Progression to tuberculosis disease increases with multiple exposures

Abstract

During a single year, a Canadian village had 34 individuals with microbiologically confirmed tuberculosis (TB) among 169 people with a new infection (20%). A contact investigation revealed multiple exposures for each person. We investigated whether the intensity of exposure might contribute to this extraordinary risk of disease.We carried out a case-control study using a public health database. Among those with a new infection, 34 had culture-confirmed TB (cases) and 118 did not progress to disease (controls). 17 patients with probable disease were excluded. Contact investigation data were utilised to tabulate the number of potential sources (total exposures). Generalised estimating equations with a logit link were used to identify associations between exposures and progression, and to investigate other potential risk factors.The median (interquartile range) number of total exposures was 15 (3-23) for cases and 3 (2-12) for controls (p=0.001). The adjusted OR for disease was 1.11 (95% CI 1.06-1.16) per additional exposure, corresponding to an OR of 3.4 for disease when comparing the medians of 15 versus 3 total exposures. This association increased when restricting to tuberculin skin test conversions.Increased exposure could be a marker of greater risk of progression to TB disease. Therefore, this risk may not be transportable across epidemiologic settings with variable exposure intensities.

Multiple Exposures, Reinfection, and Risk of Progression to Active Tuberculosis

“…The increase in disease risk with increasing exposures found in (2) could be explained by an independent effect of each infection on risk of progression (reinfection model), or by a high risk of progression once the cumulative number of exposures exceeds a threshold value (threshold model). Previous work has shown that the effect of increasing disease risk with increasing exposures persists after adjustment for a number of covariates (2). Here we found that the threshold model conferred an improved fit relative to the increasing-risk model, providing evidence for a threshold effect.…”

“…Based on previously published TB models (8), we developed binomial risk models for TB infection during the first year after infection and fit these models to data from a TB contact investigation in a largely Inuit (>90%) village of 933 people. Study design and data collection are described in detail elsewhere (2,13). Data consisted of the number of TB exposures for 149 recently-infected individuals, and whether these individuals developed active TB disease within the following year.…”

“…In Canada, TB disease rates are considerably higher in the Inuit population than in the non-aboriginal Canadian-born population (1). A recent study reported on a TB outbreak in a largely Inuit village, in which 34 out of 149 individuals (23%) progressed to active disease within a year following recent infection (2). This risk is much higher that the oft cited 2-5% risk of progression following recent infection (3)(4)(5).…”

“…However, the dynamic impact of numerous, tightly spaced exposures and attendant infections has not been explored. Using binomial risk models (9) and data from (2), we evaluated two plausible mechanisms by which reinfection during this early, critical period can account for the increased disease risk: an independent effect of each infection on risk of progression (reinfection model) and a high risk of progression once the cumulative number of exposures exceeds a threshold value (threshold model). Contacts of TB cases producing aerosols with high burdens of Mycobacterium tuberculosis are more likely to become cases (10)(11)(12), suggesting an inoculum effect as a plausible biological mechanism for both these models.…”

Previously undetected superspreading ofMycobacterium tuberculosisrevealed by deep sequencing

“…This outbreak, which has been previously described, 4,7 comprised 50 microbiologically-confirmed cases of TB who were diagnosed in a single Inuit community between 2011-2012 - a rate of 5,359/100,000 for that year. Genomic epidemiology analyses using sequencing depths of ∼50x that are standard in such work, identified multiple clusters of transmission in this outbreak, 4 however, there was insufficient genetic variation detected to infer precise person-to-person transmission events within these subgroups.…”