Smoking and HIV associated with subclinical tuberculosis: analysis of a population-based prevalence survey.

BACKGROUND: Despite multiple tuberculosis (TB) prevalence surveys reporting a relatively high frequency of bacteriologically confirmed, active TB among individuals reporting no typical symptoms of disease, our understanding of this phenomenon is limited.OBJECTIVE: To quantify the epidemiological burden and estimate associations between individual-level variables and this "subclinical" presentation.METHODS: We performed a secondary analysis of TB prevalence survey data from the South African communities of the Zambia, South Africa Tuberculosis and AIDS Reduction trial. Generalized estimating equations were used to estimate the association between individual-level demographic, behavioral, socio-economic, and medical variables and the risk of bacteriologically positive TB among participants not reporting any symptoms consistent with active TB.RESULTS: The crude prevalence of TB was 2222.1 cases per 100 000 population (95% CI 2053.4-2388.5); 44.7% (295/660) of all documented prevalent cases of TB were subclinical. Current tobacco smoking (OR 2.37, 95% CI 1.41-3.99) and HIV-positive status (OR 3.26, 95% CI 2.31-4.61) were significantly associated with subclinical TB.CONCLUSION: Individuals who smoke or have HIV may be at increased risk of active TB and not report typical symptoms consistent with disease. This suggests possible shortcomings of symptom-based case finding which may need to be addressed in similar settings.

Secondary analysis of tuberculosis stigma data from a cluster randomised trial in Zambia and South Africa (ZAMSTAR).

SETTING: Zambian and South African TB and HIV Reduction (ZAMSTAR) cluster-randomised trial (CRT) communities, 2006-2009. OBJECTIVES: To develop TB stigma items, and evaluate changes in them in response to a household intervention aimed at reducing TB transmission and prevalence but not tailored to reduce stigma. DESIGN: TB stigma was measured at baseline and 18 months later among 1826 recently diagnosed TB patients and 1235 adult members of their households across 24 communities; 12 of 24 communities were randomised to receive the household intervention. We estimated the impact of the household intervention on TB stigma using standard CRT analytical methods. RESULTS: Among household members, prevalence of blame and belief in transmission myths fell in both study arms over time: adjusted prevalence ratios (aPRs) comparing the household intervention with the non-household intervention arm were respectively 0.61 (95%CI 0.26-1.44) and 0.77 (95%CI 0.48-1.25) at 18-month follow-up. Among TB patients, at baseline a low percentage experienced social exclusion and poor treatment by health staff and a relatively high percentage reported 'being made fun of', with little change over time. Disclosure of TB status increased over time in both study arms. Internalised stigma was less prevalent in the household arm at both baseline and follow-up, with an aPR of 0.85 (95%CI 0.41-1.76). Variability in stigma levels between countries and across communities was large. CONCLUSION: Robust TB stigma items were developed. TB stigma was not significantly reduced by the household intervention, although confidence intervals for estimated intervention effects were wide. We suggest that stigma-specific interventions are required to effectively address TB stigma.

Coverage of clinic-based TB screening in South Africa may be low in key risk groups.

The South African Ministry of Health has proposed screening all clinic attendees for tuberculosis (TB). Amongst other factors, male sex and bar attendance are associated with higher TB risk. We show that 45% of adults surveyed in Western Cape attended a clinic within 6 months, and therefore potentially a relatively high proportion of the population could be reached through clinic-based screening. However, fewer than 20% of all men aged 18-25 years, or men aged 26-45 who attend bars, attended a clinic. The population-level impact of clinic-based screening may be reduced by low coverage among key risk groups.

Le Ministère de la Santé d'Afrique du Sud a proposé de dépister la tuberculose (TB) chez tous les patients visitant un centre de santé. Parmi d'autres facteurs, le sexe masculin et la fréquentation des bars sont associés à un risque plus élevé de TB. Nous montrons que 45% des adultes dépistés dans la province du Cap Ouest s'étaient rendus dans un centre de santé au cours des 6 derniers mois et c'est pourquoi une proportion relativement élevée de la population pourrait être atteinte à travers un dépistage en centre de santé. Cependant, moins de 20% de tous les hommes âgés de 18­25 ans, ou des hommes âgés de 26­45 ans qui fréquentent les bars, se rendent dans un centre de santé. L'impact sur la population de ce type de dépistage pourrait donc être réduit par une faible couverture parmi les groupes à risque majeur.

El Ministerio de Salud de Suráfrica propuso una detección sistemática de la tuberculosis (TB) a todas las personas que acudían a los consultorios. Entre los factores asociados con un mayor riesgo de padecer TB están el sexo masculino y la frecuentación de bares. El presente artículo pone de manifiesto que 45% de los adultos encuestados en la Ciudad del Cabo había acudido a un establecimiento de salud en los últimos 6 meses, por lo cual se pudo llegar a una proporción relativamente alta de la población mediante esta detección sistemática. Sin embargo, menos del 20% de todos los hombres entre los 18 y los 25 años, o entre los 26 y los 45 años de edad que frecuenta los bares, acudió a los establecimientos de salud. La repercusión a escala de la población de una detección sistemática realizada en los consultorios podría verse atenuada por una baja cobertura de los grupos más vulnerables.

Health care workers' gender bias in testing could contribute to missed tuberculosis among women in South Africa.

SETTING: Eight communities with high tuberculosis (TB) prevalence, Western Cape, South Africa. OBJECTIVE: To identify sex differences in TB health-seeking behaviour and diagnosis in primary health care facilities and how this influences TB diagnosis. DESIGN: We used data from a prevalence survey among 30,017 adults conducted in 2010 as part of the Zambia, South Africa Tuberculosis and AIDS Reduction (ZAMSTAR) trial. RESULTS: A total of 1670 (5.4%) adults indicated they had a cough of ⩾2 weeks, 950 (56.9%) of whom were women. Women were less likely to report a cough of ⩾2 weeks (5.1% vs. 6.4%, P < 0.001), but were more likely to seek care for their cough (32.6% vs. 26.9%, P = 0.012). Of all adults who sought care, 403 (80.0%) sought care for their cough at a primary health care (PHC) facility (79.0% women vs. 81.4% men, P = 0.511). Women were less likely to be asked for a sputum sample at the PHC facility (63.3% vs. 77.2%, P = 0.003) and less likely to have a positive sputum result (12.6% vs. 20.7%, P = 0.023). CONCLUSION: The attainment of sex equity in the provision of TB health services requires adherence to testing protocols. Everyone, irrespective of sex, who seeks care for a cough of ⩾2 weeks should be tested.

Symptom reporting among prevalent tuberculosis cases who smoke, are HIV-positive or have hyperglycaemia.

Data from a tuberculosis (TB) prevalence survey conducted in 24 communities in Zambia and the Western Cape, South Africa, January-December 2010, were analysed to determine the influence of smoking, hyperglycaemia and human immunodeficiency virus (HIV) infection on TB symptom reporting in culture-confirmed TB cases. Of 123 790 adults eligible for enrolment, 90 601 (73%) consented and 64 463 had evaluable sputum samples. ORs and 95%CIs were calculated using a robust standard errors logistic regression model adjusting for clustering at community level. HIV-positive TB cases were more likely to report cough, weight loss, night sweats and chest pain than non-HIV-positive TB cases. TB cases who smoked or had hyperglycaemia did not report symptoms differently from cases without these comorbidities.

Des données émanant d';une enquête de prévalence de la tuberculose (TB), réalisées dans 24 communautés de Zambie et de la province du Cap Ouest en Afrique du Sud de janvier à décembre 2010, ont été analysées afin de déterminer l'influence du tabagisme, de l'hyperglycémie et de l'infection par le virus de l'immunodéficience humaine (VIH) sur l'expression des symptômes de TB dans des cas de TB confirmés par la culture. Sur 123 790 adultes éligibles, 90 601 (73%) ont consenti et 64 463 ont produit des échantillons de crachats utilisables. Les OR et les IC à 95% ont été calculés grâce à un modèle de régression logistique des erreurs standard robustes en ajustant le regroupement au niveau des communautés. Les patients atteints de TB et VIH positifs rapportaient plus volontiers de la toux, une perte de poids, des sueurs nocturnes et des douleurs thoraciques que les patients non VIH positifs. Par contre, les cas de TB qui fumaient ou présentaient une hyperglycémie ne signalaient pas de symptômes différents des patients qui n'avaient pas ces co-morbidités.

En el presente estudio se analizaron los datos de una encuesta reciente sobre la prevalencia de tuberculosis (TB), que se había llevado a cabo en 24 comunidades en Zambia y en el Cabo Occidental de Suráfrica de enero a diciembre del 2010, con el fin de determinar la influencia del tabaquismo, la hiperglucemia y la infección por el virus de la inmunodeficiencia humana (VIH) sobre los síntomas de TB que refieren los pacientes con diagnóstico de TB confirmada por cultivo. De los 123 790 adultos aptos para el estudio, 90 601 dieron su consentimiento (73%) y 64 463 aportaron muestras de esputo adecuadas. Se calculó el cociente de posibilidades y el intervalo de confianza del 95% mediante un modelo consistente de regresión logística y errores estándar, tras corregir la agregación de los datos por comunidad. Fue más frecuente que los pacientes con TB que eran positivos frente al VIH refirieran tos, pérdida de peso, sudoración nocturna y dolor torácico que los pacientes sin esta coinfección. No hubo diferencia en los síntomas comunicados por los pacientes tuberculosos que fumaban o que presentaban hiperglucemia, en comparación con los pacientes sin estas enfermedades concurrentes.

Unusual sex differences in tuberculosis notifications across Pakistan and the role of environmental factors.

In developing countries, only one-third of new tuberculosis cases notified are from women. It is not clear whether tuberculosis incidence is lower in women than men, or whether notification figures reflect under-detection of tuberculosis in women. Pakistan, however, presents an unusual pattern of sex differences in tuberculosis notifications. While 2 of the 4 provinces (Sindh and Punjab) report more notifications from men (female to male ratios 0.81 and 0.89 respectively in 2009), the other 2 provinces (Khyber-Pakhtunkhwa and Balochistan) consistently report higher numbers of smear-positive tuberculosis notifications from women than men (1.37 and 1.40). No other country is known to have such a large variation in the sex ratios of notifications across regions. Large variations in female to male smear-positive notification ratios in different settings across a single country may indicate that environmental factors, rather than endogenous biological factors, are important in influencing the observed sex differences in tuberculosis notifications.

Systematic screening for active tuberculosis: rationale, definitions and key considerations.

The impact of current interventions to improve early detection of tuberculosis (TB) seems to have been saturated. Case detection trends have stagnated. TB incidence is falling in most settings worldwide, but the rate of decline is far lower than expected. There is growing evidence from national TB prevalence surveys and other research of a large pool of undetected TB in the community. Intensified efforts to further break down access barriers and scale up new and rapid diagnostic tools are likely to improve the situation. However, will these be enough? Or do we also need to reach out more towards people who do not actively seek care with well-recognisable TB symptoms? There have recently been calls to revisit TB screening, particularly in high-risk groups. The World Health Organization (WHO) recommends screening for TB in people with human immunodeficiency virus infection and in close TB contacts. Should other risk groups also be screened systematically? Could mass, community-wide screening, which the WHO has discouraged over the past four decades, be of benefit in some situations? If so, what screening tools and approaches should be used? The WHO is in the process of seeking answers to these questions and developing guidelines on systematic screening for active TB. In this article, we present the rationale, definitions and key considerations underpinning this process.

The measurement of household socio-economic position in tuberculosis prevalence surveys: a sensitivity analysis.

OBJECTIVE: To assess the robustness of socio-economic inequalities in tuberculosis (TB) prevalence surveys. DESIGN: Data were drawn from the TB prevalence survey conducted in Lusaka Province, Zambia, in 2005-2006. We compared TB socio-economic inequalities measured through an asset-based index (Index 0) using principal component analysis (PCA) with those observed using three alternative indices: Index 1 and Index 2 accounted respectively for the biases resulting from the inclusion of urban assets and food-related variables in Index 0. Index 3 was built using regression-based analysis instead of PCA to account for the effect of using a different assets weighting strategy. RESULTS: Household socio-economic position (SEP) was significantly associated with prevalent TB, regardless of the index used; however, the magnitude of inequalities did vary across indices. A strong association was found for Index 2, suggesting that the exclusion of food-related variables did not reduce the extent of association between SEP and prevalent TB. The weakest association was found for Index 1, indicating that the exclusion of urban assets did not lead to higher extent of TB inequalities. CONCLUSION: TB socio-economic inequalities seem to be robust to the choice of SEP indicator. The epidemiological meaning of the different extent of TB inequalities is unclear. Further studies are needed to confirm our conclusions.

Factors influencing sex differences in numbers of tuberculosis suspects at diagnostic centres in Pakistan.

SETTING: DOTS-reporting tuberculosis (TB) diagnostic centres across Pakistan. OBJECTIVES: To quantitatively investigate the influence of diagnostic centre characteristics on the number of female and male TB suspects registered at diagnostic centres. DESIGN: Ten districts were selected across the four provinces of Pakistan. Data were collected on male and female TB suspects in all diagnostic centres within each district. A structured questionnaire was used to collect data on characteristics of the diagnostic centres. Multiple linear regression analysis was conducted to evaluate the influence of each characteristic on sex differences in the numbers of suspects. RESULTS: Two diagnostic centre characteristics were associated with higher numbers of female than male TB suspects: catering to the local catchment area (P = 0.001) and being accessible on foot (P = 0.002). The following characteristics were associated with higher numbers of male than female TB suspects: being open after 2 pm (P = 0.041), having more than five doctors working at the centre (P = 0.019), and having more than 100 suspects registered per quarter (P = 0.008). CONCLUSIONS: Smaller, local diagnostic centres that are accessible on foot registered more female than male TB suspects. More centralised facilities located further from homes, larger facilities and those with evening opening hours registered more male than female suspects.

Decentralised care for the management of child contacts of multidrug-resistant tuberculosis.

SETTING: Cape Town, South Africa. OBJECTIVE: To determine the number of multidrug-resistant tuberculosis (MDR-TB) child contacts routinely identified by health services, and whether a model of decentralised care improves access. METHODS: All MDR-TB source cases registered in Cape Town from April 2010 to March 2011 were included. All child contacts assessed at hospital and outreach clinics were recorded from May 2010 to June 2011. Electronic probabilistic matching was used to match source cases with potential child contacts; the number of children accessing decentralised (Khayelitsha) and hospital-based care was compared. RESULTS: Of 1221 MDR-TB source cases identified, 189 (15.5%) were registered in Khayelitsha; 31 (16.4%) had at least one child contact assessed. In contrast, 95 (9.2%) of the 1032 source cases diagnosed in the other Cape Town subdistricts (hospital-based care) had at least one child contact assessed (P = 0.003). Children in Khayelitsha were seen at a median of 71 days (interquartile range [IQR] 37-121 days) after source case diagnosis compared to 90 days (IQR 56-132 days) in other subdistricts (P = 0.15). CONCLUSION: Although decentralised care led to an increased number of child contacts being evaluated, both models led to the assessment of a small number of potential child MDR-TB contacts, with considerable delay in assessment.

Why do child contacts of multidrug-resistant tuberculosis not come to the assessment clinic?

BACKGROUND: Local policy advises that children exposed to multidrug-resistant tuberculosis (MDR-TB) should be assessed in a specialist clinic. Many children, however, are not brought for assessment. METHODS: Focus group discussion was used to design appropriate questionnaires. From 1 September 2011, the first 50 children referred to the specialist paediatric MDR-TB clinic, Cape Town, South Africa, and who attended their clinic appointment, were recruited. The first 50 children who were referred but who did not attend were concurrently identified, traced and recruited. Differences in group characteristics were compared. RESULTS: The median age of the children was 35 months: 48 (48%) were boys, 4 (4%) were human immunodeficiency virus infected and 47 (47%) were of coloured ethnicity. Factors significantly associated with non-attendance at the MDR-TB clinic were: Coloured ethnicity (OR 2.82, 95%CI 1.21-6.59, P = 0.01), the mother being the source case (OR 3.78, 95%CI 1.29-11.1, P = 0.02), having a smoker resident in the house (OR 2.37, 95%CI 1.01-5.57, P = 0.04), the time (P = 0.002) and cost (P = 0.03) required to get to the specialist clinic, and fear of infection whilst waiting to be seen (OR 2.45, 95%CI 1.07-5.60, P = 0.03). CONCLUSIONS: Reasons for non-attendance at paediatric MDR-TB clinic appointments are complex and are influenced by demographic, social, logistical and cultural factors.

Delays in seeking treatment for symptomatic tuberculosis in Sabah, East Malaysia: factors for patient delay.

SETTING: The state of Sabah contributes one third of the tuberculosis (TB) cases in Malaysia. OBJECTIVE: To collect information on factors that affect the time period from the onset of symptoms to first contact with health care providers, whether private or government. DESIGN: A cross-sectional study using a pre-tested questionnaire was conducted among 296 newly registered smear-positive TB patients in 10 districts in Sabah. Univariable and multivariable analyses were used to determine which risk factors were associated with patient delay (>30 days) and 'extreme' patient delay (>90 days). RESULTS: The percentage of patients who sought treatment after 30 and 90 days was respectively 51.8% (95%CI 45.7-57.9) and 23.5% (95%CI 18.6-29.0). The strongest factors associated with patient delay and 'extreme' patient delay was when the first choice for treatment was a non-government health facility and in 30-39-year-olds. 'Extreme' patient delay was also weakly associated, among other factors, with comorbidity and livestock ownership. CONCLUSION: Delay and extreme delay in seeking treatment were more common when the usual first treatment choice was a non-government health facility. Continuous health education on TB aimed at raising awareness and correcting misconceptions is needed, particularly among those who use non-government facilities.

Human immunodeficiency virus associated tuberculosis more often due to recent infection than reactivation of latent infection.

BACKGROUND: It is unclear whether human immunodeficiency virus (HIV) increases the risk of tuberculosis (TB) mainly through reactivation or following recent Mycobacterium tuberculosis (re)infection. Within a DNA fingerprint-defined cluster of TB cases, reactivation cases are assumed to be the source of infection for subsequent secondary cases. As HIV-positive TB cases are less likely to be source cases, equal or higher clustering in HIV-positives would suggest that HIV mainly increases the risk of TB following recent infection. METHODS: A systematic review was conducted to identify all studies on TB clustering and HIV infection in HIV-endemic populations. Available individual patient data from eligible studies were pooled to analyse the association between clustering and HIV. RESULTS: Of seven eligible studies, six contributed individual patient data on 2116 patients. Clustering was as, or more, likely in the HIV-positive population, both overall (summary OR 1.26, 95%CI 1.0-1.5), and within age groups (OR 1.50, 95%CI 0.9-2.3; OR 1.00, 95%CI 0.8-1.3 and OR 2.57, 95%CI 1.4-5.7) for ages 15-25, 26-50 and >50 years, respectively. CONCLUSIONS: Our results suggest that HIV infection mainly increases the risk of TB following recent M. tuberculosis transmission, and that TB control measures in HIV-endemic settings should therefore focus on controlling M. tuberculosis transmission rather than treating individuals with latent M. tuberculosis infection.

Potential utility of empirical tuberculosis treatment for HIV-infected patients with advanced immunodeficiency in high TB-HIV burden settings.

The human immunodeficiency virus (HIV) and HIV-associated tuberculosis (TB-HIV) epidemics remain uncontrolled in many resource-limited regions, especially in sub-Saharan Africa. The scale of these epidemics requires the consideration of innovative bold interventions and 'out-of-the-box' thinking. To this end, a symposium entitled 'Controversies in HIV' was held at the 40th Union World Conference on Lung Health in Cancun, Mexico, in December 2009. The first topic debated, entitled 'Annual HIV testing and immediate start of antiretroviral therapy for all HIV-infected persons', received much attention at international conferences and in the literature in 2009. The second topic forms the subject of this article. The rationale for the use of empirical TB treatment is premised on the hypothesis that in settings worst affected by the TB-HIV epidemic, a subset of HIV-infected patients have such a high risk of undiagnosed TB and of associated mortality that their prognosis may be improved by immediate initiation of empirical TB treatment used in conjunction with antiretroviral therapy. In addition to morbidity and mortality reduction, additional benefits may include prevention of nosocomial TB transmission and TB preventive effect. Potential adverse consequences, however, may include failure to consider other non-TB diagnoses, drug co-toxicity, compromised treatment adherence, and logistical and resource challenges. There may also be general reluctance among national TB programmes to endorse such a strategy. Following fruitful debate, the conclusion that this strategy should be carefully evaluated in randomised controlled trials was strongly supported. This paper provides an in-depth consideration of this proposed intervention.

Tuberculosis among community-based health care researchers.

BACKGROUND: Occupational tuberculosis (TB) in hospital-based health care workers is reported regularly, but TB in community-based health care researchers has not often been addressed. OBJECTIVE: To investigate TB incidence in health care researchers in a high TB and human immunodeficiency virus prevalent setting in the Western Cape, South Africa. The health care researchers were employed at the Desmond Tutu TB Centre, Stellenbosch University. METHODS: A retrospective analysis was performed of routine information concerning employees at the Desmond Tutu TB Centre. The Centre has office-based and community-based employees. RESULTS: Of 180 researchers included in the analysis, 11 TB cases were identified over 250.4 person-years (py) of follow-up. All cases were identified among community-based researchers. TB incidence was 4.39 per 100 py (95%CI 2.45-7.93). The standardised TB morbidity ratio was 2.47 (95%CI 1.25-4.32), which exceeded the standard population rate by 147%. CONCLUSIONS: TB incidence in South Africa was 948 per 100,000 population per year in 2007; in the communities where the researchers worked, it was 1875/100,000. Community-based researchers in the study population have a 2.34 times higher TB incidence than the community. It is the responsibility of principal investigators to implement occupational health and infection control guidelines to protect researchers.

Children's role in enhanced case finding in Zambia.

OBJECTIVE: To evaluate information dissemination by children and attitudes among children towards a school-based tuberculosis (TB) reduction strategy that asked children to address TB symptoms, testing and stigma in their homes. SETTING AND DESIGN: Qualitative research was conducted with schoolchildren before, and 2 years into, an intervention to promote early detection of TB using sputum microscopy in Zambia. The baseline study in 2005 involved 38 children at five sites. The evaluation in 2008 included 209 children in schools at four sites. Research with schoolchildren included discussions, drawings, role plays and narratives. RESULTS: The baseline study revealed children's enthusiasm to learn about TB and the human immunodeficiency virus (HIV), but it also revealed children's anxieties about the possible conflicts related to discussing HIV and TB with adults. Children in the evaluation demonstrated more accurate knowledge about TB and HIV than in the baseline study. Children were enthusiastic about discussing TB and HIV at home. Their responses suggested that they did so with respect and adult approval, circumventing the intergenerational conflict expected during the baseline study. CONCLUSION: The present study demonstrates that schoolchildren have a role to play in enhanced case finding. Schoolchildren are already familiar with TB in areas of high burden, but they need more information about the link between TB and HIV and about antiretroviral treatment.

Intra-assay reliability and robustness of QuantiFERON(R)-TB Gold In-Tube test in Zambia.

BACKGROUND: Interferon-gamma (IFN-gamma) release assays (IGRAs), such as the QuantiFERON-TB Gold In-Tube test (QFT-GIT), are becoming a preferred method for diagnosis of tuberculosis (TB) infection in many industrialised countries. However, data on the effectiveness of IGRAs in high TB-HIV (human immunodeficiency virus) endemic and resource-limited settings, such as Zambia, are limited. OBJECTIVE: To determine the intra-assay reliability and robustness of QFT-GIT in a field setting in Zambia. DESIGN: During July-October 2007, 109 adult smear-positive TB patients were recruited to determine QFT-GIT reliability and the effect of a 24-h delay in incubation. Two simulated laboratory experiments were also performed using 9-14 volunteers, to explore the effect of power outages during incubation and storage temperature of collection tubes on IFN-gamma responses. RESULTS: QFT-GIT intra-assay concordance was 91.7% (kappa = 0.8). Discordance was observed for nine patients, of whom six were HIV-positive. There was evidence of an association between HIV status and discordant results (OR 1.98, 95%CI 1.06-3.67, P = 0.03). A 24-h delay in incubation changed results for 25 of the 109 (22.9%) patients. Power outages that altered incubation time reduced IFN-gamma responses. CONCLUSION: Although QFT-GIT seems reliable in this setting, we have identified operational factors that affect its robustness. These factors may influence the effectiveness of this test in similar resource-limited settings.

Extra-pulmonary and smear-negative forms of tuberculosis are associated with treatment delay and hospitalisation.

SETTING: Adult patients with tuberculosis (TB) recruited at the chest clinic of the University Teaching Hospital in Lusaka, Zambia, from 2003 to 2004. OBJECTIVE: To identify factors associated with delayed treatment or hospitalisation. DESIGN: A cross-sectional survey of newly identified adult patients with TB. RESULTS: A total of 223 patients were included in the analysis. Patients with smear-negative disease were 2.6 times more likely to be hospitalised than those with smear-positive disease (95%CI 1.28-5.30), while patients with extra-pulmonary disease were 3.42 times more likely to be hospitalised than those with pulmonary disease (95%CI 1.75-6.66). Patients with smear-negative disease were 2.81 times more likely to have experienced overall delay than those with smear-positive disease (95%CI 1.20-6.66). DISCUSSION: This analysis has demonstrated that patients with extra-pulmonary or smear-negative disease are significantly more likely to be hospitalised. Patients with smear-negative disease are also more likely to have experienced treatment delay. These data reinforce the urgent need for more robust diagnostic tests, particularly for smear-negative and extra-pulmonary disease. As these forms of disease are more likely to be associated with the human immunodeficiency virus (HIV), the data support earlier diagnosis and treatment of HIV infection.