TABLE 1.

Study characteristics of the studies included determining the association between diabetes mellitus and tuberculosis in sub‐Saharan Africa

First author, (year)	Country, setting	Study period	Study design	Sample size	DM ascertainment	TB ascertainment	Primary comparison	Age mean/median (sd/IQR)	HIV+venumber (%)	Variables adjusted for
Kubjane et al. (2020) [21]	South Africa, hospital	July 2013 ‐ August 2015	Cohort	850	FBG ≥7 mmol/L, HbA1c ≥6.5% or self‐reported	Pulmonary TB, determined by GeneExpert	Patients presenting to the clinic with respiratory symptoms with a negative GeneExpert and resolution of symptoms within 3 months without TB treatment	38 (31–47)	519 (61.1)	Age, sex, HIV, hypertension, household size, income, previous miner, previous prisoner, marital status, work status
Sinha et al. (2018) [25]	South Africa, community	2010–2015 and 2015–2016	Cross‐sectional	7708	RBG >11.0 mmol/l or self‐reported	Pulmonary TB, determined by presence of one or more of the following TB symptoms: cough of any duration, fever of any duration, weight loss, night sweats	All participants without DM	42.6 (20.5)	837 (10.9)	Age, sex, HIV, receipt monthly grant, access to tap water, access to toilet, access to solar/electric energy
Lawson et al. (2017) [26]	Nigeria, hospital	NR	Cross‐sectional	663	HbA1c >6.4% or self‐reported in interview	Pulmonary TB, determined by sputum culture	Patients presenting to the clinic with presumptive TB (cough >2 weeks) without DM	37.8 (12.6)	184 (45.9)	Age, sex, HIV status
Boillat‐Blanco et al. (2016) [22]	Tanzania, hospital	June 2012 – December 2013	Case‐control	1035	Repeated measure FCG ≥7.0 mmol/l, OGTT ≥11.1 mmol/l, HbA1c ≥6.5% or history of and treatment for DM	New active TB, determined by sputum smear microscopy, chest radiography or clinical diagnosis	Sex and age‐matched controls selected from adults accompanying patients other than the included patients	36.3 (12.5)	232 (22.7)	Age, sex, BMI, HIV, socio‐economic status
Senkoro et al. (2016) [23]	Tanzania, setting not reported	NR	Case‐control	7163	Self‐reported	Pulmonary TB, confirmed with positive sputum culture or at least 2 smear positive results for AFB or one smear positive for AFB and chest X‐ray	All participants with presumptive TB who are bacteriologically negative and a random sample of people without presumptive TB	38.5 (17.5)	313 (5.2)	Age, sex, history of previous TB, BMI, HIV
Bailey et al. (2016) [27]	Zambia and South Africa, community	January 2010 – December 2010	Cross‐sectional	90,601	RBG >11 mmol/l	Pulmonary TB, determined by sputum culture, confirmed with RNA sequencing	All participants without DM	30 a	6517 (7.2)	Age, sex, household economic position, education, BMI, HIV status, geographical location
Haraldsdottir et al. (2015) [24]	Guinea‐Bissau, community	July 2010 – July 2011	Case‐control	700	RBG ≥7 mmol/l at inclusion confirmed with 2 FBG >7 mmol/l or registered at DM clinic	Pulmonary TB, determined by sputum smear microscopy or chest radiography plus relevant, signs, symptoms and chest radiography changes after ineffective antibiotic treatment	Non‐TB controls, identified by random selection of houses in the study area	26.5 a	NR	Age, sex, BMI
Bates et al. (2012) [28]	Zambia, hospital	September 2010 – December 2011	Cross‐sectional	964 (275 with NCD)	DM as admission diagnosis to hospital	Pulmonary TB, determined by sputum microscopy and culture	Participants with a NCD (except DM) as admission diagnosis	35 (28–43)	606 (67.3)	Age, HIV
Faurholt‐Jepsen et al. (2011) [20]	Tanzania, hospital and community	April 2006 – January 2009	Case‐control	1221	FBG > 6 mmol/L or OGTT > 11 mmol/L	Pulmonary TB, confirmed with sputum smear microscopy and sputum culture	Randomly selected sex and age‐matched controls living in same neighbourhood	34.3 (12.0)	382 (33.1) a	Age, sex, HIV, socio‐demography b

Abbreviations: DM, diabetes mellitus; TB, tuberculosis; FBG, fasting blood glucose; HbA1c, glycosylated haemoglobin; HIV, human immunodeficiency virus; RBG, random blood glucose; NR, not reported; FCG, fasting capillary glucose; OGTT, oral glucose tolerance test; BMI, body mass index; AFB, acid fast bacilli; NCD, non‐communicable disease.

^a The age mean/median was not reported by the study authors, but calculated by the researchers for the purpose of this review.

^b This study also presented a model which additionally adjusted for serum alpha‐1‐acid glycoprotein, because it was uncertain whether this was a confounder or whether it was on the pathway between DM and TB risk, the model that did not control for this was chosen.