Table 3.

Characteristics of antioxidants’ effects on risk of type 2 diabetes mellitus in meta-analysis of observational studies

Author	Type of antioxidants	meta-analyzed studies (n)/ Population	Participants			Event (n)	Follow-up	Outcome measures	Significant results	Quality assessment/ adjusted
			Total population (n)	Age (yr)	Sex
Vitamin D
Rafiq et al. 2021 [80]	Serum level VD	40/ T2DM, non-DM	UN	> 18y	Both	UN	UN	Correlation between serum VD and Insulin Resistance	↑insulin resistance with ↓ VD	Yes/ NO
Dai et al. 2019 [81]	Serum level VD	7/ DF, DM	1115	Mean: 49.6–70.2	Both	40–289	UN	Serum VD, Association of serum VD with DF	↓Serum VD < 50 nmol/l in DF, and DM, ↑Risk DF (OR: 3.22 in VD < 25 nmol/l)	Yes/ NO
Zhang et al. 2019 [82]	Serum level VD	13/ T2DM with—without DPN (case–control studies)	2814	54- 66	Both	UN	NA	correlation of VD and DPN in T2DM	↓VD serum in DPNcompared to non-DPN, RR VD deficiency for DPN: 1.07	Yes/NO
Rafiq et al. 2018 [11]	Serum level VD	45/ T2DM, non-DM	UN	> 18 yr	Both	UN	UN	Correlation between serum VD and T2DM	↑T2DM risk in diabetics and non-DM with hypo-D	Yes/ NO
Ekmekcioglu et al. 2017 [83]	Serum level VD	28/ General population, Renal transplanted	132,325	UN	Both	UN	UN	Association of serum VD with T2DM	Dose–response U-Shape association, RR: 0.66 highest vs. lowest serum VD	UN/ Yes
Lucato et al. 2017 [84]	Serum level VD	9/ Older population	28,258	67.7	Both	2863	7.3	Association of hypo-VD with T2DM	RR lower serum level vs. higher: 1.17	Yes/ Yes
Luo et al. 2017 [85]	Serum level VD	15/ T2DM with/without DR	17,664	> 18 yr	UN	3455	UN	Association serum VD with DR	OR VD deficiency with DR: 2.03, Pooled difference VD in DR vs. non-DR: -1.71 ng/ml	Yes/Yes
Qu et al. 2017 [86]	Serum level VD	10/ T2DM	1468	UN	UN	NA	NA	Association serum VD with DPN	↓Serum VD in Caucasian DPN vs. non-DPN, OR VD deficiency and DPN in Asian: 1.22	Yes/ NO
Liang et al. 2016 [87]	Serum level VD	11/ T1DM, T2DM	6851	12–73	Both	2292	NA	Association serum VD with DR	OR VD deficiency with DR in DM: 1.51, but in T2DM: 1.15	UN/ Yes
Derakhshanian et al. 2015 [44]	Serum level VD	6/ T1DM, T2DM	3700	9–75	UN	NA	NA	Association serum VD with DN	↑Risk DN (OR: 1.80 in VD < 20 ng/ml or VD deficient)	Yes/ NO
Lv et al. 2015 [88]	Serum level VD	6/ T2DM	1484	54.6–63.3	Both	NA	NA	Association VD deficiency with DPN	↓VD serum in DPN, ↑Risk DPN with VD deficiency; OR: 2.88	Yes/ Yes
Afzal et al. 2013 [89]	Serum level VD	16/ General population	9841	47–64	Both	810	29 yr	T2DM risk	↑Odds ratio for T2DM: 1.50 for bottom vs. top quartile VD	Yes/ Yes
Khan et al. 2013 [12]	Serum level or dietary intake of VD	14/ General population	190,626	25–79	Both	9399	10 yr	T2DM risk	↓19% risk T2DM (top 3rd vs. bottom 3rd)	Yes/ Yes
Song et al. 2013 [90]	Serum level VD	21/ Population-based	76,220	30–79	Both	4996	22 yr	T2DM risk	↓Risk T2DM (RR:0.96 per 10 nmol/l increment in serum 25-OHD, RR: 0.62 highest vs. lowest serum 25-OHD)	UN/ Yes
Forouhi et al. 2012 [91]	Serum level VD	11/ General population	59,325	40–76	Both	3612	1.3–22 yr	T2DM risk	↓VD in T2DM vs. non-DM, RR:0.59 highest vs. lowest level VD	Yes/ Yes
Mitri et al. 2011 [46]	Serum level VD, VD intake	8/ General population	238,423	30–75	Both	7915	1.3–22 yr	T2DM risk	↓13% risk T2DM by VD intake > 500 IU/d vs. intake  < 200 IU/d, ↓risk T2DM 43% highest serum level vs. lowest	Yes/ Yes
Pittas et al. 2007 [92]	Serum level VD	4/ General population	8758	> 20	Both	UN	UN	Association with T2DM	OR: 0.36 highest vs. lowest serum VD in non-blacks	UN/ Yes
	VD + Ca intake	4/ General population	203,402	39–55		9653			OR: 0.82 highest vs. lowest by VD + Ca
Alpha lipoic acid
Muley et al. 2014 [93]	ALA intake	6/ General population	153,751	35–75	Both	NA	4-16 yr	T2DM risk	↓T2DM risk (RR: 0.90) by two fold increase intake ALA	Yes/ Yes
Selenium
Wang et al. 2016 [94]	Serum level of Sel	5/ Community population	13,460	Mean: 45.2–77.7	Both	NA	NA	OR association with T2DM	U shape association of serum Sel < 97.5, > 132.5 µg/L with T2DM	Yes/ NO
Copper
Qiu et al. 2017 [95]	Plasma and serum level of Cu	15/T1DM, T2DM, Healthy	1640	5–75	Both	UN	UN	Determine level of Cu	↑Cu in plasma and serum DM vs. healthy	Yes/ Yes
Zinc
Pearsey et al. 2020 [96]	Zinc-alpha2-glycoprotein (ZAG)	3/ T2DM, PCOS, MetS, prediabetes	3127	36–52	Both	NA	NA	Relationship between ZAG, and FBS, BS2hpp, HbA1c, fasting insulin, HOMA	↓ZAG in dysglycaemia compared to metabolically healthy obese	Yes/Yes
									subgroup analysis: only sig. diff. for PCOS
Fernandez-Cao et al. 2019 [97]	Zn intake	12/ Urban, Rural population	405,466	18–84	Both	27,643	5–24 yr	T2DM risk	↓13% or ↓41% T2DM risk by moderated high intake in general or in rural	Yes/ Yes
	Zn serum level	8/ General population	10,151	40–92		3331	4.6–20		↑64% risk of T2DM by elevated serum Zn
Fernandez-Cao et al. 2018 [98]	Zn intake	6/ T2DM, Healthy	685	25–72	Both	NA	NA	Relationship with T2DM	↓Zn intake in T2DM with complications	Yes/ Yes
	Zn whole blood level	5/T2DM, Healthy	1416	25–75					↓Zn for each year T2DM: 732.6 + (-77.88)*duration of T2DM (yr)
Several antioxidants
Hamer et al. 2007 [99]	Highest intake of VE, VC, Flavonoids, Carotenoids, Lycopene/ lowest intake	9/ General population	139,793	18–74	Both	8813	13 yr	T2DM risk	↓13% risk T2DM (mainly by VE)	Yes/ Yes

Legend: T2DM Type 2 diabetes mellitus; MetS Metabolic syndrome; IR Insulin resistance; Sel Selenium, Zn Zinc; Cu Copper; DR Diabetic retinopathy; VC Vitamin C; VE Vitamin E; VD Vitamin D; UN Unknown; Yr Year; IR Insulin resistance; NA Not applicable; ALA Alpha lipoic acid; Sel Selenium; DPN Diabetic peripheral neuropathy; Ca Calcium; DF Diabetic foot; DN diabetic nephropathy; OR Odds ratio; RR Relative risk