Vitamin D and Calcium Attenuate Bone Loss With Antiretroviral Therapy Initiation: A Randomized Trial

Abstract

Background: Antiretroviral therapy initiation for HIV-1 infection is associated with 2% to 6% loss of bone mineral density (BMD).

Objective: To evaluate the effect of vitamin D3 plus calcium supplementation on bone loss associated with antiretroviral therapy initiation.

Design: 48-week prospective, randomized, double-blind, placebo-controlled study. (ClinicalTrials.gov: NCT01403051).

Setting: 39 AIDS Clinical Trials Group units.

Patients: Adults with antiretroviral therapy-naive HIV.

Measurements: BMD by dual-energy x-ray absorptiometry, 25-hydroxyvitamin D levels, and other laboratory assessments.

Results: 165 eligible patients were randomly assigned (79 received vitamin D3 plus calcium and 86 received placebo). The study groups were well-balanced at baseline: 90% were men, 33% were non-Hispanic black, and the median CD4 count was 0.341 × 109 cells/L. At 48 weeks, the percentage of decline in total hip BMD was smaller in the vitamin D3 plus calcium group than in the placebo group: Medians were -1.36% (interquartile range [IQR], -3.43% to 0.50%) and -3.22% (IQR, -5.56% to -0.88%), respectively (P = 0.004). Similar results were seen at the lumbar spine. At 48 weeks, 90% of patients achieved HIV-1 RNA levels less than 50 copies/mL. Levels of 25-hydroxyvitamin D3 increased with vitamin D3 plus calcium but not with placebo: Median change was 61.2 nmol/L (IQR, 36.4 to 94.3) versus 1.7 nmol/L (IQR, -13.2 to 10.7) (P < 0.001). Overall, 103 patients (62%) reported 1 or more adverse event, with similar distribution between groups; no cases of hypercalcemia and 1 case of nephrolithiasis were reported in the placebo group.

Limitation: No international sites were included, and follow-up was only 48 weeks.

Conclusion: Vitamin D3 plus calcium supplementation mitigates the BMD loss seen with initiation of efavirenz/emtricitabine/tenofovir disoproxil fumarate.

Figure 1

Details and Disposition of Study Participants. No footnotes.

Figure 2. 25(OH) Vitamin D3 Levels at Baseline, 24 and 48 Weeks

25(OH) vitamin D3 levels at baseline, weeks 24 and 48 are presented in by study group. Data is presented as median value in ng/mL with error bars representing first and third quartiles (25^th and 75^th percentiles). From entry to weeks 24 and 48, 25(OH)D₃ levels did not significantly change in the placebo group (median change (Q1,Q3): −0.6 (−5.9, 5.0) and 0.7 (−5.3, 4.3) ng/mL, respectively) but levels increased significantly in the VitD/Cal group (median change: 28.6 (15.0, 38.0) and 24.5 (14.6, 37.8) ng/mL, respectively; p<0.001 at both time points). The consequence of these changes between group differences in the 25(OH)D₃ distributions was significant at both weeks 24 and 48 ( p<0.001 for both).

Figure 3. Percent Change in Bone Mineral Density from Baseline to week 48

Percent changes in bone mineral density from baseline to week 48 are presented in this figure: total hip on the left and lumbar spine on the right. For both groups there was a significant decline in median percent BMD change (p<0.001 for both sites and both groups). The percentage decline in BMD at total hip was significantly smaller in the VitD/Cal group (−1.46%; −3.16, −0.40%) compared to the placebo group (−3.19%; −5.12, −1.02%) (p= 0.001). The difference in BMD at lumbar spine between arms was of similar magnitude as the total hip, it did not achieve statistical significance: VitD/Cal group (−1.41%; −3.78, 0.00%) compared to the placebo group (−2.91%; −4.84, −1.06%) (p=0.085).

Figure 4. Percent Change in Bone Turnover Markers and PTH at 24 and 48 Weeks

Percent changes in biomarkers related to bone turnover from baseline to week 24 and 48 are presented in this figure: N-terminal propeptides of procollagen type 1 (P1NP), C-telopeptide (CTX), and parathyroid hormone (PTH) are presented in the top, middle, and bottom panels, respectively. Data is presented as median percent increase from baseline with error bars representing first and third quartiles (25^th and 75^th percentiles. In the upper and middle panels, P1NP and CTX increased significantly from baseline to week 24 and 48 in both groups (p< 0.001, for all). At week 24, the between group differences were also statistically significant (p= 0.002 for P1NP, p=0.005 for CTX) but the differences were not statistically significant at week 48 (p> 0.088 for both). The lower panel displays the percent change in PTH. In the VitD/Cal group, there was no significant change in PTH at week 24 or 48 (p>0.31 for both), while the placebo group experienced a significant increase at both 24 and 48 weeks (p< 0.001, for both). The between group differences were statistically significant for both time points (p=0.011 at 24 weeks and p=0.004 at 48 weeks)