TB disease yield from household contact screening of tuberculosis index patients; a cohort study from Karachi, Pakistan

Abstract Nearly 40% of people affected by TB in Pakistan are not diagnosed each year. Guidelines recommend screening household contacts however, not all index patients or contacts are eligible. Therefore, many contacts who may have TB disease, remain unscreened. We conducted a prospective cohort study under programmatic conditions in Karachi, Pakistan from January 2018 - December 2019, to screen all household contacts of all TB index patients. We disaggregated these according to guidelines into eligible (those with bacteriologically confirmed pulmonary TB or children <5 years) or ineligible (those with clinically diagnosed or extrapulmonary TB ≥5 years) index patients, and eligible (children <5 years or symptomatic individuals) or ineligible (asymptomatic individuals ≥5 years) contacts. We calculated TB disease yields for different groups of index patients and contacts. Out of 39,168 household contacts from 6,362 index patients, 21,035 completed clinical assessments for TB disease, and 416 were diagnosed with all forms TB. Household contacts of clinically diagnosed pulmonary TB patients were 26% more likely to be diagnosed with TB compared to the household contacts of bacteriologically confirmed pulmonary TB (adjusted Odds Ratio 1.26 [1.01 – 1.59] p-value:0.03). The yield of TB disease among child contacts (3.4%) was significantly higher than the yield among adult contacts (0.5%) ( p -value:<0.001). Broadening TB contact screening guidelines to include clinically diagnosed and extrapulmonary index patients ≥5 years could double the number of patients detected at a similar level of effort.