Vitamin D supplements were taken by 61.4% of our patients. VitD levels were deficient (<20 ng/dl) in 107 (38.5%) patients, insufficient (20–30 ng/dl) in 96 (34.5%) patients and adequate (>30 ng/dl) in 75 (27.0%) patients. Patients with deficient vitD levels were significantly younger than patients with adequate (36.96±7.64 yrs vs  41.26±8.19 yrs; P<0.0001) as well as insufficient vitD levels (36.96±7.64 yrs vs  39.63±9.39 yrs; P=0.042) while no difference was detected in male to female ratio.

MIO (global heart T2*<20 ms) was significantly more frequent in patients with deficient VitD than in patients with adequate and insufficient VitD levels (Figure 1A). Patients with deficient vitD levels had a significant higher risk of MIO than patients with adequate vitD levels (odds ratio-OR=20.62, 95%CI=2.67-153.72; P=0.004) and patients with insufficient VitD levels (OR=8.49, 95%CI=2.46-29.29; P=0.001). At receiver operating characteristic curve analysis, a vitD level≤ 17.3 ng/dl predicted MIO with a sensitivity of 81.5% and a specificity of 75.3% (P<0.0001). The area under the curve was 0.79 (95% confidence interval: 0.74–0.84) (Figure 1B).

Compared to patients with normal VitD levels, patients with deficient vitamin levels showed significantly higher LV end-diastolic volume index (87.61±19.19 ml/m2 vs 79.17±14.27 ml/m2; P=0.003) (Figure 2A) and LV mass index (63.39±13.55 g/m2 vs 57.45±12.90 g/m2; P=0.006) (Figure 2B). The LV ejection fraction tended to be lower in patients with VitD deficiency, but no statistical difference versus the other groups was detected.