Hypovitaminosis D and obesity have concomitantly reached epidemic levels worldwide [1
]. Clinical evidence reported that hypovitaminosis D represents an increasingly widespread phenomenon in patients with obesity [3
] and it is considered a clinical syndrome due to low serum 25-hydroxyvitamin D (25OHD) levels. [1
]. It is well-established that 25OHD levels are about 20% lower in subjects with obesity than people of normal weight [5
], and the prevalence of 25OHD deficiency is greater in populations with obesity [6
Several mechanisms have been hypothesized to explain the hypovitaminosis D in obesity, such as lower dietary vitamin D intake, lesser skin exposure to sunlight (due to reduced outdoor physical activity), reduced intestinal absorption, and accumulation of vitamin D in adipose tissue [7
]. Different studies also indicate a strong negative association between hypovitaminosis D and several obesity-related diseases, including type 2 diabetes, cardiovascular disease, sleep disturbance related to obstructive sleep apnoea syndrome, and neurological diseases [9
], suggesting that the 25OHD deficiency may play a crucial role in the pathogenesis of these diseases. In the context of weight excess, on the one hand, recent studies showed a causal role of obesity in determining 25OHD deficiency [14
]. On the other hand, evidence has reported that obesity is associated with hypovitaminosis D [14
] and suggested that low vitamin D status could affect health risks in obesity [9
Interestingly, diet play a key role in obesity, and different dietary strategies, including MD, can reduce weight [16
]. In addition, excess body weight significantly increases the risk of hypovitaminosis D, a well-recognized common feature of individuals with obesity [1
]. The Mediterranean Diet (MD) is characterized by plant-based foods, including legumes, green-leafy vegetables, fruits, and nuts; high consumption of fish and low of red meat; moderate consumption of dairy products; and low consumption of red wine with meals (if already consuming alcohol). In addition, it includes the use of extra virgin olive oil as the main source of dietary fat. Indeed, it is the synergistic consumption of these Mediterranean foods which exemplifies the antioxidant and anti-inflammatory nature of this diet pattern [18
It has been well reported that the MD prevents weight gain and obesity [16
], and this has been attributed to its antioxidant and anti-inflammatory properties [16
]. In addition, Castro-Quezada I et al. has recently reported that a higher adherence to the MD was associated with a higher prevalence of subjects showing adequate intake of nutrients, including vitamin D [23
]. Other previous studies demonstrated that subjects with higher adherence to the MD also had higher dietary vitamin D intake [24
In particular, Marventano S et al. reported the difference of nutrient intake according to quartiles of adherence to the MD and showed that dietary vitamin D intake was significantly different between the lowest and the highest quartile (4.4 vs. 6.4 µg, p
< 0.001; respectively) [25
]. Of interest, in Italy, the recommendation in both males and females is 15 µg/day (18–54 age) [27
], an amount that is usually not achieved either from Western diets [28
] or from a high adherence to the MD [25
Only the HELENA study of 2330 European adolescents reported that adherence to the MD evaluated by the adapted MD quality index (KIDMED) was associated not only with better nutrient intakes, but also with higher levels of some nutritional biomarkers, including 25OHD [29
]. Otherwise, to the best of our knowledge, no studies evaluated the association between the adherence to the MD and 25OHD levels in adults.
Of interest, it is well known that non-modifiable factors, including gender, influence food preferences and consumption amounts [30
]. In particular, scientific evidence reported that females have a dietary profile characterized by a higher intake of fruits and vegetables than males, who consume more meat, eggs, and dairy products [31
]. Beyond individual food choices, recent studies have reported that females showed an overall higher adherence to the MD compared with their male counterparts [33
]. In addition, results of a large epidemiological study, The Nurses’ Health Study by Heidemann C et al., clearly indicated that greater adherence to a “prudent pattern,” characterized by a high intake of vegetables, fruit, legumes, fish, poultry, and whole grains, was related to a lower risk of cardiovascular and total mortality in healthy women [36
]. Very recently, we have reported that 25OHD levels were lower in females than in males across all body mass index (BMI) categories [37
]. As far as we know, evidence on a gender difference in the association between a healthy eating pattern, such as MD, and 25OHD levels, is still lacking. Considering the influence of the adherence to the MD on the weight excess and 25OHD levels, and the potential effects of gender on these variables, we designed the present cross sectional, single centre study to investigate the possible gender-specific association between adherence to the MD and 25OHD levels in a sample of adult population across BMI categories.
Study population consisted of 617 participants; 296 were males and 321 females, matched by age and BMI. Age, anthropometric characteristics, adherence to the MD, and 25OHD levels in the whole study population and stratified by gender, are given in Table 1
. Females have a higher PREDIMED score than males (p
= 0.001), and according to PREDIMED categories, a greater percentage of males were low adherent to the MD compared to their female counterparts (p
= 0.003). The 25OHD levels were higher in males than in females (p
= 0.01), and a higher percentage of males had sufficient 25OHD levels (>30 ng/mL) than their female counterparts (p
The PREDIMED score and 25OHD levels in the population study across BMI categories are shown in Table 2
. As reported, when stratifying the sample population according to the BMI categories, there were significant differences in both PREDIMED score and 25OHD levels; in particular, both PREDIMED score and 25OHD levels decreased significantly along with the increase of BMI (p
When analysing the response frequency of dietary components included in the PREDIMED questionnaire in detail, we found that the males consumed more extra-virgin olive oil (p
= 0.025) and red/processed meats (p
< 0.001), as compared with the females (Table 3
and Figure 2
report BMIs (Figure 1
a and Figure 2
a) and PREDIMED scores (Figure 1
b and Figure 2
b), in males and females, respectively, across 25OHD categories. The same figures report BMI (Figure 1
c and Figure 2
c) and 25OHD levels (Figure 1
d and Figure 2
d), in males and females, respectively, across PREDIMED categories. In detail, stratifying the sample population according to 25OHD categories, BMI decreased and PREDIMED score increased significantly along with the increase in 25OHD levels (p
< 0.001), in both males and females, respectively (Figure 1
a and Figure 2
a). Likewise, dividing the sample population according to PREDIMED categories, BMI decreased and 25OHD levels increased significantly, along with the increase in the adherence to the MD (p
< 0.001), in both males and females, respectively (Figure 1
b and Figure 2
reported the descriptive data of PREDIMED score and 25OHD categories according to PREDIMED categories and stratified by gender. PREDIMED scores were higher in females than in males in the low and high PREDIMED categories (p
< 0.001 and p
= 0.03, respectively).
Both PREDIMED score and 25OHD levels were negatively correlated with BMI (r = −0.62 and r = −0.67, p < 0.001; respectively), but 25OHD levels only showed a negative correlation with age (r = −0.09, p = 0.02). The PREDIMED score was positively correlated with 25OHD levels (p < 0.001). This association was maintained after adjusting for age and BMI, in both males (r = 0.21, p < 0.001) and females (r = 0.30, p < 0.001).
The results of the bivariate proportional OR model used to assess the association of 25OHD levels with PREDIMED categories are reported in Table 5
. The 25OHD levels presented the highest OR values in the category low adherence vs. high adherence to the MD, in both genders (OR 1.21 and OR 1.31, in males and females, respectively).
To compare the relative predictive power of adherence to the MD associated with the 25OHD levels, we performed a multiple linear regression analysis using a model that included PREDIMED score, sex, age, and BMI. Using this model, BMI entered at the first step (p
< 0.001), followed by PREDIMED score, sex, and age. Results are reported in Table 6
ROC analysis was then performed to determine the cut-off value of PREDIMED score predictive of 25OHD levels. In particular, PREDIMED score >5 in males (p
< 0.001, AUC 0.76, standard error 0.03, 95% CI 0.71 to 0.81; Figure 3
), and >7 in females (p
< 0.001, AUC 0.85, standard error 0.02, 95% CI 0.81 to 0.89; Figure 4
) could serve as thresholds for 25OHD levels above the median (16.2 ng/mL and 16.9 ng/mL, in males and females, respectively).
In this cross sectional, single center study, we reported a novel positive association between the adherence to the MD and the 25OHD levels, independent of BMI, in an adult population stratified according to gender across BMI categories. This stratification allowed us to confirm the presence of a gender difference either in the 25OHD levels or the adherence to the MD; although 25OHD levels were higher in males than females, 69.7% were deficient. In particular, compared to males, females exhibited higher adherence to the MD, despite lower 25OHD levels, with a lower percentage of 25OHD sufficiency as well. As expected, PREDIMED scores and 25OHD levels were the lowest in study participants with the highest grades of obesity (grade III). Nevertheless, the joint evaluation of PREDIMED score and 25OHD levels across different BMI categories in both genders had not previously been reported. As the highest values of PREDIMED score were reported in the subjects with the highest 25OHD levels and the lowest BMIs, these results suggested the key role of the adherence to the MD in both 25OHD levels and BMI. Based on the ROC curve analysis, the most sensitive and specific cut-off values for the PREDIMED score to predict the highest 25OHD levels were >5 and >7, in males and females, respectively. To date, to the best of our knowledge, this is the first study investigating gender differences in the relationship between adherence to the MD and 25OHD levels in a sample of an adult population across BMI categories.
Robust scientific evidence showed that 25OHD levels are lower in subjects with obesity than in normal-weight subjects [60
]. The underlying mechanism of low 25OHD levels in subjects with obesity may be due to a number of factors, including low dietary vitamin D intake, little sun exposure, and low outdoor activity. In addition, accumulation of vitamin D in fat mass has been proposed as a potential mechanism [63
]. Another hypothesis considers the result of volumetric dilution of vitamin D in the large adipose stores [64
], as reported in women with obesity compared to normal-weight controls [65
]. A recent meta-analysis of randomized controlled trials showed an overall inverse relationship between 25OHD levels and BMI in studies of non-diabetic subjects (r = −0.152, 95% = −0.187 to −0.116, p
< 0.001), concluding that the vitamin D deficiency was associated with an increase in BMI [4
Vitamin D receptors and vitamin D metabolizing enzymes are expressed in both adipocytes and inflammatory cells infiltrating adipose tissue depots. Beyond an anti-adipogenic effect, their presence also suggests an important role for vitamin D in regulating inflammatory pathways [66
]. In this context, several piece of evidence showed that low vitamin D status was associated via low grade chronic inflammation, with consequent increased risk of metabolic and cardiovascular diseases [69
]. Beyond its anti-inflammatory effects, the link between low 25OHD levels and risks of several diseases could be also partially explained by the antioxidant effects of vitamin D [71
]. Of interest, in obese children, Codoñer-Franch P. et al. evidenced that a low vitamin D status was associated with high levels of oxidative stress biomarkers, including malondialdehyde, myeloperoxidase, 3-nitrotyrosine, interleukin (IL)-6, and soluble vascular cell adhesion molecule-(sVCAM)-1 [74
]. According to these results, other studies confirmed the links among obesity, low vitamin D status, and increased oxidative stress [2
The anti-inflammatory and antioxidant effects of the MD are equally well-known [76
]. The MD pattern is the traditional diet of some of the countries of the Mediterranean basin that gathers the following characteristics: high consumption of whole grain, legumes, fruits, vegetables, nuts, and seeds; moderate intake of fish and shellfish, white meat, and eggs; and dairy products, red meat, processed meats, and foods rich in sugars in relatively small amounts. Moderate intake of wine, especially red wine with meals and extra virgin olive oil as the main source of fat, are also recommended [78
]. From a nutrient point of view, MD is characterized by a relatively high fat consumption (40%–50% of total daily calories), by which between 15% and 25% of calories are represented by monounsaturated fatty acids, and ≤8% of calories are comprised of saturated fatty acids [81
]. The favorable effects on health status of the MD are to be attributed also to high consumption of dietary fiber, food with low glycemic index, and glycemic load [83
Among favorable effects of vitamin D, several studies reported that subjects with higher adherence to the MD also have higher dietary vitamin D intake [24
], but only one study in adolescents reported that the adherence to the MD was positively associated with 25OHD levels [29
]. However, in their study, the authors underlined a need to increase the knowledge on the nutritional adequacy of vitamins and the MD [23
]. In addition, it is well known that gender is a well-known non-modifiable factor largely influencing food preferences and that women are more inclined to a Mediterranean dietary profile [31
]. In our study, we confirmed that adult females had significantly lower 25OHD levels than adult males. However, it is evident from the gender difference in the association between the adherence to the MD and 25OHD levels in males, that despite a lower adherence to the MD compared to females, males have higher 25OHD levels. Although 25OHD levels were higher in males than females, 69.7% were deficient. As these results were independent of BMI, we hypothesized that the well-known sex difference in body fat and body fat distribution could likely account for this gender difference in the association between the two variables. Indeed, very recently we have demonstrated that females had significantly lower 25OHD concentrations than males among the different classes of BMI and that this sex difference was mainly explained by the higher fat mass percentage in women than men [37
]. However, the different intakes of the foods of the MD between females and males could provide a further explanation. In our study the extra virgin olive oil consumption was different in males than in females, and it is known that dietary fats may change vitamin D absorption [84
]. In particular, it has been reported that an increased consumption of monounsaturated fatty acids-rich oils, such as olive oil, may improve the bioavailability of vitamin D [85
Of interest, although there is some disagreement [19
], both a meta-analysis by Esposito et al. including 3436 participants [20
], and a more recent systematic review of five randomized clinical trials indicated that a MD, in combination with physical activity, MD is effective at inducing weight loss, especially when it was energy restricted, and lasted longer than six months [16
]. In particular, the high content of polyphenols in the MD could exert favorable its effects on weight loss through different mechanisms, including activation of β-oxidation, prebiotic effects for gut microbiota, induction of satiety, and modulation of both brown and white adipose tissue [19
Taking into account the independent roles of adherence to the MD and vitamin D in obesity, the translational potential of this study suggests that the high adherence to the MD, likely due its anti-inflammatory and antioxidant properties, might contribute to offerring a unique advantage in the management of subjects with low vitamin D status and obesity. In addition, gender-specific cut-off values for the PREDIMED score might help to predict the risk of the lowest levels of 25OHD levels in adults.
We are aware that there are some limitations in the current study.
First, the cross-sectional design of this study did not allow for any statements on a causal relationship between adherence to the MD and 25OHD levels. Moreover, the suggested cut-offs values of PREDIMED score should be viewed with caution until results of studies in larger patient populations have become available to perform an appropriate cross-validation.
Second, the possible underlying inflammatory and antioxidant status linking PREDIMED score, 25OHD levels, and obesity, should be better investigated by measuring both serum inflammatory and antioxidant biomarkers, such as IL-6, c-reactive protein levels, superoxide dismutase, glutathione peroxidase, catalase, and malondialdehyde; thus, our hypothesis of a synergistic effect of adherence to the MD and 25OHD levels on obesity remains largely speculative.
Third, we did not include the measurements of the body composition in this study.
Fourth, the adherence to the MD was assessed with the 14-items questionnaire of the PREDIMED study [52
]. This questionnaire is less expensive, less time-demanding, and requires less collaboration from participants than the usual full-length food frequency questionnaire or other more comprehensive methods [52
], with which it was previously validated [86
]. In addition, the PREDIMED questionnaire allows one to provide feedback to the participant immediately after the interview is completed. However, the design of this questionnaire of 14-items does not allow for an evaluation of total energy and dietary intake, and particularly, dietary vitamin D intake. This evaluation must be done with the seven-day food records, as it is the gold standard and allows for accurate measurements of the real dietary and macronutrient intakes compared to those that were obtained by retrospective food frequency questionnaires [87
]. Therefore, we could only speculate that the observed lower adherence to the MD in males compared to females, despite their higher 25OHD levels, could be the epiphenomenon of the well-known sex difference in body fat and body fat distribution.
Fifth, sunlight exposure time have not been sufficiently evaluated. Sun is the major causal factor for melanoma, and melanoma incidence rates in Italian men (40.4 per
100,000) are much higher than in Italian women (11.0 per
]. This would therefore indicate that sun exposure in Italian males is higher than in Italian females. Sun exposure is also the main predictor of circulating 25OHD levels [59
]; only a small proportion of circulating 25OHD levels are attributable to diet (excluding supplementation) [7
]. In this study, we did not evaluate the sun-exposure habits. Thus, the role of sun exposure in influencing the difference in 25OHD levels between males and females can not be excluded.
Nevertheless, a major strength of this study includes the very large sample of subjects stratified by gender—making possible the comparisons between the study variables across different categories of BMI, PREDIMED, and 25OHD independently of gender. In addition, our study was based on a single clinical center, with a possibility of selection bias in the results.
However, the single-center study allowed us to increase the homogeneity of the sample as we included subjects evaluated in the same season and living in the same geographical area around Naples metropolitan area, Italy (latitude 40° 49′ N; elevation 17 m), so with the same effect of latitude on vitamin D levels, and likely, with similar nutrient availability. Finally, the PREDIMED questionnaire was face-to-face administered and not self-reported in order to minimize any bias related to the filling of the questionnaire, and to avoid the inter-operator variability, only one expert Nutritionist evaluated the anthropometric measures and administered the PREDIMED questionnaire.