The Coronary Artery Calcification in Type 1 Diabetes (CACTI) study population has been described previously (⁴). In brief, 652 men and women with type 1 diabetes and 764 nondiabetic control subjects aged 19–56 years, with no history of CAD, were enrolled in a prospective follow-up of the development and progression of CAC. Patients with type 1 diabetes had long-standing diabetes (mean duration 23 years, range 4–52 years) on enrollment. Control subjects had not been diagnosed with diabetes of any type, had fasting blood glucose in the normal range, and were generally spouses, friends, or neighbors of the patients. All subjects provided informed consent, and the study was approved by the Colorado Multiple Institutional Review Board.

Participants were examined at a baseline visit in 2000–2002. Data used in these analyses are from the second (3-year) and third (6-year) follow-up visits, both conducted approximately 2.5 ± 0.4 years after each prior visit. For these analyses, all subjects with type 1 diabetes who had complete data at the 3-year visit were included. A total of 374 non-Hispanic white subjects with type 1 diabetes were analyzed. Anthropometric measurements were obtained, including height, weight, minimal waist circumference, and hip circumference. BMI was calculated in kilograms/meters squared. Resting systolic blood pressure and fifth-phase diastolic blood pressure were measured three times while the patients were seated after a 5-min rest, and the second and third measurements were averaged. Hypertension was defined as a blood pressure ≥140/90 mmHg or participant receiving current antihypertensive treatment. Current medication, tobacco use, and family medical history were obtained by standardized questionnaire. Vitamin D and calcium intake were obtained by a validated food frequency questionnaire (Harvard 1988) and included the use of supplements.

Characteristics of study participants at the 3-year visit are summarized in Table 1. Age- and sex-adjusted mean 25[OH]D levels for type 1 diabetes subjects were 35.4 ± 0.7 ng/mL. Normal (>30 ng/mL), insufficient (20–30 ng/mL), and deficient (<20 ng/mL) 25-[OH]D levels were present in 65%, 25%, and 10% of the patients with diabetes, respectively.

Polymorphisms in the VDP and VDR genes were associated with levels of 25[OH]D after adjusting for age, sex, hours of daylight, AER, total vitamin D intake, and total calcium intake (Fig. 1). Levels of 25[OH]D were higher among individuals with the CC (37.6 ± 1.0 ng/mL) genotype at VDP 420 compared with those with the CA (34.3 ± 1.2 ng/mL, P = 0.03) and AA (32.3 ± 2.8 ng/mL, P = 0.08) genotypes. At VDP 416, levels of 25[OH]D were significantly higher among individuals with the TT (38.4 ± 1.0 ng/mL) genotype compared with those with the TG (35.1 ± 1.2 ng/mL, P = 0.04) and GG (32.7 ± 1.1 ± 1.7 ng/mL, P = 0.007) genotypes. Of the polymorphisms studied in the VDR gene, only those at VDR M1T showed significant associations with level of 25[OH]D. Levels of 25[OH]D were significantly higher among individuals with the TT (40.0 ± 1.7 ng/mL) genotype at VDR M1T compared with those with the CT (36.0 ± 1.1 ng/mL, P = 0.05) and CC (33.7 ± 1.0 ng/mL, P = 0.002) genotypes.

Lower levels of vitamin D as a continuous measure were significantly associated with presence of CAC at the 3-year visit, adjusting for age, sex, and hours of daylight (odds ratio [OR] = 0.98 [95% CI 0.96–1.00; P = 0.02]). Additional adjustment for BMI, HDL-cholesterol, LDL-cholesterol, and triglycerides attenuated these results (OR = 0.99 [0.97–1.00]; P = 0.21).

Vitamin D deficiency (25[OH]D <20 ng/mL) was associated with the presence of CAC at the 3-year visit (OR = 3.3 [95% CI 1.6–7.0; P = 0.002]), adjusting for age, gender, and hours of daylight (Table 2). Similar results were found when examining the same models with vitamin D insufficiency (<30 ng/mL) (OR = 1.8 [1.1–3.0]). Additional adjustment for BMI, HDL-cholesterol, LDL-cholesterol, and triglycerides did not significantly affect the association with vitamin D deficiency and CAC (OR = 2.4 [1.1–5.3]. The addition of factors that could affect vitamin D levels (AER, vitamin D intake, and calcium intake) attenuated the results only slightly (2.2 [0.8–5.8]). Adjustment for VDP 420, VDP 416, VDR M1T, and VDR Bsm1 did not change these findings. No polymorphisms were associated with the presence of CAC at the 3-year visit.

In the longitudinal analyses of the relationship between vitamin D deficiency and progression of CAC, we first explored if the effect of vitamin D deficiency differed between persons free of CAC and those who already had detectable CAC at the 3-year visit. The interaction term was borderline significant (P = 0.06), indicating that vitamin D-deficient subjects without preexisting CAC at the 3-year visit were more likely to develop CAC than those with sufficient vitamin D levels, whereas vitamin D-deficient subjects with CAC already present did not experience more CAC progression than those with sufficient vitamin D levels.

The multivariate model that takes into account the interactions among vitamin D deficiency, presence/absence of CAC at 3-year visit, and VDR M1T genotype is shown in Table 3. Vitamin D deficiency predicted development of CAC among subjects free of CAC at the 3-year visit. Among those, vitamin D deficiency was a stronger risk factor for CAC in the presence of the VDR M1T CC genotype (OR = 6.5 [1.1–40.2], P = 0.04) than in the presence of the CT or TT genotype (OR = 1.6 [0.3–8.6], P = 0.57). Vitamin D deficiency did not predict progression of CAC among subjects with preexisting CAC at the 3-year visit.

The results of this study are consistent with the growing body of evidence for the role of vitamin D deficiency in development of coronary atherosclerosis. They are also internally consistent with regard to cross-sectional and longitudinal analyses. The novel findings include demonstration of a potential gene–environment interaction effect between the VDR M1T CC genotype and vitamin D deficiency on CAC development. Furthermore, these results indicate that vitamin D deficiency may be of particular importance during the early events leading to development of coronary calcification.

The only other prospective study (The Multi-Ethnic Study of Atherosclerosis [MESA]) has also found that lower 25[OH]D levels predicted incident but not prevalent CAC (¹⁷). Both studies used similar methods to measure CAC. 25[OH]D was measured by radioimmunoassay in MESA and by high-performance liquid chromatography in our study; however, both studies found similar prevalence of vitamin D deficiency (10% with <20 ng/mL in our study and 10% with <15 ng/mL in MESA). Dietary vitamin D intake was available in CACTI and included in our analyses, but not in the report from MESA. In contrast with the MESA study population, the CACTI participants analyzed all had type 1 diabetes. Furthermore, the age of vitamin D-deficient CACTI participants was lower (40.1 ± 8.9 years) than that of those in the MESA study (62.1 ± 10.3 years). Despite these differences in CACTI and MESA study populations, results were remarkably similar, indicating a robust role of vitamin D deficiency in CAC development.

Our study population was limited to non-Hispanic white subjects, whereas the MESA study (¹⁷) had a more diverse study population. African Americans and Hispanics have an increased prevalence of vitamin D deficiency; however, the MESA study did not find any differences by race (¹⁷). Results of our study demonstrate that the effects of vitamin D deficiency on CAC may be modified by genetic factors, especially the VDR M1T polymorphism. In our population, the CC genotype, associated with the lowest levels of serum 25[OH]D, conferred increased risk of CAC development indicating a possible pathway. This finding also is consistent with a previously reported association between the CC genotype and acute coronary syndrome (²⁰). Future studies should expand our understanding of the genetic architecture of vitamin D metabolism and its implications to atherosclerosis.

Our study has some limitations that should be noted. Because we measured 25[OH]D levels only at the 3-year visit, we do not know whether this measure is a stable indicator of overall vitamin D status. We only studied subjects with type 1 diabetes, so we are uncertain if these findings differed between subjects with and without type 1 diabetes. Our study was also underpowered to test gene–environment interactions of modest magnitude, so it is possible that such interactions exist but were not detected in this study.

In the CACTI study, the association of vitamin D deficiency with prevalent CAC was independent of known CAD risk factors, including confounders such as BMI and mediators such as lipids. This suggests that vitamin D may not be related to CAD through common pathways with these traditional CAD risk factors but rather through a unique biologic mechanism. Vitamin D regulates the renin-angiotensin system (¹⁴) and may lower cardiovascular risk through this mechanism. Cardioprotective effects of 1,25[OH]₂D treatment have been demonstrated in animal models (²¹–²³). Our results suggest that vitamin D deficiency may be involved in the initiation of coronary calcification process. Vitamin D has an effect on antigen-presenting cells, such as dendritic cells and macrophages, and treatment with 1,25[OH]₂D can reduce foamy macrophages and suppress cholesterol uptake (²⁴).

In conclusion, this study has strengthened the evidence for the role of vitamin D deficiency in the development of coronary atherosclerosis, especially during the early events leading to development of coronary calcification. The novel findings include demonstration of this effect in patients with type 1 diabetes and evidence for a gene–environment interaction between the VDR M1T CC genotype and vitamin D deficiency on CAC development.

This study was supported by the National Institutes of Health, National Heart, Lung, and Blood Institute grants R01-HL61753 and R01-HL079611, Diabetes Endocrinology Research Center Clinical Investigation Core P30-DK57516, and American Diabetes Association Takeda postdoctoral fellowship 7-09-CVD-06 (to J.K.S.-B.).

The study was performed at the Adult General Clinical Research Center at the University of Colorado Denver Anschutz Medical Center supported by the National Institutes of Health Grant M01-RR000051 at the Barbara Davis Center for Childhood Diabetes in Denver, Colorado and at the Colorado Heart Imaging Center in Denver, Colorado. Genetic typing was performed by Drs. Suzanne Cheng, Teodorica Bugawan, and Henry Erlich at Roche Molecular Systems. No other potential conflicts of interest relevant to this article were reported.

K.A.Y. researched data, contributed to discussion, wrote the article, and reviewed and edited the article. J.E.H. researched data, contributed to discussion, and reviewed and edited the article. J.K.S.-B. researched data, contributed to discussion, and reviewed and edited the article. M.R. researched data, contributed to discussion, and reviewed and edited the article. S.K.G. researched data, contributed to discussion, and reviewed and edited the article. R.G.N. researched data, contributed to discussion, and reviewed and edited the article. P.A.G. researched data, contributed to discussion, and reviewed and edited the article. D.T. contributed to discussion.

Parts of this study were presented in abstract form at the 69th Scientific Sessions of the American Diabetes Association, New Orleans, Louisiana, 5–9 June 2009.