Correspondence

therapy era), coupled with a poorly chest radiographs, 39 of 53 (74%) The Review by Alimuddin Zumla sensitive test (Ziehl–Neelsen had abnormalities consistent and colleagues1 of new anti- microscopy). Indeed, in individuals with previous infection, including tuberculosis treatments draws who are smear-negative, Xpert bronchiectasis and atelectasis attention to progress in tuberculosis MTB/RIF has a negative predictive resulting from scarring or fibrosis. drug development during the past value of 93%2 (vs 87% for all patients In high-HIV burden settings with decade, but also acknowledges that with suspected tuberculosis given a low diagnostic capacity, retreatment the treatment pipeline is inadequate smear, with sensitivity of 40% and of tuberculosis in the modern era to control tuberculosis and that prevalence of 20%), leaving some represents a common pathway international cooperation is needed doubt as to the true rule-out value for individuals with chronic lung to bring new treatment regimens. for a single test. In this context, disease who remain symptomatic This last point is key, and will not prescription of clinician behaviour after repeated interactions with happen without a new public health- within diagnostic randomised the public health system. In this driven framework for tuberculosis controlled trials can be ethically group, empiric treatment exposes research and development, or problematic.3 However, as countries patients to drug toxicities and without adequate funding. increasingly achieve widespread increased health-care costs without Although the registration of antiretroviral coverage leading benefit, imposing unnecessary and two new drugs for treatment of to true decreases in tuberculosis potentially substantial costs on multidrug-resistant tuberculosis is prevalence, burdens on health national tuberculosis programmes. welcome, today’s isolated research systems related to overdiagnosis of Continued movement towards and development process has not tuberculosis could increase . universal antiretroviral coverage delivered the new, shorter, more Patients with history of and widespread access to sensitive tolerable regimens urgently needed previous tuberculosis treatment diagnostic tests might ultimately to improve treatment outcomes.2 account for about 12% of global regain the confidence of clinicians No clinical studies or available data tuberculosis cases (nearly 700 000 working in resource-constrained exist on combination of the two people in 2012)4 and are notified settings that when a test result is new tuberculosis drugs, so potential from an even larger pool of negative tuberculosis treatment can new regimens incorporating these patients with suspected drug- safely be withheld. new classes of drugs are many years resistant tuberculosis, many of We declare no competing interests. away, and few regimens in planned whom undergo several courses clinical trials meet Médecins Sans of tuberculosis treatment. From *John Z Metcalfe, Peter Mason, Frontières’ ideal criteria for an Stanley Mungofa, Charles Sandy, November, 2011, to November, eff ective, all-oral, 6 month regimen Philip C Hopewell 2013, we did a prospective obser- for multidrug-resistant tuberculosis.2 vational study in Harare, Zimbabwe, Division of Pulmonary and Critical Care Medicine, Even more worryingly, with no San Francisco General Hospital, Curry incorporating Xpert MTB/RIF, International Tuberculosis Center, University of tuberculosis compounds in phase 1 microscopic observation drug California, San Francisco, CA, USA (JZM, PCH); clinical development3 the pipeline susceptibility (MODS), solid (LJ) Biomedical Research and Training Institute, will not be able to deliver the fl ow Harare, Zimbabwe (PM); Harare City Health and liquid (MGIT) culture into the Department, Harare, Zimbabwe (SM); and of new compounds necessary to diagnostic assessment for people National Tuberculosis Control Program, Harare, combat this rapidly mutating disease with suspected drug-resistant Zimbabwe (CS) in future. tuberculosis. Among sympto- 1 Theron G, Peter J, Dowdy D, Langley I, Further complicating this dire Squire SB, Dheda K. Do high rates of matic individuals registered as re- empirical treatment undermine the situation is the fact that tuberculosis treatment cases, 118 of 328 cases potential effect of new diagnostic tests for research and development funding tuberculosis in high-burden settings? (36%) did not test positive for Lancet Infect Dis 2014; 14: 527–32. decreased by US$30·4 million Mycobacterium tuberculosis, despite 2 Steingart KR, Schiller I, Home DJ, Pai M, in 2012, with private sector this extensive testing. Patients Boehme CC, Dendukuri N. Xpert(R) MTB/RIF investments dropping by 22% and assay for pulmonary tuberculosis and were, on average, aged 39 years rifampicin resistance in adults. Cochrane only 32% of the necessary funding for (SD 11), 86 of 118 (73%) were Database Syst Rev 2014; 1: CD009593. tuberculosis drug development met.4 HIV-infected, and 68 of 86 (79%) 3 Bossuyt PM, Reitsma JB, Linnet K, Moons KG. The present research and develop- Beyond diagnostic accuracy: the clinical were enrolled in antiretroviral utility of diagnostic tests. Clin Chem 2012; ment framework is clearly not programmes. 50 of 118 individuals 58: 1636–43. meeting the needs of patients, and, 4 World Health Organization. Global (42%) had been initiated on anti- tuberculosis report 2013. http://www.who. with the large-scale withdrawal of tuberculosis medicines at least int/tb/publications/global_report/en/ the private sector from tuberculosis twice before, and, of those with (accessed February 17, 2014). research and development, new

www.thelancet.com/infection Vol 14 September 2014 795 Correspondence

models of innovation must be HIV infection and Buruli 1511 Buruli ulcer cases were explored. included, of whom 78% (n=1177) Médecins Sans Frontières, in ulcer in Africa were PCR-confi rmed. HIV testing was consultation with other stakeholders, In a recent Personal View on Buruli positive in 34 of all 1511 patients has developed the “3P Project” ulcer, Daniel O’Brien and colleagues1 (2·3%) and in 25 of 500 patients (Push, Pull, Pool),5 an incentive highlighted the urgent need for (5·0%) aged 15–49 years. This framework for tuberculosis drug research on HIV–Buruli ulcer co- proportion is significantly higher regimen development. The 3P Project infection. Whether HIV infection is a than is the 1·1% estimate reported aims to deliver aff ordable, eff ective risk factor for Buruli ulcer is unknown, for this age group in Benin in 2012 new regimens for tuberculosis because the few studies published (p<0·0001).7 More than 70% (n=24) more efficiently, through an show contrasting results. In Ghana, of patients with HIV developed open, collaborative approach to HIV was not associated with Buruli severe Buruli ulcer, compared with drug development, and through ulcer (six of 116 Buruli ulcer cases 50% (n=723) of HIV-negative novel approaches to finance and with HIV versus one of 116 controls, patients (odds ratio [OR] 2·77, 95% CI coordination of research and p=0·89), whereas a significant 1·32–6·33; p=0·006). A focus on PCR- development, including push funding association was found in Benin (11 confi rmed Buruli ulcer cases further (ie, through grants), pull funding of 426 versus two of 613, p=0·003).2,3 validated this finding (OR 2·59, (ie, through milestone prizes), and This discrepancy could result from 1·06–7·27; p=0·037). The eff ect of HIV pooling of intellectual property to HIV being differentially associated on Buruli ulcer severity was driven ensure open collaborative research with subgroups of Buruli ulcer, such mainly by an increased frequency and fair licensing for high-quality as severe Buruli ulcer. Sporadic case of large or oedematous lesions in low-cost production of the final reports described patients with patients with HIV and Buruli ulcer. The products. HIV with both severe and non- eff ect of HIV on Buruli ulcer severity We agree with Zumla and col- severe evolution, for example.4,5 was driven mainly by an increased leagues that the specialty cannot be However, as emphasised by O’Brien frequency of large or oedematous complacent, and that to speed up and colleagues, no published lesions in patients with HIV and regimen development and re-engage epidemiological data exist on the Buruli ulcer (OR for large lesions in the private sector in tuberculosis association between HIV and severity patients with Buruli ulcer with and research and development novel of Buruli ulcer. without HIV 2·32, 95% CI 1·16–4·76, approaches are necessary. We did a large cohort study in a p=0·0174; OR for oedematous lesions We declare no competing interests. highly endemic region of Benin, 1·93, 0·94–3·86, p=0·0740; OR and addressed the question of the for bone lesions 0·96, 0·15–3·24, Manica Balasegaram, eff ect of HIV on the severity of Buruli p=0·95; and OR for multifocal *Grania Brigden ulcer. Clinical and laboratory data lesions 1·55, 0·25–5·29, p=0·58). [email protected] were prospectively obtained from all Adjustment for age and sex did not Access Campaign, Médecins Sans Frontières, consecutive patients with Buruli ulcer modify the results. 78 Rue de Lausanne, Geneva, Switzerland seen between 2005 and 2011 at the Although HIV-Buruli ulcer co- 1 Zumla A I, Gillespie S H, Hoelscher M, et al. New antituberculosis drugs, regimens, and Centre de Dépistage et de Traitement infection is a rare clinical event, our adjunct therapies: needs, advances, and de l’Ulcère de Buruli (CDTUB) in Pobe, data support O’Brien and colleagues’ future prospects. Lancet Infect Dis 2014; 14: 327–40. Benin. Clinical Buruli ulcer cases were hypothesis of a significant effect 2 Brigden G, Nyang’wa BT, du Cros P, et al. confi rmed by PCR and tested for HIV of HIV infection on Buruli ulcer Principles for designing future regimens for infection with two serological tests. severity. We also report evidence multidrug-resistant tuberculosis. 6 Bull World Health Org 2014; 92: 68–74. As suggested in our study, severe suggestive of a higher incidence of 3 Stop TB Partnership, Working Group on Buruli ulcer was defi ned as presenting HIV infection in patients with Buruli New TB Drugs. Drug Pipeline. 2014. ulcer compared with the general http://www.newtbdrugs.org/pipeline.php with oedematous, bone, large (accessed March 28, 2014). (≥15 cm in diameter), or multifocal population, although the absence 4 Treatment Action Group. 2013 report on lesions. The eff ect of HIV on Buruli of local controls in our design tuberculosis research funding trends, 2005–2012. 2013. http://www. ulcer severity was tested by logistic calls for further confirmation. This treatmentactiongroup.org/tbrd2013 regression, as implemented in the evidence off ers insights of profound (accessed March 28, 2014). glm function of the R software. Access significance with regards to Buruli 5 Médecins Sans Frontières Access Campaign. PUSH, PULL, POOL: accelerating to the data registry was approved by ulcer physiopathology, and opens innovation and access to medicines for the institutional review board of the new avenues for the development tuberculosis. 2014.http://www.msfaccess. org/push-pull-pool (accessed March 28, CDTUB and the national Buruli ulcer of novel preventive and therapeutic 2014). control authorities. strategies.

796 www.thelancet.com/infection Vol 14 September 2014