Abstract
BACKGROUND: Limited data exists on long-term stroke trends in low- and middle-income countries, including Pakistan, despite its growing burden. Understanding these trends is crucial for informing healthcare strategies and public health interventions. This study examines 20-year trends in stroke characteristics, risk factors, and mortality at a tertiary care center in Pakistan. METHODS: In this retrospective study, a secondary analysis was conducted using clinically digitized records from the Aga Khan University Hospital. Patients aged ≥18 years, admitted with stroke, and discharged between January 1, 1999, and December 31, 2018, were included. Stroke subtypes were classified using ICD-9 CM codes as ischemic stroke, intracerebral hemorrhage (ICH), or transient ischemic attack (TIA) and cross checked against discharge diagnosis and neuroimaging reports to validate case identification. The retrospective, single-center design and possible variability in coding accuracy should be considered limitations when interpreting results. RESULTS: The final cohort included 12,837 patients: 71.4 % ischemic stroke, 21.2 % ICH, and 7.4 % TIA. The median age was 62 years (IQR: 52-70). Stroke admissions rose over time (n = 1975 in 1999-2002 vs. n = 3829 in 2015-2018). Stroke prevalence increased among patients aged 18-40 (p = 0.005) and > 80 (p = 0.002). Risk factors including hypertension, diabetes, atrial fibrillation, smoking, and carotid artery stenosis rose significantly (all p < 0.001). The most frequent comorbidity cluster was hypertension and diabetes (26.3 %), followed by hypertension, diabetes, and dyslipidemia (5.6 %). Unadjusted in-hospital mortality was 9.1 %, declining from 10.4 % to 9.1 % (p = 0.022), primarily due to decreased ICH mortality (24.1 % to 16.4 %, p < 0.001). Mortality predictors included age > 80 (OR 1.70), ICH (OR 3.21), early admission years, complicated diabetes, atrial fibrillation, and multiple comorbidities. CONCLUSION: Stroke patterns are shifting toward younger age and greater risk factor clustering, though in-hospital mortality has declined. Continued surveillance and targeted interventions are essential.