Trilateral Overlap Between Tuberculosis, Diabetes and HIV-1 in a High Burden African 5 6 Setting: Implications for TB Control
Total Page:16
File Type:pdf, Size:1020Kb
Page 89 of 114 European Respiratory Journal 1 2 3 4 Trilateral overlap between tuberculosis, diabetes and HIV-1 in a high burden African 5 6 setting: Implications for TB control. 7 8 9 10 11 Tolu Oni 1,2 MD(Res), Natacha Berkowitz 1,2 MBChB, Mmamapudi Kubjane 1 MPH, Rene 12 13 Goliath 2 BSc(Hons), Naomi S Levitt**3 MD, Robert J Wilkinson**2,4,5 PhD. 14 15 16 1. Division of Public Health Medicine, School of Public Health & Family Medicine, 17 18 19 University of Cape Town, Cape Town 7925, South Africa. 20 21 2. Wellcome Centre for Infectious Disease Research in Africa, Institute of Infectious Disease 22 23 24 and Molecular Medicine, University of Cape Town, Cape Town 7925, South Africa. 25 26 27 3. Division of Diabetes and Endocrinology, Department of Medicine, Groote Schuur 28 29 Hospital, Cape Town, 7925, South Africa and Chronic Disease Initiative for Africa. 30 31 32 4. The Francis Crick Institute, London NW1 2AT, United Kingdom. 33 34 5. Department of Medicine, Impreial College London W2 1PG, United Kingdom. 35 36 37 38 39 40 *Corresponding author: Associate Professor Tolu Oni, Room 4.41, Entrance 5, Level 4, 41 42 Falmouth building, Faculty of Health Sciences, University of Cape Town, Anzio road, 43 44 45 Observatory, Cape Town 7925. Tel: +27 21 650 1299. Email: [email protected] 46 47 48 **Joint Senior Authors 49 50 Key words: tuberculosis (TB), diabetes mellitus (DM), impaired glucose regulation (IGR), 51 52 53 determinants, HIV-1, sub-Saharan Africa, South Africa. 54 55 56 [Word count, Body:] 3000 words 57 58 59 60 1 European Respiratory Journal Page 90 of 114 1 2 3 4 5 6 Take home message 7 8 9 DM/TB association is significant in HIV-infected, but more accurate glycaemia markers at 10 11 TB diagnosis are needed. 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 2 Page 91 of 114 European Respiratory Journal 1 2 3 Abstract 4 5 Background 6 7 8 The diabetes (DM) burden is growing in countries where tuberculosis and HIV-1 remain 9 10 major challenges, threatening tuberculosis control efforts. This study determined the 11 12 13 association between tuberculosis and diabetes / impaired glucose regulation (IGR) in the 14 15 context of HIV-1. 16 17 18 Methods 19 20 A cross-sectional study was conducted at a TB clinic in Cape Town. Participants were 21 22 23 screened for DM and IGR, using fasting plasma glucose, oral glucose tolerance test and 24 25 HbA1c. 26 27 28 Results 29 30 31 414 TB and 438 non-TB participants were enrolled. In multivariable analysis, diabetes was 32 33 associated with tuberculosis (Odds ratio: 2.4; 95% CI 1.3 – 4.3 p=0.005); with 14% 34 35 population attributable risk fraction. but this association varied by diagnostic test (driven by 36 37 HbA1c). The association remained significant in HIV-1-infected (Odds ratio: 2.4; 95% CI 38 39 40 1.1 – 5.2; p=0.030). A high prevalence of IGR (65.2% amongst tuberculosis cases) and a 41 42 significant association with tuberculosis (Odds ratio: 2.3; 95% CI: 1.6 – 3.3; p<0.001) was 43 44 also found. 45 46 Conclusions 47 48 49 Diabetes and IGR prevalence was high and associated with tuberculosis, particularly in HIV- 50 51 1-infected persons, highlighting the importance of DM screening. The variation in findings 52 53 54 by diagnostic test highlights the need for better glycaemia markers to inform screening in the 55 56 context of tuberculosis and HIV-1. 57 58 59 60 3 European Respiratory Journal Page 92 of 114 1 2 3 Introduction 4 5 Tuberculosis (TB) remains a health challenge in most low and middle-income countries 6 7 (LMIC) including sub-Saharan Africa (SSA). Poverty, smoking, alcoholism, human 8 9 10 immunodeficiency virus (HIV-1) and diabetes mellitus (DM) are known drivers of the TB 11 12 epidemic.[1] DM triples the risk of TB;[2] and is associated with adverse outcomes.[3,4] The 13 14 TB/DM association is increasingly important in LMIC experiencing a growing burden of 15 16 non-communicable disease. According to the 2013 Global Burden of Disease Study, HIV-1, 17 18 19 TB and DM rank first, third and ninth among the top ten causes of life years lost in South 20 21 Africa (SA).[5] 22 23 24 The prevalence of DM among TB patients varies by geographic region and background DM 25 26 population in the general population, with a higher prevalence reported in South Asian 27 28 (range: 25·3-44%)[6-10] and Latin American (range: 14-39%)[11,12] populations. In SSA, 29 30 DM prevalence in TB patients ranges from 8·5% to 16·4%,[13-16] although a large 31 32 33 proportion of DM in SSA remains undiagnosed.[17] Of note, most studies have used fasting 34 35 or random blood glucose to diagnose DM, which may underestimate the true prevalence. A 36 37 study in SA estimated the general population prevalence of DM and impaired glucose 38 39 tolerance to be 12·3% and 11·2% respectively.[18] 40 41 42 The increased risk of TB associated with DM is well documented. However, this risk is 43 44 variable dependent on the population studied. A systematic review that reported a 3-fold 45 46 47 increased risk of TB in DM patients[2] did not include studies from SSA where there are 48 49 competing TB risk factors such as HIV-1. Recent studies in African settings showed 50 51 increased odds of TB among DM patients with the exception of one study,[19] possibly due 52 53 to the low background prevalence of DM. A Ugandan study found HIV-1-infected TB 54 55 patients to have reduced odds of DM compared to HIV-1 uninfected,[20] and a Tanzanian 56 57 58 study showed a stronger TB/DM association observed among HIV-1-uninfected (OR = 4.23, 59 60 4 Page 93 of 114 European Respiratory Journal 1 2 3 95% CI 1.54 – 11.57) compared to HIV-1-infected (OR = 0.14, 95% CI 0.01 – 1.81) 4 5 individuals.[21] The reasons for this apparent incongruity in the associations between TB, 6 7 DM and HIV-1 are unclear as there are only a limited number of published studies.[22] 8 9 10 11 12 13 The only SA study that assessed the prevalence of DM, using the oral glucose tolerance test, 14 15 among TB patients was conducted in 1980 when SA was 42% urbanised (compared to 62% 16 17 18 urbanised today);[16] and reported a 2·1% prevalence. SA has the largest antiretroviral 19 th 20 therapy (ART) program globally,[23] and is ranked as having the 6 highest TB incidence 21 22 globally.[24] Given this HIV-1/TB burden, reducing the TB burden in South Africa is a 23 24 priority in the global fight against TB. The strong association between HIV-1 and TB, 25 26 27 dysglycaemic effects of some ART drugs, especially protease inhibitors and nonnucleoside 28 29 reverse transcriptase inhibitors ,[25, 26] and the emerging DM epidemic highlight the 30 31 importance of investigating the association between TB, DM, and HIV-1 in this setting, the 32 33 results of which will have implications for TB control strategies. This study investigated the 34 35 prevalence of DM and impaired glucose regulation (IGR), the association between TB and 36 37 38 DM/IGR; and the population attributable risk of TB due to DM in South Africa. 39 40 41 42 43 Methods 44 45 Study setting and population 46 47 In South Africa, TB patients in the public sector are largely screened and treated in TB 48 49 50 clinics. This study was conducted at the largest TB clinic in Khayelitsha, a peri-urban 51 52 township of 390 000 predominantly black Africans, in Cape Town, Western Cape province. 53 54 In this province, DM, HIV-1 and TB rank first, second and third leading causes of death, 55 56 respectively.[27] The 2012 HIV-1 antenatal prevalence in Khayelitsha was 34% (95% CI 57 58 59 60 5 European Respiratory Journal Page 94 of 114 1 2 3 31.0-36.6%) (unpublished data from the 2012 Western Cape Department of Health Antenatal 4 5 Survey) and the 2015 TB case notification rate was 917 per 100 000 population (Virginia de 6 7 Azevedo, personal communication), with a 60% HIV-1/TB co-infection rate.[28] 8 9 10 Study design and sampling 11 12 13 We conducted a cross-sectional study on consecutive patients with respiratory symptoms 14 15 presenting to the clinic from July 2013-August 2015. Patients were eligible if they provided 16 17 consent, were 18 years or older, and had not received more than 48 hours of TB 18 19 chemotherapy. Those who were critically ill and in need of emergency clinical care were 20 21 ineligible as they were too physically unwell to give informed consent. The study was 22 23 24 approved by the University of Cape Town Human Research Ethics Committee (HREC REF: 25 26 403/2011). Assuming a 3-fold higher DM prevalence in HIV-1-uninfected TB cases, 7% DM 27 28 prevalence in HIV-1-uninfected, 70% of TB cases HIV-1 co-infected and 80% power, the 29 30 study aimed to recruit 400 TB and 400 non-TB participants.