DEATHS FROM PULMONARY TUBERCULOSIS IN A LOW-INCIDENCE COUNTRY

Objectives, To study the validity of official mortality statistics reg arding deaths from pulmonary tuberculosis, and to identify factors con tributing to death. Design. A retrospective study. Setting. Cases were enrolled from the data of the Central Bureau of Statistics from where the official mortality statistics are issued, the National Tuberculos is Register and all Autopsy Registers in the region. Subjects. Case an d autopsy reports from all patients who died from active pulmonary tub erculosis in two Norwegian counties between 1977 and 1989. Main outcom e measures. Patients identified from all three registers with active p ulmonary tuberculosis, concomitant diseases/risk-factors, chest X-rays , symptoms, number of patients investigated and treated for tuberculos is, duration from hospital admission until start of treatment and/or d eath. Results. Ninety-six patients, median age 75 years, died from pul monary tuberculosis, 51 without treatment. Thirty-four patients had no t been registered at the Central Bureau of Statistics. Thirty-nine pat ients had cough on admission. Weight loss and generalized malaise occu rred just as frequently. Forty-two patients had infiltrates on chest X -ray located elsewhere than in the apical region. In 42 patients, no d iagnostic tests for tuberculosis were performed, The median length of stay in hospital was 24 days before death in the untreated group, and 21 days before start of treatment in the treated group. Conclusion. Re liable figures of patients who died from pulmonary tuberculosis could not be obtained from the official statistics because of under-notifica tion and erroneous codification of diseases. Deaths occurred mainly be cause the diagnosis was established too late: in half of the patients at autopsy. Eighty-one patients had concomitant diseases known to lowe r resistance against tuberculosis. Lack of diagnostic suspicion may ha ve been caused by nonspecific symptoms and atypical chest X-ray findin gs.