Background and aim: Severe acute respiratory infection (SARI) is a syndromic concept widely used for respiratory disease surveillance and burden estimation. In settings where routine clinical case definition fields are not consistently recorded, ICD-10 discharge diagnoses can be used as a pragmatic proxy to quantify inpatient burden and severe outcomes. The aim is to describe the 2025 inpatient burden, outcomes, sex distribution, and seasonal patterns of selected respiratory diseases commonly used to represent SARI in Sri Lanka. Method: The eIMMR (electronic Indoor Morbidity and Mortality Register) data extracted from January–December 2025, containing monthly aggregated inpatient admissions by discharge outcome and sex for four ICD-10 groupings aligned to SARI: influenza (J09–J11), pneumonia (J12–J18), other acute lower respiratory infections (ALRI; J20–J22), and ARDS (J80). A hierarchical SARI grouping was defined for analysis and reporting. The annual admissions and deaths, case fatality proportion (CFP; deaths/admissions), transfers, leave against medical advice (LAMA), sex distribution, and peak months were summerized. Results: In 2025, the selected SARI-coded respiratory diagnoses accounted for 232,679 admissions and 6,432 in-hospital deaths (overall CFP 2.8%); 49.2% of admissions were female. Other ALRI (J20–J22) contributed 83.9% of admissions (195,237) but 12.6% of deaths (812; CFP 0.4%). Pneumonia (J12–J18) accounted for 14.0% of admissions (32,543) yet 86.6% of deaths (5,573; CFP 17.1%). Influenza (J09–J11) comprised 4,700 admissions (CFP 0.3%). ARDS (J80) was uncommon (199 admissions) but had high CFP (16.6%). Admissions peaked in April–May (seasonal maximum May: 29,908) and showed a secondary rise in October–November. Conclusion: In 2025, SARI-proxy inpatient burden in Sri Lanka was dominated by ALRI volume, while mortality was overwhelmingly concentrated among pneumonia-coded admissions. Strengthening pneumonia prevention, early severity recognition, oxygen/critical care readiness, and improving etiologic diagnosis and coding completeness are key priorities for reducing SARI-associated mortality and improving surveillance precision.