Vitamin D and Risk for Type 2 Diabetes in People With Prediabetes: A Systematic Review and Meta-analysis of Individual Participant Data From 3 Randomized Clinical Trials

We read with great interest the article by Pittas and colleagues evidencing that vitamin D decreases the risk of diabetes in patients with prediabetes [1]. We would like to emphasize another effective strategy, namely L-Arginine, which a clinical trial with a 9-year follow-up had proven to significantly reduce the risk of diabetes in patients with glucose intolerance [2]. The main advantages of L-Arginine vs Vitamin D include: the virtual absence of adverse effects, widely reported instead in vitamin D supplementation (including hypercalcemia, cardiac arrhythmias, and increased risk of falls, especially at high doses); a pathophysiological rationale, lying on the attenuation of endothelial function and oxidative stress.

Equally important, a recent meta-analysis substantiated that COVID-19 significantly increases the risk of diabetes [3]. Fully in line with our view, another double-blind randomized placebo-controlled clinical trial demonstrated that L-Arginine has beneficial effects in patients affected by COVID-19, reducing the length of hospitalization, the respiratory support, and the inflammation triggered by the cytokine storm [4]. Instead, Vitamin D supplementation was shown to be not associated to any COVID-19 risk reduction [5].

References

1-Pittas, A. G. et al. Vitamin D and Risk for Type 2 Diabetes in People With Prediabetes : A Systematic Review and Meta-analysis of Individual Participant Data From 3 Randomized Clinical Trials. Ann Intern Med 2023. doi: 10.7326/M22-3018.

2-Monti, L. D. et al. Decreased diabetes risk over 9 year after 18-month oral L-arginine treatment in middle-aged subjects with impaired glucose tolerance and metabolic syndrome (extension evaluation of L-arginine study). Eur J Nutr 2018;57:2805-2817.

3-Zhang, T. et al. Risk for newly diagnosed diabetes after COVID-19: a systematic review and meta-analysis. BMC Med 2022;20:444.

4-Trimarco, V. et al. Beneficial Effects of L-Arginine in Patients Hospitalized for COVID-19: New Mechanistic Insights from a Randomized Clinical Trial. Pharmacol Res 2023. doi: 10.1016/j.phrs.2023.106702.

5-Jolliffe, D. A. et al. Effect of a test-and-treat approach to vitamin D supplementation on risk of all cause acute respiratory tract infection and covid-19: phase 3 randomised controlled trial (CORONAVIT). BMJ 2022;378:e071230.

The systematic review and meta-analysis by Pittas et al., Vitamin D and Risk for Type 2 Diabetes in People With Prediabetes, demonstrated that higher-dose oral vitamin D (VitD) supplementation is associated with a reduced risk of T2DM conversion. Of interest, the authors report that this effect may be greater in “leaner” individuals, especially those with a BMI <31.3. To explain this variance, the authors chiefly speculate that obesity suppresses the activation of key CYPs (CYP2R1 and CYP27B1), which jointly function to endogenously convert VitD₃ into its biologically active form, 1,25-dihydroxy VitD₃. While this analysis from Pittas et al. is thoughtful, the explanation offered for the differential effects on younger, leaner subjects is likely misattributed to CYP regulation.

In both relative and absolute numbers, immune-mediated destruction in those diagnosed with T2DM is significant, with a median of ~10% of subjects in large, representative “T2DM cohorts” testing positive for one or more auto-antibodies including glutamate decarboxylase (GAD), islet cytoplasmic antibodies (ICA), insulinoma-associated protein 2 (IA-2A), or zinc transporter 8 (ZnT8). In general, such are given a diagnosis of Latent Autoimmune Diabetes (LADA).

Several prominent population studies have affirmed this finding. The UK Prospective Diabetes Study (1997) tested 3672 white patients with T2DM, aged 25 to 65 years, for autoantibodies (GAD, ICA); in newly diagnosed subjects, 10% tested positive for one or both.¹ Similarly, a 2013 European cohort of 6156 patients reported a positivity rate of >9.7% for at least one autoantibody (GAD, IA-2A, ZnT8A), finding that younger subjects that were leaner, insulin-dependent, and female were more likely to be diagnosed with LADA.² More recently, the Norwegian HUNT Study (2018) documented >7-8% positivity (GAD) in 2002 individuals recently diagnosed with diabetes.³ Altogether, there is likely a distinct subpopulation of those diagnosed with T2DM that have clinical markers of autoimmunity and, therefore, islet destruction that contributes to insulin-dependence.

VitD’s role in the primary prevention of autoimmune-associated disease (AAD), such as LADA, has demonstrated compelling results. Famously, in a trial of 38 recently diagnosed T1DM subjects, Gabbay et. al (2012) reported that in those randomized to 2000 IU of VitD per day, only 18.5% progressed to undetectable C-peptide levels versus  62.5% of controls.4

We applaud Pittas et al. for this systematic review and meta-analysis, while cautiously advocating for a more nuanced understanding of the differential effects on subpopulations of pre-diabetics at risk for T2DM, potentially driven by a distinct autoimmune pathology and attenuated by higher doses of VitD supplementation.  

References

1. Turner R, Stratton I, Horton V, et al. UKPDS 25: autoantibodies to islet-cell cytoplasm and glutamic acid decarboxylase for prediction of insulin requirement in type 2 diabetes. UK Prospective Diabetes Study Group [published correction appears in Lancet 1998 Jan 31;351(9099):376]. Lancet. 1997;350(9087):1288-1293. doi:10.1016/s0140-6736(97)03062-6

2. Zinman B, Kahn SE, Haffner SM, et al. Phenotypic characteristics of GAD antibody-positive recently diagnosed patients with type 2 diabetes in North America and Europe. Diabetes. 2004;53(12):3193-3200. doi:10.2337/diabetes.53.12.3193

3. Hjort R, Ahlqvist E, Carlsson PO, et al. Overweight, obesity and the risk of LADA: results from a Swedish case-control study and the Norwegian HUNT Study. Diabetologia. 2018;61(6):1333-1343. doi:10.1007/s00125-018-4596-0

4. Gabbay MAL, Sato MN, Finazzo C, Duarte AJS, Dib SA. Effect of Cholecalciferol as Adjunctive Therapy With Insulin on Protective Immunologic Profile and Decline of Residual β-Cell Function in New-Onset Type 1 Diabetes Mellitus. Arch Pediatr Adolesc Med.2012;166(7):601–607. doi:10.1001/archpediatrics.2012.164

We thank Jacob Hands and colleagues for their insightful comment on our article. Our hypothesis that suppression of CYP2R1 in people with obesity leading to reduced production of blood 25-hydroxyvitamin D can explain why vitamin D seems to work better in leaner people with prediabetes does not preclude other explanations. There are several lines of evidence supporting a beneficial role for vitamin D in pancreatic beta cell function, either directly by binding to β-cell vitamin D receptors or indirectly via regulating calcium flux through the β-cell, which is important for calcium-mediated insulin secretion (1). Hence, vitamin D may work better for diabetes prevention in leaner individuals because they are more likely to have predominantly impaired pancreatic β-cell function leading to insulin deficiency. Indeed, in the Diabetes Prevention with active Vitamin D (DPVD) study, eldecalcitol had a favorable effect on diabetes risk among those with insufficient insulin secretion (2). We agree with Hands and colleagues that pancreatic β-cell autoimmunity, which may be present in more than 10% of people diagnosed with type 2 diabetes (3), leading to pancreatic β-cell destruction is one pathway towards insulin deficiency, and high-dose vitamin D may attenuate the autoimmune-mediated pathology. Since none of the three trials measured circulating β-cell autoimmune markers, we cannot test the hypothesis that vitamin D benefited predominantly participants with prediabetes and evidence of β-cell autoimmunity. This would be a compelling area for future research to better identify sub-populations at risk for diabetes that are most likely to benefit from vitamin D.

References

1-Pittas, A.G. et al. The role of vitamin D and calcium in type 2 diabetes. A systematic review and meta-analysis. J Clin Endocrinol Metab. 2007;92(6):2017-29.

2-Kawahara, T. et al. Effect of active vitamin D treatment on development of type 2 diabetes: DPVD randomised controlled trial in Japanese population. BMJ 2022;377:e066222.

3-Brooks-Worrell, B. et al. Islet Autoimmunity is Highly Prevalent and Associated With Diminished beta-Cell Function in Patients With Type 2 Diabetes in the Grade Study. Diabetes 2022;71(6):1261-71.

Pittas et al. show the beneficial effect of vitamin D supplementation in lowering diabetes risk in pre-diabetes individuals, although the summary statistics reported in their meta-analysis apparently hide that ethnocultural background is an important modifier for the effect estimates. Of the three original trials pooled in the meta-analysis, only the one performed in a Japanese population indicates significant benefit (hazard ratio [HR]: 0.69 [0.5-0.95]); the two other trials in the US (0.88 [0.75-1.04]) and Norway (0.90 [0.69-1.18]) are null (1).

In our recent meta-analysis of 68 RCTs on vitamin D supplementation and cardiometabolic risk factors (2), ethnocultural differences were evidenced in that vitamin D decreased glycated hemoglobin in non-Westerners (-0.36 [-0.64 to -0.07] %, presented as weighted mean difference [95% confidence intervals]), but not in Westerners (0.02 [-0.04 to 0.08] %). No effect was observed in fasting blood glucose (non-Westerners, -0.25 [-0.57 to 0.06] mmol/L; Westerners, WMD: -0.06 [-0.14 to 0.02] mmol/L) and fasting blood insulin (non-Westerners, -6.53 [-17.16 to 4.10] pmol/L; Westerners, 0.47 [-4.70 to 5.65] pmol/L). Also, vitamin D increased high-density lipoprotein cholesterol in non-Westerners (0.06 [0.01 to 0.10] mmol/L), but not in Westerners (-0.01 [-0.04 to 0.03] mmol/L). Vitamin D decreased systolic (-1.71 [-3.29 to -0.14] mmHg) and diastolic blood pressure (-1.31 [-2.50, -0.11] mmHg) in non-Westerners; in contrast, no effect on blood pressure was observed in systolic (0.07 [-1.33, 1.47] mmHg) and diastolic blood pressure (0.01 [-0.65, 0.67] mmHg) in Westerners.

These ethnocultural differences have important implications. First, non-Westerner appear to have lower baseline 25-hydroxyvitamin D (e.g., Japan: 20.9±6.1 ng/mL, Norway: 24.0±8.8 ng/mL, and US: 28.0±10.2 ng/mL (1). VitD had a superior T2D-protective effect in those whose baseline levels of 25-hydroxyvitamin D <12 ng/mL) (HR: 0.58 [0.35 to 0.97]) (1), which was supported by the finding that the cardiovascular protective effect of calcium can only be observed in Asian populations with lower to moderate calcium intake, but not in American or European populations with high calcium intake (3). Second, the synthesis and metabolism of vitamin D can be affected by magnesium levels (4), as ethnic differences in magnesium intake and VitD concentrations have been reported (5). The magnesium intake and baseline magnesium level were unknown in trials performed by Pittas and colleagues (1). In conclusion, the ethnocultural background may modify the supplemental effect of vitamin D on cardiometabolic outcomes in diverse populations reflecting the baseline levels of vitamin D, calcium, and magnesium levels and their interactions.

 REFERENCES

1. Pittas AG, Kawahara T, Jorde R, Dawson-Hughes B, Vickery EM, Angellotti E, et al. Vitamin D and Risk for Type 2 Diabetes in People With Prediabetes: A Systematic Review and Meta-analysis of Individual Participant Data From 3 Randomized Clinical Trials. Ann Intern Med. 2023;176(3):355–63.
2. An P, Wan S, Luo Y, Luo J, Zhang X, Zhou S, et al. Micronutrient Supplementation to Reduce Cardiovascular Risk. J Am Coll Cardiol. 2022;80(24):2269–85.
3. Larsson SC, Orsini N, Wolk A. Dietary calcium intake and risk of stroke: a dose-response meta-analysis,,. Am J Clin Nutrition. 2013;97(5):951–7.
4. Liu S, Liu Q. Personalized magnesium intervention to improve vitamin D metabolism: applying a systems approach for precision nutrition in large randomized trials of diverse populations. Am J Clin Nutrition. 2018;108(6):1159–61.
5. Ford ES, Ajani UA, McGuire LC, Liu S. Concentrations of serum vitamin D and the metabolic syndrome among U.S. adults. Diabetes Care 2005;28:1228-30.

Disclosures:

None.

We thank Pang An and colleagues for their comments. Although one’s ethnocultural background that typically refers to a person’s cultural identify (e.g., language, customs including diet) contributes to health and disease, including in determining one’ vitamin D status, the available data do not support the notion that the effect of vitamin D on reducing diabetes risk is modified by ethnocultural differences.

When inferring the significance of the results in the three individual vitamin D and diabetes prevention trials, the authors allude to statistical significance, which is defined based on an arbitrary dichotomous threshold (e.g., 95% confidence interval not crossing 1.00). It is increasingly appreciated that studies (and results) are not positive (“significant”) or negative (“null”) but they report a range of possible effects based on the observed data; hence, statistical significance should not be the only criterion to conclude that the intervention was effective (1, 2). In their primary analyses, the three trials reported remarkably similar effect sizes (DPVD study, 0.87 [95%CI, 0.67 to 1.17]; D2d study, 0.88 [0.75 to 1.04]; Tromsø study, 0.90 [0.69 to 1.18], indicating an effect of vitamin D on reducing diabetes risk without evidence of heterogeneity among populations in different countries (Japan, US and Norway, respectively). The observed effect in each trial missed statistical significance by conventional criteria because the observed relative risk reductions were smaller than each trial was powered to detect (36% in DPVD, 25% in D2d and 30% in Tromsø). When we combined individual participant data, we improved the precision around the true effect size of vitamin D for diabetes risk reduction (relative risk reduction 0.85 [0.75 to 0.96]) and found no evidence of differences across studies (interaction p-value 0.99)(3).

The differences in the mean blood 25(OH)D levels between the three trials are likely due to differences in the study design (e.g., the DPVD study did not allow vitamin D outside of the study; the D2d allowed up to 1000 IU/day), in the use of supplements in each country (highest in the US) and/or timing of the study (e.g., the Tromsø study was completed before the rise in the popularity of vitamin D supplements). Finally, the DPVD study enrolled people with prediabetes and relatively low BMI, and vitamin D appears to work better for diabetes prevention in leaner individuals regardless of race or ethnocultural background (3, 4).

REFERENCES

1. Guallar E, Goodman SN, Localio AR, Stephens-Shields AJ, Laine C. Seeing the Positive in Negative Studies. Ann Intern Med. 2023;176(4):561-2.
2. Amrhein V, Greenland S, McShane B. Scientists rise up against statistical significance. Nature. 2019;567(7748):305-7.
3. Pittas AG, Kawahara T, Jorde R, Dawson-Hughes B, Vickery EM, Angellotti E, et al. Vitamin D and Risk for Type 2 Diabetes in People With Prediabetes : A Systematic Review and Meta-analysis of Individual Participant Data From 3 Randomized Clinical Trials. Ann Intern Med. 2023;176(3):355-63.
4. Chatterjee R, Davenport CA, Vickery EM, Johnson KC, Kashyap SR, Leblanc ES, et al. Effect of intratrial mean 25(OH)D concentration on diabetes risk, by race and weight: an ancillary analysis in the D2d randomized study cohort. Am J Clin Nutr. 2023 (in press).