In Malawi, it has been the practice for several years to obtain sputum for
smear microscopy of acid-fast bacilli (AFB) from all patients with extrapul
monary tuberculosis (EPTB). We audited this practice, and determined in pat
ients aged greater than or equal to 15 years (i) the proportion of EPTB pat
ients who had sputum smears examined, (ii) the number of sputum smears exam
ined per patient, and (iii) the proportion of patients with EPTB who had sp
utum samples smear positive for AFB. Forty-one hospitals (3 central, 22 dis
trict and 16 mission) performing smear microscopy and registering EPTB pati
ents were visited in 1998 and 1999, and a retrospective and prospective stu
dy was carried out using TB registers and laboratory sputum registers. In t
he retrospective study, 1124 (69%) of the 1637 patients with EPTB had sputu
m smears examined; 988 (88%) of the 1124 submitted 3 sputum specimens. In t
he prospective study, 2026 (84%) of the 2411 patients with EPTB had sputum
smears examined: 94% of the 2026 submitted 3 sputum specimens. In both stud
ies, high rates of sputum submission were found in patients with pleural ef
fusion, miliary TB, lymphadenopathy and pericardial effusion. In the prospe
ctive study, only 34 (1.7%) EPTB patients submitting sputum were smear posi
tive, and the proportion who were smear positive exceeded 3% only in patien
ts with lymphadenopathy, miliary TB and TB meningitis. As a result of this
study, the Malawi TB Control Programme has changed its policy, and now only
insists on sputum-smear examination if patients with EPTB have a cough for
>3 weeks. These policy changes will be audited by further operational rese
arch.