Effect of Tranexamic Acid Administration on Remote Cerebral Ischemic Lesions in Acute Spontaneous Intracerebral Hemorrhage: A Substudy of a Randomized Clinical Trial

Key Points

Question  Does intravenous tranexamic acid that is administered after acute spontaneous intracerebral hemorrhage (ICH) increase diffusion-weighted imaging hyperintense lesions?

Findings  In this substudy of a randomized clinical trial involving 219 individuals with acute spontaneous ICH, the prevalence or number of diffusion-weighted imaging hyperintense lesions on brain MRI scans within 2 weeks did not increase in those who received tranexamic acid compared with placebo within 8 hours of acute spontaneous ICH.

Meaning  Findings of this substudy suggest that tranexamic acid is unlikely to increase cerebral ischemic events in acute spontaneous ICH.

Importance  Hyperintense foci on diffusion-weighted imaging (DWI) that are spatially remote from the acute hematoma occur in 20% of people with acute spontaneous intracerebral hemorrhage (ICH). Tranexamic acid, a hemostatic agent that is under investigation for treating acute ICH, might increase DWI hyperintense lesions (DWIHLs).

Objective  To establish whether tranexamic acid compared with placebo increased the prevalence or number of remote cerebral DWIHLs within 2 weeks of ICH onset.

Design, Setting, and Participants  This prospective nested magnetic resonance imaging (MRI) substudy of a randomized clinical trial (RCT) recruited participants from the multicenter, double-blind, placebo-controlled, phase 3 RCT (Tranexamic Acid for Hyperacute Primary Intracerebral Hemorrhage [TICH-2]) from July 1, 2015, to September 30, 2017, and conducted follow-up to 90 days after participants were randomized to either the tranexamic acid or placebo group. Participants had acute spontaneous ICH and included TICH-2 participants who provided consent to undergo additional MRI scans for the MRI substudy and those who had clinical MRI data that were compatible with the brain MRI protocol of the substudy. Data analyses were performed on an intention-to-treat basis on January 20, 2020.

Interventions  The tranexamic acid group received 1 g in 100-mL intravenous bolus loading dose, followed by 1 g in 250-mL infusion within 8 hours of ICH onset. The placebo group received 0.9% saline within 8 hours of ICH onset. Brain MRI scans, including DWI, were performed within 2 weeks.

Main Outcomes and Measures  Prevalence and number of remote DWIHLs were compared between the treatment groups using binary logistic regression adjusted for baseline covariates.

Results  A total of 219 participants (mean [SD] age, 65.1 [13.8] years; 126 men [57.5%]) who had brain MRI data were included. Of these participants, 96 (43.8%) were randomized to receive tranexamic acid and 123 (56.2%) were randomized to receive placebo. No baseline differences in demographic characteristics and clinical or imaging features were found between the groups. There was no increase for the tranexamic acid group compared with the placebo group in DWIHL prevalence (20 of 96 [20.8%] vs 28 of 123 [22.8%]; odds ratio [OR], 0.71; 95% CI, 0.33-1.53; P = .39) or mean (SD) number of DWIHLs (1.75 [1.45] vs 1.81 [1.71]; mean difference [MD], −0.08; 95% CI, −0.36 to 0.20; P = .59). In an exploratory analysis, participants who were randomized within 3 hours of ICH onset or those with chronic infarcts appeared less likely to have DWIHLs if they received tranexamic acid. Participants with probable cerebral amyloid angiopathy appeared more likely to have DWIHLs if they received tranexamic acid.

Conclusions and Relevance  This substudy of an RCT found no evidence of increased prevalence or number of remote DWIHLs after tranexamic acid treatment in acute ICH. These findings provide reassurance for ongoing and future trials that tranexamic acid for acute ICH is unlikely to induce cerebral ischemic events.

Trial Registration  isrctn.org Identifier: ISRCTN93732214

The presence of diffusion-weighted imaging hyperintense lesions (DWIHLs) that are spatially remote to acute spontaneous intracerebral hemorrhage (ICH) has been reported in 11% to 40% (with a mean of 20%) of individuals, with a similar prevalence in ICH that is attributed to arteriosclerosis (hypertensive arteriopathy) and cerebral amyloid angiopathy (CAA).^1-16 Remote DWIHLs found in the context of acute ICH are typically asymptomatic, although they have been associated with lower baseline Glasgow Coma Scale score,³ worse 3-month outcomes,^6,10 recurrent ICH,¹⁴ and cerebrovascular events or vascular death.⁷

The cause of DWIHLs in ICH has not been definitively established, although they are believed to be associated with ischemia. In CAA, histopathological and longitudinal magnetic resonance imaging (MRI) scans have indicated that DWIHLs are attributable to microinfarcts.^17,18 A meta-analysis found an association between the presence of DWIHLs in ICH and white matter hyperintensities, previous ICH, microbleed burden, and greater decrease in systolic blood pressure within the first 24 hours after ICH, but not other vascular risk factors or antithrombotic treatment.¹⁶ The association of acute blood pressure reduction with underlying small vessel disease provides a plausible mechanism for the development of acute ischemic foci.

Tranexamic acid is a synthetic lysine analog that competes with lysine residues on fibrin for the binding of plasminogen, potently inhibiting the interaction of plasmin with fibrin and preventing the dissolution of fibrin clot.¹⁹ Large randomized trials of tranexamic acid showed reduced risk of death from bleeding in traumatic^20,21 and postpartum²² hemorrhage. Tranexamic acid has been evaluated as a hemostatic agent in spontaneous ICH.^23-25 The Tranexamic Acid for Hyperacute Primary Intracerebral Hemorrhage (TICH-2) trial, an international placebo-controlled phase 3 randomized clinical trial of tranexamic acid administration within 8 hours of spontaneous ICH,²⁶ found decreases in hematoma expansion and early deaths, although the primary outcome of functional status at 3 months was not substantially different between those who were randomized to receive tranexamic acid and those who were randomized to receive placebo.²⁴ Because of the probable ischemic basis of DWIHLs in people with acute ICH, tranexamic acid administration could potentiate the risk for DWIHLs by, for example, preventing the dissolution of microemboli arriving in the cerebral microvasculature or of thrombus forming spontaneously on dysfunctional endothelium of cerebral arterioles affected by small vessel disease. The TICH, TICH-2, and STOP-AUST (the Spot Sign and Tranexamic Acid On Preventing Intracerebral Hemorrhage Growth—Australasia Trial) trials found no increase in ischemic stroke or symptomatic vascular occlusive complications after tranexamic acid use.^23-25 However, DWIHLs in ICH are typically asymptomatic, requiring MRI for diagnosis, and the possibility remains that tranexamic acid administration could increase subclinical cerebral ischemic events. In this prospective nested MRI substudy (protocol in Supplement 1²⁷) of the TICH-2 trial, we aimed to establish whether tranexamic acid, compared with placebo, increased the prevalence or number of remote cerebral DWIHLs within 2 weeks of ICH onset.

Participants in the TICH-2 trial were recruited between March 1, 2013, and September 30, 2017, from 124 participating centers (in the United Kingdom, Denmark, Georgia, Hungary, Italy, Malaysia, Poland, Republic of Ireland, Switzerland, and Turkey), according to the inclusion and exclusion criteria for the trial²⁶ (eTable 1 in Supplement 2). These participants were randomized 1:1 to receive either intravenous tranexamic acid (1 g in 100-mL intravenous bolus loading dose, followed by 1 g in 250-mL infusion over 8 hours) or placebo (0.9% saline using identical administration regime) treatment. Brain MRI scans, including DWI, were performed within 2 weeks. The nested MRI substudy commenced on July 1, 2015, and continued to the close of the TICH-2 trial on September 30, 2017. Recruitment of TICH-2 trial participants for the nested MRI substudy was allowed at any time between the initial TICH-2 trial recruitment and up to 7 days after randomization. The additional inclusion and exclusion criteria for the MRI substudy are listed in eTable 1 in Supplement 2. We included in the substudy participants who consented to undergo additional MRI scans for the substudy and those who had MRI scan data that were compatible with the brain MRI protocol of the substudy (Figure 1). Blinding to treatment randomization was maintained throughout the course of both the TICH-2 trial and MRI substudy. Written informed consent was obtained from participants in the MRI substudy. The MRI substudy obtained ethical approval from Nottingham 2 National Health Service Research Ethics Committee. We followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.

Baseline clinical variables used in this analysis were collected as part of the TICH-2 trial²⁶ and included Glasgow Coma Scale score; National Institutes of Health Stroke Scale score; prestroke modified Rankin Scale score; measurement of systolic and diastolic blood pressure; previous antiplatelet therapy or statin use; and history of stroke, transient ischemic attack (TIA), ischemic heart disease, or thromboembolism. Clinical outcome variables were specified for the TICH-2 trial²⁶ and included functional status at 90 days after trial enrollment, as assessed by the modified Rankin Scale; death; and prespecified safety outcomes, including ischemic stroke or TIA, reported up to 90 days after trial enrollment.

Data were collected on race and ethnicity in the TICH-2 trial to allow a prespecified subgroup analysis. Race and ethnicity data were obtained from the health records of the TICH-2 trial sites and were either investigator observed or self-reported. The following race and ethnicity categories were identified: Asian (including South Asian, East Asian, and any other Asian background), Black, White, and other (including Arab, Hispanic, mixed ethnic backgrounds, and any other ethnic backgrounds not otherwise specified).

The MRI protocol²⁷ comprised a core sequence set that included axial diffusion-weighted imaging (DWI), axial T2-weighted imaging, T2*-weighted gradient echo imaging, T2-weighted fluid attenuated inversion recovery (T2-FLAIR), and 3-dimensional T1-weighted volume. Susceptibility weighted imaging was specified as an optional additional sequence because it was not routinely available at all sites at the start of the substudy. Compliance with the MRI protocol was assessed on an individual basis using an automated report generator. For the substudy, we included analyzable data sets that were derived from the DWI, T2-weighted imaging, T2-FLAIR, T2*-weighted gradient echo imaging, and susceptibility weighted imaging.

The operational definition for DWIHL was a focal hyperintense parenchymal lesion on the b = 1000 diffusion-weighted image that was confirmed as low diffusion on the derived apparent diffusion coefficient maps and was spatially remote (>20 mm) from the index ICH.² Three of our certified neuroradiologists (R.G., D.S., and R.A.D.) performed DWIHL analyses. Lesions were labeled manually by the first neuroradiologist and independently labeled by the second neuroradiologist using a semiautomated lesion finding tool (eMethods 1 and eFigure 1 in Supplement 2). The third neuroradiologist adjudicated any lesions when a discrepancy emerged between the first and second neuroradiologists, either accepting or rejecting the candidate lesions. Therefore, all accepted DWIHLs achieved consensus (reached independently) between at least 2 neuroradiologists.

Baseline hematoma volume and hematoma expansion were derived from the prerandomization and 24-hour CT scan by semiautomated segmentation with manual editing as required.²⁸ Other baseline imaging features (ie, presence of intraventricular hemorrhage and hematoma location) were adjudicated by expert readers, consisting of 4 certified neuroradiologists (including R.G., A.M.C., and R.A.D.) and 1 vascular neurologist, on baseline CT.²⁶ These expert readers adjudicated brain MRI scans for the presence of chronic infarcts (areas of T2 or T2-FLAIR hyperintensity that were consistent with infarction without diffusion restriction); rated mass effect as none, moderate (effacement of overlying sulci and/or distortion of adjacent ventricle), or severe (complete effacement of adjacent ventricle and/or overt midline shift); rated white matter hyperintensities according to a visual rating scale²⁹; and applied the Modified Boston Criteria³⁰ to classify the likelihood that a participant has coexistent CAA. For the substudy, the outcome of the CAA classification was dichotomized into probable CAA and a combined group of possible CAA and not CAA, on the basis of probable CAA being the key diagnostic threshold for clinical practice and research.³¹ One of us (R.A.D.) manually labeled cerebral microbleeds on T2*-weighted gradient echo imaging or susceptibility weighted imaging (eMethods 2 and eFigure 2 in Supplement 2) as either definite or possible according to the Microbleed Anatomical Rating Scale.³² Total cerebral microbleed number (a combination of definite and possible labels) was used in the substudy.

The basis for the sample size estimate has been published previously.²⁷ In brief, assuming a 10% relative increase (from 20% to 22%) in the prevalence of DWIHLs in the tranexamic acid group, a sample size of 264 participants with analyzable DWI data was required to test the primary hypothesis (power = 80%; α = .05%). Allowing for approximately 5% imaging data rejection rate, the target recruitment was set at 280 participants.

The primary outcome measure (presence of DWIHL after ICH) was compared between the tranexamic acid and placebo groups using binary logistic regression with adjustment for baseline covariates: age, time from ICH onset to randomization, stroke severity (National Institutes of Health Stroke Scale score), mean systolic blood pressure, previous antiplatelet treatment, and hematoma volume.²⁷ These covariates were selected to align with those in the TICH-2 trial.³³ However, since the publication of the MRI substudy protocol, a meta-analysis has identified an association between the presence of DWIHLs and white matter hyperintensities, previous ICH, and presence and number of microbleeds.¹⁶ Therefore, in an explanatory analysis, we repeated the binary logistic regression with adjustment for these additional baseline covariates.

Baseline and clinical characteristics were compared using χ²; an unpaired, 2-tailed t test; and Mann-Whitney test. Heterogeneity of the treatment effect on the primary outcome was assessed in subgroups by adding an interaction term in an adjusted binary logistic regression model. Other secondary outcome measures from the TICH-2 trial were analyzed using multiple linear regression for continuous outcomes, ordinal logistic regression for ordered categorical outcomes, binary logistic regression for binary outcomes, and Cox proportional hazards regression model for time-to-event data. All regression analyses were adjusted for the covariates.

Findings were presented as number (%) for count data, mean (SD) for continuous data, and median (IQR) for categorical data. All analyses were conducted on an intention-to-treat basis. The nominal level of significance for all analyses was a 2-sided P < .05. No adjustment was made for multiplicity of testing. Two of us (S.P. and L.J.W.) performed the statistical analyses in accordance with the published protocol using SAS, version 9.4 (SAS Institute, Inc), and MATLAB, version R2018a (MathWorks). Data were analyzed on January 20, 2020.

Of the 253 TICH-2 trial participants who underwent MRI scans within the substudy time window, 34 were excluded before unblinding. A total of 219 participants (mean [SD] age, 65.1 [13.8] years; 126 men [57.5%] and 93 women [42.5%]) with analyzable DWI data sets from 50 centers participating in the TICH-2 trial were included (Figure 1). Of these participants, 96 (43.8%) were randomized to receive tranexamic acid and 123 (56.2%) were randomized to receive placebo. There were no between-group differences in demographic characteristics, baseline clinical features, baseline hematoma volume, and distribution or imaging markers of cerebral small vessel diseases (Table 1). The median (IQR) interval time from randomization to MRI scan was 4.0 (2.1-5.7) days, with no difference between the tranexamic acid and placebo groups (3.9 [2.5-5.4] days vs 4.0 [2.1-5.9] days; P = .82). The MRI platforms and acquisition parameters for included MRI data are presented in eTables 2 and 3 in Supplement 2. Comparison of baseline characteristics of the TICH-2 trial participants and the MRI substudy participants is presented in eTable 4 in Supplement 2.

Overall prevalence of DWIHLs in the MRI substudy population was 21.9% (48 of 219). Of the total of 127 individual DWIHLs detected, 86.6% (110 of 127) were in the cerebral lobes, 5.5% (7 of 127) were in the deep supratentorial gray matter (basal ganglia and thalami), 5.5% (7 of 127) were in the cerebellum, and 2.4% (3 of 127) were in the brainstem (Figure 2). Of the 110 individual lobar DWIHLs, 16.4% (18 of 110) were cortical and 83.6% (92 of 110) were in lobar white matter.

There was no increase in the prevalence of DWIHLs between the tranexamic acid and placebo groups (20 of 96 [20.8%] vs 28 of 123 [22.8%]; odds ratio [OR], 0.71; 95% CI, 0.33-1.53; P = .39) (Table 2). For participants with DWIHLs, there was no statistically significant difference in the mean (SD) number of DWIHLs per participant between the tranexamic acid and placebo groups (1.75 [1.45] vs 1.81 [1.71]; mean difference [MD], −0.08; 95% CI, −0.36 to 0.20; P = .59). When the analysis was repeated to include covariates that were associated with the presence of DWIHLs in a meta-analysis,¹⁶ the results were similar: no evidence of increased DWIHL prevalence (OR, 0.84; 95% CI, 0.34-2.05; P = .70) or number (MD, −0.20; 95% CI, −0.53 to 0.23; P = .22) was found in the tranexamic acid group.

Exploratory subgroup analysis (Figure 3) identified possible interactions between clinical or imaging variables and group randomization, which affected the presence of DWIHLs. A significant interaction (P = .01) between CAA status and treatment group affecting DWIHL prevalence was found, such that participants without probable CAA appeared less likely to have DWIHLs if they were randomized to receive tranexamic acid vs placebo (68 of 96 [70.8%] vs 92 of 123 [74.8%]; OR, 0.34; 95% CI, 0.13-0.87), whereas participants with probable CAA appeared to be more likely to have DWIHLs if they were randomized to receive tranexamic acid vs placebo (28 of 96 [29.2%] vs 31 of 123 [25.2%]; OR, 3.76; 95% CI, 0.92-15.39). A significant interaction (P = .002) between presence of chronic infarcts and treatment group affecting DWIHL prevalence was found, such that participants with chronic infarcts appeared to be less likely to have DWIHLs if they were randomized to tranexamic acid (38 of 96 [39.6%] vs 53 of 123 [43.1%]; OR, 0.19; 95% CI, 0.05-0.72), whereas those without chronic infarcts showed no treatment group effect on DWIHL prevalence (58 of 96 [60.4%] vs 70 of 123 [56.9%]; OR, 1.81; 95% CI, 0.70-4.70). A significant interaction (P = .02) between time from onset to randomization and treatment group affecting DWIHL prevalence was found, such that participants who were randomized within 3 hours of ICH onset appeared to be less likely to have DWIHLs if they were randomized to tranexamic acid group vs placebo group (25 of 96 [26.0%] vs 46 of 123 [37.4%]; OR, 0.12; 95% CI, 0.01-1.05), whereas those who were randomized after 3 hours showed no treatment group effect on DWIHL prevalence (71 of 96 [74.0%] vs 77 of 123 [62.6%]; OR, 1.05; 95% CI, 0.46-2.73). There was no evidence of interactions between treatment randomization and other demographic characteristics or baseline clinical and imaging features that affected the presence of DWIHLs.

The incidence of new clinical ischemic stroke in the MRI substudy population in the 90 days after randomization was low at 3.2%, but a higher incidence was found in participants with DWIHLs compared with participants without DWIHLs (5 of 48 [10.4%] vs 2 of 171 [1.2%]; P = .02). No TIAs were recorded in the substudy participants. There was no association between the presence of DWIHLs and either the primary outcome (functional status at day 90) or any of the secondary outcome measures of the TICH-2 trial (eTable 5 in Supplement 2).

In this prospective nested MRI substudy, we found no increase in remote DWIHLs in participants with spontaneous ICH who were treated with tranexamic acid compared with those who received placebo. Both the overall prevalence and the number of DWIHLs in participants with lesions were nonsignificantly lower in the tranexamic acid group. This finding is consistent with the report from the TICH-2 trial that there was no increase in clinical ischemic stroke or TIA in the tranexamic acid group compared with the placebo group.²⁴ Together, these results can provide reassurance to researchers in ongoing and future studies of tranexamic acid in acute spontaneous ICH³⁴ that a 1-g intravenous bolus followed by an 8-hour 1-g infusion of tranexamic acid does not appear to increase cerebral ischemic events. If subsequent trials provided support for the use of tranexamic acid for ICH in clinical practice, the present substudy, we believe, provides safety evidence to support clinical use.

The overall prevalence of DWIHLs lesions in this substudy (21.9%) aligned with the prevalence identified in a meta-analysis.¹⁶ Although the rate of clinically apparent ischemic stroke after randomization was higher in participants with DWIHLs, the absence of an overt clinical correlate in most of these participants (approximately 90%) confirmed previous reports that these lesions were typically asymptomatic. In contrast to previous studies,^6,10 this substudy found no association between the presence of DWIHLs and death or functional outcomes.

Results of the exploratory analysis we conducted suggested that having CAA-related ICH may confer an increased risk for DWIHLs in people who receive tranexamic acid compared with placebo. A previous study found a 3-fold increase in DWIHL prevalence after CAA compared with non–CAA-related ICH,² although this result was not confirmed in a subsequent meta-analysis, which found no difference in DWIHL prevalence between CAA and hypertension-related ICH.¹⁶ Kimberly and colleagues³⁵ identified a DWIHL prevalence of 15% in people with probable CAA without acute ICH, indicating a propensity for DWIHL development in people with CAA. The finding of a possible potentiating effect of tranexamic acid on DWIHL in people with probable CAA warrants further prospective testing.

We found evidence that participants with chronic infarcts had reduced DWIHL risk if they were randomized to the tranexamic acid group vs the placebo group and that there was no interaction between white matter hyperintensities and the effect of tranexamic acid on DWIHL prevalence. These findings were contrary to our expectation that people with these preexisting ischemic lesions might have increased risk for developing DWIHLs after tranexamic acid administration.

The exploratory analysis identified that randomization within 3 hours of ICH onset may confer a lower risk for DWIHL in those who received tranexamic acid rather than placebo, but randomization had no effect on DWIHL prevalence after 3 hours of onset. The basis for this interaction was unclear. In the TICH-2 trial, time from onset to randomization had no apparent effect on the primary outcome.²⁴ In other contexts, however, such as traumatic brain injury and postpartum hemorrhage, improved outcomes have been shown after the administration of tranexamic acid within 3 hours of onset.³⁶ This analysis of interactions was exploratory and based on a small sample size, which limited statistical power and precision to detect effects in subgroups. Further research is required to confirm and explore the basis for the observed interactions.

This study has several strengths. First, it was a prospective nested substudy in a large, double-blind, placebo-controlled randomized clinical trial. To our knowledge, this substudy was the first to examine the effect of tranexamic acid administration on MRI-detected ischemic lesions in people with acute ICH. Second, DWIHLs were independently identified according to predefined criteria by 2 experienced neuroradiologists, who were blinded to the treatment randomization, using manual and semiautomated labeling. This process helped reduce the observer biases associated with 1 method alone. A third experienced neuroradiologist provided independent adjudication of the discrepancies.

This study also has several limitations. First, it included fewer participants (n = 219) than the target sample size (n = 264). Underrecruitment of participants for the MRI substudy was multifactorial, but feedback from the TICH-2 sites indicated that potential recruits were often too unwell, could not provide consent, or declined to undergo the MRI scan.³⁷ These factors were supported by the finding that participants in the MRI substudy were younger, had less severe stroke, had lower rate of total anterior circulation stroke presentation, and had smaller baseline hematoma volume. As predefined in the substudy protocol,²⁷ the MRI data we analyzed came from individuals who specifically consented to participate in the MRI substudy and from TICH-2 trial participants who had clinical MRI scans that were consistent with the timing and scan parameters of the substudy. The inclusion of the latter group allowed us to mitigate the underrecruitment to the MRI substudy.

Second, although this substudy was nested in a large trial, recruitment was allowed after randomization and hence treatment randomization could be influenced by postrandomization survival and patient decision-making. Therefore, the substudy population should be considered as a nonrandom sample of the main TICH-2 trial population, with a bias toward younger age, smaller ICH, and less severe stroke syndromes. We accepted this limitation because the prerandomization requirement for additional information and participation consent in the TICH-2 trial would have delayed randomization and administration of the intervention, which could have been detrimental to the trial. Third, decrease in blood pressure before the MRI (from baseline) was not collected as a variable in the TICH-2 trial. This oversight is relevant because previous studies^4-6,9,12 and a meta-analysis¹⁶ found that early reduction in blood pressure was associated with the presence of DWIHL.

This nested MRI substudy of a randomized clinical trial found no evidence of increased prevalence or number of DWIHLs in participants with acute ICH who were treated with tranexamic acid compared with the placebo group, providing reassurance to researchers conducting current and future trials that tranexamic acid in acute ICH is unlikely to increase cerebral ischemic events. An exploratory analysis showed that participants with probable CAA may be at an increased risk for developing DWIHLs after receiving tranexamic acid compared with placebo, but the presence of chronic infarcts and shorter time from ICH onset to randomization were associated with reduced risk for DWIHL after tranexamic acid treatment. These effects should be confirmed in larger prospective trials of tranexamic acid in acute ICH.

Accepted for Publication: December 31, 2021.

Published Online: March 21, 2022. doi:10.1001/jamaneurol.2022.0217

Corresponding Author: Rob A. Dineen, PhD, Radiological Sciences, Mental Health and Clinical Neuroscience, University of Nottingham, Queen’s Medical Centre, Derby Road, Nottingham NG7 2UH, United Kingdom (rob.dineen@nottingham.ac.uk).

Author Contributions: Prof Dineen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Sprigg, Roberts, Werring, Al-Shahi Salman, Bath, Dineen.

Acquisition, analysis, or interpretation of data: Pszczolkowski, Woodhouse, Gallagher, Swienton, Law, Casado, Al-Shahi Salman, England, Morgan, Bath, Dineen.

Drafting of the manuscript: Pszczolkowski, Sprigg, Woodhouse, Gallagher, Law, Roberts, Dineen.

Critical revision of the manuscript for important intellectual content: Pszczolkowski, Woodhouse, Swienton, Law, Casado, Roberts, Werring, Al-Shahi Salman, England, Morgan, Bath, Dineen.

Statistical analysis: Pszczolkowski, Woodhouse, Dineen.

Obtained funding: Roberts, England, Bath, Dineen.

Administrative, technical, or material support: Pszczolkowski, Woodhouse, Dineen.

Supervision: Dineen.

Other - MRI physics advice: Morgan.

Conflict of Interest Disclosures: Prof Werring reported receiving personal fees from Bayer, Alnylam, and Novo Nordisk outside the submitted work. Prof Bath reported receiving personal fees for service on the steering committee or advisory board of Diamedica, Phagenesis, and Moleac outside the submitted work. No other disclosures were reported.

Funding/Support: This substudy was funded by grant PG/14/96/31262 from the British Heart Foundation. The TICH-2 Trial was funded by grant 11_129_109 from the National Institute for Health Research Health Technology Assessment program.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 3.