Risk Factors in the progression from tuberculosis infection to disease

Abstract

Tuberculosis (TB) is a two-stage disease, acquisition of infection and progression to disease. A complex interaction exists between the individual and their environment that determines who acquires infection and who progresses to disease. According to TB literature, 10% of individuals with infection will develop TB disease (1;2). Tuberculosis has been described a disease of poverty, but other factors may be important. The contribution of both individual measures, such as ethnic origin, gender and age and area-level measures, or socio-economic factors, to this two-stage process is not well understood. Understanding tuberculosis epidemiology and identifying those at risk for developing TB is important for effectively controlling the disease. The objective of this study was to determine the individual (age, gender, ethnic origin, geographic location) and area-level measures (income, home ownership, housing density, education, and employment) that contribute to the progression from tuberculosis infection to disease. Data from all Canadian-born Caucasians, Status Indians, and non-Status Indians and Metis, with an initial positive tuberculin skin test (TST) documented in the Saskatchewan TB Control database from January 1, 1986 to January 31, 2002 was analyzed. Exclusion criteria included any previous BCG vaccination, treatment for latent TB infection, or missing data. Individual data was obtained from the TB Control database. Area-level measures were obtained by matching individual postal codes with Canada census data to obtain information from enumeration areas. Outcome was time to TB disease at > 1 month following a documented positive tuberculin skin test. Analysis was completed using Cox regression proportional hazards model. 7588 individuals with a positive tuberculin skin test were included in the study and of these 338 (4.5%) developed TB disease. Thirty-four out of 4140 (0.8%) of Caucasians, 183 out of 2649 (6.9%) of Status Indians and 121 out of 799 (15.1%) non-Status Indians and Metis developed TB. The rate of progression to TB was 5.6/1000 person years for the entire study population. The incidence for Caucasians was 0.9/1000 person years, 7.7/1000 person years for Status Indians and 16.0/1000 person years for non-Status Indians and Metis. In the Cox regression model, including individual and area-level measures, the risk factors association with the progression to TB was age and ethnic origin (< 19 years of age HR 3.7, 95% CI 2.8 - 4.8 compared to > 19 years and ethnic origin HR 5.1, 95% CI 3.0 - 8.6 for Status Indians and HR 7.4, 95% CI 4.1-13.3 for non-Status Indians and Metis both compared to Caucasians). No socio-economic factor was consistently associated with progression to disease. We have found that age and ethnic origin are associated with an increased risk of TB infection progressing to disease. The differences in TB rates between Saskatchewan Caucasians and Status Indians, non-Status Indians and Metis can be explained by Grigg's natural history curve of TB epidemiology within a population (3). The Aboriginal population of Saskatchewan is much earlier in its epidemic resulting in higher disease rates compared to the Caucasian population. Identifying those at risk of developing TB and understanding the determinants of TB epidemiology are important for establishing successful TB control programs.