Table 2. Baseline characteristics of 17 case–control and 3 cross–sectional studies that reported on the association between TB and DM and that were included in the meta-analyses.
| First author, year | Country | Study period | Study location | Study population | DM ascertainment | TB ascertainment | TB cases | Controls | Adjusted effect size OR (95% CI) |
TB incidence/ 100,000 p–y1 | Matched/Adjusted variables |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Case-control | |||||||||||
| Alisjahbana et al (2006) [10] | Indonesia | March, 2001– March, 2005 | Central Jakarta | Cases: TB-patients aged >15 years from outpatients TB-clinics. Controls: TB-free individuals from TB cases communities | DM ascertained by FBG ≥126 mg/dl after stopping taking anti-diabetic agents for 48 hours & FBG were considered impaired for >110 and <126 mg/dl, in accordance with WHO criteria | Pulmonary TB ascertained by clinical presentation & chest X–ray examination confirmed by microscopic detection of AFB | 454 | 556 | 4.70 (2.70–8.10) | 128 | Matched by: sex, age (±10%), & residential location. Adjusted for: age, sex, BMI, income, number of individuals per household, & presence of TB contact in family or household |
| Lai et al (2014) [50] | Taiwan | 1998–2011 | General population | Cases: newly diagnosed TB-patients aged ≥20 years selected from the National Health Insurance Program database. Controls: TB-free individuals from same database | DM ascertained by ICD–9 codes from medical records | Pulmonary TB ascertained by ICD–9–010, 011, 012, 018 codes from medical records | 11,366 | 45,464 | 1.46 (1.38–1.54) | 64.89 in 1998 | Matched by: age, sex, & index year of TB diagnosis. Adjusted for: age, COPD, pneumoconiosi, chronic kidney disease, & chronic liver disease |
| 75 in 2002 | |||||||||||
| 67 in 2003 | |||||||||||
| 74 in 2004 | |||||||||||
| 72.5 in 2005 | |||||||||||
| 67 in 2006 | |||||||||||
| 63 in 2007 | |||||||||||
| 62 in 2008 | |||||||||||
| 89 in 2010 | |||||||||||
| 54.4 in 2011[65] | |||||||||||
| Ku et al (2013) [21] | South Korea | 1985–2012 | Severance hospital, Ajou University hospital, & Wonju Christian hospital & Seoul medical center | Cases: HIV-1-infected TB-patients aged ≥18 years. Controls: HIV–1–infected TB–free individuals | DM ascertained from patient’s medical records | All TB ascertained by isolation of Mycobacterium tuberculosis or demonstration of AFB from a clinical specimen or in a histopathological lesion when culture was not available in a patient with signs or symptoms compatible with TB, or evidence of resolution of disease where treatment with two or more anti-TB medications had been prescribed and follow-up had been instigated, excluding AFB TB–positive patients who were finally diagnosed with non-TB mycobacterial infection | 170 | 340 | 1.53 (0.74–3.14) | 119.3 | Matched by: HIV status, & CD4+ T–cell count at first visit & the date of first visit. Adjusted for: age & sex3 |
| Leegaard et al (2011) [51] | Denmark | January, 1980–December, 2008 | Northern Danish population | Cases: first time hospital contact with principal diagnosis of TB obtained from DNRP who lived in Northern Denmark for ≥6 months since the TB diagnosis date. Controls: TB-free individuals from Danish civilian registration system | DM ascertained by in- or outpatient hospital contact involving diabetes, any use of oral anti-diabetes drugs or insulin, at least one visit to a chiropodist for diabetes foot care, at least five glucose-related services in general practice in 1 year, or at least two glucose-related services each year during 5 subsequent years. Patients under 30 on therapy = Type 1DM, rest Type 2DM. HbA1c where available for a subset of controls and cases | All TB ascertained by ICD–8: 010–019; ICD–10: A15–A19 codes. A subset microbiologically confirmed TB | 2,950 | 14,274 | 1.18 (0.96–1.45) | 7 | Matched by: age (±5 years), sex, country of origin, & place & length of residence in Denmark (±1 year). Adjusted for: age, sex, country of origin, place & length of residence in Denmark, comorbidities (myocardial infarction, congestive heart failure, peripheral vascular disease, CVD, dementia, COPD, connective tissue disease, ulcer disease, mild liver disease, hemiplegia, moderate to severe renal disease, any tumor, leukemia, lymphoma, metastatic solid tumor, & HIV/AIDS), alcoholism-related disorders, marital status, number of children <15 years, & degree of urbanization |
| Type 1 DM | 2.59 (0.44–15.29) | ||||||||||
| Type 2 DM | 1.17 (0.95–1.44) | ||||||||||
| Jurcev-Savicevic et al (2013) [52] | Croatia | 2006–2008 | Seven Croatian counties | Cases: TB-patients aged ≥15 years. Controls: TB-free individuals from database of general practitioners in each TB-case county who had not developed TB in a two-month period | DM ascertainment by self-report | Pulmonary TB ascertained by bacterial positive culture among cases | 300 | 300 | 2.38 (1.05–5.38) | 23 | Matched by: age (±2 years), sex, & county of residence that had no history of TB from the database of general practitioners. Adjusted for: age, sex, BMI, country of birth of parents, education, household equipment, employment, smoking status, contact with TB, & malignant disease |
| Jick et al (2006) [53] | UK | 1990–2001 | General population | Cases: first–time TB–patients obtained from General Practice Research Database. Controls: TB–free individuals from same database | DM ascertained by presence of anti-DM medication prior to TB index date | All TB ascertained by prescription of at least 3 anti-TB medications for at least 6 months | 497 | 1,966 | 3.80 (2.30–6.10) | 3 | Matched by: age, sex, geography, index date, & medical history. Adjusted for: age, sex, index date, amount of computerized medical history, glucocorticoid use, smoking, BMI, pulmonary disease, & use of anti-rheumatic / immunosuppressive agents |
| Pablos-Mendez et al (1997) [9] | USA | 1991 | Civilian hospitals in California | Cases: TB–patients. Controls: TB–free patients with primary discharge diagnosis of deep venous thrombosis of the legs, pulmonary embolism, or acute appendicitis | DM ascertained from medical charts coded as ICD–9 250.0–250.9 | All TB coded as ICD–9 010 to 018 | 5,290 | 37,366 | 1.53 (0.81–2.90) | 17.32 | Matched by: age & race. Adjusted for: race-specific aORs for age, sex, poor education, median income, health insurance, HIV-related conditions, chronic renal insufficiency, alcohol-related conditions, & drug use were pooled using random-effects model5 |
| Type 1 DM | 1.40 (0.83–2.35) | ||||||||||
| Type 2 DM | 1.02 (0.63–1.66) | ||||||||||
| Perez et al (2006) [11] | USA | 1999–2001 | 15 Texas/Mexico border counties | Cases: TB-patients aged ≥15 years from Texas hospitals discharge database. Controls: TB-free patients aged ≥15 years with deep venous thrombosis, pulmonary embolism, or acute appendicitis from same database. Excluding HIV cases | DM ascertained from medical chart coded as ICD–9: 250.0–250.9 | All TB coded as ICD–9 code 010–018. TB codes were sought in the admitting diagnosis, principal diagnosis, and eight other variables with diagnosis codes | 3,847 | 66,714 | 1.75 (1.32–2.33) | Mexico-borders counties: 13.1. Non-Mexico borders counties: 6.6 | Region-specific aORs for age, sex, race/ethnicity, insurance type/status, any type of cancer, chronic renal failure, nutrition deficit, income, & education were pooled using random–effects model5 |
| Corris et al (2012) [54] | USA | 1976–1980 | General population | Cases: TB–patients aged 20–74 years. Controls: TB–free individuals aged 20–74 years. Cases & controls were a cross-sectional sample from the second National Health and Nutrition Examination Survey included weighted civilian non–institutionalized US population | DM was ascertained by self-report to questions “Do you have diabetes or sugar diabetes?, "Have you ever been told by a doctor that you have borderline diabetes?" and/or a 75g OGTT test at current WHO cut points (OGTT takes precedence) | All TB ascertained by self-reported ever received diagnosis of TB from a doctor regardless of whether they still had it | 166 | 15,191 | 2.31 (1.36–3.93) | 11 in 19824 | Adjusted for: age, race, poverty index, BMI, household contact with TB, & cigarette smoking status |
| Buskin et al (1994) [55] | USA | 1988–1990 | Seattle/King county TB clinic | Cases: TB–patients, residents of King County aged >17 years seeking care at a TB clinic in Washington. Controls: active TB–free, residents of King County aged >17 years seeking care at a TB clinic in Washington | DM ascertained by self-reported history of DM taken from the questionnaire 1.7 months after the date of TB onset | All TB ascertained by CDC, 1990 criteria that emphasize laboratory confirmation of Mycobacterium tuberculosis and other specified criteria [66] | 151 | 545 | 1.70 (0.70–4.30) | 9 | Adjusted for: age |
| Brassard et al (2006) [25] | USA | September, 1998–December, 2003 | General population | Cases: TB–infected rheumatoid arthritis patients aged ≥18 years. Controls: TB–free rheumatoid arthritis patients aged ≥18 years. Cases & controls selected from the PharMetrics database with ≥1 prescription for antirheumatic medication | DM ascertained from medical chart coded as ICD-9 250.0–250.9 | All TB ascertained from medical chart using ICD–9 code 010–018 codes | 386 | 38,600 | 1.50 (1.15–1.90) | 5.6 | Matched by: date of cohort entry. Adjusted for: age, sex, silicosis, chronic renal failure, hemodialysis, solid organ transplant, head & neck cancer, NSAIDs, steroids, Cox-2 inhibitors. Adjusted effect estimate reported in the previous review [13] |
| Mori et al (1992) [56] | USA | January, 1983–December, 1989 | Shannon county, South Dakota | Cases: American Oglala Sioux Indian TB–patients >18 years. Controls: TB–free individuals with positive TST before the median date of diagnosis of TB, August 1, 1986. Obtained from Oglala Sioux Indians from Indian health service hospital database | DM ascertained by anti-DM oral treatment (hypoglycemic agents or insulin); or ≥11.1 mmol/l at screening or ≥7.8 mmol/l FBG | Cases: clinically diagnosed TB from Indian health service and clinical charts based on the State Health Department definition of active TB. TB type not specified. Controls: positive TST ascertained from their medical records | 46 | 46 | 5.20 (1.22–22.10) | 90.9 in Shannon county | Matched by: age & residence. Adjusted for: sex, alcohol abuse, & isoniazid therapy for >6 months. Cases and controls were not significantly different according age |
| Viney et al (2015) [57] | Republic of Kiribati | June, 2010–March, 2012 | Residents of South Tarawa city | Cases: TB–patients >18 years from the National TB Control Center and the National TB Laboratory. Controls: TB–free individuals >18 years (members of the same community without symptoms of TB) | DM ascertained by HbA1c ≥6.5% mmol/mol or self-reported DM with a treatment by a clinician | Cases: all TB ascertained by bacteriological, clinical and radiological criteria assessed by experienced physicians. Controls: all TB ascertained by TB-symptoms (cough >2 weeks, fever, nights sweats, weight loss), confirmed by TST | 275 | 499 | 2.80 (2.00–4.10) | 429 | Adjusted for: age & sex |
| Coker et al (2006) [58] | Russia | January, 2003–December, 2003 | Residents in the city of Samara | Cases: newly diagnosed adult TB–patients at any of city’s specialist TB clinics and recruited to a WHO DOT program. Controls: TB–free general residents of the Samara city | Method of DM ascertained was an unclear | Pulmonary TB ascertained by positive bacterial culture | 334 | 334 | 7.83 (2.37–25.89) | 1182 | Matched by: age & sex. Adjusted for: age, sex, relative with TB, alcohol, drinking raw milk, assets, number of cohabitating person, employment, smoking, financial security, illicit drugs, & imprisonment |
| Faurholt-Jepsen et al (2011) [59] | Tanzania | April, 2006–January, 2009 | Four major health facilities in Mwanza city | Cases: TB–patients aged ≥15 years, excluding pregnant or lactating women, patients terminally ill from TB or HIV, patients suffering from other severe diseases, & non–residents of Mwanza City. Controls: TB–free aged ≥15 years, with no history of TB in the household members and no evidence of active TB (cough, intermittent fevers, excessive night sweating in the past two weeks, and unexplained weight loss in the past month) | DM ascertained by either FBG >6mmol/L or OGTT >11mmol/L according to WHO guidelines, for both cases and controls | Pulmonary TB ascertained by initial diagnosis with sputum positive microscopy based on three sputum samples (‘‘spot-morning-spot”), with an additional early morning sputum sample was collected for Mycobacterium tuberculosis culture, for both cases and controls | 803 | 350 | 2.13 (1.37–3.31) | 504 in 2006 | Matched by: residence, sex, age (± 5 years), not pregnant or lactating, not terminally ill from TB or HIV, not suffering from other diseases, & resident of Mwanaza city. Adjusted for: HIV–status stratum-specific aORs for age, sex, religion, marital status, & occupation were pooled using random–effects model5 |
| 452 in 20092 | |||||||||||
| Wu et al (2007) [60] | Taiwan | January, 2002–December, 2004 | Chang Gung Memorial Hospital, Keelung | Cases: TB–patients with lower respiratory tract infection or who had been in contact with TB patients. Controls: non–TB pneumonia patients who did not meet the criteria for TB | DM ascertained from medical records | Pulmonary TB ascertained by positive sputum culture for Mycobacterium tuberculosis | 264 | 438 | 3.43 (2.16–5.46) | 75 in 2002 | Adjusted for: age, sex, pneumoconiosis, bronchiectasis, liver cirrhosis, haemodialysis, & lung cancer |
| 67 in 2003 | |||||||||||
| 74 in 2004 [65] | |||||||||||
| Rosenman and Hall, (1996) [61] | USA | January, 1985–May, 1987 | New Jersey Department of Health | Cases: male TB-patients aged ≥35 years who speak English, excluding HIV positive and/or foreign born cases. Controls: TB-free individuals registered at the New Jersey Department of Health | DM ascertained by self-report | All TB ascertained by positive mycobacterium tuberculosis culture, or who had a physician diagnosis of pulmonary TB with multidrug anti-TB medication | 148 | 290 | 1.16 (0.58–2.32) | 9.5 | Matched by: age (±5 years), gender, & race. Adjusted for: age, sex, & race |
| Cross-sectional | |||||||||||
| Goldhaber-Fiebert et al (2011) [62] | Multi-center WHO survey | 2002–2003 | 46 countries | General population | DM ascertained by self-report, based on positive response to the question "Have you ever been diagnosed with diabetes (high blood glucose)?" | All TB ascertained by self-reported symptoms of active TB, based on positive response to two questions "Over the last 12 months, have you had blood in your phlegm or have you coughed blood?" & "Over the last 12 months, have you experienced cough lasting over 3 weeks?" | 124,545 | 1,744 | 1.81 (1.37–2.39) | - | Age, sex, BMI, schooling in years, smoking & length of being daily smoker, urban and rural residence, at least 1 drink per day, number of household members, number of individuals per room, & SES based on different household’s assets |
| Marks et al (2011) [12] | USA | 2000–2005 | General population | Civilians, non-institutionalized household residents aged ≥18 years selected from six national health insurance data bases | DM ascertained by self-report | All TB ascertained by self-report | 190,350 | 668 | 1.40 (1.00–2.00) | 7 in 2000 | Age, sex, race/ethnicity, foreign birth, high school drop-out, history of homelessness or incarceration, ever cancer diagnosis, current cigarette smoking, past year alcohol abuse, no health insurance, & ever HIV testing |
| 5 in 20052 | |||||||||||
| Wang et al (2013) [63] | China | September, 2010– December, 2012 | TB clinics and neighboring communities in Linyi city | TB and non-TB patients with and without DM, excluding HIV positive patients | DM ascertained by FBG ≥7 mmol/L | Pulmonary TB ascertained by sputum smear positive; if sputum smears were negative and chest radiograph was compatible with active pulmonary TB, the patient was diagnosed as smear negative pulmonary TB | 13,057 | 6,382 | 3.17 (1.14–8.84) | 78 in 2010 | Age, sex, BMI, family history of DM, annual income, education level, smoking, alcohol consumption, outdoor activity, & marital status |
| 73 in 20122 | |||||||||||
1 Background TB incidence per 100,000 person–year during the same year or closest year to the survey.
2 Data retrieved from (http://www.cdc.gov.tw/uploads/files/201407/103228a0-fadd-47b0-b056-8dedda9fce1d.pdf); (file:///C:/Users/rha2006/Downloads/%253f44CurrentStatusofTuberculosisinTaiwan%20(1).pdf).
3 Adjusted estimate provided by author.
4 Data obtained from external source; the World Bank records (http://data.worldbank.org/indicator/SH.TBS.INCD?end=2014&start=1990) and the WHO TB country profiles (http://www.who.int/tb/country/data/profiles/en/).
5 Pooling was done by the present study team and was not reported in the original study.
TB: tuberculosis; DM: diabetes mellitus; OR: odds ratio; aOR: adjusted odds ratio; HbA1c: glycated haemoglobin (measure of serum glucose levels over time in humans); DNRP: Danish National Registry of Patients; AFB: acid–fast bacilli; COPD: chronic obstructive pulmonary disease; TST: tuberculin skin test; SES: socio–economic status; HIV: human–immunodeficiency virus; BMI: body mass index; ICD–9: International Statistical Classification of Diseases and Related Health Problems 9th edition; WHO: World Health Organization; CDC 1990: 1990 Case Definition for Tuberculosis by Center for Disease Control (US); NSAID: non–steroidal anti–inflammatory drug; CVD: cardiovascular diseases.