Chest physiotherapy for collecting sputum samples from HIV-positive patients suspected of having tuberculosis.

SETTING

A public referral hospital in Sao Paulo, Brazil.

OBJECTIVE

To evaluate chest physiotherapy as a means of obtaining sputum samples from human immunodeficiency virus (HIV) positive in-patients suspected of having pulmonary tuberculosis (TB).

DESIGN

A prospective study.

MATERIAL AND

METHODS

Five consecutive samples were collected from 132 patients using the 'spontaneous' technique (ST) on day 1, slow expiration with the glottis open in a lateral posture ('expiration lente totale glotte ouverte en infralatéral', ELTGOL) on day 2, ST on day 3, sputum induction with hypertonic saline (SIHS) on day 4 and ST on day 5. Samples were processed for acid-fast bacilli (AFB) smear and seeded onto Löwenstein-Jensen medium.

RESULTS

Mycobacteria were recovered from 34 patients (25.8%). Nine (26.5%) of the strains were identified as mycobacteria other than Mycobacterium tuberculosis. AFB smear sensitivity was higher in ELTGOL samples than in ST or SIHS samples (52.9% vs. 32.4% and 29.4%), although the difference among the three was not significant (P = 0.098). In culture, the three ST samples proved significantly more sensitive (P = 0.05).

CONCLUSIONS

Physiotherapy shows promise as a technique for obtaining sputum from HIV-positive patients, and AFB testing of a single sample presents high sensitivity. However, this does not preclude the routine collection of three samples, as TB cannot be ruled out before the culture results are known.