Multidomain Autonomic Dysfunction assessed with COMPASS-31 among patients with type 2 diabetes in Zanzibar: prevalence and associated factors

Background Autonomic dysfunction is an important but frequently under-recognized complication of type 2 diabetes mellitus (T2DM), particularly in settings where objective autonomic testing is not routinely available. The Composite Autonomic Symptom Score-31 (COMPASS-31) provides a practical symptom-based approach for assessing autonomic dysfunction across multiple domains. This study compared socioeconomic, lifestyle, and clinical characteristics among adults with T2DM in Zanzibar according to COMPASS-31–defined autonomic dysfunction and examined factors associated with its presence. Methods This hospital-based cross-sectional study included 665 adults with T2DM attending the diabetes clinic at Mnazi Mmoja Hospital, Zanzibar. Autonomic symptoms were assessed using COMPASS-31, including the orthostatic intolerance, gastrointestinal, and bladder domains. Participants were classified as having autonomic dysfunction when they met the prespecified threshold for dysfunction in at least one assessed domain. Cardiovascular autonomic neuropathy (CAN) was assessed separately using standardized cardiovascular autonomic reflex tests performed with an automated CAN Analyzer and was not included in the COMPASS-31 composite outcome. Modified Poisson regression with robust variance estimation was used to calculate adjusted prevalence ratios (aPRs) and 95% confidence intervals (CIs). Results Overall, 309 participants (46.5%) had COMPASS-31–defined autonomic dysfunction. Orthostatic intolerance was the most common domain, affecting 271 participants (40.8%), followed by gastrointestinal dysfunction in 177 (26.6%) and urinary dysfunction in 123 (18.5%). Separately, 305 participants (45.9%) met the analyzer-based criteria for CAN. In the multivariable analysis, autonomic dysfunction was more prevalent among participants without a family history of diabetes (aPR = 1.37, 95% CI: 1.16–1.61; p < 0.001), those with diabetes duration of ≥ 10 years (aPR = 1.45, 95% CI: 1.14–1.75; p < 0.001), those not using oral hypoglycemic agents (aPR = 1.66, 95% CI: 1.39–1.97; p < 0.001), and those who rarely consumed vegetables or consumed one portion per day (aPR = 1.79, 95% CI: 1.37–2.34; p < 0.001). Conclusion Approximately half of the adults with T2DM in this tertiary-care population had COMPASS-31–defined autonomic dysfunction. Higher prevalence was associated with no family history of diabetes, longer diabetes duration, non-use of oral hypoglycemic agents, and lower vegetable intake. These findings highlight the potential value of symptom-based autonomic assessment and support further prospective research to better understand the clinical and lifestyle factors associated with autonomic dysfunction among adults with T2DM in Zanzibar.