InternaƟ onal Union Against Tuberculosis and Lung Disease Public Health Action Health soluƟ ons for the poor VOL 3 SUPPLEMENT 1 PUBLISHED 4 NOVEMBER 2013

EDITORIAL Taking on the diabetes-tuberculosis epidemic in India: paving the way through operational research Srinath Satyanarayana,1 Ajay M. V. Kumar,1 Nevin Wilson,1 Anil Kapur,2 Anthony D. Harries,3,4 Rony Zachariah5 http://dx.doi.org/10.5588/pha.13.0039

n September 2011, a national stakeholders meeting of the two diseases, treatment of persons with dual AFFILIATIONS 1 International Union 5 Iwas held in Delhi, India, to discuss how to move for- disease and prevention of TB in persons with DM. For Against Tuberculosis and ward with bi-directional screening of tuberculosis (TB) each of these challenges, the knowledge gaps are high- Lung Disease (The Union), South-East Asia Office, and diabetes mellitus (DM). Agreement was reached lighted along with the research questions that need to New Delhi, India about how to 1) implement screening at hospitals and be answered if care and control of the dual burden of 2 World Diabetes Founda- tion, Gentofte, Denmark peripheral health institutions, 2) monitor and record disease are to be achieved. The second review article 3 The Union, Paris, France the process and outcomes of screening for each indi- looks at existing and new technologies for screening 4 London School of Hygiene vidual patient and 3) report on aggregate data at quar- and diagnosing type 2 DM that may be more suitable & Tropical Medicine, London, UK terly intervals. The meeting was followed by training for TB patients in low- and middle-income countries.6 5 Médecins Sans Frontières, of health care personnel involved in the work, imple- As pointed out, these new technologies should be low Medical Department, Operational Research Unit, mentation of screening and a reconvening of im- cost, rapid, easy to use, non-invasive, requiring mini- Brussels Operational plementing partners to share data and discuss chal- mal additional infrastructure and able to differentiate Center, Luxembourg, Luxembourg lenges. This body of work culminated in two published between transient and longer term hyperglycaemia. papers presenting the process and aggregate data on Several tools in development, such as point-of-care CORRESPONDENCE e-mail: ssrinath@theunion. 1,2 bi-directional screening of TB and DM in India. glycated haemoglobin and glycated albumin assays, org; [email protected] These pilot projects, conducted within the routine non-invasive advanced glycation end (AGE) product health services, produced good quality evidence that readers and sudomotor function-based screening de- has led to changes in policy and practice. A policy di- vices, are discussed. rective has been issued that all patients registered with The eight operational research papers assess 1) bi- TB in India should be screened for DM. Patient TB directional screening of the two diseases in one facil- treatment cards and TB registers have been modifi ed to ity,7 2) screening of DM patients for TB in one facility,8 accommodate these new parameters, recording whether and 3) screening of TB patients for DM in the other fa- screening has taken place for DM, whether the patient cilities,9–14 with one of these facilities also evaluating has received a diagnosis of DM and the results of blood treatment outcomes.13 A few key messages that are glucose measurements during TB treatment. A Minis- consistent across sites emerge. First, the yield of diabe- try of Health Training Manual on screening TB pa- tes was high among TB patients, with higher yields tients for DM has also been developed for health care seen among patients aged more than 35–40 years, pa- workers in the fi eld.3 Importantly, a policy directive tients with smear-positive pulmonary TB, current ciga- from the Directorate General of Health Services now rette smokers and those with recurrent TB. The pro- links India’s Revised National TB Control Programme portion with newly diagnosed DM as a result of blood to the National Programme for Non-Communicable test screening was higher among TB patients man- Diseases (NCD) at the sub-centre level so that data on aged in peripheral health facilities compared to ter- TB patients screened for DM are reported to the NCD tiary care centres, highlighting the need to prioritise programme. The NCD programme is in a nascent phase active screening efforts at the peripheral level. Second, in the country, but it is also evolving rapidly, and this the yield of TB among DM patients was relatively low, model of convergence, adapting and incorporating the and further research is required to optimise the screen- DOTS framework with its ‘cohort reviews’, is impor- ing criteria and diagnostic algorithms for diagnosing tant for this evolution. This bi-directional approach TB. One study showed that DM patients who were will be the fi rst of its kind to be implemented at na- male, older, had a longer duration of DM, required tional level, with such initiatives recently being en- combined oral hypoglycaemic drugs and insulin medi- dorsed by the 66th World Health Assembly Resolution cation and had poorly controlled DM were more likely on 25 May 2013.4 to have TB.7 Third, while the results reported are use- In this context, the current supplement on DM and ful, the one study that assessed treatment outcomes TB is timely and important. It consists of two review was not adequately powered to answer the question articles and eight operational research papers. The fi rst about whether DM adversely affects outcomes.12 There review article provides an up-to-date international per- was a statistically non-signifi cant trend towards failure spective on the epidemiology of and interaction be- of DM-TB patients to smear convert at 2 months, but tween DM and TB, and examines three important op- this whole area requires adequately powered, prospec- PHA 2013; 3(S1): S1–S2 erational challenges for care—bidirectional screening tive cohort research. © 2013 The Union Public Health Action Editorial S2

The interaction between DM and TB is rapidly becoming a hot 6 Adepoyibi T, Weigl B, Greb H, Neogi T, McGuire H. New screening technolo- topic for research, with projects examining the biological and mo- gies for type 2 diabetes mellitus appropriate for use in tuberculosis patients. Public Health Action 2013; 3 (Suppl): S10–S17. lecular reasons for the linkages and addressing questions about 7 Prakash B C, Ravish K S, Prabhakar B, et al. Tuberculosis-diabetes mellitus bi- how best to manage and integrate care. These operational research directional screening at a tertiary care centre, South India. Public Health Ac- studies from various sites in India will begin to pave the way to- tion 2013; 3 (Suppl): S18–S22. 8 Kumpatla S, Sekar A, Achanta S, et al. Characteristics of patients with diabe- wards a better understanding of the two diseases, in addition to tes screened for tuberculosis in a tertiary care hospital in South India. Public better care and, ultimately, better health outcomes. Health Action 2013; 3 (Suppl): S23–S28. 9 Dave P, Shah A, Chauhan M, et al. Screening patients with tuberculosis for References diabetes mellitus in Gujarat, India. Public Health Action 2013; 3 (Suppl): S29–S33. 1 India Tuberculosis-Diabetes Study Group. Screening of patients with tubercu- 10 Naik B, Kumar A M V, Satyanarayana S, et al. Is screening for diabetes among losis for diabetes mellitus in India. Trop Med Int Health 2013; 18: 636–645. tuberculosis patients feasible at the fi eld level? Public Health Action 2013; 3 2 India Diabetes Mellitus-Tuberculosis Study Group. Screening of patients with (Suppl): S34–S37. diabetes mellitus for tuberculosis in India. Trop Med Int Health 2013; 18: 11 Nair S, Kumari A K, Subramonianpillai J, et al. High prevalence of undiag- 646–654. nosed diabetes among tuberculosis patients in peripheral health facilities in 3 Central Tuberculosis Division, Ministry of Health and Family Welfare, Gov- Kerala. Public Health Action 2013; 3 (Suppl): S38–S42. ernment of India. Screening of tuberculosis patients for diabetes mellitus— 12 Achanta S, Tekumalla R R, Jaju J, et al. Screening tuberculosis patients for di- training module for the staff of the Revised National Tuberculosis Control abetes in a tribal area in South India. Public Health Action 2013; 3 (Suppl): Programme. New Delhi, India: Directorate General of Health Services, Minis- S43–S47. try of Health and Family Welfare, Government of India, 2013. 13 Khanna A, Lohya S, Sharath B N, Harries A D. Characteristics and treatment 4 World Health Organization. Sixty-Sixth World Health Assembly, Agenda response in patients with tuberculosis and diabetes mellitus in Delhi, India. Item 13. Follow-up to the Political Declaration of the High-level Meeting of Public Health Action 2013; 3 (Suppl): S48–S50. the General Assembly on the Prevention and Control of Non-Communicable 14 Jali M V, Mahishale V K, Hiremath M B, et al. Diabetes mellitus and smoking Diseases. A66/A/CONF./1 Rev.1 (25 May 2013). Geneva, Switzerland: WHO, among tuberculosis patients in a tertiary care centre in Karnataka, India. 2013. http://apps.who.int/gb/e/e_wha66.html Accessed August 2013. Public Health Action 2013; 3 (Suppl): S51–S53. 5 Harries A D, Satyanarayana S, Kumar A M V, et al. Epidemiology and inter- action of diabetes mellitus and tuberculosis and the challenges for care: a re- view. Public Health Action 2013; 3 (Suppl): S3–S9.

Public Health Action (PHA) The voice for operational research. e-ISSN 2220-8372 Published by The Union (www.theunion.org), PHA provides a platform to Editor-in-Chief: Donald A Enarson, MD, Canada fulfi l its mission, ‘Health solutions for the poor’. PHA publishes high-quality Contact: [email protected] scientifi c research that provides new knowledge to improve the accessibility, PHA website: http://www.theunion.org/index.php/en/journals/pha equity, quality and effi ciency of health systems and services. Article submission: http://mc.manuscriptcentral.com/pha InternaƟ onal Union Against Tuberculosis and Lung Disease Public Health Action Health soluƟ ons for the poor VOL 3 SUPPLEMENT 1 PUBLISHED 4 NOVEMBER 2013

REVIEW ARTICLE Epidemiology and interaction of diabetes mellitus and tuberculosis and challenges for care: a review A. D. Harries,1,2 S. Satyanarayana,3 A. M. V. Kumar,3 S. B. Nagaraja,4,5 P. Isaakidis,6 S. Malhotra,7 S. Achanta,4 B. Naik,4 N. Wilson,3 R. Zachariah,8 K. Lönnroth,9 A. Kapur10 http://dx.doi.org/10.5588/pha.13.0024 still exacts a huge toll, especially among the poorest AFFILIATION 1 International Union The global burden of diabetes mellitus (DM) is immense, 3 people on the globe. Several key obstacles continue to Against Tuberculosis with numbers expected to rise to over 550 million by thwart control efforts: many people are infected with and Lung Disease (The 2030. Countries in Asia, such as India and China, will bear Mycobacterium tuberculosis and are at risk of developing Union), Paris, France the brunt of this unfolding epidemic. Persons with DM 2 London School of active TB during their lifetime; many vulnerable peo- Hygiene & Tropical have a significantly increased risk of developing active tu- ple with presumptive TB do not have access to afford- Medicine, London, UK berculosis (TB) that is two to three times higher than in 3 The Union South-East able, high-quality TB diagnostic and treatment services; Asia Office, New Delhi, persons without DM. This article reviews the epidemiol- multidrug-resistant TB (MDR-TB, defi ned as M. tuber- India ogy and interactions of these two diseases, discusses how 4 Office of the WHO culosis resistant to at least isoniazid and rifampicin) is Representative in India, the World Health Organization and International Union a serious threat in some settings; and in Africa, espe- World Health Against Tuberculosis and Lung Disease developed and Organization, cially in the southern part of the continent, HIV/AIDS New Delhi, India launched the Collaborative Framework for the care and fuels an already burgeoning epidemic. Other risk fac- 5 Department of control of TB and DM, and examines three important Community Medicine, tors have also emerged in recent years as important Employees State challenges for care. These relate to 1) bi-directional screen- determinants of the TB epidemic, one of which is DM. Insurance Corporation ing of the two diseases, 2) treatment of patients with dual (ESIC) Medical College, In this review, we discuss 1) the global epidemiol- Bangalore, India disease, and 3) prevention of TB in persons with DM. For ogy of and evidence for interaction of the two diseases, 6 Médecins Sans Frontières, each area, the gaps in knowledge and the priority re- Brussels Operational DM and TB; 2) the new collaborative framework for Centre, Mumbai, India search areas are highlighted. Undiagnosed, inadequately the care and control of TB and DM; and 3) challenges 7 All India Institute of treated and poorly controlled DM appears to be a much Medical Sciences, for care. New Delhi, India greater threat to TB prevention and control than previ- 8 Medical Department, ously realised, and the problem needs to be addressed. Operational Research Unit, Médecins Sans Prevention of DM through attention to unhealthy diets, GLOBAL EPIDEMIOLOGY Frontières, Brussels sedentary lifestyles and childhood and adult obesity must OF DIABETES MELLITUS Operational Centre, be included in broad non-communicable disease preven- AND TUBERCULOSIS Luxembourg, Luxembourg tion strategies. This collaborative framework provides a Diabetes mellitus 9 Stop TB Department, template for action, and the recommendations now need World Health DM is a chronic condition that occurs when the body Organization, Geneva, to be implemented and evaluated in the field to lay down Switzerland a firm foundation for the scaling up of interventions that cannot produce enough insulin or cannot use it effec- 10 World Diabetes tively. As a result, a person with DM does not utilise Foundation, Gentofte, work and are effective in tackling this dual burden of Denmark disease. glucose properly, and glucose circulates in the blood at high levels (hyperglycaemia), causing tissue damage CORRESPONDENCE A D Harries over time. DM was fi rst recognised in 1500 BC by the Old Inn Cottage hronic, non-communicable diseases such as car- ancient Egyptians, and the term diabetes mellitus was Vears Lane Colden Common diovascular disease, diabetes mellitus (DM), can- used by Greek physicians to describe persons with uri- C Winchester SO21 1TQ, UK cer and chronic obstructive pulmonary diseases have nary frequency and in whom the urine also tasted Tel: (+44) 1962 714 29 4 e-mail: adharries@ emerged as the next twenty-fi rst century global epi- sweet. According to the World Health Organization theunion.org demic, and have already become the leading causes of (WHO), three common forms of DM account for the 1 KEY WORDS death and disability worldwide. Among these diseases, majority of cases: type 1, type 2 and gestational dia- DM; TB; DM-TB interaction; the global burden of DM is immense. In 2012, there betes mellitus (GDM); their defi nitions are given in Ta- epidemiology; challenges were an estimated 371 million people living globally ble 1.5 People with high blood glucose levels that are for care with DM, with numbers expected to rise to 552 mil- below the thresholds required for a diagnosis of DM lion by 2030.2 Over half of these patients are undiag- are said to have pre-diabetes. This may occur either as nosed, and complications due to diabetes are a major impaired fasting glucose or as impaired glucose toler- cause of disability and reduced quality of life. ance (IGT), and these people have an increased risk of Tuberculosis (TB) is a major communicable disease type 2 DM. which, along with the human immunodefi ciency vi- People with DM may suffer from a number of seri- Received 3 May 2013 rus and acquired immune-defi ciency syndrome (HIV/ ous health problems, such as cardiovascular disease, Accepted 15 July 2013 AIDS) and malaria, is responsible for considerable mor- renal disease (nephropathy), eye disease (retinopathy), bidity and mortality worldwide. Although TB control nerve damage (neuropathy), diabetic foot and infec- PHA 2013; 3(S1): S3–S9 has come a long way in the last 20 years, the disease tions that include TB. DM may be diagnosed on the © 2013 The Union Public Health Action Interaction of DM and TB S4

TABLE 1 Classification of diabetes mellitus* TABLE 2 World Health Organization diagnostic criteria ACKNOWLEDGEMENTS for DM or pre-diabetes* A workshop was convened in Classification Definition Delhi, India, for the purpose of writing the papers that are DM Pre-diabetes Type 1 DM Type 1 diabetes mellitus encompasses the published in this supple- ment. The workshop was run majority of cases that are primarily due to Fasting plasma glucose ⩾7.0 mmol/l 6.1–6.9 mmol/l by the Centre for Opera- pancreatic islet beta-cell destruction, ⩾126 mg/dl 110–125 mg/dl tional Research, International attributable either to an autoimmune Union Against Tuberculosis 2 h plasma glucose ⩾11.1 mmol/l 7.8–11.0 mmol/l process or to an unknown cause and Lung Disease (The after OGTT ⩾200 mg/dl 140–199 mg/dl (idiopathic). These patients are prone to Union), Paris, France; The Union South-East Asia ketoacidosis and require insulin injections Random plasma glucose ⩾11.1 mmol/l ⩾ Office, New Delhi, India; the for survival. It does not include cases with 200 mg/dl Operational Research Unit, beta-cell destruction or failure to which Glycosylated haemoglobin ⩾6.5% Médecins Sans Frontières, specific causes can be assigned (e.g., cystic Luxembourg; the World fibrosis, mitochondrial defects, etc.) * Adapted from World Health Organization and International Diabetes Health Organization Country Office in India, New Delhi, Federation.6 Type 2 DM Type 2 is the most common form of India; the All India Institute DM = diabetes mellitus; OGTT = oral glucose tolerance test. diabetes mellitus. It is characterised by of Medical Sciences, New disorders of insulin action and insulin Delhi, India; and ESIC Medical College, Bangalore, secretion, either of which may be the India. predominant feature. Both are usually Funding for the workshop present at the time that this form of The majority of people with DM are in the age and open access publication diabetes manifests clinically. By definition, group 40–59 years, with little difference in terms of sex. was received from the World the specific reasons for the development of Diabetes Foundation, The mean age at onset of type 2 DM in LMICs is at Gentofte, Denmark. these abnormalities are not yet known. least a decade earlier than in industrialised countries. The views expressed in this Gestational DM Gestational diabetes is carbohydrate Given the association between DM and lifestyles such article are the sole responsi- intolerance resulting in hyperglycaemia of bility of the authors and do as unhealthy diet and physical inactivity, there are more not necessarily reflect the variable severity, with onset or first people with DM in urban areas than in rural areas, with positions and views of their recognition during pregnancy. It does not affiliated institutions. exclude the possibility that the glucose the divide in 2030 estimated at 314 million to 143 mil- Conflict of interest: none intolerance may antedate pregnancy but lion, respectively. It is estimated that about 50% of peo- declared. had been previously unrecognised. The ple who have DM, mostly those with type 2 disease, definition applies irrespective of whether or not insulin is used for treatment or are undiagnosed. In 2012, an estimated 4.8 million whether the condition persists after people aged 20–79 years died from DM, with little dif- pregnancy. Gestational DM usually resolves ference between males and females. LMICs account after pregnancy, but mothers and babies for 88% of all premature mortality due to DM.2 both have a higher risk of developing Asia is the global region most affected by DM.2,9 type 2 DM later in life. The disease develops at a younger age than in white * Adapted from World Health Organization.5 populations of European descent, and associated car- = DM diabetes mellitus. diovascular disease is common in young Asian people. China and India are the two countries with the highest basis of fasting plasma glucose, 2-hour plasma glucose prevalence of DM, with respectively 92.3 million and following a 75 g oral glucose tolerance test (OGTT), 63 million people aged 20–79 years estimated to have random plasma glucose in a patient with classic the disease in each country. In India, DM prevalence hyperglycaemic symptoms or glycosylated haemoglo- peaks at 60–69 years, while in China it peaks later, at bin (HbA1c). The different cut-off values used by the 70–89 years. Other high-burden Asian countries include WHO, which are generally those used in most low- Indonesia, Japan, Pakistan, Bangladesh, Malaysia and and middle-income countries (LMICs), are shown in the Philippines. Type 2 DM in Asia is often associated Table 2.6 In recognition of the widespread use of capil- with a strong family history of diabetes, with most lary sampling in these countries, fasting values for Asian patients having a fi rst-degree relative with the venous and capillary plasma glucose are regarded as disease. It is unclear at present whether this strong identical,6 although not all front-line clinicians are family history is genetic and/or whether it is a result prepared to accept this stance. of shared risk behaviour and early life programming. Recent, up-to-date estimates of the worldwide DM burden are provided by the International Diabetes Tuberculosis Federation in their Diabetes Atlas.2 Data for this are TB is an infectious disease of equally great antiquity as supplemented through epidemiological studies, health DM, with evidence of spinal involvement being found examination surveys, population-based prevalence in Egyptian mummies dating back to several thousand studies and WHO STEPwise surveillance studies.7,8 In years BC. Although common in various populations the 2012 update to the Atlas, it was estimated that 371 over the centuries, it acquired notoriety during the million people worldwide had DM. About 80% live in Industrial Revolution in Europe as ‘the captain of all LMICs, and, if the trends of the past 10–15 years con- these men of death’.10 The discovery of the tubercle tinue, over 550 million people will have DM by 2030. bacillus in 1882 by Robert Koch, the documentation This equates to one in ten globally, although there will of the close link between TB and poverty, undernutri- be marked differences from country to country, de- tion and poor living conditions, the discovery and use pending on factors such as urbanisation and lifestyle. of anti-tuberculosis drugs between the 1940s and 1970s In 2012, a further 280 million people worldwide had and the introduction of combination chemotherapy IGT, with the number expected to rise to 398 million to cure disease and prevent the development of drug by 2030. resistance all led to modern-day control efforts. Public Health Action Interaction of DM and TB S5

In 1994, the WHO developed a standardised framework for TB 20 million lives, and as a consequence of such efforts, the incidence control, branded as ‘DOTS’.11 The crucial components of DOTS of and mortality from TB have been falling for several years. are: 1) standardised case fi nding and registration, 2) standard- Despite the progress made, TB continues to be a major public ised anti-tuberculosis treatment regimens for new and previously health problem in many LMICs. In 2011, there were an estimated treated disease, and 3) monitoring, recording and reporting of 8.7 million cases of TB (13% co-infected with HIV) and 1.4 mil- cases and treatment outcomes using internationally agreed defi ni- lion TB-related deaths.15 The burden of TB is highest in Asia and tions.12 The Stop TB Partnership, which has been in existence for Africa, with India and China together accounting for almost 40% nearly a decade and has over 700 partners worldwide, endorsed of the world’s TB cases; India has an estimated 2.2 million cases two epidemiological targets in 2000, linked to the Millennium per annum and China about 1.0 million. Globally, 5.8 million TB Development Goals. These targets are by 2015 to reduce TB preva- cases were notifi ed and reported to the WHO in 2011, leaving lence and death rates by 50% compared with 1990 levels, and by 2.9 million patients unaccounted for. The emergence of MDR-TB 2050 to eliminate TB as a public health problem (defi ned as an in- has been of major global concern: of the 310 000 estimated new cidence of less than 1 TB case per million people per year). To MDR-TB cases in 2011, only 19% were notifi ed and even fewer achieve these targets, a Stop TB Strategy was agreed upon in 2006 were treated according to international recommendations.15 (Table 3) and launched along with a Global Plan to Stop TB.13,14 About one third of the world’s population is infected with M. tu- The Stop TB Strategy is focused on the expansion of high quality berculosis, of whom about 5–10% will develop active TB at some DOTS, but at the same time addresses other important issues time in their lives, the greatest risk being within the fi rst few years such as HIV-associated TB, drug-resistant TB, engagement of all of infection. Infants and young children up to the age of 5 years public and private health care providers, health systems strength- are at relatively high risk, while those aged 5–15 years are rela- ening, patient-centred approaches and programme-based opera- tively resistant; the risk then rises again through adolescence, re- tional research. mains stable during adulthood and increases once again in the el- From 1995, DOTS-based programmes were initiated worldwide, derly. Other risk factors, including DM, are shown in Table 4. and by 2011, 204 countries and territories were using DOTS and HIV/AIDS is the strongest risk factor, with the incidence of TB in- reporting data to the WHO. During this time, the global case detec- creasing as the CD4 lymphocyte count declines.16 Although at the tion rate of TB rose from 15% to 67%, and the treatment success individual level the risk of developing TB is considerably lower in rate in new smear-positive pulmonary TB rose from 77% to 87%.15 people with DM compared with those who have HIV, the much It is estimated that 51 million people have been successfully treated larger and rapidly growing pool of people with DM makes the for TB in countries that have adopted the DOTS strategy, saving global and population-attributable fraction of TB due to DM simi- lar to that seen with HIV.3

TABLE 3 The 2006 Stop TB Strategy* INTERACTION BETWEEN DIABETES MELLITUS 1 Pursue high quality DOTS expansion and enhancement AND TUBERCULOSIS • Secure political commitment with adequate and sustained funding • Ensure case detection through quality-assured bacteriology Evidence of interaction • Provide standardised treatment, with supervision and patient support The association between DM and TB was suggested as far back as • Ensure an effective drug supply and management system • Monitor and evaluate performance and impact Roman times and, since then until recently, anecdotal reports and case studies have kept the link alive. As effective treatment for 2 Address TB-HIV, MDR-TB and needs of poor, vulnerable populations • Implement and scale up collaborative TB-HIV activities both DM and TB became available and successful public health • Scale up the prevention and control of MDR-TB measures were put in place to control TB, the association between • Address the needs of prisoners, refugees and other high-risk groups the two diseases was perceived to be less relevant, especially as TB and situations became relatively rare in high-income countries where DM was 3 Contribute to health systems strengthening based on primary prevalent, while DM was believed to be a minor problem in poor health care countries where TB was endemic. • Actively participate in efforts to improve health policies, human resources, financing, management, service delivery and The situation has changed dramatically in the last decade, with information systems recognition of the enormous and unfolding epidemic of DM in • Strengthen infection control in health services and LMICs, a slower decline in global TB incidence rates than would congregate settings be expected from epidemiological modelling, and a rekindling of • Upgrade laboratory networks and implement the Practical Approach interest in the association between DM and TB. The interaction to Lung Health • Adapt successful approaches and innovations from other fields between the two diseases was brought to global attention by three important publications in 2007 and 2008. Catherine Stevenson 4 Engage all care providers • Involve public-public and public-private mix (PPM) approaches • Promote the use of International Standards for Tuberculosis Care (ISTC) TABLE 4 Risk factors for the development of active TB 5 Empower people with TB and communities through partnership • HIV/AIDS • Pursue advocacy, communication and social mobilisation • Other causes of immune suppression (e.g., treatment with • Foster community participation in TB care and prevention corticosteroids) • Promote use of the Patient’s Charter for Tuberculosis Care • Silicosis 6 Enable and promote research • Malnutrition • Conduct programme-based operational research • Indoor air pollution • Advocate research to develop new diagnostics, drugs and vaccines • Cigarette smoking • Harmful alcohol use and other drug abuse * Adapted from World Health Organization.13 • Diabetes mellitus TB = tuberculosis; HIV = human immunodeficiency virus; MDR-TB = multidrug- resistant tuberculosis (Mycobacterium tuberculosis resistant to at least isoniazid and TB = tuberculosis; HIV = human immunodeficiency virus; AIDS = acquired immune- rifampicin). deficiency syndrome. Public Health Action Interaction of DM and TB S6 and colleagues undertook a Medline literature search of studies TABLE 5 Collaborative activities to reduce the dual burden of published after 1995 that quantifi ed the association between DM diabetes and TB* and TB.17 They identifi ed nine studies that all showed signifi cant 1 Establish mechanisms for collaboration and clinically important associations, with the increase in the risk • Set up means of coordinating diabetes and TB activities (odds) of TB varying from 1.5 to 7.8 for those with DM. Limita- • Conduct surveillance of TB disease prevalence among people with tions were that most studies had not measured or controlled ade- diabetes in medium and high TB burden settings quately for potential confounders. The same group further as- • Conduct surveillance of diabetes prevalence in TB patients in all countries sessed through an epidemiological model the potential impact of • Conduct monitoring and evaluation of collaborative diabetes and TB DM as a risk factor for incident pulmonary TB in India.18 In 2000, activities it was estimated that DM accounted for 14.8% of pulmonary TB 2 Detect and manage TB in patients with diabetes cases (uncertainty range 7.1–23.8%) and 20.2% of smear-positive • Intensify detection of TB among people with diabetes pulmonary TB cases (uncertainty range 8.3–41.9%). A systematic • Ensure TB infection control in health care settings where diabetes is review and meta-analysis by Jeon and Murray used various data- managed • Ensure high-quality TB treatment and management in people bases to identify observational studies dating back to 1965 that with diabetes had reported an age-adjusted quantitative estimate of the associa- 3 Detect and manage diabetes in patients with TB 19 tion between DM and active TB. Thirteen observational studies • Screen TB patients for diabetes met the selection criteria, with 1 786 212 participants and 17 698 • Ensure high-quality diabetes management among TB patients TB cases. Across the three cohort studies, the relative risk of TB in * Adapted from World Health Organization and International Union Against Tubercu- DM patients was 3.1 (95% confi dence interval [CI] 2.3– 4.3), and losis and Lung Disease.26 in the case control studies the odds ratios (ORs) ranged from 1.16 TB = tuberculosis. to 7.83. Further reviews have confi rmed these fi ndings and have suggested that overall the risk of TB in persons with DM is two to CHALLENGES FOR CARE FOR DUAL DISEASE three times higher.20,21 Within the Framework for Care and Control of DM and TB,26 Is the interaction between diabetes mellitus there are three main challenges, which are discussed below. and tuberculosis biologically plausible? The answer is ‘yes’. DM increases the general risk of infection, but Bi-directional screening of diabetes and tuberculosis the precise mechanisms by which DM predisposes to TB are still in routine settings not clear and require further research. Unlike the situation with The framework provisionally recommends screening for DM in all HIV infection, in which cell-mediated immunity is gradually people diagnosed with TB, and screening for TB in people with compromised by progressive depletion and dysfunction of CD4 DM in countries with a high prevalence of TB (>100 per 100 000 T-lymphocytes, DM impairs cell-mediated immunity by impairing population). The latter recommendation has recently been re- the function and activation of macrophages, monocytes and lym- peated in the WHO’s 2013 guidelines on systematic screening for phocytes, with additional potential roles played by pulmonary active TB in risk groups.27 microangiopathy, renal dysfunction and vitamin defi ciencies.17 In A systematic review of bi-directional screening for DM and TB addition, patients with uncontrolled hyperglycaemia appear to be in 2009 using strict inclusion criteria identifi ed 12 studies on at higher risk for TB than those with controlled blood glucose lev- screening people with DM for TB and 18 studies on screening TB els, suggesting that uncontrolled hyperglycaemia is an important patients for DM.28 In persons with DM, screening for TB showed determinant in this interaction.22,23 high rates of TB, ranging from 1.7% to 36%, with the screening yield being highly dependent on the underlying TB prevalence Collaborative Framework for Care and Control and on the severity of DM. Because DM is much more common of Diabetes and Tuberculosis than TB, in patients with TB the screening for DM yielded a high In December 2009, an expert meeting was held in Paris to review prevalence of DM, ranging from 1.9% to 35%. the evidence of the association between DM and TB,24 to defi ne Studies in the last 2–3 years have shown a high frequency of the research agenda needed to reduce the dual burden of disease,25 TB and DM co-morbidity in many areas of the world. Unpub- and to assess whether the evidence was strong enough to formu- lished reports show a very high prevalence of DM in people with late international guidelines for the care and control of dual dis- TB in the South Pacifi c of between 40% and 45%. In South India, ease. After much refl ection, it was decided that the currently avail- one study in the state of Karnataka showed a DM prevalence of able evidence to support specifi c recommendations, although 32% in TB patients;29 in the state of Kerala, 44% of TB patients incomplete, was suffi cient to develop a provisional framework for had DM,30 and in the state of Tamil Nadu, 25% of TB patients had action, and that this should be used to stimulate further research DM and a further 25% had IGT.31 On the Texas-Mexico border, to strengthen the evidence base for appropriate interventions. the prevalence of DM among TB patients was 39% in Texas and This framework was launched in August 2011, and it now serves 36% in Mexico.32 In a recent study from Tanzania, the prevalence as a guide to help policy makers and implementers move forward of DM in TB patients was 16.7% compared with 9.4% amongst to combat the looming epidemic of DM and TB.26 The framework controls, with a stronger association among non-HIV-infected includes provisional recommendations (pending better evidence) compared with HIV-infected patients.33 A similar case-control packaged under three main themes: 1) the establishment of mech- study in Pakistan reported a DM prevalence of 16.0% and an IGT anisms for collaboration, 2) the detection and management of TB prevalence of 34% among TB patients, compared with 7% and in patients with DM, and 3) the detection and management of 28%, respectively, in the general community.34 DM in patients with TB (Table 5).26 Operational and other research There is little published information on whether DM has any is encouraged to build the evidence base, which will be reviewed effect on TB case notifi cations or estimated case numbers at coun- in 2015 to determine whether recommendations at that time can try level. An ecological association between changes in DM and be made more defi nitive. TB prevalence across 46 countries has been observed,35 and an Public Health Action Interaction of DM and TB S7 analysis in India has suggested that increased DM prevalence be- disease (age, sex, body mass index), the effects of co-morbid dis- tween 1998 and 2008 contributed to an increase in the total num- ease and associated exposures such as alcohol and smoking, and ber of TB cases which exceeded the rate of population growth how dual care can best be integrated within the same health facil- during the same time period.36 It will be important to monitor ity or clinic while at the same time paying attention to good TB these associations more closely in the future, especially in coun- infection control. tries in Asia with escalating DM epidemics. At the health facility level, there are various challenges with Treatment of patients with dual disease bi-directional screening. In TB patients, DM cannot be recognised A systematic review of studies from 1980 to 2010 focused on the clinically and some form of blood glucose measurement is there- treatment of patients with dual disease and assessed the results of fore required to determine its presence or absence. How this is sputum culture conversion at 2–3 months of anti-tuberculosis best done in a routine situation still needs to be clarifi ed. In India treatment, death during treatment, and relapse of TB after success- and China, two large operational studies showed that TB patients ful completion of treatment.46 There were nine studies assessing could be screened for DM at the time of registration by asking the infl uence of DM on prolonging culture positivity at 2–3 months fi rst about the presence or absence of known DM, and, in those of treatment, with six studies reporting relative risks (RRs) of >2, denying any known disease, using random blood glucose mea- and three reporting RRs of <1. The risk of death during TB treat- surements to identify those at risk, followed by fasting blood glu- ment was assessed in 23 studies, with a pooled and signifi cantly cose measurements in those needing further screening.37,38 In higher RR of 1.89 (95%CI 1.52–2.36) in patients with DM. Four of both countries, the large majority of patients were willing to be these studies adjusted for age and other potential confounders screened, and a total of 12–13% of TB patients screened had DM, and found a pooled OR of 4.95 (95%CI 2.69–9.10), suggesting with previously unrecognised new disease being diagnosed in 3% that patients who die during anti-tuberculosis treatment have of patients in China and 5% of patients in India based on fasting other strong risk factors for death, such as HIV and co-morbidities, blood glucose values. The majority of these patients were referred which reduce the impact of DM in the unadjusted analyses. Five for diabetes care. Screening of TB patients is thus useful in help- studies assessed the risk of TB relapse, with a pooled and increased ing to identify the large proportion of persons with DM who have signifi cant RR of 3.89 (95%CI 2.43–6.23) in patients with DM. undiagnosed disease in the community and who would benefi t Whether these patients experienced a recurrence of the former in- from earlier diagnosis and treatment. fection (true relapse) or re-infection with a new strain of M. tuber- Many questions, however, need to be answered. For example, culosis is not known. No consistent associations were found be- previous studies have shown that it may be more reliable to screen tween DM and drug-resistant TB, although this area requires more for DM later in the course of anti-tuberculosis treatment rather detailed and prospective research. than at the start, because as a chronic infectious disease TB may Current international guidelines disseminated by the WHO elevate blood glucose levels, resulting in false-positive diagno- recommend that treatment for new cases of TB be standardised ses.39–42 On the other hand, in a large DM prevalence study in for 6 months with rifampicin-based TB treatment regimens, re- China,7 it was found that screening with fasting blood glucose gardless of other co-morbidities such as HIV infection or DM.12 missed nearly half of the DM patients diagnosed with a 2 h 75 g Whether longer or different treatment in TB patients with DM OGTT. The latter test is suitable for research surveys but inappro- has a benefi cial effect on treatment outcomes and whether it re- priate for screening individuals within routine general health ser- duces the risk of recurrent disease is not known, and needs pro- vices. These issues highlight the need at the screening level for spective research. Similarly, whether more aggressive management better, simpler and non-expensive point-of-care glucose measure- and control of DM might improve the treatment response to TB ment tests. remains unclear. There are many other unanswered questions that In India and China, pilot studies assessed the screening of DM include the infl uence of poor DM control on death and recurrent patients for TB with a traditional symptom-screen approach every TB, the best oral hypoglycaemic drugs to use with anti-tuberculosis time the patient came to the clinic, and referral of those with a drugs, drug-drug interactions and overlapping toxicities, the tim- positive screen for TB investigations.43,44 This approach was feasi- ing and aetiology of death in DM patients, the reasons for recur- ble and resulted in high detection rates of TB that varied from rent disease, and interventions that may reduce the frequency of 300 to 800 per 100 000 people screened per quarter in China to these adverse events. Moreover, as second-line anti-tuberculosis 600–950/100 000 in India. However, several operational chal- treatment is increasingly being offered to patients in settings with lenges were identifi ed that need to be overcome if such screening a high MDR-TB burden, we need to study the challenges of co- is to be rolled out on a larger scale. These include 1) the reluc- a dministration of second-line anti-tuberculosis agents with oral tance of busy DM doctors to take on the additional work needed hypoglycaemic drugs. These questions are best answered through to screen for and monitor this infectious disease, 2) the low sensi- prospective studies, such as a recently published study from south- tivity of current TB diagnostic approaches that rely on sputum ern Mexico, which found that patients with DM and TB had smear examination and chest radiography, and 3) the absence of higher rates of delayed sputum conversion, treatment failure and structured recording or reporting systems in most DM clinics, recurrent TB after treatment, with one fi fth of the recurrence due making it diffi cult to have reliable denominators for calculating to new infection.47 While there is no direct evidence that early TB case detection rates. In particular, symptom screening has low diagnosis and optimal treatment of DM in people with TB im- sensitivity, especially for the detection of TB early in the course of proves TB treatment outcomes, such an effect is plausible. More- the disease. Screening by chest radiography would give a higher over, considering the high prevalence of undiagnosed DM in peo- yield,45 but the cost and logistical challenges would also be greater. ple with TB, screening and proper management of DM in TB More research is needed to determine the most appropriate patients should be pursued to improve general health outcomes. screening and diagnostic algorithm and the cost-effectiveness of different approaches. Preventing tuberculosis in people with diabetes With both bi-directional screening systems, more information Two studies conducted before the 1970s showed that active TB in is also needed about what phenotypes are especially at risk of dual people with DM could be prevented with TB chemoprophylaxis.48,49 Public Health Action Interaction of DM and TB S8

However, both studies were methodologically fl awed, and the true 10 Harries A D, Dye C. Tuberculosis. Ann Trop Med Parasitol 2006; 100: 415– 431. benefi t of chemoprophylaxis remains unknown. The question can 11 World Health Organization. Global tuberculosis programme. Framework for effective tuberculosis control. WHO/TB/94.179. Geneva, Switzerland: WHO, only be properly answered through a randomised controlled trial. 1994. As poor glucose control is associated with a higher risk of devel- 12 World Health Organization. Treatment of tuberculosis: guidelines for na- oping TB, it is plausible that improved DM management in itself tional programmes. 4th ed. WHO/HTM/TB/2009.420. Geneva, Switzerland: WHO, 2010. is a means of preventing TB disease in a person with latent tu- 13 World Health Organization. The Stop TB Strategy. Building on and enhancing berculous infection. However, no trial has been conducted to DOTS to meet the TB-related Millennium Development Goals. WHO/HTM/ evaluate such an effect. Finally, it will be important to ascertain TB/2006.368. Geneva, Switzerland: WHO, 2006. whether DM clinics are areas where TB transmission occurs, and, 14 Stop TB Partnership/World Health Organization. Global plan to stop TB 2006– 2015. WHO/HTM/STB/2006.35. Geneva, Switzerland: WHO, 2006. if so, to determine the measures that need to be put in place to 15 World Health Organization. Global tuberculosis report 2012. WHO/HTM/ prevent this spread of TB. TB/2012.6. Geneva, Switzerland: WHO, 2012. 16 Lawn S D, Harries A D, Williams B G, et al. Antiretroviral therapy and the control of HIV-associated tuberculosis. Will ART do it? Int J Tuberc Lung Dis CONCLUSION 2011; 15: 571–581. 17 Stevenson C R, Critchley J A, Forouhi N G, et al. Diabetes and the risk of tu- In LMICs, and especially in Asia, the DM epidemic is growing rap- berculosis: a neglected threat to public health. Chronic Illn 2007; 3: 228–245. idly. There is now strong evidence that there is an important asso- 18 Stevenson C R, Forouhi N G, Roglic G, et al. Diabetes and tuberculosis: the impact of the diabetes epidemic on tuberculosis incidence. BMC Public ciation between DM and TB and that this association results in Health 2007; 7: 234. poor TB treatment outcomes. Undiagnosed, inadequately treated 19 Jeon C Y, Murray M B. Diabetes mellitus increases the risk of active tubercu- and poorly controlled DM appears to be a much greater threat to losis: a systematic review of 13 observational studies. PLOS Med 2008; 5: e152. 50,51 TB care and prevention than previously realised, and this needs 20 Dooley K E, Chaisson R E. Tuberculosis and diabetes mellitus: convergence to be tackled. Upstream prevention of DM through attention to of two epidemics. Lancet Infect Dis 2009; 9: 737–746. healthier diets, more physical activity and reduction of childhood 21 Ruslami R, Aarnoutse R E, Alisjahbana B, van der Ven A J, van Crevel R. Im- and adult obesity must be included in broad non-communicable plications of the global increase of diabetes for tuberculosis control and pa- tient care. Trop Med Int Health 2010; 15: 1289–1299. disease prevention strategies. Simple point-of-care diagnostic tests 22 Leung C C, Lam T H, Chan W M, et al. Diabetic control and risk of tubercu- are needed for DM, and simple, standardised and affordable inter- losis: a cohort study. Am J Epidemiol 2008; 167: 1486–1494. ventions to treat and monitor established DM and IGT that are 23 Harries A D, Lin Y, Satyanarayana S, et al. The looming epidemic of diabetes- associated tuberculosis: learning lessons from HIV-associated tuberculosis. easy to access must be scaled up in all countries. Adapting DOTS Int J Tuberc Lung Dis 2011; 15: 1436–1444. principles may help to bring a more structured approach to DM 24 Ottmani S-E, Murray M B, Jeon C Y, et al. Consultation meeting on tubercu- care delivery, something that is currently lacking in most high losis and diabetes mellitus: meeting summary and recommendations. Int J 52,53 Tuberc Lung Dis 2010; 14: 1513–1517. DM burden countries. Heightened clinical suspicion for TB is 25 Harries A D, Murray M B, Jeon C Y, et al. Defi ning the research agenda to re- needed for people with DM, especially among those who live in duce the joint burden of disease from diabetes mellitus and tuberculosis. Trop TB-endemic areas, and systematic screening should be considered. Med Int Health 2010; 15: 659–663. However, more research is needed to establish appropriate TB 26 World Health Organization/International Union Against Tuberculosis and Lung Disease. Provisional collaborative framework for care and control of screening eligibility criteria and algorithms in people with DM. A tuberculosis and diabetes. WHO/HTM/TB/2011.15. Geneva, Switzerland: framework has now been established for reducing the dual bur- WHO, 2011. den of disease, and pilot studies have started to show how to im- 27 World Health Organization. Systematic screening for active tuberculosis: principles and recommendations. WHO/HTM/TB/2013.04. Geneva, Switzer- plement some of the screening and monitoring initiatives. A col- land: WHO, 2013. lection of evidence is needed to bring about the scaling up of 28 Jeon C Y, Harries A D, Baker M A, et al. Bi-directional screening for tubercu- interventions that work and that clearly make a difference. losis and diabetes: a systematic review. Trop Med Int Health 2010; 15: 1300– 1314. 29 Gupta S, Shenoy V P, Bairy I, Srinivasa H, Mukhopadhyay C. Diabetes melli- References tus and HIV as co-morbidities in tuberculosis patients of rural South India. 1 Lozano R M, Naghavi M, Foreman K, et al. Global and regional mortality from J Infect Public Health 2011; 4: 140–144. 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis 30 Balakrishnan S, Vijayan S, Nair S, et al. High diabetes prevalence among tu- for the Global Burden of Disease Study. Lancet 2012; 380: 2095–2128. berculosis cases in Kerala, India. PLOS ONE 2012; 7: e46502. 2 International Diabetes Federation. IDF diabetes atlas. 5th ed, 2012 update. 31 Viswanathan V, Kumpatla S, Aravindalochanan V, et al. Prevalence of diabetes Unwin N, Whiting D, Guariguata L, et al., eds. Brussels, Belgium: Interna- and pre-diabetes and associated risk factors among tuberculosis patients in In- tional Diabetes Federation, 2012. http://www.idf.org/diabetesatlas/5e/Update dia. PLOS ONE 2012; 7: e41367. 2012 Accessed August 2013. 32 Restrepo B I, Camerlin A J, Rahbar M H, et al. Cross-sectional assessment re- 3 Lönnroth K, Castro K G, Chakaya J M, et al. Tuberculosis control and elimi- veals high diabetes prevalence among newly-diagnosed tuberculosis cases. nation 2010–2050: cure, care, and social development. Lancet 2010; 375: Bull World Health Organ 2011; 89: 352–359. 1814–1829. 33 Faurholt-Jepsen D, Range N, Praygod G, et al. Diabetes is a risk factor for 4 Polonsky K S. The past 200 years in diabetes. N Engl J Med 2012; 367: 1332– pulmonary tuberculosis: a case-control study from Mwanza, Tanzania. PLOS 1340. ONE 2011; 6: e24215. 5 World Health Organization, Department of Non-communicable Disease Sur- 34 Codlin A, Nadeem A, Lotia I, et al. Diabetes, pre-diabetes and tuberculosis in veillance. Defi nition, diagnosis and classifi cation of diabetes mellitus and its an Asian mega-city: Karachi, Pakistan. Int J Tuberc Lung Dis 2012; 16 complications. Report of a WHO Consultation. WHO/NCD/NCS/99.2. Ge- (Suppl 1): S338–S339. [Abstract] neva, Switzerland: WHO, 1999. 35 Goldhaber-Fiebert J D, Jeon C Y, Cohen T, Murray M B. Diabetes mellitus 6 World Health Organization/International Diabetes Federation. Defi nition and tuberculosis in countries with high tuberculosis burdens: individual and diagnosis of diabetes mellitus and intermediate hyperglycaemia. Sum- risks and social determinants. Int J Epidemiol 2011; 40: 417– 428. mary of technical report and recommendations. Report of a WHO/IDF con- 36 Dye C, Trunz B B, Lönnroth K, Roglic G, Williams B G. Nutrition, diabetes sultation. Geneva, Switzerland: WHO, 2006. and tuberculosis in the epidemiological transition. PLOS ONE 2011; 6: 7 Yang W, Lu J, Weng J, et al. Prevalence of diabetes among men and women e21161. in China. N Engl J Med 2010; 362: 1090–1101. 37 Li L, Lin Y, Mi F, et al. Screening of patients with tuberculosis for diabetes 8 Danaei G, Finucane M M, Lu Y, et al. National, regional, and global trends in mellitus. Trop Med Int Health 2012; 17: 1294–1301. fasting plasma glucose and diabetes prevalence since 1980: systematic analy- 38 India Tuberculosis-Diabetes Study Group. Screening of patients with tuber- sis of health examination surveys and epidemiological studies with 370 culosis for diabetes mellitus in India. Trop Med Int Health 2013; 18: 636–645. country-years and 2.7 million participants. Lancet 2011; 378: 31– 40. 39 Kishore B, Nagrath S P, Mathur K S, Hazra D K, Agarwal B D. Manifest, chem- 9 Ramachandran A, Ma R C W, Snehalatha C. Diabetes in Asia. Lancet 2010; ical and latent chemical diabetes in pulmonary tuberculosis. J Assoc Physi- 375: 408– 418. cians India 1973; 21: 875–881. Public Health Action Interaction of DM and TB S9

40 Goyal B N, Nigam P, Dubey A L, Joshi L D, Saxena H N. Study of the diabetic 47 Jimenez-Corona M E, Cruz-Hervert L P, Garcia-Garcia L, et al. Association of status in pulmonary tuberculosis. J Diabetic Association India 1978; 18: 191– diabetes and tuberculosis: impact on treatment and post-treatment out- 197. comes. Thorax 2013; 68: 214–220. 41 Singh M M, Biswas S K, Shah A. Impaired glucose tolerance in active pulmo- 48 Pfaffenberg R, Jahler H. [Isoniazid and recurrence of tuberculosis in diabet- nary tuberculosis. Indian J Tuberc 1984; 31: 118–121. ics]. Z Tuberk 1958; 111: 167–173. [German] 42 Oluboyo P O, Erasmus R T. The signifi cance of glucose intolerance in pulmo- 49 Lesnichii A V, Karpina L Z. [Experience with the chemoprophylaxis of pul- nary tuberculosis. Tubercle 1990; 71: 135–138. monary tuberculosis in diabetes mellitus patients]. Probl Tuberk 1969; 47: 43 Lin Y, Li L, Mi F, et al. Screening patients with diabetes mellitus for tubercu- 1–3. [Russian] losis in China. Trop Med Int Health 2012; 17: 1302–1308. 50 Sullivan T, Amor Y B. The co-management of tuberculosis and diabetes: chal- 44 India Diabetes Mellitus-Tuberculosis Study Group. Screening of patients with lenges and opportunities in the developing world. PLOS Med 2012; 9: diabetes mellitus for tuberculosis in India. Trop Med Int Health 2013; 18: e1001269. 646–654. 51 Kapur A, Harries A D. The double burden of diabetes and tuberculosis—pub- 45 van’t Hoog A H, Langendam M W, Mitchell E, et al. A systematic review of lic health implications. Diabetes Res Clin Pract 2013; Jan 7: pii: S0168- the sensitivity and specifi city of symptom- and chest-radiography screening 8227(12)00497-4. for active pulmonary tuberculosis in HIV-negative persons and persons with 52 Harries A D, Jahn A, Zachariah R, Enarson D. Adapting the DOTS framework unknown HIV status. Report–Version March 2013. World Health Organiza- for tuberculosis control to the management of non-communicable diseases tion, Geneva, Switzerland: WHO, 2013. http://who.int/tb/Review2Accuracy in sub-Saharan Africa. PLOS Med 2008; 5: e124. ofscreeningtests.pdf Accessed 7 August 2013. 53 Khader A, Farajallah L, Shahin Y, et al. Cohort monitoring of persons with 46 Baker M A, Harries A D, Jeon C Y, et al. The impact of diabetes on tuberculosis diabetes mellitus in a primary health care clinic for Palestine refugees in Jor- treatment outcomes: a systematic review. BMC Med 2011; 9: 81. dan. Trop Med Int Health 2012; 17: 1569–1576.

Le fardeau mondial du diabète sucré (DM) est immense et l’on chaque secteur, les déficiences en matière de connaissances et les s’attend à ce que le nombre de cas augmente d’ici 2030 jusqu’à plus zones prioritaires de recherche sont soulignées. Un diabète sucré non de 550 millions. Les pays d’Asie, comme l’Inde et la Chine, devront diagnostiqué, traité de manière inadéquate et médiocrement con- supporter le poids principal de cette épidémie en expansion. Le risque trôlé semble une menace beaucoup plus importante pour la préven- de développer une tuberculose (TB) active est significativement accru tion et la lutte contre la TB qu’on ne l’avait pensé précédemment, et chez les patients atteints de DM : il est de deux à trois fois supé- ce problème doit être abordé. La prévention du diabète grâce à une rieur à celui des personnes sans DM. Cet article fait la revue de attention portée aux régimes inadéquats, au style de vie sédentaire et l’épidémiologie et des interactions de ces deux maladies, discute la à l’obésité de l’enfant et de l’adulte doit être incluse dans de larges façon dont l’Organisation Mondiale de la Santé et l’Union Internatio- stratégies de prévention des maladies non transmissibles. Le réseau nale Contre la Tuberculose et les Maladies Respiratoires ont élaboré et de collaboration fournit un modèle d’action et les recommandations lancé le réseau de collaboration pour les soins et la lutte contre la TB doivent à présent être mises en œuvre et évaluées sur le terrain afin et le DM et examine trois défis importants pour les soins de ces affec- de donner un fondement solide à l’extension d’interventions qui tions. Ceux-ci sont en relation avec 1) un dépistage bidirectionnel des fonctionnent et sont efficientes pour lutter contre le double fardeau deux maladies, 2) le traitement des patients atteints des deux mala- de ces maladies. dies, et 3) la prévention de la TB chez les sujets atteints de DM. Pour

La carga mundial de morbilidad por diabetes sacarina (DM) es con- 3) la prevención de la TB en las personas diabéticas. Se destacaron las siderable y se prevé que se sobrepasen los 550 millones de personas deficiencias en los conocimientos y los dominios prioritarios de investi- en el 2030. Países de Asia como la India y la China soportarán la gación en cada aspecto. La DM no diagnosticada, tratada inadecua- mayor parte de la carga de esta epidemia en expansión. Las personas damente o mal equilibrada constituye una amenaza a la prevención y que padecen DM presentan un riesgo considerable de contraer tu- la lucha contra la TB, que es más determinante de lo que se conside- berculosis (TB) activa, el cual es de dos a tres veces mayor que el raba y es preciso atenderla. La prevención de la DM mediante la cor- riesgo de las personas que no sufren DM. En el presente artículo se rección de los regímenes poco saludables, los estilos de vida sedenta- consideran las características epidemiológicas de ambas enferme- rios y la obesidad infantil se debe incorporar a las estrategias globales dades, se analizan la elaboración y la puesta en marcha del marco de prevención de las enfermedades no transmisibles. El marco con- conjunto de atención y de lucha contra la TB y la DM de la Orga- junto de atención ofrece un modelo para la acción y es preciso poner nización Mundial de la Salud y la Unión Internacional contra la Tuber- en práctica sus recomendaciones y evaluarlas en el terreno, a fin de culosis y las Enfermedades Respiratorias y se examinan además tres sentar unas bases firmes a la ampliación de escala de las intervenciones dificultades importantes que plantea la atención. Estos problemas que dan resultados y abordan eficazmente esta carga de morbilidad hacen referencia a: 1) el cribado bidireccional de ambas enfermedades, doble. 2) el tratamiento de pacientes aquejados de ambas enfermedades, y

Public Health Action (PHA) The voice for operational research. e-ISSN 2220-8372 Published by The Union (www.theunion.org), PHA provides a platform to Editor-in-Chief: Donald A Enarson, MD, Canada fulfi l its mission, ‘Health solutions for the poor’. PHA publishes high-quality Contact: [email protected] scientifi c research that provides new knowledge to improve the accessibility, PHA website: http://www.theunion.org/index.php/en/journals/pha equity, quality and effi ciency of health systems and services. Article submission: http://mc.manuscriptcentral.com/pha InternaƟ onal Union Against Tuberculosis and Lung Disease Public Health Action Health soluƟ ons for the poor VOL 3 SUPPLEMENT 1 PUBLISHED 4 NOVEMBER 2013

REVIEW ARTICLE New screening technologies for type 2 diabetes mellitus appropriate for use in tuberculosis patients T. Adepoyibi,1 B. Weigl,1 H. Greb,1 T. Neogi,1,2 H. McGuire1 http://dx.doi.org/10.5588/pha.13.0036 rate, rapid, non-invasive and cost-effective point-of- AFFILIATIONS 1 PATH, Washington, DC, Type 2 diabetes mellitus (DM), which is epidemic in low- care (POC) diagnostic and monitoring tests—including and middle-income countries (LMICs), may threaten USA measurements of blood glucose and blood glycated 2 University of Washington gains made in tuberculosis (TB) control, as DM is both a hemoglobin (A1c)—was acknowledged in 2011 at a con- School of Medicine, major risk factor for developing active TB and it can lead Department of Family sultation meeting of global experts on TB and DM.10 Medicine, Seattle, to adverse TB treatment outcomes. Despite World Health Technologies for the detection of diabetes face spe- Washington, USA Organization guidance that all TB patients should be cial challenges in their use among TB patients.9 As CORRESPONDENCE screened for DM, most facilities in LMICs that manage TB with other infectious diseases, TB may temporarily ele- Temitope Adepoyibi patients do not currently perform screening for DM, due PATH vate blood glucose levels, resulting in false-positive 455 Massachusetts Ave in part to the cost and complexity involved. DM screen- DM diagnoses.9,11 Current DM detection tools are not Washington, DC 20001, USA ing is further complicated by the presentation of transient Fax: (+1) 202 457 1466 optimized for LMICs, where the majority of TB pa- e-mail: [email protected] hyperglycemia in many TB patients, as well as differences tients are found,4 and are often inaccurate, expensive, in diabetes risk factors (e.g., body mass index) between ACKNOWLEDGEMENTS invasive and inconvenient.12 Multiple tests are often The authors acknowledge TB patients and the general public. In this article, we re- required to establish a diagnosis, as is referral to a spe- the work of J K Drake, PATH. view existing and new technologies for DM screening Conflict of interest: none cialist center for confi rmatory diagnosis.13 These cir- declared. that may be more suitable for TB patients in LMICs. Such cumstances present special challenges for TB patients methods should be rapid, they should not require fasting, KEY WORDS who are already burdened with the fi nancial costs as- DM; TB; technology; and they should allow the provider to differentiate between sociated with 6–24 months of TB treatment, depend- screening transient and longer-term hyperglycemia, using inexpen- ing on whether drug resistance is present or not. A sys- sive tools that require little training and no specialized in- tematic review found that the costs (both direct and frastructure. Several methods that are currently under de- indirect) associated with TB treatment in sub-Saharan velopment, such as point-of-care glycated hemoglobin Africa can be as much as 10 times the average annual and glycated albumin assays, non-invasive advanced gly- income for TB patients in the poorest 20% of the pop- cation end-product readers, and sudomotor function-based ulation, and for many households such costs are con- screening devices, offer interesting performance charac- sidered catastrophic.14 teristics and warrant evaluation in populations with TB. Nonetheless, screening and treatment for hypergly- cemia and DM is essential, due to the potential impact on TB treatment outcomes and the effectiveness of TB ype 2 diabetes mellitus (DM) triples the risk of de- medications.15,16 Individuals with elevated blood glu- Tveloping tuberculosis (TB), and rates of TB are cose may have delayed or impaired recovery, face higher in people with DM than in the general popula- worse treatment outcomes, or may even require al- tion.1,2 DM is also associated with adverse TB treat- tered doses or schedules of TB medications.15,16 ment outcomes.2 With the epidemiological transition This article builds on the recent publication from underway in many countries around the world, cur- Weigl and Drake in Point of Care,12 and will outline a rent predictions indicate that the prevalence of DM proposed target product profi le for a POC DM test for will reach 552 million by 2030.3 Approximately 80% use in TB patients. Optimal test characteristics, includ- of these cases will be in low- and middle-income coun- ing accuracy, suitability for use in current TB program tries (LMICs), where TB prevalence is high.4 Although structures, patient acceptability, suitability for use given the links between diabetes and TB have been known TB patient infection status, convenience and cost will for decades,5,6 the unprecedented global rise in DM led be considered. The current pipeline for POC DM screen- the World Health Organization (WHO) and the Inter- ing technologies appropriate for LMICs will then be national Union Against Tuberculosis and Lung Disease examined to identify the candidate technologies that to issue a global recommendation in 2011 that all TB best meet the optimal test characteristics as described, patients should be screened for DM and vice versa,4 and which may require further evaluation within TB and countries including China and India have com- populations. menced TB-DM screening programs.7–9 Despite this, Existing technologies and approaches for Received 31 May 2013 global response to the crisis has been hampered by a Accepted 3 August 2013 lack of knowledge regarding the most appropriate diabetes screening in tuberculosis patients screening methods and technologies to use in TB set- According to the WHO, DM is diagnosed by the pres- PHA 2013; 3(S1): S10–S17 tings.4 The need to develop and evaluate more accu- ence of one of the following:17 a fasting plasma glucose © 2013 The Union Public Health Action DM screening for TB patients S11 level ⩾ 7.0 mmol/l (126 mg/dl); plasma glucose ⩾ 11.1 mmol/l score may lead to a biochemical test.19, 28 The benefi ts of using DM (200 mg/dl) 2 h after a 75 g oral glucose load in a glucose toler- risk scores to evaluate DM risk in TB patients in the context of TB ance test; symptoms of hyperglycemia and casual plasma glu- programs include the fact that they are relatively easy to adminis- cose ⩾ 11.1 mmol/l (200 mg/dl); glycated hemoglobin (HbA1c) ter (TB program staff could potentially be trained relatively easily ⩾ 6.5%.18 A clinical DM diagnosis also requires retesting on an- to administer a DM risk score), promote recommended guidelines other day by any of the above methods, unless the patient is relating to targeted biochemical screening, and are low cost in symptomatic and the plasma glucose is unequivocally elevated.18 themselves.28 One disadvantage associated with using DM risk It is important to note the overlap between DM screening and scores in the context of TB programs is the fact that, of the seven diagnosis, as various screening and diagnosis algorithms may be risk scores recommended for adaptation in routine clinical prac- employed.19 The WHO recommends screening TB patients for DM tice,29 all were developed in North America or Europe, and none using existing DM country guidelines.4 Table 1 shows DM screen- of the 22 countries that account for 90% of the global TB burden ing and diagnosis guidelines from a selection of the 22 countries is in Europe or North America.30 This raises questions regarding with the highest TB burdens in the world from which informa- the applicability of the risk scores in these countries. In addition, tion regarding national DM guidelines was available. 20–27 not enough is known about how active TB disease infl uences the Screening approaches for DM in TB patients—as in the general phenotypic characteristics associated with DM as measured by population—may include questionnaire-based risk assessments standard risk scores; for example, TB-DM co-morbid patients may (used to create risk scores) as well as blood, urine, and non-invasive have, on average, a lower BMI or lower hip-to-waist ratio than tests that determine the levels of biochemical markers or physio- their TB-negative DM patient counterparts. This article proposes logical responses correlated with diabetes risk.19,28 that more research is required to develop a risk score optimized for use in TB populations. Diabetes risk assessments DM risk assessment questionnaires commonly include the follow- Biochemical tests ing components: age; sex; family history; biometric measurements The most commonly used biochemical diabetes screening and di- such as body mass index (BMI), waist circumference, and waist- agnostic tests are blood or plasma glucose tests, variants of which to-hip ratio; and history of hypertension.19,28 Risk of DM is gener- include random plasma glucose (RPG), random blood glucose, ated through a scoring system related to each of these attributes, fasting plasma glucose (FPG), fasting blood glucose (FBG) and the and a biochemical test may be included automatically, or a high oral glucose tolerance (OGTT) and challenge tests.17 Glucose

TABLE 1 DM screening and diagnosis guidelines from selected high TB burden countries20–27

Country DM screening DM diagnostic criteria Bangladesh20 Venous RPG At least two of: More than one characteristic and sign of DM Venous FPG >7.0 mmol/l Urine glucose Venous RPG taken at least 2 h after eating or after taking 75 g glucose ⩾11.1 mmol/l Presence of diabetic retinopathy, or Random sample on more than one occasion >11.1 mmol/l Brazil21 Risk score FBG >126 mg/dl, and 2 h after ingestion of 75 g Capillary glucose test anhydrous glucose, ⩾200 mg/dl + symptoms FBG ⩾126 mg/dl on more than one occasion FBG >126 mg/dl + symptoms China22 RBG FBG ⩾7.0 mmol/l on at least two occasions Ethiopia23 Capillary glucose test At least two of the following on subsequent days: FBG ⩾126 mg/dl RBG ⩾200 mg/dl 2 h oral glucose test ⩾200 mg/dl, as confirmed on a subsequent day by any one of the three methods India24 Capillary glucose test, 2 h post-glucose load ⩾200 mg/dl random Risk score FBG ⩾ 126 mg/dl Kenya25 Capillary glucose test Venous FPG >7.8 mmol/l on more than one occasion Plasma glucose >11.1 mmol/l in symptomatic patients South Africa26 Risk score In patients without symptoms, at least two of the following: FPG FPG ⩾7.0 mmol/l 2 h plasma glucose ⩾11.1 mmol/l RPG RPG ⩾11.1 mmol/l A1c A1c >6.5% One of the above + symptoms Uganda27 Risk score Blood glucose Urine glucose Blood glucose A1c A1c

DM = diabetes mellitus; TB = tuberculosis; RPG = random plasma glucose; FBG = fasting blood glucose; RBG = random blood glucose; FPG = fasting plasma glucose; A1c = blood glycated hemoglobin. Public Health Action DM screening for TB patients S12 testing may be conducted via venous blood usually drawn for lab- the case of RPG, capillary glucose samples give values that are oratory instrument-based testing, or via capillary blood—or fi nger- higher than venous sample values, and therefore value conver- prick—for glucometer testing using portable POC devices. FPG sion must be performed.18 This complication is in addition to the and RPG are single-point glucose measurements taken in blood poor and variable sensitivity and specifi city of RPG of respectively after fasting or at random times, respectively, whereas the OGTT 40–79% and 66–96%.32 measures the response to a glucose challenge by measuring blood A1c has advantages over both FPG and RPG in that the test glucose levels before and after receiving it (the fi rst glucose mea- does not require fasting and is generally more sensitive than FPG, surement being an FPG).28 The OGTT is considered the gold stan- at 78–81% sensitivity, although its specifi city is 79–84%.32 A1c is dard diagnostic test for DM,17 although its inconvenience (pa- especially sensitive in the detection of early type 2 DM in at-risk tients must wait for 2 h) and cost makes FPG preferable in clinical subjects.35 The main drawbacks in the use of A1c for TB patients practice settings.12,28 Other common tests include urine glucose are the lack of standardization and the cost; the consumables re- and A1c measurements.19 Although the WHO has advised that quired for POC-compatible A1c instruments, and the instruments A1c can also be considered a diagnostic test for DM, in addition themselves, are cost-prohibitive in many LMICs.12 In addition, to its use as a glycemic control monitoring test for diagnosed DM race, ethnicity and hemoglobin variants may all impact on A1c patients,31 of the countries listed in Table 1, only South Africa and readings, as is the case with conditions such as anemia (including Uganda use it for either purpose. The usefulness of urinary glu- that caused by malaria) and vitamin B12 defi ciency.36,37 Taking cose as a screening test for undiagnosed DM is limited because of these factors into consideration, there is a paucity of data regard- its low sensitivity, which some studies show to range from 21% to ing A1c performance from the high TB burden countries featured 64%, although specifi city is high, at >98%.19 in Table 1. Despite the low sensitivity of urinary glucose testing, it There are notable disadvantages with each of these currently may be useful in LMICs where no other procedure is possible.19 used and recommended tests in relation to the TB care context. Because of the importance of high sensitivity in a DM screening FPG and OGTT both require patient preparation in the form of test for TB patients, however, urinary glucose testing is considered fasting, and in the case of OGTT, extended clinic waiting times to have serious limitations. and multiple blood draws are required.12,32 These approaches The transitory hyperglycemia that many TB patients exhibit therefore reduce the ability to test TB patients opportunistically as due to the nature of the disease often resolves upon commence- they come into contact with the health system, thus delaying di- ment of TB treatment.4 To this end, to accurately represent true agnosis and resulting in increased risk of serious consequences, as DM presence or risk, a DM screening test for TB patients should TB patients with DM are more likely to experience delays in spu- represent a deviation of actual blood glucose from normal values tum culture conversion, increased case fatality rates during treat- averaged over a time period long enough to distinguish between ment, and increased relapse rates of TB after successful comple- temporary elevation in blood glucose brought on by active TB. tion of treatment.2,33 In many of the countries listed in Table 1, The biochemical tests commonly used in the high TB burden DM screening is performed via FPG or RPG, using a POC glucom- countries shown in Table 1 face disadvantages in this regard. FPG eter device on a capillary specimen. The WHO has recognized the and RPG are single-point glucose measurements and are therefore widespread use of capillary sampling in LMICs, although venous unable to identify chronically elevated blood glucose, and OGTT plasma glucose is considered the standard measurement and re- measures the response to a glucose challenge.12 Of the commonly porting method.18 Although venous and capillary measurements available tests, A1c is the most appropriate for identifying chronic give the same fasting result, unfortunately the most cost-effective hyperglycemia, as it represents blood glucose levels averaged over and commonly used glucometer devices exhibit universally low 3 months,36 although as discussed, the test is cost-prohibitive for accuracy, often due to incorrect and/or infrequent calibration.34 many LMICs.12 Moreover, although it has a specifi city of >90%, the sensitivity of Table 2 reviews key attributes and performance of these vari- FPG is only 40–65%.32 RPG does not require fasting and is there- ous tests for DM screening and diagnosis.17,35,38 The selection of fore more conducive to opportunistic screening of TB patients, a screening tool can be limited by cost, complexity of use and a but as with FPG, POC glucometer devices are commonly used. In lack of trained personnel, and, in some cases, a lack of availability.

TABLE 2 Common DM screening and diagnostic tests17,35,38

Test FPG, FBG OGTT RPG, RBG Urine A1c Measures Glucose post 8 h fast Glucose after fasting Glucose regardless of Presence of glucose % of Hb that is glycated + 2 h post-glucose load when person last ate in urine Specimen Capillary venous blood Capillary venous blood Capillary venous blood Voided mid-stream Capillary venous blood urine Test Glucometer laboratory Glucometer laboratory Glucometer laboratory Dipstick Glucometer laboratory measurement measurement measurement measurement Performance 40–65% sensitivity, 96.8% max sensitivity, 40–79% sensitivity, 21–64% sensitivity; 78–81% sensitivity, >90% specificity17 90.8% max specificity38 66–96% specificity17 specificity >98%17 79–84% specificity35 Normal range FPG ⩽99 mg/dl 2 h PGL ⩽139 mg/dl RPG <200 mg/dl 0–15 mg/dl A1c <5.7% (7.0 mmol/l) (7.0 mmol/l) (11/1 mmol/l) Pre-diabetes FPG = 100–125 mg/dl 140–199 mg/dl Cannot be used to >15 mg/dl A1c = 5.7–6.4%. (5.6–6.9 mmol/l); IFG (7.8–11.0 mmol/l); IFG diagnose pre-diabetes Distinguishes No No No No Yes transient hyper- glycaemia

DM = diabetes mellitus; FPG = fasting plasma glucose; FBG = fasting blood glucose; OGTT = oral glucose tolerance test; RPG = random plasma glucose; RBG = random blood glucose; A1c = blood glycated hemoglobin; PGL = post-glucose load; IFG = impaired fasting glucose. Public Health Action DM screening for TB patients S13

Typically the urine glucose test strips are the least expensive, be able to be deployed within optimized algorithms that may in- while the hemoglobin A1c tests are the most expensive screen- clude non-device-based screening approaches, and algorithm re- ing option.39 search and development should be prioritized in this regard. This article proposes the following ideal attributes for a DM POC screen- Tuberculosis program requirements ing test for use in TB patients: TB patients in the public sector are usually diagnosed and man- • Ability to distinguish between short-, medium- and longer-term aged via two mechanisms: 1) through a vertical TB program struc- glycemic control with a single measurement at a single point in ture, via TB clinics and microscopy centers/laboratories staffed time. To avoid false-positive results, determining multiple pa- by specialized personnel, as is the case in India,40 a country that rameters correlated to chronic hyperglycemia is especially im- a ccounted for 26% of global TB cases in 2011,30 or 2) in an inte- portant in TB patients, as active TB can result in transient hy- grated manner, where TB is one of numerous conditions managed perglycemia.7 Such a test would be useful not only for as part of a package of primary health care interventions overseen screening but also for glycemic control monitoring in patients by multidisciplinary staff within one facility, as in Mexico.41 Often already identifi ed as having prediabetes or DM. a combination of the two systems is present, in that the TB clinic • Low cost. A benchmark cost for comparison would be US$1, and microscopy center/laboratory are physically co-located within which is comparable to that of the glucose test strips used for the primary health care structure, and staff are shared between RPG, FPG, or OGTT tests.12 Cost will be a critical consideration 42 the two, as in Tanzania. To reduce the burden of referral on TB in the operationalization of the WHO’s recommendation to patients (even within the same structure), an optimal screening screen all TB patients for DM, and resource-poor TB programs test is one that can be used safely and with acceptance by the and nascent, underfunded non-communicable disease programs cadres of clinical and laboratory staff who currently manage TB pa- will have to work together to identify additional budgetary allo- tients, integrated within their current systems and using existing cations for screening and ongoing care and management of infrastructure, such as ambient temperature storage and adequate TB-DM patients. waste disposal for portable glucometer test strips. A recent study • Ability to be administered without fasting. Ideally, there should highlighted that a waste management system in TB microscopy be no preparation required for the patient, to avoid the need 43 centers existed in only 50% of the 22 high-burden TB countries. for repeat visits. Fasting may present an additional burden for The indoor temperature always remained below 35ºC in only 23% TB patients in particular, who already face signifi cant loss of 43 of the countries, but 82% had gloves in stock. income and inconvenience due to the repeated interactions The DM burden in the 22 high-burden TB countries continues with the health system required by 6–24 months of treatment. to grow,3 and it is estimated that globally between 30% and 90% • Minimal waiting time for results. The presence of DM increases of those with DM are undiagnosed.28 In examining the various the risk of poor treatment outcomes in TB patients.2 It is there- disadvantages associated with current DM screening tests relating fore imperative that DM in TB patients be identifi ed and man- to their use in TB patients, it is clear that to address the WHO’s aged without delay. recommendation for the universal screening of TB patients for • High sensitivity. Although a screening test/device should ide- DM, two distinct screening needs emerge: 1) a POC DM assay that ally have as high a sensitivity and specifi city as possible, trade- requires no patient preparation, is acceptably accurate, sensitive offs always have to be made with respect to test function and and specifi c, is low cost, requires minimal infrastructure and is sensitivity and specifi city. A screening tool should primarily easy to use; and 2) an assay that differentiates transient hypergly- serve a ‘rule-out’ function, and have as high a sensitivity as cemia from pre-diabetes and DM. possible to ensure that patients are not mistakenly declared negative,46 and, in the case of a TB patient, falsely determined TARGET PRODUCT PROFILE FOR A DIABETES to be DM-negative, risking adverse treatment outcomes. POINT-OF-CARE SCREENING TEST FOR • Simplicity of use. The reality of many settings is that the man- TUBERCULOSIS PATIENTS agement of TB patients is left to lower cadres of health workers. To reduce the burden of referral on TB patients, an optimal DM A target product profi le (TPP) is simply a statement of the essen- screening device should be simple enough for use by the health tial attributes of a clinically and commercially successful product; care and laboratory staff currently providing care to TB patients. many TPPs, however, are not widely publicized,44 despite the fact • Non-invasive/minimally invasive. The priority for any screening that they can be very valuable in terms of evaluating the appro- device should be to minimize patient inconvenience and dis- priateness of existing technologies, and in guiding discovery and comfort by requiring only non-invasive or minimally invasive development activities.45 A TPP can stimulate innovation and de- sample collection. velopment by reducing the risks for companies that are considering • Minimal requirement for associated infrastructure. Given the in- developing products for underserved markets such as those found frastructure constraints experienced by TB microscopy centers in many high-burden TB countries, by providing a consensus doc- in the 22 high-burden TB countries,43 an ideal test/device ument containing optimal target product specifi cations developed would require no maintenance or calibration, and would have by experts and informed by clinical needs assessments.44,45 no temperature or storage control requirements. There should In identifying a TPP for a DM POC screening test for TB pa- not be any need for reagents that are not included with the tients, there is signifi cant overlap with potential TPPs for general disposable or kit. populations in LMIC settings. Indeed, the TPP proposed here builds • Indigenous manufacture and distribution. Distribution and mar- upon the TPP for a DM POC screening test for low-resource set- keting costs for glucose tests can comprise more than 50% of tings described by Weigl and Drake.12 Although there are overlaps, the total end-user costs.12 Any optimal test/device should be there is utility in identifying the attributes of a DM screening test able to be manufactured in high TB burden countries and dis- that can be easily deployed within TB program structures, poses tributed using existing supply chains, to keep costs to a mini- minimal burden upon TB patients, and is appropriate given the mum. The cost savings associated with producing a DM test physiological attributes of this population. Appropriate tests should close to or within a target country can have a huge impact. Public Health Action DM screening for TB patients S14

DIABETES SCREENING TECHNOLOGY of a chain of chemical reactions after an initial glycation reac- PIPELINE—ARE THERE CANDIDATES tion.50,51 Specifi c and accurate measurement of AGEs requires gra- SUITABLE FOR TUBERCULOSIS PATIENTS? dient high-pressure liquid chromatography analysis or gas or liquid chromatography-mass spectrometry, which is not available in most The pipeline for DM detection technologies is relatively robust, clinical settings.51 For this reason, non-invasive tools using skin although product development priorities have been shaped by autofl uoresence quantifi cation that correlate with levels of tissue 12 the traditional markets in high-income settings. This has meant AGEs have been developed that take advantage of the characteris- that many promising candidate technologies are not fully opti- tic fl uorescence of AGEs in the skin.49 Due to the fact that the gly- mized for use in LMIC settings such as those in Table 1. An exten- cated products (the major contribution in fl uorescence comes from sive review of the literature has shown no previous studies that fl uorescent AGEs linked mostly to collagen, but also to other pro- have examined the attributes of technologies currently in the teins and lipids) remain present in the dermal layer for a long pe- pipeline for applicability in TB programs, given the particular riod of time, measurements from AGE readers represent averages of characteristics required. Among the promising novel DM screen- hyperglycemia over a long period of time (average 12 months);51 ing technologies currently under investigation and development, they therefore present an opportunity for use in TB patients, given some are in routine use, but none has been widely evaluated for the need to distinguish medium- and long-term glycemic control use in TB patient populations. Table 3 provides a summary of novel from short-term, infection-induced hyperglycemia. DM screening technologies identifi ed, and indicates whether they Two non-invasive AGE devices are currently at an advanced meet the proposed TPP requirements described earlier.35,47–49 A stage of development, by the companies DiagnOptics Technologies summary of each category of test is also provided below. BV (AGE Reader™; Groningen, The Netherlands) and Veralight Point-of-care compatible A1c readers (SCOUT DS®; Miraculins Inc., Winnipeg, MB, Canada).12 While and low-cost test disposables the capital cost of their devices is currently quite high, they are A1c readers are one of a class of DM screening devices that deter- simple to use, and there are no consumables and little to no main- mine glucose in protein-bound form. Glucose not only exists in tenance, although electricity is required. They do not require pa- free forms, it also persists in protein-bound forms in blood and tient fasting or preparation and are non-invasive, and both pro- other tissues (e.g., glycated hemoglobin and glycated albumin in vide results in a few short minutes after the patient places the 12 blood, and advanced glycation end products in skin). After the underside of the forearm on the device. Although more studies ® initial formation of a Schiff base by glucose and protein mole- are required, one study demonstrated the SCOUT DS to have cules, an irreversible reaction forms a stable ketoamine. As the re- 74.7% sensitivity, and the same study calculated the sensitivity action is irreversible, in vivo protein half-life determines the aver- differential to mean that the device would detect 28.8% more in- aging time of glycemic control. Although A1c readers are now dividuals in the OGTT-defi ned positive screening class than FPG 49 POC-compatible, accurate and provide blood glucose levels aver- testing and 17.1% more than A1c testing. Despite these promis- aged over a 3-month period, their suitability for use in TB programs ing results, more information is needed regarding the performance is limited by the fact that they are invasive (requiring a fi nger- of AGE readers in the 22 high TB burden countries, given poten- prick sample) and require expensive disposables that require re- tial variations in performance due to genetic and socio-cultural frigeration. Low-cost A1c test disposables are required.12 variations, skin color, skin dryness, or the use of topical creams and substances. Overall, in addition to their non-invasiveness and Autofluorescence-based readers operational advantages, AGE readers offer the potential for very This category of device measures skin fl uorescence due to the low-cost screening. They do not require disposables, are easy to accumulation of DM-related advanced glycation end products operate by minimally trained users, and have relatively low oper- (AGEs) in the dermal collagen, and biomarkers of metabolism and ating costs. The initial instrument costs could be amortized very oxidative stress to determine a subject’s risk of having undiag- quickly in settings with high-patient throughput, and with suffi - nosed pre-diabetes and DM.50 AGEs are thought to play a central ciently high volumes per test, costs might ultimately be much role in the pathogenesis of DM complications, and are the result lower than A1c per-test costs.

TABLE 3 Novel DM screening methods with potential utility in TB populations35,47–49

Glycated albumin Sudomotor 1,5-anhydroglucitol and fructosamine function readers POC A1c readers AGE readers Distinguishes transient 1–3 months GA = 1 month Yes 3 months 12 months hyperglycemia F = mixed Low cost No Yes Low per test cost, Low device cost, Low per test cost, high device cost high per test cost high device cost No fasting Yes Yes Yes Yes Yes <5 min results No Yes Yes Yes Yes High sensitivity Yes 64.1%47 75%48 78–81%35 74.7%49 Simple use No No Yes Yes Yes Non-invasive No No Yes No Yes No calibration No No Yes No Yes No minimum operating YesNoNoNoNo temperature POC No Yes Yes Yes Yes Other Early in development Early in development Early in development Early in development pipeline pipeline pipeline pipeline

DM = diabetes mellitus; TB = tuberculosis; POC = point-of-care; A1c = blood glycated hemoglobin; AGE = advanced glycation end product; GA = glycated albumin; F = fructosamine. Public Health Action DM screening for TB patients S15

Sudomotor function devices landscape of DM screening technologies and an analysis of the The EZSCAN Sudomotor reader (Impeto Medical, Paris, France) is opportunities and challenges associated with each. It is important an alternative DM screening device based on electrochemical po- to note, however, that while better screening tools can improve tential measurements across hand and foot skin surfaces to deter- and increase the diagnosis of DM among TB patients, TB patients mine emergent neuropathy.52 Small fi ber neuropathies are com- must also have access to treatment, monitoring and ongoing care mon in people with insulin resistance and prediabetes.48 EZSCAN for DM and support to successfully complete their TB treatment. is a dynamic test equivalent to a stress test, which measures the References capacity of the sweat glands to release chloride ions in response to electrochemical activation.48 A low voltage of variable ampli- 1 Jeon C Y, Murray M B. Diabetes mellitus increases the risk of active tubercu- losis: a systematic review of 13 observational studies. PLOS Med 2008; 5: tude is applied to electrodes on the skin in regions with a high e152. density of sweat glands (palms, feet), and the electrical potential 2 Baker M A, Harries A D, Jeon C Y, et al. Systematic review: the impact of dia- difference caused by the electrochemical reaction on these elec- betes on tuberculosis treatment outcomes. BMC Med 2011; 9: 81. trodes is measured.48 The information is then used to determine 3 International Diabetes Federation. The global burden. IDF diabetes atlas. 5th ed. Unwin N, Whiting D, Guariguata L, et al., eds. Bussels, Belgium: Inter- 53 the patient’s cardiometabolic risk. The test does not require pa- national Diabetes Federation, 2012. http://www.idf.org/diabetesatlas/5e/the- tient preparation, such as fasting, and does not require blood global-burden Accessed August 2013. drawing. Results are available within 2 min, and clinical studies 4 World Health Organization/International Union Against Tuberculosis and Lung Disease. Provisional collaborative framework for care and control of have shown their sensitivity and specifi city to be as high as 75% tuberculosis and diabetes. WHO/HTM/TB/2011.15. Geneva, Switzerland: and 100%, respectively.48,53 More studies are required to assess the WHO, 2011. http://whqlibdoc.who.int/publications/2011/9789241502252_ utility of the test for TB patients. eng.pdf Accessed August 2013. 5 Dooley K E, Chaisson R E. Tuberculosis and diabetes mellitus: convergence Other promising biomarkers of two epidemics. Lancet Infect Dis 2009; 9: 737–746. 6 Restrepo B I. Convergence of the tuberculosis and diabetes epidemics: re- The ability to distinguish between short-, medium- and longer- newal of old acquaintances. Clin Infect Dis 2007; 45: 436– 438. term glycemic control with a single measurement at a single point 7 Li L, Lin Y, Mi F, et al. Screening of patients with tuberculosis for diabetes in time is a key priority for any DM screening device intended for mellitus in China. Trop Med Int Health 2012 July 25. Epub ahead of print. 8 India Tuberculosis–Diabetes Study Group. Screening of patients with tubercu- use in TB populations. To this end, a possible solution may be the losis for diabetes mellitus in India. Trop Med Int Health 2013; 18: 636–645. development of a multivalent platform that contains a set of ana- 9 Balakrishnan S, Vijayan S, Nair S et al. High diabetes prevalence among tu- lytical targets from glycated products selected to demonstrate a berculosis cases in Kerala, India. PLOS ONE 2012; 7: e46502. 10 Ottmani S-E, Murray M B, Jeon C Y, et al. Consultation meeting on tubercu- range of half-lives, allowing the differentiation between short-, losis and diabetes mellitus: meeting summary and recommendations. Int J medium- and long-term glycemic control. Promising biomarkers Tuberc Lung Dis 2010; 14: 1513–1517. are currently being evaluated, for example, that identify physio- 11 Basoglu O K, Bacakoglu F, Cok G, Sayiner A, Ates M. The oral glucose toler- logical changes due to elevated glucose levels early in pregnancy, ance test in patients with respiratory infections. Monaldi Arch Chest Dis 1999; 54: 307–310. 12 which result in transitory hyperglycemia. The main biomarker 12 Weigl B, Drake J K. Developing an adaptable set of point-of-care diabetes being investigated in this regard is glycated albumin (GA), which screening technologies for low-resource settings. Point Care 2012; 12: 33– 40. remains in circulation approximately one third as long as A1c; a 13 Whiting D R, Hayes L, Unwin N C. Diabetes in Africa. Challenges to health care for diabetes in Africa. J Cardiovasc Risk 2003; 10: 103–110. GA reading will thus represent the effects of elevated glucose over 14 Barter D M, Agboola S, Murray M B, Barnighausen T. Tuberculosis and pov- an average of 1 month. 54 erty: the contribution of patient costs in sub-Saharan Africa—a systematic GA is currently being investigated by PATH in relation to a ges- review. BMC Public Health 2012; 12: 980. tational DM project, in addition to fructosamine and 1,5-anhydro- 15 Kapur A, Harries A D. The double burden of diabetes and tuberculosis—public health implications. Diabetes Res Clin Pract 2013, Jan 7. E-pub ahead of print. 12 glucitol (1,5-AG). Fructosamine measures a mixture of glycated 16 Jeon C Y, Murray M B, Baker M A. Managing tuberculosis in patients with proteins that exhibit both longer and shorter half-lives than GA, diabetes mellitus: why we care and what we know. Expert Rev Anti Infect and 1.5-AG has a very short half-life, averaging glycemic control Ther 2012; 10: 863–868. 55 17 World Health Organization. Defi nition, diagnosis and classifi cation of dia- for a 1–3 month period. Other markers that have thus far mainly betes mellitus and its complications. Report of a WHO consultation. Part 1: been investigated in relation to gestational DM, but may have diagnosis and classifi cation of diabetes mellitus. WHO/NCD/NCS/99.2. Ge- utility for DM screening in TB patients, include cytokines, chemo- neva, Switzerland: WHO, 1999. http://whqlibdoc.who.int/hq/1999/WHO_ kines, hormones and transcription factors stimulated by the AGE/ NCD_NCS_99.2.pdf Accessed August 2013. 18 World Health Organization. Defi nition and diagnosis of diabetes mellitus 56 receptor for advanced glycation end-product signaling pathway. and intermediate hyperglycemia: report of a WHO/IDF consultation. Ge- The benefi ts of a multivalent diagnostic platform are many, neva, Switzerland: WHO, 2006. http://www.idf.org/webdata/docs/WHO_ and additional parameters of relevance to TB patients, such as the IDF_defi nition_diagnosis_of_diabetes.pdf Accessed August 2013. human immunodefi ciency virus, could be tested on such a screen- 19 World Health Organization. Screening for type 2 diabetes: report of a World Health Organization and International Diabetes Federation meeting. WHO/ ing platform. NMC/MNH/03.1. Geneva, Switzerland: WHO, 2003. http://www.who.int/ diabetes/publications/en/screening_mnc03.pdf Accessed August 2013. 20 World Health Organization/Bangladesh Institute of Research and Rehabilita- tion in Diabetes, Endocrine and Metabolic Disorders. Guidelines for care of CONCLUSION type 2 diabetes mellitus in Bangladesh. Dhaka, Bangladesh: BIRDEM, 2003. http://www.slideshare.net/roger961/guidelines-for-care-of-type-2-diabetes- With an estimated 8.7 million new cases of TB in 2011, and the mellitus-in-bangladesh Accessed August 2013. majority of these in LMICs,30 the ability to implement the WHO’s 21 Ministry of Health, Brazil. [Plan for reorganization of care for hypertension recommendation that all TB patients be screened for DM will re- and diabetes mellitus]. Brasilia, Brazil: MOH, 2001. http://bvsms.saude.gov. br/bvs/publicacoes/miolo2002.pdf Accessed August 2013. [Portuguese] quire the identifi cation of technologies that are appropriate for 22 Chinese Diabetes Society. [Clinical practice recommendations for the Chi- TB patients and current TB program structures. Research and ad- nese]. Beijing, China: CDS, 2010. http://cdschina.org/news_show.jsp?id=402. vocacy are necessary to ensure that evidence is generated regarding html Accessed August 2013. [Chinese] the potential for the DM technologies currently in the pipeline to 23 Food, Medicines and Healthcare Administration and Control Authority. Minimum standards for health centres. Addis Ababa, Ethiopia: Ethiopian be used in TB populations, given current program structures. This Federal Ministry of Health, 2010. article provides a fi rst step towards understanding the current 24 National Programme for Prevention and Control of Cancer, Diabetes, Public Health Action DM screening for TB patients S16

Cardiovascular Diseases and Stroke, Directorate General of Health Services, glucose tolerance tests for diagnosis of gestational mellitus. J Med Assoc Thai Ministry of Health and Family Welfare. Operational guidelines. New Delhi, 2011; 94: 540Y544. India: Government of India, 2008–2009. http://health.bih.nic.in/Docs/Guide 39 Bennett C M, Guo M, Dharmage S C. HbA(1c) as a screening tool for detec- lines-NPCDCS.pdf Accessed August 2013. tion of type 2 diabetes: a systematic review. Diabet Med 2007; 24: 333–343. 25 Ministry of Health, Kenya. Clinical guidelines for diagnosis and treatment of 40 Chauhan L S, Agarwal S P. Tuberculosis control in India. Chapter 3. Revised common conditions in Kenya. Nairobi, Kenya: MoH, 2002. http://apps.who. National Tuberculosis Control Program. Directorate General of Health Ser- int/medicinedocs/documents/s16427e/s16427e.pdf Accessed August 2013. vices, Ministry of Health and Family Welfare. New Delhi, India: Government 26 Society for Endocrinology, Metabolism and Diabetes of South Africa. The of India, 2005. http://tbcindia.nic.in/pdfs/Tuberculosis%20Control%20in 2012 SEMDSA guideline for the management of type 2 diabetes mellitus: %20India-Final.pdf Accessed August 2013. summary. Sandton, South Africa: SEMDSA, 2012. http://www.semdsa.org.za/ 41 Frenk J. Bridging the divide: global lessons from evidence-based health pol- images/guideline_2013_new.pdf Accessed August 2013. icy in Mexico. Lancet 2006; 368: 954–961. 27 Ministry of Health, Republic of Uganda. Uganda clinical guidelines: na- 42 Kwesigabo G, Mwangu M A, Kakoko D C, et al. Tanzania’s health system and tional guidelines on management of common conditions. Kampala, Re- workforce crisis. J Pub Health Pol 2012; 33: S35–S44. public of Uganda: Government of Uganda, 2010. http://health.go.ug/docs/ 43 Denkinger C, Nicolau I, Ramsay A, Chedore P, Pai M. Are peripheral micros- ucg_2010.pdf Accessed August 2013. copy centres ready for next generation molecular TB diagnostics? Eur Respir J 28 International Diabetes Federation. Clinical guidelines taskforce: global 2013; 42: 544–547. guidelines for type 2 diabetes. Brussels, Belgium: International Diabetes Fed- 44 Weigl B, Gaydos C A, Kost G, et al. The value of clinical needs assessments eration, 2005. http://www.idf.org/webdata/docs/IDF%20GGT2D.pdf Accessed for point-of-care diagnostics. Point Care 2012; 11: 108–113. August 2013. 45 Pai N P, Vadnais C, Denkinger C, Engel N, Pai M. Point-of-care testing for in- fectious diseases: diversity, complexity, and barriers in low- and middle- 29 Noble D, Mathur R, Dent T, Meads C, Greenhalgh T. Risk models and scores income countries. PLOS Med 2012; 9: e1001306. for type 2 diabetes: systematic review. BMJ 2011; 343: d7163. 46 Lee W-C. Selecting diagnostic tests for ruling out or ruling in disease: the use 30 World Health Organization. Global tuberculosis report, 2012. WHO/HTM/ of the Kullback-Leibler distance. Int J Epidemiol 1999; 28: 521–525. TB/2012.6. Geneva, Switzerland; WHO, 2012. http://www.who.int/tb/ 47 Dominiczak M H, Macrury S M, Orrell J M, et al. Long-term performance of publications/global_report/en/ Accessed August 2013. the fructosamine assay. Ann Clin Biochem 1988; 25: 627–633. 31 World Health Organization. Use of glycated haemoglobin (HbA1c) in the di- 48 Tavee J, Zhou L. Small fi ber neuropathy: a burning problem. Clev Clin J Med agnosis of diabetes mellitus. Abbreviated report of a WHO consultation. 2009; 76: 297–305. WHO/NMH/CHP/CPM/11.1. Geneva, Switzerland: WHO, 2011. http://www. 49 Maynard J D, Rohrscheib M, Way J F, Nuyen C M, Ediger M N. Noninvasive who.int/diabetes/publications/diagnosis_diabetes2011/en/ Accessed August type 2 diabetes screening: superior sensitivity to fasting plasma glucose and 2013. A1c. Diabet Care 2007; 30: 1120–1124. 32 American Diabetes Association. Diagnosis and classifi cation of diabetes mel- 50 Bucala R, Cerami A. Advanced glycosylation: chemistry, biology, and impli- litus. Diabet Care 2011; 33 (Suppl): S62–S69. cations for diabetes and aging. Adv Pharmacol 1992; 23: 1–34. 33 Guler M, Unsal E, Dursun B, Aydln O, Capan N. Factors infl uencing sputum 51 Ahmed N, Thornalley P J. Quantitative screening of protein biomarkers of smear and culture conversion time among patients with new case pulmo- early glycation, advanced glycation, oxidation and nitrosation intracellular nary tuberculosis. Int J Clin Pract 2007; 61: 231–235. and extracellular proteins by tandem mass spectrometry multiple reaction 34 D’Orazio P, Burnett R W, Fogh-Andersen N, et al. Approved IFCC recommen- monitoring. Biochem Soc Trans 2003; 31: 1417–1422. dation on reporting results for blood glucose (abbreviated). Clin Chem 2005; 52 Schwarz P E H, Brunswick P, Calvet J H. EZSCAN, a new technology to detect 51: 1573–1576. diabetes risk. Br J Diabetes Vasc Dis 2011; 11: 204–209. 35 Perry C R, Shankar R R, Fineberg N, McGill J, Baron A D. HbA1c measure- 53 Mayaudon H, Miloche P O, Bauduceau B. A new method for assessing sudo- ment improves the detection of type 2 diabetes in high-risk individuals with motor function: relevance in type 2 diabetes. Diabetes Metab 2010; 36: 450– nondiagnostic levels of fasting plasma glucose. Diabet Care 2001; 24: 465– 454. 471. 54 Koga M, Kasayama S. Clinical impact of glycated albumin as another glyce- 36 Gallagher E J, Bloomgarden Z T, Le Roith D. Review of hemoglobin A1c in mic control marker. Endocr J 2010; 57: 751–762. the management of diabetes. J Diabet 2009; 1: 9–17. 55 Behan K J, Merschen J. HbA1c does not always estimate average glucose. 37 Roberts W L, De B K, Brown D, et al. Effects of hemoglobin C and S traits on Clin Lab Sci 2011; 24: 71–77. eight glycohemoglobin methods. Clin Chem 2002; 48: 383–385. 56 Bierhaus A, Humpert P M, Morcos M, et al. Understanding RAGE, the receptor 38 Hansarikit J, Manotaya S. Sensitivity and specifi city of modifi ed 100-g oral for advanced glycation end products. J Mol Med (Berl) 2005; 83: 876–886.

L’épidémie de diabète sucré (DM) de type 2 dans les pays à revenus sons les technologies existantes et nouvelles pour le dépistage du DM faibles et moyens (LMIC) peut constituer une menace pour les progrès qui pourraient être les plus applicables aux patients TB dans les LMIC. de la lutte contre la tuberculose (TB), car le DM est à la fois un facteur De telles méthodes devraient être rapides, elles devraient ne pas exiger majeur de risque et de développement d’une TB active et peut aussi le jeûne et elles devraient permettre aux pourvoyeurs de soins de di- entrainer des résultats défavorables du traitement de la TB. En dépit stinguer entre des hyperglycémies transitoires et de plus longue durée de la directive de l’Organisation mondiale de la Santé selon laquelle au moyen d’outils peu coûteux, n’exigeant que peu de formation et tous les patients TB devraient faire l’objet d’un dépistage pour le DM, aucune infrastructure spécialisée. Différentes méthodes sont actuelle- la plupart des services des LMIC qui traitent les patients TB ne ré- ment en cours de développement, tels que les tests sur l’hémoglobine alisent pas actuellement le dépistage du DM, en partie en raison du glycosylée aux lieux de soins et sur l’albumine glycosylée, les lecteurs coût et de la complexité qu’il implique. Le dépistage du DM est par des produits finaux d’une glycation avancée non-invasive et les outils ailleurs compliqué par l’existence d’une hyperglycémie transitoire de dépistage basés sur la fonction sudomotrice ; elles offrent des car- chez beaucoup de patients TB ainsi que par les différences de facteurs actéristiques intéressantes de performance et méritent une évaluation de risque de DM (par exemple l’indice de masse corporelle) entre dans les populations TB. les patients TB et la population générale. Dans cet article, nous révi-

La diabetes (DM) de tipo 2 presenta características epidémicas en los de los establecimientos que atienden a estos pacientes en los LMIC países con ingresos bajos e intermedios (LMIC) y puede poner en no cumplen con esta práctica, en parte debido a los costos y a la peligro los avances alcanzados en materia de control de la tuberculo- complejidad de la misma. La detección de la DM se complica además sis (TB); la DM constituye un factor de riesgo importante de padecer por la hiperglucemia transitoria que suele observarse en muchos paci- la enfermedad TB activa y también puede tener consecuencias desfa- entes TB y por las diferencias en los factores de riesgo de DM, pre- vorables en el desenlace del tratamiento antituberculoso. Pese a la re- sentes en los pacientes con TB y la población general, por ejemplo comendación de la Organización Mundial de la Salud de realizar la el índice de masa corporal. En el presente artículo se analizan las técni- detección sistemática de la DM en todos los pacientes TB, la mayoría cas existentes y los nuevos métodos de detección sistemática de la Public Health Action DM screening for TB patients S17

DM que pueden ser más adaptados a los pacientes con TB de los pruebas de hemoglobina glucosilada y albúmina glucosilada realizadas LMIC. Estos métodos deben ser rápidos, no deben precisar el estado en el punto de atención, los lectores no invasivos de productos finales de ayuno y deben permitir al profesional de salud diferenciar entre de la glucosilación avanzada y los dispositivos de detección basados la hiperglucemia transitoria y la hiperglucemia de largo plazo, medi- en la función sudomotora; estos métodos ofrecen características de ante la utilización instrumentos de bajo costo, que exijan poco en- rendimiento interesantes y merecen una evaluación en las poblaciones trenamiento y no necesiten infraestructuras especializadas. En la actuali- de pacientes con diagnóstico de TB. dad, se encuentran en curso de desarrollo varios métodos como las

Public Health Action (PHA) The voice for operational research. e-ISSN 2220-8372 Published by The Union (www.theunion.org), PHA provides a platform to Editor-in-Chief: Donald A Enarson, MD, Canada fulfi l its mission, ‘Health solutions for the poor’. PHA publishes high-quality Contact: [email protected] scientifi c research that provides new knowledge to improve the accessibility, PHA website: http://www.theunion.org/index.php/en/journals/pha equity, quality and effi ciency of health systems and services. Article submission: http://mc.manuscriptcentral.com/pha InternaƟ onal Union Against Tuberculosis and Lung Disease Public Health Action Health soluƟ ons for the poor VOL 3 SUPPLEMENT 1 PUBLISHED 4 NOVEMBER 2013

Tuberculosis-diabetes mellitus bidirectional screening at a tertiary care centre, South India B. C. Prakash,1,2 K. S. Ravish,2 B. Prabhakar,1,2 T. S. Ranganath,2 B. Naik,3 S. Satyanarayana,4 P. Isaakidis,5 A. M. V. Kumar4 http://dx.doi.org/10.5588/pha.13.0032 25% to as high as 44%.8–10 A pilot project was initiated AFFILIATIONS Setting: Tuberculosis (TB) and diabetes mellitus (DM) 1 Bowring and Lady Curzon in India in 2012 to assess the feasibility of bidirec- Hospital, Bangalore, India clinics at Bowring and Lady Curzon Hospital, a tertiary tional screening. Bowring and Lady Curzon Hospital, 2 Bangalore Medical College care centre in Bangalore, India. Bangalore, the teaching hospital of Bangalore Medical & Research Institute, Obejctive: To assess the feasibility and results of TB-DM Bangalore, India College and Research Institute in the South Indian 3 World Health Organization bidirectional screening. State of Karnataka, was one of eight tertiary health fa- Country Office in India, Methods: A descriptive study conducted from 1 March New Delhi, India cilities that participated in this pilot project. We report 4 International Union to 30 September 2012, in which all TB patients were as- on the feasibility and results of bidirectional screening Against Tuberculosis and sessed for DM and vice versa. Fasting blood glucose val- Lung Disease, South-East of TB and DM at this centre. Asia Office, New Delhi, ues of ⩾126 mg/dl and 110–125 mg/dl were considered India as DM and pre-diabetes, respectively. 5 Médecins Sans Frontières, METHODS Operational Research Unit, Results: Of 510 TB patients, 32 (6.3%) had been previ- Luxembourg ously diagnosed with DM. Screening among the remain- Design CORRESPONDENCE ing 478 patients yielded 15 (2.9%) with pre-diabetes and This is a descriptive study. K S Ravish 15 (2.9%) newly diagnosed cases of DM. A higher preva- Assistant Professor ⩾ Department of Community lence of DM was found among patients aged 40 years, Setting Medicine patients with pulmonary TB and smokers. Of the 47 TB- Bangalore Medical College Bowring and Lady Curzon hospital is a 570-bed multi- & Research Institute DM patients, 45 were enrolled in DM care. Of 1670 DM speciality teaching hospital in Bangalore. The out- Fort, K R Road patients followed up in DM clinics, 45 already had TB. Bangalore 560002, India patient department caters for an average of 900 patients Tel: (+91) 98450 47773 Among the remaining 1625 patients screened, 152 (9%) per day. The hospital has a separate TB clinic and a Fax: (+91) 80267 04342 had symptoms suggestive of TB; two of these were found e-mail: ravish7474@gmail. DM clinic. TB patients diagnosed at different clinics com to have the disease. and hospital departments are referred to the TB clinic Conclusion: Bidirectional screening for DM and TB is fea- KEY WORDS for treatment per the Revised National TB Control Pro- TB; DM; bidirectional sible and produces a high yield for DM among TB patients. gramme (RNTCP) guidelines.11 An average of 1000 TB screening; India; feasibility The yield of TB among DM patients was low and needs patients are treated in the TB clinic each year. The DM future research using new, improved TB diagnostic tools. clinic is open 1 day a week and is visited by an average of ∼100 DM patients per week.

ith an annual tuberculosis (TB) incidence of Study population W2.2 million cases (range 2.0–2.5 million) and an All consecutively diagnosed TB patients aged ⩾15 years estimated 63 million people living with diabetes melli- who attended the hospital’s TB clinic from 1 March to tus (DM), India has the highest TB burden and second 30 September 2012 were screened for DM. At the DM highest DM burden in the world.1–3 Nearly half of DM clinic, all DM patients aged ⩾15 years were screened patients do not know their status, and a further 77 mil- for TB over the same period. lion people are estimated to have impaired glucose tol- erance and are at higher risk of becoming diabetic.2–4 Screening procedures DM is known to increase the risk of active TB ap- The screening procedures and the recording and re- proximately three fold, and contributes to adverse TB porting tools used were as per the protocols developed treatment outcomes such as death, treatment failure for the pilot project. The details of these procedures and relapse.5–6 The World Health Organization (WHO) have been described in detail elsewhere.12–13 Briefl y, all and the International Union Against Tuberculosis and of the patients at the TB clinic were asked about their Lung Disease (The Union) have launched a new ‘Col- history of DM. Patients not aware of their status were laborative framework for the care and control of dia- offered random blood glucose (RBG) testing. Venous betes and tuberculosis’, with one of the main activities blood samples were collected from the patients and being the routine implementation of bidirectional blood glucose was tested by the hexokinase method. If screening of the two diseases.7 Bidirectional screening RBG was ⩾110 mg/dl, the patients were offered fasting aims at early detection of TB among DM patients and blood glucose (FBG) testing. The cut-offs used in the < Received 22 May 2013 vice versa, and assists with an integrated approach to study were as follows: FBG 110 mg/dl was consid- Accepted 3 July 2013 the management of the co-morbidity. ered normal; FBG 110–125 mg/dl was diagnosed as pre- Available data from various sites in India show that diabetes; and FBG ⩾ 126 mg/dl was diagnosed as DM.14 PHA 2013; 3(S1): S18–S22 the prevalence of DM among TB patients ranges from Patients with known DM and newly diagnosed DM were © 2013 The Union Public Health Action Bidirectional TB-DM screening S19 referred to the diabetes clinic for further management. TABLE 1 Screening TB patients for DM at Bowring and ACKNOWLEDGEMENTS The authors acknowledge Testing for human immunodefi ciency virus (HIV) in- Lady Curzon Hospital, Bangalore, India, March– their Dean/Director, Banga- fection was offered routinely to TB patients; informa- September 2012 lore Medical College and tion on HIV status was extracted from the TB registers. Research Institute, and the Indicator n (%) Medical Superintendent, At the DM clinic, patients were asked whether they Bowring and Lady Curzon were under TB treatment. Those who were not were Patients with TB registered over the study period 510 hospital for their support. A workshop was convened in asked about specifi c symptoms and signs indicative of Patients with a known diagnosis of DM 32 (6.3) Delhi, India, for the purpose TB (cough for ⩾2 weeks, prolonged fever, weight loss, Patients needing to be screened with RBG 478 of writing the papers that are published in this supple- loss of appetite, enlarged glands). Patients who gave Patients screened with RBG 478 (100) ment. The workshop was run an affi rmative answer for any of these symptoms were ⩾ by the Centre for Opera- Patients with RBG 110 mg/dl 110 tional Research, International referred to the TB clinic for further evaluation. At the Patients screened with FBG 109 (99.1) Union Against Tuberculosis TB clinic, these patients underwent investigations in- Patients with FBG <110 79 and Lung Disease (The Union), Paris, France; The cluding sputum smear microscopy, chest radiography Patients with FBG 110–125 15 (2.9) ⩾ Union South-East Asia Office, and other investigations per RNTCP guidelines. Type Patients with FBG 126 mg/dl New Delhi, India; the Opera- (newly diagnosed with DM) 15 (2.9) tional Research Unit, of TB was classifi ed per the national guidelines, which Patients with known or newly diagnosed DM 47 (9.2) Médecins Sans Frontières, are in line with WHO recommendations.11 Luxembourg; the World Patients with known and newly diagnosed DM Health Organization Country referred to DM care 47 (100) Office in India, New Delhi, Data collection and validation India; the All India Institute Patients with known or newly diagnosed DM At the TB clinic, an additional TB-DM register was used of Medical Sciences, New reaching DM care 45 (95.7) Delhi, India; and ESIC to record data on the following variables: age, sex, cur- Medical College, Bangalore, rent smoker (defi ned as a person who had smoked to- TB = tuberculosis; DM = diabetes mellitus; RBG = random blood India. glucose; FBG = fasting blood glucose. Funding for the workshop bacco at least once during the last 3 months), residence and open access publication (urban or rural), HIV status, RBG and FBG levels, dates was received from the World Diabetes Foundation, of TB diagnosis, RBG and FBG tests and referral and DM. Screening among the remaining 478 patients Gentofte, Denmark. enrolment into DM care. At the DM clinic, a separate yielded 15 (2.9%) and 15 (2.9%) pre-diabetes and new Conflict of interest: none declared. treatment card for each patient was used to record DM cases, respectively. Of the 47 known/newly diag- data about the patient’s DM history and current DM nosed patients, 45 (96%) were registered at the DM status, screening for TB symptoms, the result of screen- clinic for further management. Almost all patients un- ing and the result of the investigations. The existing derwent RBG testing within one day of the date of TB staff of the TB and DM clinics in the routine health diagnosis; >95% of the patients eligible for FBG un- care setting were trained and involved in the study; no derwent the test within 2 days of the RBG test. additional manpower or infrastructure was used. Su- The demographic and clinical characteristics of TB pervision and site visits were undertaken by staff from patients associated with DM prevalence are shown in The Union and the RNTCP during the study, and all Table 2. DM prevalence was signifi cantly higher among the records were checked for completeness, consis- tency and accuracy. TABLE 2 Characteristics of TB patients screened for Data analysis and statistics DM at Bowring and Lady Curzon Hospital, Bangalore, India, March–September 2012 The individual patient data were double-entered into EpiData software, version 3.1 (EpiData Association, Total Patients Odense, Denmark), validated and analysed. The data TB patients with DM were summarised using frequencies and proportions. Characteristic n n (%) P value Differences between groups were compared using the Total 510 47 (9.2) χ² test. A P value of <0.05 was considered statistically Age, years signifi cant. The number needed to test (NNT, defi ned <40 319 12 (3.8) <0.001 ⩾ as reciprocal proportion of the sum of new DM and pre- 40 191 35 (18.3) diabetes cases) was calculated, disaggregated by demo- Sex Male 316 32 (10.1) 0.36 graphic and clinical characteristics. Female 194 15 (7.7) Ethics approval Residence Urban 393 38 (9.7) 0.52 We obtained administrative approval from the head of Rural 117 9 (7.7) the Bangalore Medical College and Research Institute, Smoking status* Bangalore. The entire protocol was reviewed and ap- Smoker 137 23 (16.8) <0.001 proved by the Ethics Advisory Group of The Union. Non-smoker 372 23 (6.2) HIV status Positive 83 8 (9.6) 0.54 RESULTS Negative 423 38 (9.0) Screening tuberculosis patients Unknown 4 1 (25.0) for diabetes mellitus Type of TB Pulmonary 205 30 (14.6) <0.001 The results of DM screening among TB patients are sum- Extra-pulmonary 305 17 (5.6) marised in Table 1. The median (interquartile range) age * Not recorded for one TB patient. of the TB patients was 35 (25–45) years. Of 510 TB pa- TB = tuberculosis; DM = diabetes mellitus; HIV = human immuno- tients, 32 (6.3%) had been previously diagnosed with deficiency virus. Public Health Action Bidirectional TB-DM screening S20

TABLE 3 Number needed to screen to find a new case of DM and TB clinic for further evaluation. All those referred reached the TB pre-diabetes among TB patients at Bowring and Lady Curzon clinic and underwent evaluation. Among the 152 symptomatic Hospital, Bangalore, India, March–September 2012 patients, two new smear-positive TB patients were diagnosed and initiated on TB treatment. The NNT to detect one person with A B C ∼ Patients with New Pre-diabetes symptoms suggestive of TB was 10, and for detection of one TB unknown DM cases cases case it was ∼812. DM status diagnosed diagnosed NNT Characteristic n n n (A/B+C) DISCUSSION Total TB patients 478 15 15 16 Age, years The fi ndings of this study provide important insights into the <40 312 5 5 31 programme management of TB-DM co-morbidity, and have sev- ⩾40 166 10 10 8 eral health policy implications. Sex First, we implemented bidirectional screening for TB and DM Male 294 10 11 14 using existing resources and staff, thus indicating that this is fea- Female 184 5 4 20 sible. Almost all TB patients underwent screening for DM and vice Residence versa. One of the primary reasons for the low loss to follow-up Urban 367 12 12 15 Rural 111 3 3 18 was the close proximity of the TB and DM clinics. The process of Smoking status* screening yielded additional new cases of DM and TB, although Smoker 122 8 5 9 the NNT to detect one new TB case among DM patients was high Non-smoker 355 6 10 22 compared with the NNT to detect DM among TB patients. HIV status Second, about 10% of TB patients attending the TB clinics had Positive 77 2 1 25 DM. This is lower than the prevalence reported from other sites in Negative 397 12 14 15 South India, where it ranged from 25% to 44%.8–10 It should be Unknown 4 1 0 40 noted that the median age of our study population was lower Type of TB than that in the other studies, and it is well known that the prev- Pulmonary 184 9 8 11 Extra-pulmonary 294 6 7 23 alence of diabetes increases with age. The tests and criteria used for the diagnosis of DM in our study (RBG followed by FBG) dif-

*Not recorded for one TB patient. fered from those used in other studies.8–9 The sensitivity of FBG is DM = diabetes mellitus; TB = tuberculosis; NNT = number needed to test; HIV = h uman immunodeficiency virus. lower than that of the 75 g oral glucose tolerance test, and this may have resulted in underestimation of the true proportion of DM in our population. Furthermore, the screening yielded 15 pa- TB patients aged ⩾40 years, among smokers, and in patients with tients with pre-diabetes who are at higher risk of developing type pulmonary TB (PTB). 2 DM in the future. These patients could be targeted for counsel- The NNT to fi nd a new case of DM and/or pre-diabetes is sum- ling and other preventive services. Another reason is probably the marised in Table 3. The overall NNT in this cohort of TB patients higher proportion of extra-pulmonary TB patients observed in our was 16. However, the NNT was lower among TB patients aged study, who are known to have a lower prevalence of DM.15 ⩾40 years, smokers, and individuals with PTB. Third, nearly two thirds of all identifi ed DM patients knew their status prior to screening. This could be because of the nature Screening diabetes mellitus patients of the study setting, i.e., a tertiary care hospital where patients have for tuberculosis the opportunity to periodically undergo various investigations, The results of TB screening among DM patients are shown in Ta- including those for DM. Given the ease and speed of performing ble 4. Overall, 1670 DM patients were seen at least once during DM tests, the results are usually available earlier than those for TB this period. Of these, 45 were already known TB patients: 17 new tests. This might also have contributed to the higher proportion smear-positive PTB, 9 new smear-negative PTB, 16 new extra- of previously known DM. Nevertheless, the early identifi cation of pulmonary TB, 1 new-others and 2 treatment after default. Of the patients with co-morbidity, especially among the newly diagnosed remaining 1625 patients screened for TB, 152 (9.3%) were found cases, helped us to link these patients to appropriate DM care, to have symptoms suggestive of TB, and they were referred to the which could lead to improved TB treatment outcomes. Fourth, nearly 10% of the DM patients screened (at least once during the study period) in the DM clinics had symptoms sugges- TABLE 4 Screening of DM patients for TB at Bowring and Lady tive of TB. This is higher than in the general health facility set- Curzon Hospital, Bangalore, India, March–September 2012 ting, where it is estimated that ∼2–3% of patients have TB symp- Indicator n (%) toms.16–17 However, we found very few TB cases among them. While the exact reasons for the low yield are not clear, it could be Patients seen in the DM clinic during this period 1670 Patients already diagnosed with TB elsewhere 45 due to the fact that DM patients were managed by individual phy- Patients screened at least once for TB symptoms in sicians at diagnosis, and only after glycaemic control was achieved this period 1625 were they referred to DM clinic for periodic visits and collection Of those screened, patients with a positive TB of free drugs. Patients attending DM clinics were therefore more symptom screen 152 (9.3) Patients with a positive TB symptom screen referred for likely to have adequate glycaemic control and could thus have TB investigations 152 had a lower risk of having TB.15 The other reason for the low yield Patients diagnosed with TB after referral for investigations 2 (0.1) could be the lack of availability of state-of-the-art diagnostic tools, Patients identified with TB (known and new) 47 (2.8) such as nucleic acid amplifi cation tests or culture. Future studies Patients with TB receiving treatment 47 should focus on the use of these tools to evaluate whether this DM = diabetes mellitus; TB = tuberculosis. could increase the yield of TB among screened DM patients. Public Health Action Bidirectional TB-DM screening S21

One limitation of our study was that we were not able to ascer- 6 Jeon C Y, Murray M B. Diabetes mellitus increases the risk of active tuberculo- tain whether a high FBG in patients with TB was indicative of true sis: a systematic review of 13 observational studies. PLOS Med 2008; 5: e152. 7 World Health Organization/International Union Against Tuberculosis and DM or of infection-induced hyperglycaemia. This requires periodic Lung Disease. Provisional collaborative framework for care and control of blood glucose testing over the course of TB treatment, which was tuberculosis and diabetes. WHO/HTM/TB/2011.15. Geneva, Switzerland: beyond the scope of the current study. Further research is needed WHO, 2011. 8 Vishwanathan V, Kumpatla S, Aravndalochanan V, et al. Prevalence of diabe- to ascertain this and the optimum timing of DM screening among tes and prediabetes and associated risk factors among tuberculosis patients TB patients. The other related limitation was that we could not in India. PLOS ONE 2012; 7: e41367. use a glycosylated haemoglobin test for the diagnosis of DM, and 9 Balakrishnan S, Vijayan S, Nair S, et al. High diabetes prevalence among this might have led to an under-diagnosis of pre-diabetes and DM. tuberculosis cases in Kerala, India. PLOS ONE 2012; 7: e46502. 10 Gupta S, Shenoy V P, Bairy I, Srinivasa H, Mukhopadhyay C. Diabetes melli- However, as the glycosylated haemoglobin test is expensive, its tus and HIV as co-morbidities in tuberculosis patients of rural South India. use in programme conditions needs further evaluation. J Infect Public Health 2011; 4: 140–144. In conclusion, our study showed that bidirectional screening 11 Central Tuberculosis Division, Directorate General of Health Services, Min- istry of Health & Family Welfare. Technical and operational guidelines for for DM and TB was feasible, with a high yield of DM among TB tuberculosis control, Revised National Tuberculosis Control Programme. New patients. Screening TB patients for DM could be an effi cient tool Delhi, India: Government of India, 2005. http://www.tbcindia.nic.in/pdfs/ for the programme management of TB-DM co-morbidity. The Technical%20&%20Operational%20guidelines%20for%20TB%20Control. yield of TB among DM patients was low and needs future research pdf Accessed August 2013. 12 India Tuberculosis-Diabetes Study Group. Screening of patients with tuber- using new, improved TB diagnostic tools. culosis for diabetes mellitus in India. Trop Med Int Health 2013; 18: 636–645. 13 India Diabetes-Tuberculosis Study Group. Screening of patients with diabe- References tes mellitus for tuberculosis in India. Trop Med Int Health 2013; 18: 646–654. 14 Directorate General of Health Services, Ministry of Health & Family Welfare. 1 World Health Organization. Global tuberculosis report 2012. WHO/HTM/ Operational guidelines. National Programme for Prevention and Control of TB/2012.6. Geneva, Switzerland: WHO, 2012. www.who.int/tb/publications/ Cancer, Diabetes, Cardiovascular Disease and Stroke (NPCDCS). New Delhi, global_report/gtbr12_main.pdf Accessed August 2013. India: Government of India, 2008–2009. http://health.bih.nic.in/Docs/Guide 2 International Diabetes Federation. IDF diabetes atlas. 5th ed. Unwin N, lines/ Guidelines-NPCDCS.pdf Accessed August 2013. Whiting D, Guariguata L, et al., eds. Brussels, Belgium: International Diabe- 15 Leung C C, Lam T H, Chan W M, et al. Diabetic control and risk of tubercu- tes Federation, 2012. http://www.eatlas.idf.org Accessed August 2013. losis: a cohort study. Am J Epidemiol 2008; 167: 1486–1494. 3 Ramachandran A, Ma R C W, Snehalatha C. Diabetes in Asia. Lancet 2010; 16 Directorate General of Health Services, Ministry of Health & Family Welfare. 375: 408– 418. Training module for medical practitioners. Revised National TB Control Pro- 4 Danaei G, Finucane M M, Lu Y, et al. National, regional, and global trends in gramme. New Delhi, India: Government of India, 2010. http://www.tbcindia. fasting plasma glucose and diabetes prevalence since 1980: systematic analysis nic.in/pdfs/Training%20Module%20for%20Medical%20Practitioner.pdf Ac- of health examination surveys and epidemiological studies with 370 country- cessed August 2013. years and 2.7 million participants. Lancet 2011; 378: 31– 40. 17 Santha T, Garg R, Subramani R, et al. Comparison of cough of 2 and 3 weeks 5 Stevenson C R, Critchley J A, Forouhi N G, et al. Diabetes and the risk of tu- to improve detection of smear-positive tuberculosis among out-patients in berculosis: a neglected threat to public health. Chronic Illn 2007; 3: 228–245. India. Int J Tuberc Lung Dis 2005; 9: 61–68.

Contexte : Dispensaires de tuberculose (TB) et de diabète (DM) à (2,9%) et de 15 nouveaux cas de diabète (2,9%). On a trouvé une l’Hôpital Bowring et Lady Curzon, un centre de soins tertiaires à prévalence plus élevée de DM chez les patients de ⩾40 ans, chez les Bangalore, Inde. patients atteints de TB pulmonaire et chez les fumeurs. Sur 47 patients Objectif : Evaluer la faisabilité et les résultats d’un dépistage bidirec- TB-DM, 45 ont été pris en charge pour leur DM. Sur 1670 patients tionnel TB-DM. DM suivis dans les dispensaires du DM, 45 avaient déjà une TB. Parmi Méthodes : Il s’agit d’une étude descriptive menée entre le 1er mars les 1625 restants et dépistés, 152 (9%) souffraient de symptômes et le 30 septembre 2012 pendant laquelle les patients TB ont fait suggestifs de TB, parmi lesquels deux patients souffraient de TB. l’objet d’une évaluation pour DM et vice versa. Un glucose sanguin à Conclusion : Le dépistage bidirectionnel pour DM et TB est réalisable jeun ⩾126 mg/dl et a été considéré comme DM et un glucose sanguin et donne un rendement élevé pour le DM chez les patients TB. Le à jeun de 110–125 mg/dl comme prédiabète. rendement de TB parmi les patients DM est faible et nécessite des Résultats : Sur 510 patients TB, 32 (6,3%) avaient été diagnostiqués recherches ultérieures utilisant des outils améliorés pour le diagnostic antérieurement comme diabétiques. Le dépistage parmi les 478 pa- de la TB. tients TB restants a eu un rendement de 15 sujets prédiabétiques

Public Health Action Bidirectional TB-DM screening S22

Marco de referencia: Los consultorios de tuberculosis (TB) y diabetes (2,9%) y de 15 casos nuevos de DM (2,9%). Se observó una prevalen- (DM) en el Hospital Bowring y Lady Curzon, un centro de atención cia de DM más alta en los pacientes ⩾40 años de edad, los pacientes terciaria de Bangalore en la India. con TB pulmonar y los fumadores. Cuarenta y cinco de los 47 pacientes Objetivo: Se buscó evaluar la factibilidad del cribado bidireccional con TB y DM se inscribieron en el programa de atención de la DM. de la TB y la DM y los resultados de esta intervención. De los 1670 pacientes atendidos en las consultas de DM, en 45 se Métodos: Fue este un estudio descriptivo realizado entre el 1° de había establecido ya el diagnóstico de TB. De los 1625 pacientes re- marzo y el 30 de septiembre del 2012, en el cual se investigó en to- stantes que participaron en el cribado, 152 presentaban signos indic- dos los pacientes TB el diagnóstico de DM y vice versa. Una glucemia ativos de TB (9 %) y en dos de ellos se confirmó el diagnóstico. en ayunas igual o superior a 126 mg/dl determinó el diagnóstico de Conclusión: La detección sistemática bidireccional de la DM y la TB es DM y entre 110 y 125 mg/dl definió el diagnóstico de prediabetes. factible y ofrece un alto rendimiento diagnóstico de casos de DM en Resultados: De los 510 pacientes con TB, 32 contaban con un di- los pacientes que sufren TB. El rendimiento diagnóstico del cribado de agnóstico previo de DM (6,3%). Con el cribado de los 478 pacientes la TB en los pacientes DM fue bajo y se precisan nuevas investigaciones restantes se estableció el diagnóstico de 15 casos de prediabetes que utilicen instrumentos diagnósticos de la TB nuevos y mejorados.

Public Health Action (PHA) The voice for operational research. e-ISSN 2220-8372 Published by The Union (www.theunion.org), PHA provides a platform to Editor-in-Chief: Donald A Enarson, MD, Canada fulfi l its mission, ‘Health solutions for the poor’. PHA publishes high-quality Contact: [email protected] scientifi c research that provides new knowledge to improve the accessibility, PHA website: http://www.theunion.org/index.php/en/journals/pha equity, quality and effi ciency of health systems and services. Article submission: http://mc.manuscriptcentral.com/pha InternaƟ onal Union Against Tuberculosis and Lung Disease Public Health Action Health soluƟ ons for the poor VOL 3 SUPPLEMENT 1 PUBLISHED 4 NOVEMBER 2013

Characteristics of patients with diabetes screened for tuberculosis in a tertiary care hospital in South India S. Kumpatla,1 A. Sekar,1 S. Achanta,2 B. N. Sharath,2,3 A. M. V. Kumar,4 A. D. Harries,5,6 V. Viswanathan1 http://doi.dx.org/10.5588/pha.13.0035 screening of the two diseases.11 However, screening AFFILIATIONS Setting: Tertiary care hospital for diabetes mellitus (DM) 1 MV Hospital for Diabetes methods, recording and reporting for the two diseases and Prof M Viswanathan in Tamil Nadu, South India. in routine health care settings have not been well de- Diabetes Research Centre, Chennai, India (World Objective: To compare the socio-demographic, clinical termined, and operational research is needed to pro- and biochemical characteristics in DM patients with and Health Organization vide better information in this area.12 Based on these Collaborating Centre for without tuberculosis (TB). recommendations, a standardised procedure for bi- Research, Education and Design: A descriptive study involving a review of rou- Training in Diabetes) directional screening, a monitoring tool and a quarterly 2 World Health Organization tinely maintained records to capture the results of screen- system of recording and reporting were recently devel- Country Office in India, ing of DM patients for TB between March and December New Delhi, India oped and implemented in eight tertiary centres and 3 Department of Commu- 2012. DM patients were first asked whether they already more than 60 peripheral health facilities across India. nity Medicine, Employees had TB, and if not they were screened for TB symptoms, State Insurance Corpora- Our hospital was one of the eight tertiary health care tion (ESIC) Medical followed by investigations for and possible diagnosis facilities that participated in the pilot screening of DM College, Bangalore, India of TB. 4 International Union patients for TB. Evaluation of this pilot in 2012 by the Against Tuberculosis and Results: Of 7083 DM patients, 38 already had TB. Of the India Diabetes Mellitus–Tuberculosis Study Group Lung Disease ( The Union), remainder, 125 (1.8%) had TB symptoms; 71 were inves- South-East Asia Office, showed that it is feasible to screen DM patients for TB New Delhi, India tigated and 12 were newly diagnosed with TB. Of the 50 within the routine setting, resulting in high rates of 5 The Union, Paris, France 6 London School of Hygiene TB patients, 64% had smear-positive pulmonary TB (PTB). 13 detection of TB. & Tropical Medicine, DM-TB patients were older, and had lower education Despite these good results, there is a paucity of in- London, UK level and economic status, a higher frequency of alcohol formation on the association of socio-demographic CORRESPONDENCE use, lower body mass index, a longer duration of DM, a characteristics and clinical features in DM patients Vijay Viswanathan MV Hospital for Diabetes & greater likelihood of receiving insulin and poorer glycae- with TB. In the present study, we therefore aimed to mic control. Prof M Viswanathan Diabetes describe the screening of DM patients for TB and the Research Centre Conclusion: Screening of DM patients for TB was feasible socio-demographic characteristics, clinical features and No. 4 in a tertiary care hospital. The yield of new TB cases was Main Road, Royapuram, biochemical variables of DM patients in relation to the Chennai–600013, India + low and merits further investigation. Socio-demographic diagnosis of TB in a tertiary care hospital for DM in Tel: ( 91) 44259 54913 and clinical characteristics were different in patients with Fax (+91) 44259 54919 South India. e-mail: drvijay@mvdiabetes. DM and TB compared to those with DM only. com

KEY WORDS METHODS DM; TB; India; screening

iabetes mellitus (DM) has become a global epi- Study design Ddemic, especially in low- and middle-income This was a descriptive study involving a review of rec- countries, where 80% of DM-related mortality is esti- ords maintained during the pilot screening of DM pa- mated to occur.1 Currently, there are more than 61 mil- tients for TB. lion people living with DM in India.2 In a similar vein to DM, about one third of the world’s population is Setting currently infected with Mycobacterium tuberculosis, The study was conducted at the MV Hospital for Dia- and approximately 8.8 million new cases of active TB betes, a 100-bed tertiary care hospital for DM in the are identifi ed globally each year.3 India also has a huge state of Tamil Nadu, South India. More than 200 000 TB burden, with an estimated 2.3 million new cases DM patients have been registered in care at the hospi- every year.4 tal since its opening, and 100–200 patients visit the There is good evidence that the risk of TB among hospital every day on an out-patient basis. Patients at- people with DM is three times higher than in those tending the hospital and suspected of having DM are without DM,5 and patients with both DM and TB have screened using the 2 h 75 g oral glucose tolerance test. poorer TB treatment outcomes.6–8 India, which has a The diagnosis of DM is based on previous DM history high dual burden of DM and TB,9,10 could benefi t if or on the WHO’s criteria for the classifi cation of glu- patients were screened early for either disease. The cose intolerance.14 Fasting and postprandial samples World Health Organization (WHO) and the Interna- are collected from known cases of DM. tional Union Against Tuberculosis and Lung Disease For the TB screening, all DM patients presenting to Received 24 May 2013 (The Union) launched the ‘Collaborative framework the out-patient department were asked whether they Accepted 17 July 2013 for the Care and Control of Diabetes and Tuberculo- had already been diagnosed with TB and were on TB sis’, with one of several important recommendations treatment. If the answer was yes, this was recorded PHA 2013; 3(S1): S23–S28 being the routine implementation of bidirectional and the patient was not asked again about TB until © 2013 The Union Public Health Action Screening DM patients for TB S24 completion of TB treatment. If the answer was no, the smoking in the last 3 months) and alcohol consump- ACKNOWLEDGEMENTS The authors acknowledge patient was screened for symptoms by trained staff, tion (60 ml of alcohol daily). Occupation status was the help rendered by R Priya- based on the Revised National TB Control Programme classifi ed as skilled (carpenter, painter, electrician, fi t- darshini, Krishna, M Rajalak- 13,15 shmi, V Arulmozhi and (RNTCP) guidelines. Briefl y, patients with cough ter, etc.), unskilled (farmer, labourer), business, and C Deepika for data collec- for ⩾2 weeks or any suspicion of active pulmonary TB ‘others’ for categories such as retired people, home- tion. They also acknowledge (PTB) or extra-pulmonary TB were categorised as hav- makers and the unemployed. Economic status was Selvan and A Vigneswari for conducting data analysis. ing presumptive TB and were further investigated to classifi ed as low (family income US$400 per month); 2) clinical A workshop was convened in DM clinic and transported to the government-run features: family history of DM, weight and height (for Delhi, India, for the purpose of writing the papers that are microscopy centre (1.5 km away) for sputum smear body mass index [BMI]), duration of DM and current published in this supple- microscopy by Ziehl-Neelsen staining.15 Patients with medication for DM; and 3) blood glucose measure- ment. The workshop was run by the Centre for Opera- negative sputum smears or extra-pulmonary TB sus- ments performed at the time of TB screening, including tional Research, International pects underwent appropriate investigations such as fasting and postprandial glucose in mg/dl and glyco- Union Against Tuberculosis and Lung Disease (The chest radiography to confi rm TB. Those subsequently sylated haemoglobin (HbA1c) in %. Plasma glucose was Union), Paris, France; The diagnosed with TB were referred to the RNTCP for TB estimated using the glucose oxidase peroxidase method. Union South-East Asia Office, New Delhi, India; the Opera- treatment. All patient data were recorded on treatment HbA1c was estimated using the high-performance tional Research Unit, cards and captured in an electronic database. liquid chromatography method with Bio-Rad Variant Médecins Sans Frontières, Luxembourg; the World The TB screening process started in March 2012 Turbo equipment (Bio-Rad Laboratories, Hercules, CA, Health Organization Country and was performed when the patient visited the clinic. USA; Appendix Table A). Data were also collected on Office in India, New Delhi, Screening was done on every patient visit. For the pur- the TB screening process, diagnosis of TB, type of TB India; the All India Institute of Medical Sciences, New pose of this study, however, we only describe the re- and referral for TB care. Delhi, India; and ESIC sults of screening on the fi rst visit. Medical College, Bangalore, India. Analysis and statistics Funding for the workshop Study population Data were extracted from the electronic database and and open access publication was received from the World All DM patients aged ⩾15 years attending the MV analysed using SPSS (Statistical Package and Service Diabetes Foundation, hospital for their routine DM care and screened for TB Solutions, version 16.0, SPSS Inc, Chicago, IL). The Gentofte, Denmark. Conflict of interest: none between March and December 2012 were included in fl ow of patients from screening to diagnosis of TB was declared. the study. described, and the socio-demographic characteristics, clinical features and biochemical variables of DM pa- Data variables and sources of data tients without TB symptoms (DM only) and with TB Data variables included: 1) socio-demographic charac- (previously known and newly diagnosed DM-TB) were teristics: DM registration number, age, sex, residence, evaluated. Patients with symptoms of TB who were ei- education, occupation and socio-economic status, ther not investigated or not diagnosed with TB and smoking (current smoker was defi ned as a history of patients with missing data were not included in this

FIGURE Flow chart showing the results of screening DM patients for TB in a tertiary care hospital in South India, March–December 2012. DM = diabetes mellitus; TB = tuber- culosis; M = male; F = female. Public Health Action Screening DM patients for TB S25 comparative analysis. Mean and standard deviations (SD) were RESULTS calculated. Continuous variables such as age, BMI and duration of DM were converted to categorical variables and compared The TB screening process is shown in the Figure. Of 7083 DM u sing the χ² test where appropriate. Levels of signifi cance were patients screened for TB, (mean age 54 [SD 11.3] years, 4255 [60%] set at 5%. male), 38 had a known history of TB and were on treatment (26 smear-positive PTB, 8 smear-negative PTB, 2 extra-pulmonary Ethics approval TB and 2 with no information on TB type; 33 new and 5 retreat- Ethics approval for the study was obtained from the Institutional ment patients). There were 6920 patients with no symptoms of Ethics Committee of MV Diabetes Research Centre and The Union TB and 125 (1.8%) who screened positive for TB symptoms. Of Ethics Advisory Group. these 125 patients, 71 (57%) submitted two sputum specimens for smear examination and were investigated for TB, and 12 were diagnosed with TB (6 new smear-positive PTB, 4 new smear- TABLE 1 Types and categories of TB in DM patients screened for negative PTB and 2 extra-pulmonary TB). The clinical and socio- TB in a tertiary care hospital between March and December 2012 demographic characteristics of the 71 patients examined for TB did not differ substantially from those of the 54 patients not in- Patients vestigated (Table A). In all, there were 50 patients (0.7%) with TB Category and type of TB n among the DM patients; the categories and types of TB are shown New 45 in Table 1. Notably, 32 patients (64%) had smear-positive PTB. Smear-positive PTB 27 Smear-negative PTB 12 The socio-demographic characteristics of DM-TB and DM-only Extra-pulmonary TB 4 patients are shown in Table 2. A signifi cantly higher proportion Not recorded 2 of DM-TB patients were male, older and had a lower education Retreatment 5 level and socio-economic status than the DM-only patients. More Relapse smear-positive PTB 1 patients in the DM-TB group consumed alcohol compared to the Failure smear-positive PTB 1 DM-only group. The percentage of subjects defi ned as current Return after default smear-positive PTB 3 cigarette smokers was also higher in the DM-TB group than that TB = tuberculosis; DM = diabetes mellitus; PTB = pulmonary tuberculosis. in the DM-only group, but the difference was not statistically signifi cant. The clinical features of the DM-TB and DM-only patients are TABLE 2 Socio-demographic characteristics of DM patients shown in Table 3. There were signifi cant differences between the with and without TB* in a tertiary care hospital in South India, DM-TB and DM-only groups: there was a stronger family history March–December 2012 of DM in the DM-only group; the mean BMI was lower in the DM DM DM-TB group; and the DM-TB group had a higher proportion of patients patients patients with a duration of DM >10 years, and who were on com- with TB with no TB* bined oral medication and insulin. Characteristic n (%) n (%) P value Biochemical variables in DM-TB patients and DM-only patients Total 47 (100) 6113 (100) are shown in Table 4. Although there were no differences in fasting Sex and postprandial plasma glucose between the two groups, more in- < Male 39 (83) 3544 (58) 0.001 dividuals had HbA1c ⩾9% in the DM-TB than the DM-only group. Female 8 (17) 2569 (42) Age, years, mean ± SD 57.4 ± 12.7 53.8 ± 11.2 0.03 Age, years TABLE 3 Clinical features in DM patients with and without TB* in a 15– 44 5 (10.6) 1135 (18.6) 0.23 tertiary care hospital in South India, March–December 2012 45–64 32 (68.1) 3956 (64.7) 0.74 ⩾ 65 10 (21.3) 1022 (16.7) 0.52 DM DM Residence patients patients Urban 27 (57.4) 3383 (55.3) 0.77 with TB with no TB* Rural 20 (42.6) 2730 (44.7) Characteristic n (%) n (%) P value Educational status Total 47 (100) 6113 (100) No school 8 (17) 1182 (19.3) 0.83 < Primary/high school 31 (66) 2555 (41.8) <0.001 Family history of DM 22 (46.8) 4740 (77.5) 0.001 Technical/diploma 2 (4.3) 964 (15.8) 0.05 BMI, kg/m2, mean ± SD 22.8 ± 5.03 27.3 ± 4.45 <0.001 University degree 6 (12.8) 1412 (23.1) 0.13 <18 8 (17.0) 49 (0.8) 0.001 18–23 19 (40.4) 865 (14.2) <0.001 Occupation > < Skilled 5 (10.6) 741 (12.1) 0.93 23 20 (42.6) 5199 (85) 0.001 Unskilled 16 (34) 517 (8.5) <0.001 Duration of DM, years, ± ± ± < Business 3 (6.4) 974 (15.9) 0.11 mean SD 10.9 8.7 5.93 5.99 0.001 < < Other 23 (48.9) 3881 (63.5) 0.05 5 17 (36.2) 3793 (62.0) 0.001 5–10 6 (12.8) 1135 (18.6) 0.41 Economic status >10 24 (51.1) 1185 (19.4) <0.001 Low 13 (27.7) 1783 (29.2) 0.95 Middle 33 (70.2) 3433 (56.2) 0.07 Current DM treatment High 1 (2.1) 897 (14.7) 0.03 Diet only 0 184 (3.0) NA Oral medication only 8 (17.0) 3465 (56.7) <0.001 Current smoking 7 (14.9) 589 (9.6) 0.22 Insulin only 4 (8.5) 98 (1.6) <0.01 Alcohol consumption 13 (27.7) 627 (10.3) <0.001 Oral plus insulin 35 (74.5) 2366 (38.7) <0.001

* Those with no symptoms suggestive of TB. * Those with no symptoms suggestive of TB. DM = diabetes mellitus; TB = tuberculosis; SD = standard deviation. DM = diabetes mellitus; TB = tuberculosis; BMI = body mass index. Public Health Action Screening DM patients for TB S26

TABLE 4 Biochemical variables in DM patients with and without ease, an association that has previously been reported as a risk TB* in a tertiary care hospital of South India, March–December 2012 factor for TB.24 The strengths of this study were that a large number of DM pa- DM DM patients patients tients were screened, the registration of patients was consecutive with TB with no TB* and robust, allowing a denominator for the study, and the record- (N = 47) (N = 6920) ing and reporting system using the established electronic database ± ± mean SD mean SD worked well. There were a number of limitations and challenges. Characteristic or n (%) or n (%) P value First, over 40% of patients with symptoms suggestive of TB were Fasting plasma glucose, reluctant to give sputum specimens for reasons that are currently ± ± mg/dl 189.4 75.1 170.4 69.7 0.17 unclear and may be related to stigma or disbelief that they might Postprandial plasma have TB, and this requires further prospective research using qual- glucose, mg/dl 302.9 ± 124 265 ± 111 0.05 Glycosylated haemoglobin† 9.2 ± 2.1 8.5 ± 2.1 0.03 itative methods. Second, we are reporting only on the results of <7 7 (15.9) 1234 (24.4) 0.26 the fi rst screening, and it is possible that patients returning to the 7–8.9 13 (29.5) 1988 (39.3) 0.24 clinic and being screened again may be found to have TB on sub- ⩾ 9 24 (54.5) 1834 (36.3) 0.02 sequent occasions. This is also the subject of further research, as * Those with no symptoms suggestive of TB. we will be prospectively following up this cohort of 7000 patients † Data available for 44 DM patients with TB and 5056 DM patients without TB. over several years to determine the yield of TB screening at each = = TB tuberculosis; DM diabetes mellitus. subsequent clinic visit. Third, we relied on symptom screening and sputum smear examination, both of which lack sensitivity for diagnosing TB. Screening by chest radiography might give a DISCUSSION higher yield, but the cost and logistical challenges would also be greater. More research is needed to determine the most appropri- This study shows that the fi rst-time screening of DM patients for ate screening and diagnostic algorithm for DM patients attending TB is feasible in a routine health care setting of a tertiary care hos- clinics, and also whether it is cost-effective to use newer technol- pital for DM in South India. Over a period of 9 months, during ogy such as Xpert® MTB/RIF (Cepheid, Inc, Sunnyvale, CA, USA) which DM patients attending the clinic were screened, we de- for diagnosing TB.25 Advocacy and the political will to deploy tected 50 patients with TB, giving a case rate of 706 per 100 000 more effective and affordable point-of-care diagnostics at DM screened patients. Moreover, 64% of the patients diagnosed had clinics might assist in the diagnosis and management of both of smear-positive disease and were therefore highly infectious. Most these diseases. This is urgently required if we are to move forward of the patients had been diagnosed outside the DM clinic and in making TB screening routine in DM clinics. Fourth, the lower were already on treatment at the time of screening. This is proba- mean BMI in the DM-TB group compared to the DM-only group bly due to the very good geographical coverage of the RNTCP and should be interpreted with caution, given the cross-sectional na- good access to community TB case fi nding services in India.16 The ture of the study, and it is not clear whether TB led to low BMI or relatively low yield in detecting new TB cases merits further inves- whether TB occurred as a result of a low BMI. Fifth, a multiple re- tigation. However, if we exclude those already diagnosed with TB, gression model would have strengthened the conclusions of the the case rate for newly diagnosed cases would be 169 per 100 000 study, but due to the small number of TB-DM cases, we felt it pru- patients, which is still higher than the total TB case notifi cation dent to limit our analysis to bivariate associations. rate in India. In terms of policy, the screening of TB patients for DM is rela- DM patients with TB tended to be male, older, have low BMI tively straightforward, and in pilot studies in different states in and were more likely to consume tobacco and alcohol than pa- India this has resulted in a high rate of detection of DM in TB pa- tients without TB. These fi ndings might be expected given the ep- tients (10–15%).26 A decision was thus made by the Ministry of idemiology of TB. Rieder has reported that age and sex are strong Health to scale up screening of DM in TB patients countrywide, determinants of TB, with the highest risks being found in elderly with a national training manual and revised registers and treat- people,17 and this was evident in the current study.. Systematic re- ment cards to support this activity. However, screening of DM pa- views have shown that undernutrition, smoking, diabetes and al- tients for TB is less easy, with pilot studies in India documenting cohol misuse are individual risk factors that can double or triple challenges such as poor registration of DM patients, a reluctance the risk of developing active TB.18 A mathematical modelling by DM doctors to undertake this additional work and systemati- analysis of the effects of smoking on TB infection and mortality cally record every screening event, and resistance of patients to projected that smoking would produce an excess of 18 million TB submit sputum specimens.13 Our current study also confi rms cases and 40 million deaths from TB between 2010 and 2050.19 It these fi ndings. Importantly, an electronic recording system can be was also reported that smoking could delay the attainment of used to reliably collect and report on the necessary data. Paper- the Stop TB Partnership target of reducing TB mortality by 50% based systems could probably not work long-term in this type of from 1990 to 2015.19 A large part of the TB burden in India can be setting with patients in chronic care, and there is a need for elec- attributed to smoking (40%)20 and DM (15%).21 Alcohol misuse tronic health systems, as shown elsewhere in DM clinics in Africa and DM are predicted to increase in low- and middle-income and the Middle East.27,28 Further research is needed to better un- countries, and might be crucial factors in the coming decades.22,23 derstand some of these challenges, and in particular the value of More patients with both DM and TB came from the unskilled repeated screening and use of point-of-care TB diagnostics at each workforce sector with lower socio-economic status, more had had subsequent clinic visit. This study shows the importance of good their metabolic disease for >10 years, were on combination ther- collaboration between communicable and non-communicable apy with oral and insulin medication and their glycosylated hae- disease programmes. moglobin levels were signifi cantly higher, indicating poorer gly- In conclusion, the study highlights that screening of DM pa- caemic control. These data all indicate that DM patients with TB tients for TB is feasible in the routine health care setting of a terti- were poorer and had long-standing, severe and uncontrolled dis- ary care hospital. DM-TB patients were older, had lower education Public Health Action Screening DM patients for TB S27

levels and socio-economic status, a higher frequency of smoking tional Diabetes Federation, 2012. http://www.eatlas.idf.org/diabetesatlas/5e/ and alcohol use, a longer duration of DM, a greater likelihood of update2012 Accessed August 2013. 3 World Health Organization. Global tuberculosis control, 2011. WHO/HTM/ being on oral medication and insulin and lower BMI and poorer TB/2011.16. Geneva, Switzerland: WHO, 2011. glycaemic control. 4 World Health Organization. Global tuberculosis report, 2012. WHO/HTM/ TB/2012.6. Geneva, Switzerland: WHO, 2012. http://www.who.int/tb/data Accessed August 2013. APPENDIX 5 Jeon C Y, Murray M B. Diabetes mellitus increases the risk of active tubercu- losis: a systematic review of 13 observational studies. PLOS Med 2008; 5: e152. TABLE A Data variables 6 Alisjahbana B, Sahiratmadja E, Nelwan E J, et al. The effect of type 2 diabetes mellitus on the presentation and treatment response of pulmonary N = 125 (M:F 79:46) tuberculosis. Clin Infect Dis 2007; 45: 428– 435. DM patients DM patients 7 Restrepo B I, Fisher-Hoch S P, Smith B, et al. Mycobacterial clearance from with symptoms with symptoms sputum is delayed during the fi rst phase of treatment in patients with suggestive of suggestive of diabetes. Am J Trop Med Hyg 2008; 79: 541–544. TB who were TB who were 8 Wang C S, Yang C J, Chen H C, et al. Impact of type 2 diabetes on manifesta- tions and treatment outcome of pulmonary tuberculosis. Epidemiol Infect investigated not investigated = = 2009; 137: 203–210. n 71 (46:25) n 54 (32:22) 9 Viswanathan V, Kumpatla S, Aravindalochanan A, et al. Prevalence of dia- n (%) n (%) P value betes and pre-diabetes and associated risk factors among tuberculosis pa- Socio-demographic tients in India. PLOS ONE 2012; 7: e41367. characteristics 10 Balakrishnan S, Vijayan S, Nair S, et al. High diabetes prevalence among tu- Age, years, mean ± SD 57.1 ± 11.3 56 ± 11.1 0.58 berculosis cases in Kerala, India. PLOS ONE 2012; 7: e46502. Residence 11 World Health Organization/International Union Against Tuberculosis and Rural 24 (33.8) 17 (31.5) 0.85 Lung Disease. Provisional collaborative framework for care and control of Urban 47 (66.2) 37 (68.5) tuberculosis and diabetes. WHO/HTM/TB/2011.15. Geneva, Switzerland: WHO, 2011. Educational status 12 Harries A D, Murray M B, Jeon C Y, et al. Defi ning the research agenda to re- No schooling 5 (7) 4 (7.4) 0.75 duce the joint burden of disease from diabetes mellitus and tuberculosis. Primary/high school 51 (71.8) 41 (75.9) 0.76 Trop Med Int Health 2010; 15: 659–663. Technical/diploma 6 (8.5) 5 (9.3) 0.87 13 India Diabetes Mellitus–Tuberculosis Study Group. Screening of patients University degree 9 (12.7) 4 (7.4) 0.51 with diabetes mellitus for tuberculosis in India. Trop Med Int Health 2013; Occupation 18: 646–654. Skilled 6 (8.5) 2 (3.7) 0.48 14 World Health Organization. Defi nition, diagnosis and classifi cation of diabe- Unskilled 14 (19.7) 7 (13) 0.45 tes mellitus and its complications. Report of WHO consultation. Part 1: Business 11 (15.5) 12 (22.2) 0.47 Diagnosis and classifi cation of diabetes mellitus. WHO/NCD/NCS/99.2. Ge- Other 40 (56.3) 33 (61.1) 0.72 neva, Switzerland: WHO, 1999. http://whqlibdoc.who.int/hq/1999/who_ncd_ Socio-economic status (USD) ncs_99.2.pdf Accessed August 2013. Low (<200) 5 (7) — 15 Central Tuberculosis Division, Revised National Tuberculosis Control Pro- Middle (200–400) 60 (84.5) 47 (87) 0.89 gramme, Directorate General of Health Services, Ministry of Health and High (>400) 6 (8.5) 7 (13) 0.60 Family Welfare. Technical and operational guidelines for tuberculosis con- Current smoking 7 (9.9) 4 (7.4) 0.76 trol. New Delhi, India: Government of India, 2005. Alcohol consumption 9 (12.7) 4 (7.4) 0.39 16 Sachdeva K S, Kumar A, Dewan P, Kumar A M V, Satyanarayana S. New vision for Revised National Tuberculosis Control Programme (RNTCP): universal Clinical features access—‘Reaching the un-reached’. Ind J Med Res 2012; 135: 690–694. Positive family history 17 Rieder H. Epidemiologic basis of tuberculosis control. Paris, France: Interna- of DM 52 (73.2) 36 (66.7) 0.44 tional Union Against Tuberculosis and Lung Disease, 1999. BMI, kg/m2, mean ± SD 24.6 ± 4.8 27.3 ± 4.9 0.002 18 Lönnroth K, Castro K G, Chakaya J M, et al. Tuberculosis control and elimi- Duration of DM nation 2010–50: cure, care and social development. Lancet 2010; 375: 1814– <5 years 29 (40.8) 21 (38.9) 0.97 1829. 5–10 years 16 (22.5) 18 (33.3) 0.25 19 Basu S, Stuckler D, Bitton A, Glantz S A. Projected effects of tobacco smoking >10 years 26 (36.6) 15 (27.8) 0.39 on worldwide tuberculosis control: mathematical modelling analysis. BMJ Current DM treatment 2011; 343: d5506. 20 Hassmiller K. The impact of smoking on population level tuberculosis out- Diet only 1 (1.4) — comes. TSRU progress report, 2007. The Hague, The Netherlands: KNCV, Oral medication only 28 (39.4) 24 (44.4) 0.7 2007. Insulin only 2 (2.8) 2 (3.7) 0.82 21 Stevenson C R, Forouhi N G, Roglic G, et al. Diabetes and tuberculosis: the Oral plus insulin 40 (56.3) 28 (51.9) 0.75 impact of the diabetes epidemic on tuberculosis incidence. BMC Public Biochemical variables Health 2007; 7: 234. Fasting plasma glucose, 22 Dooley K E, Chaisson R E. Tuberculosis and diabetes mellitus: convergence mg/dl 165.8 ± 67.6 172 ± 82 0.687 of two epidemics. Lancet Infect Dis 2009; 9: 737–746. Postprandial plasma 23 World Health Organization, Department of Mental Health and Substance glucose, mg/dl 238.9 ± 90.1 265.6 ± 92.5 0.12 Abuse. Global status report on alcohol 2004: country profi les Uganda. Ge- Glycosylated haemoglobin neva, Switzerland: WHO, 2004. % 9.1 ± 2.3 8.6 ± 1.9 0.19 24 Restrepo B I, Fisher-Hoch S P, Pino P A, et al. Tuberculosis in poorly con- trolled type 2 diabetes: altered cytokine expression in peripheral white blood M = male; F = female; DM = diabetes mellitus; TB = tuberculosis; SD = standard cells. Clin Infect Dis 2008; 47: 634–641. deviation. 25 Boehme C C, Nabeta P, Hillemann D, et al. Rapid molecular detection of tu- berculosis and rifampin resistance. N Engl J Med 2010; 363: 1005–1015. 26 India Tuberculosis–Diabetes Study Group. Screening of patients with tuber- culosis for diabetes mellitus in India. Trop Med Int Health 2013; 18: 636– 645. References 27 Allain T J, van Oosterhout J J, Douglas G P, et al. Applying lessons learnt from the ‘DOTS’ tuberculosis model to monitoring and evaluating persons 1 World Health Organization. Diabetes. Fact sheet no. 312. Updated March with diabetes mellitus in Blantyre, Malawi. Trop Med Int Health 2001; 16: 2013. WHO Media Centre: WHO, 2013. http://www.who.int/mediacentre/ 1077–1084. factsheets/fs312/en/index.html Accessed August 2013. 28 Khader A, Farajallah L, Shahin Y, et al. Cohort monitoring of persons with 2 International Diabetes Federation. IDF diabetes atlas. 5th ed, 2012 update. diabetes mellitus in a primary healthcare clinic for Palestine refugees in Jor- Unwin N, Whiting D, Guariguata L, et al., eds. Brussels, Belgium: Interna- dan. Trop Med Int Health 2012; 17: 1569–1576. Public Health Action Screening DM patients for TB S28

Contexte : Hôpital de soins tertiaires pour le diabète sucré (DM) à investigués et 12 nouveaux cas de TB ont été diagnostiqués. Une TB Tamil Nadu, Inde du Sud. pulmonaire à frottis positif existait chez 64% de l’ensemble des 50 pa- Objectif : Comparer les caractéristiques socio-démographiques, cli- tients TB. Les patients DM-TB étaient plus âgés, bénéficiaient d’une niques et biochimiques chez les patients DM atteints ou non de tu- éducation et d’un statut socio-économique moins favorables, con- berculose (TB). naissaient une fréquence plus élevée d’utilisation d’alcool, avaient un Schéma : Etude descriptive impliquant une révision des dossiers en- indice de masse corporelle plus bas et une durée plus longue du DM tretenus en routine entre mars et décembre 2012 pour y prélever les ainsi qu’une susceptibilité plus grande de recevoir de l’insuline et un résultats du dépistage chez les patients DM à la recherche d’une TB. contrôle moins bon de la glycémie. On a d’abord demandé aux patients DM s’ils avaient déjà été diag- Conclusion : Le dépistage de la TB chez les patients DM est réalisable nostiqués avec une TB et dans la négative, on les a dépistés à la dans un hôpital de soins tertiaires. Le rendement en nouveaux cas de recherche de symptômes de TB, puis investigués et diagnostiqués en TB est bas et mérite d’être investigué davantage. Les caractéristiques matière de TB. socio-démographiques et cliniques sont différentes chez les patients Résultats : Sur 7083 patients DM, 38 souffraient déjà de TB. Parmi les atteints de DM-TB par comparaison avec ceux atteints seulement restants, 125 (1,8%) étaient atteints de symptômes de TB, 71 ont été de DM.

Marco de referencia: Un hospital de atención terciaria de la diabetes y se detectaron 12 casos nuevos de TB. De los 50 pacientes que (DM) en Tamil Nadu, en el sur de la India. padecían TB, el 64% correspondió a una TB pulmonar con bacilosco- Objetivo: Comparar las características sociodemográficas, clínicas y pia positiva. Los pacientes con ambas enfermedades eran mayores, bioquímicas entre los pacientes con diagnóstico de DM que padecen poseían un menor grado de instrucción, presentaban una situación tuberculosis (TB) y los pacientes que sufren exclusivamente de DM. económica menos favorable, mayor frecuencia de consumo de alcohol, Métodos: Fue este un estudio descriptivo, que consistió en examinar un índice de masa corporal más bajo, una evolución más prolongada los expedientes clínicos corrientes, a fin de investigar los resultados de la DM, mayor probabilidad de estar recibiendo insulina y exhibían del cribado de la TB en los pacientes con diagnóstico de DM entre un equilibrio menos eficaz de la glucemia. marzo y diciembre del 2012. Inicialmente se preguntó a los pacientes Conclusión: Es posible practicar la detección de la TB en los pacientes DM sobre el antecedente personal de TB y en los casos negativos, se DM en un hospital de atención terciaria. Se obtuvo un bajo rendi- buscaron los síntomas indicativos de la enfermedad y se practicó la miento diagnóstico de casos nuevos de TB, lo cual justifica nuevas investigación y el diagnóstico de la TB. investigaciones. Al comparar los pacientes que padecen DM única- Resultados: De los 7083 pacientes con diagnóstico de DM, 38 tenían mente con los pacientes que padecen ambas enfermedades se observa- un antecedente de TB y del resto de pacientes, 125 presentaban sínto- ron diferencias en las características sociodemográficas y clínicas. mas indicativos de la enfermedad (1,8%), se investigaron 71 pacientes

Public Health Action (PHA) The voice for operational research. e-ISSN 2220-8372 Published by The Union (www.theunion.org), PHA provides a platform to Editor-in-Chief: Donald A Enarson, MD, Canada fulfi l its mission, ‘Health solutions for the poor’. PHA publishes high-quality Contact: [email protected] scientifi c research that provides new knowledge to improve the accessibility, PHA website: http://www.theunion.org/index.php/en/journals/pha equity, quality and effi ciency of health systems and services. Article submission: http://mc.manuscriptcentral.com/pha InternaƟ onal Union Against Tuberculosis and Lung Disease Public Health Action Health soluƟ ons for the poor VOL 3 SUPPLEMENT 1 PUBLISHED 4 NOVEMBER 2013

Screening patients with tuberculosis for diabetes mellitus in Gujarat, India

P. Dave,1 A. Shah,2 M. Chauhan,1 A. M. V. Kumar,3 A. D. Harries,4,5 S. Malhotra,6 K. Pujara,1 P. Patel,1 M. Mane,1 A. Thakkar,1 S. Bharaswadkar,2 B. N. Sharath,7 S. Achanta2 http://dx.doi.org/10.5588/pha.13.0027 with an expected rate of increase of 2.0 per 1000 pop- AFFILIATIONS 1 Department of Health, 4 Setting: Anklav Tuberculosis Unit, Anand, Gujarat, India. ulation per year. Government of Gujarat, Objective: To determine in a cohort of TB patients 1) the DM has been shown to be an independent risk factor Gandhinagar, India 2 World Health Organization 5 prevalence of diabetes (DM) and impaired fasting glucose for TB, and there is evidence showing high DM preva- Country Office for India, (IFG), 2) the time taken for diagnosis, 3) demographic lence in TB patients in community-based studies mainly New Delhi, India 3 International Union 6–8 and clinical factors associated with DM and IFG, and 4) the from the southern part of India. DM is said to account Against Tuberculosis and number needed to screen (NNS) for diagnosing new for about 10% of the population-attributable fraction Lung Disease (The Union), South-East Asia Office, 9 cases of DM and IFG. of TB cases globally, although this will vary from coun- New Delhi, India Design: Descriptive study. TB patients registered be- try to country depending on the degree of overlap be- 4 The Union, Paris, France 5 Department of Infectious tween January and September 2012 were asked whether tween DM and TB in the population. It has been sug- and Tropical Diseases, they had a history of DM. Those with unknown DM were gested by modelling analysis that DM may account for London School of Hygiene & Tropical Medicine, 10 tested for random and fasting blood glucose (FBG). FBG 15% of all TB and 21% of smear-positive TB cases. London, UK of ⩾126 mg/dl and 110–125 mg/dl were considered in- This strong association raises the question as to 6 All India Institute of Medical Sciences, dicative of DM and IFG, respectively. whether TB patients should be routinely screened for New Delhi, India Results: Of 556 TB patients, 553 (99%) were assessed: DM. In this regard, a systematic review of screening 7 Employees State Insurance Corporation (ESIC) 36 (6.5%) had diabetes (14 had been previously diag- conducted in multiple settings showed that the yield of Medical College, nosed DM and 22 were newly diagnosed), and 39 (7%) DM from screening patients with TB varied from 1.7% Bangalore, India had IFG. The median (interquartile range) time to DM to 36%, with this proportion being infl uenced by the CORRESPONDENCE diagnosis was 5 (1–17) days. Age ⩾35 years was associ- underlying burden of TB and the severity of DM.11 The Amar Shah World Health Organization ated with DM. The NNS was 25 and 14 for one new case presence of concomitant DM amongst TB patients is Country Office for India of DM and IFG, respectively, with a lower NNS in males, also associated with a higher risk of poor treatment out- 5 Jivraj Mehta Bhavan Sector-10 12 those aged ⩾35 years, those with smear-positive pulmo- comes in terms of treatment failure, relapse and death. Gandhinagar, Gujarat, India nary TB, retreatment patients and smokers. Screening for DM in TB patients could thus potentially Tel: (+91) 94281 00915 e-mail: dr.amar.mph@ Conclusion: This pilot project shows that it is feasible identify DM cases early; this might lead to early link- gmail.com and valuable to screen patients with TB for DM in a rou- age to DM care,13 which in turn might improve treat- KEY WORDS 14 tine setting, resulting in earlier identification of DM and ment outcomes. TB; DM; pre-diabetes; opportunities for better management of comorbidity. There has been an international call for investment screening; Gujarat in collaborative care directed towards TB and DM.15 Appreciating the need to implement this joint frame- work in India, a consensus was reached among multi- uberculosis (TB) continues to be a major health ple stakeholders, including the national authorities, Tproblem in India, with an estimated 3.1 million to initiate screening for DM among TB patients in pilot prevalent cases and an average annual mortality of settings across India. The results of this countrywide 0.32 million.1 This high national burden of TB is made screening have been reported in terms of aggregate potentially worse by other diseases, such as human data,13 and this has encouraged the belief that DM immunodefi ciency virus/acquired immune-defi ciency screening and care should be integrated into India’s syndrome (HIV/AIDS), in some districts that can lead Revised National TB Control Programme (RNTCP). to higher rates of disease, increased transmission of in- We report the implementation fi ndings and lessons fection and poor treatment outcomes. learnt in Gujarat, one of the network of pilot sites With India in transition in terms of socio-economic where the feasibility of and yield from introducing DM development and lifestyle changes, there is also a ris- screening for TB patients under routine programme ing epidemic of diabetes mellitus (DM). It is estimated settings were assessed. The specifi c objectives of this that there are over 60 million people currently with DM study were to determine 1) the prevalence of DM and in the country, and almost half are undiagnosed, with impaired fasting glucose (IFG) in a cohort of TB pa- the number expected to increase to 80 million by tients, 2) the time intervals for the diagnosis of DM 2,3 Received 14 May 2013 2030. This increase in the prevalence of DM has and IFG, 3) demographic and disease-related factors Accepted 27 June 2013 been noted in both urban and rural areas. A recent re- associated with DM and IFG, and 4) the number needed view reported that the prevalence of DM ranged from to screen (NNS) for diagnosing patients with newly di- PHA 2013; 3(S1): S29–S33 3% to 12% across different rural areas of the country, agnosed DM and IFG. © 2013 The Union Public Health Action Screening of TB patients for DM S30

METHODS FBG 110–125 mg/dl (6.1–6.9 mmol/l) indicates IFG; ACKNOWLEDGEMENTS < < A workshop was convened in and FBG 110 mg/dl ( 6.1 mmol/l) is normal. Delhi, India, for the purpose Study setting and population of writing the papers that are Anand, a district in Gujarat State in western India, has Provision of care for diabetes mellitus published in this supple- ment. The workshop was run a population of 2.1 million. It has 11 community TB patients with FBG of ⩾126 mg/dl were diagnosed by the Centre for Opera- health centres, 43 primary health centres and two ter- with presumptive DM and referred to the diabetes ser- tional Research, International Union Against Tuberculosis tiary care centres, in the form of one private medical vices in public hospitals with DM facilities for a defi ni- and Lung Disease (The college and one civil hospital. For its TB control activi- tive diagnosis and enrolment in care. Three DM health Union), Paris, France; The Union South-East Asia Office, ties, the district is covered by the RNTCP, which has care facilities are distributed equally within the TU, New Delhi, India; the Opera- been operating since 2004 along the same basic princi- one of which is at the private medical college. tional Research Unit, Médecins Sans Frontières, ples as the World Health Organization’s (WHO’s) DOTS The state level offi cer and district TB offi cer were Luxembourg; the World strategy. Under the RNTCP, the district has four Tuber- trained at national level in January 2012, after which Health Organization Country Office in India, New Delhi, culosis Units (TUs), geographical areas defi ned as sub- district and sub-district level staff were trained using a India; the All India Institute district level programme management units, each training module.18 The DM testing kits were supplied of Medical Sciences, New at the end of February 2012. Delhi, India; and ESIC covering a population of 250 000–500 000, with TB di- Medical College, Bangalore, agnostic and treatment services being delivered through India. a network of primary, secondary and tertiary health Study participants Funding for the workshop and open access publication care facilities. The district also has 23 designated mi- All TB patients (adults and children) registered from was received from the World croscopy centres—1 per 0.1 million population—for January to September 2012 were enrolled into the Diabetes Foundation, Gentofte, Denmark. performing quality assured sputum microscopy, and screening programme. DM screening was initiated in Conflict of interest: none 33 sputum collection and transportation centres situ- the facilities from 1 March 2012. declared. ated in the remotest parts of the district where no other health care facility or easy transport system is Data collection and data variables available. An additional TB-DM register was developed and used In the last 5 years, the district has consistently to capture the screening results. All sites were visited achieved the twin RNTCP objectives, i.e., treatment by the study investigators at monthly intervals to success rates of ⩾85% among new smear-positive cases check for completeness, consistency and accuracy of and detection rates of ⩾70% of estimated cases. In data, and any issues relating to recording and report- 2012, 14 963 patients with presumptive TB were in- ing were resolved. Variables related to the study objec- vestigated, and 2999 TB patients were placed on stan- tives were sourced from the TB-DM register and TB dardised treatment. register. The implementation site for the current descriptive study was the Anklav TU in the Anand district. This Data entry analysis and statistics TU covers a population of 0.57 million, 85% of whom All data from the TB-DM register were double entered live in rural areas. The TU has 16 government health into an electronic database created using EpiData en- facilities in the form of peripheral health institutions try software version 3.1 (EpiData Association, Odense, (PHIs), where this project was implemented. Denmark), validated for discrepancies and corrected. Additional variables such as age, sex, type of TB and Screening procedures for diabetes mellitus HIV status, were captured in MS Excel® (Microsoft, All health facilities were provided with DM testing kits Palisade Corp, Newfi eld, NY, USA) and merged into containing glucometers and test strips. Laboratory the EpiData database before analysis. The data was an- technicians were instructed to perform DM screening alysed using EpiData analysis software (version as soon as a patient was confi rmed as having TB. The V2.2.2.180). Statistical differences between patient RNTCP diagnostic algorithm was used for the diagno- characteristics among prevalent DM cases and patients sis of TB.16 In some patients where the diagnosis was with no DM were evaluated using the χ² test or Fisher’s made through sputum collection centres, screening was exact test, as appropriate. Confi dence intervals of 95% performed during outreach sessions (‘Mamata Day’) were used; the level of signifi cance was set at ⩽0.05. organised in the village locally by a laboratory techni- cian. These sessions were held for other primary care Ethics activities into which the DM screening activity was in- Ethics approval was obtained from the Ethics Advisory tegrated to enhance TB screening coverage within the Group of the International Union Against Tuberculosis district. and Lung Disease, Paris, France. Programmatic approval At all sites, patients were fi rst asked whether they was granted for the study at state and national level. had previously been diagnosed with DM. Patients with a previous diagnosis were referred back for DM care so RESULTS that their blood glucose levels could be controlled. Those with no known diagnosis of DM underwent a Of the 556 TB patients (median age 35 years, inter- random blood glucose (RBG) test based on a capillary quartile range [IQR] 25–50 years) registered, the major- blood sample with a glucometer, followed by an FBG ity were screened for DM (Table 1). Of these patients, test at the next visit if the RBG was ⩾110 mg/dl. The 36 (6.5%) were found to have DM: 14 (2.5%) had a diagnosis of DM followed national guidelines, with previous diagnosis of DM and 22 (4%) were newly FBG cut-off thresholds as recommended by the WHO.17 diagnosed. The median FBG in newly diagnosed DM Briefl y, FBG ⩾126 mg/dl (⩾7 mmol/l) indicates DM; patients was 248 mg/dl (IQR 153–337 mg/dl); 39 (7.0%) Public Health Action Screening of TB patients for DM S31

TABLE 1 Screening of patients with TB for DM in Anklav TABLE 2 Interval between start of TB treatment and diagnosis Tuberculosis Unit, Gujarat, January–September 2012 of DM in TB patients in Anklav Tuberculosis Unit, Gujarat, January–September 2012 Indicator n % Quarter 2 Quarter 3 Patients with TB registered over the 3 quarters 556 duation, days duration, days Patients with a known diagnosis of DM 14 2.5 median [IQR] median [IQR] Patients needing to be screened with RBG 542 Patients screened with RBG 539 99.4 TB treatment start date to RBG 2 [0–24] 2 [0–7] Patients with RBG >110 mg/dl, needing to be screened RBG to FBG 2 [1–12] 3 [2–5] with FBG 229 Patients screened with FBG 227 99.1 TB = tuberculosis; DM = diabetes mellitus; IQR = interquartile range; RBG = ran- Patients with FBG ⩾126 mg/dl (newly diagnosed dom blood glucose; FBG = fasting blood glucose. with DM) 22 4.1 Patients with FBG ⩾110 to <126 mg/dl (impaired fasting glucose) 39 7.0 Patients with known and newly diagnosed DM 36 6.5 trict; these had to be procured through external sources, and they Patients with known and newly diagnosed DM referred arrived towards the end of the fi rst quarter. For the two quarters for DM care 34 94.4 combined, the median (IQR) time interval from TB treatment to Patients with known or newly diagnosed DM who reached DM care 32 94.1 DM diagnosis was 5 days (1–17)—with a median of 2 days between TB treatment initiation and RBG, and a further 2 days between = = = = TB tuberculosis; DM diabetes mellitus; RBG random blood glucose; FBG RBG and FBG. fasting blood glucose. The number of TB patients with DM and the NNS for diagnos- ing DM and IFG are shown in Table 3. The only characteristic as- had IFG. Most DM patients were referred to and enrolled into sociated with increased DM prevalence was age ⩾35 years (P = DM care; no action was taken for patients with IFG. Of the 74 pa- 0.001). There were no characteristics associated with IFG. Among tients with extra-pulmonary TB, 54 had lymphadenopathy, 18 those screened, 18 (3.2%) were children. None of these was diag- pleural effusion, 1 abdominal TB, 1 bone TB; of these, 3 had DM. nosed with DM. The NNS for all TB patients was 25 for one new The time intervals between the start of TB treatment and the DM case, and 14 for one new case of IFG. The NNS for identifying RBG and FBG measurements are shown in Table 2. The glucometer a new case of DM was lower in males, those aged ⩾35 years, those and test strips were not available in the health care facility, as the with smear-positive pulmonary TB, those on a re-treatment regi- National Programme for Prevention and Control of Cancer, Diabe- men, smokers and those who were HIV-negative. The median age tes and Stroke (NPCDCS) had not been implemented in the dis- of the HIV-infected subjects was 30 years (IQR 20.5–38.0).

TABLE 3 TB patients NNS to identify new case of DM, one new case with IFG, and both DM and IFG, in Anklav Tuberculosis Unit, Gujarat, January–September 2012

TB TB NNS for patients patients New NNS for NNS for both registered with DM DM IFG new DM new IFG DM and IFG Characteristic n n n n n n n Total 556 36 22 39 25 14 9 Age, years <35 236 4 2 12 118 20 17 ⩾35 320 32 20 27 16 12 7 Sex Male 371 27 18 25 21 15 9 Female 185 9 4 14 46 13 10 Initial sputum smear result* Smear-positive 364 26 17 28 21 13 8 Smear-negative 120 8 4 6 30 20 12 Type of TB† New case 429 24 15 34 29 13 9 Retreatment 126 12 7 5 18 25 11 Current smoking status‡ Current smoker 89 9 9 11 10 8 4 Non-smoker 462 25 13 28 36 17 11 Disease classification Pulmonary 482 34 21 36 23 13 8 Extra-pulmonary 74 2 1 3 74 25 19 HIV status Negative 532 35 22 38 24 14 9 Positive 24 1 0 1 — 24 24

* Initial sputum results were not recorded for 65 extra-pulmonary TB patients and 7 pulmonary TB patients. † One transfer-in patient not included. ‡ Defined as having smoked tobacco at least once in the last month; data not available for 5 patients. TB = tuberculosis; NNS = number needed to screen; DM = diabetes mellitus; IFG = impaired fasting glucose; HIV = human immunodeficiency virus. Public Health Action Screening of TB patients for DM S32

DISCUSSION study also indicates the importance of training peripheral health care workers in the simple management of DM so that patients do In the setting of a TU in India with a predominantly rural popula- not have to make long journeys for specialist care and follow-up. tion, screening of TB patients for DM and IFG was feasible, with nearly 100% of patients agreeing to be tested, and over 13% being References diagnosed with either DM or IFG. The time from the start of TB 1 World Health Organization. Global tuberculosis report 2012. WHO/HTM/ treatment to diagnosis was <1 week, which was short, consider- TB/2012.6. Geneva, Switzerland: WHO, 2012. http://www.who.int/tb/ ing that this was a two-step process requiring RBG to identify publications/global_report/en/ Accessed 9 August 2013. 2 International Diabetes Federation. IDF diabetes atlas, 5th ed. Unwin N, those at risk and FBG to establish the diagnosis. The baseline Whiting D, Guariguata L, et al., eds. Brussels, Belgium: International Diabetes characteristic signifi cantly associated with DM was age, with pa- Federation, 2011. http://www.eatlas.idf.org/ Accessed 9 August 2013. tients aged ⩾35 years having an increased prevalence of disease. 3 Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of diabetes: No case of DM or IFG was found among children suffering from estimates for the year 2000 and projections for 2030. Diabetes Care 2004; 27: 1047–1053. TB. This suggests that child TB patients do not need to be screened 4 Misra P, Upadhyay R P, Misra A, Anand K. A review of the epidemiology of for DM. The NNS to identify one new case of DM was 25, with diabetes in rural India. Diabetes Res Clin Pract 2011; 92: 303–311. fewer needed for certain categories of patient, including men, 5 Jeon C Y, Murray M B. Diabetes mellitus increases the risk of active tubercu- ⩾ losis: a systematic review of 13 observational studies. PLOS Med 2008; 5: those aged 35 years, those with smear-positive and recurrent e152. disease and smokers. 6 Gupta S, Shenoy V P, Bairy I, Srinivasa H, Mukhopadhyay C. Diabetes melli- At 6.5%, the prevalence of DM in TB patients in our TU was tus and HIV as co-morbidities in tuberculosis patients or rural South India. J Infect Public Health 2011; 4: 140–144. low compared with other pilot sites in India.13 This is probably 7 Balakrishnan S, Vijayan S, Nair S, et al. High diabetes prevalence among explained by the fact that our patients came mainly from rural tuberculosis cases in Kerala, India. PLOS ONE 2012; 7: e46502. areas, with a relatively low median age. A study in the urban pop- 8 Viswanathan V, Kumpatla S, Aravindalochanan V, et al. Prevalence of diabetes ulation of Ahmedabad, Gujarat, found a 14% prevalence of type 2 and pre-diabetes and associated risk factors among tuberculosis patients in India. PLOS ONE 2012; 7: e41367. 19 DM and 6% IFG. Due to the young age of the HIV-infected indi- 9 Lönnroth K, Castro K G, Chakaya J M, et al. Tuberculosis control and elimi- viduals, the incidence of pre-existing DM in this group was rela- nation 2010–50: cure, care, and social development. Lancet 2010; 375: tively low. 1814–1829. 10 Stevenson C R, Forouhi N G, Roglic G, et al. Diabetes and tuberculosis: the There are concerns that TB may induce infection-related impact of the diabetes epidemic on tuberculosis incidence. BMC Public hyperglycaemia,20–22 and that investigations for DM should be Health 2007; 7: 234. conducted later during the course of TB treatment to avoid false- 11 Jeon C Y, Harries A D, Baker M A, et al. Bi-directional screening for tubercu- positive diagnoses. In our study, the median FBG in newly diag- losis and diabetes: a systematic review. Trop Med Int Health 2010; 15: 1300– 1314. nosed DM patients was almost 250 mg/dl, which is twice as high 12 Zhang Q, Xiao H, Sugawara I. Tuberculosis complicated by diabetes mellitus as the cut-off threshold. This suggests that even if stress-induced at Shanghai Pulmonary Hospital in China. Jpn J Infect Dis 2009; 62: 390– hyperglycaemia is present, it is important to screen early during 391. 13 India Tuberculosis-Diabetes Study Group. Screening of patients with tuber- TB treatment so that appropriate measures can be taken to reduce culosis for diabetes mellitus in India. Trop Med Int Health 2013; 18: 636–645. the high blood glucose levels. A caveat is that we do not currently 14 Ruslami R, Aarnoutse R E, Alisjahbana B, van der Ven A J, van Crevel R. know whether good blood glucose control early on in the course Implications of the global increase of diabetes for tuberculosis control and of TB treatment affects treatment outcomes. patient care. Trop Med Int Health 2010; 15: 1289–1299. 15 World Health Organization/International Union Against Tuberculosis and The strengths of this study are that we implemented screening Lung Disease. Provisional collaborative framework for care and control of within the routine system with existing staff. Furthermore, with tuberculosis and diabetes. WHO/HTM/TB/2011.15. Geneva, Switzerland: just one day of training, clinical and nursing staff were able to fol- WHO, 2011. 16 Central TB Division, Directorate General of Health Services, Ministry of low the diagnostic algorithm and record appropriate data in the Health & Family Welfare, Government of India. Managing tuberculosis in devised formats. There were, however, some challenges. First, as the your area—a training course (module 1–4). New Delhi, India: Government NPCDCS had not been implemented in the district, the glucometer of India, 2011. http://www.tbcindia.nic.in/pdfs/Module%201%20to%204.zip/ Accessed 9 August 2013. and test strips had to be procured through external sources. Sec- 17 World Health Organization. Defi nition and diagnosis of diabetes mellitus ond, for the same reason, there was no free supply of oral hypo- and intermediate hyperglycaemia. Summary of Technical Report and Rec- glycaemic drugs, and some patients had to pay for these as out-of- ommendations. Report of a WHO/IDF consultation. Geneva, Switzerland: pocket expenses. Third, the fact that it was a rural area meant that WHO, 2006. 18 Central TB Division, Directorate General of Health Services, Ministry of only one specialised doctor was available to manage clinically com- Health & Family Welfare, Government of India. Screening of tuberculosis plicated DM-TB patients, and patients needed special counselling patients for diabetes mellitus: pilot project. Training module for RNTCP to ensure that they visited the specialist for regular follow-up. staff. New Delhi, India: Government of India, 2012. http://www.tbcindia. nic.in/pdfs/TB-DM%20Training%20Module_Final.pdf/ Accessed 9 August In conclusion, this study demonstrates the feasibility and value 2013. of screening TB patients for DM in a predominantly rural area. 19 Nayak H K, Vyas S, Solanki A, Tiwari H. Prevalence of type 2 diabetes in ur- The information from this study is also useful for national scale- ban population of Ahmedabad, Gujarat. Indian J Medical Specialities 2011; up. It shows that simple diagnostic technology should be in place 2: 101–105. 20 Goyal B N, Nigam P, Dubey A L, Joshi L D, Saxena H N. Study of the diabetic at the start of screening activities, and that DM drugs should be status in pulmonary tuberculosis. J Diabetes Assoc India 1978; 18: 191–197. available free of charge for patients to avoid out-of-pocket ex- 21 Kishore B, Nagrath S P, Mathur K S, Hazra D K, Agarwal B D. Manifest, penses; both of these issues highlight the importance of scaling chemical and latent chemical diabetes in pulmonary tuberculosis. J Assoc Physicians India 1973; 21: 875–881. up DM screening in parallel with the scale-up of the NPCDCS pro- 22 Oluboyo P O, Erasmus R T. The significance of glucose intolerance in pulmo- gramme and/or the scale-up of state health systems resources. Our nary tuberculosis. Tubercle 1990; 71: 135–138. Public Health Action Screening of TB patients for DM S33

Contexte : Unité de Tuberculose (TB) d’Anklav, Anand, Gujarat, Inde. évaluation ; 36 (6,5%) souffraient de diabète (14 dont le DM était Objectif : Déterminer dans une cohorte de patients TB, 1) la préva- déjà connu et 22 nouvellement diagnostiqués) et 39 (7%) souffraient lence du diabète (DM) et les détériorations du glucose à jeun (IFG), d’IFG. La durée médiane (IQR) avant le diagnostic a été de 5 jours (1– 2) la durée avant le diagnostic, 3) les facteurs démographiques et 17). Un âge ⩾ 35 ans a été en association avec le DM. Le NNS était cliniques associés à DM et IFG, et 4) le nombre de sujets à dépister respectivement de 25 et de 14 pour un nouveau cas de DM et d’IFG, (NNS) pour diagnostiquer un nouveau cas de DM et d’IFG. avec un NNS plus faible chez les hommes, dans le groupe d’âge Schéma : Etude descriptive : on a interrogé les patients TB enregistrés ⩾ 35 ans, dans les TB pulmonaires à frottis positif, chez les patients entre janvier et septembre 2012 au sujet d’antécédents de DM. Chez en retraitement et chez ceux fumant actuellement. ceux dont le DM n’était pas connu le glucose a été mesuré au hasard Conclusion : Ce projet pilote démontre qu’il est faisable et valable de et à jeun (FBG). On a considéré respectivement comme DM et IFG un dépister le DM chez les patients atteints de TB dans un contexte de FBG de ⩾126 mg/dl et 110–125 mg/dl. routine, ce qui entraine une identification plus précoce du DM et des Résultats : Sur 556 patients TB, 553 (99%) ont fait l’objet d’une occasions d’une meilleure prise en charge de cette co-morbidité.

Marco de referencia: La Unidad de Tuberculosis (TB) de Anklav del Resultados: Se examinaron 553 de los 556 pacientes con TB (99%); distrito de Anand, en el estado de Gujarat, en la India. se diagnosticó DM en 36 pacientes (6,5%; 14 de ellos conocían el Objetivo: Determinar en una cohorte de pacientes TB los siguientes diagnóstico y 22 fueron casos nuevos) y 39 pacientes presentaron aspectos: 1) la prevalencia de diabetes (DM) y alteración de la glu- una IFG (7%). La mediana del lapso hasta el diagnóstico de DM fue cemia en ayunas (IFG, prediabetes), 2) el lapso necesario hasta el 5 días (IQR a 17). La edad ⩾35 años se asoció con el diagnóstico de diagnóstico, 3) los factores demográficos y clínicos que se asocian DM. El NNS a fin de detectar un caso de DM fue 25 y para un caso con la DM y la IFG, y 4) el número de personas que se deben examinar de IFG fue 14. Esta cifra fue inferior en los hombres, en las personas (NNS) con el fin de diagnosticar un caso nuevo de DM y IFG. ⩾35 años de edad, en los casos de TB con baciloscopia positiva, en Méthodos: Fue este un estudio descriptivo, en el cual se interrogó a los pacientes en retratamiento de la TB y en los fumadores actuales. los pacientes registrados por TB entre enero y septiembre del 2012 Conclusión: El presente proyecto piloto pone en evidencia que es sobre sus antecedentes de DM. A los pacientes que desconocían su posible y útil practicar una detección sistemática de la TB y la DM en situación con respecto a la DM se practicaron pruebas de glucemia la práctica corriente y que con esta intervención se logra una detección casual (RBG) y glucemia en ayunas (FBG). Se consideró como predia- más temprana de la DM y se ofrecen ocasiones de mejorar el trata- betes una IFG de 110 a 125 mg/dl y se estableció el diagnóstico de miento de esta morbilidad asociada. DM con una FBG de ⩾126 mg/dl.

Public Health Action (PHA) The voice for operational research. e-ISSN 2220-8372 Published by The Union (www.theunion.org), PHA provides a platform to Editor-in-Chief: Donald A Enarson, MD, Canada fulfi l its mission, ‘Health solutions for the poor’. PHA publishes high-quality Contact: [email protected] scientifi c research that provides new knowledge to improve the accessibility, PHA website: http://www.theunion.org/index.php/en/journals/pha equity, quality and effi ciency of health systems and services. Article submission: http://mc.manuscriptcentral.com/pha InternaƟ onal Union Against Tuberculosis and Lung Disease Public Health Action Health soluƟ ons for the poor VOL 3 SUPPLEMENT 1 PUBLISHED 4 NOVEMBER 2013

Is screening for diabetes among tuberculosis patients feasible at the fi eld level? B. Naik,1 A. M. V. Kumar,2 S. Satyanarayana,2 M. D. Suryakant,3 N. M. V. Swamy,4 S. Nair,1 P. Isaakidis,5 A. D. Harries6,7

http://dx.doi.org/10.5588/pha.13.0022 and TB recommends bidirectional screening of the AFFILIATIONS 1 World Health Organization Setting: Seventeen peripheral health institutions (PHI) in two diseases.6 Kolar district (population: 0.5 million), South India. Country Office for India, A countrywide project was conducted in 2012 to New Delhi, India Objective: To assess the feasibility and results of screen- better understand the optimal screening procedures, 2 International Union ing patients with tuberculosis (TB) for diabetes mellitus Against Tuberculosis and implementation, monitoring, results and challenges Lung Disease (The Union), (DM) at peripheral level. of screening patients with TB for DM within eight South-East Asia Office, Design: From January to September 2012, all TB patients New Delhi, India tertiary health care settings and more than 60 periph- 3 State Tuberculosis Centre, were assessed for DM. Those with unknown DM status eral health institutions (PHIs). The aggregate data of Ministry of Health and were screened for the disease (free of charge) by trained Family Welfare, Govern- the project have been reported elsewhere.7,8 In this ar- ment of Karnataka, Banga- laboratory technicians at each PHI, using a glucometer sup- ticle, we present on a more individual basis the feasi- lore, India 4 District Tuberculosis plied by the national programme on a capillary blood sam- bility and results of screening TB patients for DM in ⩾ Centre, Ministry of Health ple. Those with fasting blood glucose (FBG) 126 mg/dl 17 PHIs in a district of South India, and the challenges and Family Welfare, (⩾7 mM) were diagnosed as DM-positive. Government of Karnataka, faced during implementation. Kolar, India Results: Of 362 TB patients, 358 (99%) were assessed 5 Médecins Sans Frontières, for DM and 62 (17.1%) had the diseases—53 (14.6%) Mumbai, India 6 The Union, Paris, France had a previous history of DM and 9 (2.9%) were newly METHODS 7 London School of Hygiene diagnosed. All new DM patients were enrolled into DM & Tropical Medicine, care. Higher DM prevalence was found among TB patients Design London, UK aged ⩾40 years, smokers and those with smear-positive This was a descriptive study involving the implemen- CORRESPONDENCE Balaji Naik pulmonary TB. To detect a new case of DM, the number tation of DM screening procedures for TB patients in Office of the Joint Director (TB) needed to screen (NNS) among TB patients was 40. routine programmatic settings. Lady Willingdon State TB Centre Conclusion: Screening of TB patients for DM was feasi- 4th Main Road ble and effective in a peripheral setting. The availability Setting Sampangiramanagara of trained laboratory technicians and free services at Bangalore 560 027, India Bangarpet (population 0.5 million) is one of the Tu- Tel: (+91) 93410 60148 e very PHI made the intervention feasible. The study has e-mail: drbalajinaik@ berculosis Units (TU) in the Kolar district in the south- contributed towards a national policy decision in this gmail.com ern part of India. TB diagnosis and treatment ser- regard. KEY WORDS vices are delivered in a decentralised manner at all the tuberculosis; diabetes 17 PHIs of the TU. Under India’s Revised National TB m ellitus; screening; op erational research; India Control Programme (RNTCP), a PHI is defi ned as any he world is facing a new epidemic of non- health facility with a sanctioned medical offi cer posi- Tc ommunicable diseases due to various factors tion. All presumptive TB cases are investigated for the such as urbanisation, a sedentary lifestyle, rising obe- disease at the designated microscopy centres. After sity and other lifestyle factors. In 2012, there were an being diagnosed with TB, patients receive intermit- estimated 371 million people living globally with dia- tent treatment three times a week for a period of 6– betes mellitus (DM), with numbers expected to rise to 9 months, delivered under direct observation in ac- 552 million by 2030.1 India is badly affected by the cordance with national guidelines.9 TB patients are DM epidemic, with over 60 million people estimated registered and followed up until completion of treat- to have the disease.1 ment as per national guidelines. Previous studies have shown that people with DM While facilities for DM screening are available at all have a higher risk of developing tuberculosis (TB) PHIs, treatment services are provided only by the spe- compared to those who do not have diabetes.2,3 Recent cialists at the district level hospital and are continued studies in India have also found a high prevalence of in the PHI closest to the patient. The average distance DM in TB patients.4,5 With a high DM burden in India, from the PHIs to the district hospital is 35 km. Patients routine screening of TB patients for DM would appear had to travel to the specialist hospital at their own ex- to be worthwhile at all levels of care, including periph- pense. The drugs for treating DM were available in the Received 3 May 2013 eral health facilities. A recently launched World Health general pharmacy and dispensed by the hospital phar- Accepted 28 May 2013 Organization–International Union Against Tuberculo- macist on prescription by the treating doctor. All ser- sis and Lung Disease (The Union) Framework for Col- vices for the diagnosis and treatment of DM were pro- PHA 2013; 3(S1): S34–S37 laborative Activities to reduce the dual burden of DM vided free of charge. © 2013 The Union Public Health Action Field screening for diabetes in TB patients S35

Study population determined. All variables were described as propor- ACKNOWLEDGEMENTS tions, and differences between groups were compared A workshop was convened in All TB patients registered from January to September Delhi, India, for the purpose χ2 2012 in 17 PHIs in Bangarpet TU were included in the for statistical significance using the test or Fisher’s of writing the papers that are exact test, as applicable. P values <0.05 were consid- published in this supple- study. ment. The workshop was run ered statistically significant. by the Centre for Opera- tional Research, International Diabetes mellitus screening, variables Union Against Tuberculosis and data collection Ethics approval and Lung Disease (The The protocol was reviewed and approved by The Union), Paris, France; The The methods of screening, diagnosis and referral of TB Union South-East Asia Office, patients with DM to diabetes care have been described Union Ethics Advisory Group, Paris, France. New Delhi, India; the Opera- tional Research Unit, 8 in detail elsewhere. Briefl y, TB patients were asked Médecins Sans Frontières, if they had a history of DM. Among those with un- Luxembourg; the World RESULTS Health Organization Country known DM status, a random blood glucose (RBG) test Office in India, New Delhi, was offered, followed by a fasting blood glucose (FBG) There were 362 patients registered in the TU, of whom India; the All India Institute ⩾ ⩾ of Medical Sciences, New test if the RBG was 110 mg/dl ( 6.1 mM). Blood glu- 68% were male. The median age of the cohort was Delhi, India; and ESIC cose was assessed using a glucometer supplied by the 40 years (interquartile range [IQR] 27–56), and the me- Medical College, Bangalore, dian age of females was 30 years (IQR 21–45) com- India. National Programme for Prevention and Control of Funding for the workshop Cancer, Diabetes, Cardiovascular Disease and Stroke pared to 45 years (IQR 32–60) among males. The re- and open access publication sults of screening are presented in Table 1. Of the 362 was received from the World (NPCDCS) on a capillary blood sample. Those with an Diabetes Foundation, FBG ⩾126 mg/dl (⩾7 mM) were diagnosed with DM, TB patients, 358 (99%) were assessed for DM. More Gentofte, Denmark. and those with an FBG between 110 and 125 were di- than 75% of the TB patients were found to have had Conflict of interest: none declared. agnosed with pre-diabetes (impaired fasting glucose), their RBG tested within 3 days of starting TB treatment, in accordance with national guidelines.10 and among those eligible for FBG, more than 75% had Screening was performed by PHI laboratory tech- had their FBG tested within 6 days of the RBG. Among nicians trained by NPCDCS district nodal offi cers as the 358 patients, 62 (17.1%) were found to have DM: part of the strategy adopted by the NPCDCS to screen 53 (14.6%) had a previous history of DM and 9 (2.9%) all people aged >30 years at the point of primary con- were newly diagnosed. In addition, 11 patients were di- tact with any health care facility. No additional staff agnosed with pre-diabetes. All newly diagnosed DM were deployed for this project. Variables of interest to patients were referred and enrolled into DM care. the study were sourced from the TB register and the Further analysis showed that the prevalence of TB-DM register. DM among TB patients aged ⩾40 years was signifi - cantly higher than among patients aged <40 years Data analysis and statistics (Table 2). There was a higher prevalence of DM in new Data were entered in duplicate into an EpiData soft- smear-positive pulmonary TB patients compared with ware package (EpiData entry version 3.1; EpiData As- new extra-pulmonary TB cases, and in smokers com- sociation, Odense, Denmark, http://www.epidata.dk), pared to non-smokers. the databases were compared and discrepancies re- solved by checking the original register. The propor- tion of DM cases that were ‘newly diagnosed’ was cal- TABLE 2 Prevalence of DM among TB patients according to culated. To find one additional DM case, the number different variables, Kolar District, India, January–September 2012 needed to be screened (NNS) among TB patients was TB Patients patients with DM Characteristic n n (%) P value TABLE 1 Screening patients with TB for DM at 17 peripheral health institutions of the Bangarpet Total 362 62 (17.1) Tuberculosis Unit, Kolar, India, January–September 2012 Age, years <40 173 10 (5.8) <0.01 Parameter n (%) ⩾40 189 52 (27.5) Patients registered with TB 362 Sex Patients with known diagnosis of DM 53 (14.6) Male 246 43 (17.5) 0.79 Number needed to be screened with RBG 309 Female 116 19 (16.4) Patients screened with RBG 305 (98.7) Type of TB Patients with RBG ⩾110 mg/dl (⩾6.1 mM) New smear-positive and needing to be screened with FBG 101 pulmonary TB 162 36 (22.2) <0.01 Patients screened with FBG 92 New smear-negative FBG results pulmonary TB 45 7 (15.6) 0.20 FBG <110 mg/dl (<6.1 mM) New extra-pulmonary TB 95 8 (8.4) Reference FBG ⩾110 to 125 mg/dl (⩾6.1 mM to 72 Retreatment 60 11 (18.3) 0.07 6.9 mM) 11 Current smoker* FBG ⩾126 mg/dl (⩾7 mM) 9 Yes 157 34 (21.7) 0.04 Newly diagnosed with DM* 9 (2.9) No 205 28 (13.7) Patients with known or newly diagnosed DM 62 (17.1) HIV status Patients with newly diagnosed DM enrolled Positive 22 1 (4.5) 0.15 in DM care 9 (100) Negative 340 61 (17.9)

* Percentage of screened patients with TB with a new diagnosis of DM. * Current smoker = smoked in the last month. TB = tuberculosis; DM = diabetes mellitus; RBG = random blood glu- DM = diabetes mellitus; TB = tuberculosis; HIV = human immunodeficiency cose; FBG = fasting blood glucose; mM = millimoles. virus. Public Health Action Field screening for diabetes in TB patients S36

TABLE 3 NNS to find an additional case of DM among TB patients. This higher prevalence may be attributed to selection patients, according to different variables, Kolar District, India, bias due to hospital-based sampling.12 Previous studies from other January–September 2012 South Indian states, such as Kerala and Tamil Nadu, reported a prevalence of 44% and 25% respectively.4,5 As has been reported a b a/b > New DM in earlier studies, patients aged 40 years and those with smear- Unknown among those positive pulmonary TB had a higher prevalence of DM.4 Nearly DM status screened NNS 44% of the TB patients were smokers, and DM prevalence was Characteristic n n N higher among smokers than non-smokers. As smoking and DM Total 362 9 40 are risk factors for unfavourable TB treatment outcomes and Age, years health outcomes in general, this group needs special attention <40 173 5 35 and should be linked to smoking cessation services in addition to ⩾40 189 4 47 DM care. We are planning to prospectively follow the whole co- Sex hort of this study population to examine the joint impact of DM Male 246 8 31 and smoking on TB treatment outcomes. Female 116 1 116 Among the DM cases found in our study, the proportion of Type of TB previously diagnosed DM was high, at 86%, compared to about New smear-positive pulmonary TB 162 7 23 New smear-negative pulmonary TB 45 0 NA 50% in other studies of South India.4,5 This could be attributed to New extra-pulmonary TB 95 0 NA the fact that Kolar is one of the 100 districts in the country where Retreatment 60 2 30 the NPCDCS has launched a programme for early diagnosis of Current smoker* chronic non-communicable diseases by screening all persons aged Yes 157 6 27 >30 years at the point of primary contact with any health care fa- No 205 3 68 cility.6 It is worth mentioning here that the RNTCP was quick to HIV status take advantage of the momentum of the screening programme Negative 340 8 43 Positive 22 1 22 launched by the NPCDCS and extend it to TB patients, with bene- fi ts for both national programmes. This universal screening strategy = *Current smoker smoked in the last month. was one of the reasons for the relatively high NNS found in our NNS = number needed to screen; DM = diabetes mellitus; TB = tuberculosis; NA = not applicable; HIV = human immunodeficiency virus. study. Despite this fact, several new cases of DM and pre-diabetes were still found during our screening. Studies have shown that pa- tients with pre-diabetes have a higher risk of developing type 2 DM, To detect a new case of DM, the NNS among TB patients was and they also need to be targeted for preventive interventions.13 40 (Table 3). Among the new smear-positive TB patients, the NNS Several challenges were noted in the supply chain manage- was 23 compared to 30 among previously treated TB cases. ment of glucometer strips, and some instances of strip shortages Some challenges were noted during the implementation of the and expiry were experienced during fi eld visits. The fact that labo- screening process. There was a shortage of test kits (strips for the ratory staff were using the glucometer strips after expiry indicates glucometer) for a brief period of 8 days, and the laboratory tech- a need for training. The NPCDCS is still in its infancy, and mecha- nicians used the expired strips, re-calibrating the glucometers. nisms need to be put in place to ensure uninterrupted supplies for This was later rectifi ed by the district programme manager of the smooth implementation of screening. NPCDCS, who obtained supplies of fresh batches of test strips. One of the limitations of our study is that the results obtained by the expired strips could not be excluded from the analysis, as DISCUSSION we could not ascertain the exact number of tests performed with the expired strips or on whom they were performed in that pe- This is part of the fi rst nationwide study from India reporting on riod. Another limitation is the lack of confi rmatory repeat FBG the feasibility and outcomes of routine screening of patients with testing before making a diagnosis of DM, as recommended by the TB for DM. The results of this current study suggest that it is feasi- American Diabetes Association.13 We did not do this in our study, ble to implement routine screening of TB patients for DM at the as this recommendation is not included in the national guidelines peripheral level. Near-universal screening of TB patients shows that in India. We are not aware whether this was done at the point of testing for DM is acceptable among TB patients. The availability enrolment into DM care and therefore cannot comment on the of staff trained in using glucometers supplied by the NPCDCS and implications. of free DM diagnostic and treatment services helped to make this This study contributed to a policy decision made by the intervention feasible, with no requirement for additional re- RNTCP in India to routinely offer DM screening for all TB patients. sources. While most of the TB patients underwent RBG testing A plan for recording, reporting and monitoring this important in- within 3 days of starting TB treatment, nearly 8% of those with tervention has been prepared, and a training manual for health RBG levels >110 mg/dl (6.1 mM) who were eligible for FBG did care providers has been drafted. There is now a need to monitor not return. The exact reasons for this are not clear, although it the roll-out of the intervention and lessons need to be learnt could be due to implementation problems in the initial stages of through fi eld-based operational research. It is time to convert pol- the pilot project. icy to practice and ensure implementation as per guidelines. The overall prevalence of DM among TB patients was high, at 17%, although lower than in other parts of South India.4,5 The ac- References tual prevalence could be much higher, as we used FBG to diag- 1 Unwin N, Whiting D, Guariguata L, et al., eds. IDF diabetes atlas. 5th ed. Brus- nose DM and it is well known that FBG is not as sensitive as the sels, Belgium: International Diabetes Federation, 2011. http://www.eatlas. 75 g oral glucose tolerance test (OGTT) and that it may miss cases idf.org Accessed 6 August 2013. 2 Stevenson C R, Critchley J A, Forouhi N G, et al. Diabetes and the risk of tu- 11 of impaired glucose tolerance. A hospital-based study conducted berculosis: a neglected threat to public health. Chronic Illn 2007; 3: 228– in Karnataka State reported a DM prevalence of 30% among TB 245. Public Health Action Field screening for diabetes in TB patients S37

3 Jeon C Y, Murray M B. Diabetes mellitus increases the risk of active tubercu- 9 Central TB Division. Technical and operational guidelines for tuberculosis losis: a systematic review of 13 observational studies. PLoS Med 2008; 5: control. New Delhi, India: Directorate of Health Services, Ministry of Health e152. and Family Welfare, 2005. http://tbcindia.nic.in/pdfs/Technical%20&%20 4 Balakrishnan S, Vijayan S, Nair S, et al. High diabetes prevalence among tu- Operational%20guidelines%20for%20TB%20Control.pdf Accessed 6 August berculosis cases in Kerala, India. PLoS ONE 2012; 7: e46502. 2013. 5 Viswanathan V, Kumpatla S, Aravindalochanan V, et al. Prevalence of diabe- 10 National Programme for Prevention and Control of Cancer, Diabetes, CVD tes and pre-diabetes and associated risk factors among tuberculosis patients and Stroke. Operational guidelines. New Delhi, India: Directorate General in India. PLoS ONE 2012; 7: e41367. of Health Services, Ministry of Health and Family Welfare, Government of 6 Stop TB Department, World Health Organization, International Union India. http://health.bih.nic.in/Docs/Guidelines/Guidelines-NPCDCS.pdf Ac- Against Tuberculosis and Lung Disease. Provisional collaborative framework cessed 6 August 2013. for care and control of tuberculosis and diabetes. WHO/HTM/TB/2011.15. 11 Mannucci E, Ognibene A, Sposato I, et al. Fasting plasma glucose and gly- Geneva, Switzerland: WHO, The Union, 2011. http://whqlibdoc.who.int/ cated haemoglobin in the screening of diabetes and impaired glucose toler- publications/2011/9789241502252_eng.pdf Accessed 6 August 2013. ance. Acta Diabetol 2003; 40: 181–186. 7 India Diabetes-Tuberculosis Study Group. Screening of patients with diabetes 12 Gupta S, Shenoy V P, Bairy I, Srinivasa H, Mukhopadhyay C. Diabetes melli- mellitus for tuberculosis in India. Trop Med Int Health 2013; 18: 646–654. tus and HIV as co-morbidities in tuberculosis patients of rural south India. 8 India Tuberculosis-Diabetes Study Group. Screening of patients with tuber- J Infect Public Health 2011; 4: 140–144. culosis for diabetes mellitus in India. Trop Med Int Health 2013; 18: 636– 13 American Diabetes Association. Diagnosis and classifi cation of diabetes mel- 645. litus. Diabetes Care 2006; 29 (Suppl 1): S43–S48.

Contexte : Dix-sept institutions de santé périphériques (PHI) du Dis- et 62 (17,1%) souffraient de DM ; chez 53 (14,6%) le DM était déjà trict de Kolar (population: 0,5 million d’habitants), Inde du Sud. connu antérieurement et chez 9 (2,9%) le diagnostic de DM était Objectif : Evaluer la faisabilité et les résultats du dépistage au niveau nouveau. Tous les nouveaux patients DM ont été enrôlés pour des périphérique du diabète sucré (DM) chez les patients atteints de tu- soins du DM. Une prévalence plus élevée de DM a été observée parmi berculose (TB). les patients TB âgés de ⩾40 ans, parmi les fumeurs et chez ceux at- Schéma : Au cours de la période de janvier à décembre 2012, on a teints d’une TB pulmonaire à frottis positif. Pour détecter un nouveau évalué tous les patients TB en matière de DM. Ceux dont le statut cas de DM, le nombre de personnes à dépister (NNS) chez les pa- DM était inconnu ont été dépistés pour le DM (gratuitement) par un tients TB est de 40. technicien entrainé de laboratoire dans chaque PHI sur un échantillon Conclusion : Le dépistage du DM chez les patients TB est réalisable et de sang capillaire, au moyen d’un glycomètre fourni par le pro- efficient dans un contexte périphérique. La disponibilité de tech- gramme national. Ceux dont le glucose sanguin à jeun (FBG) était niciens entrainés de laboratoire et le service gratuit dans chaque PHI ⩾126 mg/dl (⩾7 mM) ont été diagnostiqués comme DM. ont rendu l’intervention réalisable. Cette étude a contribué à une dé- Résultats : Sur 362 patients TB, 358 (99%) ont été évalués pour DM cision politique nationale dans ce domaine.

Marco de referencia: Diecisiete establecimientos sanitarios periféri- Resultados: De los 362 pacientes con TB se investigó la DM en 358 cos en el Distrito de Kolar, que cuenta con una población de medio (99%) y se estableció el diagnóstico en 62 casos (17,1%), de los millón de habitantes en el sur de la India. c uales 53 (14,6%) conocían su diagnóstico y en nueve (2,9%) se Objetivo: Evaluar la factibilidad de introducir una detección si- trató de un diagnóstico nuevo. Todos los pacientes con diagnóstico stemática de la diabetes (DM) en los pacientes con tuberculosis (TB) reciente se inscribieron en el programa de atención de la DM. La y analizar los resultados que aportaría esta iniciativa en el nivel mayor prevalencia de DM se observó en los pacientes TB a partir de periférico. los 40 años de edad, fumadores y que presentaban una TB pulmonar Métodos: Entre enero y septiembre del 2012 se investigó en todos con baciloscopia positiva. Con el fin de detectar un nuevo caso de los pacientes con TB el diagnóstico de DM. En las personas que DM fue necesario practicar el cribado en 40 pacientes TB. desconocían su situación frente a la DM, un técnico de laboratorio Conclusión: Practicar la detección sistemática de la DM en los paci- capacitado practicó el examen diagnóstico en cada establecimiento entes con diagnóstico de TB constituye una estrategia factible y eficaz (sin costo alguno), en una muestra de sangre capilar, con el uso de en un centro periférico de salud. La presencia de técnicos de labora- un glucómetro que suministraba el programa nacional. Se estableció torio capacitados y de servicios sin costo alguno en cada estableci- el diagnostico de DM en todos los pacientes con una glucemia en miento periférico hizo posible la intervención. El estudio contribuyó a ayunas ⩾126 mg/dl (⩾7 mM/dl). una decisión de política nacional en este sentido.

Public Health Action (PHA) The voice for operational research. e-ISSN 2220-8372 Published by The Union (www.theunion.org), PHA provides a platform to Editor-in-Chief: Donald A Enarson, MD, Canada fulfi l its mission, ‘Health solutions for the poor’. PHA publishes high-quality Contact: [email protected] scientifi c research that provides new knowledge to improve the accessibility, PHA website: http://www.theunion.org/index.php/en/journals/pha equity, quality and effi ciency of health systems and services. Article submission: http://mc.manuscriptcentral.com/pha InternaƟ onal Union Against Tuberculosis and Lung Disease Public Health Action Health soluƟ ons for the poor VOL 3 SUPPLEMENT 1 PUBLISHED 4 NOVEMBER 2013

High prevalence of undiagnosed diabetes among tuberculosis patients in peripheral health facilities in Kerala S. Nair,1 A. K. Kumari,1 J. Subramonianpillai,2 D. S. Shabna,2 S. M. Kumar,2 S. Balakrishnan,3 B. Naik,3 A. M. V. Kumar,4 P. Isaakidis,5 S. Satyanarayana4 http://dx.doi.org/10.5588/pha.13.0037 TB patients, with a recent study reporting a prevalence AFFILIATIONS 1 Department of Pulmonary Setting: Two tertiary care hospitals and 12 peripheral 9 of 44%. Medicine, Medical health institutions (PHIs) in Trivandrum, Kerala, India. As standardised procedures for DM screening College, Trivandrum, Objective: To determine factors associated with the prev- among TB patients for use in hospitals and TB units Thiruvananthapuram, alence of diabetes mellitus (DM) among tuberculosis (TB) India (TUs)—a sub-district level tuberculosis programme 2 Directorate of Health patients and examine differences in the proportion of management unit covering a population of 500 000, Services, Thiruvanan- new DM cases among TB patients diagnosed at tertiary thapuram, Kerala, India made up of varying numbers of peripheral health in- 3 World Health Organization care centres and PHIs. stitutions (PHIs)—were not available in India, a moni- Country Office in India, Design: A descriptive study: TB patients diagnosed dur- New Delhi, India toring tool linked to the TB registration and quarterly 4 International Union ing March–September 2012 were screened for known reporting system was developed and implemented, Against Tuberculosis and DM. Those with unknown DM status were tested for ran- Lung Disease, South-East starting in multiple centres in India in 2012. The fi nd- Asia Office, New Delhi, dom blood glucose and fasting blood glucose (FBG); FBG ings of this pilot study10 showed that the screening India ⩾ 126 mg/dl was diagnosed as new DM. 5 Médecins Sans Frontières, procedures worked well, and the overall prevalence Operational Research Unit, Results: Of 920 TB patients, 689 (72%) were male and of DM in over 8000 screened TB patients was 13%. Luxembourg the mean (standard deviation) age was 47.6 (16.4) years. Trivandrum, in Kerala, was one of the sites that imple- CORRESPONDENCE Of these, 298 (32.4%) were diabetic: 235 (26%) had pre- mented screening at both hospital and fi eld level in Sanjeev Nair viously known DM and 63 (7%) were newly diagnosed. Department of Pulmonary the TU, allowing for comparison of the yield from the Medicine During the screening at PHIs and tertiary care hospitals, different levels of screening. Medical College respectively 30/183 (16.4%) and 33/737 (4.5%) were Trivandrum 695011, India This study aimed to determine factors associated Tel: (+91) 94463 27204 newly diagnosed with DM (OR 3.71; 95%CI 2.17–6.32). with the diagnosis of DM during screening of TB pa- (+91) 47123 27204 Overall, age >50 years and pulmonary tuberculosis were e-mail: drsanjeevnair@ tients under programme conditions. Specifi c objectives gmail.com independently associated with a higher prevalence of were to determine 1) the prevalence of DM, 2) the de- diabetes. mographic and clinical factors associated with preva- KEY WORDS Conclusion: As nearly one in three TB patients had DM, TB; DM; screening; RNTCP; lence of DM, and 3) whether the yield of newly diag- Kerala we recommend that TB patients should be routinely nosed DM was different in tertiary care facilities screened for DM in Kerala. As the proportion of new DM compared to that in PHIs. was higher among TB patients diagnosed at PHIs, we would recommend that specific attention and investment be directed to PHIs. METHODS Study design This descriptive study was part of a national level im- ndia, the country with the highest number of tuber- plementation project on bidirectional TB-DM screen- Iculosis (TB) cases in the world (an estimated 2.2 mil- ing performed within the routine health services in lion cases in 2011),1 also has a very high burden of di- India in 2012.10 abetes mellitus (DM), with an estimated 63 million cases in 2012.2 In India, 15% of pulmonary tuberculo- Study setting sis (PTB) cases have been estimated to be attributable The study was conducted in Trivandrum City in Ker- to DM.3 DM also affects TB treatment outcomes and ala, a state in the south of India. In 2012, 1750 smear- may delay sputum conversion, increase case fatality positive TB patients were diagnosed and 2676 TB rates during treatment, lead to increased failure rates patients (including pulmonary smear-positive, pulmo- in non-drug-resistant cases and also increase relapse nary smear-negative and extra-pulmonary TB) were rates of TB after successful completion of treatment.4,5 registered for treatment in Trivandrum district.8 Accord- Given this scenario, it has been felt that active screen- ing to national guidelines,11 TB patients are classifi ed ing for DM in TB patients may allow identifi cation of as PTB based on sputum smear examination for acid- previously undiagnosed DM and improve TB treat- fast bacilli (AFB) or, in those who are sputum smear- ment outcomes through improved DM care. negative, by chest radiograph within 2 weeks of the Diabetes is common in Kerala, a state in South India negative result being obtained. Extra-pulmonary TB Received 27 May 2013 with a population of 34.6 million, with an estimated cases are diagnosed by the treating doctor based on a Accepted 8 July 2013 community prevalence of 16–20%.6,7 The state treated combination of radiology, histopathology, demon- 25 917 TB patients in 2012, including all types of TB.8 stration of AFB by smear or culture and/or clinical fea- PHA 2013; 3(S1): S38–S42 Kerala is known to have a high prevalence of DM in tures. All TB patients are treated with a short-course, © 2013 The Union Public Health Action DM in TB patients in Kerala S39 fully intermittent (thrice weekly), standardised treat- study tool, by the fi eld-level worker of the Revised Na- ACKNOWLEDGEMENTS The authors acknowledge ment regimen containing rifampicin throughout, de- tional TB Control Programme (RNTCP), and were the assistance and support livered under direct observation. cross-checked by the District TB Offi cer of Trivandrum, provided by the staff of the State Tuberculosis Cell, The three centres that took part in the study in the Medical Offi cer of the Medical College RNTCP cell District Tuberculosis Officers, Trivandrum included two tertiary care hospitals and a and the principal investigator. The main outcomes for the Medical College Revised National TB Control TU consisting of 12 PHIs. The two hospitals were the analysis were the number and proportion (with 95% Programme Cell and their Government Medical College Hospital, Trivandrum, confi dence interval [CI]) of TB patients with a diag- staff in the State of Kerala and the Chest Disease Hospital, Pulayanarkotta. The nosis of DM (previously known and newly diagnosed), in the process of data collection. Government Medical College Hospital is a 1916-bed, stratifi ed by age, sex, place of diagnosis and type of TB. A workshop was convened in multi-specialty teaching hospital, with a daily out- Bivariate analysis was conducted to examine whether Delhi, India, for the purpose of writing the papers that are patient load of 1800 and approximately 850 cases of those factors were associated with the prevalence of published in this supple- TB diagnosed annually. The Chest Disease Hospital, DM. Odds ratios (OR) with 95%CI were used to mea- ment. The workshop was run by the Centre for Opera- Pulayanarkotta, which caters for patients with respira- sure associations. Multivariate analysis using logistic tional Research, International tory diseases, including TB, has 508 beds and a daily regression was conducted to calculate the adjusted OR. Union Against Tuberculosis and Lung Disease ( The out-patient load of about 60; some 325 TB patients Only those variables that were statistically signifi cant Union), Paris, France; The are diagnosed and treated there every year. As TB care (P < 0.05) in bivariate analysis were considered for Union South-East Asia Office, New Delhi, India; the Opera- in India is mainly domiciliary, TB patients are referred multivariate analysis. EpiData Version 3.5.3 (EpiData tional Research Unit, to peripheral centres to continue their treatment, and Association, Odense, Denmark) was used to analyse Médecins Sans Frontières, Luxembourg; the World the hospitals provide in-patient care only for the man- the data. Health Organization Country agement of complications or co-morbidities. The fi eld- Office in India, New Delhi, Ethics approval India; the All India Institute level programme management unit that participated of Medical Sciences, New in the study was the Trivandrum DTC (district TB cen- The study was approved by the Institutional Ethics Delhi, India; and ESIC tre) TU (covering a population of 0.5 million), which Committee of Government Medical College, Trivan- Medical College, Bangalore, India. consists of 12 health care facilities, providing diagnosis drum. As this was a pilot project aiming to test the fea- Funding for the workshop of TB and directly observed treatment for TB patients. sibility of DM screening among TB patients with a view and open access publication was received from the World This unit registered 413 TB patients in 2012. to learning lessons for national scale-up approved by Diabetes Foundation, Trivandrum is the southernmost district of Kerala, RNTCP authorities, the requirement for individual in- Gentofte, Denmark. Conflict of interest: none with both a rural and an urban population. Patients are formed consent was waived by the Institutional Ethics declared. referred to the tertiary care hospitals from the PHIs or Committee. All diagnosed DM patients were linked to from private hospitals. Smaller PHIs are used by a pre- DM care. Permission to use, report and publish the dominantly rural population, whereas the tertiary care collected data in the pilot study was also obtained from hospitals receive patients from both urban and rural the International Union Against Tuberculosis and Lung areas. All of these hospitals provide free or subsidised Disease Ethics Advisory Group, Paris, France. care for patients. As a pulmonary care centre, the Chest Disease Hospital deals more with PTB, whereas the Gov- ernment Medical College Hospital has various special- TABLE 1 Demographic and clinical profile of TB patients screened and prevalence of DM among these by site, Trivandrum, Kerala, ity departments providing care for extra-pulmonary India, March–September 2012 TB patients. All patients CDH MCH DTC TU Study participants Category n (%) n (%) n (%) n (%) P value All TB patients aged ⩾15 years diagnosed consecutively Total 920 (100) 313 (100) 424 (100) 183 (100) at the two hospitals, or those registered for treatment at the Trivandrum DTC TU, between March and Sep- Age, years >50 407 (44) 173 (55) 143 (34) 91 (50) <0.01 tember 2012 made up the study population. The Chest ⩽50 513 (56) 140 (45) 281 (66) 92 (50) Disease Hospital, Pulayanarkotta, started patient re- Sex cruitment on 15 March 2012 and the Medical College Male 670 (73) 264 (84) 274 (65) 132 (72) <0.01 Hospital, Trivandrum, on 19 March 2012. DTC Trivan- Female 250 (27) 49 (16) 150 (35) 51 (28) drum TU started recruiting cases from 1 April 2012. Current smoker Yes 370 (41) 117 (38) 174 (42) 79 (41) 0.29 No 531 (59) 194 (62) 236 (58) 101 (59) Data collection and analysis Type of TB The methods of screening, recording and reporting Pulmonary 545 (59) 245 (78) 167 (39) 133 (73) <0.01 have been described in detail elsewhere.10 Briefl y, all Extra-pulmonary 375 (41) 68 (22) 261 (69) 50 (27) TB patients were asked whether they had DM, and DM those with unknown DM status underwent random Yes 298 (32) 127 (43) 117 (28) 54 (30) <0.01 blood glucose (RBG) testing. If RBG was ⩾110 mg/dl, No 622 (68) 186 (60) 307 (72) 129 (71) the patient was screened for fasting blood glucose New DM (among ⩾ DM cases only) (FBG). If FBG was 126 mg/dl, DM was diagnosed. Ve- New 63 (21) 27 (21) 6 (5) 30 (56) <0.01 nous blood samples were collected for blood sugar Previously level measurement at the participating centre. Data on diagnosed 235 (80) 100 (79) 111 (95) 24 (44) age, sex, smoking status (current smoker was defi ned TB = tuberculosis; DM = diabetes mellitus; CDH = Chest Disease Hospital; MCH = as a person who smoked in the last month), RBG and Medical College Hospital; DTC TU = district TB centre TB unit (consisting of 21 pe- FBG values were collected in the TB-DM register, the ripheral health institutions). Public Health Action DM in TB patients in Kerala S40

TABLE 2 Factors associated with DM (previously and newly pulmonary TB, females and those aged ⩽50 years. Of the 920 TB diagnosed) prevalence among TB patients, Trivandrum, Kerala, patients screened, 298 (32%) were diagnosed with DM. Of these, 63 India, March–September 2012 (21%) were newly diagnosed DM cases, whereas 235 (79%) had previously known DM. The proportion of new DM was higher TB Patients patients with DM Adjusted among TB patients diagnosed at PHIs compared to those at ter- Category n n (%) OR (95% CI) OR (95%CI)* tiary care hospitals. Factors associated with a higher prevalence of DM among the Age, years > >50 418 182 (45) 2.8 (2.1–3.7)† 2.5 (1.9, 3.3)† TB patients are shown in Table 2. Age 50 years, male sex, PTB ⩽50 530 116 (22) and current smoking were associated with higher prevalence of Sex DM on bivariate analysis. In multivariate modelling, only age Male 689 233 (34) 1.5 (1.1–2.1)† 1.0 (0.7, 1.5) >50 years and PTB were found to be independently associated Female 259 65 (25) with DM. Type of TB Factors associated with the prevalence of new (previously un- † † Pulmonary 567 216 (38) 2.3 (1.7–3.1) 1.9 (1.4, 2.7) diagnosed) DM among the TB patients are shown in Table 3. Di- Extra-pulmonary 383 82 (21) agnosis in a PHI, age >50 years and PTB were found to be inde- Current smoker pendently associated with new DM after multivariate modelling. Yes 377 144 (38) 1.6 (1.2–2.1)† 1.3 (0.96, 1.9) No 546 152 (28) Site of diagnosis DISCUSSION PHI 183 54 (30) 0.9 (0.6–1.2) Tertiary care 737 244 (33) We found a high prevalence of DM among TB patients treated in Trivandrum; it was signifi cantly higher among those aged * Adjusted for age, sex, type of TB and smoking status. > † P < 0.05. 50 years and in those with PTB. The prevalence of DM among DM = diabetes mellitus; TB = tuberculosis; OR = odds ratio; CI = confidence interval; the TB patients was higher in Trivandrum than in the other sites = PHI peripheral health institution. that took part in the pilot study (personal communication, SS): the overall prevalence of diabetes in TB patients in the nationwide pilot 10 RESULTS was 13%, whereas that in the Trivandrum centres was 32.4%. More than three quarters of the TB patients with DM knew A total of 948 TB patients were screened in all three sites. Of these, their DM status before the screening; however, this proportion 920 (97%) fi nished the entire screening process, completing all was lower among TB patients screened at PHIs. Among the TB pa- their required blood measurements. These 920 patients were in- tients screened at peripheral TU level, more than half the cases of cluded in the analysis. Of these, 670 (72%) were male and the DM diagnosed were new cases. This probably indicates that pa- mean (standard deviation) age was 47.8 (16.2) years. The median tients reaching the major hospitals have multiple opportunities (interquartile range) age was 49 (37–59) years. The profi le of the for screening before the diagnosis of TB, whereas patients diag- patients screened and the overall prevalence of DM among TB pa- nosed with TB at peripheral level are likely to miss a diagnosis of tients, desegregated by age, sex, smoking status and type of TB DM if they do not undergo active screening after TB diagnosis. and for each site, are shown in Table 1. TB patients diagnosed at The high prevalence of DM among TB patients in Trivandrum the Medical College Hospital had a higher proportion of extra- could be explained by the high prevalence of DM in the general population in Kerala, which, as reported by previous studies among the adult population, is 16–20%,6,7 and is much higher TABLE 3 Factors associated with proportion of new DM among TB than in other states in India. TB patients in Kerala are also older patients screened, Trivandrum, Kerala, India, March–September 2012 than in other Indian states, which could be another reason for Patients the high prevalence of DM in TB patients. The prevalence of TB is with newly lower in Kerala than in the rest of the country.12,13 Kerala also has TB diagnosed a lower prevalence of human immunodefi ciency virus (HIV) infec- patients DM Adjusted Category n n (%) OR (95%CI) OR (95%CI)* tion in the general population compared to other states of South India: the prevalence of HIV in the general population of India is Age, years 0.31% compared to 0.19% in Kerala.14 DM could therefore be one >50 407 41 (10.1) 2.5 (1.5– 4.3)† 2.2 (1.3, 3.8)† ⩽50 513 22 (4.3) of the main factors driving the TB epidemic in Kerala, unlike other high TB burden countries.15 Sex Male 668 48 (7.2) 1.2 (0.7–2.2) 0.98 (0.5, 2.0) A study looking at the prevalence of DM in TB patients in Ker- Female 252 15 (6) ala showed a high prevalence of DM, with 44% of TB patients Type of TB having the disease.9 However, this prevalence was determined by Pulmonary 545 50 (9.2) 2.8 (1.5–5.3)† 2.1 (1.1, 4.1)† glycated haemoglobin (HbA1c) testing, which is not feasible in Extra-pulmonary 375 13 (3.5) fi eld conditions. The use of the HbA1c measurement could also Current smoker be the reason for the higher diagnosis of DM in TB patients. This Yes 370 28 (7.6) 1.2 (0.7–1.9) 0.9 (0.5, 1.7) study also showed that almost half the DM patients diagnosed on No 531 35 (6.6) screening were unaware of their DM status at the time of screen- Site of diagnosis ing, and they would have remained undiagnosed if the screening PHI 183 30 (16.4) 4.2 (2.5–7.1)† 3.7 (2.2, 6.4)† Tertiary care 737 33 (4.5) had not been undertaken. Factors associated with a diagnosis of DM in TB patients in that study were male sex and age >50 years.

*Adjusted for age, sex, type of TB, smoking status and site of diagnosis. The strengths of this study are that a large number of TB pa- † P < 0.05. DM = diabetes mellitus; TB = tuberculosis; OR = odds ratio; CI = confidence interval; tients were consecutively screened in the study sites in Trivan- PHI = peripheral health institution. drum and are thus representative. Trivandrum was the only one Public Health Action DM in TB patients in Kerala S41 of two cities where screening was performed at both tertiary care be aware of their status; screening for diabetes is thus an urgent hospital and PHI level, allowing a comparison to be made be- requirement and a minimum standard of care for TB patients, par- tween the two different sites of screening of the yield and propor- ticularly at fi eld level. tion of new DM diagnosed in the same population. Also, the use of a screening tool developed at national level allows for a com- References parison of the proportion of DM diagnosed in Trivandrum to that 1 World Health Organization. Global tuberculosis report, 2012. WHO/HTM/ in other states in India. Possible limitations of the study include TB/2012.6. Geneva, Switzerland: WHO, 2012. the lack of information on other possible confounding factors, as 2 International Diabetes Federation. IDF diabetes atlas. 5th ed. Unwin N, the information collected was limited to that included in the Whiting D, Guariguata L, et al., eds. Brussels, Belgium: International Dia- screening tool. Only patients notifi ed to the RNCTP were included betes Federation, 2012. http://www.idf.org/diabetesatlas Accessed August 2013. in the screening. There is a possibility that some patients, particu- 3 Stevenson C R, Forouhi N G, Roglic G, et al. Diabetes and tuberculosis: the larly those with extra-pulmonary TB being treated with individu- impact of the diabetes epidemic on tuberculosis incidence. BMC Public alised regimens outside the RNTCP, were not included in the study. Health 2007; 7: 234. This study has major policy implications. First, it demonstrates 4 Baker M A, Harries A D, Jeon C Y, et al. The impact of diabetes on tuberculosis treatment outcomes: a systematic review. BMC Med 2011; 9: 81. the feasibility of screening for DM in TB patients using a standard 5 Balakrishnan S, Prema J, Sunilkumar M, et al. Diabetes mellitus increases risk screening tool. Second, it reiterates the high prevalence of DM in of failing treatment in drug-susceptible TB patients. Int J Tuberc Lung Dis TB patients in Kerala, highlighting once again the importance of 2011; 15 (Suppl): S348. DM as a risk factor for TB in Kerala. Third, it demonstrates that 6 Raman Kutty V, Joseph A, Soman C R. High prevalence of type 2 diabetes in an urban settlement in Kerala, India. Ethn Health 1999; 4: 231–239. screening for DM in TB patients at fi eld level results in a higher 7 Menon V U, Kumar K V, Gilchrist A, et al. Prevalence of known and unde- proportion of previously undiagnosed DM patients being detected, tected diabetes and associated risk factors in central Kerala—ADEPS. Diabe- which shows the importance of early screening of TB patients at tes Res Clin Pract 2006; 74: 289–294. fi eld level. The RNTCP can also support India’s non-communicable 8 Central TB Division, Directorate General of Health Services, Ministry of Health and Family Welfare. TB India 2013. Revised National TB Control Pro- disease programme by early diagnosis of DM in a group of pa- gramme annual status report. New Delhi, India: Government of India, 2013. tients who would otherwise be undiagnosed, linking them to care http://www.tbcindia.nic.in/pdfs/TB%20India%202013.pdf Accessed Septem- and with a consequent reduction in morbidity and mortality. All ber 2013. of this requires close coordination between the RNTCP and the 9 Balakrishnan S, Vijayan S, Nair S, et al. High diabetes prevalence among tuberculosis cases in Kerala, India. PLOS ONE 2012; 7: e46502. National Non-Communicable Disease Control Programme. 10 India Tuberculosis-Diabetes Study Group. Screening of patients with tuber- This study reinforces the view that DM screening in TB pa- culosis for diabetes mellitus in India. Trop Med Int Health 2013; 18: 636– tients should be part of the minimum standard of care offered to 645. all TB patients in Kerala and that it needs to be a high priority 11 Central TB Division, Directorate General of Health Services, Ministry of Health and Family Welfare. Technical and operations guidelines for tubercu- area for implementation. The state of Kerala has since taken a pol- losis control. New Delhi, India: Government of India, 2005. http://tbcindia. icy decision to screen all of its TB patients for DM. The govern- nic.in/pdfs/Technical%20&%20Operational%20guidelines%20for%20TB% ment of India subsequently also made this a policy for the rest 20Control.pdf Accessed August 2013. 12 Kumar S, Radhakrishna, Chadha V K, et al. Prevalence of tuberculous of the country. In addition, steps have been taken to link the i nfection among school children in Kerala. Indian J Tuberc 2009; 56: 10–16. RNTCP and the National Non-Communicable Disease Control 13 Chadha V K, Agarwal S P, Kumar P, et al. Annual risk of tuberculous infec- Programme at PHI level in Kerala, allowing for DM screening and tion in four defi ned zones of India: a comparative picture. Int J Tuberc Lung improved care of TB patients diagnosed with DM. Dis 2005; 9: 569–575. 14 National AIDS Control Organization, Ministry of Health and Family W elfare. State fact sheets. March 2012. New Delhi, India: Government of India, 2012. CONCLUSION http://www.nacoonline.org/NACO/Quick_Links/Publication/State_Fact_ Sheets/Fact_Sheets/State_Fact_Sheet_-_March_2012/ Accessed August 2013. This study confi rms the high prevalence of DM in TB patients in 15 Lönnroth K, Castro K, Chakaya J M, et al. Tuberculosis control 2010–2050: Kerala and that a signifi cant proportion of DM patients may not cure, care and social change. Lancet 2010; 375: 1814–1829.

Contexte : Deux hôpitaux de soins tertiaires et 12 institutions péri- étaient diabétiques ; 235 (26%) souffraient d’un DM déjà connu et phériques de santé (PHI) à Trivandrum, Kerala, Inde. 63 (7%) ont été nouvellement diagnostiqués. Dans les PHI, 30/183 Objectif : Déterminer les facteurs en association avec la prévalence (16,4%) ont été nouvellement diagnostiqués comme diabétiques au du diabète sucré (DM) chez les patients tuberculeux ( TB) et examiner cours du dépistage et dans les hôpitaux de soins tertiaires 33/737 la différence des proportions de nouveaux DM entre les patients TB (4,5%) l’ont été également (OR 3,71 ; IC 95% 2,17–6,32). Au total, diagnostiqués dans les centres de soins tertiaires et dans les PHI. un âge >50 ans et la TB pulmonaire sont en association indépen- Schéma : Etude descriptive. On a dépisté un diabète déjà connu chez dante avec une prévalence plus élevée du DM. les patients TB diagnostiqués pendant la période de mars à septembre Conclusion : La DM existe chez près d’un patient TB sur trois. Dès 2012, et chez les patients dont le statut de DM était inconnu on a lors, nous recommandons le dépistage en routine du DM chez les pa- mesuré le glucose sanguin au hasard et le glucose sanguin à jeun tients TB de Kerala. La proportion de nouveaux cas de DM étant plus (FBG) ; les sujets dont le FBG était ⩾ 126 mg/dl ont été diagnostiqués élevée chez les patients TB diagnostiqués dans les PHI, nous recom- comme nouveaux cas de DM. mandons dès lors qu’une attention et un investissement spécifiques Résultats : Sur 920 patients TB, il y a eu 689 hommes (72%). Leur soient portés à ce problème dans les PHI. âge moyen était de 47,6 années (16,4). D’entre eux 298 (32,4%)

Public Health Action DM in TB patients in Kerala S42

Marco de referencia: Dos hospitales de atención terciaria y 12 cen- 16,4 años). De estos pacientes, 298 eran diabéticos (32,4%), de ellos tros periféricos de atención de salud de Trivandrum, en el estado de 235 conocían su diagnóstico (26%) y se diagnosticaron 63 casos Kerala en la India. nuevos (7%). La proporción de casos nuevos diagnosticados en los Objetivo: Determinar los factores que se asocian con la prevalencia centros periféricos fue de 30 en 183 pacientes (16,4%) y en los hos- de diabetes (DM) en los pacientes aquejados de tuberculosis (TB) y pitales de atención terciaria fue de 33 en 737 (4,5%) (OR 3,71; examinar las diferencias en la proporción de casos nuevos de DM di- IC95% de 2,17 a 6,32). En general, una edad >50 años y la presencia agnosticados en los establecimientos de atención terciaria y en los de TB pulmonar fueron factores asociados de manera independiente centros periféricos de salud. con una mayor prevalencia de DM. Métodos: Fue este un estudio descriptivo; se interrogó a los pacien- Conclusión: Cerca de uno de cada tres pacientes TB presentaba DM; tes con diagnóstico de TB establecido entre marzo y septiembre del por esta razón se recomienda la detección sistemática de la DM en 2012 sobre el antecedente personal de DM y a los pacientes que los pacientes con diagnóstico de TB en Kerala. La proporción de casos desconocían su situación frente a la DM se practicaron pruebas de nuevos de DM fue mayor en los pacientes TB diagnosticados en los glucemia casual y glucemia en ayunas (FBG); el diagnóstico de un caso centros periféricos de atención sanitaria, por lo cual se preconiza nuevo de diabetes se definió con una FBG ⩾ 126 mg/dl. prestar una mayor atención a este aspecto y realizar inversiones espe- Resultados: De los 920 pacientes con TB, 689 eran de sexo mascu- cíficas en estos centros. lino (72%) y la mediana de la edad fue 47,6 años (desviación estándar

Public Health Action (PHA) The voice for operational research. e-ISSN 2220-8372 Published by The Union (www.theunion.org), PHA provides a platform to Editor-in-Chief: Donald A Enarson, MD, Canada fulfi l its mission, ‘Health solutions for the poor’. PHA publishes high-quality Contact: [email protected] scientifi c research that provides new knowledge to improve the accessibility, PHA website: http://www.theunion.org/index.php/en/journals/pha equity, quality and effi ciency of health systems and services. Article submission: http://mc.manuscriptcentral.com/pha InternaƟ onal Union Against Tuberculosis and Lung Disease Public Health Action Health soluƟ ons for the poor VOL 3 SUPPLEMENT 1 PUBLISHED 4 NOVEMBER 2013

Screening tuberculosis patients for diabetes in a tribal area in South India S. Achanta,1 R. R. Tekumalla,2 J. Jaju,1 C. Purad,1 R. Chepuri,1 R. Samyukta,3 S. Malhotra,4 S. B. Nagaraja,1,5 A. M. V. Kumar,6 A. D. Harries7,8 http://dx.doi.org/10.5588/pha.13.0033 showed that of 8109 TB patients who were assessed for AFFILIATIONS Setting: Ten peripheral health institutions of a tribal 1 World Health Organization DM, 1084 (13%) were found to have DM, based on t uberculosis unit, Saluru, Vizianagaram District, South Country Office in India, fasting blood glucose (FBG) measurements. Based on New Delhi, India India. 2 District TB Centre, Ministry these data, a policy decision was made by India’s Re- Objective: To assess among tuberculosis (TB) patients: of Health and Family vised National TB Control Programme (RNTCP) to im- Welfare, Government of 1) the feasibility of screening for diabetes mellitus (DM), Andhra Pradesh, plement countrywide screening of TB patients for 2) the prevalence of DM, 3) the demographic and clinical Visakhapatnam, India DM.11 One limitation of this study, however, was that 3 State TB Cell, Directorate features associated with DM, and 4) the number needed General of Health Services, it published only aggregate data from all sites, and to screen (NNS) to find one new case of DM. Ministry of Health and may have missed site-specifi c variations, and other Family Welfare, Govern- Design: Cross-sectional study: all TB patients registered socio-demographic and clinical factors affecting the ef- ment of Andhra Pradesh, from January to September 2012 were screened for DM Hyderabad, India fectiveness and feasibility of the screening programme. 4 All India Institute of using a screening questionnaire and random blood glu- We therefore analysed individual patient data and de- Medical Sciences, New cose, followed by fasting blood glucose (FBG) measure- Delhi, India scribed the effectiveness of screening all TB patients for 5 Department of Commu- ments using a glucometer. DM was diagnosed if FBG was DM in one tribal TU. nity Medicine, Employees ⩾126 mg/dl. State Insurance Corpora- The tribal area is remote and diffi cult to access due tion (ESIC) Medical Results: Of 381 patients, 374 (98%) were assessed for to poor connectivity and lack of other basic infrastruc- College, Bangalore, India DM, suggesting feasibility of screening, and 19 (5%) were 6 International Union ture. Indicators relating to literacy, economic status, found to have DM (12 were newly diagnosed and 7 had Against Tuberculosis and social status and access to health care services are poor Lung Disease ( The Union), a previous diagnosis of DM). The only characteristic asso- South-East Asia Office, among tribes compared to the general population.12 ciated with DM was age ⩾40 years. The NNS to detect a New Delhi, India The tribes are in transition from a forest-centred way 7 The Union, Paris, France new case of DM among all TB patients was 31; among 8 London School of Hygiene of life to a rural, settled farming lifestyle. Given the those aged ⩾40 years, the NNS was 20, and among cur- & Tropical Medicine, different lifestyle, more access to unprocessed, fi bre- London, UK rent smokers it was 21. rich foods, including fruit and vegetables in their diet Conclusion: Screening of TB patients for DM was fea- CORRESPONDENCE and greater daily physical activity, we hypothesized S Achanta sible and effective, and this should inform national scale- District TB Centre that the prevalence of DM would be considerably up. Other key considerations include the continued pro- Government Hospital for lower in tribal areas when compared with the rest of Chest and Communicable vision of free TB-DM screening, with co-location and the country. The specifi c objectives of the study were to Diseases integration of services. Pedda Waltair assess, among a cohort of TB patients: 1) the feasibility Visakhapatnam Andhra of screening for DM, 2) the prevalence of DM, 3) the Pradesh, India 530017 e-mail: [email protected]; demographic and clinical features associated with DM, [email protected] ndia has the highest burden of tuberculosis (TB) in and 4) the number needed to screen (NNS) to fi nd one KEY WORDS Ithe world,1 and an estimated 63 million people liv- new case of DM among TB patients.13,14 TB; bidirectional screening; ing with diabetes mellitus (DM).2 Evidence suggests DM; India; NNS that the DM population has a signifi cantly increased METHODS risk of developing active TB (two or three times higher than in those without DM).3–6 Three recently pub- Study design lished epidemiologic studies in South India in about This was a descriptive study of all TB patients attend- 1500 patients with TB found a high prevalence of DM: ing the study TU. about 25% in Tamil Nadu,7 32% in Karnataka8 and 44% in Kerala.9 Setting A landmark project conducted by the International The study was conducted in Saluru TU (a geographical Union Against Tuberculosis and Lung Disease (The area defi ned under the RNTCP as a sub-district-level Union), the World Health Organization (WHO) and the programme management unit, covering a population national authorities in India on bidirectional screen- of 250 000), with TB diagnostic and treatment services ing of TB and DM, modelled on the Collaborative being delivered through a network of primary, second- Framework for care and control of tuberculosis and di- ary and tertiary health care facilities. Saluru TU is a abetes, was implemented in eight tertiary centres and sub-division of one of the initial 30 districts identifi ed more than 60 peripheral health facilities in eight tuber- in India for piloting the roll-out of non-communicable 10 Received 22 May 2013 culosis units (TUs), including Vizianagaram. The India disease prevention and control activities through the Accepted 2 July 2013 Tuberculosis–Diabetes Study Group (ITDG) assessed the National Programme for Prevention of Cancer, Diabetes, feasibility and results of screening TB patients for DM Cardiovascular Disease and Stroke, with community- PHA 2013; 3(S1): S43–S47 with pooled data from the project sites.11 The study based screening of all individuals aged >30 years for © 2013 The Union Public Health Action TB-DM screening in a tribal TU, India S44

DM.15 The TU is termed tribal because >70% of the errors. Both databases were compared and discrepan- ACKNOWLEDGEMENTS The authors acknowledge population belong to one of the Scheduled Tribes. cies resolved by referring to the original data collec- the assistance provided by Screening and diagnosis of DM followed national tion sheets. All analyses were done using EpiData anal- the staff of the State Tuber- culosis (TB) Cell, the National guidelines, which stipulate that FBG should be used ysis software, Version 2.2.2.180. A descriptive analysis Programme for Prevention of with cut-off thresholds in line with those recommended was performed to determine DM prevalence. Compari- Cancer, Diabetes, Cardiovas- 16 ⩾ ⩾ cular Disease and Stroke, by the WHO. Briefl y, FBG 126 mg/dl ( 7 mM) indi- sons were then made between TB patients with and Peripheral Health Institution cates DM; FBG 110–125 mg/dl (6.1–6.9 mM) indicates without DM using the χ² test. Levels of signifi cance Medical Officers and their impaired fasting glucose (IFG); and FBG < 110 mg/dl were set at 5%. The NNS to diagnose an additional staff, and the staff of the Revised National Tuberculosis (<6.1 mM) is normal. case of DM is the reciprocal of the proportion of newly Control Programme of TB patients were categorised as ‘new’ and ‘previ- detected DM cases. We adhered to STROBE guidelines Saluru, Vizianagaram District of the State of Andhra ously treated’ based on their past history of TB treat- for reporting observational studies in writing this Pradesh, in the process of ment. The diagnosis of new smear-positive or new manuscript.18 data collection. The authors also express their gratitude smear-negative TB was based on quality assured smear to all the TB patients of microscopy and chest radiography, while that of extra- Ethics approval Saluru, whose participation in the study made this pulmonary TB was based on a combination of clini- Local administrative approval was obtained from the research possible. cal, radiological and histopathological evidence per District TB Centre, Vizianagaram, for conducting the A workshop was convened in Delhi, India, for the purpose 17 RNTCP guidelines. study. Confi dentiality was assured, as data collection of writing the papers that are For the screening process in TB patients, all patients sheets were maintained securely by programme staff published in this supple- ment. The workshop was run attending the TU were asked by laboratory technicians and electronic databases contained no personal identi- by the Centre for Opera- or attending nurses in the health facility whether they fi ers. Ethics approval was obtained from the Ethics Ad- tional Research, International Union Against Tuberculosis had already been diagnosed with DM. Those who con- visory Group of The Union. and Lung Disease ( The fi rmed that they had DM were referred to the nearest Union), Paris, France; The public health facility offering DM care. Per protocol, Union South-East Asia Office, RESULTS New Delhi, India; the Opera- TB patients with known DM status were not tested tional Research Unit, again. In those with no known diagnosis of DM, ran- The screening process for DM is shown in Table 1. Of Médecins Sans Frontières, Luxembourg; the World dom blood glucose (RBG) tests were performed on cap- 381 TB patients, 374 (98%) were assessed for DM and Health Organization Country illary blood using glucometers, followed by FBG test- 19 (5.1%) were found to have the disease. Of these, 7 Office in India, New Delhi, India; the All India Institute ing by the same method at the next visit if the RBG (1.9%) had a previously known diagnosis of DM and of Medical Sciences, New was ⩾110 mg/dl. If patients with TB were found to have 12 (3.2%) were newly detected based on blood glucose Delhi, India; and ESIC ⩾ Medical College, Bangalore, an FBG of 126 mg/dl, they were diagnosed as having measurements. IFG was diagnosed in 32 (8.5%) pa- India. presumptive DM and referred to DM services for defi n- tients. Of all the TB-DM cases, four known DM cases Funding for the workshop and open access publication itive diagnosis and enrolment in care. Those with RBG had already been enrolled in DM care; the remaining was received from the World of between 110 mg/dl and 126 mg/dl were considered 15 TB-DM cases were referred for DM care (3 with Diabetes Foundation, Gentofte, Denmark. to have IFG, and no specifi c action was taken, apart known DM and 12 with new DM), of whom 12 were The authors alone are from informing and counselling the patients. Screen- enrolled in care. responsible for the views expressed in this publication ing for DM and TB and all care for diagnosed cases and they do not necessarily were provided free of charge. represent the decisions or TABLE 1 DM screening among TB patients registered policies of the World Health Study population and study period under the RNTCP in Saluru TU, Vizianagaram, Organization. South India, January–September 2012 Conflict of interest: none The study was conducted from January to September declared. 2012. All TB patients registered in the TU at the 10 pe- Indicator n (%) ripheral health institutions (PHIs) formed the study A Total patients registered with TB 381 population. B Number and proportion of (A) assessed for DM as per protocol 374 (98.2) Data variables, collection and validation C Number and proportion of (B) with a known diagnosis of DM 7 (1.9) The data variables relating to the study objectives were D Patients needing to be screened for RBG 367 sourced from the TB register and an additional TB-DM E Number and proportion of (D) screened register that was developed and used to record data with RBG 367 (100) ⩾ from the screening questionnaire used for the purpose F Patients with RBG 110 mg/dl and needing to be screened for FBG 117 of the pilot study. These data were extracted to pre- G Number and proportion of (F) screened tested, structured collection sheets, which were checked with FBG 115 (98.3) for completeness and consistency by TB laboratory su- H Patients with FBG ⩾126 mg/dl (newly pervisors of the programme once a week, and by the diagnosed for DM) 12 I Patients with IFG (blood glucose levels principal investigator once a fortnight. All staff in- 110–126 mg/dl) 32 volved in data collection, data validation and data entry J Number and proportion of (B) with known were trained in performing the respective procedures and newly diagnosed DM 19 (5.1) using the study protocol and data collection sheets. K Number and proportion of ( J) referred for DM care 15 (78.9) Data entry, analysis and reporting L Number and proportion of ( J) who reached DM care 12 (63.2) The data were double-entered by two data entry opera- = = = tors into a pre-designed data entry form using EpiData DM diabetes mellitus; TB tuberculosis; RNTCP Revised National Tuberculosis Control Programme; TU = tuberculosis unit; RBG = software, Version 3.1 (EpiData Association, Odense, random blood glucose; FBG = fasting blood glucose; IFG = impaired Denmark), with inbuilt checks to minimise data entry fasting glucose. Public Health Action TB-DM screening in a tribal TU, India S45

TABLE 2 Demographic and clinical characteristics associated DISCUSSION with DM in TB patients in Vizianagaram, South India, January–September 2012 A low prevalence of diabetes has been reported among tribal and rural populations in India.19 ‘Lifestyle’ diseases are increasing TB patients steadily, however, with the gradual domestication of tribal areas with DM and the increased infl ux of tribes into the mainstream popula- (known TB patients and newly tion. Sparse data are available on the prevalence of DM among TB without DM diagnosed) patients, and little is known about the effectiveness of screening Characteristic n (%) n (%) Total P value all TB patients among the tribal population. This is to our knowl- Total 355 19 374 edge the fi rst study from India to examine the screening process for DM among TB patients in a peripheral TU in a tribal area. Age, years <40 185 (52.1) 4 (21.1) 189 <0.01 This study confi rms that, given the setting of a tribal TU in ⩾40 170 (47.8) 15 (78.9) 185 India, screening of TB patients for DM can be effectively imple- Sex mented within the existing framework of health care delivery. Male 221 (62.2) 15 (78.9) 236 0.14 Questions about previous DM diagnosis and blood tests for those Female 134 (37.7) 4 (21.1) 138 who have no known previous history of DM can be undertaken Type of TB reliably, and the recording and monitoring system was shown to New 308 (86.7) 18 (94.7) 326 0.48 work well. The overall prevalence of DM in this tribal population, Previously treated 47 (13.2) 1 (5.2) 48 both known and newly detected, was approximately 5%. This was HIV test low compared to the overall prevalence of 9% DM among TB pa- Reactive 25 (7.04) 0 25 0.62 tients in TUs found by the ITDG,11 indicating that DM prevalence Non-reactive 314 (88.4) 17 (89.5) 331 is heterogeneous across geographical areas and various health set- Unknown 16 (4.5) 2 (10.5) 18 tings in India. Nevertheless, the screening strategy in our study Smoking status Current smoker 80 (22.5) 5 (26.3) 85 0.77 detected more new DM cases than previously diagnosed DM cases. Not current smoker 275 (77.4) 14 (73.6) 289 This probably indicates a lack of awareness about DM, poor access to services for managing DM in these populations, and the poten- = = = DM diabetes mellitus; TB tuberculosis; HIV human immunodeficiency virus. tial of the screening strategy for detecting new DM cases. Similar to fi ndings by a study in China,20 nearly 10% of new cases were diagnosed as having IFG, an indicator for a future high Characteristics associated with DM are shown in Table 2. Age risk of DM or stroke.21,22 Co-existing TB disease could be responsi- >40 years was the only factor signifi cantly associated with a ble for stress-induced hyperglycaemia in some of these patients, higher prevalence of DM, and no differences were observed with but this is also probably an early indicator of a future risk of DM. respect to sex, smoking status, human immunodefi ciency virus Lifestyle modifi cation measures and health promotion strategies (HIV) status or type of TB. Of the 19 TB-DM cases, 12 were new may help to prevent DM in most of these patients,22 and this is smear-positive TB cases, 4 new smear-negative cases, 2 new extra- something to be considered for future management, as little ac- pulmonary cases and one was a previously treated extra-pulmonary tion was taken in our study apart from informing and counselling case (data not shown). the patients. All of the newly detected TB-DM patients were re- The NNS to detect a new case of DM is shown in Table 3. The ferred to DM care services in the nearest public health facilities, NNS was 31 for all patients; among those aged ⩾40 years, it was and although specifi c data were not captured in this study, the 20, and among smokers it was 21. majority of the known cases of DM were already enrolled in care. Our study found that the prevalence of DM in TB patients was signifi cantly associated with age ⩾40 years, but did not vary by TABLE 3 Number needed to screen to detect a new case sex or type of TB. HIV positivity was associated with absence of of DM among TB patients in Vizianagaram, South India, DM, and no HIV patient was found to have DM in our study, sim- January–September 2012 ilar to fi ndings elsewhere.23 A sub-analysis of the NNS to detect Patients Patients newly one new case of DM showed that it might be more cost-effective assessed diagnosed with DM to screen patients aged ⩾40 years and possibly current smokers, for DM (FBG ⩾126 mg/dl) but this requires more formal, prospective research. Characteristic n n (%) NNS* The strengths of the study are that it was implemented with- Total 374 12 (3.2) 31 out any additional resources within the existing health care system Age, years and with minimum training needs. The screening of patients was <40 189 3 (1.6) 63 well accepted in the community, with almost 100% of registered ⩾ 40 185 9 (4.9) 20 TB patients undergoing the screening process that was offered. Sex High acceptability in the community with minimum additional Male 236 10 (4.2) 24 resources deployed for the screening process suggests that the Female 138 2 (1.4) 69 study is feasible. Type of TB There are some limitations to the study. First, decisions on the New 326 12 (3.7) 27 Previously treated 48 0 NA diagnosis of DM were based on capillary blood glucose levels, as Smoking status venous blood glucose measurements were not available in the PHIs. Current smoker 85 4 (4.7) 21 However, the WHO states that capillary and venous blood sam- Not current smoker 289 8 (2.8) 36 ples should be regarded as almost identical.16 Second, the skill lev- els of health staff in correctly using this technology and the qual- * Rounded to nearest integer. DM = diabetes mellitus; TB = tuberculosis; FBG = fasting blood glucose; NNS = ity control of the glucometers were not assessed during the study. number needed to screen; NA = not applicable. Third, the measurements of RBG and FBG may be inaccurate in Public Health Action TB-DM screening in a tribal TU, India S46

diagnosing DM. Rapid swings in blood glucose levels can occur 9 Balakrishnan S, Vijayan S, Nair S, et al. High diabetes prevalence among tu- with RBG and FBG measurements.24–26 It has also been accepted berculosis cases in Kerala, India. PLOS ONE 2012; 7: e46502. 10 World Health Organization/International Union Against Tuberculosis and that FBG alone may fail to diagnose approximately 30% of cases of Lung Disease. Provisional collaborative framework for care and control of previously undiagnosed DM, relative to those assessed by a 75 g tuberculosis and diabetes. WHO/HTM/TB/2011.15. Geneva, Switzerland: oral glucose tolerance test,27,28 which may have to be performed WHO, 2011. 11 India Tuberculosis–Diabetes Study Group. Screening of patients with tuber- when indicated in certain cases to avoid delayed diagnosis of true culosis for diabetes mellitus in India. Trop Med Int Health 2013; 18: 636–645. DM. Glycosylated haemoglobin (HbA1c) provides a more stable 12 M Gopinath Reddy, K Anil Kumar. Political economy of tribal development: measure of blood glucose levels over 2 to 3 months,24–26 but this a case study of Andhra Pradesh. Working paper no. 25. Hyderabad, India: form of measurement was too expensive and was not used in the Centre for Economic and Social Studies, 2010. http://www.cess.ac.in/cess- home/ wp/WP_85.pdf Accessed August 2013. current study. Fourth, the timing of DM screening during TB treat- 13 Naik B, Kumar Mv A, Lal K, et al. HIV prevalence among persons suspected ment can also affect blood glucose readings, and studies have shown of tuberculosis: policy implications for India. J Acquir Immune Defi c Syndr that the prevalence of hyperglycaemia decreases over time during TB 2012; 59: e72–76. 29–32 14 Achanta S, Kumar A M, Nagaraja S B, et al. Feasibility and effectiveness of treatment. However, as the effect of transient hyperglycaemias provider initiated HIV testing and counseling of TB suspects in Vizianaga- on TB treatment outcomes is not fully understood, we would argue ram District, South India. PLOS ONE 2012; 7: e41378. that it is better to screen patients at the earliest opportunity so that 15 Directorate General of Health Services, Ministry of Health and Family Welfare, we can try to diagnose and treat DM early in the course of TB treat- National Programme for Prevention and Control of Cancer, Diabetes, Cardio- vascular Disease and Stroke. Operational guidelines. New Delhi, India: Gov- ment. Assessing the effect of better DM control on TB treatment out- ernment of India, 2010. http://health.bih.nic.in/Docs/Guidelines-NPCDCS. comes and on the reported risk of recurrent TB33 was beyond the pdf Accessed August 2013. scope of this study but should be the subject of future research. 16 World Health Organization. Defi nition and diagnosis of diabetes mellitus and intermediate hyperglycaemia. Summary of Technical Report and Rec- There are some key issues for policy makers to consider before ommendations. Geneva, Switzerland: WHO, 2006. http://whqlibdoc.who. implementing this strategy more widely. First, the high accept- int/publications/2006/9241594934_eng.pdf Accessed September 2013. ability of DM testing among TB patients in our study was facili- 17 Central Tuberculosis Division, Revised National Tuberculosis Control Pro- tated by the widespread free availability of DM testing services at gramme, Directorate General of Health Services, Ministry of Health and Family Welfare. Technical and operational guidelines for tuberculosis con- all 10 PHIs. Availability of free DM testing services at all peripheral trol. New Delhi, India: Government of India, 2005. health facilities is probably a prerequisite for the successful imple- 18 Von Elm E, Altman D G, Egger M, Pocock S J, Gotzsche P C, Vandenbroucke mentation of this screening strategy. Second, the total number of J P. STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting obser- DM screening tests performed will increase substantially as a re- vational studies. Bull World Health Organ 2007; 85: 867–872. sult of the roll-out of this strategy, and both national programmes 19 Gupta R, Misra A. Type 2 diabetes in India: regional disparities. Br J Diabetes need to plan for enhanced procurement and supply chain man- Vasc Dis 2007; 7: 16–19. agement. Third, it is acknowledged worldwide that timely referral 20 Li L, Lin Y, Mi F, et al. Screening of patients with tuberculosis for diabetes mellitus in China. Trop Med Int Health 2012; 17: 1294–1301. of cases for proper care and management is vital in co-morbid 21 Lee M, Saver J L, Hong K S, Song S, Chang K H, Ovbiagele B. Effect of pre- conditions such as HIV-TB,34 and the co-location and integration diabetes on future risk of stroke: meta-analysis. BMJ 2012; 344: e3564. of services is key to successful programme collaboration.35 Simi- 22 Perreault L, Pan Q, Mather K J, Watson K E, Hamman R F, Kahn S E, for the Diabetes Prevention Program Research Group. Effect of regression from pre- larly, the integration of DM and TB management services must be diabetes to normal glucose regulation on long-term reduction in diabetes considered as scale-up takes place, and this could mark the begin- risk: results from the Diabetes Prevention Program Outcomes Study. Lancet ning of strong cooperation and collaboration between communi- 2012; 379: 2243–2251. 23 Faurholt-Jepsen D, Range N, Praygod G, et al. Diabetes is a risk factor for cable and non-communicable disease programmes. pulmonary tuberculosis: a case-control study from Mwanza, Tanzania. PLOS In conclusion, our study found that the screening of TB pa- ONE 2011; 6: e24215. tients for DM was feasible and effective, and in a tribal TU where 24 International Expert Committee. International expert committee report on prevalence rates of DM might be expected to be lower than in ur- the role of the A1c assay in the diagnosis of diabetes. Diabetes Care 2009; 32: 1327–1334. ban populations, we still detected one in 20 cases of DM. Age was 25 Kumar P R, Bhansali A, Ravikiran M, et al. Utility of glycated hemoglobin in a factor signifi cantly associated with the prevalence of DM, and if diagnosing type 2 diabetes mellitus: a community-based study. J Clin resources are limited, it might be worth focusing the screening of Endocrinol Metab 2010; 95: 2832–2835. ⩾ 26 World Health Organization. Use of glycated haemoglobin (HbA1c) in the di- TB patients on those aged 40 years. agnosis of diabetes mellitus: abbreviated report of a WHO consultation. WHO/NMH/CHP/CPM/11.1. Geneva, Switzerland: WHO, 2011. http://www. who.int/cardiovascular_diseases/report-hba1c_2011_edited.pdf Accessed Au- References gust 2013. 1 World Health Organization. Global tuberculosis report 2012. WHO/HTM/ 27 Yang W, Lu J, Weng J, et al. Prevalence of diabetes among men and women TB/2012.6. Geneva, Switzerland: WHO, 2012. in China. N Engl J Med 2010; 362: 1090–1101. 2 International Diabetes Federation. IDF diabetes atlas. 5th ed, 2012 update. 28 Ramachandran A, Snehalatha C, Latha E, Vijay V. Evaluation of the use of Unwin N, Whiting D, Guariguata L, et al., eds. Brussels, Belgium: Interna- fasting plasma glucose as a new diagnostic criterion for diabetes in Asian In- tional Diabetes Federation, 2012. http://www.idf.org/diabetesatlas/5e/Update dian population. Diabetes Care 1998; 21: 666–667. 2012 Accessed August 2013. 29 Kishore B, Nagrath S P, Mathur K S, Hazra D K, Agarwal B D. Manifest, chem- 3 Stevenson C R, Critchley J A, Forouhi N G, et al. Diabetes and the risk of tu- ical and latent chemical diabetes in pulmonary tuberculosis. J Assoc berculosis: a neglected threat to public health. Chronic Illn 2007; 3: 228–245. Physicians India 1973; 21: 875–881. 4 Jeon C Y, Murray M B. Diabetes mellitus increases the risk of active tuberculo- 30 Goyal B N, Nigam P, Dubey A L, Joshi L D, Saxena H N. Study of the diabetic sis: a systematic review of 13 observational studies. PLOS Med 2008; 5: e152. status in pulmonary tuberculosis. J Diabetes Assn India 1978; 18: 191–197. 5 Dooley K E, Chaisson R E. Tuberculosis and diabetes mellitus: convergence 31 Singh M M, Biswas S K, Shah A. Impaired glucose tolerance in active pulmo- of two epidemics. Lancet Infect Dis 2009; 9: 737–746. nary tuberculosis. Indian J Tuberc 1984; 31: 118–121. 6 Ruslami R, Aarnoutse R E, Alisjahbana B, van der Ven A J A M, Van Crevel R. 32 Oluboyo P O, Erasmus R T. The signifi cance of glucose intolerance in pulmo- Implications of the global increase of diabetes for tuberculosis control and nary tuberculosis. Tubercle 1990; 71: 135–138. patient care. Trop Med Int Health 2010; 15: 1289–1299. 33 Baker M A, Harries A D, Jeon C Y, et al. The impact of diabetes on tuberculo- 7 Viswanathan V, Kumpatla S, Aravindalochanan V, et al. Prevalence of dia- sis treatment outcomes: a systematic review. BMC Med 2011; 9: 81. betes and pre-diabetes and associated risk factors among tuberculosis patients 34 Lawn S D, Wood R. Timing of antiretroviral therapy or HIV-1-associated in India. PLOS ONE 2012; 7: e41367. tuberculosis. N Eng J Med 2012; 366: 474. 8 Gupta S, Shenoy V P, Bairly I, Srinivasa H, Mukhopadhyay C. Diabetes mel- 35 Howard A A, El-Sadr W M. Integration of tuberculosis and HIV services in litus and HIV as co-morbidities in tuberculosis patients of rural South India. sub-Saharan Africa: lessons learned. Clin Infect Dis 2010; 50 (Suppl): S238– J Infect Public Health 2011; 4: 140–144. S244. Public Health Action TB-DM screening in a tribal TU, India S47

Contexte : Dix institutions périphériques de santé d’une unité tribale Resultants : Sur 381 patients, 374 (98%) ont fait l’objet d’une évalua- de tuberculose (TB), Saluru, district de Vizianagaram, Inde du Sud. tion pour DM, ce qui suggère la faisabilité du dépistage, et chez 19 Objectif : Evaluer parmi les patients TB : 1) la faisabilité du dépistage (5%) on a trouvé un DM (12 nouveaux diagnostics et 7 diagnostics de diabète (DM), 2) la prévalence du DM, 3) les caractéristiques dé- antérieurs de DM). La seule caractéristique associée avec le DM est mographiques et cliniques associées au DM, et 4) le nombre de sujets l’âge ⩾40 ans. Le NNS pour détecter un nouveau cas de DM pour à dépister (NNS) pour trouver un nouveau cas de DM. l’ensemble des patients TB est de 31 ; parmi ceux âgés de ⩾40 ans, le Schéma : Etude transversal : on a recherché le DM chez tous les pa- NNS est de 20 et chez les fumeurs actuels il est de 21. tients TB enregistrés entre juillet et septembre 2013 au moyen d’un Conclusion : Le dépistage du DM chez les patients TB est réalisable et questionnaire de dépistage et d’un prélèvement au hasard du glucose efficient, et ceci devrait influencer une extension nationale. D’autres sanguin suivi par des mesures du glucose à jeun au moyen d’un glu- considérations-clé comportent la fourniture continue d’un dépistage comètre. On a diagnostiqué le DM lorsque le glucose sanguin à jeun gratuit TB-DM avec une co-localisation et une intégration des était ⩾126 mg/dl. services.

Marco de referencia: Diez establecimientos periféricos de atención en 374 (98%), lo cual indica la factibilidad del cribado y se estableció de salud de una unidad de tuberculosis (TB) tribal de Saluru en el dis- el diagnóstico DM en 19 pacientes (5%), de los cuales 12 fueron casos trito Vizianagaram del sur de la India. nuevos y 7 personas conocían su diagnóstico. La única característica Objective: Evaluar los siguientes aspectos en los pacientes tratados asociada con el diagnóstico de DM fue la edad ⩾40 años. El NNS con por TB: 1) la factibilidad del cribado de la diabetes sacarina (DM), el fin de detectar un caso de DM en todos los pacientes con TB fue 2) la prevalencia de DM, 3) las características demográficas y clínicas 31; en los pacientes de ⩾40 años de edad, el NNS fue 20 y en los fu- que se asocian con la DM, y 4) el número de personas que se debe madores actuales fue 21. examinar (NNS) con el fin de detectar un caso nuevo de DM. Conclusión: La detección sistemática de la DM en los pacientes con Métodos: Se llevó a cabo un estudio transversal en el cual se prac- TB constituye una medida factible y eficaz; estos resultados servirán ticó una detección sistemática de la DM en todos los pacientes regi- para documentar las iniciativas nacionales de ampliación de escala de strados con TB entre enero y septiembre del 2012, mediante un cues- estas intervenciones. Otros aspectos que merecen atención especial tionario y una prueba de glucemia casual, seguida de una glucemia son la provisión continua de detección de la TB y la DM sin costo al- en ayunas, con el uso de un glucómetro. Se definió el diagnóstico de guno en un mismo centro y la integración de los servicios de atención DM con un resultado de la glucemia en ayunas ⩾126 mg/dl. de ambas enfermedades. Resultados: De los 381 pacientes tratados por TB, se investigó la DM

Public Health Action (PHA) The voice for operational research. e-ISSN 2220-8372 Published by The Union (www.theunion.org), PHA provides a platform to Editor-in-Chief: Donald A Enarson, MD, Canada fulfi l its mission, ‘Health solutions for the poor’. PHA publishes high-quality Contact: [email protected] scientifi c research that provides new knowledge to improve the accessibility, PHA website: http://www.theunion.org/index.php/en/journals/pha equity, quality and effi ciency of health systems and services. Article submission: http://mc.manuscriptcentral.com/pha InternaƟ onal Union Against Tuberculosis and Lung Disease Public Health Action Health soluƟ ons for the poor VOL 3 SUPPLEMENT 1 PUBLISHED 4 NOVEMBER 2013

Characteristics and treatment response in patients with tuberculosis and diabetes mellitus in New Delhi, India A. Khanna,1 S. Lohya,1 B. N. Sharath, 2 A. D. Harries 3,4 http://dx.doi/10.5588/pha.13.0025 15 February to 30 September 2012 were included in AFFILIATIONS Diabetes mellitus (DM) is known to increase the risk of 1 Lok Nayak Chest Clinic the study. (TB), Lok Nayak Hospital, tuberculosis (TB) and adversely affect TB treatment out- Patients were fi rst screened verbally for known DM. New Delhi, India comes. A descriptive study was carried out in registered 2 ESIC Medical College, In those with unknown DM status, blood tests were Bangalore, India TB patients screened for DM at Lok Nayak Hospital, New carried out using glucometers and test strips as previ- 3 International Union Delhi, India. Of 458 TB patients, 66 (14%) had DM. In ously described.6 Patients were diagnosed with DM if Against Tuberculosis and ⩾ Lung Disease, Paris, France those with dual disease, age 40 years, smear-positive fasting blood glucose was ⩾ 126 mg/dl (7 mmol/dl), in 4 London School of Hygiene pulmonary TB and recurrent TB were significantly more line with World Health Organization guidelines.8 Re- & Tropical Medicine, London, UK common. There was no effect of DM on TB treatment gardless of DM status, all TB patients received stan- outcomes, although there was a trend towards smear dardised TB treatment in accordance with national CORRESPONDENCE Ashwani Khanna non-conversion at 2 months. Screening for DM works guidelines,9 and were followed until the end of treat- Lok Nayak Chest Clinic (TB) well, and certain patient characteristics are more com- ment. Treatment outcomes were monitored through Lok Nayak Hospital, E-202 Greater Kailash -1 mon in those with dual disease. registers and treatment cards, and reported in standard New Delhi 110048, India fashion according to the guidelines.9 For the purpose Tel: (+91) 98 101 64127 e-mail: drashwani.khanna@ of the study, treatment outcomes were categorised as gmail.com ndia, a middle-income country, has the highest tu- successful (cured with negative sputum smear at the ACKNOWLEDGEMENTS Iberculosis (TB) burden in the world, with an esti- end of treatment, and treatment completed with no A workshop was convened in mated 2.3 million cases annually.1 The country is expe- smears performed) and other (died, lost to follow-up, Delhi, India, for the purpose of writing the papers that are riencing urbanisation, ageing and changes in lifestyle, failed treatment, and transferred out with no outcome published in this supple- and these are associated with an escalating epidemic reported). All screening and treatment were provided ment. The workshop was run by the Centre for Operational of diabetes mellitus (DM); in 2012, there were an esti- free of charge. Research, International Union mated 63 million prevalent cases.2 There is now good The data were sourced from TB treatment cards, TB Against Tuberculosis and Lung Disease (The Union), evidence that people with DM have 2–3 times the risk registers and TB-DM registers set up for the purpose of Paris, France; The Union of developing active TB compared with those who do the pilot study. Data were extracted from the registers South-East Asia Office, New 3,4 Delhi, India; the Operational not have DM. Not only may this lead to an increase into paper-based forms and double entered into Epi- Research Unit, Médecins in the TB burden, but patients with dual disease ap- Data version 3.1 (EpiData Association, Odense, Den- Sans Frontières, Luxem- mark, http://www.epidata.dk). Patients were grouped bourg; the World Health pear to have an increased frequency of adverse TB treat- Organization Country Office ment outcomes, with delayed sputum culture conver- according to whether or not they had DM, and cate- in India, New Delhi, India; sion, an increased risk of death during anti-tuberculosis gorical variables such as baseline characteristics and the All India Institute of Medical Sciences, New Delhi, treatment and an increased risk of recurrent disease treatment outcomes were compared using the χ² test, India; and ESIC Medical after successful completion of treatment. 5 with odds ratios and 95% confi dence intervals as ap- College, Bangalore, India. Funding for the workshop In 2012, a series of pilot studies was conducted in In- propriate, with levels of signifi cance set at 5%. and open access publication dia on bi-directional screening for DM and TB in a rou- Ethics approval was obtained from the institutional was received from the World Diabetes Foundation, tine setting, with aggregate data showing that screen- ethics committee of Maulana Azad Medical College, Gentofte, Denmark. ing in both directions was both feasible and effective.6,7 New Delhi, and the International Union Against Tu- Conflict of interest: none declared. No information, however, was presented on whether berculosis and Lung Disease Ethics Advisory Group, DM affects patient outcomes in a routine setting. We Paris, France. KEY WORDS tuberculosis; diabetes therefore decided to assess whether sputum smear mellitus; India; recurrent TB; conversion and treatment outcomes were affected by TB treatment outcomes DM status in a hospital setting in Delhi, India. RESULTS Of the 458 registered TB patients, 226 (49%) were male. METHODS The mean age of the patients was 32 years (inter- quartile range 20– 42). Of 66 patients (14.5%) diagnosed A descriptive study was carried out using the records with DM, 40 (60%) had a previous diagnosis. The base- and reports of India’s Revised National TB Control line characteristics, sputum smear status during treat- Programme. The setting was a chest clinic at a tertiary ment and treatment outcomes for TB patients with care teaching hospital (Lok Nayak) in New Delhi, and without DM are shown in the Table. In terms of which has been implementing TB control activities for baseline characteristics, age ⩾ 40 years, smear-positive Received 9 May 2013 the last 10 years. The clinic has a catchment popula- pulmonary TB (PTB) and recurrent TB were signifi cantly Accepted 20 June 2013 tion of 0.5 million, and was one of the collaborating more common in DM patients, while extra-pulmonary centres for the TB-DM bi-directional pilot project.6,7 TB was signifi cantly less common. There were no other PHA 2013; 3(S1): S48–S50 All adult TB patients diagnosed and registered from differences. An increased proportion of patients with © 2013 The Union Public Health Action Screening TB patients for diabetes S49

TABLE Baseline characteristics, smear conversion and treatment PTB and recurrent TB. Older age might be expected, because type outcomes of TB patients with and without diabetes registered at 2 DM disease tends to occur in older subjects. The reasons for the Lok Nayak Hospital, Delhi, India, in 2012 higher prevalence of PTB are unclear, but may be related to pul- monary microangiopathy and defi cient activation of pulmonary TB patients TB patients with DM without DM macrophages, which are reported to occur in DM and may predis- (n = 66) (n = 392) pose patients to TB.3 An increased risk of recurrent TB disease Characteristic n (%) n (%) OR (95%CI)* among DM patients has previously been reported.5 In a recent Sex prospective study in Mexico this association was confi rmed, and Male 40 (61) 186 (47) was also shown to be due to reactivation of the same strain of My- Female 26 (39) 206 (53) cobacterium tuberculosis in 80% of cases and reinfection with a dif- Age, years ferent strain in 20%.10 More work is needed in India to under- < † 40 12 (18) 315 (80) 18.4 (9.3–36) stand the reasons for recurrent TB in patients with DM. ⩾40 54 (82) 77 (20) We found no signifi cant effect of DM on sputum smear con- Disease classification version or treatment outcomes, although an increased proportion Pulmonary smear-positive 38 (57) 151 (39) 2.2 (1.3–3.7)‡ Pulmonary smear-negative 17 (26) 80 (20) of patients with DM remained sputum smear-positive at 2 months. Extra-pulmonary 11 (17) 161 (41) 0.3 (0.1–0.6)† Delays in sputum culture conversion have been reported in asso- Type of TB ciation with DM, although not consistently, and this also requires New 43 (66) 302 (77) further study.5 Previously treated 23 (34) 90 (23) 1.8 (1.0–3.1)‡ The strength of the study was that it was conducted in a pro- HIV status gramme setting with no additional resources apart from govern- Positive 0 17 (4) ment funds. Its limitations relate to any record review where data Negative 66 (100) 370 (94) Unknown 0 10 (2) may be inaccurate and, in this study, the small sample group of patients with dual disease. Smoking status Smoker 9 (14) 49 (13) In conclusion, screening of TB patients for DM worked well in Non-smoker 57 (86) 343 (87) this tertiary hospital setting. The study also identifi ed certain pa- Sputum smear conversion tient characteristics that may be more prevalent in those with DM, End of intensive phase thus allowing targeting of screening if resources are constrained. Converted 28 (74) 122 (83) Not converted 9 (24) 18 (12) References Unknown 1 (2) 7 (5) End of treatment 1 World Health Organization. Global tuberculosis report 2012. WHO/HTM/ Converted 32 (88) 127 (91) TB/2012.6. Geneva, Switzerland: WHO, 2012. Not converted 2 (6) 3 (2) 2 International Diabetes Federation. IDF diabetes atlas. 5th ed, 2012 update. Unknown 2 (6) 9 (7) Unwin N, Whiting D, Guariguata L, et al., eds. Brussels, Belgium: Interna- tional Diabetes Federation, 2012. http://www.idf.org/diabetesatlas/5e/Update Treatment outcomes 2012 Accessed 7 August 2013. Treatment success 58 (88) 362 (92) 3 Stevenson C R, Critchley J A, Forouhi N G, et al. Diabetes and the risk of tu- Other outcomes 8 (12) 30 (8) berculosis: a neglected threat to public health. Chronic Illn 2007; 3: 228– 245. * OR only shown for comparisons where there were significant differences. 4 Jeon C Y, Murray M B. Diabetes mellitus increases the risk of active tubercu- † P < 0.001. losis: a systematic review of 13 observational studies. PLOS Med 2008; 5: ‡ < P 0.05. e152. = = = = TB tuberculosis; DM diabetes mellitus; OR odds ratio; CI confidence interval; 5 Baker M A, Harries A D, Jeon C Y, et al. The impact of diabetes on tuberculosis = HIV human immunodeficiency virus. treatment outcomes: a systematic review. BMC Med 2011; 9: 81. 6 India Tuberculosis-Diabetes Study Group. Screening of patients with tuber- culosis for diabetes mellitus in India. Trop Med Int Health 2013; 18: 636– DM who had their sputum smears examined did not experience 645. smear conversion at the end of the initial phase of treatment, but 7 India Diabetes Mellitus-Tuberculosis Study Group. Screening of patients with diabetes mellitus for tuberculosis in India. Trop Med Int Health 2013; 18: this did not reach statistical signifi cance when compared with pa- 646–654. tients who did not have DM. Final treatment outcomes did not 8 World Health Organization. Defi nition and diagnosis of diabetes mellitus differ between the groups. and intermediate hyperglycaemia. Summary of Technical Report and Rec- ommendations. Geneva, Switzerland: WHO, 2006. 9 Central Tuberculosis Division, Revised National Tuberculosis Control Pro- DISCUSSION gramme. Directorate General of Health Services, Ministry of Health and Fam- ily Welfare. Technical and operational guidelines for tuberculosis control. New In this observational study, it was feasible to routinely screen all Delhi, India: Government of India, 2005. 10 Jimenez-Corona M E, Cruz-Hervert L P, Garcia-Garcia L, et al. Association of TB patients for DM. Among patients with DM, there was a higher diabetes and tuberculosis: impact on treatment and post-treatment out- prevalence of older individuals and patients with smear-positive comes. Thorax 2013; 68: 214–220.

On sait que le diabète sucré (DM) augmente le risque de tuberculose rechute de TB sont significativement plus courants. Il n’y a pas d’effet (TB) et influe de façon défavorable les résultats du traitement de la TB. du DM sur les résultats du traitement de la TB, quoiqu’il y ait une ten- On a mené une étude descriptive chez les patients TB enregistrés et dé- dance vers la non-négativation du frottis à 2 mois. Le dépistage du pistés pour DM à l’Hôpital Lok Nayak de New Delhi, Inde. Sur 458 pa- DM fonctionne correctement et certaines caractéristiques du patient tients TB, 66 (14%) souffraient de DM. Chez ceux atteints des deux sont plus courantes chez ceux atteints des deux maladies. maladies, un âge ⩾40 ans, une TB pulmonaire à frottis positif et une

Public Health Action Screening TB patients for diabetes S50

Se conoce que la presencia de diabetes sacarina (DM) aumenta el dades fue significativamente más frecuente que tuvieran una edad de riesgo de contraer la tuberculosis (TB) y ejerce un efecto desfavorable ⩾40 años, que padecieran TB pulmonar con baciloscopia positiva y sobre el desenlace del tratamiento antituberculoso. Se llevó a cabo recaídas de la TB. No se presentó ningún efecto de la DM sobre el un estudio descriptivo de los pacientes tuberculosos registrados, en desenlace terapéutico, aunque se observó una tendencia a la falta de quienes se investigó el diagnóstico de DM en el Hospital Lok Nayak conversión de la baciloscopia a los 2 meses de tratamiento. El cribado de Nueva Delhi en la India. De los 458 pacientes registrados con TB, de la DM es eficaz y algunas características son más frecuentes en los 66 presentaban DM (14%). En los pacientes con ambas enferme- pacientes que padecen ambas enfermedades.

Public Health Action (PHA) The voice for operational research. e-ISSN 2220-8372 Published by The Union (www.theunion.org), PHA provides a platform to Editor-in-Chief: Donald A Enarson, MD, Canada fulfi l its mission, ‘Health solutions for the poor’. PHA publishes high-quality Contact: [email protected] scientifi c research that provides new knowledge to improve the accessibility, PHA website: http://www.theunion.org/index.php/en/journals/pha equity, quality and effi ciency of health systems and services. Article submission: http://mc.manuscriptcentral.com/pha InternaƟ onal Union Against Tuberculosis and Lung Disease Public Health Action Health soluƟ ons for the poor VOL 3 SUPPLEMENT 1 PUBLISHED 4 NOVEMBER 2013

Diabetes mellitus and smoking among tuberculosis patients in a tertiary care centre in Karnataka, India M. V. Jali,1 V. K. Mahishale,2 M. B. Hiremath,1 S. Satyanarayana,3 A. M. V. Kumar,3 S. B. Nagaraja,4 P. Isaakidis5 http://dx.doi.org/10.5588/pha.13.0031 consecutively during the period February to Septem- AFFILIATIONS Diabetes mellitus (DM) and smoking are risk factors for 1 KLES Diabetes Centre, ber 2012 who were referred to the DOTS centre for TB KLES Dr Prabhakar Kore adverse outcomes in the treatment of tuberculosis (TB). treatment were included in the study. All eligible pa- Hospital & Medical In a tertiary care hospital at Belgaum in the South Indian tients were asked about their history of DM. If there Research Centre, Belgaum, ⩾ Karnataka, India State of Karnataka, all TB patients aged 18 years con- was no known history of DM, patients underwent ran- 2 Department of Pulmonary secutively diagnosed from February to September 2012 dom blood glucose (RBG) tests using a glucometer. Pa- Medicine, KLES Dr Prabha- were evaluated for DM and smoking. Of 307 TB patients, kar Kore Hospital & tients with RBG levels ⩾ 110 mg/dl (>6.1 mM) were Medical Research Centre, 35.5% were found to have DM, 9.8% were current smok- requested to undergo a fasting blood glucose (FBG) Belgaum, Karnataka, India ers, and 3.6% had DM and were also smokers. Measures 3 International Union test. Patients with FBG levels ⩾ 126 mg/dl (⩾7 mM) Against Tuberculosis and to assess and address both these factors need to be taken and between 110 and 125 mg/dl (6.1–7.0 mM) were Lung Disease, South-East into account during TB treatment. Asia Office, New Delhi, diagnosed with DM and IFG, respectively, and referred India for diabetes care. Simultaneously, TB patients with a 4 Employees State Insurance Corporation (ESIC) Post history of at least one episode of tobacco smoking in Graduate Institute of iabetes mellitus (DM) and tobacco smoking are the last 3 months were identifi ed as ‘current smokers’. Medical Science and Research and Model Dindependent risk factors for adverse tuberculosis The procedure for screening, recording (in a separate Hospital, Rajaji Nagar, (TB) treatment outcomes such as relapse, treatment register) and reporting (on a standard quarterly report- Bangalore, India 5 Médecins Sans Frontières, 1,2 6 failure and death. Individuals with DM who smoke ing form) has been described in detail elsewhere. Mumbai, India ⩾1 pack of cigarettes daily are at particularly high risk CORRESPONDENCE 3 Data variables, data entry and analysis of death from TB. It has been recommended that TB M V Jali patients should be routinely screened for both of these The following data were extracted from the registers: KLES Dr Prabhakar Kore Hospital & Medical Research risk factors and that, if present, they should be ad- age, sex, type of TB, RBG levels, FBG levels, and history Center dressed to improve TB treatment outcomes.3 of at least one episode of smoking in the last 3 months. Belgaum, Karnataka, India Tel: (+91) 831 2473777 India, with an annual incidence of 2.2 million TB The data were entered into a pre-structured data entry Fax: (+91) 831 2470732 cases, has the highest TB burden in the world.4 Nearly form created in EpiData Version 3.1 (EpiData Associa- e-mail: [email protected]; 63 million people (∼8% of the adult population aged tion, Odense, Denmark) and analysed. The data were medicaldirector@klehospital. org ⩾20 years) are estimated to have DM, and 120 million summarised by frequencies and proportions. are estimated to be tobacco smokers (∼14% of the adult KEY WORDS DM; TB; smoking; population).5 Previous studies have estimated the prev- Ethics approval prevalence; India alence of diabetes and smoking among TB patients in- This project protocol was reviewed and approved by dependently of each other, but none of these studies the International Union Against Tuberculosis and Lung has screened for both conditions in the same setting Disease Ethics Advisory Group. and reported on the co-existence of both of these risk 5 factors among TB patients. These data will be useful RESULTS to assess the extent of DM care and smoking cessation services required during the course of TB treatment. In Of 307 patients referred to the DOTS centre, 247 this study we report the prevalence of impaired fasting (80.5%) underwent all appropriate blood glucose tests glucose (IFG), DM and current smoking in a cohort of and screening for current smoking status. About two TB patients diagnosed at a medical college teaching thirds (64.2%) were male. Overall, 7.8% of patients had hospital in the South Indian state of Karnataka. IFG, 35.5% had DM, 9.8% were current smokers and 3.6% both had DM and were current smokers (Table). METHODS The prevalence of DM and smoking was higher among the younger age groups. Study setting KLES Dr Prabhakar Kore Hospital is a 2400-bed hospital DISCUSSION in the Belgaum district of Karnataka State in South India. TB patients diagnosed in the hospital are sent from var- Despite India’s good Revised National TB Control Pro- ious departments to the DOTS centre for TB treatment. gramme, DM and tobacco consumption have the po- tential to hamper TB control efforts. Research studies Study design, study population, screening Received 23 May 2013 published recently in South India have found a high Accepted 13 June 2013 procedures, recording and reporting prevalence of DM among TB patients of approximately This was a cross-sectional study involving record 25% in the State of Tamil Nadu and 44% in the State PHA 2013; 3(S1): S51–S53 review. All TB patients aged ⩾18 years diagnosed of Kerala.6–8 These studies have shown that in addition © 2013 The Union Public Health Action DM and smoking in TB patients S52

TABLE Prevalence of IFG, DM and smoking among a cohort of ACKNOWLEDGEMENTS A workshop was convened in TB patients in Belgaum, India, January–September 2012 Delhi, India, for the purpose of writing the papers that are Smoking DM + published in this supple- Total IFG DM only only smoking ment. The workshop was run Characteristic n n (%)* n (%)* n (%)* n (%)* by the Centre for Opera- tional Research, International Total 307 24 (7.8) 109 (35.5) 30 (9.8) 11 (3.6) Union Against Tuberculosis and Lung Disease (The Age, years Union), Paris, France; The 18–24 35 0 5 (14.3) 6 (17.1) 3 (8.6) Union South-East Asia 25–34 84 4 (4.8) 13 (15.5) 12 (14.3) 2 (2.4) Office, New Delhi, India; the 35–44 81 9 (11.1) 36 (44.4) 9 (11.1) 1 (1.2) Operational Research Unit, Médecins Sans Frontières, 45–54 42 10 (23.8) 33 (78.6) 1 (2.4) 3 (7.1) Luxembourg; the World 55–64 43 0 13 (30.2) 1 (2.3) 2 (4.7) Health Organization Country ⩾65 22 1 (4.6) 9 (40.9) 1 (4.6) 0 Office in India, New Delhi, India; the All India Institute Sex of Medical Sciences, New Male 197 17 (8.6) 73 (37.1) 30 (15.2) 11 (5.6) Delhi, India; and ESIC Female 110 7 (6.4) 36 (32.7) 0 0 Medical College, Bangalore, India. Type of TB Funding for the workshop New smear-positive 160 8 (5.0) 57 (35.6) 14 (8.8) 5 (3.1) and open access publication New smear-negative 49 3 (6.1) 29 (59.2) 2 (4.1) 1 (2.0) was received from the World New extra-pulmonary 36 3 (8.3) 9 (25.0) 3 (8.3) 1 (4.2) Diabetes Foundation, Retreatment cases 55 4 (7.3) 13 (23.6) 10 (18.2) 4 (7.3) Gentofte, Denmark. Conflict of interest: none Multidrug-resistant TB declared. (resistant to isoniazid and rifampicin) 7 6 (85.7) 1 (14.3) 1 (14.3) 0

* Row percentages with denominator in Total column. IFG = impaired fasting glucose; DM = diabetes mellitus; TB = tuberculosis. to DM, tobacco smoking needs to be addressed during we feel with hindsight that we should also have docu- TB treatment.9 Of note, all of the current smokers in the mented information on chewable forms of tobacco study were male, highlighting the cultural tradition in and used standardised defi nitions. this part of the country for women not to smoke. Independent international guidelines exist for the CONCLUSIONS AND management of co-morbid conditions, such as DM, RECOMMENDATIONS tobacco smoking, human immunodefi ciency virus infection, etc., among tuberculosis patients.10–12 De- It is feasible to screen for DM and tobacco smoking pending on the local situation, national TB pro- among TB patients in tertiary care settings. We recom- grammes in high TB burden countries need to adopt mend that TB and DM clinics and smoking cessation and incorporate these guidelines and integrate the im- services develop links towards an integrated, compre- plementation and monitoring of the management of hensive package of services that could lead to im- co-morbidities. This must begin at least in the tertiary proved care and better patient outcomes. care centres, where the resources and expertise are References available and, based on the experiences gained, imple- mentation can then be extended to the peripheral 1 Jee S H, Golub J E, Jo J, et al. Smoking and risk of tuberculosis incidence, mortality, and recurrence in South Korean men and health facilities. women. Am J Epidemiol 2009; 170: 1478–1485. The strength of this study is that we implemented 2 Restrepo B I, Fisher-Hoch S P, Crespo J G, et al. Type 2 diabetes and documented the screening of both factors within and tuberculosis in a dynamic bi-national border population. Epidemiol Infect 2007; 135: 483– 491. the routine hospital system, with no special budget 3 Reed G W, Choi H, Lee S Y, et al. Impact of diabetes and smok- allocated to support these activities. With nearly 80% ing on mortality in tuberculosis. PLOS ONE 2013; (2): e58044. of the TB patients successfully undergoing screening 4 Whiting D R, Guariguata L, Weil C, Shaw J. IDF diabetes atlas: global estimates of the prevalence of diabetes for 2011 and for both conditions, it appears that screening is feasible. 2030. Diabet Res Clin Pract 2011; 94: 311–321. About 20% were not screened, and the exact reasons 5 World Health Organization. Global tuberculosis report 2012. were not clear. As our facility is a tertiary care centre Geneva, Switzerland: WHO, 2012. http://www.who.int/tb/ publications/global_report/en/ Accessed August 2013. with many specialised services, we were able to link 6 India Tuberculosis-Diabetes Study Group. Screening of patients patients with these co-morbidities to appropriate care. with tuberculosis for diabetes mellitus in India. Trop Med Int The limitations of the study were that 1) it was im- Health 2013; 18: 636–645. 7 Vijay V, Satyavani K, Vigneswari A, et al. Prevalence of diabe- plemented in a tertiary care centre and the fi ndings tes and pre-diabetes and associated risk factors among tuber- may not be generalisable, and 2) we defi ned a current culosis patients in India. PLOS ONE 2012; 7: e41367. smoker as anyone who had smoked at least one tobacco 8 Balakrishnan S, Vijayan S, Nair S, et al. High diabetes preva- product within the last 3 months, and documented lence among tuberculosis cases in Kerala, India. PLOS ONE 2012; 7: e46502. this status. In India, tobacco products are either 9 Slama K, Chiang C-Y, Enarson D A, et al. Tobacco and tubercu- chewed or smoked—both of which are equally danger- losis: a qualitative systematic review and meta-analysis. Int J ous to personal health—and the consumption pattern Tuberc Lung Dis 2007; 11: 1049–1061. 10 World Health Organization/International Union Against of these two forms of tobacco varies widely across the Tuberculosis and Lung Disease. Provisional collaborative country. As we mentioned the integration of services, framework for care and control of tuberculosis and diabetes. Public Health Action DM and smoking in TB patients S53

WHO/HTM/TB/2011.15. Geneva, Switzerland: WHO, 2011. http://whqlibdoc. land: WHO, 2007. http://www.who.int/tobacco/resources/publications/tb_tob_ who.int/publications/2011/9789241502252_eng.pdf Accessed August 2013. control_monograph/en/ Accessed August 2013. 11 World Health Organization/International Union Against Tuberculosis and 12 World Health Organization. WHO policy on collaborative TB/HIV activities: Lung Disease. A WHO/The Union monograph on TB and tobacco control. guidelines for national programmes and other stakeholders. WHO/HTM/ Joining efforts to control two related global epidemics. Geneva, Switzer- TB/2012.1, WHO/HIV/2012.1. Geneva, Switzerland: WHO, 2012.

Le diabète sucré (DM) et le tabagisme sont des facteurs de risque de matière de DM et de tabagisme. Sur 307 patients TB, on a trouvé un mauvais résultats dans le traitement de la tuberculose (TB). Dans un DM chez 35,5%, un tabagisme actuel chez 9,8% et chez 3,6%, à la hôpital de soins tertiaires à Belgaum dans l’Etat de Karnataka en Inde fois un DM et un tabagisme actuel. Il est indispensable d’incorporer du Sud, tous les patients TB âgés de ⩾18 ans et diagnostiqués con- des mesures d’évaluation et de réaction à ces deux facteurs au cours sécutivement entre février et septembre 2012 ont été évalués en du traitement de la TB.

La diabetes sacarina (DM) y el tabaquismo representan factores de entre febrero y septiembre del 2012. De los 307 pacientes TB, se di- riesgo de alcanzar desenlaces terapéuticos desfavorables en el trata- agnosticó DM en el 35,5%, el 9,8% eran fumadores actuales y el miento de la tuberculosis (TB). En un hospital de atención terciaria de 3,6% eran diabéticos y fumadores actuales. Es necesario incorporar Belgaum, en el estado de Karnataka del sur de la India, se investigaron medidas de evaluación y de respuesta a estos dos factores durante el el tabaquismo y la DM en todos los pacientes consecutivos de tratamiento de la TB. ⩾18 años de edad, en quienes se estableció el diagnóstico de TB

Public Health Action (PHA) The voice for operational research. e-ISSN 2220-8372 Published by The Union (www.theunion.org), PHA provides a platform to Editor-in-Chief: Donald A Enarson, MD, Canada fulfi l its mission, ‘Health solutions for the poor’. PHA publishes high-quality Contact: [email protected] scientifi c research that provides new knowledge to improve the accessibility, PHA website: http://www.theunion.org/index.php/en/journals/pha equity, quality and effi ciency of health systems and services. Article submission: http://mc.manuscriptcentral.com/pha