Table 1. Baseline characteristics of 23 cohort studies, prospective and retrospective, 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 | Sample size | TB cases | Adjusted effect size (95% CI) | TB incidence/100,000 p–y1 | Adjusted variables |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Prospective | |||||||||||
| Kim et al (1995) [8] | South Korea | 1988–1990 | Authorized hospitals in South Korea | Civil servants examined by Korean Medical Insurance Corporation who claimed health insurance for TB | DM ascertained by glucose level of ≥119 mg/dl at screening followed by FBG ≥150 mg/dl & PPBG ≥180 mg/dl | Pulmonary TB, bacteriologically ascertained | 814,713 | 5,105 | RRs: 4.97 (3.68–6.70) | 306 | Sex-specific stratum crude RRs were pooled using random-effects model2 |
| Leung et al (2008) [35] | China | January, 2000–December, 2005 | 18 elderly health service centers | Elderly people aged ≥65 years | DM ascertained by HbA1c ≥7% at enrollment for those with known history of DM based on a physician diagnosis | Culture confirmed pulmonary and extra-pulmonary TB | 42,116 | 326 | HR: 2.69 (1.94–3.72) | 90 | Age, sex, alcohol use, BMI, marital status, smoking, language, education, housing, working status, public means-tested financial assistance status, CVD, hypertension, COPD/asthma, malignancy, recent weight loss of 5% within 6 months, hospital admission within 12 months, & activities of daily living scores |
| Active TB | HR: 2.56 (1.95–3.35) | ||||||||||
| Pulmonary TB | HR: 2.80 (2.11–3.70) | ||||||||||
| Extrapulmonary TB | HR: 0.88 (0.35–2.20) | ||||||||||
| John et al (2001) [36] | India | 1986–1999 | Christian Medical College and Hospital at Vellore in southern India | Renal allograft recipients | DM ascertained by FBG >120 mg/dl, or 2-hours PPBG 200>mg/dl, or two elevated levels of either measurement from medical records | All TB ascertained from medical records based on X-ray, AFB, gastric juice, bronchoalveolar specimen, or histopathology | 1,251 | 166 | HR: 2.24 (1.38–3.65) | 1683 | Age, chronic liver disease, deep mycoses, cytomegalovirus, Pneumocystis carinii pneumonia, nocardia, prednisolone plus azathioprine, & cyclosporine use |
| Cegielski et al (2012) [23] | USA | 1971–1992 | General population | Civilian, non-institutionalized adults aged 25–74 years recorded in the First National Health and Nutrition Examination Survey (NHANES I) | DM ascertained by self-report. NHANES I questionnaire asked respondent 'Has a doctor ever told you that you have any of the following conditions, and if so, do you still have it? Diabetes? to assess DM status | All TB. 21 TB cases were ascertained by self-report, the rest based on the ICD-9–010–018 and 137, excluding TB exposure without disease (ICD-9-V01.1), primary infection without disease (ICD-9-010.0), TST positivity without diseases (ICD-9-795.5) & subjects who had TB before NHANES I | 14,189 | 61 | HR: 7.58 (2.94–19.49) | 15.43 | Age, sex, & BMI |
| Retrospective | |||||||||||
| Chung et al (2014) [22] | Taiwan | 1997–2010 | General population | Newly diagnosed TB patients from the Taiwan’s National Health Insurance Research Database and non-TB subjects from general population | DM ascertained by ICD-9-CM 250 codes from medical records | All TB ascertained by receiving medical care at least three times, including out-patient visits and/or hospitalizations, for a principal diagnosis of TB based on ICD–9–CM 011–018 codes | 50,840 | 10,168 | RRs: 1.55 (1.47–1.64) | 72.5 | Age & sex5 |
| Ou et al (2012) [37] | Taiwan | January, 1997–December, 2006 | General population | Kidney transplant recipients identified from the Taiwan’s National Health Insurance Database | DM ascertained from National Health Insurance Database | Newly diagnosed all TB ascertained by ICD–010–018 codes validated by the use of at least 2 anti-TB medications | 4,554 | 109 | OR: 1.42 (0.96–2.09) | 67.4 | Age, sex, COPD, autoimmune disease, cirrhosis, hepatitis C virus infection, HIV, cyclosporine, & mycophenolate mofetil |
| Chen et al (2013) [5] | China | 2006–2008 | General population in rural areas | Residents of Danyang county of Jiangsu province and Xiangtan county of Hunan province | DM ascertained by self-reported history of DM by answering the question ‘‘Has a doctor ever told you that you have diabetes?” | All TB ascertained by sputum smear positive (including scanty positive) or sputum culture-positive for mycobacterium tuberculosis | 177,529 | 117 | RRs: 2.43 (0.84–7.00) | 59.74 in Danyang county. 101.1 in Xiangtan county | County-specific aRRs for sex, age, marital status, occupation, & educational level were pooled using random-effects model2 |
| Pealing et al (2015) [38] | United Kingdom | 1990–2013 | Clinical practice research data linked to the hospital episode statistics | DM cohort: patients with first recorded diagnosis for DM (type 1 and 2) aged ≥5 years.DM-free cohort: patients who did not have a prevalent diagnosis of DM on the matched index date | DM ascertained by HbA1c >7.5% mmol/mol | All TB ascertained by ICD–10 codes. Prescriptions for anti-TB drugs were not used in developing or later validating cases of TB identified by diagnostic codes. Only one TB case occurred in T1DM cases | 6,941,000 | 969 | RR: 1.30 (1.01–1.66) | 13.9 in 2012 | Age, smoking, alcohol use, BMI, ethnicity, & index of multiple deprivations. DM & non-DM subjects matched by age ±5yeras, sex, & general practice |
| Moran-Mendoza et al (2010) [6] | Canada | 1990–2000 | General population in British Columbia | Contacts of active TB cases recorded at the Division of TB control at the British Columbia Center for Diseases Control, excluding contacts of HIV infection cases or previous active TB cases | DM ascertained from databases, but unclear which databases | All TB by smear and/or culture positive for tubercle bacilli, histopathological diagnosis, or clinical and radiological diagnosis of active TB, with complete treatment response, when smears and cultures were negative | 33,146 | 228 | HR: 1.76 (0.54–5.75) | 72 | Age, sex, malignancy, corticosteroids, alcohol, malnutrition, closeness of TB contact, TST size in millimeter, intravenous drug use, ethnicity, SES, recent arrival from country with high TB prevalence, residents/employees in prisons, nursing homes or homeless shelters, chest X-ray compatible with previous TB, & previous BCG vaccination. Adjustment was done with robust variance estimation |
| Baker et al (2012) [39] | Taiwan | August, 2001–December, 2004 | General population | Taiwanese adults aged ≥12 years interviewed during the Taiwan’s 2001 National Health Interview Survey (NHIS) | Treated DM ascertained by ≥2 outpatient ICD–9–CM codes for DM, ≥1 inpatient ICD–9–CM code for DM, or prescription of anti-DM medications for ≥28 days during the study period or ≥2 prescriptions | All TB ascertained if all of the following criteria reported in NHIS database: ≥1 medical visit during the follow-up period with an ICD–9–CM code for TB (codes 010–018); a prescription for ≥2 anti-TB medications for >28 days during the study period; and no finding of a misdiagnosis of TB during the study period on the basis of later diagnosis of non-TB mycobacterial infection, lung cancer, or TB infection without evidence of disease | 17,715 | 57 | HR: 2.60 (1.34–5.03) | 73 | Age, sex, income, employment, alcohol use, education, BMI, living in a crowded home, receipt of government subsidy, residence in an indigenous community, hypertension, heart disease, & lung disease |
| Kuo et al (2013) [40] | Taiwan | 2000–2011 | General population | Patients aged ≥18–≥70 years with type 2 DM matched by sex, year of birth, and month and year of first diagnosis at enrollment with patients without DM or TB recorded in the Taiwan’s National Health Insurance Research Database representing about 5% of Taiwan’s population, excluding HIV cases | DM ascertained by ICD–9–250 (excluding 2501) with continuous prescriptions of anti-DM medications for ≥60 days | All TB ascertained by ICD–9–010–018 codes with continuous prescriptions of anti-TB medications for ≥60 days at least one year after DM code | 253,349 | 5,013 | HR: 1.31 (1.23–1.39) | 73 [65] | Age, sex, bronchiectasis, asthma, & COPD |
| Hu et al (2014) [41] | Taiwan | January, 1998–December, 2009 | General population | Patients receiving dialysis recorded in the Taiwan’s National Health Insurance Research Database, representing about 5% of Taiwan’s population in 2000 | DM ascertained by ICD–9–250 or A181 | All TB ascertained by ICD–9–010–018 or A02 codes & ≥2 anti-TB medications for >28 days | 20,655 | 287 | HR: 1.36 (1.05–1.76) | 64.89 in 1998. | Age, sex, hypertension, silicosis, COPD, connective tissue diseases, & malignancy co-morbidities |
| 75 in 2002 | |||||||||||
| 67 in 2003 | |||||||||||
| 74 in 2004 | |||||||||||
| 72.5 in 2005 | |||||||||||
| 67 in 2006 | |||||||||||
| 63 in 2007 | |||||||||||
| 62 in 2008 [65]4 | |||||||||||
| Lee et al (2013) [42] | Taiwan | 1996–2007 | General population | Subjects with and without COPD disease matched in age (within 5 years), sex, and time of entering the Longitudinal Health Insurance Database-2005 recorded in the National Health Insurance program database that covers more than 95% residents of Taiwan since 1996 | DM ascertained from Longitudinal Health Insurance Database–2005 | All TB ascertained by at least two ambulatory visits or one inpatient record with a compatible diagnosis (ICD–9–CM codes 010–012, and 018, and A-codes A020, A021), plus at least one prescription consisting of ≥3 anti-TB. There should be a prescription of at least 2 anti-TB drugs simultaneously for ≥ 120 days during a period of 180 days | 23,594 | 674 | HR: 1.25 (1.02–1.53) | 64.89 in 1998 | Age, sex, oral corticosteroids, inhaled corticosteroids & oral β-agonists |
| 75 in 2002 | |||||||||||
| 67 in 2003 | |||||||||||
| 74 in 2004 | |||||||||||
| 72.5 in 2005 | |||||||||||
| 67 in 2006 | |||||||||||
| 63 in 2007 [65] 4 | |||||||||||
| Wu et al (2011) [43] | Taiwan | January, 2000–December, 2007 | General population | New onset cancer patients and cancer-free patients recorded in the Taiwan’s’ National Health Insurance Database in 2005, matched by sex and age | DM ascertained by ICD–9–250 | All TB ascertained by ICD-9-010–018 with prescription history of treatment with isoniazid | 82,435 | 694 | HR: 1.38 (1.17–1.62) | 64.89 in 1998 | Age, sex, chronic renal failure, autoimmune diseases, COPD, aerodigestive tract and lung cancers, haematological cancers, & other major/less common cancers |
| 75 in 2002 | |||||||||||
| 67 in 2003 | |||||||||||
| 74 in 2004 | |||||||||||
| 72.5 in 2005 | |||||||||||
| 67 in 2006 | |||||||||||
| 63 in 2007 [65] 4 | |||||||||||
| Chen et al (2006) [24] | Taiwan | January, 1983–December, 2003 | Hospitals | Renal transplant recipients in Taichung | DM ascertained from medical records | All TB ascertained either by positive culture, presence of caseating or non–caseating granuloma in biopsy specimens taken from involved tissue and responsive to treatment, or typical chest X–ray finding or clinical presentation consistent with TB, without microbiological or pathological confirmation but with favorable response to anti-TB treatment | 756 | 29 | RRs: 3.07 (1.14–8.26) | 66.67 | Age, sex, dialysis duration, hepatitis B virus infection, hepatitis C virus infection, graft rejection >3, & immunosuppressive medications. Adjusted effect estimate reported in the previous review [13] |
| Rungruanghiranya et al (2008) [7] | Thailand | January, 1992–December, 2007 | Nationwide | Renal transplant recipients | DM ascertained from case medical records | All TB ascertained by one or more of: AFB in body fluid smears, TB-polymerase chain reaction, and/or growth in various culture specimens; histopathology examination of tissue specimens showing either AFB or granulomatous inflammation; response to TB treatment in patients with typical radiographic findings consistent with TB, or those who had fever of unknown origin despite negative results of extensive investigations | 233 | 9 | OR: 3.59 (0.74–17.35) | 142 in 2005 | Age & sex5 |
| Demlow et al (2015) [44] | USA | 2010–2012 | California department of public health | Non-institutionalized TB cases with and without DM aged ≥18 years | DM ascertained based on history of DM gathered from medical records or healthcare provider, excluding pre-DM, borderline DM, self-reported DM, or gestational DM | All TB ascertained based on information gathered from local TB control programs from medical records or a health care provider | 27,797,000 | 6,050 | RRs: 2.18 (1.79–2.66) | 4.8 | Age & birth location-specific stratified crude RRs were pooled using random-effects model2 |
| Suwanpimolkul et al (2014) [45] | USA | April, 2005–March, 2012 | San Francisco TB control sections | All individuals seeking medical care who had final diagnosis of TB, latent TB (LTB), or no evidence of TB or LTB. DM in TB patients was assessed versus DM in individuals with LTB | DM status reported by patient ascertained from medical records based on the screening policies of the San Francisco TB control sections | All TB ascertained by Standards of the American Thoracic Society and Centers for Disease Control and Prevention | 5,162 | 791 | OR: 1.81 (1.37–2.39) | 2.8 | Age & place of birth |
| Kamper-Jorgensen et al (2015) [20] | Denmark | January, 1995–December, 2009 | General population | Entire Danish population | DM ascertained from Danish National Diabetes Register including blood glucose testing, foot treatment, or purchase of anti-DM drugs | All TB ascertained according to the WHO definitions. TB is diagnosed on the basis of microbiology and/or laboratory results, or solely on clinical evaluation. In Denmark, around 70–75% of all notified cases are verified using culture | 77,935 | 6,468 | RR: 1.60 (1.43–1.79) | 73 | Age & sex5 |
| Young et al (2012) [46] | England | ORLS1: 1963–1998.ORLS2: 1999–2005 | Admissions records in all NHS hospitals in defined populations in the former Oxford NHS region | DM cohort: all forms of DM first record on file for each individual with DM. Reference cohort: people with various common orthopedic, dental, ENT and other relatively minor disorders | DM ascertained by ICD7 260, ICD8 250, ICD9 250, ICD10 E10-E14 codes | All TB ascertained by ICD7 001–019, ICD8 010–019, ICD9 010–018, 137, ICD10 A15–A19, B90 codes | 837,399 | 7,996 | RR: 2.02 (1.35–3.04) | 56 in 1964 | Age in 5–years band, sex, time period, & district of residence adjusted ORLS1 & ORLS2 survey rounds-specific RRs were pooled using random-effects model2 |
| 26 in 1974 | |||||||||||
| 13 in 1984 | |||||||||||
| 4 in 1994 | |||||||||||
| 5 in 2004 [65] 4 | |||||||||||
| Dobler et al (2012) [47] | Australia | January, 2001–December, 2006 | General population | Residents of Australia | DM ascertained from medical records per the National Diabetes Services Scheme | Culture–positive TB ascertained based on state and territory TB notification records. | 19,855,283 | 6,276 | RRs: 1.49 (1.05–2.11) | 5.8 | Age, sex, indigenous status, & TB incidence in country of birth |
| All TB | RRs: 1.48 (1.04–2.10) | ||||||||||
| All TB in insulin users | RRs: 2.27 (1.41–3.66) | ||||||||||
| Culture–positive TB in insulin users | RRs: 2.55 (1.62–4.01) | ||||||||||
| Shen et al (2014) [48] | Taiwan | 2002–2011 | General population | Type 1 DM patients aged <40 years identified from the Registry of Catastrophic Illnesses Patient database & non-type 1 DM cohort identified from the Longitudinal Health Insurance Database in 2000 | Newly diagnosed type 1 DM ascertained by ICD–9 250.x1 & 250.x3 codes from data recorded in the Registry of Catastrophic Illnesses Patient database | All TB ascertained by ICD–9–CM codes from medical records | 25,975 | 59 | HR: 4.23 (2.43–7.36) | 53 | Age, sex, chronic liver infection, chronic kidney infection, & previous infections |
| Dyck et al (2007) [49] | Canada | January 1986–December 2001 for TB case; January 1991–December 1995 for DM survey | General population | Registered American Indians and other Saskatchewans aged ≥20 years selected from population-based health databases in Saskatchewan | DM ascertained by ICD–9: 250 codes from medical charts | All TB cases aged ≥20 years reported to Saskatchewan Health | 791,673 | 1,118 | RR: 1.00 (0.69–1.44) | 43.8 | Age, race, & sex stratum-specific crude ORs were pooled using random–effects model2 |
| Other6 | |||||||||||
| Ponce-De-Leon (2004) [64] | Mexico, state of Veracruz | March 1995–April 2003 for TB case; 2005 for DM survey | General population | Non–institutionalized civilians | DM ascertained by a previous diagnosis of a physician; or FBG ≥126 mg/dl or random blood glucose ≥200 mg/dl | All TB ascertained by positive AFB or positive culture | 21,230 | 581 | RR: 6.00 (5.00–7.20) | 28 | Age & sex–standardized for the adult population of the study area |
1 Background TB incidence per 100,000 person–year during the same year or closest year to the survey.
2 Pooling was done by the present study team and was not reported in the original study.
3 Data obtained from external source; the World Bank records (http://www.cdc.gov/tb/statistics/tbcases.htm) and the WHO TB country profiles (http://www.who.int/tb/country/data/profiles/en/)
4 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).
5 Adjusted estimate provided by author.
6 Study by Ponce-De-Leon A., et al [64] neither categorized as prospective, retrospective, cross–sectional, or case–control study.
TB: tuberculosis; DM: diabetes mellitus; HbA1c: glycated haemoglobin (measure of serum glucose levels over time in humans); PPBG: postprandial blood glucose; AFB: acid–fast bacilli; COPD: chronic obstructive pulmonary disease; TST: tuberculin skin test; HIV: human–immunodeficiency virus; RRs: relative risk; OR: odds ratio; HR: hazard ratio; RR: rate ratio; aOR: adjusted odds ratio; aRRs: adjusted relative risk; BMI: body mass index; BCG: bacilli Calmette–Guérin; 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).