Global prevalence of diabetes in active tuberculosis: a systematic review and meta-analysis of data from 2·3 million patients with tuberculosis

Abstract

Background: Although diabetes and poor glycaemic control significantly increase the risk of tuberculosis and adversely affect tuberculosis treatment outcomes, the global burden of diabetes in the context of tuberculosis remains unknown. We did a systematic review and meta-analysis to estimate the prevalence of diabetes among patients with tuberculosis at global, regional, and country levels.

Methods: We searched PubMed, Excerpta Medica Database, Web of Science, and Global Index Medicus to identify studies published between Jan 1, 1986, and June 15, 2017, on the prevalence of diabetes in patients with active tuberculosis, with no language restrictions. Criteria to diagnose tuberculosis and diabetes concurred with WHO guidelines. Methodological quality of eligible studies was assessed, and random-effect models meta-analysis served to obtain the pooled prevalence estimate of diabetes among patients with active tuberculosis, globally. Heterogeneity (I²) was assessed via the χ² test on Cochran's Q statistic. This study is registered with PROSPERO, number CRD42016049901.

Findings: We screened 7565 records of which 200 studies (2 291 571 people with active tuberculosis) were included in meta-analyses. The pooled prevalence of diabetes was 15·3% (95% prediction interval 2·5-36·1; I² 99·8%), varying from 0·1% in Latvia to 45·2% in Marshall Islands. Subgroup and metaregression analyses for identifying sources of heterogeneity showed that four International Diabetes Federation (IDF) regions (North America and Caribbean [19·7%], western Pacific [19·4%], southeast Asia [19·0%], Middle East and North Africa [17·5%]) had significantly higher prevalence estimates than the three others (Africa [8·0%], South and Central America [7·7%], and Europe [7·5%]; p<0·0001). Additionally, the prevalence increased with age, in men, and in countries with low tuberculosis burden. The prevalence of diabetes was decreased in countries that had low incomes and low Human Development Index scores. The form of tuberculosis infection and presence of HIV seemed not to affect the prevalence of diabetes among patients with active tuberculosis.

Interpretation: This study suggests a high burden of diabetes among patients with active tuberculosis, with disparities according to age, sex, regions, level of country income, and development. Cost-effective strategies to curb the burden of diabetes among patients with active tuberculosis are needed.

Funding: None.