoint was an intention-to-treat analysis evaluating whether mean utility-weighted modified Rankin Scale (uw-mRS) score at 180 days in the MIPS group was superior to that of the medical management group, using a Bayesian analysis with a posterior probability of success >97.5%. The study enrolled 300 patients with 69% lobar and 31% anterior basal ganglia, a baseline hematoma volume of 50.5 mL, and a mean last known well to surgery time of 16.6 hours. The median evacuation percentage was 87.7%, and the proportion of patients in whom a final hematoma volume of ≤15 mL was achieved was 72.7%. The 30-day mortality rate was 11.3% in the MIPS group and 18.7% in the medical management group. The proportion of patients achieving a uw-mRS score of 0–3 at 6 months was 50% in the MIPS group and 41% in the medical management group for a posterior probability of 0.9813, which met criteria supporting superiority of surgical evacuation over medical management.

There were three unusual but well-supported features in the trial design that merit discussion. First, the primary outcome measure used was the uw-mRS score, which is a composite score of the mRS score and a patient-reported outcome measure. The mRS 0–6 scale is converted to a 0–1 scale with each step in the score weighted by the results of the patient-reported outcome measure for that group. Although an uncommon outcome measure and one that has not been validated in an ICH population, the uw-mRS has strong support among stroke researchers. Importantly, this scale converts an ordinal scale with fixed relative mathematical value between each mRS grade to a scale that seeks to value the difference more appropriately between each mRS grade. Essentially, by capturing patients’ input regarding their perceived value of each mRS grade, the uw-mRS makes it clear that the relative difference between an mRS of a 0 to a 1 is not the same as the difference between a 3 and a 4, and therefore should be weighted equivalently in the statistical analysis. This has important statistical implications.

Second, the study used an adaptive trial design including prespecified criteria to stop enrolling one of the hematoma locations (anterior basal ganglia [ABG] vs lobar) if the interim analysis suggested futility in continuing to enroll that group. At 175 patients, the interim analysis did suggest futility in the ABG cohort, leading to cessation of enrollment for these patients. This mid-trial adaptation limits the claim of the final outcome to lobar hemorrhages alone and does not conclusively inform us of the impact of MIPS on ABG hemorrhages.

Finally, the study used a Bayesian statistical analysis, presenting a posterior probability rather than the common P value. While unusual, each of these design features is evidence-based, broadly accepted, and should not limit acceptance of the trial results, although they may limit attempts to compare ENRICH’s results with those of other studies.

Numerous questions remain, for which the full manuscript is eagerly anticipated. Is there additional information that might explain why an effect was not seen in the ABG cohort? How often did rebleeding occur and was there any correlation between time to evacuation and rebleeding? Were there other subgroups that appeared to benefit more or less from the procedure? Was there a correlation between time to evacuation and procedural or clinical outcome measures? The anticipated economic outcomes will also help to promote or limit the broad application of this procedure, particularly in cost-conscious or resource-poor settings. Furthermore, numerous questions remain to be answered that ENRICH was not designed to address. How do results achieved with this minimally invasive technique compare with other techniques and devices? What role did the use of a parafascicular approach play in the outcome? How will these results affect the standard of care when surgical nihilism has dominated the care of patients with ICH for so many years?

In 2015, after the presentation of MR CLEAN, six additional randomized clinical trials were halted and rapidly published—all of which demonstrated the benefit of thrombectomy over medical management.6–12 That wave of trials differed from earlier negative studies by improved patient selection, earlier time to treatment, improved technology, and improved procedural proficiency. ENRICH has demonstrated that these differentiating features are as relevant to hemorrhagic stroke as they were to ischemic stroke 8 years ago. If hemorrhagic stroke treatment continues to recapitulate thrombectomy’s path, ENRICH’s success is not final, but rather the beginning of an era that brings increasingly effective ICH treatments.