Antiretroviral Switch Linked to Drop in Bone Density



SEATTLE – Switching to an antiretroviral regimen that contains tenofovir may adversely affect the skeletal health of HIV-positive patients, according to a new analysis reported at the Conference on Retroviruses and Opportunistic Infections.

The analysis came from the Bone Biomarker Substudy of the multicenter PREPARE trial, in which HIV-positive patients taking first-line zidovudine-lamivudine (AZT-3TC) who had virologic suppression were randomized to either continue on that therapy or switch to tenofovir-emtricitabine (TDF-FTC).

The 54 patients in the substudy underwent serial measurements of bone biomarkers and had dual-energy x-ray absorptiometry (DXA) scanning at baseline and 48 weeks.

Main results showed that the group that switched therapy had increases in markers of bone turnover and lost 2% of the bone mineral density (BMD) in their lumbar spine, reported lead investigator Dr. Aoife G. Cotter of University College Dublin. In contrast, those who continued on the same therapy did not have significant changes in these measures.

"The impact of these changes on future and overall risk of fracture is unknown, and prospective longitudinal studies are required to answer this question," she commented. Also, "while we show that tenofovir affects bone metabolism in vivo, whether this is a direct effect of tenofovir on bone cells versus an indirect effect of tenofovir on bone via the renal tubules remains to be answered."

Session attendee Dr. Todd Brown of Johns Hopkins University, Baltimore, noted that tenofovir has heterogeneous effects on bone across patients, and it would be nice to have a marker for early identification of patients who will lose BMD. "One way to do that potentially is with bone markers, so I was wondering if early absolute levels or early changes in bone markers predicted bone loss with tenofovir," he said.

The investigators did not assess that association, but a similar study looking at regimen simplification, called the STEAL study, did and found no correlation, according to Dr. Cotter. "However, when you look at our medians [for biomarker levels] plotted over time, the changes really weren’t maximal until at least 24 weeks, so we would need to have looked at changes up until at least 24 weeks to see if they correlated," she added.

Another attendee, Dr. Michael Yin of Columbia University Medical Center in New York, asked, "Are there plans to follow up with another DXA at a later date, because this is quite similar to the loss that you experience with initiation of ART, although in a different setting. It would be interesting to see whether or not the bone loss continues." Additionally, he wondered if the investigators were looking at other biomarkers.

"There are no plans that I know of to repeat DXA scans on these subjects or to increase the duration of follow-up," Dr. Cotter replied. "We do, however, have samples left and are currently performing vitamin D and parathyroid hormone analysis."

Adult HIV-positive patients were eligible for the PREPARE trial if they had been on a zidovudine-lamivudine–containing regimen as their first treatment for more than 2 years and had had a viral load of fewer than 50 copies of HIV RNA/mL for more than 6 months.

They had a median age of about 46 years, and the majority were male and white. They had been on zidovudine and lamivudine (brand names Retrovir and Epivir) for a median of roughly 5.5 years.

Results showed that compared with the group that continued on zidovudine-lamivudine, the group that switched to tenofovir plus emtricitabine (brand names Viread and Emtriva) had greater changes between baseline and week 48 in three biomarkers of bone turnover: type 1 collagen cross-linked C-telopeptide (P = .0005), osteocalcin (P less than .0001), and procollagen type 1 N-terminal propeptide (P less than .0001).

BMD of the lumbar spine fell by 2.04% from baseline in the group that switched (P = .01) but changed little in the group that continued on the original regimen, resulting in a significant difference between them (P = .03). Neither group had a significant change from baseline in BMD of the femoral neck.

The changes in levels of osteocalcin and procollagen type 1 N-terminal propeptide over the 48-week period were significantly inversely correlated with loss of BMD in the lumbar spine.

In a multivariate analysis adjusted for baseline biomarker level, sex, and ethnicity, the switch from zidovudine-lamivudine to tenofovir-emtricitabine was a significant independent predictor of changes in levels of all three biomarkers.

Among various study limitations, Dr. Cotter acknowledged, was "an absence of data on some classical risk factors, such as smoking and significant use of steroids." Also, "our follow-up was limited to 48 weeks; it would have been interesting to see what happened beyond that," she commented.

Next Article: