Effect of Vitamins D2 and D3 Supplement Use on Serum 25OHD Concentration in Elderly Women in Summer and Winter
- First Online:
- Cite this article as:
- Rapuri, P., Gallagher, J. & Haynatzki, G. Calcif Tissue Int (2004) 74: 150. doi:10.1007/s00223-003-0083-8
- 190 Views
Vitamin D2 and D3 are generally considered equipotent in humans. A few studies have reported that serum 25OHD levels are higher in vitamin D3- compared with vitamin D2-supplemented subjects. As both vitamin D2 and D3 supplements are commonly used by elderly in United States, in the present study we determined the effect of self-reported vitamin D2 and vitamin D3 supplement use on serum total 25OHD levels according to season in elderly women aged 65–77 years. Serum total 25OHD levels were determined in winter and summer in unsupplemented women (N = 307) and in women who reported taking vitamin D2 (N = 56) and vitamin D3 (N = 55) supplements by competitive protein binding assay. In vitamin D2-supplemented women, the contribution of vitamin D2 and D3 to the mean serum total 25OHD level was assessed by HPLC. In summer, there were no significant differences in the mean total serum 25OHD levels (ng/ml) among the vitamin D2 (32 ± 2.1), vitamin D3 (36.7 ± 1.95), and unsupplemented (32.2 ± 0.95) groups. In winter, the mean serum total 25OHD levels were higher in women on vitamin D2 (33.6 ± 2.34, P < 0.05) and vitamin D3 (29.7 ± 1.76, NS) supplements compared with unsupplemented women (27.3 ± 0.72). In vitamin D2-supplemented women, about 25% of the mean serum total 25OHD was 25OHD2, in both summer and winter. Twelve percent of unsupplemented women and 3.6% of vitamin D-supplemented women had a mean serum total 25OHD level below 15 ng/ml in winter. In elderly subjects, both vitamin D2 and Vitamin D3 supplements may contribute equally to circulating 25OHD levels, with the role of vitamin D supplement use being more predominant during winter.
Keywords25-Hydroxyvitamin D25OHDHPLCVitamin D supplementCompetitive protein binding assay25OHD2
Vitamin D deficiency is a risk factor for bone loss and fracture [1, 2, 3]. Elderly, especially homebound and institutionalized, are more prone to vitamin D deficiency  because of both a decrease in cutaneous production [5, 6, 7] and in some reduced dietary intake. Treatment of vitamin D deficiency can result in a decrease in the incidence of hip fracture [4, 8, 9], and a daily supplement with 800 IU (20 µg) of vitamin D may reduce the risk of osteoporotic fracture in elderly with low blood levels of vitamin D . A greater vitamin D intake from diet and supplements has also been associated with less bone loss in older women . In the United States, about 50% of women above age 50 years use vitamin and mineral supplements . Recently, Margiloff et al.  observed that self-reported vitamin D supplement use is a positive determinant of wintertime serum 25OHD levels.
Vitamin D exists in two forms, cholecalciferol (vitamin D3) and ergocalciferol (vitamin D2) Cholecalciferol is synthesized endogenously in the skin from its prohormone, 7-dehydrocholesterol, mediated by UVB radiation. Ergocalciferol, on the other hand, is plant-derived and is formed exogenously by irradiation of ergosterol. Vitamin D supplements use either cholecalciferol or ergocalciferol as their source of vitamin D. While some of the commonly used vitamin supplements in United State, like Centrum, Walgreens, Osco multivitamin, use vitamin D2 as their source of vitamin D, some, like UNICAP vitamins, Citracal, Theragram M, use vitamin D3.
Vitamins D2 and D3 have generally been considered equally bioactive in mammals. However, differential response to vitamins D2 and D3 in different animal species has been shown by some researchers [14, 15, 16]. In humans, also, it was originally thought that both vitamin D2 and vitamin D3 follow the same metabolic pathway and are equally bioactive in normal subjects [17, 18]. Later studies suggested differential response to vitamins D2 and D3 in humans as well. Clinical trials of D2 and D3 in patients on anticonvulsants indicate that treatment with vitamin D3 produced higher serum concentrations of the major vitamin D metabolites than treatment with vitamin D2 given at the same dosage [19, 20]. Trang et al.  recently reported that supplementation with vitamin D3 increased serum 25OHD levels 1.7 times higher than that seen with vitamin D2 supplementation. Earlier studies by Chapuy et al. [4, 22] also reported a two-fold difference between equal doses of vitamin D2 and vitamin D3 in increasing the serum 25OHD levels. Furthermore, there are few studies that examined the relationship between vitamin D supplement use and season [23, 24, 25, 26, 27] and none have examined the effect of the type of vitamin D supplement and season on serum 25OHD levels.
As there is some evidence of differential response to vitamins D2 and D3 in humans, and some of the vitamin D supplements commonly used by the elderly have vitamin D2 as their source of vitamin D, in the present study we determined serum 25OHD, parathyroid hormone (PTH), and bone remodeling markers in self-reported vitamin D2, and vitamin D3 supplement users and also in unsupplemented women in winter and summer. We determined circulating 25OHD levels in all subjects by competitive binding assay, which does not distinguish between 25OHD2 and 25OHD3. In women taking vitamin D2 supplements we determined the quantitative contribution of 25OHD2 and 25OHD3 to the serum total 25OHD levels by high-pressure liquid chromatography (HPLC).
Materials and Methods
The baseline information of 489 women (age range = 65–77 years) enrolled in an osteoporosis intervention trial called STOP IT (Sites Testing Osteoporosis Prevention/ Intervention Treatment) was used for this analysis. The study was intended to test the efficacy of three therapies in reversing bone loss in proximal femur and spine compared with placebo. The details of the enrollment of the subjects into the STOP IT study have been described earlier . Briefly, all the women between ages 65 and 77 years in the geographical area of Omaha and the surrounding districts were contacted through the use of newspaper advertisements, telephone calls, and direct mailing. Approximately 8005 women were contacted and about 1905 agreed to come in for a preliminary screening. Women were excluded if they had severe chronic illness, had primary hyperparathyroidism or active renal stone disease, or were on certain medications that affect bone and mineral metabolism. To complete eligibility, femoral neck BMD had to be within the normal range (± 2 SD) for their age. Four hundred eighty-nine women satisfied eligibility and were enrolled into the study. Only the data of the women who entered the STOP IT trial (N = 489) have been included in the present study. Of the 489 women, 472 were Caucasian, 11 were Black, 4 were Hispanic, 1 was Asian, and 1 was of mixed race.
The dietary intake data were collected using 7-day food diaries. The average daily calcium and vitamin D intake were calculated by a dietician, using the Food Processor 11+ (version 5.1) nutrition and diet analysis system (Esha Research, Salem, OR). The information about the multivitamin supplement use was obtained by a questionnaire administered to the subjects by the study personnel. The form of vitamin D (either D2 or D3) used in the supplements was obtained from the package inserts and also cross-checked with the Drug Information Center, Immanuel Hospital, Omaha, NE. Of the 489 women, 307 were not taking any vitamin D supplements and 182 were taking a vitamin D supplement. Of the 182 women, 56 were taking a vitamin D2 supplement and 55 were taking a vitamin D2 supplement. Of the rest of 71 women, some were taking both vitamin D2 and vitamin D3 supplements while others were taking a supplement whose source of vitamin D could not be determined. These women were not included in the analyses. In women taking vitamin D2 supplements, about 73% were taking Centrum multivitamin, 18% were taking Walgreen’s multivitamin, and the remaining 9% were taking Osco, Universal life, or Century Vite brand of multivitamins. In women taking vitamin D3 supplements, 31% constituted Theragram multivitamin users, 22% were taking Oyster shell calcium with vitamin D, 13% were taking Cod liver oil, 7% were using Unicap multivitamin, and the remaining 27% were using other multivitamins with vitamin D3. Women taking a vitamin D2 supplement were taking about 401 IU/day, while women taking a vitamin D3 supplement were getting varying amounts of vitamin D3 (200 IU, 250 IU, or 400 IU) averaging to about 465 IU/day. Serum 25OHD measurements were done by competitive protein binding assay (CPBA) on all 418 subjects included in the analyses. In all the women taking a vitamin D2 supplement and in a randomly selected sample (N = 60) from unsupplemented group, the contribution of 25OHD2 and 25OHD3 to the serum total 25OHD levels was determined by HPLC. The protocol was approved by Creighton University Institutional Review Board.
Fasting blood samples and 24-hour urine were obtained from the subjects. Blood samples were allowed to clot and centrifuged at 4°C for 15 minutes at 2056g to separate serum. The serum was stored frozen at −70°C until analysis.
Serum 25OHD Measurement
Serum total 25OHD measurement was performed by both CPBA and HPLC after precipitating plasma proteins with acetonitrile.
Competitive protein binding assay
Serum total 25OHD levels were measured by CPBA  after extraction and purification of serum on Sep-Pak C-18 and silica cartridges (Waters Associates, Milford, MA) . Briefly, after precipitating plasma proteins with acetonitrile, the supernate was backwashed with potassium phosphate (0.4 M, pH 10.5) to enhance the solubilization of lipids. The samples were then extracted on a reverse phase Sep Pak C-18 columns. The acetonitrile fraction containing the vitamin D metabolites was taken through a normal phase extraction with a silica Sep-Pak cartridge, where the 25OHD was eluted with 96:4 hexane:isopropanol. 25OHD was quantitated by CPBA using normal rat serum as the source of binding protein. The minimum detection limit for the assay was about 12.5 nmol/l (5 ng/ml) and the interassay variation was 5%.
Serum total 25OHD levels were estimated by HPLC after prepurification of the acetonitrile extract of serum by solid-phase extraction using “Bond-Elut LRC” C18/OH cartridges and Sep-Pak silica cartridges (Waters Associates) . HPLC of the purified extract was performed using a Shimadzu LC-10 system with Shimadzu LC-10AT pump, GT-104 degasser, Sil 10A Injector (autosampler) with sample cooler, CTO-10A column oven and SCL-10A system controller. The separation of 25OHD2 and 25OHD3 metabolites was achieved on a 0.45 ID × 25 cm (5 µm) Zorbax SIL column using hexane/isopropanol (97/3) at a flow rate of 2 ml/min. 25OHD2 and 25OHD3 were detected using a Shimadzu SPD-10A UV–Vis detector and the data were analyzed using the CLASS-VP chromatography data system. The serum total 25OHD levels were computed by measuring both 25OHD2 and 25OHD3. The minimum detection limit for the assay was about 2.5 ng/ml and the interassay variation was less than 1%.
Serum PTH, Serum Osteocalcin, and Urine N-Telopeptide Measurement
Serum intact PTH (iPTH) was measured with Allegro immunoradiometric assay (Nichols Institute, San Juan Capistrano, CA). The interassay variation was 5% and the limit of detection for the assay was 1 ng/l (1 pg/ml). Serum concentrations of osteocalcin were determined by radioimmunoassay (Incstar Corp.). The limit of detection was 0.78 µg/l and the interassay variation was 5%. Urine collagen crosslinks were measured by enzyme-linked immunosorbent assay (Osteomark International, Seattle, WA) as N-telopeptides, a marker for bone type I collagen. The lower limit of detection was 20 nmol bone collagen equivalents (BCE), and the interassay variation was 6%. The data are expressed as nmol BCE/mmol of creatinine.
All analyses were done using SPSS for Windows (version 11.0, SPSS Inc., Chicago, IL). The characteristics and serum 25OHD levels among the unsupplemented, vitamin D2-supplemented, and vitamin D3-supplemented women were compared using a one-way analysis of variance (ANOVA). Pearson’s correlation coefficient and simple linear regression methods were used to assess the relationship between the two methods of 25OHD estimation, dietary vitamin D intake and serum 25OHD, and serum 25OHD2 and serum 25OHD3.
The characteristics of the subjects of unsupplemented, vitamin D2-supplemented and vitamin D3-supplemented groups are given in Table 1. There were no significant differences among the groups with respect to age, weight, and dietary vitamin D intake. The calcium intake was higher in women taking a vitamin D2 supplement compared with women not taking any vitamin D supplement. The vitamin D supplement use and the duration of the vitamin D supplement use were not different between the two vitamin D groups. Furthermore, the characteristics of the random sample of subjects of the unsupplemented group (N = 60) were not different from those of the entire group (N = 307) (Table 1, parentheses of unsupplemented group). Of the 489 elderly subjects enrolled in the original study, 182 were taking a vitamin D supplement, of which we could identify 56 as receiving vitamin D2 and 55 as receiving vitamin D3 as their vitamin D source. The mean serum total 25OHD measured by CPBA in vitamin D2- and D3-supplemented groups was 33.0 ± 1.56 ng/ml and 33.6 ± 1.44 ng/ml, respectively, values which were significantly (P < 0.05) higher than in the unsupplemented group (29.3 ± 0.53 ng/ml) (Table 2). The mean serum total 25OHD determined by HPLC in women receiving vitamin D2 supplement (31.6 ± 1.35 ng/ml) was also higher when compared with the mean serum total 25OHD determined in a random sample of women not receiving any vitamin D supplement (27.7 ± 1.21 ng/ml). The serum total 25OHD measurements done by CPBA were not different from those measured by HPLC. The overall correlation between CPBA and HPLC for the serum total 25OHD was about 0.91 (P < 0.001). There were no significant differences in serum PTH, osteocalcin, and urine N-telopeptide levels among the unsupplemented, vitamin D2- and vitamin D3-supplemented women (Table 2).
Characteristics of the subjects in unsupplemented, vitamin D2 supplemented and vitamin D3 supplemented groups
71 ± 0.20 (71.8 ± 0.45)
72 ± 0.4
72 ± 0.5
160 ± 0.34 (159.5 ± 0.82)
159 ± 0.89
159 ± 0.91
70 ± 0.72 (67.7 ± 1.7)
67 ± 1.67
68 ± 1.48
Dietary calcium intake (mg/d)
696 ± 16 (735 ± 35.5)
806 ± 35a
755 ± 45
Dietary vitamin D intake (ID/d)
142 ± 4.9 (148 ± 11.6)
130 ± 10.0
128 ± 12.2
Supplemental vitamin D intake (IU/d)
401 ± 19.7
465 ± 43.7
Vitamin D supplement use duration (d)
2029 ± 388
2046 ± 391.1
Serum 25OHD concentration determined by CPBA and HPLC in vitamin D2, D3 supplemented and unsupplemented women
Total serum 25OHD (CPBA) (ng/ml)
29.3 ± 0.53
33.0 ± 1.56a
33.6 ± 1.44a
Serum 25OHD2(HPLC) (ng/ml)b
0.92 ± 0.41
8.4 ± 0.80a
Serum 25OHD3 (HPLC) (ng/ml)b
26.8 ± 1.22
23.2 ± 1.61
Serum PTH (pg/ml)
37.7 ± 0.80
34.5 ± 2.2
36.2 ± 2.0
Serum osteocalcin (ng/ml)
3.8 ± 0.07
3.7 ± 0.19
3.5 ± 0.16
Urine NTx:creatinine ratio (nmol BCE:mmol Cr)
51.6 ± 1.57
52.4 ± 3.35
45.5 ± 3.35
Total serum 25OHD (CPBA) (ng/ml)
32.2 ± 0.95
32.0 ± 2.1
36.7 ± 1.95
Serum 25OHD2 (HPLC) (ng/ml)c
0.48 ± 0.048
8.29 ± 1.10
Serum 250HD3 (HPLC) (ng/ml)c
30.0 ± 2.32
23.75 ± 2.17
Serum PTH (pg/ml)
38.6 ± 1.62
35.1 ± 4.04
34.3 ± 2.5
Serum osteocalcin (ng/ml)
3.8 ± 0.12
3.8 ± 0.35
3.26 ± 0.28
Urine NTx:creatinine ratio (nmol BCE:mmol Cr)
47.3 ± 2.48
51.9 ± 5.8
36.8 ± 2.80
Total serum 25OHD (CPBA) (ng/ml)
27.3 ± 0.72
33.6 ± 2.34a
29.7 ± 1.76
Serum 25OHD2 (HPLC) (ng/ml)d
1.24 ± 0.59
8.48 ± 1.19a
Serum 25OHD3 (HPLC) (ng/ml)d
25.2 ± 1.60
23.4 ± 2.4
Serum PTH (pg/ml)
37.8 ± 1.04
34.5 ± 2.68
38.0 ± 3.55
Serum osteocalcin (ng/ml)
3.84 ± 0.10
3.78 ± 0.24
3.65 ± 0.28
Urine NTx:creatinine ratio (nmol BCE:mmol Cr)
55.5 ± 2.31
55.0 ± 4.31
52.9 ± 6.89
When the data were analyzed according to season, in the summer months (May–September) the mean serum total 25OHD levels were not different among the unsupplemented, vitamin D2- and vitamin D3-supplemented groups (Table 2). In the winter months (December–March), mean serum total 25OHD in vitamin D2- and vitamin D3-supplemented groups were higher by 23% (P < 0.05) and 10%, respectively, compared with the unsupplemented group (Table 2). In both summer and winter, serum PTH, osteocalcin, and urine N-telopeptides were not different among the three groups (Table 2).
About 78% of women taking a vitamin D2 supplement had appreciable amounts of serum 25OHD2, while subjects not taking any vitamin D supplement had negligible amounts. In women taking vitamin D2 supplements, approximately 25% of the serum total 25OHD was contributed by 25OHD2, which remained the same during both summer and winter (Table 2); the rest of the circulating serum 25OHD could be accounted for by the vitamin D stores from sunlight and dietary vitamin D3. Figure 1 gives the correlation plot between dietary vitamin D intake and serum 25OHD levels in unsupplemented women. It is apparent from the y intercept that summer to winter differences in unsupplemented women (P < 0.001) is mainly due to sunlight exposure while the contribution from diet was not different between winter and summer. However, in women taking either a vitamin D2 or a D3 supplement, this seasonal effect is greatly reduced. In subjects taking vitamin D2 supplement, an inverse relationship was observed between serum 25OHD2 and 25OHD3 levels (Fig. 2a). In women with a low concentration of 25OHD3, the contribution of 25OHD2 to the serum total 25OHD level is higher. In vitamin D2-supplemented women, the contribution of 25OHD2 to the total serum 25OHD pool is more significant in winter than in summer (Fig. 2b, c). About 20% of vitamin D2-supplemented women would have been classified as vitamin D deficient if their serum 25OHD2 levels were not accounted for the serum total 25OHD content.
Vitamin D deficiency (<15 ng/ml) was marginally low (4.5%) in the vitamin D-supplemented groups vitamin D2 and D3 groups combined) compared with elderly subjects not taking any vitamin supplements (7%). When the data were analyzed according to season, about 12% of women not taking any vitamin D supplements recruited during winter had serum 25OHD levels below 15 ng/ml, while only 3.6% of women taking a vitamin D supplement had mean serum 25OHD levels below 15 ng/ml. In summer-recruited women, the prevalence of vitamin D deficiency (15 ng/ml) was about 1% in unsupplemented women and 5% in vitamin D-supplemented women.
In the present study, we report that in elderly women both vitamin D2 and vitamin D3 supplements appear to contribute equally to the circulating 25OHD level. We also report that the effect of vitamin D supplementation is reflected mostly in winter, with serum 25OHD levels in vitamin D2- and vitamin D3-supplemented women being higher than in unsupplemented women in winter months, significantly so only for vitamin D2 group.
In this study, we measured serum total 25OHD levels by CPBA, and the contribution of 25OHD2 and 25OHD3 to the serum total 25OHD levels was estimated by HPLC. We observed a good correlation (r = 0.91, P < 0.001) between CPBA and HPLC methods for serum total serum 25OHD measurement. Similar correlation between the two methods has been reported earlier [32, 33]. The mean concentrations of serum 25OHD observed in the present study agree with previously reported values determined by HPLC or CPBA  in normal subjects .
To our knowledge, this is the first study to report the effect of vitamin D2 and vitamin D3 supplement use on serum total 25OHD concentration according to season in elderly women. In the present study, we observed that in winter the mean serum 25OHD levels in vitamin D2 and vitamin D3-supplemented groups were higher (significant only for vitamin D2 group) than in the unsupplemented group. However, in summer there were no significant differences among the three groups. The lack of additive effect of vitamin D supplement use and sunlight exposure on serum 25OHD levels in summer in vitamin D-supplemented groups can be explained by the product inhibition of liver 25-hydroxylase enzyme. A decrease in 25-hydroxylase activity with vitamin D supplementation has been demonstrated in rats both in vivo and in vitro . In humans, Trang et al.  reported that the increase in serum 25OHD levels is inversely related to basal 25OHD levels. They noted a diminished ability of vitamins D2 and D3 to increase the serum 25OHD levels above a basal 25OHD level of 25 ng/ml.
There are few studies reporting the effect of vitamin D supplement use and season on serum 25OHD levels [23, 24, 25, 26, 27] and none of them investigated the role of vitamin D2 and D3 supplements separately. In a group of 333 healthy postmenopausal women, Krall et al.  reported that in low-vitamin-D intake group (200 IU), the mean serum 25OHD levels were significantly higher (47%) in August–October compared to March–May. But when the vitamin D intake was 200IU, the mean serum 25OHD did not change with season. In a double-blind placebo-controlled one-year trial in healthy postmenopausal women, DawsonHughes et al.  noted that in the placebo group the serum 25OHD level decreased 21% from summer to winter, while in the vitamin D-supplemented group (400 IU), only a 5% decrease was observed. In elderly women with mean age of 72 years, Omdahl et al.  reported about a 15% higher serum 25OHD in vitamin D-supplemented women compared with unsupplemented women in winter. However, in summer, vitamin D-supplemented women had 7% higher serum 25OHD levels compared with unsupplemented women. Thus, from our data and the studies reported in the literature, it is apparent that seasonal fluctuations in serum 25OHD levels are less common in vitamin D-supplemented subjects.
In the present study, a significant inverse correlation seen between serum 25OHD2 and 25OHD3 levels during winter but not in summer further substantiates the importance of the role of vitamin D supplement use during winter, wherein the endogenous production of vitamin D3 would be low. The observation of an inverse relationship between serum 25OHD2 and serum 25OHD3 levels in subjects receiving vitamin D2 has been reported previously, though not according to season. Clemens et al.  reported that in institutionalized elderly, subjects with 25OHD2 as the predominant circulating form had significantly lower (P < 0.01) mean serum 25OHD3 concentration. Hartwell et al.  measured serum 25OHD levels in vitamin D2- and vitamin D3-supplemented (4000 IU) normal subjects. In vitamin D2-supplemented subjects, they observed an inverse relationship between serum 25OHD2 and 25OHD3 levels and the serum total 25OHD remained unchanged with time. However, in vitamin D3-supplemented subjects, the mean serum 25OHD levels continued to increase with increasing time.
In our study, we did not find significant differences in serum PTH and bone remodeling markers between the unsupplemented and the vitamin D2- or D3-supplemented women in both summer and winter. In contrast, Dawson Hughes et al.  reported that from August to November the serum PTH levels were similar between placebo and vitamin D-supplemented groups (29 vs. 28 ng/l), while from February to May the serum PTH values were significantly (P < 0.05) higher in placebo group (32 ng/l) compared with the vitamin D group (29 ng/l). In agreement with our results, Himmelstein et al. , in a double-blind study of elderly, did not observe a significant effect of vitamin D supplement use on serum PTH and osteocalcin levels.
Vitamin D deficiency is a significant risk factor for bone loss and subsequent fracture. In the present study carried out in Omaha (41°N latitude), about 12% of unsupplemented women and 3.6% of vitamin D-supplemented women had a mean serum 25OHD level below 15 ng/ml in winter, with no significant differences between the two groups in summer. In Boston (42.2°N), it has been reported that in elderly nursing home residents about 40% had serum 25OHD levels below 10 ng/ml and about 80% had below 15 ng/ml in the winter time  which is much higher than observed in the present study. However, another recent study from Boston reported that about 13.6% of the population (age range = 1886) attending a Boston outpatient clinic had serum 25OHD levels less than 16 ng/ml in winter time , which is close to what we observed in the unsupplemented group of the present study. They also reported that vitamin D supplementation is a positive determinant of serum 25OHD concentration with about 65% of those taking a vitamin D supplement having serum 25OHD levels as high as 32 ng/ml. In another study, Webb et al.  reported that in the elderly of Massachusetts, about 24% of unsupplemented women had serum 25OHD levels of 10 ng/ml in summer and about 29% had this value in winter. In vitamin D-supplemented women, these levels occurred in only 2% and 3% in summer and winter, respectively.
Our study has certain limitations. The study is cross sectional in design and factors like time spent outdoors and travel to lower or higher latitudes during different seasons were not taken into consideration in interpreting the results. Dawson–Hughes et al. reported that travel to warmer places in winter influences the serum 25OHD levels . Though we did not find any significant difference in the dietary or vitamin D supplement intake between vitamin D2 and vitamin D3 groups, the self-reported vitamin D intake could be confounded by the compliance. Also, the self-reported duration of vitamin D supplement use varied among the 3 groups only in summer and adjusting for this did not alter the results. Prospective randomized controlled trials are further needed to confirm our findings.
In conclusion, in our study population we observed that vitamin D supplement use, both vitamins D2 and D3, may contribute equally to circulating 25OHD levels in the elderly. It is evident from our data that season will have a greater effect on serum 25OHD levels in women not taking vitamin D supplements. In agreements with this, we have earlier reported a significant decrease in serum 25OHD levels, an increase in bone marker level, and a decrease in bone mineral density during winter months in elderly women not taking any vitamin D supplements . Vitamin D supplement use raises serum 25OHD levels during winter and greatly reduces this seasonal fluctuation. Thus, vitamin D supplement use would be beneficial in maintaining optimum serum 25OHD levels in the elderly, especially during winter, which otherwise would lead to an increased risk of vitamin D deficiency leading to bone loss and subsequent fracture.
This work was supported by National Institute of Health Research grants UO1-AG10373 and RO1-AGI0358. We thank Karen A. Rafferty for her help in food diary data collection and analysis.