Delayed recognition and infection control for tuberculosis patients in the emergency department

The recent increase in tuberculosis (TB) cases may have an important effect on emergency department infection-control measures. We describe infection-control interventions for TB patients admitted through the ED and hypothesize that ED suspicion of TB is associated with more rapid isolation and treatment.

Retrospective chart review.

The ED of a 400-bed urban, university-affiliated county hospital.

Fifty-five patients with TB culture-positive and acid-fast bacillus stain-positive respiratory specimens who were evaluated in the ED during 1991 and 1992.

We identified cases from the mycobacteriology log. Demographic and historical data and time elapsed before initiation of infection-control measures and TB therapy were recorded. We assessed the relationships of individual clinical findings and the ED presumptive diagnosis of TB (predictor variables) to time elapsed before isolation and therapy (outcome variables) with the log-rank test. The median time (interquartile range) from ED registration to isolation was 8 hours (range, 3 to 13 hours). An ED presumptive diagnosis of TB was made in 71% of cases and was significantly associated with shorter time elapsed before isolation (5 hours [range, 2 to 10 hours] versus 21 hours [range, 11 to 111 hours]; P < .001) and less time elapsed before therapy (12 hours [range, 9 to 22 hours] versus 128 hours [68 to 374 hours]; P < .001). We found TB exposure, radiographic changes typical of TB, absence of HIV risk factors, presence of cough, and sputum production to be associated with more rapid isolation.

Among patients with active pulmonary TB in the ED, TB is often unsuspected and isolation measures are often not used. ED suspicion of TB is associated with more rapid isolation and treatment.