Background
Tuberculosis and HIV co-infection is one of the main drivers of poor outcome for both diseases in Zambia. HIV infection has been found to predict TB infection/disease and TB has been reported as a major cause of death among individuals with HIV. Improving case detection of TB/HIV co-infection has the potential to lead to early treatment of both conditions and can impact positively on treatment outcomes. This study was conducted in order to determine the HIV prevalence among adults with tuberculosis in a national prevalence survey setting in Zambia, 2013–2014.

Methods
A countrywide cross sectional survey was conducted in 2013/2014 using stratified cluster sampling, proportional to population size for rural and urban populations. Each of the 66 countrywide clusters represented one census supervisory area with cluster size averaging 825 individuals. Socio-demographic characteristics were collected during a household visit by trained survey staff. A standard symptom-screening questionnaire was administered to 46,099 eligible individuals across all clusters, followed by chest x-ray reading for all eligible. Those symptomatic or with x-ray abnormalities were confirmed or ruled out as TB case by either liquid culture or Xpert MTBRif performed at the three central reference laboratories. HIV testing was offered to all participants at the survey site following the national testing algorithm with rapid tests. The prevalence was expressed as the proportion of HIV among TB cases with 95% confidence limits.

Results
A total of 265/6123 (4.3%) participants were confirmed of having tuberculosis. Thirty-six of 151 TB survey cases who accepted HIV testing were HIV-seropositive (23.8%; 95% CI 17.2–31.4). The mean age of the TB/HIV cases was 37.6 years (range 24–70). The majority of the TB/HIV cases had some chest x-ray abnormality (88.9%); were smear positive (50.0%), and/or had a positive culture result (94.4%). None of the 36 detected TB/HIV cases were already on TB treatment, and 5/36 (13.9%) had a previous history of TB treatment. The proportion of TB/HIV was higher in urban than in the rural clusters. The HIV status was unknown for 114/265 (43.0%) of the TB cases.

Conclusions
The TB/HIV prevalence in the general population was found to be lower than what is routinely reported as incident TB/HIV cases at facility level. However; the TB/HIV co-infection was higher in areas with higher TB prevalence. Innovative and effective strategies for ensuring TB/HIV co-infected individuals are detected and treated early are required.