Therapeutic effects of dietary antioxidative supplements on the management of type 2 diabetes and its complications; umbrella review of observational/trials meta-analysis studies

Abstract

Background

Controversial data on the effects of vitamins (V) and nutrients on the management of type 2 diabetes mellitus (T2DM) is available. Thus, it is aimed to clarify the role of vitamins and nutrients through an umbrella review regarding the available observational/ trials meta-analyses.

Methods

All meta-analyses of observational and clinical trials conducted on the impact of vitamins and nutrients in T2DM published until 5^th June 2021 in PubMed or Web of Sciences were included in this review. Also, the meta-analysis on children, pregnant women, type 1 DM, or in vivo/in vitro studies was excluded. Search results were reported according to the Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) flowchart.

Results

The overall of 93 papers (99 studies) consisting of 75 trials and 24 observational studies were included. Most studies were conducted on the effect of VD and alpha-lipoic acid (ALA) in both genders. Consumption of VD or its analogous; 20 IU/d to 450,000 IU/once for 3 weeks to 7 years showed to have a positive effect on the parameters of glucose hemostasis. Moreover, an inverse association was observed between VD level and T2DM risk. Daily consumption of 1200 mg VC for at least 12 weeks improved lipid profile and glucose hemostasis parameters. Furthermore, VB and medications for diabetic polyneuropathy (DPN) increased nerve conduction velocity. Vitamins K and E were showen to not have significant impact on T2DM. ALA had a beneficial effect on DPN symptoms after 2–4 weeks of intake of at least 300 mg/d. T2DM risk was reduced by doubling ALA intake. The effective daily doses of chromium, zinc and coenzyme Q10 on lipid profile and glucose hemostasis parameters were > 200 mg, < 25 mg, and < 200 mg, respectively.

Conclusion

This umbrella review suggests that dietary vitamins and nutrients can result in protective impacts the complications associated with T2DM. However, due to discrepancies between the results of the trials and observational studies is essential to conduct long-term high-qualified studies to prove the beneficial therapeutic effects of the vitamins and nutrients on T2DM and its complications.

Background

Diabetes mellitus (DM) is a chronic multifactorial metabolic disorder with an increasing global prevalence, which is expected to rise from 537 million cases in 2021 to 783 million by 2045 [1]. It has been established that oxidative stress plays a crucial role in the pathogenesis of DM and its complications through cellular damage [2]. Under cellular damage condition, reactive oxygen species (ROS) are produced, which can be suppressed by the antioxidative defense system [3], and can be modified by enzymatic antioxidants such as glutathione peroxidase, or non-enzymatic antioxidants including vitamins and minerals [4].

Various studies have been conducted on patients with DM to investigate the effects of antioxidant vitamins on ROS. Also, numerous epidemiological studies showed significant negative associations between serum concentrations of these antioxidants and risk of diabetes and its complications [5, 6]. However, some findings are inconsistent, and the positive effects of antioxidants are far from exclusive [7, 8]. Consumption of vitamin D (VD) and its analogous improved glycemic status [9], however another meta-analysis study did not show the same beneficial impact [10]. Similarly, inconsistent results were reported in other meta-analysis studies of observational studies [11, 12], which assessed the effect of VD on the risk of type 2 diabetes mellitus (T2DM). The results of two meta-analyses of the trials conducted on vitamin C consumption were contradictory, too [13, 14]. A nearly 6 month trial with chromium, improved lipid profile and fasting blood sugar (FBS) [15, 16]. Still, another meta-analysis study showed that these factors were not improved after 7-month supplementation with chromium [17].

Meta-analysis studies earn the highest rank in the evidence-based approach [18]. Although several systematic reviews have been conducted on specific nutrients, an umbrella review of them can present an overview of the results and conclude the main findings. Therefore, we aimed to critically assess all available meta-analysis studies on the effects of the vitamins and nutrients for the management of DM, focusing on T2DM.

Material and methods

Data searches & study selection

All relevant meta-analyses of observational studies and clinical trials that had reviewed the effects of the most frequently studied antioxidants (vitamins A, B, C, D, E, K, and nutrients including alpha-lipoic acid (ALA), zinc, copper, selenium, chromium, and coenzyme Q10) on T2DM published until 5^th June 2021 were included in this review. To obtain all relevant studies, PubMed and Web of Science databases were systematically searched using the following search terms: 'antioxidant', 'type 2 diabetes mellitus', 'vitamin', 'ascorbic', 'tocopherol', 'carotene', 'alpha-lipoic acid', 'retinol', 'zinc', 'copper', 'chromium', 'coenzyme Q10', and their equivalents, based on MeSH terms.

Eligibility criteria for inclusion were as follows.

• Population: Any meta-analysis of observational studies and clinical trials on T2DM.

• Intervention: Studies in which participants with T2DM received vitamins A, B, C, D, E, K, or nutrients including ALA, zinc, copper, selenium, chromium, and coenzyme Q10 at any concentration and for any duration.

• Comparator: Any studies compared the above antioxidants with placebo or active control.

• Outcomes: Any studies investigated any effect of the above antioxidants on risk, incidence, association, severity, or progression of T2DM and its complications

Meta-analyses conducted on children, pregnant women, or type 1 DM were excluded from this review. We used the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analysis) flowchart [19] to report our results. After excluding the duplicate publications, the titles and abstracts of the papers were examined. Moreover, hand searching was performed for full-text retrieval of the included studies' references. Two investigators performed all steps independently, and a third reviewer resolved any disagreement. Our search was limited to English language.

Data extraction and quality assessment

The following data were extracted; first author's name, year of publication, patients' characteristics (sample size, age, and sex), type of study and number of studies included in the meta-analysis, type of the antioxidant, dose, and duration of treatment, outcome measures, significant results, relative risk (RR) of diabetes, odds ratio (OR), effective dose, and quality assessment.

Results

Overall, 93 articles were included in the present umbrella review based on the PRISMA flowchart (Fig. 1). The details of the included meta-analysis studies are shown in Tables 1, 2, and 3.

Fig. 1.

Flow diagram of the study selection process

Table 1.

Characteristics of antioxidative vitamins’ effects on type 2 diabetes mellitus and its complications in meta-analysis of clinical trials

Author	Type of antioxidant/ Control	meta-analyzed studies (n)/ Disorders	Participants			Intervention		Outcome measures	Significant results	Effective dose	Quality assessment
			Sample size (n)	Age (yr)	Sex	Dose/ Frequency	Duration
Vitamin K
Shahdadian et al. 2018 [20]	VK alone or as multivitamin/ Placebo	5/ Pre-diabetes, Healthy	533	20–80	Both	500 µg/d- 90 mg/d	1w-3 yr	FBS, FI, HOMA	NO	NA	Yes
Vitamin B
Sawangjit et al. 2020 [21]	Single Mecobalamin or combination with other treatments/ Placebo or active control	15/ DPN	1707	47–67	Both	0.5–2/d (varied mode: oral, IM, IV)	2-24w	Pain (VAS, NRS), numbness, tingling weakness, tendon reflex, nerve conduction velocity (NCV)	↑clinical therapeutic efficacy by alone or combination, ↑NCV by combination	No	Yes
Raval et al. 2015 [22]	Single or multiple VB derivatives/ Placebo or active control	9/ T1DM, T2DM	1354	30–75	Both	varied: for single VB: 500 µg/d- 300 mg/d	2-36 m	Kidney function, All-cause and Cardiovascular mortality, BP	NO	NA	Yes
Vitamin D
AlAnouti, et al. 2020 [23]	VD/Placebo	4/ MetS	303	40–65	Both	Low dose: 2000 IU/d-20000 IU/w, High dose: 50,000 IU/w	8w-1 yr	Lipid profiles, VD serum	↑TGs by low dose and high dose VD	NO	Yes
Emadzadeh et al. 2020 [24]	VD-fortified food/Placebo, Ca	11/T2DM, Healthy	238	18–75	Both	100–28,000 IU/d	8-24w	HbA1c, FBS, insulin, HOMA	↓FBS, ↓Insulin, ↓HOMA	NO	Yes
Yammine et al. 2020 [25]	VD/ (before-after, or quasi-randomized controlled trial)	4/DPN	364	52.3–62	Both	Varied, 2059/d-50000/w	8–20 w	Serum VD, HbA1c, MPQ score	↑serum VD, ↑MPQ score	NO	Yes
Zhang et al. 2020 [26]	VD2 and VD3/ Placebo	8/ Pre-DM, VD deficiency, Obesity	4896	46–62	Both	Varied, 4000 IU/d-88865 IU/w	6–60 m	New-onset T2DM, reversion prediabetes to normal	↓risk T2DM, ↑reversion prediabetes to normo-glycemia	NO	Yes
Asbaghi et al. 2019 [5]	VD3-Co-Ca/ Placebo	12/ T2DM, Healthy	4395	24–72	Both	200–7142.8 IU/d	8w-6.5 yr	FBS, HbA1c, HOMA, FI, Insulin sensitivity	Totally: ↓FBS, ↓HOMA, ↓FI, Subgroup analysis: ↓FBS by ≤ 12w treatment, ↓HOMA and ↓FI by each duration	↓FBS, and ↓HbA1c, and ↓HOMA by high dose (Ca ≥ 1000 mg/d + VD ≥ 800 IU/d)	Yes
Gupta et al. 2019 [27]	VD and its analogues/ Placebo	9/ T1DM, T2DM	734	Mean: 57	Both	varied	8w-6 m	UACR, UAER, 24 h-UP, UPCI, serum: Ca, Cr, PTH, 25-OHD, 1,25-OHD	NO	NA	Yes
Hu et al. 2019 [28]	VD3/ Placebo	19/ Community population	1374	UN	UN	2000 IU/d-300000 IU/ Once	4w-12 m	FBS, HbA1c, HOMA, FI	Totally: NO, Subgroup analysis: ↓HbA1c, ↓HOMA, ↓FI in duration < 6 m	NO	Yes
Sahebi et al. 2019 [29]	VD fortified food/ Placebo, Ca or non-fortified	37/ T1DM, T2DM, GDM	1673	UN	Both	UN	UN	FBS, HbA1c, HOMA, hs-CRP, serum VD	Totally: ↓serum VD in diabetics vs. healthy, Subgroup analysis: ↓FBS, ↓HbA1c, ↓HOMA	NO	Yes
Wang et al. 2019 [30]	Varied VD types/ Placebo, non-treatment	20/ T1DN, T2DN	1464	35–72	UN	10–7142.8 IU/d	8w-6 m	FBS, HbA1c, renal function, inflammatory biomarkers	↓24 h-UP, ↓UAER, ↓hs-CRP, ↓TNF-α, ↓IL-6	No-difference between dose or treatment duration	Yes
He et al. 2018 [31]	VD/ Placebo, Non-VD	23/ NGT, Pre-DM	UN	UN	UN	UN	UN	T2DM risk, FBS, HOMA, serum VD	Totally: NO, Subgroup analysis: RR in pre-DM: 0.84, ↓FBS in pre-DM or BMI < 25 or serum VDinsufficien,↓HOMA in sufficient serum VD, ↓T2DM risk in pre-DM or overweight	VD > 2000 IU/d for prevention T2DM	Yes
Li et al. 2018 [32]	VD3, VD2/ Placebo	20/ Community population	2703	48–67	Both	VD3: 20 IU/d-300000 IU/Once, VD2:100000 IU/Once	2-6 m	FBS, HbA1c, HOMA, FI, serum VD	Totally: ↑serum VD, ↓HOMA, Subgroup analysis: ↓FBS in Middle East, > 2000 IU/d, duration < 3 m, baseline VD serum < 50 nmol/L, ↓HbA1c in obese	> 2000 IU/d	Yes
Mansournia et al. 2018 [33]	VD3/ Placebo	33/ T2DM, GDM, Diabetes complications (mostly T2DM)	1053	20–75	Both	200 IU/d-300000 IU/once	6w-12 m	Biomarkers of inflammation and oxidative stress	Totally: ↓hs-CRP, ↓MDA, ↓NO, ↑TAC, ↑GSH Subgroup analysis: ↓hs-CRP in T2DM, or complications	NO	Yes
Tabrizi et al., 2018 [34]	VD2, VD3/ Placebo	22/ T2DM, PCOS, CKD, overweight, CVD, MetS (mostly T2DM)	1799	27–89	Both	Varied, 2000 IU/d, 40,000/w, 60,000/m, 300,000 once	4–48 w	FMD, AI, PWV	↑FMD	< 3000 IU/d for dysglicemics non-DM	Yes
Yu et al. 2018 [35]	VD/ varied (placebo, Ca, VK)	13/ T2DM	875	21–75	Both	20 IU/d-50000 IU/w	8-52w	FBS, HOMA, serum VD, hs-CRP, TNF-α, IL-6	Totally: ↓hs-CRP, ↓borderline TNF- α, Subgroup analysis: ↓hs-CRP by dose ≤ 4000 IU/d	≤ 4000 IU/d	Yes
Krul-Poel et al. 2017 [36]	VD/ Placebo	23/ T2DM	1797	44–67	Both	400 IU/d-300000 IU/once	4w-12 m	FBS, HbA1c	Totally: NO, Subgroup analysis: ↓FBS when HbA1c ≥ 8%	NO	Yes
Lee et al. 2017 [37]	VD formulations/ Placebo, non-VD	22/ T2DM	2747	Mean: 48–70	UN	400 IU/d- 450,000 IU/ Once	8w-6 yr	FBS, HbA1c	Totally: ↓HbA1c, Subgroup analysis: ↓HbA1c by dose ≤ 2000 IU/d, or duration ≤ 12w	≤ 2000 IU/d	Yes
Mirhosseini et al. 2017 [38]	VD fortified food/Placebo	24/ T2DM, obese diabetics	1528	40–67	Both	400–8500 IU/d	2 m-12 m	FBS, HbA1c, HOMA, serum VD	Totally: ↓FBS, ↓HbA1c, ↓HOMA, ↑serum VD after intake, Subgroup analysis: ↓HbA1c in non-obese, ↓HOMA in non-obese, ↓FBS &↓HOMA by serum VD ≥ 50 nmol/l	≥ 4000 IU/d	Yes
de Paula et al. 2017 [39]	VD3, VD2/ Placebo or non-VD	7/ T2DM	542	Mean: 50.7–66.8	Both	D2:100,000 IU/Once, D3: 6250–50,000 IU/w	3-4w	BP	↓DBP	NO	Yes
Wu et al. 2017 [40]	VD formulations/ Placebo or non-VD	26/ T2DM	1877	42.4–67	Both	20 IU/d- 300,000 IU/ Once	4-48w	FBS, HbA1c	Totally: ↓HbA1c, Subgroup analysis: ↓HbA1c in VD deficiency, ↓FBS in VD deficiency, ↓HbA1c in non-obese, ↓FBS by VD dose ≥ 1000 IU/d, ↓HbA1c by ≥ 12w treatment	≥ 1000 IU/d	Yes
Jafari et al. 2016 [41]	VD supplement or food fortified/ Placebo	17/ T2DM	1365	UN	UN	20–7143 IU/d	8–48	Lipid profile	Totally: ↓lipid profile, Subgroup analysis: ↓TC by base serum VD ≥ 25 nmol/l, or food fortified, or VD dose ≤ 2000 IU/d, ↓TGs by food fortified, or VD dose ≤ 2000 IU/d, ↓LDL-C at base serum VD ≥ 50 nmol/l, or ≤ 12w treatment, ↓HDL-C by VD supplement, or base serum VD > 75 nmol/l, or ≤ 12w treatment	≤ 2000 IU/d	Yes
Lee et al. 2016 [42]	VD2, VD3/ Placebo or non-VD	15/ T2DM	1134	21–75	Both	D2: 100,000 IU/Once, D3:1000 IU/d-300000 IU/Once	8w-12 m	BP	↓DBP	NO	Yes
Poolsup et al. 2016 [9]	VD or VD analogues/ Placebo or Ca or non-treatment	10/ Pre-DM	1671	20–80	Both	400 IU/d-300000 IU/Once	2 m-7 yr	FBS, HbA1c, 2 h-OGTT, HOMA	Totally: ↓FBS, ↓HbA1c, Subgroup analysis: ↓HOMA in base serum VD ≥ 50 nmol/l, ↓2 h-OGTT in base serum VD < 50 nmol/l	NO	Yes
Chokhandre et al. 2015 [43]	VD & its analogues/ Placebo	6/ T2DN	584	UN	UN	0.5 µg/d-50000 IU/w	8-24w	UACR, albuminuria, eGFR, VD serum level, Ca, HbA1c	Significant improvement in kidney function	NO	Yes
Derakhshanian et al. 2015 [44]	VD derivatives/ Placebo	4/ T2DN plus VD deficient	219	57–60	UN	20–7142.8 IU/d	6-24w	UACR	NO	NA	Yes
Seida et al. 2014 [8]	VD3/ Placebo or non-VD	35/ T2DM, NGT, Pre-DM	43,407	27–77	Both	125 IU/d-300000 IU/Once	4w-3 yr	IR, Insulin secretion, HOMA, HbA1c	Totally: NO, Subgroup analysis: ↓FBS in non-DM	NA	Yes
Zhao et al. 2014 [45]	VD3/ Placebo, blank control	20/ T2DN	1497	35–75	Both	UN	4-48w	24 h-UP, UACR, HbA1c, BP	↓24 h-UP, ↓UACR	UN	Yes
George et al. 2012 [10]	VD3 or VD derivatives/ Placebo	15/T2DM, NGT, IGT	40,479	26–77	Both	0.25 µg/d-200000 IU/d	7d-7 yr	FBS, HOMA, HbA1c	Totally: NO, Subgroup analysis: ↓FBS & ↓HOMA in DM or IGT	NO	Yes
Mitri et al. 2011 [46]	VD formulations/ Placebo	11/ T2DM, Healthy	41,070	18–79	Both	400–8571 IU/d	8w-7 yr	FBS, HbA1c, HOMA	NO	NA	Yes
Vitamin C
Namkhah et al. 2021 [47]	VC/ Placebo	15/ T2DM	872	36–72	Both	200–3000 mg/d	2-48w	Lipid profile	↓ TGs, ↓TC	NO	Yes
Tareke et al. 2021 [13]	VC/ Placebo	11/ T2DM	606	37.8–60	UN	200–2000 mg/d	4-52w	Lipid profile, FBS, HbA1c	↓TGs, ↓FBS, ↓HbA1c	≥ 1000 mg/d, > 12w, < 52.8 yr age for TGs, < 1000 mg/d, > 12w, < 52.8 yr age for LDL-C, each doses, w, age for FBS, HbA1C	Yes
Ashor et al. 2017 [14]	VC/ Placebo	22/ T1DM, T2DM, Healthy	937	22–60	Both	72–6000 mg/d	1-120d	BG, Insulin serum, HbA1c	Totally: NO, Subgroup analysis: ↓BG in T2DM, ↓BG by > 30d VC, ↓insulin serum in T2DM, ↓insulin serum in fasting	NO	Yes
de Paula et al. 2017 [39]	VC/ Placebo	2/ T2DM	65	51.8–66.5	Both	500–1500 mg/d	3-4w	BP	↓DBP	NO	Yes
Khodaeian et al. 2015 [48]	VC/ Placebo	3/ T2DM	92	39–72	Both	200–1000 mg/d	4-16w	HOMA	NO	NA	Yes
Tabatabaei-Malazy et al. 2014 [6]	VC/ Placebo	4/ T2DM	316	29–75	Both	200–1250 mg/d	4w-3 m	FBS, FI, HbA1c, lipid profile, serum VC	↓FBS, ↓HbA1c, ↓TC, ↓LDL-c	At least 1250 mg/d, and for 3 m	Yes
Vitamin E
Khodaeian et al. 2015 [48]	VE/ Placebo	8/ T1DM, T2DM	425	20–75	Both	150–800 mg/d	4-27w	HOMA	NO	NA	Yes
Xu et al. 2014 [7]	VE/ Placebo or non-VE	14/ T2DM	714	Mean age: 48.5–71.3	Both	200–1600 IU/d	6-52w	FBS, FI, BS2hpp, HbA1c, BP, lipid profiles, HOMA	Totally: NO, Subgroup analysis: ↓HbA1c by dose > 400 mg/d, duration ≥ 12w, HbA1c < 8%, FBS < 8 mmol/l, normal baseline VE serum, ↓FBS in duration ≥ 12w,↓FI by dose > 400 mg/d, FBS ≥ 8 mmol/l	> 400 mg/d	Yes
Suksomboon et al. 2011 [49]	VE, tocotrienoles/ Placebo or non-treatment	9/ T1DM, T2DM	418	21–80	UN	6 mg/kg/d-1800 mg/d	2-12 m	HbA1c	Totally: NO, Subgroup analysis: ↓HbA1c when HbA1c ≥ 8%, or low baseline VE serum	NA	Yes
Mixture antioxidants
Balbi et al. 2018 [50]	VB, VC, VD, VE, Mixture/ Placebo	12/ T2DM	1450	40–77	Both	VB: 5 mg/d, VC:200–3000 mg/d, VD:500–200,000 IU/d, VE: 150–1200 mg/d, Mixture: varied	4–24 w	FBS, HbA1c, oxidative/antioxidative biomarkers	Totally: ↓FBS, ↑GPx,↓MDA, ↓TBARS, ↑TAC, ↑SOD, Subgroup analysis: ↓FBS and ↓HbA1c by VE	NO	Yes
Bolignano et al. 2017 [51]	Alone or combination of VA, VC, VE, Sel, Zn/ placebo, or non-treatment	14/ T1DN, T2DN	4345	18.9–65.4	Both	varied	4w-4.5 yr	UAER, eGFR, serum Cr, percent of dialysis	Totally: ↓UAER, Subgroup analysis: ↓UAER by VE	NO	Yes
Kandhare et al. 2017 [52]	VC, VE, Turmeric, Green tea, pyridoxine/ Placebo, conventional treatment	12/ T1DM, T2DM	1461	UN	Both	varied	4-52w	BUN, eGFR, serum Cr, urine TGF-β, UACR	↓BUN, ↓TGF-β, ↓eGFR	NO	Yes
Khodaeian et al. 2015 [48]	VE, VC, ALA, Zn, Chr/ placebo, sole intake vitamin or mineral	5/ T2DM	256	30–75	Both	Varied	6-24w	HOMA	NO	NA	Yes
Zheng et al., 2015 [53]	Breviscapine (B, a type flavonoid) + Mecobalamin (B12)/ Mecobalamin (M)	17/ DPN	1398	47–84	Both	B: 50–150 mg/d, M: 0.5–1.5 mg/d	2-6w	MNCV, SNCV	↑MNCV, ↑SNCV	NO	Yes
Montero et al. 2014 [54]	VE, VC/ Non-VE, Non-VC	9/ T2DM	296	Mean: 47.9–64.2	Both	VC: 800-3000 mg/d VE: 537-811 mg/d	2-52w	FMD	Totally: NO, Subgroup analysis: ↓FMD in non-obese T2DM	NA	Yes
Tabatabaei-Malazy et al. 2014 [6]	VE, EPA, Zn with-without VC/ Placebo	3/ T2DM	190	29–75	Both	Varied	4w-3 m	FBS, FI, HbA1c, lipid profile, serum VC	NO	NA	Yes
Akbar et al. 2011 [55]	VC, VE/ Placebo or without intervention	14/ T1DM, T2DM	562	UN	UN	VC:100 mg-2 g/Once, VE:100-1800 IU/varied	VC:4–12 w, VE: 6–27 w	FBS, BS2hpp, HbA1c, FI	↓HbA1c	NO	Yes

Legend: Chr Chromium; VC Vitamin C; VE Vitamin E; T2DM Type 2 diabetes mellitus; Pre-DM Prediabetes; NIDDM Non-insulin dependent; W Week; FPG Fasting plasma glucose; HbA1c Glycosylated hemoglobin; TGs Triglycerides; NGT Normal glucose tolerance; IGT Impaired glucose tolerance; UN Unknown; m Month; FBS Fasting blood sugar; BS2hpp Blood sugar 2h post prandial; IS Insulin sensitivity; D Day; FI Fasting insulin; IU International unit; HOMA Homeostasis model assessment index; Yr Year; IR Insulin resistance; BP Blood pressure; NA Not applicable; CAD Coronary artery diseases; DBP Diastolic blood pressure; GPx Glutathione peroxidase, MDA Malondialdehyde; TBARS Thiobarbituric acid reactive substances; TAC Total antioxidant capacity; SOD Superoxide dismutase; Zn Zinc; ALA Alpha lipoic acid; Sel Selenium; LDL-C Low density lipoprotein cholesterol; CRP C-reactive protein; IL-6 Interleukin-6; VK vitamin K; EF Endothelial function; EPA Eicosapentaenoic acid; Ca Calcium; 2h-OGTT 2-h oral glucose tolerance test; GDM Gestational diabetes mellitus; UACR Urine albumin-to-creatinine ratio; UACR Urine albumin creatinine ratio; UAER Urinary albumin excretion rate, PTH Parathyroid hormone; UPCI Urine protein creatinine index; EF Endothelial function; DN Diabetic nephropathy; DPN Diabetic peripheral neuropathy; NCV Nerve conduction velocity; MPQ McGill Pain Questionnaire score, FMD Flow-mediated dilation; PWV Pulse wave velocity; AI Augmentation index; MNCV Motor nerve conduction velocity; SNCV sensory nerve conduction velocity

Table 2.

Characteristics of antioxidative micronutrients’ effects on type 2 diabetes mellitus and its complications in meta-analysis of clinical trials

Author	Type of antioxidant/ Control	meta-analyzed studies (n)/ Disorders	Participants			Intervention		Outcome measures	Significant results	Effective dose	Quality assessment
			Sample size (n)	Age (yr)	Sex	Dose/ Frequency	Duration
Chromium
Asbaghi et al. 2021 [15]	Chr/ Placebo	24/ T2DM	1418	35–73	Both	42–1000 µg/d	6-25w	Lipid profiles	↓TGs, ↓TC, ↑HDL-C	> 200 µg/d for TGs, TC, ≤ 200 µg/d for LDL-C	Yes
Zhao et al. 2021 [56]	Chr (Different formulations)/ Placebo	10/ T2DM	509	35–68	Both	42–1000 µg/d	90–175 d	HbA1c, FBS, Lipid profile	↓HbA1c	NO	Yes
Asbaghi et al. 2020 [57]	Chr/ Placebo	23/ T2DM	732	28–68	Both	50–1000 µg/d	4-25w	FBS, Insulin, HbA1c, HOMA	↓ FBS, ↓HbA1c, ↓HOMA, ↓Insulin	for all: > 200 and ≤ 200 mg/d	Yes
Yin et al. 2015 [58]	Chr (Different formulations)/ Placebo	14/ T2DM	875	30–83	Both	23.2–1000 µg/d	8-24w	HbA1c, FBS	Subgroup analysis: ↓FBS	No	Yes
Suksomboon et al. 2014 [16]	Mono Chr or mixed with VC, VE, or biotin/ Placebo	25/ T2DM, NIDDM, Healthy	1641	> 18–83	Both	1.28–1000 µg/d	3–24 w	FPG, HbA1c, Lipid profile	↓FPG, ↓HbA1c, ↓TGs	> 200 µg/d	Yes
Patal et al. 2010 [17]	Chr/ Placebo	6/ T2DM	671	UN	UN	200–1000 ug/d	3-7 m	FBS, FI, HbA1c, 2hpp-Bs, lipid profile	Totally: NO, Subgroup analysis: ↓FBS, ↓HbA1c	NO	Yes
Balk et al. 2007 [59]	Chr (Different formulations)/ Placebo	41/ T2DM, NGT, IGT	1198	UN	UN	1.28–1000 µg/d	3w-8 m	FBS, HbA1c, Bs2hpp, lipid profile, IS	In T2DM: ↓FBS ↓HbA1c	For FBS: 1000 µg/d, For HbA1c: 200 µg/d	Yes
Althuis et al. 2002 [60]	Chr (Different formulations)/ Placebo	15/ T2DM, IGT, Healthy	618	18–93	Both	10.8–1000 µg/d	28d-16 m	FBS, Glucose at 120 min, HbA1c, FI, Insulin at 120 min	NO	NO	Yes
Zinc
Jayawardena et al. 2021 [61]	Zn with-without other micronutrients/ Placebo	3/Pre-diabetes	265	UN	Both	20–30 mg/d	6-12 m	FBS, BS2hpp, HbA1c, HOMA, CRP, lipid profiles	↓FBS, ↓BS2hpp, ↓TC, ↓HDL-C, ↓HOMA, ↑zinc serum	No	Yes
Pompano et al. 2021 [62]	Zn/ Placebo	27/ T2DM, CVD, obese, healthy	2016	20–70	Both	9.8–75 mg/d	4-50w	FBS, HbA1c, HOMA, lipid profile	Low dose (< 25 mg/d): ↓FBS, ↓TC, ↓TGs, ↓LDL-C High dose (≥ 25 mg/d): ↓HbA1c, ↓HOMA	< 25 mg/d and ≥ 12 w	Yes
Asbaghi et al. 2020 [63]	Zn/ Placebo	9/T2DM	427	48–66	Both	30–660 mg/d	6-52w	Lipid profiles	Totally: ↓TGs, ↓TC Subgroup analyses: ↓HDL-C	For TGs, LDL-C: < 100 mg/d, For TC, HDL-C each doses	Yes
Wang et al. 2019 [64]	Zn formulations/ Placebo	32/ T2DM, Obese	1700	18–75	Both	5-660 mg/d	1-12 m	FBS, BS2hpp, HbA1c, FI, HOMA, hs-CRP	Totally: ↓FBS, ↓Bs2hpp, ↓FI, ↓HOMA, ↓HbA1c, ↓hs-CRP Subgroup analysis: ↓FBS in T2DM vs. others, ↓FBS by inorganic Zn vs. organic Zn	NO	Yes
Capdor et al. 2013 [65]	Zn formulations/ Placebo	14/ T2DM, non-DM	3978	Infants-68	Both	3-240 mg/d	1.4–390 w	FBS, HbA1c, Serum insulin, serum Zn	Totally: ↓FBS, ↓HbA1c, ↑serum Zn, Subgroup analysis: ↓FBS in T2DM and high risk for DM vs. healthy	NA	NO
Jayawardena et al. 2012 [66]	Zn formulation with-without other antioxidants/ Placebo	25/ T1DM, T2DM	1317	Mean: 4.1–72.0	UN	7.5–660 mg/d	3w-5 yr	FBS, BS2hpp, HbA1c, lipid profiles, BP	↓FBS, ↓HbA1c, ↓TC, ↓LDL-C, ↓SBP, ↓DBP	NO	UN
Coenzyme Q10
Dludla et al. 2020 [67]	CoQ10/ Placebo	12/T2DM or MetS	650	46 -63	Both	20–400 mg/d	8-24w	FBS, insulin, HbA1c, Lipid profile, BMI	↓LDL-C, ↓TC	No	Yes
Suksomboon et al. 2015 [68]	CoQ10 or CoQ10 + fenofibrate/ Placebo	7/ DM	356	UN	UN	100 mg/ twice daily or 200 mg/d	3-6 m	FBS, HbA1c, lipid profile, BP	↓TGs by CoQ10 or CoQ10 + Fenofibrate, ↓TC by CoQ10 + Fenofibrate	NA	Yes
Moradi et al. 2016 [69]	CoQ10/ UN	14/ T2DM, non-DM	920	35–70	Both	100–300 mg/d	4-25w	FBS, HbA1c, FI	Totally: ↓FBS, Subgroup analysis: ↓FBS in duration < 20w, < 200 mg/d	< 200 mg/d	UN
Alpha lipoic acid
Ebada et al. 2019 [70]	ALA/Placebo	10/ T1DM, T2DM	553	46–72	Both	300–600 mg/d	3-24w	HbA1c, FBS, Lipid profile, HOMA, GPx	↑GPx	No	Yes
Rahimlou et al. 2019 [71]	ALA/ Placebo	41/ T2DM, T1DM, Overweight, MetS,	2564	15–74	Both	300–1200 mg/d	2-192w	HbA1c, FBS, TNF-α, IL- 6, CRP, insulin, HOMA	↓HbA1c, ↓FBS, ↓TNF-α, ↓IL-6, ↓CRP	< 600 and ≥ 600 mg/d for FBS, HbA1C, CRP, IL-6	Yes
Akbari et al. 2018 [72]	ALA alone or plus nutrients/ Placebo	24/ T2DM, non-DM	1537	16–92	UN	200–1800 mg/d	2-51w	FBS, FI, HbA1c, HOMA, lipid profile	Totally: ↓FBS, ↓FI, ↓HbA1c, ↓HOMA, ↓TGs, ↓TC, ↓LDL-C, Subgroup analysis: ↓FBS: in T2DM, by 600 mg/d, ↓FI: in non-DM, > 12w treatment, ↓HbA1c: by > 600 mg/d, > 12w treatment, by ALA with/ without nutrients, ↓HOMA: in T2DM, non-DM, by ≤ 600 mg/d, > 12w treatment, ↓TGs: in T2DM, by 600 mg/d, > 12w treatment, by combination, ↓TC: in T2DM, in non-DM, ≤ 600 mg/d, > 12w treatment, by sole ALA, ↓LDL-C: in T2DM, by 600 mg/d, < 8w treatment, by sole ALA	varied	Yes
Amato Nesbit et al. 2018 [73]	ALA/ Placebo	6/ DPN	1614	UN	UN	600–1800 mg/d	3w-4 yr	DPN symptoms	↓ Pain	NO	Yes
Wang et al. 2018 [74]	ALA + Epalrestat/ Epalrestat	12/DPN	813	40–87	UN	ALA:300 or 600 mg/d Epalrestat: 50 mg/tid	14-28d	MNCV, SNCV, TSS, TCSS	Totally: effective, Subgroup analysis: After 14, 21, 28d by 300, or 600 mg ALA: ↑Median & peroneal MNCV, ↑median & peroneal SNCV, After 21d by 600 mg ALA: ↓TCSS, ↓TSS	600 mg/dALA after 14d, 28d	Yes
Cakici et al. 2016 [75]	ALA/ Placebo	6/ DPN (T1DM, T2DM)	885	56–60	UN	100–1800 mg/d	30 min IV-6 m oral	TSS	↓TSS	600 mg/d oral or IV	Yes
Snedecore et al. 2013 [76]	ALA, non-ALA/ Placebo, other treatments	58/ painful DPN	1183	Mean: 53–71	Both	Varied (ALA: 600–1800 mg/d)	4-27w	Pain reduction	NO (for ALA)	NA	Yes
Han et al. 2012 [77]	ALA/ non-ALA	15/ DPN	1058	Mean: 43.9–66.0	UN	300–600 mg/d	14-28d	Efficacy treatment, MNCV, SNCV	↑efficacy,↑MNCV, ↑SNCV	300–600 mg/d for 2-4w	Yes
Mijnhout et al. 2012 [78]	ALA/ Placebo	4/ DPN	653	18–74	UN	100–1800 mg/d	3-5w	TSS	↓TSS	600 mg/d for 3w	Yes
Ziegler et al. 2004 [79]	ALA/ Placebo	4/ DPN	1258	48–65	Both	600 mg/d IV	> 3w	TSS, NISLL	↓TSS, ↓NISLL	600 mg/d IV > 3w	Yes

Legend: Chr Chromium; VC Vitamin C; VE Vitamin E; T2DM Type 2 diabetes mellitus; NIDDM Non-insulin dependent; W Week; FPG Fasting plasma glucose; HbA1c Glycosylated hemoglobin; TGs Triglycerides; NGT Normal glucose tolerance; IGT Impaired glucose tolerance; UN Unknown; m Month; FBS Fasting blood sugar; BS2hpp Blood sugar 2 h post prandial; IS Insulin sensitivity; D Day; FI Fasting insulin; IU International unit; HOMA Homeostasis model assessment index; Yr Year; IR Insulin resistance; CoQ10 Coenzyme Q10; BP Blood pressure; NA Not applicable; CAD Coronary artery diseases; DBP Diastolic blood pressure; GPx Glutathione peroxidase, MDA Malondialdehyde; TBARS Thiobarbituric acid reactive substances; TAC Total antioxidant capacity; SOD Superoxide dismutase; Zn Zinc; ALA Alpha lipoic acid; Sel Selenium; DPN Diabetic peripheral neuropathy; MNCV Motor nerve conduction velocity; SNCV Sensory nerve conduction velocity; TCSS Toronto clinical scoring system; TSS Total symptom score; LDL-C Low density lipoprotein cholesterol; CRP C-reactive protein; IL-6 Interleukin-6; NISLL Neuropathy impairment score lower limbs; GPx Glutathione peroxidase; TNF-α Tumor necrosis factor alpha.

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

Studies' characteristics

Most of the included studies were meta-analyses of clinical trials conducted on vitamins, antioxidative nutrients, and their combinations. Twenty-four meta-analyses were performed on observational studies investigating vitamins or antioxidative nutrients. Most of the included studies were shown on both genders. Most meta-analysis studies have compared the effects of vitamins and placebos on the prevention or progression of diabetes and its complications. Quality assessment or adjustments with age and sex were performed for nearly all the of included meta-analyses. Details of the included studies are described as follows.

Influence of vitamins on the prevention or progression of DM and its complications

Overall, 61 papers (65 meta-analyses) were included in this topic; 48 and 17 studies were clinical trials and observational studies, respectively. The most commonly studied antioxidant was VD, which was examined in 45 meta-analyses including 28 meta-analyses of clinical trials, and 17 meta-analyses of observational studies. These figures were followed by eight studies on combining vitamins or other nutrients, six studies on VC, three on VE, two on VB, and one on VK in the meta-analysis of clinical trials. No meta-analyses of clinical trials or observational studies were found on either VA or beta-carotene.

The main factors that were measured from the outcome of meta-analyses on VD trials were FBS as the glycemic index, followed by fasting insulin (FI), Homeostasis Model Assessment (HOMA) index [5, 8–10, 24–26, 28, 29, 31, 32, 38, 46], renal function [27, 30, 43–45], biomarkers of inflammation or oxidative stress [29, 30, 33, 35], flow-mediated dilation (FMD) [34], and lipid profile [23, 41]. VD supplementation was associated with significant improvements in FBS, HbA1c, HOMA index, inflammatory and oxidative stress biomarkers, but a substantial acceleration in FMD. However, the effect of VD consumption on renal function or lipid profile in patients with DM was inconsistent. Also, studies that had assessed the impact of VD intake on blood pressure (BP) [39, 42] showed a significant reduction in diastolic blood pressure (DBP). Overally, the dose of the VD taken varied from 20 IU/d to 450,000 IU/once, and VD was administered for 3 weeks to 7 years. Meta-analyses of observational studies on VD had assessed the association between serum levels of VD and the risk of T2DM [11, 12, 83, 84, 89–92], or its complications; insulin resistance [80], diabetic foot (DF) [81], diabetic retinopathy (DR) [85, 87], diabetic nephropathy (DN) [44], and diabetic polyneuropathy (DPN) [82, 86, 88]. All of the studies mentioned above reported significant associations between low serum levels of VD and the risk of T2DM and its complications.

The meta-analysis of clinical trials conducted on combinations of antioxidants had assessed their effects on FBS, HbA1c, FI, HOMA index, renal function, oxidative stress biomarkers, FMD, as well as the lipid profiles [50–52, 54, 55]. Significant improvements in all of the measured factors had been observed, excluding HOMA index [48], and lipid profiles [6]. The effect of VE on the improvement of FBS, HbA1c, and renal function was more significant than the other nutrients [50, 51]. A meta-analysis study of observational studies indicated that the highest intake of antioxidant mixtures, particularly VE vs. the lowest intake, reduced the risk of T2DM by 13% [99].

The outcome measures studied in meta-analyses of clinical trials on VC were FBS, HbA1c, FI, HOMA index, lipid profile, BP, and endothelial function (EF) [6, 13, 14, 39, 47, 48]. Significant reductions in FBS, HbA1c, FI, EF, DBP, total cholesterol (TC), and low-density lipoprotein cholesterol (LDL-C) were observed in these studies [6, 13, 14, 39]. FBS was significantly reduced when the patients were treated with VC for more than 30 days. The appropriate dose of VC for improving EF was significantly higher than 500 mg/d, and FBS, HbA1c, and lipid profiles were at least 1250 mg/d.

The meta-analysis of trials conducted on VE had assessed its effect on FBS, HbA1c, FI, HOMA index, BP, and lipid profiles [7, 48, 49]. A significant improvement in HbA1c and FI was observed for doses greater than 400 mg/d. A significant reduction in FBS was seen when patients were treated with VE for ≥ 12 weeks.

Although meta-analyses of clinical trials on multiple VK or VB derivatives did not recongnize any significant improvements in FBS, FI, HOMA index, renal function, and BP [20, 22], sole intake of mecobalamin or its combinations with other treatments or flavonoids has improved nerve conduction velocity [21, 53].

The influence of antioxidative nutrients on the prevention or progression of diabetes mellitus and its complications

Thirty-four papers (34 meta-analyses) were included in this review. Of these, 27 and 7 were meta-analyses of clinical trials and observational studies. The most commonly studied antioxidant was ALA, examined in 10 meta-analyses of clinical trials and one meta-analysis of observational studies. These figures were followed by eight studies on chromium (Cr), six studies on zinc (Zn), and three studies on coenzyme Q10 (CoQ10) in meta-analyses of clinical trials. One meta-analysis of observational studies was conducted on ALA, selenium, and copper (Cu) for each of these nutrients, and three meta-analyses were performed on Zn.

Nearly all meta-analysis studies of clinical trials on ALA [73–79] had been conducted on patients with DPN. ALA showed significant beneficial effects on DPN symptoms after 14, 21, or 28 days and doses of 300 mg/d and 600 mg/d. In addition, considerable improvements in FBS, FI, HbA1c, HOMA index, TC, triglycerides (TGs), and LDL-C by ALA was seen in patients with T2DM or non-diabetic subjects [71, 72]. Beneficial effects of ALA on inflammatory and antioxidative biomarkers were observed in two meta-analyses [70, 71]. A meta-analysis of an observational study on ALA indicated a significant reduction in T2DM risk by doubling its intake [93].

The main outcome measures in meta-analyses of trials conducted on the sole intake of Cr were FBS, HbA1c, FI, and lipid profiles [15–17, 56–60]; the HOMA index was the main outcome in studies on its combinations with vitamins [48, 57]. Cr supplementation was associated with a significantly reduction of FBS, HbA1c, and TGs. The appropriate dose of Cr intake for the significant decrease in HbA1c and TGs was ≥ 200 µg/d, but the significant reduction of FBS varied from > 200 to 1000 µg/d. Combining Cr with VC, VE, ALA, and Zn did not show any beneficial effects on the HOMA index.

In a meta-analysis of clinical trials that examined the effect of Zn [61–63, 65, 66, 74], significant reductions were observed in the serum levels of FBS, FI, HbA1c, HOMA index, TC, LDL-C, and high-sensitive C-reactive protein (hs-CRP). Meta-analyses by Capdor et al., 2013 and Jayawardena et al., 2021 reported a significant rise in serum levels of Zn after supplementation and subsequent significant reductions in FBS and HbA1c. In three meta-analyses of observational studies on Zn [96–98], a significantly low intake of Zn was observed in patients with T2DM complications, and the risk of T2DM was reduced by increasing the Zn intake. No significant improvements in FBS, FI, HbA1c, HOMA index, and lipid profile were observed upon combining Zn with other nutrients or vitamins [6, 48, 51]. However, a significant reduction in urinary albumin excretion rate (UAER) was reported in combinations of Zn with other antioxidative agents [51].

The effect of CoQ10 intake was assessed in three meta-analyses of clinical trials [67–69] through the measurement of FBS, HbA1c, FI, lipid profiles, and BP. CoQ10 supplementation improved FBS at doses lower than 200 mg/d or used for less than 20 weeks. In addition, it showed beneficial effects on TC and TGs levels when combined with fenofibrate.

A meta-analysis of observational studies on selenium [74] reported a significant U-shape association between serum levels of selenium (< 97.5 and > 132.5 µg/L) and T2DM risk. The combination of selenium and other antioxidants significantly reduced UAER [51]. The meta-analysis of observational studies [95] reported a significantly higher level of Cu in the serum and plasma of patients with DM compared with healthy individuals.

Discussion

Our study revealed that low serum levels of VD and high serum levels of Cu and Zn are associated with an increased risk of T2DM and its complications. Furthermore, supplementation of vitamins C, D, E, and nutrients Cr, Zn, CoQ10, and ALA can significantly impact FBS and glycemic control. Thus, it can result in beneficial impacts in patients with diabetes with/without presenting further complications.

The beneficial effects of VD in T2DM may be partly attributed to its direct and indirect beneficial impacts on insulin resistance and improvement of insulin function. Also, VD is associated to be beneficial for the systemic inflammation caused by increased insulin secretion through its specific receptors in pancreatic β-cells, reduced activity of NF-κB, and increased expression of calbindin against the apoptosis initiated by cytokines. The other uses of VD stem from its antioxidative effects, restricting oxidative biomarkers or elevating antioxidative biomarkers [92]. Furthermore, a link between VD level and the renin-angiotensin system, endothelial function, and blood pressure was indicated, all of which can be explained by the association between low levels of VD and the risk of T2DM and its complications [8, 36, 100]. However, most of the included meta-analyses recommended no effective VD dose to prevent or treat T2DM. A meta-analysis of 12 clinical trials on 4395 samples recommended a high daily dose of VD (≥ 800 IU/d) in combination with Ca ≥ 1000 mg/d for ≤ 12 weeks for the improvement of FBS, HbA1c, and the HOMA index [5]. However, numerous studies [37, 38, 40, 41] have recommended an effective dose for VD only –without Ca- ranging from ≥ 1000 to ≤ 4000 IU/d for ≤ 12 weeks to improve FBS, HbA1c, and HOMA index in those with deficient or sufficient VD levels. He et al. 2018 and Li et al. 2018, suggested concentration of > 2000 IU/d for < 3 months duration for the prevention of T2DM in pre-diabetics patients with insufficient serum levels of VD. The effective anti-inflammatory concentration of VD reported by Yu et al. 2018 was to be ≤ 4000 IU/d. In a meta-analysis, an improvement in lipid profiles was observed through the administration of an effective dose of ≤ 2000 IU/d for ≤ 12 weeks [41]. But another meta-analysis [23] reported a significant increase in TGs levels by intake of both low and high doses of VD (20,000 and 50,000 IU/week). In conclusion, the effect of VD supplements on lipid components and lipid profile differed based on the level of VD.

Non-enzymatic antioxidants such as vitamins C and E can scavenge ROS, protect against lipid peroxidation, inhibit protein kinase-C, decrease inflammatory activity, reverse β-cell apoptosis, and ameliorate insulin resistance [6, 48, 50]. The oxidative damage to cells and vessel walls [101] induced by thiobarbituric acid reactive substances (TBARS) and malondialdehyde (MDA) as end-products of lipid peroxidation, can be decreased by inhibiting the activity of free radicals or by reacting with free radicals after supplementation with antioxidant agents [2]. The recommended minimum effective dose of VC for improving FBS, HbA1c, and lipid profile was 1250 mg/d for 3 months [6, 13]. Supplementation of VE by doses higher than 400 mg/d for ≥ 12 weeks was effective for improving HbA1c in T2DM patients with poor glycemic control or normal baseline levels of VE [7].

The beneficial effects of VK on insulin sensitivity, glucose homeostasis, and risk of diabetes have been reported through decreasing the level of cytokines and inflammatory markers [102]. It has been postulated that vitamin K-dependent bone protein osteocalcin can influence insulin sensitivity by acting directly on pancreatic β-cells, increasing their proliferation and insulin secretion. However, we found no beneficial effects of VK on FBS, FI, HOMA index, and biomarkers of inflammation in our study [20].

VB and its derivatives showed anti-inflammatory effects and are also associated with reducing advanced glycation end (AGE) products and oxidative stress in renal cells [103, 104]. Those, as mentioned earlier, can result in a decreased incidence of albuminuria/proteinuria and prove effective in reducing kidney injury in diabetes. However, this effect was not confirmed by a meta-analysis of 9 clinical trials by VB and its derivatives on renal function [22]. The beneficial effects of single mecobalamin or its combinations with other medications have shown significant improvement in motor and sensory nerve conduction velocity of patients with DPN [21, 53].

Zn is an essential trace element that plays specific critical roles in the metabolism of carbohydrates via insulin synthesis, storage, crystallization, secretion by pancreatic β-cells, function, and translocation into the cells [105]. In addition, it demonstrated to have insulin-mimetic and anti-inflammatory activity; zinc suppresses nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) and protects against apoptosis of β-cells [97]. Although some studies have reported the beneficial effects of Zn supplementation on DN, DPN, peripheral insulin sensitivity, bio-inflammatory markers, DNA damage, and lipid peroxidation [106, 107], there is currently no evidence that supports the use of Zn supplements in the prevention of T2DM [108, 109]. The reason might be the failure of observational studies to differentiate between participants who received Zn alone and those used a combination of Zn with other nutrients [110, 111]. Thus, the interaction between these nutrients and Zn could have affected Zn intake and T2DM risk. Observational studies on the relationship between serum Zn levels and T2DM in Finnish men are contradictory [112] to the conclusions observed in Russian postmenopausal women [113]. The response mechanism against Zn excess in the serum/plasma (hyperzincuria) in diabetic patients [114] is consistent with the observed direct positive relationship between serum/plasma levels of Zn and T2DM. However, a significant linear negative association was reported between serum levels of Zn and the duration of diabetes [98].

Chromium is an essential element for the metabolism of carbohydrate and lipid, and its deficiency is associated with insulin resistance as well as diabetes [115]. It has been observed that Cr can increase the number of insulin receptors and insulin binding at its site of action. It can improve insulin sensitivity and decrease triglyceride levels via enhancing free fatty acid oxidation [94, 116]. However, its impact on glucose homeostasis parameters was controversial in our included meta-analyses. The suggested effective doses of Cr to improve FBS and HbA1c were 200–1000 µg/d and 200 µg/d, respectively [16, 57, 59]. Moreover, the recommended effective dose for improvement of TGs and TC was > 200 µg/d; however, to enhance the level of LDL-C, Cr was recommended at the concentration of < 200 µg/d [15]. Dietary Cr supplements are inexpensive and are safe even at high doses [117].

Inactivity of glycerol-3-phosphate dehydrogenase (G3PD), an electron transporter from the cytosol to the mitochondria, has been reported in diabetic patients. CoQ10 is a component of the mitochondrial respiratory chain that plays a key role in the electron transport chain in the absence of G3PD. In addition, it is a potent lipophilic antioxidant capable of regenerating and recycling other antioxidants such as tocopherol and ascorbates [118]. Although it has beneficial effects on insulin secretion and insulin sensitivity, the results of two meta-analyses were controversial regarding the impact of CoQ10 on glycemic control in diabetic patients [68, 69]. The observed mechanism of the beneficial effects of CoQ10 alone or in combination with fenofibrate on TGs [68] might be related to increased lipolysis of TGs [119].

ALA is commonly used to relieve neuropathic symptoms due to its beneficial effects on DPN via improved nerve blood flow and progress of nitric oxide-mediated endothelium-dependent vasodilation and anti-inflammatory anti-thrombotic activities [77]. Its appropriate dosage in most meta-analyses is usually recommended at 300–600 mg/d for 2–4 weeks [74, 75, 77–79]. The beneficial effects of ALA on the improvement of glycemic control of FBS, HbA1c, and lipid profiles, except for high-density lipoprotein cholesterol (HDL-C), might be related to elevated glutathione levels and the activity of antioxidant enzymes, the prevention of lipid peroxidation, suppression of NF-κB activity, and preservation of β-cell function via insulin sensitivity [120, 121]. Moreover, doubling ALA intake was negatively associated with the T2DM risk [93].

As one of the essential trace elements in maintaining human health, selenium is a crucial component of glutathione peroxidases (GPx). The overexpression of GPx in the islets of Langerhans can protect these cells from oxidative stress and improve their function [122]. Furthermore, a high concentration of selenium may interfere with insulin signaling. However, selenoproteins' anti-oxidative and anti-inflammatory effects have been previously proven [123, 124]. In several studies, lower selenium levels have been detected in patients with T2DM versus healthy individuals, which might be related to the inverse association between selenium and inflammation. Conversely, a higher selenium level in patients with T2DM than in healthy individuals has been reported elsewhere [125, 126], suggesting a U-Shape association between selenium and T2DM. This U-Shape association between selenium and T2DM was confirmed in a meta-analysis of five observational studies on selenium in 13,460 participants [94].

Although Cu is a key component of copper/zinc superoxide dismutase (Cu/Zn SOD) –a scavenger of free radicals- its antioxidative effect is usually known as a pro-oxidant [127]. Induction of oxidative stress and renal dysfunction in diabetic rats and raised lipid peroxidation in patients with T2DM after Cu supplementation [127] suggests that Cu does not protect against ROS. Moreover, higher concentrations of Cu can be replace with Zn and elevate the Cu-to-Zn ratio, resulting in impaired Zn metabolism, and inflammation and aggravation of oxidative stress, consequently [128, 129]. This evidence confirms Qiu et al., 2017 findings, which found higher serum levels of Cu in diabetic patients compared with healthy subjects.

Our study has particular strengths and limitations. The study's main strength is its review of systematic reviews and meta-analyses conducted on the most frequently studied antioxidants. The main limitation of the study is the controversy between the results of some meta-analyses due to differences in several parameters, including sample size, the subjects' health status, the quality score of studies, the supplemented dose, the methods used to measure vitamins/nutrients, not having evaluated the effect of confounding factors, and having not performed subgroup analysis, thereby leading to inconsistencies among the meta-analysis.

Conclusions

This review suggests that antioxidants may present some protective effects against T2DM and its complications and be helpful in the management of diabetes. However, due to the heterogeneity of results among the observational studies and clinical trials, caution should be taken in approving their beneficial effects on T2DM.

Abbreviations

AGE

Advanced glycation end

ALA

Alpha-lipoic acid

BP

Blood pressure

CoQ10

Coenzyme Q10

Cr

Chromium

Cu

Copper

Cu/Zn SOD

Copper/zinc superoxide dismutase

DBP

Diastolic blood pressure

DF

Diabetic foot

DM

Diabetes mellitus

DN

Diabetic nephropathy

DPN

Diabetic polyneuropathy

DR

Diabetic retinopathy

EF

Endothelial function

FBS

Fasting blood sugar

FI

Fasting insulin

FMD

Flow-mediated dilation

G3PD

Glycerol-3-phosphate dehydrogenase

GPx

Glutathione peroxidases

HDL-C

High-density lipoprotein cholesterol

HOMA

Homeostasis Model Assessment

hs-CRP

High-sensitive C-reactive protein

LDL-C

Low-density lipoprotein cholesterol

MDA

Malondialdehyde

NF-κB

Nuclear factor kappa-light-chain-enhancer of activated B cells

OR

Odds ratio

PRISMA

Preferred Reporting Items for Systematic reviews and Meta-Analysis

ROS

Reactive oxygen species

RR

Relative risk

T2DM

Type 2 diabetes mellitus

TBARS

Thiobarbituric acid reactive substances

TC

Total cholesterol

TGs

Triglycerides

UAER

Urinary albumin excretion rate

VD

Vitamin D

Zn

Zinc

Authors' Contributions

OTM and SN conceived and coordinated the study. OTM, SN, MA, and BL participated in the study's design. OTM, MP, and SHM extracted data from the published articles and drafted the manuscript. SN, MA, and BL helped to edit the manuscript draft. MP, OTM, and SN critically reviewed the manuscript and helped with quality assessment. All authors read and approved the final manuscript.

Funding

This study is an in-home study without any financial support.

Data Availability

Not applicable. All data analyzed in the current systematic reviews are extracted from published articles in PubMed, Web of Science, Scopus, and Cochrane Library databases.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing Interest

All authors declare that there are no conflicts of interest.