The impact of comorbidity on mortality following in-hospital diagnosis of tuberculosis

Study objectives: Despite the availability of curative chemotherapy, mortality remains high among patients hospitalized for tuberculosis. Although the elevated mortality rate is often attributed to the presence of multidrug resistant tuberculosis (MDRTB) or concomitant infection with the HIV, other factors must be contributory, especially among the HIV-negative population. Therefore, we performed a study to define the factors associated with mortality following the in-hospital diagnosis of tuberculosis in a region with low levels of MDRTB and coinfection with HIV. Design: Retrospective cohort study. Setting: The eight hospitals in the Barnes-Jewish-Christian (BJC) Health System, which is a network of community and tertiary-care level facilities serving the St. Louis, MO, metropolitan area. Patients: All 203 patients hospitalized with culture-positive tuberculosis at one of the BJC system hospitals between 1988 and 1996. Interventions: Followup information was obtained by telephone interview and review of medical and public health records. Death was verified through a search of the death certificate registry of Missouri and the records of the Social Security Administration. Mortality was defined as death from any cause during the 14 months following the initial date of hospitalization. Measurements and results: The cumulative all-cause mortality rate for this cohort was 28.1%. The incidence of HIV positivity was 7.9% and of MDRTB was 1.5%. Multiple logistic regression analysis demonstrated that respiratory failure requiring mechanical ventilation (adjusted odds ratio [AOR] = 6.5; 95% confidence interval [CI] = 6.0 to 7.0; p < 0.001) and the presence of end-stage renal disease requiring dialysis (AOR = 7.0; 95% CI = 3.7 to 13.3; p = 0.002) were the largest contributors to mortality. Other variables independently associated with mortality included the presence of malnutrition (AOR = 3.2; 95% CI = 2.1 to 4.9; p = 0.007), age > 60 years (AOR = 3.5; 95% CI = 2.4 to 5.2; p < 0.001), drug-induced immunosuppression (AOR = 3.2; 95% CI = 1.6 to 5.2; p = 0.018), and dyspnea at the time of hospital presentation (AOR = 2.1; 95% CI = 1.4 to 3.1; p = 0.048). Overall, 45.3% of the patients had a > 7-day delay in the suspicion of the diagnosis of tuberculosis and the institution of antituberculosis therapy following hospital admission. There was no association between the presence of these delays and mortality. Conclusions: Our data suggest that the 14-month mortality rate is high among patients diagnosed as having tuberculosis during hospitalization, despite low incidences of HIV infection and multidrug resistant disease. The factors that appear to contribute to this elevated mortality rate are markers of disease chronicity and severity of not only the tuberculosis, but also of the patient's underlying health status. Thus, while HIV positivity and multidrug resistance can be important determinants of mortality in some populations, other demographic factors and comorbid conditions may play a role as well. These data also suggest that tuberculosis is often superimposed on chronic illnesses that are important determinants of patient outcomes.