A 57 year old African woman presented with a 2-week history of fever, cough, and dyspnea. She had no sick contacts, chest pain, hemoptysis or weight loss and was HIV negative. Past medical history was unremarkable. On physical examination, temperature was 100.3 F and RR was 19. Auscultatory chest findings were positive for crackles in the left upper lobe. CXR and CT scan showed evidence of necrotizing pneumonia, and a 4.4 cm cavity in the left upper lobe. A presumptive diagnosis of community acquired cavitary pneumonia was made; she was started on Ceftriaxone and Azithromycin. Patient could expectorate only 1 sputum sample, which was negative for AFB and cultures. Blood cultures, urine Legionella antigen and rapid influenza testing were negative. Bronchoscopy showed no acute inflammation and bronchial washings showed a copious number of leukocytes, but no AFB. After 4 days, leukocytes began trending downward but the low grade fever continued. Antibiotics were changed to Vancomycin and Piperacillin-Tazobactam. The patient was clinically better on Day 7 and discharged on Augmentin.
2 months later, the patient presented again with the same symptoms of fever, cough and dyspnea. Repeat CT showed a persistent left upper lobe cavity and a new cavity in the right upper lobe. Vancomycin, Piperacillin-Tazobactam and Levofloxacin were started. A more extensive workup for carcinoma, autoimmune diseases, fungal pneumonia and aspiration was performed, all of which were unremarkable. Bronchoscopic biopsy was also negative. The patient was discharged on day 6 with oral antibiotics.
After 5 months the patient was readmitted a third time with the same signs and symptoms. While thoroughly reviewing the previous workup, it was noted that 3 sputum samples had never been obtained. Sputum was subsequently induced, which was positive for AFB, and gene probes confirmed Mycobacterium Tuberculosis (MTB). Sputum cultures were positive in both solid and liquid media. The patient was treated with appropriate anti-tuberculosis drugs. She was afebrile by day 4, and discharged on day 7. The patient has been asymptomatic for more than a year now.
Discussion:
Studies in Canada and New Zealand have consistently shown that sensitivity for AFB by sputum is close to 80% in cavitary disease, whereas sensitivity of Bronchoscopy is only 65%. The exception to this is in HIV positive patients, in whom the sensitivity of bronchoscopy is higher.
The delay in diagnosis was multifactorial. The one negative sputum sample was misleading as were the negative bronchoscopy findings. Also, several antibiotics used for bacterial pneumonia, like Levofloxacin and Amoxicillin-clavulanic acid have antimycobacterial activity. Hence the patient clinically improved briefly, only to deteriorate again. Finally, the fact that 40% of hospitalized patients with community-acquired pneumonia have no specific microbiological diagnosis also led us to persist with empiric antimicrobial treatment. The dangers of this are significant- the emergence of drug-resistant MTB and also by discharging the patient, we exposed the community to a patient with active TB. The average cost per admission has been estimated at more than $5,000. Inducing sputum would have led to a more timely diagnosis and treatment.
Conclusions:
In patients in whom the clinical suspicion of Tuberculosis is high, 3 sputum samples should be obtained and tested for AFB and culture. If the patient is unable to expectorate, induced samples or even post bronchoscopy samples should be collected.