Type 2 diabetes and tuberculosis in a dynamic bi-national border population

Abstract

The epidemic of type 2 diabetes in the United States prompted us to explore the association between diabetes and tuberculosis (TB) on the South Texas-Mexico border, in a large population of mostly non-hospitalized TB patients. We examined 6 years of retrospective data from all TB patients (n=5049) in South Texas and northeastern Mexico and found diabetes self-reported by 27.8% of Texan and 17.8% of Mexican TB patients, significantly exceeding national self-reported diabetes rates for both countries. Diabetes comorbidity substantially exceeded that of HIV/AIDS. Patients with TB and diabetes were older, more likely to have haemoptysis, pulmonary cavitations, be smear positive at diagnosis, and remain positive at the end of the first (Texas) or second (Mexico) month of treatment. The impact of type 2 diabetes on TB is underappreciated, and in the light of its epidemic status in many countries, it should be actively considered by TB control programmes, particularly in older patients.

Fig. 1

Self-reported DM is more frequent among TB patients than in the general population, stratified by age. (a) Texas. –■–, Patients with self-reported DM and TB (1996–2002); - -◆- -, randomly selected participants with self-reported DM in a survey of a subset (Lower Rio Grande Valley) from the same population [5]. (b) Mexico. –■–, Patients with self-reported DM and TB in Tamaulipas, Mexico (1998–2003); - -◆- -, randomly selected participants in a population survey in Mexico (2000) [6].

Fig. 2

Differences in age distribution between diabetic and non-diabetic TB patients from Mexico and Texas. Bar graphs illustrate the number of TB patients by 5-year age intervals in Mexico and Texas with DM (■) and without DM (□).

Fig. 3

Patients with DM and TB take longer to become direct smear negative upon treatment initiation. Percent of TB patients in (a) Texas and (b) Mexico whose direct smear for acid-fast bacilli by month of antimycobacterial therapy. Percentage figures are based on the total number of patients with productive cough available for direct smear evaluation for each month. Solid line, TB patients with DM. Hatched line, TB patients without DM; vertical solid/hatched lines, 95% confidence intervals.