Diabetes in a TB and HIV-endemic South African population: Analysis of a virtual cohort using routine health data

Abstract

Background: It is widely accepted that people living with diabetes (PLWD) are at increased risk of infectious disease, yet there is a paucity of epidemiology studies on the relationship between diabetes and infectious disease in SSA. In a region with a high burden of infectious disease, this has serious consequences for PLWD.

Methods and findings: Using routinely collected longitudinal health data, we describe the epidemiology of diabetes in a large virtual cohort of PLWD who have a high burden of HIV and TB, from the Khayelitsha subdistrict in the Western Cape Province in South Africa. We described the relationship between previous TB, newly diagnosed TB disease and HIV infection on diabetes using HbA1c results as an outcome measure. The study population was predominately female (67%), 13% had a history of active TB disease and 18% were HIV positive. The HIV positive group had diabetes ascertained at a significantly younger age (46 years c.f. 53 years respectively, p<0.001) and in general had increased HbA1c values over time after their HIV diagnosis, when compared to the HIV-negative group. There was no evidence of TB disease influencing the trajectory of glycaemic control in the long term, but diabetes patients who developed active TB had higher mortality than those without TB (12.4% vs 6.7% p-value < 0.001). HIV and diabetes are both chronic diseases whose long-term management includes drug therapy, however, only 52.8% of the study population with an HIV-diabetes comorbidity had a record of diabetes treatment. In addition, the data suggest overall poor glycaemic control in the study population with only 24.5% of the participants having an HbA1c <7% at baseline despite 85% of the study population being on diabetes treatment.

Conclusion: The epidemiologic findings in this exploratory study highlight the need for further research into diabetes outcomes in a high TB and HIV burden setting and demonstrate that routine health data are a valuable resource for understanding disease epidemiology in the general population.

Fig 1. Flow chart showing the selection of the study population from the PHDC routine health data.

Fig 2. Bar graph showing new diabetes cases (bars) from January 2011 to December 2017 overlaid with line plots of new Tuberculosis (solid line) and HIV (dashed line) cases in these diabetes patients in the same time period.

Fig 3. Distribution of repeat Tuberculosis (TB) episodes in the study population before and after diabetes ascertainment.

Fig 4. Effect of Tuberculosis ascertainment on HbA1c (%) over a 5-year period.

A. HbA1c plotted by diabetes treatment i.e. on diabetes treatment (grey circle) or not on diabetes treatment (dark-red star). B. HbA1c plotted by TB episode i.e. 1 episode (grey circle), 2 episodes (pale-green square), 3 episodes (steel-blue star) or 4 episodes (dark-red diamond).

Fig 5. Effect of HIV ascertainment on HbA1c (%) over a 5-year period.

A. HbA1c plotted by diabetes treatment i.e. on diabetes treatment (grey circle) or not on diabetes treatment (dark-red star). B. HbA1c plotted by TB history i.e. never had TB (grey circle) or have had TB (dark-red star).