Meningitis diagnosis, treatment, and outcomes in rural, northern Uganda: 2015-2024.

Meningitis affects over 2.5 million people worldwide, primarily within resource-limited regions of the meningitis belt of Africa. In 2017, we implemented a cryptococcal meningitis (CM) diagnosis and treatment program (CM-DTP) designed to improve CM outcomes. In 2021, a meningitis diagnosis and treatment program (MEN-DTP) began to include all etiologies of meningitis to assess the impact of expanded diagnosis and treatment of meningitis. We conducted a retrospective study utilizing clinical records from 1443 adult patients admitted to Lira Regional Referral Hospital (LRRH) over three time periods between 2015-2024. Group 1 included 321 patients in a historical control group; Group 2 consisted of 890 patients during the CM-DTP and Group 3 included 232 patients during the MEN-DTP. Group 1 received routine meningitis care, Group 2 received testing and treatment focused on CM, and Group 3 received expanded diagnostic testing, including gram stain, culture, PCR- and antigen-based testing. Meningitis diagnosis and in-hospital mortality outcomes were assessed to evaluate our programs. A confirmed meningitis etiology was found in 13.7% in Group 1, 20.7% in Group 2, and 42.2% in Group 3. In Group 3, confirmed etiologies were identified based on culture (n = 30), Pastorex LA (n = 23), BioFire PCR (n = 56), GeneXpert Ultra (n = 14), and serum or CSF CrAg LFA (n = 53). Overall, more confirmed etiologies of meningitis were identified among people living with HIV (PLWH) (n = 319) compared to those without HIV (n = 22) in Group 3. Antibiotic use increased with pre-admission antibiotic use doubling from Group 1 to Group 3 (10.6% to 27.2%). Compared to Group 1, mortality improved in Groups 2 and 3. Overall, PLWH had increased mortality compared to those without HIV (32.4% vs. 13.3%). Introduction of molecular diagnostics increased meningitis diagnoses and improved outcomes, particularly in those with CM. MEN-DTP increased confirmed diagnoses, yet half remained undiagnosed, supporting investigation of deep sequencing technologies.