Cardiovascular

Comparative external validation of ICH prognostic scores for discharge outcomes in the RES-Q multinational quality registry.

TL;DR

All three prognostic scores demonstrated acceptable-to-excellent discriminative ability for discharge outcomes following spontaneous ICH, with ICH Functional Outcome Score consistently achieving the highest discriminative performance across all endpoints.

Key Findings

The complete-case cohort comprised 8450 patients from 56 countries with a discharge mortality of 15.1% and favourable outcome (mRS 0-3) of 56.6%.

  • Median age was 63 (IQR 20) years and 62% were male.
  • Median National Institutes of Health Stroke Scale was 11 (IQR 12), Glasgow Coma Scale 14 (IQR 4), and median haematoma volume 16 (IQR 27) mL.
  • Supratentorial ICH occurred in 82.1%, intraventricular haemorrhage in 31.5%, and pre-admission oral anticoagulant use in 4.5%.
  • Data were drawn from consecutive adults with spontaneous ICH recorded in RES-Q between September 2024 and April 2025.

All three prognostic scores showed good discrimination for discharge mortality, with ICH-FOS achieving the highest AUC.

  • AUC for mortality was 0.82 (95% CI 0.81–0.83) for ICH Score, 0.81 (95% CI 0.79–0.82) for max-ICH, and 0.84 (95% CI 0.83–0.86) for ICH-FOS.
  • Global test across all three scores for mortality discrimination was statistically significant (P < .0001).
  • Pairwise DeLong contrasts were used to compare AUCs within the same complete-case cohort.

For unfavourable functional outcome (mRS ≥ 4), ICH-FOS again demonstrated the highest discriminative performance, followed by max-ICH.

  • AUCs for unfavourable outcome were 0.77 (95% CI 0.76–0.78) for ICH Score, 0.81 (95% CI 0.81–0.82) for max-ICH, and 0.84 (95% CI 0.83–0.85) for ICH-FOS.
  • Global test was statistically significant (P < .0001).
  • The ICH Score showed the lowest AUC for this endpoint, suggesting it has relatively weaker discrimination for functional outcomes compared to mortality.

For composite poor outcome (mRS 5–6), ICH-FOS achieved the highest AUC of 0.87, with all scores performing at their best across the three endpoints.

  • AUCs for composite poor outcome were 0.83 (95% CI 0.82–0.84) for ICH Score, 0.84 (95% CI 0.83–0.85) for max-ICH, and 0.87 (95% CI 0.86–0.88) for ICH-FOS.
  • Global test was statistically significant (P < .0001).
  • ICH-FOS exceeded an AUC of 0.85 for this endpoint, reaching the threshold commonly described as excellent discrimination.

ICH-FOS consistently achieved the highest discriminative performance across all three discharge outcome endpoints, while max-ICH provided a pragmatic and robust alternative based on routinely available clinical variables.

  • ICH-FOS outperformed both ICH Score and max-ICH for mortality, unfavourable outcome, and composite poor outcome.
  • max-ICH was characterized as a 'pragmatic and robust alternative based on routinely available clinical variables.'
  • Score distributions were characterized descriptively to explore differences in granularity among the three tools.
  • Areas under the curve were compared using global tests and pairwise DeLong contrasts in the same complete-case cohort to ensure methodological rigor.

The study population represented a large, contemporary, multinational real-world cohort, addressing a gap in direct comparative external validation of ICH prognostic scores.

  • Patients were drawn from 56 countries via the RES-Q International Registry of Stroke Care Quality.
  • The authors noted that 'direct comparative validation in contemporary, multinational real-world cohorts remains scarce.'
  • The study used consecutive adults with spontaneous ICH, supporting generalizability of findings.
  • Complete-case analysis was applied, meaning patients with missing data for any score variable were excluded from the primary comparative analysis.

What This Means

This research compared the accuracy of three scoring systems used to predict outcomes for patients who have had a spontaneous brain bleed (intracerebral haemorrhage, or ICH) — specifically, whether patients would survive, have a good functional recovery, or be severely disabled or die by the time they left the hospital. The three tools studied were the ICH Score, max-ICH, and the ICH Functional Outcome Score (ICH-FOS). Data came from over 8,400 patients across 56 countries, making this one of the largest and most geographically diverse validation studies of these tools to date. All three scoring systems performed reasonably well to excellently at predicting hospital discharge outcomes, but they were not equally accurate. The ICH-FOS consistently outperformed the other two scores across every outcome measured — death, severe disability, and a combined measure of the two. The ICH Score, the oldest and simplest tool, showed particularly weaker performance when predicting functional disability (as opposed to death alone). The max-ICH score, which relies only on information routinely collected at the bedside, performed nearly as well as ICH-FOS and was highlighted as a practical alternative for clinical settings. This research suggests that clinicians and health systems using older or simpler ICH prognostic tools may benefit from adopting more comprehensive scoring systems like ICH-FOS, especially when making decisions about functional recovery. At the same time, the finding that max-ICH performs robustly using only routinely available data points makes it an attractive option in resource-limited settings. The multinational nature of this study increases confidence that these findings are broadly applicable across different healthcare contexts worldwide.

Have a question about this study?

Citation

de Souza A, Neto G, Thang N, Suwanwela N, Nulkhasanah A, Collantes M, et al.. (2026). Comparative external validation of ICH prognostic scores for discharge outcomes in the RES-Q multinational quality registry.. European stroke journal. https://doi.org/10.1093/esj/aakag105