National tuberculosis control programmes (NTPs) in sub-Saharan Africa do not routinely record or report treatment outcome data on smear-negative pulmonary tuberculosis (PTB) patients. Twelve-month treatment outcome on patients with smear-negative PTB registered in all district and mission hospitals in Malawi during the year 1995 was collected, and was compared with 8-month treatment outcome in smear-positive PTB patients registered during the same period. Of 4240 patients with smear-negative PTB, 35% completed treatment, 25% died, 9% defaulted and 7% were transferred to another district with no treatment outcome results available. In 24% of patients treatment cards were lost and treatment outcome was unknown. These results were significantly inferior to those obtained in 4003 patients with smear-positive PTB in whom 72% completed treatment, 20% died, 4% defaulted, 2% were transferred and 1% had positive smears at the end of treatment. These differences between patients with smear-negative and smear-positive PTB were similar when analysed by sex and by most age-groups. Higher mortality rates in patients with smear-negative PTB are probably attributable to advanced HIV-related immunosuppression, and higher default and treatment unknown rates probably reflect the lack of attention paid by TB programme staff to this group of patients. As a result of this country-wide study the Malawi NTP has started to record routinely the treatment outcomes of smear-negative TB patients and has set treatment completion targets of 50% or higher for this group of patients.