Country differences and determinants of yield in programmatic migrant TB screening in four European countries.

INTRODUCTION The WHO End TB Strategy emphasises early diagnosis and screening of TB in high-risk groups, including migrants. We analysed TB yield data from four large migrant TB screening programmes to inform TB policy.METHODS We pooled routinely collected individual TB screening episode data from Italy, the Netherlands, Sweden and the United Kingdom under the European Union Commission E-DETECT.TB grant, described characteristics of the screened population, and analysed TB case yield.RESULTS We collected data on 2,302,260 screening episodes among 2,107,016 migrants, mostly young adults aged 18-44 years (77.8%) from Asia (78%) and Africa (18%). There were 1,658 TB cases detected through screening, with substantial yield variation (per 100,000): 201.1 for Sweden (95% confidence intervals CI 111.4-362.7), 68.9 (95% CI 65.4-72.7) for the United Kingdom, 83.2 (95% CI 73.3-94.4) for the Netherlands and 653.6 (95% CI 445.4-958.2) in Italy. Most TB cases were notified among migrants from Asia (n = 1,206, 75/100,000) or Africa (n = 370, 76.4/100,000), and among asylum seekers (n = 174, 131.5/100,000), migrants to the Netherlands (n = 101, 61.9/100,000) and settlement visa migrants to the United Kingdom (n = 590, 120.3/100,000).CONCLUSIONS We found considerable variations in yield across programmes, types of migrants and country of origin. These variations may be partly explained by differences in migration patterns and programmatic characteristics.

Cost-effectiveness of the Xpert® MTB/RIF assay for tuberculosis diagnosis in Brazil.

INTRODUCTION: The Xpert® MTB/RIF assay is being implemented as a substitute for sputum smear microscopy (SSM) in many low and high tuberculosis (TB) burden countries, including Brazil, a country with low multidrug resistance and moderate human immunodeficiency virus co-infection rates. SETTING: Brazilian National TB Programme (NTP). OBJECTIVE AND DESIGN: We estimated the incremental cost-effectiveness ratio (ICER) of Xpert as a substitute for two SSM tests in the diagnosis of drug-susceptible TB. The costs for confirming each additional case and for avoiding treatment due to false-positive empirical diagnoses were estimated. RESULTS: The ICER was US$943 for each additional TB diagnosis and US$356 for each additional TB diagnosis with bacteriological confirmation, assuming 80% specificity of clinical diagnosis using both strategies. For every 100 000 patients with suspected TB, the NTP would spend an additional US$1.2 million per year to confirm 3344 more TB patients. The model was highly sensitive to specificity of clinical diagnosis after a negative test. CONCLUSION: Although the NTP has no threshold for cost-effectiveness, our model can provide support for decision makers in Brazil and other countries with a low prevalence of drug resistance among TB patients. Financial benefit can potentially be expected if physicians rely more on a negative Xpert result and empirical treatment is reduced.

Surgery as an Adjunctive Treatment for Multidrug-Resistant Tuberculosis: An Individual Patient Data Metaanalysis.

BACKGROUND: Medical treatment for multidrug-resistant (MDR)-tuberculosis is complex, toxic, and associated with poor outcomes. Surgical lung resection may be used as an adjunct to medical therapy, with the intent of reducing bacterial burden and improving cure rates. We conducted an individual patient data metaanalysis to evaluate the effectiveness of surgery as adjunctive therapy for MDR-tuberculosis. METHODS: Individual patient data, was obtained from the authors of 26 cohort studies, identified from 3 systematic reviews of MDR-tuberculosis treatment. Data included the clinical characteristics and medical and surgical therapy of each patient. Primary analyses compared treatment success (cure and completion) to a combined outcome of failure, relapse, or death. The effects of all forms of resection surgery, pneumonectomy, and partial lung resection were evaluated. RESULTS: A total of 4238 patients from 18 surgical studies and 2193 patients from 8 nonsurgical studies were included. Pulmonary resection surgery was performed on 478 patients. Partial lung resection surgery was associated with improved treatment success (adjusted odds ratio [aOR], 3.0; 95% confidence interval [CI], 1.5-5.9; I(2)R, 11.8%), but pneumonectomy was not (aOR, 1.1; 95% CI, .6-2.3; I(2)R, 13.2%). Treatment success was more likely when surgery was performed after culture conversion than before conversion (aOR, 2.6; 95% CI, 0.9-7.1; I(2)R, 0.2%). CONCLUSIONS: Partial lung resection, but not pneumonectomy, was associated with improved treatment success among patients with MDR-tuberculosis. Although improved outcomes may reflect patient selection, partial lung resection surgery after culture conversion may improve treatment outcomes in patients who receive optimal medical therapy.

Suspected tuberculosis case detection and referral in private pharmacies in Viet Nam.

SETTINGS: Private pharmacies in Hanoi, Viet Nam. OBJECTIVES: To explore the response of health care providers (HCPs) in private pharmacies to suspected tuberculosis (TB) patients. METHODS: A simulated patient method combined with an interview in 128 randomly selected private pharmacies and 10 private pharmacies near TB hospitals. RESULTS: In the simulated patient method and interview, respectively 59 (46%) and 70 (55%) of HCPs referred the TB suspect to general health care. Only 11 (9%) referred the simulated patient to a TB care facility. Fifty-two (42%) of the HCPs identified suspected TB from a fictitious case described on paper; 34 (27%) were aware that free treatment was provided under the National Tuberculosis Programme (NTP). Knowledge about free NTP treatment predicted a higher rate of direct referrals to TB facilities (OR 5.80, 95%CI 1.88-19.62) and greater ability to identify suspected TB from a fictitious case on paper (OR 5.14, 95%CI 2.36-11.73). Pharmacies with Good Pharmacy Practice (GPP) certification were less likely to refer simulated patients to TB facilities than non-GPP pharmacies (OR 0.10, 95%CI ≤0.01-0.79). CONCLUSIONS: Nearly half of HCPs in private pharmacies do not refer TB suspects, possibly contributing to delays in diagnosis and treatment. Knowledge about free NTP treatment predicted better performance of HCPs.

Yield of interview screening and chest X-ray abnormalities in a tuberculosis prevalence survey.

BACKGROUND: Tuberculosis (TB) prevalence surveys generally rely on a combination of screening methods to identify suspects eligible for sputum culture. OBJECTIVE: To assess the yield of screening methods applied in a recent prevalence survey in Viet Nam and estimate the proportion of TB cases missed due to incomplete participation. METHODS: TB suspects were identified based on self-reported TB history or productive cough by interview and chest X-ray (CXR). We calculated the case yield of these two screening methods by dividing the number of cases detected per method by the total number of cases detected. As not all participants underwent the full screening procedure, we recalculated the maximum yield of the screening methods using multiple imputation methods. RESULTS: The yield from screening by interview and CXR were respectively 38% and 91%. Adjusting for missing data by multiple imputation, we estimated that we missed 9.9% (95%CI 6.8-14.2) of expected TB cases. CONCLUSION: In prevalence surveys, screening by pre-structured interview is insufficient, and should be supplemented with CXR to achieve sufficient identification of TB cases. The effect of incomplete participation in the full screening procedure may be substantial and should be adjusted for in the analysis.

Operational research for improved tuberculosis control: the scope, the needs and the way forward.

The promotion of research is one of the main components of the World Health Organizations Stop TB Strategy, which includes 'programme-based operational research (OR)' and 'research on introducing new tools into practice'. The importance of OR in improving tuberculosis (TB) control was recognised a long time ago, and historical OR studies have been instrumental in the development of major strategies for TB control. Although a growing number of OR projects are being conducted in the world today, little is known about their results or their likely impact on TB control programmes. As funding organisations increasingly recognise the need for OR, we propose a rational framework to conduct OR, which covers a spectrum from local setting-oriented to international policy guiding research, and determines the relevance, replicability and generalisability of the results. OR in TB control is aimed at 1) improving programme performance; 2) assessing the feasibility, effectiveness and impact of new strategies or interventions on TB control; and 3) collecting evidence to guide policy recommendations on specific interventions. This requires strengthened capacity to plan and conduct OR in low-income countries and appropriate support to conduct both nationally and internationally led OR projects. Suggestions are made for potential steps for improved purpose-driven OR, which may help to improve TB control locally and inform policy recommendations internationally.

Measurement of exhaled nitric oxide as a potential screening tool for pulmonary tuberculosis.

BACKGROUND: There is a need for low-technology, inexpensive screening tools for active tuberculosis (TB) case finding. OBJECTIVE: to assess the potential usefulness of measuring exhaled nitric oxide (eNO). DESIGN: Cross-sectional comparison in Hanoi, Viet Nam, comparing 90 consecutive smear-positive, culture-confirmed TB patients presenting at a referral hospital with office workers (no X-ray confirming TB) at this hospital (n = 52) and at a construction firm (n = 84). eNO levels were analysed using a validated handheld analyser. RESULTS: eNO levels among TB patients (median 15 parts per billion [ppb], interquartile range [IQR] 10-20) were equal to those among construction firm workers (15 ppb, IQR 12-19, P = 0.517) but higher than those among hospital workers (8.5 ppb, IQR 5-12.5, P < 0.001). Taking the hospital workers as the comparison group, best performance as a diagnostic tool was at a cut-off of 10 ppb, with sensitivity 78% (95%CI 68-86) and specificity 62% (95%CI 47-75). Test characteristics could be optimised to 84% vs. 67% by excluding individuals who had recently smoked or consumed alcohol. CONCLUSION: While eNO measurement has limited value in the direct diagnosis of pulmonary TB, it may be worth developing and evaluating as a cost-effective replacement of chest X-ray in screening algorithms of pulmonary TB where X-ray is not available.

Interferon-gamma release assays in immigrant contacts and effect of remote exposure to Mycobacterium tuberculosis.

OBJECTIVE: To assess the association between remote exposure to tuberculosis (TB) and results of the tuberculin skin test (TST), and two interferon-gamma release assays (IGRAs)-QuantiFERON-TB Gold In-Tube (QFT-GIT) and T-SPOT.TB-in immigrant contacts of sputum smear-positive TB patients. METHODS: Immigrants aged >or=16 years in close contact with smear-positive TB patients were included. QFT-GIT and T-SPOT.TB were performed if the TST induration size was >or=5 mm. Associations between test results and origin from an endemic country were assessed. RESULTS: Of 433 close contacts, 322 (74%) had TST >or=5 mm, of whom, 282 (88%) had valid test results for all assays. Positive QFT-GIT results were obtained for 152/282 (54%) and positive T-SPOT.TB for 168/282 (60%). After adjustment for age, sex and recent contact, positive IGRA results and TST results >/=10 mm were found to be more frequent among immigrants who originated from Africa, in particular sub-Saharan Africa. CONCLUSION: When IGRAs are used to determine latent TB infection in foreign-born individuals, positive findings not only relate to recent TB infection, but also reflect prior TB exposure in the country of origin. This late reactivity will limit their usefulness in contact investigations among immigrants originating from endemic areas.

LTBI: latent tuberculosis infection or lasting immune responses to M. tuberculosis? A TBNET consensus statement.

Tuberculosis control relies on the identification and preventive treatment of individuals who are latently infected with Mycobacterium tuberculosis. However, direct identification of latent tuberculosis infection is not possible. The diagnostic tests used to identify individuals latently infected with M. tuberculosis, the in vivo tuberculin skin test and the ex vivo interferon-gamma release assays (IGRAs), are designed to identify an adaptive immune response against, but not necessarily a latent infection with, M. tuberculosis. The proportion of individuals who truly remain infected with M. tuberculosis after tuberculin skin test or IGRA conversion is unknown. It is also uncertain how long adaptive immune responses towards mycobacterial antigens persist in the absence of live mycobacteria. Clinical management and public healthcare policies for preventive chemotherapy against tuberculosis could be improved, if we were to gain a better understanding on M. tuberculosis latency and reactivation. This statement by the TBNET summarises knowledge and limitations of the currently available tests used in adults and children for the diagnosis of latent tuberculosis infection. In summary, the main issue regarding testing is to restrict it to those who are known to be at higher risk of developing tuberculosis and who are willing to accept preventive chemotherapy.

Inclusion of information on risk factors, socio-economic status and health seeking in a tuberculosis prevalence survey.

Data on socio-economic status, exposure to risk factors for tuberculosis (TB) and previous health-seeking for TB may be included in a TB prevalence survey to gain better knowledge about the distribution of TB in the population as well as a better understanding of what factors are driving the TB epidemic in a given setting. This article provides an overview of how such additional information may be collected. The article highlights the need to carefully consider the risk of jeopardising the quality of the overall survey by overburdening it with additional data collection, and concludes that additional time and resources for planning, training, logistics and supervision are required to safeguard quality. The article also discusses special considerations regarding sampling, sample size and data interpretation when including such information in a TB prevalence survey.

Tuberculosis prevalence surveys: rationale and cost.

This article is the first of the educational series 'Assessing tuberculosis (TB) prevalence through population-based surveys'. The series will give overall guidance in conducting cross-sectional surveys of pulmonary TB (PTB) disease. TB prevalence surveys are most valuable in areas where notification data obtained through routine surveillance are of unproven accuracy or incomplete, and in areas with an estimated prevalence of bacteriologically confirmed TB of more than 100 per 100,000 population. To embark on a TB prevalence survey requires commitment from the national TB programme, compliance in the study population, plus availability of trained staff and financial resources. The primary objective of TB prevalence surveys is to determine the prevalence of PTB in the general population aged >or=15 years. Limitations of TB prevalence surveys are their inability to assess regional or geographic differences in prevalence of TB, estimate the burden of childhood TB or estimate the prevalence of extra-pulmonary TB. The cost of a prevalence survey is typically US$ 4-15 per person surveyed, and up to US$ 25 per person with radiographic screening. A survey of 50,000 people, of limited precision, would typically cost US$ 200,000-1,250,000.

HIV-related mortality among tuberculosis patients in The Netherlands, 1993-2001.

Human immunodeficiency virus (HIV) infection in tuberculosis (TB) patients in The Netherlands during the period 1993-2001 was associated with an increased risk of death (adjusted odds ratio 4.71, P < 0.002). Age and sex-standardised mortality rates among HIV-infected TB patients decreased significantly over time, from 22.9% in 1993-1995 to 11.8% in 1999-2001 (P < 0.001). No such change was observed for HIV-negative patients. The decrease in mortality is likely due to the more widespread use of highly active antiretroviral therapy (HAART).

[HIV-test policy for tuberculosis patients in the Netherlands]. / Hiv-testbeleid bij tuberculosepatiënten in Nederland.

OBJECTIVE: To compare the proportion of tuberculosis patients tested for HIV infection, before and after introduction of highly active antiretroviral therapy (HAART) in the Netherlands, and to analyse predictive factors for performing an HTV-test in this population. DESIGN: Retrospective. METHOD: Whether patients had been tested for HIV, was investigated in random samples consisting of 200 patients, who were registered in the Netherlands Tuberculosis Register (NTR) in the years 1995 and 2001 respectively. RESULTS: The number of patients tested for HIV was 29 out of 84 (16%) in 1995, and 39 out of 190 (21%) in 2001 (not significant). HIV-tests had been carried out most frequently among homeless patients (71%), drug addicts (56%) and alcohol-abusing patients (60%). Significant predictive factors for HIV testing were place of residence (city), localisation of disease (pulmonary tuberculosis in combination with extrapulmonary tuberculosis) and place of origin (sub-Saharan Africa). CONCLUSION: Despite introduction of HAART during this period, in the Netherlands the proportion of tuberculosis patients tested for HIV did not significantly increase between 1995 and 2001. HIV testing was mainly limited to tuberculosis patients from risk groups.

[Limited tuberculin screening participation amongst travellers to countries with high tuberculosis incidence; reason to consider BCG vaccination for some travellers]. / Beperkte deelname aan tuberculinescreening onder reizigers naar landen met hoge tuberculose-incidentie; reden om voor sommigen BCG-vaccinatie te overwegen.

OBJECTIVE: To determine the compliance amongst Dutch travellers to high tuberculosis-incidence countries with a screening procedure involving a tuberculin skin test before and after the trip. DESIGN: Prospective study. METHOD: Nine hundred and eighty-eight tuberculin-negative Dutch people who travelled to high tuberculosis-incidence countries for 3 to 12 months were studied for their compliance with an advised screening procedure of repeat tuberculin skin testing 2 to 4 months after return. At 2 of the 4 participating health services, data were also collected on extra calls made and the pertinent time investments. RESULTS: Five hundred and ninety-nine travellers (61%) were compliant with the screening procedure. Of those for whom the data was available (n = 417), 33% (98/300) of the compliant travellers required extra calls. These took an average of 30 min per extra traveller tested as a result. Compliance varied according to health service and was better amongst travellers to Africa. In addition, non-compliance was independently associated with male sex, work being the main travel purpose, and an undecided duration of travel on departure. CONCLUSIONS: Compliance of Dutch travellers with tuberculin skin-test screening is limited, particularly if no extra calls are issued. Bacillus Calmette-Guérin vaccination appears to be preferable for travellers with undecided travel duration and persons travelling for work on a frequent basis.

Incidence and risk factors of diarrhoea in Dutch travellers: consequences for priorities in pre-travel health advice.

A cohort of 743 Dutch short-term travellers (1-6 weeks) to various (sub)tropical areas was studied to assess incidences of travellers' diarrhoea (TD) and risk factors to guide prevention policies. The occurrence of TD was ascertained retrospectively by questionnaire; independent risk factors were identified by logistic regression analysis. The overall attack rate (AR, 95% CI) of TD was 52% (49-56); 11% (9-14) reported two or more episodes. The overall incidence rate (IR) per 100 person weeks of travel (pwt) (95% CI) was 22 (20-24). IRs were highest for travellers to the Middle East (48, 33-71), lowest for South-east Asia (17, 15-20) and East Africa (18, 14-24) and intermediate for South America and West Africa (both 26, 19-36), Central America (29, 23-37) and the Indian subcontinent (32, 26-39). Compared to first episodes of TD, subsequent episodes were of longer duration and more frequently accompanied by faecal blood loss, abdominal cramps or systemic symptoms. After adjustment for travel duration and destination, independent risk factors (OR, 95% CI) for TD were recent treatment for gastrointestinal (GI) disorders (4.6, 1.2-17.2), history of GI surgery (3.9, 1.4-11.1) and, possibly, current use of medication reducing gastric acidity (6.9, 0.7-67.4). The risk was reduced for extensive travel experience (0.4, 0.3-0.7) and organized travel (0.7, 0.5-0.9). Regarding prevention and/or antibiotic self-treatment of TD, priority should be given to travellers who may suffer major health or other consequences from TD and to those with pre-existing GI disorders, particularly when visiting a high or intermediate-risk area on individual journeys with limited travel experience.