This retrospective observational target-trial emulation uses electronic health record data from the TriNetX US Collaborative Network to compare preadmission glucose-lowering treatment strategies in adults with type 2 diabetes hospitalized with pneumonia. Time zero is the date of the index pneumonia admission. The study compares patients with evidence of prescriptions for both a GLP-1 receptor agonist and an SGLT2 inhibitor in the year before admission with patients prescribed a GLP-1 receptor agonist alone, an SGLT2 inhibitor alone, or usual care. Usual care includes metformin, sulfonylureas, or DPP-4 inhibitors without either study class. The primary outcome is a composite of all-cause death, invasive mechanical ventilation, or renal replacement therapy within 30 days of admission, with the individual components and 12-month mortality, major adverse cardiovascular events, and major adverse kidney events also evaluated. Propensity-score methods are used to reduce bias from nonrandom treatment selection.

Inclusion Criteria:

• Adults aged 18 years or older.
• Type 2 diabetes mellitus
• Hospitalization with pneumonia during the study period, serving as the index admission
• Prescription of a GLP-1 receptor agonist, an SGLT2 inhibitor, or usual-care glucose-lowering therapy in the year before the index admission
• No GLP-1 receptor agonist or SGLT2 inhibitor prescription in the 6 months before the first qualifying prescription

Exclusion Criteria:

• Type 1 diabetes mellitus, or other specified diabetes types that are not type 2 diabetes
• Human immunodeficiency virus infection
• Bariatric surgery
• Solid-organ transplantation
• Pneumonia hospitalization, invasive mechanical ventilation, or renal replacement therapy in the year before the index admission
• End-stage kidney disease, dialysis dependence, ventilator dependence, or tracheostomy status
• Metastatic malignancy, lung cancer, or palliative care
• Viral pneumonia, influenza with pneumonia, or COVID-19 around the index admission
• Acute cardiovascular event in the 6 months before the index admission

Adults will be selected from the TriNetX US Collaborative Network, a distributed database of de-identified electronic health records contributed by participating healthcare organizations across multiple clinical systems and practice settings. The study population consists of adults with type 2 diabetes who were hospitalized with pneumonia and received routine clinical care in this network, identified through diagnosis records, laboratory data, encounter records, and medication prescribing data. Comparator-specific cohorts were defined within this source population according to glucose-lowering prescriptions recorded in the year before the index admission, to evaluate preadmission GLP-1 receptor agonist- and SGLT2 inhibitor-based treatment strategies.

Individual participant data will not be shared. This retrospective observational study uses de-identified electronic health record data from the TriNetX US Collaborative Network. Access to individual-level data is restricted by data use agreements, institutional policies, and privacy protections. Researchers who meet eligibility requirements may obtain access to similar de-identified data through a TriNetX license or through participating institutions.