Article Text


Tuberculosis screening: variations on a theme

Statistics from


1M. P. R. Berry, 1C. M. Graham, 1S. J. Bloch, 1T. Oni, 1K. A. Wilkinson, 3R. J. Wilkinson, 2O. M. Kon, 4J. Banchereau, 4D. Chaussabel, 1A. O’Garra. 1National Institute For Medical Research, London, UK, 2Imperial College Healthcare NHS Trust, London, UK, 3University of Cape Town, Cape Town, Republic of South Africa, 4Baylor Institute for Immunology Research, Dallas, Texas, USA

Introduction Tuberculosis (TB) is a major cause of morbidity and mortality worldwide. The immune response during TB is complex and incompletely characterised, hindering the development of new diagnostics, treatments and vaccines. Studies performed in different disease settings (intermediate vs high burden) have sometimes yielded divergent results, limiting advances in our understanding of TB. We used systems biology approaches to obtain an unbiased comprehensive survey of the host response to TB in both the UK and South Africa.

Methods We recruited 214 participants; 153 met the final inclusion criteria. Active TB patients were culture positive. Latent TB patients were positive by both tuberculin skin test (TST) and interferon γ release assay (IGRA), and healthy controls were negative by both TST and IGRA. Whole blood was collected before treatment. RNA was extracted and used for whole genome expression studies using Illumina HT-12 microarrays. This was complemented by multiplexed cytokine analysis using the MILLIPLEX Multi-Analyte Profiling system. Biological data were integrated with comprehensive clinical data including radiology. Data mining was performed using Genespring GX 7.3 and Ingenuity Pathways Analysis software in combination with a novel Genomic Modular Analysis Framework. A subset of patients was assessed at 2 and 12 months post-treatment.

Results 42 participants from London were used as a training set to explore the microarray data, using supervised and unsupervised analyses to define a transcriptional signature. Findings were then tested in an independent set of 53 participants from London, and further validated using an independent cohort of 56 South African patients. We thus identify a robust blood transcriptional signature for active TB in both intermediate and high burden settings. Analysis of blood leucocyte counts and serum cytokines, along with interrogation of gene expression data using Pathway and Modular analysis, suggests that this signature reflects changes in cellular composition and altered cytokine expression. Transcriptional profiles appeared to reflect disease status. Longitudinal analysis revealed that the signature of active TB disappears during successful treatment.

Conclusions This is the first whole-genome expression profiling study in human TB. Our findings have implications for understanding disease pathogenesis, and could yield biomarkers for diagnosis and treatment monitoring.


P. Haldar, H. Thuraisingham, W. Hoskyns, G. Woltmann. University Hospitals of Leicester, Leicester, UK

Introduction A policy of a systematic single-step T cell-based interferon γ release assay (TIGRA) with QuantiFERON-TB Gold (QTG) has been employed at Leicester to screen for latent tuberculosis infection (LTBI) in close TB contacts since January 2007. Twelve week chemoprophylaxis is offered to all healthy QTG +ve contacts <35 years old. Baseline and longitudinal data have been collected prospectively for all index cases and their associated contacts since the introduction of the policy. Here we present the results of a cohort study examining outcome with QTG testing in all contacts since 2007.

Methods Data from all recorded contacts of patients with active TB diagnosed after 1 January 2007 are included. Contacts that developed active TB during the follow-up period were labelled “converters” if no evidence of active disease was identified at screening. Conversion risk was estimated using Kaplan–Meir plots and compared between defined subgroups using the log-rank test.

Results 2204 contacts (mean age 31.6 years, 51% male) were recorded in 446 notified cases (mean 4.9 contacts per case). 1039 contacts (47.1%) underwent screening with QTG and 204 were QTG +ve (19.6%). The mean follow-up period in QTG screened and unscreened contacts was 435 (±212) days and 455 (±235) days, respectively (p = 0.88). Chemoprophylaxis was started in 173 contacts (54 QTG untested). No QTG –ve contacts received chemoprophylaxis. Twelve week adherence was achieved in 90 contacts (52%). There were 39 converters (20 in the QTG-screened group). Of these, two had received chemoprophylaxis but only one had completed 12 weeks. All converters in the QTG tested group were QTG +ve. Two-year conversion risk (95% CI) in contacts not receiving chemoprophylaxis was 1.6% (0.8% to 2.4%) in the QTG untested population and 17.2% (9.4–25%) in QTG +ve contacts (hazard ratio (HR) = 10.3 if QTG +ve). Chemoprophylaxis lowered 2-year conversion risk in QTG +ve contacts to 4.7% (2.9% to 6.4%; HR 9 if no chemoprophylaxis).

Conclusion We estimate that 1 in 7.8 contacts testing QTG +ve will develop active TB without prophylaxis. QTG testing is an effective screening tool for identifying contacts at high risk of conversion.


L. P. Ormerod. Royal Blackburn Hospital, Blackburn, UK

Background Our Local Government area has predominantly (but not exclusively) immigration from South Asia (India, Pakistan and Banglasdesh). Prior to 2006 these were screened for latent tuberculosis infection as per then Joint Tuberculosis Committee 2000 Guidance. Following the National Institute for Health and Clinical Excellence (NICE) Guidelines 2006, new entrant screening for those aged 16–35 from countries with an incidence of under 500/100 000 per annum was limited to a chest x ray only. Data presented in 2007,1 showed that 10.5% of such individuals without treatment of LTBI developed clinical tuberculosis within 10 years. With these data we obtained funding from the Primary Care Trust (PCT), under Health Improvement initiative money, to screen those aged 16–34 age with interferon γ release assay (IGRA) testing, for a 2 year period from February 2009. Our initial experience is reported.

Methods From 1 February to date (20 July 2009) new entrant TB screening for those aged 16–34 years is by IGRA test alone (Quantiferon-in-tube Gold (Cellestis: QFT)). No prior Mantoux test was used (to save staffing costs on testing and reading), and a chest x ray was only performed on those with a positive QFT to exclude active disease, and if there were significant respiratory symptoms. Those with a positive QFT were seen, examined and started on treatment of LTBI with 3 months rifampicin and isoniazid (3RH).

Results Of new entrants (0.90%) from South Asia, 39 (28.9%) were positive, 94 were negative and 2 indeterminate. Of 13 non-African asylum seekers, 6/13 (46.1%) had a positive QFT. Aggregated results are given in the table. No patient with a positive QFT has had clinical disease on x ray and examination. No patient with some symptoms but a negative QFT has had disease (n = 5).

Conclusion The initial 6 months of this 2 year pilot show a 31.1% overall rate of LTBI in our predominantly South Asian 16- to 35-year-old new entrants, who would otherwise have been missed and hence not offered LTBI treatment. This is important as in our locality >90% of cases of TB are in the South Asian population. This strategy will shortly be expanded into the other local PCT with >75% South Asian TB cases. Longer term data on notifications should show whether there has been a fall compared with the non-intervention strategy of 1989–2001.

Abstract S14 Table



R. Y. Ahmed, T. G. D. Capstick, J. P. Watson. The Leeds Teaching Hospitals, NHS Trust, Leeds, UK

Background World Health Organization guidelines for managing multidrug-resistant tuberculosis (MDR-TB) are based on expert opinion and a review of retrospective cohort data, but there is no randomised controlled trial evidence. Numbers in the UK remain low (∼1% of isolates, 40–55 patients per year).1 We are aware of no UK published data on MDR-TB outcomes.

Methods We reviewed the case records of all MDR-TB patients treated in Leeds since 1996, looking at their demographic profile, drug resistance, treatment given, the cost and duration of treatment and their outcome.

Results 28 (19 male, 9 female) patients have been treated for MDR-TB in Leeds since 1996, 23 of them since 2003. Ages ranged from 11 months to 71 years (mean 32.8 years). 12 (5 white) were born in the UK. 7 had a previous history of TB. 75% had pulmonary disease. The sensitivity profile is shown in the table. An injectable drug was used in all except one patient who was under 1 year of age. The other commonly used drugs were: fluoroquinolone (25 patients), pyrazinamide (24), ethambutol (22) and prothionamide (20). Side effects were common, requiring a change in treatment in 13 patients. Out of 28 patients, 18 successfully completed their treatment. Five were transferred out, two are still on treatment and three died of their disease. Inpatient stay ranged from 0 to 205 days (mean 66). The mean number of outpatient visits was 31 (range 16–49). The cost of medications ranged from £5893 to £187 000 with a mean of £18 450. Data on length of treatment were available on 15 of the patients who have completed their treatment. 12 received treatment for 18 months and 3 for between 12 and 18 months. Of those completing treatment, there has been no evidence of recurrence to date (range 0–57 months, mean 27 months from completion).

Conclusion Treatment of MDR-TB is costly and side effects common, but good outcomes are possible in a specialist UK centre with current treatment strategy.

Abstract S15 Table



1K. M. Scott, 1F. J. C. Millard, 1R. M. Smith, 1E. Manomano, 2M. Glenshaw, 1R. D. Barker. 1Department of Respiratory Medicine, Kings College Hospital, London, UK, 2Murambinda Mission Hospital, Buhera, Zimbababwe

Background Zimbabwe has the second highest incidence of tuberculosis (TB) in the world (782/100 000/year). This is largely driven by the HIV epidemic. Furthermore, Zimbabwe has been in financial, economic and political decline for several years. Buhera is a rural district in southern Zimbabwe with a population of ∼230 000 people. In 2003 the district medical officer of Buhera requested financial and technical assistance for the TB programme from the British Charity, TB Alert. Here we report outcomes against key performance indicators for the first 5 years of the project.

Methods Site visits were undertaken by British physicians with an interest in TB in Africa. The aims of the project were to improve access to diagnosis of TB, to improve accuracy of diagnosis through increased use of sputum smear microscopy, to improve follow-up and treatment completion and to improve access to HIV care for those co-infected. Performance indicators were agreed and included the number of patients identified in relation to those expected, sputum smear rates, HIV testing and specific anti-HIV treatment administration and treatment outcomes. Some equipment was purchased and key salaries were supported. Approximately £16 000 per annum was contributed by the charity.

Results 5989 patients were registered in the 5-year period. 51% were men. The median age was 33 years (range 1–99) and 80.3% had pulmonary tuberculosis (PTB).

In 2008, 47% of patients with PTB were smear positive and 74% of those tested were HIV positive. Of 514 patients with sputum smear-positive PTB diagnosed in 2007, 218 (42.4%) were cured, 139 (27.0%) completed treatment, 3 (0.6%) failed treatment, 45 (8.8%) died, 3 (0.6%) interrupted treatment and 70 (13.6%) transferred out. Outcomes for 36 (7.0%) were not evaluated.

Discussion There has been a significant increase in TB notifications. Despite the increased burden, several aspects of the TB service in Buhera have improved significantly. Factors associated with success include good management, a committed work force, reasonable financial support and partnership with other non-governmental organisations, especially Médecins sans Frontières. We believe the results represent an outstanding return for the investment.

Acknowledgements We would like to think TB Alert for financial and moral support.

Abstract S16 Table

Performance indicators for a tuberculosis (TB) programme in Buhera District Zimbabwe


1J. P. Crofts, 1M. Kruijshaar, 2V. Delpech, 2F. Ncube, 1D. Quinn, 1I. Abubakar. 1Health Protection Agency Centre for Infections, Respiratory & Systemic Infections Department, London, UK, 2Health Protection Agency Centre for Infections, HIV & STI Department, London, UK

Background The National Health Service (NHS) has become increasingly reliant on healthcare workers recruited from overseas. Many of the countries of recruitment have high rates of tuberculosis and HIV. Screening, of new healthcare workers (HCWs) for tuberculosis is focused on “one-off” testing. HIV-co-infected individuals often test falsely negative for latent tuberculosis.

Objectives To investigate tuberculosis and HIV co-infection among HCWs.

Methods Enhanced Tuberculosis Surveillance (ETS) collects detailed information, including occupation, on reported tuberculosis cases. To identify cases co-infected with HIV, records in the national ETS database for the years 1999–2005 were matched with records in the national HIV database (1979–2006). Patient confidentiality was protected by unlinked pseudo-anonymised data linkage. Characteristics of HCWs with tuberculosis were compared by HIV status and tested for significance using the χ2 test or the Fisher exact test where appropriate.

Results Information on occupation was available for 25 696 (54%) cases of which 1627(6%) were HCWs. Compared with all other cases, a higher proportion of HCWs were co-infected with HIV (14.2% vs 5.3%). From 1999 to 2005 the prevalence of HIV among HCWs with tuberculosis almost doubled, with a peak prevalence of 19% in 2004 (fig 1). In total 98% of HCWs with tuberculosis–HIV co-infection were born outside the UK compared with 80% of HCWs with tuberculosis only (p<0.01). Co-infected HCWs were more often from sub-Saharan Africa (97%, p<0.01); of these, >70% were diagnosed with tuberculosis ⩾2 years after arrival. HCWs co-infected with HIV were also more likely than HCWs with tuberculosis only to have pulmonary disease (71% vs 52%, p<0.01) but had similar levels of multidrug-resistant tuberculosis (2.5% vs 1.4% p = 0.25).

Conclusions The burden of tuberculosis and HIV co-infection among HCWs falls mainly upon individuals born in high prevalence countries who often develop tuberculosis some years after arrival in the UK. One-off pre-employment screening for tuberculosis may not be enough to protect patient safety. In those HCWs who are at most risk, the provision of ongoing screening should be considered, and a high index of clinical suspicion for relevant symptoms is important.

Abstract S17 Figure 1

Prevalence of HIV co-infection in tuberculosis cases, England and Wales, 1999–2005.

View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.