Intracerebral haemorrhage laterality and associations with mood and pain: a MISTIE III substudy Authors: Dawson C Cooper, Efrat Abramson, Wendy C Ziai, Matthew L Flaherty, Vishank A Shah, Radhika Avadhani, Noeleen Ostapkovich, Lourdes Carhuapoma, Issam Awad, Mario Zuccarello, Daniel Hanley, Lauren H Sansing Published: 2025 ABSTRACT AbstractBackgroundThe prognostic significance of the affected hemisphere in haemorrhagic stroke remains uncertain. We aimed to determine the relationship between the affected hemisphere (right or left) and differences in non-motor outcomes, including mood and pain, in patients with acute, supratentorial intracerebral haemorrhage (ICH). These non-motor outcomes are often overlooked in studies following ICH but impact patient recovery and well-being.MethodsA secondary prespecified analysis of the Minimally Invasive Surgery with Thrombolysis in Intracerebral Hemorrhage Evacuation (MISTIE) III study—a randomised, international, multicentre, placebo-controlled trial of participants with spontaneous, non-traumatic, supratentorial ICH of 30 mL or more that evaluated minimally invasive surgery with thrombolysis compared with standard medical care. Outcomes included EQ-5D three-level version (EQ-5D-3L, composite and individual non-motor components) and modified Rankin scale (mRS) scores at days 30, 180 and 365 post-ICH.ResultsA total of 493 participants were eligible for analysis at day 30 following ICH. In multivariable analyses, patients with right hemispheric ICH were more likely to report problems with pain and discomfort at days 30 (β=0.257 (95% CI 0.131, 0.383)), 180 (β=0.213 (95% CI 0.090, 0.336)) and 365 (β=0.209 (95% CI 0.090, 0.328)) post-ICH. Patients with right hemispheric ICH were also more likely to report problems with anxiety and depression at days 30 (β=0.160 (95% CI 0.030, 0.291)) and 180 (β=0.171 (95% CI 0.049, 0.293)) following ICH. There were no differences in mRS scores between patients with left or right-sided haemorrhages.ConclusionsRight hemispheric lesions were associated with increased reports of mood-related symptoms (depression, anxiety) and pain in patients with acute ICH over time.Trial registration numberNCT01827046.Keywords: Stroke, Hemorrhage WHAT IS ALREADY KNOWN ON THIS TOPIC WHAT THIS STUDY ADDS Through our comprehensive prespecified secondary analysis, we found that right-sided ICH was associated with increased reported problems with non-motor outcomes, including pain and mood-related symptoms and that these outcomes change dynamically over time. Through our comprehensive prespecified secondary analysis, we found that right-sided ICH was associated with increased reported problems with non-motor outcomes, including pain and mood-related symptoms and that these outcomes change dynamically over time. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These findings emphasise the importance of screening patients for non-motor sequelae following ICH, including for mood disturbances and pain at multiple time points, and that these changes may be more common in those with right-sided lesions. These findings emphasise the importance of screening patients for non-motor sequelae following ICH, including for mood disturbances and pain at multiple time points, and that these changes may be more common in those with right-sided lesions. Introduction Clinical consequences differ between stroke affecting the left or right hemisphere, owing to the functional lateralisation of the human brain. With respect to intracerebral haemorrhage (ICH), differences in motor and non-motor outcomes following stroke in the right versus left hemisphere remain unclear. Several studies report worse outcomes for those with right-sided haematomas,1,3 one study found the opposite,4 and the majority of studies report no significant difference.5,12 The variation in findings investigating laterality could be due to several factors including varied definitions of ICH location,5 non-standardised outcome measures across studies examining laterality and the reporting of clinical outcomes at different time points following ICH. Importantly, current outcome measures (eg, modified Rankin scale (mRS) and Glasgow Outcome Scale (GOS)) in studies examining laterality often focus on motor aspects of function, and neither of these measures explicitly consider traditionally lateralised functions when determining scores. While these functions, including aphasia and neglect, affect recovery and independence, the greater emphasis on motor-based function and functional independence with activities of daily living could explain the lack of significant results thus far. Indeed, a study using cognitive measures has suggested a role for lateralisation, with those with right-sided haemorrhages scoring lower on a figure-ground discrimination task compared to those with left-sided lesions.1 Because these additional domains can be further localised, they may provide valuable insight in prognosis.13 Impairments in mood are some of the most common non-motor sequelae of stroke, with 40%–75% of patients experiencing some degree of deficit.14 15 Several factors including female sex, lesion site, ICH volume and medial temporal lobe atrophy predict the degree of impairment.16 Other factors including aphasia, hypertension, atrial fibrillation and pre-existing neurodegenerative pathology are also predictive of the degree of deficit.17 18 The severity of impairment may also change over time, though longitudinal data is sparse and lacks consistency regarding clinical outcome measures across studies.19 Furthermore, lack of early recognition of non-motor deficits in patients following stroke can significantly affect subsequent recovery.20 Thus, early identification and refinement of our ability to anticipate those who may be more likely to develop impairments in mood and pain is prudent in improving outcomes following ICH. The current study aimed to investigate the relationship between ICH laterality and non-motor outcomes, including mood and pain, in patients with acute supratentorial ICH, represented by data from the Minimally Invasive Surgery with Thrombolysis in Intracerebral Hemorrhage Evacuation (MISTIE-III) clinical trial.21 Additionally, we sought to determine both early and late changes in these outcomes and their relation to ICH hemisphere, thus providing a longitudinal analysis of patient outcomes. We aimed to test the hypothesis that right-sided ICH would be associated with worse non-motor outcomes. Methods Study design We performed a secondary analysis of patients enrolled in the MISTIE III trial, an international, randomised, controlled, open-label, phase 3 clinical trial that evaluated minimally invasive surgery (MIS) with thrombolysis (n=250) compared with standard medical care (n=249). Detailed inclusion and exclusion criteria and study design are published.21 Briefly, patients aged 18 years or older with spontaneous, non-traumatic, supratentorial ICH of 30 mL or more with measurable clinical symptoms and minimal premorbid disability were randomised to MIS with intrahaematomal alteplase injections until haematoma volume was reduced to ≤15 mL (or a maximum of 9 doses) or standard medical care. The study is registered with ClinicalTrials.gov, NCT01827046. Data collection and statistical analyses All data were collected prospectively by blinded observers in the MISTIE III trial. The primary variable of interest was ICH laterality, and primary outcomes in this analysis included EQ-5D three-level version (EQ-5D-3L, composite and individual components) at 30, 180 and 365 days post-ICH. The EQ-5D-3L questionnaire comprises five dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression22 and is well established as a reliable measure following stroke.23 24 Each dimension has three levels including no problems, some problems and extreme problems. The patient is asked to indicate their health status for each dimension, resulting in a one-digit number that expresses the level for that dimension. The digits for the five dimensions were combined into a single composite score between 0 and 10, with higher scores indicating increased problems across dimensions. For comparison to motor and global functional outcomes, we analysed mRS scores at the same time points. Predefined clinical covariates included gender, race (black, white, Asian and other), ethnicity (Hispanic or non-Hispanic), premorbid mRS, randomisation group (MIS or medical arm) and ICH severity index. The ICH severity index had been previously developed by the MISTIE investigators and shown to inversely predict mRS; it is a composite score of age, Glasgow Coma Scale (GCS) at randomisation, ICH location (deep or lobar), ICH volume at stability scan prior to randomisation (dichotomised at <45 mL), the presence of intraventricular haemorrhage, medical history of diabetes and Fazekas score.21 Normally distributed variables were reported as means with interquartile ranges, while non-normally distributed variables were categorised. This categorisation included GCS as mild (13–15), moderate (9–12) or severe (3–8). Student’s t-test for continuous variables and χ2 test for categorical variables were performed for univariate analyses. The association of ICH laterality with clinical outcomes was estimated using multiple linear regression models for EQ-5D-3L and mRS scores at days 30, 180 and 365 post-ICH with the clinical covariates above. This approach allowed for statistical standardisation among all outcome measures, improving comparison across the analyses. In an exploratory analysis, we repeated the same models at days 30 and 180 but excluded patients with severe aphasia, defined by NIHSS subscore 9 being >1. Statistical analyses were performed using GraphPad Prism (V.10.1.1 for Mac, GraphPad Software, Boston, Massachusetts USA, www.graphpad.com) and R (V.4.0.1, http://www.r-project.org). The significance level for all analyses was set at p<0.05. Results Baseline characteristics Of the 499 randomised patients, 6 (1.2%) were excluded from analysis across all outcomes at day 30 due to loss of follow-up or declining further participation. Remaining patients were eligible for analysis if they had mRS and/or EQ-5D-3L scores (including composite and/or subcomponents) at any of the three time points beginning at day 30 following ICH. At day 30 post-ICH, 493 participants had an mRS score and 380 had EQ-5D-3L scores. There was a reduction in the number of completed clinical outcome measures with time following ICH, resulting in an unequal distribution of participants for each reported clinical measure. However, no significant differences were found with respect to the number of participants with left versus right hemispheric haemorrhages among clinical outcome measures at all time points (online supplemental table 1). Due to the established responsiveness of the EQ-5D-3L in stroke and our primary focus on non-motor outcomes, we report demographic characteristics of participants who had completed this measure across all study time points (table 1). Table 1Demographic characteristics of participants with EQ-5D-3L scores at days 30, 180 and 365 post-ICHCharacteristics of participants with EQ-5D-3L scoresDay 30 (n=380)Day 180 (n=381)Day 365 (n=362)Right ICH, n (%)201 (52.9)199 (52.2)187 (51.7)Surgical treatment arm, n (%)205 (53.9)205 (53.8)192 (53.0)Demographics Age at consent (years), median (IQR)62 (19.0)61 (19.0)60.5 (18.8) Female sex, n (%)147 (38.7)147 (38.6)138 (38.1) Ethnicity: Hispanic or Latino, n (%)46 (12.3)46 (12.1)45 (12.4)Race, n (%) White288 (76.0)272 (71.6)265 (73.2) Black59 (15.6)73 (19.2)66 (18.2) Asian26 (6.9)29 (7.6)26 (7.2) Other6 (1.6)6 (1.6)5 (1.4)Comorbidities, n (%) On anticoagulants21 (5.5)21 (5.5)20 (5.5) Diabetes104 (27.4)100 (26.2)94 (26.0) Hyperlipidaemia148 (38.9)138 (36.2)129 (35.6) Cardiovascular disease49 (12.9)46 (12.1)45 (12.4) Hypertension376 (96.6)368 (96.6)349 (96.4) Prior haemorrhage23 (6.1)18 (4.7)13 (3.6)Prerandomisation CT findings Stability ICH volume (mL), median (IQR)43.5 (20.5)43.3 (20.3)43.4 (19.8) Stability IVH present, n (%)149 (40.8)149 (40.8)145 (41.8) ICH deep location, n (%)228 (60.0)236 (61.9)224 (61.9)ICH location, n (%) Basal ganglia218 (57.4)226 (59.6)215 (59.4) Frontal lobe70 (18.4)69 (18.1)63 (17.4) Occipital lobe9 (2.4)7 (1.8)7 (1.9) Parietal lobe44 (11.6)44 (11.5)43 (11.9) Temporal lobe29 (7.6)25 (6.6)25 (6.9) Thalamus10 (2.6)10 (2.6)9 (2.5)Other findingsGCS at randomisation, n (%) Mild (13–15)133 (35.0)133 (34.9)129 (35.6) Moderate (9–12)169 (44.5)166 (43.6)155 (42.8) Severe (3–8)78 (20.5)82 (21.5)78 (21.5) Severity index, median (IQR)3.40 (1.9)3.9 (1.8)3.4 (1.8) NIHSS score median (IQR)12.0 (10.0)7.93 (9.0)6.0 (7.0)EQ-5D-3L, 3-Level Version of EQ-5D; GCS, Glasgow Coma Scale; ICH, intracerebral haemorrhage; IVH, intraventricular haemorrhage; NIHSS, National Institutes of Health Stroke Scale. Demographic characteristics of participants with EQ-5D-3L scores at days 30, 180 and 365 post-ICH EQ-5D-3L, 3-Level Version of EQ-5D; GCS, Glasgow Coma Scale; ICH, intracerebral haemorrhage; IVH, intraventricular haemorrhage; NIHSS, National Institutes of Health Stroke Scale. Differences between those with left- and right-sided haemorrhages who had completed the EQ-5D-3L were compared at day 30 post-ICH (table 2). Demographics of those excluded from this analysis due to incomplete outcome variable data can be found in online supplemental table 2. Of the 380 patients who completed the EQ-5D-3L at day 30, 201 (52.9%) had a right hemispheric ICH. Demographic variables including age, sex, ethnicity and race were similar among patients with left or right-sided ICH. Prerandomisation CT findings revealed similar results between hemispheric groups including location (deep vs lobar) and stability ICH volumes. Patients with right-sided ICH had higher GCS scores at randomisation among all subcategories and lower National Institutes of Health Stroke Scale (NIHSS) scores at randomisation compared with those with left-sided ICH, consistent with the prevalence of lateralised language function, leading to higher deficits on these scales. When controlling for aphasia (NIHSS subscore 9), NIHSS scores did not significantly differ between those with right versus left hemisphere ICH (mean difference 0.87, 95% (CI −0.46, 2.21); p=0.198), supporting overall similar severity between the two groups. Table 2Univariate associations of left versus right hemisphere ICH among participants with EQ-5D-3L scores at day 30CharacteristicsLeft hemisphere ICH(n=179, 47.1%)Right hemisphere ICH(n=201, 52.9%)P valueSurgical treatment arm93 (52.0)112 (55.7)0.462Demographics Age at consent (years), median (IQR)62 (18.5)61 (19.0)0.855 Female sex, n (%)69 (38.5)78 (38.8)0.959 Ethnicity: Hispanic or Latino, n (%)22 (12.3)24 (11.9)0.917Race, n (%) White132 (74.2)156 (77.6)0.277 Black27 (15.2)32 (15.9) Asian14 (7.9)12 (6.0) Other5 (2.8)1 (0.5)Comorbidities, n (%) On anticoagulants7 (3.9)14 (7.0)0.193 Diabetes50 (27.9)54 (26.9)0.816Hyperlipidaemia79 (44.1)69 (34.3)0.050 Cardiovascular disease21 (11.7)28 (13.9)0.523 Hypertension172 (96.1)195 (97.0)0.620 Prior haemorrhage11 (6.1)12 (6.0)0.943Prerandomisation CT findings Stability ICH volume (mL), median (IQR)43.4 (20.6)43.5 (19.8)0.374 Stability IVH present, n (%)72 (41.9)77 (39.9)0.703 ICH deep location, n (%)105 (58.7)123 (61.2)0.615ICH location, n (%) Basal ganglia100 (55.9)118 (58.7)0.181 Frontal lobe29 (16.2)41 (20.4) Occipital lobe2 (1.1)7 (3.5) Parietal lobe25 (14.0)19 (9.5) Temporal lobe18 (10.1)11 (5.5) Thalamus5 (2.8)5 (2.5)Other findingsGCS at randomisation Mild (13–15)32 (17.9)101 (50.2)<0.001 Moderate (9–12)103 (57.5)66 (32.8) Severe (3–8)44 (24.6)34 (16.9) Severity index, median (IQR)3.30 (1.90)3.50 (1.82)0.623 NIHSS score median (IQR)14.0 (10.3)11.0 (9.5)0.002EQ-5D-3L, 3-Level Version of EQ-5D; GCS, Glasgow Coma Scale; ICH, intracerebral haemorrhage; IVH, intraventricular haemorrhage; NIHSS, National Institutes of Health Stroke Scale. Univariate associations of left versus right hemisphere ICH among participants with EQ-5D-3L scores at day 30 EQ-5D-3L, 3-Level Version of EQ-5D; GCS, Glasgow Coma Scale; ICH, intracerebral haemorrhage; IVH, intraventricular haemorrhage; NIHSS, National Institutes of Health Stroke Scale. Right hemispheric ICH and EQ-5D-3L scores In the models examining composite EQ-5D-3L scores, right-sided ICH was significantly associated with an increased score at days 30 (β=0.561 (95% CI 0.160, 0.963)) and 180 (β=0.780 (95% CI 0.346, 1.21)). Factors associated with an increased and decreased odds of higher composite scores across all time points included black race and severity index, respectively (online supplemental table 3). Individual components of the EQ-5D-3L were also analysed. The prevalence of individuals reporting some or extreme problems with anxiety and depression decreased with time, with 57.4% at day 30, 51.5% at day 180 and 48.1% at day 365 post-ICH. Right-sided ICH was associated with a higher odds of reported problems with anxiety and depression at days 30 (β=0.160 (95% CI 0.030, 0.291)) and 180 (β=0.171 (95% CI 0.049, 0.293)). Female sex was also associated with higher odds of reported problems with anxiety and depression at days 180 (β=0.128 (95% CI 0.002, 0.254)) and 365 (β=0.156 (95% CI 0.027, 0.285)) post-ICH. The full results of the anxiety and depression subcomponent of the EQ-5D-3L across all time points are shown in table 3. In contrast to the anxiety and depression dimension, the prevalence of individuals reporting some or extreme problems with pain and discomfort varied across time with 58.4% at day 30, 60.9% at day 180 and 50.8% at day 365. Patients with right-sided ICH had increased odds of reported problems with pain and discomfort at all time points. The results of the pain and discomfort subcomponent of the EQ-5D-3L at all time points are shown in table 4. Table 3Multiple linear regression results of EQ-5D-3L anxiety and depression subscore at days 30, 180, 365VariableEstimateSE95% CIP valueDay 30 Intercept0.7240.108(0.512, 0.935)<0.0001 Surgical group−0.0550.067(−0.187, 0.077)0.411 Female0.0500.069(−0.086, 0.185)0.471 Hispanic ethnicity0.0570.105(−0.148, 0.263)0.584 Race: black0.1100.094(−0.075, 0.294)0.244 Race: Asian−0.1470.133(−0.408, 0.115)0.271 Race: other0.1040.271(−0.428, 0.637)0.701 Severity Index−0.0480.025(−0.097, 0.002)0.061 Historic Rankin0.2830.137(0.014, 0.552)0.040 Right hemisphere0.1600.067(0.030, 0.291)0.016Day 180 Intercept0.5130.105(0.307, 0.719)<0.0001 Surgical group−0.0440.062(−0.167, 0.079)0.480 Female0.1280.064(0.002, 0.254)0.047 Hispanic ethnicity0.2700.098(0.077, 0.462)0.006 Race: black0.0440.081(−0.115, 0.203)0.585 Race: Asian0.00040.119(−0.234, 0.234)0.997 Race: other−0.0050.253(−0.502, 0.492)0.984 Severity Index−0.0330.024(−0.080, 0.014)0.170 Historic Rankin0.3100.136(0.043, 0.577)0.023 Right hemisphere0.1710.062(0.049, 0.293)0.006Day 365 Intercept0.5050.107(0.295, 0.715)<0.0001 Surgical group−0.0540.064(−0.179, 0.071)0.393 Female0.1560.066(0.027, 0.285)0.018 Hispanic ethnicity0.2790.098(0.086, 0.473)0.005 Race: black0.1260.084(−0.040, 0.292)0.137 Race: Asian−0.0640.125(−0.310, 0.182)0.609 Race: other0.0110.273(−0.527, 0.549)0.967 Severity Index−0.0320.025(−0.081, 0.016)0.194 Historic Rankin0.2250.143(−0.056, 0.510)0.116 Right hemisphere0.0890.064(−0.036, 0.213)0.164EQ-5D-3L, 3-Level Version of EQ-5D. Multiple linear regression results of EQ-5D-3L anxiety and depression subscore at days 30, 180, 365 EQ-5D-3L, 3-Level Version of EQ-5D. Table 4Multiple linear regression results of EQ-5D-3L pain and discomfort subscore at days 30, 180, 365VariableEstimateSE95% CIP valueDay 30 Intercept0.5010.104(0.297, 0.704)<0.0001 Surgical group−0.0310.065(−0.159, 0.096)0.628 Female0.1030.066(−0.027, 0.234)0.120 Hispanic ethnicity−0.2310.101(−0.429, 0.033)0.022 Race: black0.1920.090(0.015, 0.370)0.034 Race: Asian0.1380.128(−0.114, 0.390)0.283 Race: other0.1100.261(−0.403, 0.623)0.673 Severity Index−0.0030.024(−0.051, 0.045)0.901 Historic Rankin0.2080.132(−0.051, 0.468)0.115 Right hemisphere0.2570.064(0.131, 0.383)<0.0001Day 180 Intercept0.6730.106(0.465, 0.880)<0.0001 Surgical group0.0200.063(−0.104, 0.144)0.754 Female0.1380.065(0.011, 0.265)0.033 Hispanic ethnicity0.0310.099(−0.163, 0.225)0.756 Race: black0.2650.081(0.105, 0.425)0.001 Race: Asian0.1140.120(−0.122, 0.350)0.343 Race: other0.1930.255(−0.308, 0.6940.449 Severity Index−0.0700.024(−0.117, 0.022)0.004 Historic Rankin0.2420.137(−0.027, 0.511)0.077 Right hemisphere0.2130.063(0.090, 0.336)0.0008Day 365 Intercept0.5950.102(0.395, 0.796)<0.0001 Surgical group−0.0090.061(−0.128, 0.110)0.885 Female0.0910.063(−0.032, 0.214)0.146 Hispanic ethnicity−0.0250.094(−0.210, 0.159)0.786 Race: black0.1510.080(−0.007, 0.309)0.062 Race: Asian−0.1030.119(−0.337, 0.132)0.390 Race: other0.3210.261(−0.192, 0.833)0.219 Severity Index−0.0650.024(−0.111 to 0.019)0.006 Historic Rankin0.3060.136(0.039, 0.574)0.025 Right hemisphere0.2090.061(0.090, 0.328)0.0006EQ-5D-3L, 3-Level Version of EQ-5D. Multiple linear regression results of EQ-5D-3L pain and discomfort subscore at days 30, 180, 365 EQ-5D-3L, 3-Level Version of EQ-5D. We then sought to determine whether the association of right-sided ICH and anxiety, depression and pain was confounded by patients with aphasia from left-sided ICH not being able to report these symptoms and caregivers not recognising or reporting them. The proportion of patients with EQ-5D-3L scores at day 30 post-ICH reported by proxy versus patient did differ significantly by hemisphere χ2 (1, n=312)=16.6, p<0.001, as expected with the lateralisation of language. However, the proportion of reported problems with pain/discomfort or anxiety/mood at day 30 following ICH did not significantly differ by whether the answers were reported by the patient or a proxy, with pain χ2 (1, n=377)=1.6, p=0.449 and anxiety χ2 (2, n=377)=2.57, p=0.276. We then further explored this by repeating the multivariable analyses on anxiety/depression and pain for days 30 and 180 but with the exclusion of patients with severe aphasia (NIHSS subscore 9 of >1). Right-sided ICH remained significantly associated with both anxiety and depression and with pain (online supplemental table 4). Right hemispheric ICH and mRS scores In the multiple linear regression model examining factors associated with mRS scores, right hemispheric ICH was not significantly associated with more severe scores at any time point. However, factors associated with worse mRS scores at day 30 were Hispanic ethnicity (β=0.27 (95% CI 0.05, 0.48)) and black race (β=0.20 (95% CI 0.002, 0.398)). Severity index (β=−0.344 (95% CI −0.388, –0.279)) and surgical group (β=−0.164 (95% CI −0.310, –0.018)) were associated with less disability and lower mRS scores. At day 180, severity index (β=−0.557 (95% CI −0.637, –0.478)), surgical group (β=−0.250 (95% CI −0.463, –0.038)) and Asian race (β=−0.666 (95% CI −1.11, –0.218)) were associated with lower mRS scores. No factors were significantly associated with mRS scores at day 180 post-ICH. At day 365 post-ICH, participants with evidence of previous disability (ie, documented historic Rankin score) were associated with higher mRS scores (β=0.559 (95% CI 0.010, 1.02)), whereas severity index (β=−0.571 (95% CI −0.658, –0.483)) and Asian race (β=−0.599 (95% CI −1.10, –0.099)) were associated with lower mRS scores. The severity index was the only factor to be associated with lower mRS scores across all time points (online supplemental table 5). Discussion In a large prospective observational cohort of patients with spontaneous, supratentorial ICH, we found multiple associations at both short-term and long-term time points between right hemispheric ICH and higher levels of reported impairments with non-motor outcomes including pain and discomfort and anxiety and depression. Patients with right hemispheric ICH were more likely to report problems with anxiety and depression at days 30 and 180 following ICH compared with those with left-sided haematomas. Right hemispheric ICH was also associated with an increase in reported problems with pain and discomfort at all time points. However, laterality was not associated with an outcome more reflective of motor function, reported here as mRS scores. Together, these findings suggest that the non-motor sequelae following ICH change over time and are more often associated in those with right-hemispheric ICH. In our analysis, the laterality of ICH did not appear to affect the level of overall disability in survivors at days 30, 180 and 365 post-ICH, reflected by mRS scores. Two previous analyses including participants from the Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial (INTERACT 1 and 2) and Virtual International Stroke Trials Archives (VISTA) found that ICH laterality did not affect the level of major disability in survivors at 90 days, measured by mRS scores.2 25 Interestingly, right hemispheric lesion was associated with an increased risk of mortality in the INTERACT participants at day 90, while no differences in mortality were found in the VISTA group at the same time point. One reason for these differences may be due to the type of stroke studied. Only 10% of participants in the VISTA analysis had a haemorrhagic stroke compared with all participants in the INTERACT analysis. Our study also demonstrated an association between right hemispheric ICH and increased reported anxiety and depression. A recent study reported a high rate of anxiety in ICH survivors, with 27% of survivors reporting anxiety at 3–5 years follow-up; however, laterality was not included in the multivariable model of factors associated with anxiety.26 Other work has reported similar findings among haemorrhagic and ischaemic stroke populations regarding poststroke depression and right-sided lesions, although using different outcome measures and endpoints.27,29 One rationale that has been used for this finding is that language impairment contributes to the inability to completely assess those with left hemisphere ICH, resulting in a sampling bias. Here, we found no significant difference in the number of participants for each analysis with regard to hemisphere. Of those excluded due to missing data from the EQ-5D-3L at analysis at day 30, 63.7% had left hemispheric lesions. Variables predictive of ICH outcome including age, NIHSS scores and severity index were similar between those with left and right-sided lesions excluded from the main analysis and rates of depression, anxiety and pain were similar between patient and proxy reporting. Additionally, to account for possible bias introduced by the left lateralisation of language impairment, in a subanalysis of patients with no or minor aphasia (ie, NIHSS subscore 9 of 0–1), right-sided ICH was associated with increased reported anxiety and depression and pain at days 30 and 180. There was an association between right hemispheric ICH and increased reported problems with pain and discomfort. Recent analyses of other large cohorts of ICH patients have also reported a high prevalence of pain (39.7%–60%) in ICH survivors, but laterality was not assessed in these studies.29 30 Strokes affecting the right hemisphere are more often associated with pain, including both thalamic and non-thalamic strokes,31 and while cortical areas are well recognised for their hemispheric specialisations, subcortical areas may also play a role in functional lateralisation. For example, in a rodent model of arthritis, Guangchen et al found differences in the processing of nociceptive information between neurons in the left and right amygdala, with right hemispheric lateralisation of pain processing.32 Other work has been done in human participants, including patients with complex regional pain syndrome. Diffusion tensor imaging revealed grey matter atrophy in the right hemispheric structures and decreased white matter connectivity in the left cingulum-callosal bundle.33 The strengths of this substudy include the large sample size and rigorous clinical trial design, enabling the collection of numerous clinical outcome measures at multiple time points without bias related to hemispheric lateralisation. Additionally, to our knowledge, this study is the first to longitudinally characterise multiple non-motor outcomes across time exclusively limited to patients with ICH. The analysis was also prespecified, reducing, but not eliminating, biased associations. However, there are also several limitations. First, patients included in this analysis were not screened prior to randomisation for problems with anxiety, depression or pain, among other measures collected by the EQ-5D-3L questionnaire. Previous MISTIE III substudy analysis has described this limitation regarding depression screening in further detail.22 Second, while many of our measures were collected directly from patients, some were collected by proxy for those whose ability to communicate was affected by varying degrees of aphasia. However, the magnitude of these differences has been estimated to be small to moderate and was therefore not included as a covariate in our analyses.34 In addition, when we explored for confounding of aphasia, our results did not change. In conclusion, those with right hemispheric ICH reported increased problems with anxiety, depression and pain compared with those with left-sided lesions. We add to the current evidence that ICH laterality may be a significant prognostic tool when evaluating patients with ICH for ongoing sequelae. Our data suggest that increased screening for these symptoms may be warranted among all patients, especially those with right hemispheric haemorrhages. Given the availability of treatment for mood disorders and pain, explicit questioning of ICH survivors and their caregivers for these symptoms can impact their treatment and overall quality of life. Furthermore, changes in mood and pain are dynamic and change over the course of recovery, warranting close monitoring over time to detect changing symptoms. Supplementary material Acknowledgements The authors would like to thank Meghana Rao and Nick Rizzuitti for their assistance in data preparation and BIOS project proposal submission. Footnotes Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Data availability free text: The datasets generated during and/or analysed during the current study are available from LHS (lauren.sansing@yale.edu) on reasonable request. Ethics approval: The Johns Hopkins BIOS Clinical Trials Coordinating Centre and institutional review board approved the de-identified analysis of previously collected patient data. Data availability statement Data are available on reasonable request. References Associated Data Supplementary Materials Data Availability Statement Data are available on reasonable request.