This study, undertaken to assess if a causal relationship links glucose homeostasis and naive T2DM to vitamin D status in obesity, documented a negative effect of vitamin D deficiency on glycated haemoglobin when the effect of age, BMI, and gender was accounted for. Our analysis did not capture a direct association between diabetes mellitus and vitamin D, possibly due to the overwhelming effects of obesity on both conditions, however patients with impaired glucose tolerance clustered more frequently with an altered vitamin D status, after controlling for potential confounding factors. The potential relationship linking vitamin D to metabolic homeostasis in obesity was substantiated by the observed association between vitamin D and insulin secretion, as well as HDL cholesterol levels.
Evidence collected until now shows that the relationship between T2DM and hypovitaminosis D is debatable, as the bulk of data is based on observational or epidemiological studies, which are useful for generating hypotheses but not for proving causality . In previous studies, lower vitamin D concentrations were independently associated with obesity, metabolic syndrome and IFG in a teenager population, with the risk of IFG being doubled in patients at the lowest compared to those at the highest quartile of vitamin D . Similarly, waist circumference, triglycerides, fasting glucose and insulin sensitivity impairment were predicted by low vitamin D levels in an adult cohort of Australian patients . Data from adults showed that hypovitaminosis D is correlated with the development of IGT , and an inverse correlation has been described between vitamin D concentrations and the risk of developing T2DM in a 22-year follow-up Finnish cohort study ; in obstructive sleep apnea syndrome-patients, vitamin D concentrations have also been inversely related to the risk of diabetes and metabolic syndrome . Likewise, a higher prevalence of hypovitaminosis D was reported in diabetic patients compared to healthy controls in a population of middle aged Caucasian men and women and in South Asian UK residents [37, 38]. Prospective studies prompted an inverse association between 25(OH)D3 levels and future glycemia and insulin resistance . Finally, vitamin D seems to play a protective role on macrovascular damage in murine models of diabetes, slowing down one of the main diabetes related complications .
To the best of our knowledge, this is the first large study investigating such relationship in a homogeneous subset of severely obese patients, subjected to screening for T2DM by different biochemical approaches. One main finding consisted in the observation that neither 25(OH)D3 concentrations were lower nor hypovitaminosis D was significantly more frequent in prediabetic and diabetic obese patients compared to those with normoglycaemia, likely due to the blunting effect of obesity on circulating vitamin D levels. At variance with previous reports [32, 34], we were unable to disclose negative effects of low 25(OH)D3 concentrations on plasma glucose levels both in fasting conditions and after the oral glucose challenge after correction for multiple variables known to affect glucose metabolism. Even so, we confirm the inverse correlation between 25(OH)D3 and glycated haemoglobin, an established marker of glucose homeostasis [30, 41, 42]. It would be therefore important to expand to obesity results of prospective cohort studies describing an inverse association between vitamin D levels and the odds of transitioning from normoglycaemia to IFG, from normoglycaemia to T2DM and from IFG to T2DM , confirmed by a recent meta-analysis . Another circumstance suggestive of this association involves the direct correlation found between 25(OH)D3 and insulin levels, both on fasting and after the OGTT. In previous studies on individuals at risk for T2DM , 25(OH)D3 concentrations were independently associated to insulin sensitivity and beta-cell function, and in subjects at risk or not for T2DM a positive correlation was found between vitamin D and early response of C-peptide and insulin levels to the oral glucose challenge . Our findings may thus support the notion that vitamin D yields regulatory effects on insulin secretion in vitro and in vivo [47–49], while being in apparent contrast with previous inconclusive studies on poorly-controlled T2DM-patients with exhausted insulin secretion . Because our patients harboured a glucose tolerance that spanned from normal to naive T2DM, and may have thus retained a superior beta-cell activity than patients with chronic T2DM, our results support the findings of Harris et al., who studied the effects of vitamin D supplementation in prediabetic obese patients and showed that the increase in 25(OH)D3 concentrations was associated with an increase of insulin secretion rate and C-peptide concentrations . In a study by Guasch et al. , high 25(OH)D3 levels were significantly associated with HDL-cholesterol levels and diabetes/hyperglycemia, but this relationship was lost after adjustment for BMI. Also, vitamin D supplementation was shown to improve the metabolic profile of diabetic Saudi individuals undergoing different therapeutic regimens, with particular effectiveness on HDL-cholesterol levels . In our analysis, the lipid profile clearly reflected vitamin D status, and high 25(OH)D3 levels were significantly associated with higher levels of HDL cholesterol, after adjustment for key confounding factors. This seems to confirm that vitamin D status is inversely related to atherogenic dyslipidemia [54, 55], and indirectly suggests that vitamin D may be independently protective against the atherogenic profile in a population at high risk for cardiovascular disease.
Finally, current results substantiate the known high prevalence of vitamin D deficiency associated with obesity; the reason for this association has not been fully defined, and is commonly attributed to vitamin D accumulation in the adipose tissue [22, 56–61]. In our experience, the cumulative prevalence of vitamin D deficiency and insufficiency was 95%. In our clinical practice, patients diagnosed with vitamin D deficiency receive cholecalciferol treatment (300.000 units p.o. for two consecutive days) and undergo regular follow-up. Also frequent was secondary hyperparathyroidism, affecting nearly 55% of subjects with vitamin D deficiency/insufficiency. These figures consolidate those found in similar studies [62, 63]. Peculiarly, we observed an inverse relationship between 25(OH)D3 and PTH levels on one hand, and calcium concentrations on the other, both confirming the secondary origin of hyperparathyroidism. In multivariate analysis, the enhancing effect of hypovitaminosis D on PTH secretion was independent of common confounding factors herein associated. It is important to note that growing attention has recently focused on PTH as being a potentially closer factor associated to metabolic abnormalities than vitamin D levels. PTH plays a role in increasing the cardiovascular risk possibly via its effects on blood pressure, insulin resistance, hyperglycaemia and low HDL-CHO levels . In a population-based cross-sectional study of US men and women, the odds ratio for metabolic syndrome increased with increasing PTH in older men only , while a survey in aging European population showed that a decreased risk for metabolic syndrome with increasing quintiles of 25(OH)D3 but not with PTH . Neither of these studies was adjusted for BMI. In another study in obese subjects, PTH was paradigmatically associated to the metabolic syndrome via other biomarkers like vitamin D and magnesium, albeit this association was only significant in women . Based on our analysis in severely obese patients, our results do not appear to confirm the predictive role of PTH on metabolic derangement when BMI, age and gender are accounted for. It would be valuable to corroborate this observation in a study encompassing a wider BMI range.
The cross-sectional design of our study represents its main limitation, since no cause-effect relationship could have been investigated. Further studies are required to better define the impact of vitamin D status on the development of glucose metabolism alterations. Furthermore we have no data on the effect of vitamin D supplementation on glycated haemoglobin. Recently, the correction of poor vitamin D status has been inversely associated with fasting insulin and HOMA-IR, in obese adolescents, but the effect on HbA1c levels has not been investigated . An interventional study should specifically address this issue. Finally, if, on one hand, the high homogeneity of our sample could be considered a strength, on the other hand it could be seen as a weakness, since the selection of severely obese patients could have masked the relationship between vitamin D status and glucose metabolism. Some of the strengths of this study should not be neglected, including the significant amount of clinical data collected in a large population of obese people, the status of previously undiagnosed glucose abnormalities, and the adjustment for common confounders such as age, BMI, and gender.