Nearly Half of Adults with Type 2 Diabetes in Zanzibar Show Multidomain Autonomic Dysfunction

Nearly half of adults living with type 2 diabetes attending a tertiary referral hospital in Zanzibar were found to have autonomic dysfunction affecting at least one organ system, according to a hospital-based cross-sectional study published in Cardiovascular Diabetology – Endocrinology Reports. Among 665 adults with type 2 diabetes mellitus attending the diabetes clinic at Mnazi Mmoja Hospital, 309 participants, or 46.5 percent, met the study’s prespecified criteria for dysfunction in at least one of three symptom domains assessed with the Composite Autonomic Symptom Score-31, or COMPASS-31, questionnaire. The finding places the burden of a frequently under-recognized diabetic complication at a level the authors describe as high within this tertiary-care population, and it arrives with a detailed map of the clinical, treatment-related, and dietary characteristics associated with the condition.

The study was conducted at the diabetes clinic of Mnazi Mmoja Hospital, the main tertiary referral hospital in Zanzibar, which serves patients predominantly from the Urban West Region of Unguja. Data collection ran from 20 December 2023 to 15 August 2024. Adults aged 18 years or older with a confirmed diagnosis of type 2 diabetes, permanent residence in Zanzibar, and current use of at least one glucose-lowering medication were eligible. The researchers excluded individuals with severe medical conditions likely to affect autonomic function independently of diabetes, including severe liver disease and chronic kidney disease, as well as pregnant women and those within 12 weeks postpartum. Participants were recruited during routine clinic visits using systematic random sampling, and written informed consent was obtained before data collection. The analysis forms part of a broader research program on autonomic dysfunction in Zanzibar, and the authors note that earlier publications reported interim, domain-specific analyses, whereas the current study used the completed dataset of 665 participants.

Autonomic symptoms were assessed using COMPASS-31, a structured symptom-based questionnaire administered face-to-face by trained research personnel. Although the full questionnaire was administered, the present analysis focused on three domains: orthostatic intolerance, gastrointestinal symptoms, and bladder symptoms. Orthostatic intolerance captures symptoms such as dizziness, light-headedness, faintness, or cognitive difficulty occurring after standing, and was scored from four items with a weighted score ranging from 0 to 40, where any score above zero indicated the presence of the condition. Gastrointestinal autonomic dysfunction was assessed from symptoms including bloating, vomiting, early satiety, abdominal discomfort, constipation, and diarrhea, with cumulative scores categorized as absent, mild, moderate, or severe. Urinary autonomic dysfunction was assessed from loss of bladder control, difficulty passing urine, and incomplete bladder emptying, using a similar severity categorization. Participants were classified as having autonomic dysfunction when dysfunction was identified in at least one of the three domains.

The domain-specific results revealed a clear hierarchy of symptom burden. Orthostatic intolerance was the most common manifestation, affecting 271 participants, or 40.8 percent. Gastrointestinal autonomic dysfunction followed in 177 participants, or 26.6 percent, and urinary autonomic dysfunction affected 123 participants, or 18.5 percent. When the number of affected domains was tallied, 28.2 percent of all participants had one affected domain, 14.1 percent had two, and 4.2 percent had all three. Separately from the COMPASS-31 composite outcome, the team assessed cardiovascular autonomic neuropathy using standardized cardiovascular autonomic reflex tests performed with an automated CAN Analyzer. Testing included resting heart rate, heart-rate response to deep breathing, heart-rate response to standing, the Valsalva manoeuvre, postural blood-pressure response, and sustained handgrip. Using this instrument-based approach, 305 participants, or 45.9 percent, met the analyzer-based criteria for cardiovascular autonomic neuropathy, with early, definite, and severe disease observed in 34.7 percent, 10.1 percent, and 1.1 percent of participants, respectively. The authors emphasize that this analyzer-defined outcome was reported descriptively and was not included in the primary COMPASS-31-based multidomain outcome.

The participant population was characterized in detail. The mean age was 56.9 years with a standard deviation of 10.8 years; 8.2 percent of participants were younger than 40 years, 51.6 percent were aged 40 to 60 years, and 40.2 percent were older than 60 years. A total of 27.1 percent had lived with diabetes for 10 years or more, while 49.1 percent, 13.4 percent, and 10.4 percent had durations of 1 to 3, 4 to 6, and 7 to 9 years, respectively. A family history of diabetes was reported by 66.0 percent of participants, and oral hypoglycemic agents were used by 58.9 percent. For vegetable intake, 51.9 percent reported rarely consuming vegetables or consuming one portion per day, while 48.1 percent consumed at least two portions per day. Fruit and vegetable intake was assessed by self-reported number of portions per day, with a portion representing the participant’s usual self-defined serving size, since a standardized portion size in grams was not specified or directly measured.

To identify factors associated with autonomic dysfunction, the researchers used modified Poisson regression with robust variance estimation, calculating adjusted prevalence ratios with 95 percent confidence intervals. All variables evaluated in the bivariable analysis were entered simultaneously into the multivariable model, irrespective of their bivariable p-values. Four factors emerged as significantly associated with a higher prevalence of autonomic dysfunction. Participants without a family history of diabetes had a higher prevalence than those with a family history, with an adjusted prevalence ratio of 1.37 (95 percent confidence interval: 1.16 to 1.61; p < 0.001). A diabetes duration of 10 years or more was associated with a higher prevalence compared with shorter durations, with an adjusted prevalence ratio of 1.45 (95 percent confidence interval: 1.14 to 1.75; p < 0.001). Participants not using oral hypoglycemic agents had a higher prevalence than those using them, with an adjusted prevalence ratio of 1.66 (95 percent confidence interval: 1.39 to 1.97; p < 0.001). Finally, participants who rarely consumed vegetables or consumed one portion per day had a higher prevalence than those consuming two or more portions per day, with an adjusted prevalence ratio of 1.79 (95 percent confidence interval: 1.37 to 2.34; p < 0.001).

Several of these associations deserve closer attention. The link between the absence of a family history of diabetes and higher autonomic dysfunction prevalence was an additional finding in the present study, and the authors suggest it may reflect differences in diabetes awareness, healthcare-seeking behavior, or earlier engagement with diabetes care, although these potential explanations were not directly assessed. Previous studies have shown that individuals with a family history of diabetes have greater awareness of diabetes-related risk and are more likely to undergo screening and adopt risk-reducing behaviors, while those without family history may underestimate risk, leading to delayed diagnosis and poorer self-management. The authors note that these findings support assessing autonomic symptoms regardless of family history of diabetes. Longer diabetes duration is consistent with evidence linking prolonged diabetes to greater complication burden and higher neuropathy prevalence, and the authors state it may reflect cumulative metabolic exposure and progressive autonomic nerve injury over time. However, they also observe that the lack of a consistent gradient across intermediate categories in their data suggests the association may not increase uniformly with diabetes duration, and the cross-sectional design does not allow identification of a specific duration threshold for autonomic nerve injury.

The association with oral hypoglycemic agent non-use and with vegetable intake was interpreted with similar caution. The authors state that the association with non-use of oral hypoglycemic agents may reflect differences in disease stage, glycemic exposure, or treatment patterns rather than a direct protective effect of the medications, noting that intensive glycemic control has been shown to reduce the risk of microvascular complications; they conclude that the association likely reflects better disease status, underscoring the importance of early diagnosis, appropriate treatment, and sustained glycemic control. On diet, the authors highlight vegetable consumption as a potentially modifiable correlate, noting that fruits and vegetables provide vitamins, antioxidants, and other bioactive compounds that may support vascular and neural health through antioxidant, anti-inflammatory, and endothelial mechanisms. Fruit intake was not significantly associated with autonomic dysfunction in the present analysis, although a previous study in the same population found that consuming two or more portions of fruit per day was associated with lower odds of gastrointestinal autonomic dysfunction, a difference the authors attribute to variation across autonomic domains and outcome definitions. The present findings particularly emphasize vegetable consumption, rather than fruit intake, as the dietary factor associated with autonomic dysfunction.

The authors also situate their prevalence estimate in a wider context, reporting that the 46.5 percent figure is comparable to global estimates and places this tertiary-care population above the commonly reported range of 20 to 40 percent for autonomic neuropathy in diabetic populations, while one previous study reported a prevalence of autonomic impairment of 73 percent among patients with type 2 diabetes. They attribute differences in prevalence across studies to variations in study populations, clinical settings, diagnostic approaches, and participant characteristics. The study’s limitations are stated plainly: the cross-sectional nature limits the ability to infer causal relationships, highlighting the need for future longitudinal studies or clinical trials to clarify underlying mechanisms; reliance on self-reported data for alcohol consumption and dietary intake may have introduced recall and social desirability bias; and the single-center design warrants caution when generalizing findings to other regions or populations with differing sociodemographic or healthcare contexts. Within those constraints, the authors conclude that the findings highlight the importance of clinical and lifestyle-related factors in autonomic dysfunction and support the need for early identification and appropriate management among patients with type 2 diabetes, with symptom-based autonomic assessment offering a practical approach in settings where specialized autonomic testing is not routinely available.

Subject of Research: Prevalence and associated factors of multidomain autonomic dysfunction assessed with COMPASS-31 among adults with type 2 diabetes in Zanzibar

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

Article References: Haji, H. T., Ali, R. M., Khamis, A. G., Ramaiya, K., Munseri, P., & Mashili, F. L. (2026). Multidomain Autonomic Dysfunction assessed with COMPASS-31 among patients with type 2 diabetes in Zanzibar: prevalence and associated factors. Cardiovascular Diabetology – Endocrinology Reports, 12(1), Article 69. https://doi.org/10.1186/s40842-026-00340-1

Image Credits: AI Generated

DOI: 10.1186/s40842-026-00340-1

Keywords: type 2 diabetes, autonomic dysfunction, COMPASS-31, cardiovascular autonomic neuropathy, orthostatic intolerance, Zanzibar, cross-sectional study, prevalence, diabetic neuropathy, vegetable intake, oral hypoglycemic agents, tertiary care