This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A comprehensive literature search was performed in PubMed, Embase, SinoMed, and the Cochrane Library for all relevant articles published from database inception to June 2025. The search strategy combined medical subject headings (MeSH) and text words related to intracerebral hemorrhage (e.g., “Intracerebral Hemorrhage,” “Cerebral Hemorrhage”) and C-reactive protein (e.g., “C-Reactive Protein,” “CRP,” “inflammatory markers”) and clinical outcomes (e.g., “prognosis,” “outcome,” “mortality,” “morbidity”). No language restrictions were applied initially. The reference lists of retrieved articles and relevant reviews were also manually screened for additional eligible studies.

Studies were included if they met the following criteria: 1) study design was an observational cohort (prospective or retrospective) or case-control study; 2) the study population consisted of adult patients (≥18 years) with a primary diagnosis of spontaneous ICH; 3) the exposure of interest was the level of CRP measured during hospitalization; 4) the study reported at least one of the primary outcomes: mortality, poor functional outcome, early hematoma growth, overall survival, delayed mobilization; and 5) the study provided sufficient data to calculate an odds ratio (OR), hazard ratio (HR), or relative risk (RR) with a corresponding 95% confidence interval (CI), or provided raw data from which these could be calculated.

Studies were excluded if they: 1) involved non-spontaneous ICH (e.g., due to trauma, tumor, aneurysm rupture, or arteriovenous malformation); 2) were case reports, case series with fewer than 20 patients, reviews, editorials, or conference abstracts; 3) did not measure CRP; or 4) did not report on the specified clinical outcomes.

Two reviewers independently screened the titles and abstracts of all identified articles. The full texts of potentially eligible articles were then retrieved and assessed for final inclusion. Any disagreements were resolved by consensus or by consulting a third reviewer.

We extracted the following data from each included study: publication year, first author, country, study design, sample size (with male count and proportion), mean age, CRP measurement timing, follow-up duration, statistical model, outcomes, study design (single-center/multicenter), and CRP cut-off values, and the adjusted effect estimates (ORs or HRs) with their 95% CIs for the association between CRP and outcomes. If a study reported multiple adjusted models, we extracted the estimate from the most fully adjusted model.

The methodological quality of the included observational studies was independently assessed by two reviewers using the Newcastle-Ottawa Scale (NOS). The NOS evaluates studies based on three domains: selection of study groups, comparability of groups, and ascertainment of exposure or outcome. Scores range from 0 to 9 stars, with studies scoring ≥ 7 considered high quality, 5 - 6 as moderate quality, and <5 as low quality.

The primary outcomes were mortality, poor functional outcome, early hematoma growth, overall survival, and delayed mobilization. We pooled the adjusted ORs from individual studies using a random-effects model (DerSimonian and Laird method), which accounts for both within-study and between-study variation. For studies reporting HRs, these were considered as approximations of ORs, a valid assumption when the outcome is relatively rare. When studies reported CRP as a continuous variable, we used the provided OR per unit or standard deviation increase. When CRP was categorized, we compared the highest versus the lowest category.

Statistical heterogeneity among studies was quantified using the I² statistic, with values of <50% and >50% indicating low and high heterogeneity, respectively. The Cochran’s Q test was also used, with a p-value < 0.10 indicating significant heterogeneity.

Potential publication bias was assessed visually by inspecting the symmetry of a funnel plot and quantitatively using Egger’s linear regression test, where a p-value < 0.05 was considered indicative of significant bias. All statistical analyses were performed using R 4.3.3 and Stata software, version 17.0.

The initial database search yielded 841 records. After removing 89 duplicates, 752 titles and abstracts were screened. Of these, 673 were excluded as they were irrelevant, were not original research (e.g., reviews, editorials), or did not meet the population or exposure criteria. The full texts of the remaining 79 articles were assessed for eligibility. A further 64 articles were excluded for various reasons, including having a population with mixed stroke types without separate data for ICH, not measuring CRP on admission, not reporting the outcomes of interest, or not providing sufficient data for meta-analysis. Ultimately, 15 studies met all inclusion criteria and were included in the systematic review and meta-analysis. The selection process is outlined in Figure 1.

The 15 included studies were published between 2011 and 2022 and comprised a total of 14,285 patients with spontaneous ICH. Of these, 11 were conducted in Asia, 3 in Europe, and 1 in Africa. All studies were observational, with 7 being retrospective cohort studies and 7 being prospective cohort studies, 1 being a pro-retrospective cohort study. The sample sizes ranged from 46 to 9589 patients. The mean or median age of patients was typically between 50 and 80 years. The definition of elevated CRP varied across studies, with cut-off values ranging from 5 mg/L to over 30 mg/L, while some studies analyzed CRP as a continuous variable. The primary outcomes were mortality, poor functional outcome, hematoma expansion, and overall survival, with follow-up periods ranging from in-hospital to 12 months. The quality of the included studies, as assessed by the NOS, was generally good, with 12 studies rated as high quality (score ≥ 7) and 3 as moderate quality (score 5 - 6). Detailed information on the included studies is provided in Table 1.