Antiplatelet therapy (APT) has been shown to protect against atherothrombosis while increasing the risk of major bleeding1. In recent decades, significant efforts have focused on identifying individuals for whom the benefits of APT in preventing a first myocardial infarction or ischemic stroke (primary prevention) outweigh the associated risks of bleeding. Early meta-analyses of primary prevention trials demonstrated that while low-dose aspirin provided modest protection against cardiovascular events, it also increased the risk of major bleeding2,3. More recent data show that the number needed to treat with aspirin to prevent one cardiovascular event over 10 years in the general population was 265, comparable to the number needed to cause one major bleeding event (210). Consequently, guidelines recommend low-dose aspirin only for individuals with high cardiovascular risk, where the potential benefits clearly outweigh the risks4,5,6.
Despite these recommendations, the evidence-based use of APT for primary prevention of cardiovascular diseases is still far from being consistently applied in clinical practice. A cross-sectional study from the U.S. National Cardiovascular Disease Registry’s Practice Innovation and Clinical Excellence (PINNACLE) registry revealed that more than 1 in 10 patients were receiving inappropriate aspirin therapy for primary prevention, with significant variation across clinical practices7. Similarly, a recent Italian study found that inappropriate APT prescription is even more prevalent among acutely hospitalized older adults (> 50%)8.
Intracerebral hemorrhage (ICH) represents the most feared complication of long-term APT and it has been shown that up to 20–30% of ICHs occur in patients receiving APT9. Furthermore, antiplatelet agents could potentially influence the prognosis of ICH by promoting the growth of the hematoma through their inhibitory effects on platelet aggregation and thus raising mortality rates and residual disability10. Nevertheless, existing studies have yielded no conclusive evidence regarding the impact of APT on both the occurrence and the prognosis of ICH, particularly in terms of disability and mortality. Prior meta-analyses indicated a higher case-fatality in patients with APT-ICH compared to non-APT-ICH9,10,11, though functional outcomes appeared comparable11. However, recent cohort studies have failed to confirm these associations12,13. In our study we aimed to provide updated figures on the incidence and prognosis of first-ever ICH occurring on prior APT over 10 years in a population-based stroke registry and to investigate the rates of inappropriate APT prescription.
Methods
Study design and population
Results of the present study were presented according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE). Our research is part of a prospective, population-based stroke registry conducted within the L’Aquila district (mean catchment population of 298,343 residents). The district is characterized by mountainous terrain and is served by 4 public hospitals, all equipped with around-the-clock access to brain CT scans. Two of these hospitals have specialized neurology departments, and one has a dedicated neurosurgical unit. Medical care within this district is provided free of charge, ensuring easy access to medical services during the acute phase of stroke. The registry adheres to epidemiological standards for studying stroke incidence14and has received approval from the Internal Review Board at the University of L’Aquila, under protocol numbers 13/2018 and 57/2019. The registry was conducted in accordance with relevant guidelines/regulations and in accordance with the Declaration of Helsinki. It encompasses all instances of cerebrovascular events occurring within the district, which are regularly reported by local physicians, verified by our research team, and followed up. Patients receive treatment in accordance with routine clinical protocols, as well as national and international guidelines. For our investigation, we specifically recorded cases of intracerebral hemorrhage (ICH) that occurred in the L’Aquila district over a ten-year period, spanning from January 1, 2011, to December 31, 2020. The diagnosis of ICH was confirmed based on the presence of localized neurological deficits accompanied by concurrent evidence of intraparenchymal bleeding as shown in brain imaging15. We included only patients experiencing their first-ever ICH, excluding those with a history of previous strokes/TIA and those with hemorrhagic transformations of cerebral infarctions. ICH cases occurring on anticoagulation were excluded when the last intake of anticoagulant occurred within the last 48 h of ICH. Similarly, we excluded patients with primary subdural or epidural hematomas, traumatic ICH, or hemorrhage related to a tumor were also excluded.
Case-finding procedures
We monitored both inpatient and outpatient healthcare services to detect any ICH event. All cases were initially recognized by a senior physician within 7 days of symptoms onset and subsequently verified by a consulting neurologist. Admission and discharge records were also examined, as well as records from emergency, neuroradiology, neurophysiology, and neurosonology services. Additionally, we conducted a thorough review of patient records that exhibited symptoms potentially indicative of differential diagnoses, including transient ischemic attack (TIA), dizziness, vertigo, confusion, seizures, headaches, and transient global amnesia. We extended our surveillance to encompass neighboring hospitals, rehabilitation centers, and long-term care facilities. On a monthly basis, we scrutinized death certificates, and we included clinical data for all patients who died with an ICH diagnosis, whether their information was not already included in the registry. Comprehensive identification and follow-up was ensured by hot and cold pursuit.
Data collection and follow-up
Demographic and clinical information was systematically collected by reviewing medical records and securely stored in a computerized database in a completely anonymous form, by utilizing Research Electronic Data Capture (REDCap). We recorded data regarding medical history, cardiovascular and neurological assessments. Additionally, we assessed the clinical severity at the onset of ICH using the National Institutes of Health Stroke Scale (NIHSS) score. Furthermore, we evaluated disability or dependence in daily activities by means of the modified Rankin scale (mRS) score. Specifically, we adjudicated the mRS score both prior to the occurrence of the index event and at the time of discharge, with the latter evaluation separately reported for the overall population and for those who survived at discharge. All outcomes were assessed by the treating physician without blinding to APT status. Definitions of vascular risk factors are detailed in Supplemental Table 1.
The diagnosis of ICH was confirmed through non-contrast computed tomography (NCCT) scans of the brain at the time of inclusion. We retrospectively assessed the volume and location of the ICH based on the initial available brain NCCT. ICH volumes were estimated by a single operator using the ABC/2 method16. ICH location was categorized according to the anatomical site as lobar, deep, infratentorial (brainstem or cerebellum) or mixed (i.e., very large ICH extending into both lobar and non-lobar areas). Any conflict was resolved through consensus before starting the primary analyses. Hematoma expansion was adjudicated retrospectively when ICH volume increased ≥ 6 ml from the first to the second available brain NCCT17.
APT use
APT-ICH was defined as an ICH occurring in patients on single or dual treatment with aspirin, clopidogrel, ticlopidine, or dipyridamole. We included only patients who developed ICH within 96 h of the last APT intake, based on the maximum time required to recover a normal platelet function after discontinuing aspirin18.
Based on the adherence to the 2021 European Society of Cardiology (ESC) guidelines for cardiovascular prevention in clinical practice6, (Supplemental Table 2) we categorized patients with APT-ICH into two groups: recommended (R-) and not recommended (NR-) APT. R-APT-ICH was defined if any of the following conditions were met: (1) a history of prior myocardial infarction or revascularization; (2) symptomatic lower extremity artery disease; (3) diabetes mellitus and a high risk of ASCVD. Given that our data collection did not encompass laboratory results, including cholesterol levels, we were unable to calculate the 10-year risk of ASCVD. Therefore, we determined that APT was appropriate for individuals with diabetes mellitus if they presented with at least two of the following conditions: age ≥ 65 years, hypertension, dyslipidemia, current cigarette smoking, or a history of heart failure. Lastly, since we included only first-ever strokes, there were no cases on APT for the secondary prevention of ischemic stroke or TIA.
Statistical analysis
We provided an analysis of the global incidence rate based on catchment population for non-APT and APT-ICH and its trend over the 10-year observation period. To assess the incidence trend and determine confidence intervals (CIs) for the incidence rates, we employed Poisson regression analysis. The Incidence Rate Ratios (IRRs) were calculated based on a Poisson distribution, under the assumption of constant event-rates. We provided separate figures for patients with NR-APT-ICH. Descriptive statistics are provided in terms of absolute numbers with accompanying percentages or as mean values ± standard deviation (SD), as appropriate. Continuous and categorical variables were compared using either the Wilcoxon test or the Pearson 2 test, respectively. We established a two-sided statistical significance level at P < 0.05.
We reported 30-day and 1-year case-fatality rates as numbers and percentages with associated confidence intervals (CIs). The overall survival following the index event was estimated using Kaplan-Meier curves, and differences between various groups were assessed using the log-rank test. Univariate estimates of hazard ratios for factors influencing the 30-day and 1-year case-fatality rates were calculated using Cox regression analysis, which included age, sex, risk factors, APT intake, ICH volume and location, intraventricular expansion. Subsequently, we performed a multivariate Cox regression analysis including variables with statistical significance < 0.05 to identify independent predictors of 30-day and 1-year case-fatality in the overall population.
To address missing data in risk factors, we cross-referenced patients’ treatment records to enhance data completeness. For Cox regression analyses involving continuous variables, missing values were imputed using the median of the respective variables, where applicable, to maintain consistency in the analyses.
All statistical analyses were performed using R software, version 4.2.
link
