Tuberculosis 2013 : 3 Tuberculosis comorbidity with communicable and non-communicable diseases : integrating health services and control eff orts

Recent data for the global burden of disease refl ect major demographic and lifestyle changes, leading to a rise in noncommunicable diseases. Most countries with high levels of tuberculosis face a large comorbidity burden from both non-communicable and communicable diseases. Traditional disease-specifi c approaches typically fail to recognise common features and potential synergies in integration of care, management, and control of non-communicable and communicable diseases. In resource-limited countries, the need to tackle a broader range of overlapping comorbid diseases is growing. Tuberculosis and HIV/AIDS persist as global emergencies. The lethal interaction between tuberculosis and HIV coinfection in adults, children, and pregnant women in sub-Saharan Africa exemplifi es the need for well integrated approaches to disease management and control. Furthermore, links between diabetes mellitus, smoking, alcoholism, chronic lung diseases, cancer, immunosuppressive treatment, malnutrition, and tuberculosis are well recognised. Here, we focus on interactions, synergies, and challenges of integration of tuberculosis care with management strategies for non-communicable and communicable diseases without eroding the functionality of existing national programmes for tuberculosis. The need for sustained and increased funding for these initiatives is greater than ever and requires increased political and funder commitment.


Introduction
In recent decades, the prevalence of non-com municable diseases has risen, an increase attributable to historic and projected demographic shifts of the world's population, together with urbanisation and accom panying lifestyle changes. 1 Countries of low and middle income now have a large burden of non-communicable disease, which overlaps with the unfi nished agenda of communicable diseases; the risk factors of poverty, unhealthy lifestyles, tobacco use, and alcohol misuse are common to both categories of disease. 2-6 People living with chronic communicable diseases such as tuber culosis and HIV/ AIDS are most likely to develop comorbidity with noncommunicable diseases. More over, coexisting communicable and non-communicable diseases augment the risk or eff ect of the other. Health-care systems in resourcelimited settings are poorly equipped to deal with this double burden of disease, 7 and disease-specifi c health-care approaches do not represent the most effi cient response. 8,9 Therefore, traditional approaches to global health need to be reassessed, with greater emphasis on multidisciplinary collaboration and integrated strategies.
In the 2010 Global Burden of Disease study, 2-4 the number of deaths per year (1980-2010) from 235 causes were estimated for 187 countries. In 2010, 52·8 million deaths were recorded worldwide, of which 24·9% were attributable to communicable diseases related to maternal, neonatal, and nutritional causes, a fall from 34·1% of 46·5 million deaths in 1990. 3 These data show a continued shift from communicable to non-communicable diseases and from premature deaths to years lived with disability. However, disease projections that refl ect ever decreasing rates of communicable diseases fail to appreciate their highly dynamic and unpredictable nature. Globalisation, increased popu la tion density, and environmental disruption present a fertile ground for emergence of new com municable diseases. 10 Movement of people and growing drug resistance are important factors hampering global tuberculosis control eff orts. [11][12][13][14][15] In sub-Saharan Africa, communicable diseases and maternal, neonatal, and nutritional disorders remain the dominant causes of disease burden. 3,4 Deaths from HIV/ AIDS increased from 0·30 million in 1990 to 1·5 million in 2010, and malaria mortality rose by an estimated 19·9% since 1990, to 1·17 million deaths in 2010. WHO estimates show that, in 2011, 8·7 million new cases of tuberculosis were recorded and 1·4 million people died from the disease, including 350 000 deaths associated with HIV coinfection. 16 Sub-Saharan Africa, India, and China have the highest number of tuberculosis cases, 16 and multidrug-resistant tuberculosis has become widespread in eastern Europe. 17 Here, we focus on the close relationship between tuberculosis and non-communicable disease, and tuberculosis and other communicable diseases, in terms of comorbidity, causation, and health care. We also investigate commonalities, synergies, and challenges for integration of tuberculosis care and control eff orts and make proposals for change.

Factors aff ecting comorbidity
Identifi cation of the best strategies for prevention and management of overlapping disease burdens supports progress towards attainment of the Millennium Develop mental Goals (MDGs). For example, after correction for health system variables, lower burdens of non-commu nicable diseases and HIV infection were associated with much greater progress towards MDG goals for child mortality and tuberculosis than were gains in gross domestic product. 18 Although social determinants of health are the main factors that drive the current tuberculosis pandemic, 19,20 other risk factors for Mycobacterium tuberculosis infection and progression to active tuber culosis disease are well known (panel 1). Figure 1 shows the geographical variation in burden of tuberculosis and associated comorbidities.

HIV-associated tuberculosis
Tuberculosis and HIV/AIDS are both chronic diseases and continue as global emergencies. Tuberculosis is a leading cause of death in people coinfected with HIV. 16 The lethal bidirectional interaction between these two diseases in sub-Saharan Africa exemplifi es the need for well integrated approaches to disease identifi cation, management, and control. HIV infection increases the risk of tuberculosis by 20-40-fold, and tuberculosis accounts for 25% of HIV-related deaths worldwide-the single biggest cause. 16 The HIV epidemic fuelled the threefold to fi vefold rise in tuberculosis notifi cation rates between 1990 and 2005 in sub-Saharan Africa, which bears 80% of the global burden of HIV-associated tuberculosis. Thus, health systems are generally weakest where the disease burden is greatest. In the mid-1990s, early application of WHO's global DOTS (directly observed treatment, short-course) strategy, without explicit tuberculosis-HIV collaborative activities, proved inadequate to stem the tuberculosis epidemic in settings with a high prevalence of HIV. 21 WHO published a comprehensive three-pronged strategy for collaborative tuberculosis-HIV activities in 2004, which was updated in 2012. 21,22 First, mechanisms for delivery of integrated tuberculosis and HIV services need to be established and strengthened at all levels of the health service. Second, the burden of tuberculosis in people living with HIV needs to be reduced by implementation of early combination anti retroviral therapy together with isoniazid preventive therapy, intensifi ed case-fi nding, and improved infection control. Third, the burden of HIV in patients with suspected or diagnosed tuberculosis needs to be reduced through provider-initiated HIV testing and counselling, HIV prevention services for those who test negative, and optimised case management for those who test positive.
Young children from HIV-aff ected households are at high risk of tuberculosis exposure. 23 Tuberculosis remains a major unrecognised cause of death in young children-both HIV-infected and HIV-uninfected-in endemic areas. 24,25 Young children are also at risk of HIV infection through the infected mother. 26 Active screening of household contacts to rule out tuberculosis and provide preventive treatment to vulnerable individuals, particularly young children, could be linked to active HIV case-fi nding by simple point-of-care testing. 26 Close links with maternal and child health programmes are essential to ensure universal tuberculosis and HIV screening of pregnant mothers, implementation of eff ective strategies to prevent mother-to-child HIV transmission, and early initiation of combined antiretroviral therapy in all HIV-infected infants. 27 Adverse drug reactions complicate concurrent treatment of HIV/AIDS and tuberculosis. 27 Although efavirenz-based combined antiretroviral regi mens can be used with standard tuberculosis drugs, use of protease inhibitors with rifampicin is problematic because rifampicin signifi cantly reduces con centrations in serum of protease inhibitors; use of dose-adjusted lopinavir and ritonavir with rifampicin, or substitution of rifampicin with rifabutin, is needed. 28 Study fi ndings show that early initiation of combined antiretroviral therapy in people with low CD4+ T-cell counts reduces mortality and

Key messages
• Comorbidity of tuberculosis with non-communicable diseases (eg, diabetes mellitus, malnutrition, smoking-related and alcohol-related diseases) and other communicable diseases (eg, HIV/AIDS) is prevalent in regions of the world that are highly endemic for tuberculosis; integrated and effi cient responses are required to tackle these comorbidities in resource-poor countries • Wider integration of tuberculosis control activities presents opportunities to investigate commonalities and potential synergies in management and control of communicable and non-communicable diseases, which have been overlooked by traditional vertical approaches, without sacrifi cing recent gains in health care or eroding the functionality of existing programmes • Early initiation of antiretroviral therapy should be a priority for HIV-infected individuals with newly diagnosed tuberculosis; bidirectional screening for diabetes mellitus and tuberculosis is needed • Levels of unsuspected comorbidity are high; patients with concurrent communicable or non-communicable diseases, or pregnant women attending antenatal clinics who have concurrent active or subclinical tuberculosis, are frequently overlooked • Effi cient integration of tuberculosis services should be encouraged while maintaining some vertical elements to secure essential functions, such as drug supply, monitoring, assessment, and national surveillance, to a point at which the national tuberculosis programme could be restricted to a small central team doing general oversight Series AIDS-defi ning disorders. [29][30][31] Thus, WHO recommends such regimens should be started as soon as possible and within the fi rst 8 weeks of tuberculosis treatment initiation. 32 This recommen dation requires a patientcentred approach tailored to national and local health infrastructure, such that patients with HIV-associated tuberculosis can receive treatment for both diseases at the same health facility, either in diff erent clinics or in the same clinic but always in a coordinated and integrated manner. Furthermore, in a systematic review 33 of studies looking at the eff ect of combined antiretroviral therapy on the incidence of tuberculosis in adults with HIV infection, combined regimens were associated strongly with a reduction in the incidence of tuberculosis across all CD4+ T-cell strata, and early initiation of combined antiretroviral therapy was proposed as a key component of national strategies to control HIV-associated tuberculosis.

Series
Millions of HIV-infected people are at risk of communicable lung diseases. In view of the eff ectiveness of modern combined antiretroviral therapy regimens, HIV infection has become a chronic illness. Noncommunicable diseases aff ecting the lung (such as chronic obstructive pulmonary disease [COPD]) and metabolic disorders (such as diabetes mellitus) are on the rise. Increasingly, the health-care needs of people on combined antiretroviral drugs resemble those of individuals with chronic non-communicable disorders. Since vertical programmes are diffi cult to sustain when health systems are under-resourced and strained, calls have been made to focus care at the primary level for people with chronic diseases. 34

Tuberculosis and non-communicable diseases
Non-communicable diseases were not included explicitly during formulation of the MDGs. This oversight was partly corrected when the UN published a declaration on the prevention and control of non-communicable diseases, 1 which was supported by all 193 member states. Despite intercountry heterogeneity resulting from variable demographic and socioeconomic transition, most non-communicable and communicable diseases are associated with a strong social gradient, mostly aff ecting poor and marginalised populations.
Historically, eff orts to control communicable and noncommunicable diseases had little in common and tended to emphasise diff erences rather than similarities. 8 However, despite being transmissible, management and control of chronic infectious diseases such as tuberculosis has more in common with non-communicable diseases than with acute communicable diseases (table). Synergies that could be investigated in public health eff orts include joint health promotion strategies and reciprocal screening and coordinated management programmes (panel 2). Non-communicable diseases interact adversely with tuberculosis by increasing both individual vulnerability to disease and the likelihood that the epidemic will be sustained within a population. 35 Individual vulnerability is aff ected by the number and severity of comorbid disorders, whereas the total burden of comorbidity determines vulnerability at the population level.

Diabetes mellitus and tuberculosis
The link between diabetes mellitus and tuberculosis has long been recognised, but the looming threat of these convergent global epidemics has only recently been appreciated. [35][36][37][38][39] Study fi ndings consistently show a twofold to threefold higher risk of developing tuberculosis in patients with diabetes mellitus   39 In a study from the Indian state of Tamil Nadu, nearly 50% of patients with tuberculosis had either diabetes mellitus (25·3%) or prediabetes (24·5%). 39 In a Chinese survey of 8886 registered tuberculosis cases, 1090 (12·4%) patients had diabetes mellitus and 575 (7·8%) had impaired fasting glucose. 40,41 Prevalence of diabetes mellitus was signifi cantly higher in patients with tuber culosis in urban areas (14·0%) versus rural populations (10·6%), and it was higher among people with tuberculosis than in the general population in both urban and rural areas.
Africa also has rising rates of diabetes mellitus, particularly in urban populations. 42,43 In Zambia, unsuspected tuberculosis among adult inpatients was signifi cantly associated with diabetes mellitus (odds ratio 6·6, 95% CI 1·7-25·3). 44 In Hispanic populations in the USA, diabetes mellitus has long been recognised as a factor contributing to high prevalence of tuberculosis. 45 People with diabetes also show impaired sputum conversion and cure rates on tuberculosis treatment, with increased risk of death and relapse. [46][47][48] To address the double burden of tuberculosis and diabetes mellitus and the absence of international guidelines, WHO and the International Union Against Tuberculosis and Lung Disease developed a collaborative framework for care and control of these diseases, 49 which sets outs out principles for bidirectional screening and coordinated management of the two diseases. The framework is based on data from systematic reviews on the diagnosis and management of diabetes and tuberculosis. Similarly, the International Diabetes Federation recognises the link between communicable and noncommunicable dis eases and supports the reorientation and integration of the health workforce and services to simultaneously address this double burden of disease. 50 An important limitation of dual screening for tuberculosis in patients with diabetes mellitus, and vice versa, is costeff ective ness. WHO recommends screening for tuberculosis only if the estimated prevalence of the disease in the population is more than 100 cases per 100 000 population, which implies screening around 100 people to detect one case of tuberculosis. However, good examples of integration of diabetes mellitus and tuberculosis screening come from China. 40,41 Risk of metabolic disease is increased in HIV-infected patients on lifelong combined antiretroviral therapy. In one study, 51 the incidence of diabetes rose with cumulative exposure to combined antiretroviral therapy, and stavudine and zidovudine were signifi cantly associated with diabetes risk. In a prospective study, 52 use of a protease inhibitor was associated with a threefold increased risk of incident diabetes mellitus. This fi nding is a particular concern in sub-Saharan Africa, where fi rst-line antiretroviral therapy typically uses lower cost drugs that often have more metabolic complications. 42 Even with increased availability of cheaper, less toxic, generic fi xed-drug

Malnutrition
Countries in epidemiological transition are aff ected adversely by both overnutrition and undernutrition. An individual who is either moderately to severely underweight or micronutrient-defi cient is at increased vulnerability to develop tuberculosis, 53-56 although evidence for the detrimental eff ect of milder nutrient defi ciencies is less robust. A reanalysis was done of US National Health and Nutrition Examination Survey data gathered in 1971-75 and matched to tuberculosis outcomes in 1982-92. 57 Disease rates of 24·7 per 100 000 person-years (95% CI 13·0-36·3) were recorded in normal-weight individuals. After controlling for demographic, socioeconomic, and medical characteristics, adjusted hazard ratios in people who were underweight, over weight, and obese were 12·4 (95% CI 5·8-27·0), 0·3 (0·1-0·6), and 0·20 (0·1-0·6), respectively. 57 The fact that malnutrition is very common in many parts of the world accounts for its large contribution to the attributable risk for tuberculosis. The eff ects of malnutrition, notably in utero and in early life, also take a major toll in the form of obesity, diabetes mellitus, hypertension, and heart disease in later life, 57 thus serving to reinforce the double burden of disease in poorer countries and among disadvantaged subpopulations of wealthier nations.
Rates of malnutrition, anaemia, and micronutrient defi ciencies are amplifi ed by chronic worm infestation, although the contribution is poorly quantifi ed. 58 Worm infestation might also increase vulnerability to tuberculosis via immune-mediated mechanisms, but the exact pathways await better characterisation. 59 Excessive exposure to mycotoxins (afl atoxins and fumonisins) is a particular risk factor in underdeveloped rural areas, where food is not stored in temperature-controlled and humidity-controlled environments. Fungal contamination of stored food could result in immunosuppression or various systemic adverse eff ects 60 that could increase tuberculosis vulnerability.

Tobacco, alcohol, and drug dependency and tuberculosis
Cigarette smoking, including passive smoking, has been associated consistently with an increased risk of M tuber culosis infection, subsequent disease develop ment, and poor treatment outcomes. 61,62 The mechanisms have not been elucidated fully, but airway immunity is aff ected negatively by cigarette smoke inhalation, and we are only starting to appreciate the importance of local immune responses within the lung in the pathogenesis of tuberculosis. 63,64 In China, fi ndings of a multiple risk factor modelling study assessing the eff ects of smoking and solid-fuel use on COPD, lung cancer, and tuber culosis showed that complete gradual cessation of smoking and solid-fuel use by 2033 could avoid 26 million deaths from COPD and 6·3 million deaths from lung cancer and would reduce the projected incidence of tuberculosis in 2033 by 14-52% if 80% DOTS coverage were sustained. 65 WHO and the Inter national Union Against Tuberculosis and Lung Disease have developed a framework for coordinated action on tuberculosis and tobacco control. 66 Misuse of alcohol and other addictive substances also augments the likelihood of developing tuberculosis, although separation of its independent eff ect from associated malnutrition, other comorbidities, and increased tuberculosis exposure is diffi cult within the social context. 67,68 Alcohol misuse is associated with poor treatment adherence and a signifi cantly amplifi ed risk of relapse and death during and after tuberculosis treatment. 67,68 Diagnosis and treatment of harmful drinking and alcohol misuse disorders should be part of basic clinical care for people with tuberculosis, particularly in settings with high alcohol consumption. More research is needed to assess the eff ect of such interventions on tuberculosis outcomes. 69 In a review of autopsies done on people who died with tuberculosis in London, UK, 35 of 46 cases had comorbidities, mainly hepatitis C virus and HIV infections, cancer, cardiovascular disease, and COPD. 70 Hepatitis C infection is especially prevalent in injecting drug users. Control measures for tobacco, alcohol, and drug misuse are an obvious priority for improved public health.

Chronic lung disease and tuberculosis
Chronic lung disease refl ects the cumulative eff ect of multiple lung insults, including inhalation of cigarette and solid-fuel smoke, silica, and other air pollutants, and allergens, together with recurrent chest infections. 71 Modelling the eff ects of cigarette smoking and solid-fuel exposure in China, 65 and the well-known comorbidity between pulmonary silicosis and tuberculosis, 72-74 shows how chronic lung disease increases vulnerability to tuberculosis, and the cycle is mutually reinforcing. 75 Tuberculosis, COPD, and smoking have major implications for global lung health. 73 Chronic lung disease 74 accounts for some of the substantial diff erences in life expectancy noted within Europe, emphasising the importance of integrated lung health initiatives such as the practical approach to lung health (PAL) 75 and the European respiratory roadmap. 76

Concomitant tuberculosis in immunocompromised people
Tuberculosis is seen increasingly in patients with medical conditions and those who are receiving treat ment that compromises the immune system. 77-81 WHO estimates Series that nearly 2 billion people globally have latent M tuberculosis infection, 79 and these individuals are at risk of developing active tuberculosis disease when the body's immune system is perturbed. Apart from immunosuppression due to HIV, any immunocomprom ised state increases susceptibility to either development of active tuberculosis after infection or reactivation of latent infection. These include physiological (very young and elderly people), pathological (malignant dis ease), therapeutic (immunosuppressive treatment), and chronic disease states (panel 1). Targeted biological agents that block tumour necrosis factor α have transformed treatment of chronic infl ammatory, rheumatological, and autoimmune diseases. However, reactivation of latent M tuberculosis infection in patients using these drugs is leading increasingly to comorbidity of active tuberculosis and non-commu nicable diseases. 78 Although this occurrence is not a major public health issue, it has important implications for management of individual patients in countries with low endemicity for tuberculosis and a large migrant population from highly endemic regions harbouring latent infection. 81

Cancer and tuberculosis associations
A growing number of bacterial and parasitic infections are associated with development of cancer. 82 Three main interactions have been noted between tuberculosis and cancer. First, tuberculosis increases lung cancer risk. 83 Second, cancers can promote reactivation of latent M tuberculosis infection. 84 Third, immunosuppression attributable to cancer treatment can reactivate latent tuberculosis infection. 85 In a com parison of 4480 Taiwanese adults with newly diagnosed tuberculosis and more than 700 000 patients without tuberculosis, 84 after adjustment for COPD and smoking-related cancers, lung cancer incidence was 11 times higher in people with tuberculosis. Signifi cant risk factors were oral, nasopharyngeal, oesophageal, lung, and haematological cancers (eg, non-Hodgkin lymph oma and leukaemia).

Subclinical and undiagnosed tuberculosis
One of the challenges at busy referral centres with a high turnover of patients, in countries highly endemic for tuberculosis, is that the immediate focus of the admitting doctor is to deal with the acute medical or obstetric condition that necessitated admission. Many patients with communicable or non-communicable diseases who could have concurrent active overt or subclinical tuberculosis are overlooked. High levels of unsuspected tuberculosis comorbidity with other non-communicable and communicable diseases have been reported among general internal medicine inpatients. In a study from Zambia, 44 all patients who could produce a sputum sample were tested for tuber culosis, irrespective of their reason for admission. The presence of unsuspected active pulmonary tuberculosis indicated a large number of undiagnosed cases in the community. Tuberculosis screening services are needed at all points of care for noncommunicable and com municable diseases.

Maternal and child health and tuberculosis
Several maternal illnesses and lifestyle factors during pregnancy (eg, smoking, malnutrition, substance misuse) can aff ect the health of an unborn child and contribute to poor intrauterine growth and low birthweight, subsequently increasing the risk of noncommunicable disease in adulthood. Babies born to malnourished or anaemic mothers are at increased risk of perinatal death and are likely to have higher rates of metabolic disease in adulthood. The in-utero and nutritional environment is important because it establishes susceptibility to metabolic diseases in later life. Intrauterine growth restriction followed by accelerated postnatal growth is associated with obesity, type 2 diabetes, and other features of the metabolic syndrome. 86 Pregnant women and young children are more susceptible to development of active tuberculosis than other groups. 87 At all levels of health care in sub-Saharan Africa, diagnosis of tuberculosis in women is easily overlooked. Furthermore, the low sensitivity of sputum smear microscopy and the time and infrastructure constraints of mycobacterial culture mean that not all tuberculosis cases are identifi ed. In sub-Saharan Africa, tuberculosis causes 15-34% of non-obstetric maternal deaths. [88][89][90] Maternal and childhood tuberculosis are linked epidemio logically. [91][92][93] In HIV-infected women, maternal tuber culosis was linked independently to increased risk of mother-to-child-transmission of HIV. Risk of culture-confi rmed tuberculosis is more than 20-fold higher in HIV-infected versus HIV-uninfected infants 94 living in tuberculosis-endemic areas.
In paediatric autopsy studies, 24,95 tuberculosis was a major cause of death, although comorbidity with other lung infections was common. Tuberculosis can be easily over looked in women, particularly those who are HIVinfected, who access routine maternal and child health services, because non-specifi c tuberculosis symp toms such as fatigue or night sweats are common in pregnancy, childbirth, and postnatally. 87 Antenatal and postnatal clinics should screen mothers and children for relevant communicable and non-communicable diseases.

Other associations between tuberculosis and disease
Links between cytomegalovirus, Chlamydia pneu monia, M tuberculosis, and non-communicable diseases such as coronary artery atherosclerosis and diabetes have been studied and debated intensively. 96,97 Although tuberculosis has been linked to coronary artery atherosclerosis and acute myocardial infarction, 98 a causative link has yet to be established. Blood-borne and sexually transmitted infections, including viral hepatitis, are relevant to tuberculosis risk, since aff ected patients could also have Series HIV or latent M tuberculosis infection. Integrated models of care for tuberculosis, HIV, and sexually transmitted infections have particular relevance in settings in which overlapping disease burdens aff ect the same sexually active population. 99 Guidelines for prevention and control of tuberculosis, HIV infection, viral hepatitis, and sexually transmitted infections, and for people using illicit drugs, are available. [99][100][101] Despite a striking decline in recent years, malaria remains a major cause of chronic anaemia in tropical areas, particularly aff ecting young children and pregnant women. Diabetes mellitus in adulthood has also been associated with an increased risk of malaria. 102 Induction of the hepatic cytochrome P450 enzyme system by rifampicin results in a halving of the concentrations in serum of artemisinin and its derivatives, complicating the concurrent management of tuberculosis and acute malaria. 103 Leprosy can present as multisystem illness at any noncommunicable disease clinic. 104 It remains a major challenge in isolated pockets within the Asia-Pacifi c, Africa, and South America regions. Although control eff orts are combined in many countries, programme integration is rarely optimised. Active house-to-house case-fi nding eff orts to identify minimally symptomatic leprosy patients in some Pacifi c island nations presents an opportunity for expanded interventions, including child hood immunisation eff orts and screening for tuberculosis, diabetes mellitus, and hypertension.
Data from the 1918 infl uenza epidemic show a clear association between increased risk for tuberculosisrelated disease and death during this period. 105 Infl uenza and other acute respiratory infections are likely to augment vulnerability to tuberculosis 106 and could partly account for some of the seasonal variability noted in temperate regions. Findings of a randomised controlled trial to assess the eff ect of pneumococcal vaccination showed decreased rates of culture-confi rmed and clinically diagnosed tuberculosis in vaccinated children, indicating the intriguing possibility that pneumococcal coinfection might increase children's vulnerability to develop active tuberculosis. 107

Community vulnerability
More than 90% of immune-competent individuals who acquire M tuberculosis infection never progress to active tuberculosis, 23 indicating a high level of herd immunity (fi gure 2). Control eff orts focus mainly on early identifi cation, isolation, and treatment of active disease as a means of interrupting the transmission cycle. Although this work makes an important contribution to reduce transmission within communities, it has been inadequate to eliminate the continuing epidemic. 108,109 Tuberculosis provides a barometer of poverty and deprivation, since its incidence is driven by socioeconomic factors, poor access to and delivery of health services, inconsistent treatment practices, HIV, and migration from countries that are highly endemic for tuberculosis.
In the WHO European region, researchers investigated the prospective association of a nation's wealth, level of egalitarianism, migration rate, health-related lifestyle, and social capital with tuber culosis incidence and prevalence over a 10-year period (2000-09). 110 Nearly 50% of tuberculosis variation was accounted for by a nation's wealth and level of egalitarianism. National income levels per person and income inequality are important predictors for tuber culosis incidence and prevalence in the WHO European region. 111 How can vulnerability be reduced at the population level? Development of an eff ective vaccine is proving especially challenging in view of the fi ne immune balance that is needed to contain M tuberculosis infec tion. 63,64 A complementary strategy would be to reduce factors that increase vulnerability at the community level, such as malnutrition, cigarette smoking, indoor air pollution, diabetes mellitus, HIV-related immune compromise, and alcohol misuse (fi gure 2). 112,113 Tuberculosis and social determinants of health Calls for unconventional approaches to control tuberculosis are not new. 114 Biomedical approaches to curing tuberculosis, which fail to account for social circumstances that predispose people to disease, will have limited eff ect. More recently, a shift has taken place in the policy discourse towards a social determinants model of health, [111][112][113][114][115] which recognises that health inequalities arise directly from the unequal distribution of resources at a global, national, and local level 115 and are shaped by the social structures and economic  Series environments in which people are "born, grow, live, work, and age". 116 This observation is especially apparent with respect to diff erences in life expectancy, as high as 29 years between countries across the world (eg, Botswana and Japan) 117 and 17 years between the richest and poorest parts of London, UK. 118 These diff erences cannot be attributed to biology alone. Susceptibility to disease and the ability to access health care and sustain a course of treatment is also related to inequalities of wealth, power, and status, and yet many public health interventions aim to address individual lifestyle factors rather than the underlying causes of health inequalities.
Accessible health care is an important aspect of tuberculosis case-fi nding, yet the inverse care law 119 suggests the availability of good health care is inversely related to the needs of the population served. If people in greatest need are less likely to access appropriate health care, this factor will have important implications for the organisation and delivery of tuberculosis services for socially and economically marginalised populations. At an operational level, responsive outreach services delivered at scale and integrated into existing health systems, with the ability to address the health and social care needs of specifi c populations in distinct contexts, are likely to entail mixed economies of care, 108 which might raise concerns about fragmentation of services.
Although integrated delivery of care for communicable and non-communicable diseases has several advantages, evidence is scanty for its eff ectiveness 120 and how best to achieve it. 121 Vertical service delivery models target very specifi c health outcomes, frequently diverting resources away from comprehensive primary care services that are already constrained. Adopting a so-called diagonal approach to scaling up primary women-centred health care-since many diseases aff ecting women are interrelated-has been suggested. 122 Service confi gurations based on MDGs 4, 5, and 6 (actions on maternal and child mor tality, HIV, malaria, and tuberculosis) could incorporate services for non-communicable diseases (heart disease, obesity, smoking, diabetes,

Panel 3: Priorities for integration of interventions for communicable and non-communicable diseases
Series cancer), which are the leading causes of death in women aged 15-44 years. However, services organised around maternal health could disadvantage older people and women of non-child-bearing age and those with specifi c non-com municable diseases. Models of service provision, which off er outreach and social support through networks of formal health, social care, and third-sector organisations coordinated by outreach and link working staff , have proved popular when clinical caseloads comprise patients with complex health and social care needs, including homelessness and substance and alcohol misuse in high-income settings. 123 Evidence supports use of mobile radiography in combination with models of care that aim to ameliorate the social risk factors associated with non-adherence to treatment. 124 Such models can be provided as a centralised service, working across areas or integrated into individual tuberculosis clinics. A more contextualised approach to risk management and targeting of resources might be needed. Study fi ndings on the adverse health outcomes of crisis populations 125 suggest a need to attend to the type of risk environments 126 that contribute to excess risk of tuberculosis, such as overcrowding in urban environments, poor nutrition, previous health status before displacement, and dis persed populations, all of which make identifi cation, monitoring, and treatment of tuberculosis and other associated communicable and non-communicable diseases diffi cult. A lack of specifi city can serve to homogenise risk based on group identity (eg, asylum seekers) and to stigmatise those very groups.
Action across the social gradient, which is proportionate to the level of disadvantage, is needed for both preventative and curative services. 112 In relation to noncommunicable diseases, perspective is needed that links the eff ect of early lifestyle experiences on later adult life and highlights the cumulative eff ects of disadvantage throughout life. Thus, structural interventions are needed outside formal health-care sectors, 127 such as programmes for urban regeneration, job creation, and social protection initiatives. In this current climate of austerity, fi nding ways to fund such initiatives in countries of low income and with a high disease burden is diffi cult, particularly when substantial political commitment and redistri bution of resources is needed. Although health services remain largely focused on curative services, further investment in eff ective prevention measures remains important because these will be more cost eff ective in the long run.

Conclusions
In the current poor global economic climate, we must avoid destructive competition between communicable and non-communicable diseases for the limited funds available for health services. Integrated solutions should seek to unlock potential commonalities and synergies and optimise scarce resources. Effi cient integration of tuberculosis services with those for non-communicable and other communicable diseases should be a priority (panel 3) while maintaining some vertical elements to secure essential functions, such as drug supply, monitoring and assessment, and national surveillance, to a point at which a small central team could oversee the national tuberculosis programme. Appropriate balance should be realised, based on the best solution in a particular situation, and should not be an issue of integration or not, or horizontal versus vertical. Focus on local ownership is especially important, as is increasing equitable domestic investment in universal health coverage and sustainable health services. Rates of tuberculosis are a useful indication of socioeconomic inequality and deprivation. Meeting the social and medical care needs of vulnerable com munities needs a cohesive response from many groups, including the private sector. 128 Prevention and management of chronic communicable and non-communicable diseases is a litmus test for healthsystems strengthening. 120 Integrated strategies should be designed to build capacity and encourage local ownership to develop contextualised solutions to key health challenges (panel 3). The need for continued and increased funding is greater than ever before. Countries of low and middle income need to augment international political and fi nancial commitment to tackle these global health issues with increased mobilisation of domestic resources.

Contributors
AZu and MS initiated the article. AZu and BM developed the initial, subsequent, and fi nal drafts of this article. All authors contributed to the writing and fi nalisation of the article.

Confl icts of interest
We declare that we have no confl icts of interest.

Search strategy and selection criteria
We searched PubMed and Google Scholar (Jan 1, 1995, to Dec 31, 2012) and the Cochrane Library and Embase (Jan 1, 2001, to Dec 31, 2012) with the terms: "tuberculosis", "Mycobacterium tuberculosis", "Communicable Diseases", "Non-Communicable Diseases", "Global Burden of Disease", "Diabetes", "COPD", "HIV", "malnutrition", "chronic disease", "chronic lung disease", "smoking", "immunosuppression", and "social determinants". We added to this strategy by searching publications on tuberculosis from the WHO STOP TB department and the International Union Against Tuberculosis and Lung Disease. We also reviewed studies cited by reports identifi ed by this search strategy and selected those we judged relevant. Some review articles are cited to provide readers with more detail and references than our report can accommodate. We only included articles published in English.