We thank Gavriilaki et al for their interest in our recent study exploring the association of vitamin D with cardiometabolic risk factors in RA.

As we outlined in our discussion, the association of vitamin D deficiency with obesity, a known risk factor for cardiovascular disease and metabolic syndrome, is well established. In our study, we did observe lower 25-hydroxyvitamin D (25[OH]D) levels in those with higher BMI and we therefore adjusted for BMI in our regression model. This adjustment resulted in attenuation of the association between 25(OH)D levels and several cardiometabolic risk factors, suggesting that the association of 25(OH)D levels with these metabolic risk factors could be partially explained by obesity.

In our cohort of RA patients, we observed no association between 25(OH)D levels and hypertension. While an inverse association between vitamin D levels and hypertension has been observed in the general population, our results were replicated in a recent publication evaluating vitamin D and metabolic syndrome in RA (Baker JF, Mehta NN, Baker DG, Toedter G, Shults J, Von Feldt JM, et al. Vitamin D, metabolic dyslipidemia, and metabolic syndrome in rheumatoid arthritis. Am J Med 2012;125:1036.e9–15). Both studies had a relatively small study sample; thus, whether any relationship between vitamin D levels and hypertension in RA exists should be further investigated in larger studies.

Although we did not exclude patients with renal insufficiency from our cohort, the prevalence of significant renal dysfunction was low, considering only 15 patients (8%) had a creatinine clearance <60 ml/minute/1.73 m2 (the creatinine clearance measurement was based on the Cockcroft-Gault equation corrected for actual body surface area). Among these 15 patients, none had a creatinine clearance <40 ml/minute/1.73 m2. While creatinine clearance was modestly inversely associated with 25(OH)D levels (Spearman's ρ −0.182, P = 0.010), it was not associated significantly with the cardiometabolic risk factors that were associated with 25(OH)D levels in the final adjusted models of our study. Therefore, further adjustment for creatinine clearance would not alter the association between 25(OH)D levels and the cardiometabolic factors reported.

We agree that our study should be considered a hypothesis-generating study. Because it was cross-sectional, the temporality of the association and causality could not be established. Thus, other well-designed prospective studies are required to further validate these observations. Whether vitamin D has a therapeutic role in the management of cardiometabolic risk factors in RA remains to be established.