What is in a cough?

The aims of this talk are to: (a) provide an overview of our method of collecting, quantifying, and sizing infectious aerosols of Mycobacterium tuberculosis; and (b) to review data indicating that cough aerosol cultures of M. tuberculosis are the best predictors of infection and incident disease among household contacts of persons with active tuberculosis (TB). New infection was defined as tuberculin skin test conversion.

We developed a cough aerosol sampling system by placing two Andersen cascade impactor viability samplers inside a cylinder into which patients cough via connector tubing. We recruited sputum acid fast bacilli (AFB) smear-positive patients from the tuberculosis clinic and wards at Mulago Hospital in Kampala, Uganda. Patients were asked to cough as strongly and frequently as they comfortably could for two 5-min sessions of coughing.

In a cohort of 96 sputum culture positive index TB cases, 43 (45%) produced culture-positive cough aerosols. Household contacts of TB patients who produced high aerosols (⩾10colony forming units (CFU)) were more likely to have a new infection compared with contacts with low aerosol CFU (1-9CFU) and aerosol-negative cases (69%, 25%, and 30%, respectively, p=0.009). In adjusted multivariate analyses, high cough aerosols were the only predictor of new TB infection (odds ratio [OR] 4.81; 1.20-19.23). In a follow-up (median 3.9years) of this cohort, 369 (84%) of the contacts could be traced; eight (2%) had developed TB disease. Incident TB disease was associated with larger bacillary load in sputum measured by days to positive in liquid culture (OR 7.9; 0.7-70.5), exposure to a high-aerosol TB case patient (OR 6.0, 1.4-25.2) and marginally to HIV infection in the contact (OR 7.2; 0.7-70.5). Cough aerosol studies of TB patients in Brazil and South Africa are ongoing and appear to be finding similar proportions of cough aerosol cultures among TB patients (personal communication).

Cough aerosol cultures of M. tuberculosis are the best predictors of infectiousness and predict incident TB disease among sputum smear-positive patients in Uganda. We propose that cough aerosol cultures are a better surrogate of inhaled dose than sputum smear.