One-Year Outcome Trajectories and Factors Associated with Functional Recovery Among Survivors of Intracerebral and Intraventricular Hemorrhage With Initial Severe Disability

Key Points

Question  Is inclusion of hospital events, interventions, and responses to treatment in the first month after severe intracerebral hemorrhage (ICH) or intraventricular hemorrhage (IVH) associated with improved prediction of functional outcome trajectory among survivors with initial poor outcome?

Findings  In this post hoc longitudinal analysis of the Clot Lysis: Evaluating Accelerated Resolution of Intraventricular Hemorrhage phase 3 trial (CLEAR-III) and the Minimally Invasive Surgery Plus Alteplase for Intracerebral Hemorrhage Evacuation (MISTIE-III) phase 3 trial, more than 40% of patients with poor outcome at day 30 recovered to good outcome by 1 year. Inclusion of hospital events and hematoma volume reduction was significantly associated with enhanced discrimination of eventual functional recovery at 1 year.

Meaning  The findings indicate that hospital events significantly influenced long-term functional recovery after ICH and IVH, suggesting that aggressive ICH and IVH resolution and prevention of systemic complications and cerebral ischemic injury may promote better outcomes among patients with ICH and IVH with high clinical severity.

Importance  Patients who survive severe intracerebral hemorrhage (ICH) and intraventricular hemorrhage (IVH) typically have poor functional outcome in the short term and understanding of future recovery is limited.

Objective  To describe 1-year recovery trajectories among ICH and IVH survivors with initial severe disability and assess the association of hospital events with long-term recovery.

Design, Setting, and Participants  This post hoc analysis pooled all individual patient data from the Clot Lysis: Evaluating Accelerated Resolution of Intraventricular Hemorrhage phase 3 trial (CLEAR-III) and the Minimally Invasive Surgery Plus Alteplase for Intracerebral Hemorrhage Evacuation (MISTIE-III) phase 3 trial in multiple centers across the US, Canada, Europe, and Asia. Patients were enrolled from August 1, 2010, to September 30, 2018, with a follow-up duration of 1 year. Of 999 enrolled patients, 724 survived with a day 30 modified Rankin Scale score (mRS) of 4 to 5 after excluding 13 participants with missing day 30 mRS. An additional 9 patients were excluded because of missing 1-year mRS. The final pooled cohort included 715 patients (71.6%) with day 30 mRS 4 to 5. Data were analyzed from July 2019 to January 2022.

Exposures  CLEAR-III participants randomized to intraventricular alteplase vs placebo. MISTIE-III participants randomized to stereotactic thrombolysis of hematoma vs standard medical care.

Main Outcomes and Measures  Primary outcome was 1-year mRS. Patients were dichotomized into good outcome at 1 year (mRS 0 to 3) vs poor outcome at 1 year (mRS 4 to 6). Multivariable logistic regression models assessed associations between prospectively adjudicated hospital events and 1-year good outcome after adjusting for demographic characteristics, ICH and IVH severity, and trial cohort.

Results  Of 715 survivors, 417 (58%) were male, and the overall mean (SD) age was 60.3 (11.7) years. Overall, 174 participants (24.3%) were Black, 491 (68.6%) were White, and 49 (6.9%) were of other races (including Asian, Native American, and Pacific Islander, consolidated owing to small numbers); 98 (13.7%) were of Hispanic ethnicity. By 1 year, 129 participants (18%) had died and 308 (43%) had achieved mRS 0 to 3. In adjusted models for the combined cohort, diabetes (adjusted odds ratio [aOR], 0.50; 95% CI, 0.26-0.96), National Institutes of Health Stroke Scale (aOR, 0.93; 95% CI, 0.90-0.96), severe leukoaraiosis (aOR, 0.30; 95% CI, 0.16-0.54), pineal gland shift (aOR, 0.87; 95% CI, 0.76-0.99]), acute ischemic stroke (aOR, 0.44; 95% CI, 0.21-0.94), gastrostomy (aOR, 0.30; 95% CI, 0.17-0.50), and persistent hydrocephalus by day 30 (aOR, 0.37; 95% CI, 0.14-0.98) were associated with lack of recovery. Resolution of ICH (aOR, 1.82; 95% CI, 1.08-3.04) and IVH (aOR, 2.19; 95% CI, 1.02-4.68) by day 30 were associated with recovery to good outcome. In the CLEAR-III model, cerebral perfusion pressure less than 60 mm Hg (aOR, 0.30; 95% CI, 0.13-0.71), sepsis (aOR, 0.05; 95% CI, 0.00-0.80), and prolonged mechanical ventilation (aOR, 0.96; 95% CI, 0.92-1.00 per day), and in MISTIE-III, need for intracranial pressure monitoring (aOR, 0.35; 95% CI, 0.12-0.98), were additional factors associated with poor outcome. Thirty-day event-based models strongly predicted 1-year outcome (area under the receiver operating characteristic curve [AUC], 0.87; 95% CI, 0.83–0.90), with significantly improved discrimination over models using baseline severity factors alone (AUC, 0.76; 95% CI, 0.71-0.80; P < .001).

Conclusions and Relevance  Among survivors of severe ICH and IVH with initial poor functional outcome, more than 40% recovered to good outcome by 1 year. Hospital events were strongly associated with long-term functional recovery and may be potential targets for intervention. Avoiding early pessimistic prognostication and delaying prognostication until after treatment may improve ability to predict future recovery.

Intracerebral hemorrhage (ICH) prognostication is historically performed on admission, and most models predict short-term outcomes.^1-3 However, survivors of severe ICH often have initial considerable disability, with a median 90-day modified Rankin Scale score (mRS) of 5⁴ and limited understanding of future recovery. In a meta-analysis of population-based studies, 12% to 39% of ICH survivors gained functional independence beyond 1 year.⁵ However, most studies describing long-term outcomes in ICH are single-center studies^4,6-9 and rarely describe long-term recovery among ICH survivors with initial severe disability.⁶ For these patients and their caregivers, long-term prognostication and decision-making have important implications. Therefore, the goal of this study was to describe 1-year outcome trajectories among patients with ICH and severe disability (mRS 4 to 5) at day 30.

Most ICH prognostication models include baseline ICH severity factors,^1-3,10,11 not accounting for comorbidities, hospital interventions, and complications. Other factors of potential importance include leukoariosis,¹² temporal changes in hematoma volume,¹³ and intraventricular hemorrhage (IVH) severity.¹⁴ IVH grading scales often only incorporate baseline IVH-volume,^15-17 not accounting for IVH expansion,¹⁸ volume reduction, or hydrocephalus,¹⁹ which may also impact recovery. We hypothesized that including comorbidities, hospital procedures, medical complications, and temporal hematoma evolution would be significantly associated with improved prediction of 1-year recovery among survivors of ICH and IVH with initial disability compared with models using only baseline characteristics. Understanding the impact of hospital events on long-term recovery among ICH survivors might help identify important targets for interventions to improve ICH outcomes as well as assist in counseling patients and their families regarding potential for future recovery.

We conducted a post hoc longitudinal analysis of prospectively collected participant data from the Clot Lysis: Evaluating Accelerated Resolution of Intraventricular Hemorrhage phase 3 trial (CLEAR-III)²⁰ and the Minimally Invasive Surgery plus Alteplase for Intracerebral Hemorrhage Evacuation (MISTIE-III) trial.²¹ Parent trial protocols were approved by the institutional review boards of each trial site, and written informed consent was obtained prior to trial enrollment. This study was approved by our local institutional review board and performed in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

CLEAR-III randomized 500 patients with spontaneous obstructive IVH and small to moderate supratentorial ICH (<30 mL) to receive intraventricular thrombolysis vs placebo.²⁰ MISTIE-III randomized 499 patients with spontaneous large supratentorial ICH (≥30 mL) without obstructive IVH to receive either minimally invasive stereotactic thrombolysis of the hematoma or standard medical treatment.²¹ Both trials were neutral for primary end point of improved functional outcome at 180 days and 1 year, respectively. However, both trials demonstrated a significant reduction in mortality for the treatment vs control groups at 180 and 365 days. In the current study, we included all patients with ICH and IVH from both trials who survived and had mRS 4 to 5 at day 30.

Baseline characteristics included age, sex, race, ethnicity, stroke-related comorbidities, Glasgow Coma Scale score, and National Institutes of Health Stroke Scale score. Race and ethnicity were included to account for racial or ethnic variances in ICH outcomes and were recorded by trial investigators from review of medical records. Hematoma volumes were measured on noncontrast computed tomography scans performed upon admission, at ICH and IVH stability, at end of treatment, and at day 30 from enrollment in both clinical trials. End of treatment was defined as 24 hours after the last study treatment or placebo in CLEAR-III. In MISTIE-III, end of treatment time was defined as 24 hours after the last study treatment in the surgical arm, and the median time to end of treatment in the surgical arm was used as end of treatment for the medical arm. ICH volumes were calculated using semiautomated planimetry and read by a single neuroradiologist blinded to treatment and outcomes. IVH severity was graded using modified Graeb score in CLEAR-III.¹⁵ ICH resolution was defined as 0-mL ICH volume and IVH resolution as either 0–modified Graeb score or 0-mL IVH volume. Midline shift was measured at septum pellucidum and pineal gland. Obstructive hydrocephalus was defined as third ventricular obstruction by hematoma and/or IVH.²² Hydrocephalus was measured on admission and at day 30 via computed tomography using the Evans index in CLEAR-III.²³ Leukoaraiosis was graded on admission computed tomography by 2 independent reviewers (N.U. and B.M.H.) using the van Sweiten Scale²⁴ in CLEAR-III. In MISTIE-III, leukoaraiosis was graded on magnetic resonance imaging, using the Fazekas scale.²⁵ Severe leukoaraiosis was defined as van Sweiten Scale score 3 or higher, deep Fazekas scale score 2 to 3, and/or total Fazekas scale score greater than 3. In CLEAR-III, intracranial and cerebral perfusion pressure were measured every 4 hours after extraventricular drain insertion. Only 72 (14.4%) patients in MISTIE-III underwent intracranial pressure monitoring, and these data were excluded. Adverse events and hospital complications were adjudicated centrally during the treatment phase in both trials, and serious adverse events were also assessed by an independent safety committee until the end of follow-up.

The primary outcome measure was 1-year mRS. Blinded mRS assessments were performed in both trials at days 30, 180, and 365. Secondary outcomes included 1-year mortality, withdrawal of life-sustaining treatment, home discharge, and European Quality-of-Life Visual Analog Scale (EQ-VAS) score.²⁶ EQ-VAS score was self-reported by survivors at days 30, 180, and 365.

Patients were dichotomized into 2 groups based on 1-year mRS outcome: good (mRS 0 to 3) vs poor (mRS 4 to 6). Traditionally, mRS 0 to 2 is most closely associated with functional independence^27,28 and has been used previously to define good or excellent functional outcome in clinical trials. However, in this analysis, mRS 3 was also classified as good outcome, as both CLEAR-III²⁰ and MISTIE-III²¹ used this dichotomization to determine treatment efficacy. Additionally, more recent literature suggests that survivors of moderate to severe stroke or ICH survivors with mRS 2 vs mRS 3 report a similar quality of life and have identical direct valuation of their current health status in standard-gamble utility.^29-31 In addition, considerable interrater disagreement when classifying mRS 2 and 3³² suggests that these categories may have similar functional status.

Median time to home, death, and withdrawal of life-sustaining treatment were calculated using time-to-event analyses. mRS and EQ-VAS score trajectories were assessed using mixed-effects linear regression, treating mRS and EQ-VAS scores as outcome, day of assessment (30, 180, and 365) as fixed effects, and patients as random effects.

Factors associated with functional recovery at 1 year were assessed by bivariate analyses comparing all variables between good vs poor 1-year outcome groups in each trial and in the combined cohort. We used the Wilcoxon rank sum test or t test for continuous variables and χ² test or Fisher exact test for categorical variables. Multivariable logistic regression evaluated the association between variables of interest prior to day 30 and 1-year good outcome for each trial independently and for the combined cohort. Covariates were selected using a significance of P < .20 in bivariate analysis with backward elimination of covariates with P > .10 in final models. Age, sex, race, and ethnicity were chosen a priori as universal confounders and included in models regardless of P value. In models for each independent trial, we adjusted for assignment to treatment group and, for the combined cohort of patients, we adjusted for the trial cohort. Receiver operating characteristic curve and area under the receiver operating characteristic curve (AUC) were compared between full models and models using baseline predictors only (ie, age, Glasgow Coma Scale score, ICH volume, presence of IVH [IVH volume for CLEAR-III], and ICH location [deep location for MISTIE-III and thalamic location for CLEAR-III]). Ordinal logistic regression for 1-year mRS was not included in final results as models violated the proportional odds assumption in likelihood-ratio and Brant tests. In log-ordered analyses, covariates were assessed at each mRS threshold and remained similar, with a few covariates demonstrating variability in adjusted odds ratios across different mRS thresholds. Cox proportional hazards regression assessed factors associated with 1-year mortality. All patients surviving to 1 year were censored. Sensitivity analyses were conducted to assess impact of missing outcomes and data points. All statistical analyses were performed using Stata version 17.0 (StataCorp), and were 2-tailed with significance set at P ≤ .05.

Of 999 patients, 115 (11.5%) died and 147 (14.7%) had good outcome (mRS 0 to 3) at day 30, while 22 (2.2%) lacked mRS data. The final pooled cohort consisted of 715 patients with poor outcome (mRS 4 to 5) at day 30 (218 [29.5%] with mRS 4 and 497 [69.5%] with mRS 5) (Figure 1). The mean (SD) age was 60.3 (11.7) years, and 417 participants (58%) were male. Overall, 174 participants (24.3%) were Black, 491 (68.6%) were White, and 49 (6.9%) were of other races (including Asian, Native American, and Pacific Islander, consolidated owing to small numbers); 98 (13.7%) were of Hispanic ethnicity.

Of these, 308 patients (43%; 153 [45.1%] in CLEAR-III and 155 [41.2%] in MISTIE-III) recovered to good outcome at 1 year. Median (IQR) improvement in mRS was 1 (0-2). A 2-point improvement or greater occurred in 214 patients (30%) and a 1-point improvement occurred in 248 (34.7%) (Figure 2; eFigure 1 in the Supplement).

By 1 year, 462 patients (64.6%) had returned home at a median (IQR) of 98 (52-302) days postictus. Among 308 patients recovering to good outcome by 1 year, 294 (95.4%) returned home, as did 168 of 407 patients (41%) who had persistent poor outcome at 1 year (eFigures 2 and 3 in the Supplement). Differences by good and poor 1-year outcome groups are shown in Table 1 and eTable 1 in the Supplement.

Day 30 and 1-year mRS were missing in 13 and 9 patients, respectively. Patients lacking mRS data were younger, did not have diabetes or hyperlipidemia, and were less likely to have chronic hydrocephalus but were more likely to have new symptomatic intracranial hemorrhage within the first 30 days (eTable 2 in the Supplement). Adjusted linear predictions of mRS over 1 year for good vs poor 1-year outcome groups are shown in Figure 3. For good 1-year outcome group models, improvement in mean mRS was significant between all time points (day 30 vs day 180; day 180 vs day 365) in mixed linear models. Good 1-year outcome groups had a 2-point improvement in mean mRS (eFigure 4 in the Supplement).

In addition to 129 patients (18%) who died, 77 of 586 survivors (13.1%) did not report EQ-VAS score at follow-up (day 180 and/or day 365). These patients were more likely to have higher baseline National Institutes of Health Stroke Scale score, lower Glasgow Coma Scale score, severe leukoaraiosis, left-hemispheric lesions, and a worse clinical course with longer ventilatory support, need for gastrostomy, and persistent poor outcomes (eTable 3 in the Supplement). In mixed linear models, EQ-VAS score was significantly lower in the poor 1-year outcome group throughout follow-up, but both groups were associated with a statistically significant upward trajectory in EQ-VAS scores (Figure 3).

Covariates associated with recovery to good outcome in both trials are shown in Table 2. In the combined cohort of patients with ICH and IVH and poor outcome at day 30, diabetes (adjusted odds ratio [aOR], 0.50; 95% CI, 0.26-0.96), National Institutes of Health Stroke Scale score (aOR per 1 point, 0.93; 95% CI, 0.90-0.96), severe leukoaraiosis (aOR, 0.30; 95% CI, 0.16-0.54), pineal gland shift (aOR per 1 mm, 0.87; 95% CI, 0.76-0.99), persistent hydrocephalus at day 30 (aOR, 0.37; 95% CI, 0.14-0.98), new ischemic stroke (aOR, 0.45; 95% CI, 0.21-0.96), and gastrostomy (aOR, 0.30; 95% CI, 0.17-0.50) were associated with lack of recovery. Radiographic ICH (aOR, 1.83; 95% CI, 1.08-3.08) and IVH (aOR, 2.20; 95% CI, 1.03-4.70) resolution by day 30 were associated with recovery to good outcome after adjusting for baseline ICH and IVH characteristics, demographic characteristics, and trial cohort. In the CLEAR-III model, cerebral perfusion pressure less than 60 mm Hg (aOR, 0.30; 95% CI, 0.13-0.71), sepsis (aOR, 0.05; 95% CI, 0.00-0.80), prolonged mechanical ventilation (aOR, 0.96; 95% CI, 0.92-1.00, per -day), and in MISTIE-III, need for intracranial pressure monitoring (aOR, 0.35; 95% CI, [0.12-0.98]) were additional factors associated with lack of recovery to good outcome. The day 30 model for the combined cohort (AUC, 0.87; 95% CI, 0.83-0.90) showed significantly improved discrimination for 1-year outcome compared with the baseline model (AUC, 0.76; 95% CI, 0.71-0.80; P < .001). The CLEAR-III day 30 model (AUC, 0.90; 95% CI, 0.86-0.93) showed improved discrimination compared with the baseline model (AUC, 0.80; 95% CI, 0.75-0.85; P < .001). In MISTIE-III, the day 30 model (AUC, 0.87; 95% CI, 0.83-0.91; P < .001) vs the baseline model (AUC, 0.79; 95% CI, 0.74-0.84; P < .001) (eFigure 5 in the Supplement). Notably, 286 patients were excluded in the combined model because of missing data points. A comparison of these patients is included in eTable 4 in the Supplement.

Multivariable logistic regression for other mRS dichotomizations, including recovery to mRS 0 to 2 and mRS improvement of 1-point or greater, revealed similar covariates (eTables 5-7 in the Supplement), with a notable exception of ICH resolution by day 30, which was independently associated with recovery to mRS 0 to 3 but not with mRS 0to 2.

Seventy-seven patients in CLEAR-III (22.7%) and 52 in MISTIE-III (13.8%) with poor day 30 outcome died by 1 year (total deaths = 129 [18%]). Median (IQR) time to death was 85 (48-180) days postictus. Index ICH led to delayed deaths in 19 patients (15%). Twelve deaths (9%) occurred because of other neurological events (brain death, herniation, recurrent hemorrhage, or hydrocephalus) and 98 (76%) because of nonneurological causes. Withdrawal of life-sustaining treatment occurred in 40 patients (31%) at a median (IQR) time of 73 (25-188) days after day 30.

In Cox proportional hazards regression for the combined cohort, 1-year mortality was associated with age (hazard ratio [HR] per year, 1.05; 95% CI, 1.03-1.08), female sex (HR male, 0.63; 95% CI, 0.43-0.94), diabetes (HR, 1.75; 95% CI, 1.09-2.79), severe leukoaraiosis (HR, 1.94; 95% CI, 1.30-2.92), pulmonary edema (HR, 3.04; 95% CI, 1.44-6.42), sepsis (HR, 2.88; 95% CI, 1.02-8.09), and lack of IVH resolution (HR, 0.52; 95% CI, 0.30-0.91) by day 30 after adjusting for ICH severity (eTable 8 in the Supplement).

In sensitivity analyses conducted for missing outcomes, models were largely unchanged after accounting for patients lacking mRS data (eTables 9 and 10 in the Supplement). Most covariates were unchanged in sensitivity analysis for the combined model after including all patients, except for IVH volume and pineal shift, which were no longer statistically significant (eTable 11 in the Supplement). Similarly, covariates in the Cox regression model for 1-year mortality were also unchanged in sensitivity analysis (eTable 12 in the Supplement) (eFigures 6 and 7 in the Supplement).

In this post hoc analysis of patients with ICH and IVH and initial severe disability, more than 40% of patients recovered to good outcome by 1 year. Prediction of long-term functional recovery was significantly associated with improvement after inclusion of more granular baseline characteristics (ie, diabetes, leukoaraiosis, IVH volume, and pineal shift) and hospital events, including cerebral hypoperfusion, ischemic stroke, sepsis, persistent hydrocephalus, need for supportive interventions (mechanical ventilation and gastrostomy), and hematoma evolution. Significant reduction in hematoma and IVH volumes in the first month were strongly associated with long-term functional recovery, highlighting the importance of minimally invasive interventions to reduce blood volume.

The combined cohort experienced higher rates of good 1-year outcome than previously reported,^2,33 despite the inclusion of only patients with high-severity ICH and IVH. Outcomes from these clinical trials may be biased by rigid exclusion criteria, including baseline Glasgow Coma Scale score less than 4, fixed pupils, infratentorial extension, and baseline mRS greater than 1, limiting generalizability. However, outcomes were less impacted by early withdrawal of life-sustaining treatment, which occurred in only 88 of 999 patients (8.8%) in the first month, compared with 20% to 30% in prior studies.^10,33 Patients with ICH who are maximally treated have favorable outcomes at a similar rate,¹⁰ highlighting the possible negative impact of early pessimistic prognostication on future recovery.³⁴

Most ICH outcome prognostication scales only include admission characteristics and typically predict short-term outcomes.^2,3 Previous prediction scores have not integrated hospital events and therapeutic interventions that also impact outcome. Thus, existing scales have lower discrimination for 1-year outcome among patients with ICH who receive maximal treatment.¹⁰ Hence, by delaying prognostication until after the treatment phase and incorporating hospital events, our study provides a novel approach to ICH outcome prognostication, emphasizing the need to promote maximal treatment of patients with ICH and IVH. These findings also identify targets for therapeutic interventions in the short term to promote long-term recovery.

Among comorbidities, type 2 diabetes was associated with persistent poor 1-year outcome. No associations were noted between diabetes and short-term functional outcomes in prior community-based stroke registry studies,^35,36 but diabetes is associated with long-term mortality³⁷ and long-term functional decline among ICH survivors.³⁸ Diabetes is associated with recurrent ischemic stroke in ICH survivors,³⁹ which may explain the association between ICH and poor long-term outcome. Severe leukoaraiosis was also associated with persistent poor outcome at 1 year in this analysis, further supporting its association with short-term outcomes after ICH.^12,40

Among hospital events, need for intracranial pressure monitoring (MISTIE-III) and cerebral perfusion pressure less than 60 mm Hg (CLEAR-III) were associated with persistent poor outcome. In CLEAR-III, any cerebral perfusion pressure less than 60 mm Hg negatively influenced functional recovery⁴¹ even though intracranial pressure elevation did not, similar to findings from a prior MISTIE-III analysis, which also reported cerebral perfusion pressure as an independent predictor of poor long-term outcome.⁴² This suggests that cerebral hypoperfusion may be more critical than intracranial hypertension, raising concerns about aggressive blood pressure reduction in patients with large hematomas.^43-45 Ischemic stroke in the first month was also associated with lack of functional recovery, which is associated with leukoaraiosis, high-volume hemorrhage, and rapid blood-pressure reduction.^44,45 Thus cerebral ischemic injury prior to and following ICH appears to be an important contributor to persistent poor long-term outcomes among survivors³⁸ and may be driven by underlying comorbidities (eg, diabetes and leukoaraiosis) and secondary injury due to cerebral hypoperfusion.⁴⁶

Gastrostomy and prolonged ventilatory support were independently associated with poor 1-year outcome, even after adjusting for ICH and IVH severity. This may suggest that other factors driving the need for prolonged supportive care, such as nosocomial infections,⁴⁷ may also contribute to poor functional recovery. Specifically, need for supportive interventions may indicate impaired protective (eg, cough and gag) reflexes, which are associated with increased long-term mortality in general stroke populations.⁴⁸ Among medical complications, sepsis and pulmonary edema were associated with persistent poor long-term outcome and reduced long-term survival after ICH, confirming that hospital complications impact recovery well beyond hospital discharge.^49-52

Association of intraventricular thrombolysis with persistent poor outcomes in CLEAR-III likely reflects a higher proportion of patients with mRS 5 in the alteplase group (alteplase, 17% vs placebo, 9%; P = .007).²⁰ However, after adjusting for alteplase use, complete resolution of IVH and ICH by day 30 was independently associated with functional recovery at 1 year. IVH clearance was also associated with improved long-term survival. These findings support associations between hematoma volume reduction and 180-day good outcomes in both CLEAR-III²⁰ and MISTIE-III,²¹ consistent with the established effects of alteplase on mortality and supporting the need for continued efforts to develop interventions that will assist in timely reduction of IVH and ICH volumes.

Despite a significantly lower quality of life reported among patients who did not functionally recover by 1 year, there was a statistically significant upward trajectory in EQ-VAS scores of these patients throughout the one-year follow-up period.⁵³ Additionally, nearly half of the survivors with persistent poor 1-year outcome were able to return home. These findings emphasize that an acceptable quality of life may be achieved even among survivors who do not make good functional recovery. The strengths of our study include a large cohort of prospectively monitored patients with high-severity ICH or IVH, well-defined inclusion criteria, preestablished time points for neuroimaging, adjudicated and prospectively collected hospital events, prolonged follow-up, and serial blinded mRS and quality-of-life assessments.

There are several limitations to this study warranting discussion. First, generalizability of the findings may be limited owing to the strict inclusion criteria of the clinical trials. Second, testing associations between a large number of covariates and 1-year outcomes may have been impacted by type I error. While we demonstrated that 30-day event-based models performed better than admission criteria–based models, we acknowledge that their performance was not compared with that of models using factors at other time points between admission and day 30. Also, given that patients were enrolled over an 8-year period, temporal advances in ICH management could have influenced outcomes through the study period. Additionally, combining CLEAR-III and MISTIE-III cohorts has limitations owing to significant differences in patient characteristics and trial interventions. However, to address this we adjusted for trial cohort in the combined model and created separate models for each trial. Missing variables led to exclusion of many patients in the multivariable analysis, especially for the combined cohort. Thus, the analysis may have been affected by an omitted-variable bias. However, in sensitivity analysis, most covariates remained unchanged after including all patients. Similarly, 2.2% of patients were lost to follow-up owing to missing mRS data. Quality-of-life data may have been impacted by a reporting bias, given that patients that did not report EQ-VAS score, had a worse clinical course, and had more left hemispheric lesions, suggesting that aphasia and poor functional and cognitive recovery may have impaired their ability to self-report quality of life. Ascertainment of leukoaraiosis in CLEAR-III and MISTIE-III used different methods; however, significant association between van Sweiten Scale and Fazekas scale has been reported previously.^54,55 Owing to these limitations, models from this post hoc analysis of clinical trials need independent prospective validation in external data sets.

In this cohort study of patients with high-severity ICH and IVH enrolled in clinical trials, nearly half of survivors with poor outcome in the early phase recovered to a good functional outcome by 1 year. Discrimination of an upward trajectory compared with stable or downward trajectory was associated with significant improvement when including factors not included in traditional models, such as preexisting conditions, hospital events, and responses to therapy. These findings support longer evaluation periods to provide better communication about long-term recovery after severe ICH and IVH and identify important targets of improvement in ICH care that may enhance long-term recovery.

Accepted for Publication: May 13, 2022.

Published Online: July 25, 2022. doi:10.1001/jamaneurol.2022.1991

Corresponding Author: Wendy C. Ziai, MD, MPH, Division of Neurocritical Care, Department of Neurology, Johns Hopkins University School of Medicine, 600 N Wolfe St, Phipps 455, Baltimore, MD 21287 (weziai@jhmi.edu).

Author Contributions: Drs Shah and Ziai had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Shah, Awad, Hanley, Ziai.

Acquisition, analysis, or interpretation of data: Shah, Thompson, Yenokyan, Acosta, Avadhani, Dlugash, McBee, Li, Hansen, Ullman, Falcone, Awad, Ziai.

Drafting of the manuscript: Shah, Dlugash, Ziai.

Critical revision of the manuscript for important intellectual content: Shah, Thompson, Yenokyan, Acosta, Avadhani, McBee, Li, Hansen, Ullman, Falcone, Awad, Hanley, Ziai.

Statistical analysis: Shah, Thompson, Yenokyan, Acosta, Avadhani, Dlugash, Li, Ziai.

Obtained funding: Awad.

Administrative, technical, or material support: Acosta, Li, Ullman, Hanley.

Supervision: Awad, Ziai.

Conflict of Interest Disclosures: Dr Shah reported serving on the editorial board for the Neurohospitalist Journal. Dr Thompson reported grants from Johns Hopkins University during the conduct of the study. Dr Yenokyan reported grants from National Institute of Neurological Disorders and Stroke during the conduct of the study and grants from National Center for Advancing Translational Sciences outside the submitted work. Dr Acosta reported being employed by Rad AI outside the submitted work. Dr McBee reported grants from National Institute of Neurological Disorders and Stroke during the conduct of the study. Dr Awad reported grants from National Institute of Neurological Disorders and Stroke during the conduct of the study and grants from Be Brave for Life Foundation and StrideBio, Inc; serving as a consultant and advisory board chairman for Neurelis and Recursion Pharma; and personal fees from Medicolegal for consulting outside the submitted work. Dr Hanley reported grants from National Institute of Neurological Disorders and Stroke (MISTIE III and CLEAR III grants) during the conduct of the study and grants from the Department of Defense personal fees for medicolegal consulting from Neurotrope and Neurelis outside the submitted work. Dr Ziai reported grants from National Institutes of Health during the conduct of the study and personal fees from C. R. Bard DMC and personal fees from Neurocritical Care Associate Editor outside the submitted work. No other disclosures were reported.

Additional Information: Individual deidentified participant data from CLEAR-III and MISTIE-III may be obtained by submitting a formal proposal to the trial steering committees or to the National Institute of Neurological Disorders and Stroke trial repository.