Chinese Multicenter Cohort Study of Intracerebral Hemorrhage

Overview

Spontaneous intracerebral hemorrhage is a severe form of stroke associated with high mortality and long-term disability. The Chinese Multicenter Cohort Study of Intracerebral Hemorrhage (CASA) is a prospective observational patient registry designed to enroll approximately 10,000 adults with imaging-confirmed spontaneous intracerebral hemorrhage and 2,000 adult healthy controls at more than 30 hospitals in China. The study will collect standardized clinical, laboratory, neuroimaging, lifestyle, and treatment data together with ethically approved blood, urine, and other biospecimens. Participants with intracerebral hemorrhage will be followed at 3 months, 6 months, and annually for up to 10 years. The primary objective is to characterize and identify determinants of the full distribution of modified Rankin Scale scores at 90 days. Secondary objectives include evaluating 90-day unfavorable functional outcome defined as mRS 4-6, recurrent intracerebral hemorrhage and mortality, identifying clinical and molecular biomarkers, and developing and externally evaluating multimodal prognostic models. The study does not assign treatment; all clinical care is determined by treating clinicians.

CASA is a multicenter, prospective, observational cohort and patient registry coordinated by Xiangya Hospital, Central South University. Consecutive adults with spontaneous intracerebral hemorrhage confirmed by computed tomography and/or magnetic resonance imaging are identified within 30 days after symptom onset. A healthy comparison cohort is recruited from participating health examination programs and communities. The planned sample comprises 10,000 participants with intracerebral hemorrhage and 2,000 healthy controls. At baseline, trained investigators obtain demographic characteristics, medical and family history, prior medications, vascular risk factors, smoking, alcohol use, dietary and sleep habits, vital signs, neurological severity scores, functional status, laboratory tests, electrocardiography, vascular examinations, treatments, complications, and discharge information. Original computed tomography, computed tomography angiography, magnetic resonance imaging, magnetic resonance angiography, diffusion-weighted imaging, susceptibility-weighted imaging, and digital subtraction angiography are collected when performed as part of clinical care. Imaging is stored in Digital Imaging and Communications in Medicine format. Biospecimens are processed under standardized operating procedures, assigned coded identifiers, and stored centrally for genomic, transcriptomic, proteomic, metabolomic, lipidomic, epigenomic, and other prespecified analyses, within the scope of the ethics-approved consent. Participants with intracerebral hemorrhage are followed at 3 months, 6 months, and annually from year 1 through year 10. Follow-up records functional outcome using the modified Rankin Scale, recurrent symptomatic intracerebral hemorrhage, and vital status. Outcomes are obtained through clinic review, telephone follow-up, family contact, and linkage to available health insurance and death-registration data. Participants not reached after at least three attempts on different dates in different months within a 12-month period are classified as lost to follow-up after additional tracing efforts. The primary outcome is the ordinal distribution of modified Rankin Scale scores from 0 (no symptoms) to 6 (death) at 90 days. The key secondary/supportive outcome is 90-day unfavorable functional outcome, defined as mRS 4-6; this binary endpoint is also used in the protocol's minimum sample-size calculation for prognostic modelling. Other secondary analyses evaluate later mRS distributions, all-cause mortality, recurrent symptomatic intracerebral hemorrhage, hematoma expansion, perihematomal edema, and biomarker-outcome associations. Prediction models will be developed using clinical, imaging, and molecular data. Sites in Hunan, Jiangxi, and Hubei provinces are planned as the development cohort, and sites in other provinces as a geographically external evaluation cohort. Model performance will include discrimination, calibration, overall accuracy, and clinical utility. The registry uses a centralized electronic data capture system, prespecified data definitions, logic and range checks, staff training, source-data review, periodic monitoring, and repeat assessment in a random 3% sample at participating sites. Biospecimen collection, transport, processing, and storage follow harmonized operating procedures with temperature monitoring and chain-of-custody records. The study began before public registration. This retrospective registration is based on protocol version 1.0 dated 9 December 2023, initially approved on 25 January 2024 (No. 202401026). The original protocol prespecified mRS follow-up and used mRS 4-6 in its prognostic-model sample-size calculation but did not label one outcome as primary. Before PRS submission, the steering committee formalized the 90-day ordinal mRS distribution as the primary outcome and mRS 4-6 at 90 days as the key secondary/supportive outcome. The decision, its date and the absence of outcome-informed changes should be recorded in a signed protocol clarification or statistical analysis plan.