Daily cough frequency in tuberculosis and association with household infection

Although cough in tuberculosis (TB) is presumed to be important for transmission, there is little objective supporting evidence.

To describe 24-h cough frequency in a group with TB, and investigate associations with household rates of infection.

Patients with a new diagnosis of pulmonary TB underwent 24-h cough frequency measurement at or just before initiation of anti-tuberculosis treatment. A group with latent Mycobacterium tuberculosis infection (LTBI) acted as controls. Rates of infection among household contacts of sputum smear-positive TB were measured using the interferon-gamma release assay and the tuberculin skin test, and compared with variables relating to the contacts themselves, and to the index case, including cough frequency.

Daily cough frequency in TB patients (n = 44) was variable (geometric mean [GM] 174, interquartile range [IQR] 68-475 coughs/24 h), higher than in LTBI (n = 17; GM 19 coughs/24 h, IQR 8-53; P < 0.001), and higher during the day than overnight (GM 8.9 coughs/h, IQR 4.1-19.0 vs. GM 2.9 coughs/h, IQR 0.7-13.4; P < 0.0001). Also, 24-h cough frequency in TB was associated with sputum smear status (P = 0.040), but not smoking (P = 0.475). Multivariable logistic regression confirmed that infection in contacts was independently associated with index case sputum smear grade (P = 0.014) and cough frequency (P = 0.022).

Measurement of 24-h cough frequency in pulmonary TB helps predict infectiousness and transmission patterns.