Seattle, Washington
March 4–7, 2019


Conference Dates and Location: 
March 4–7, 2018 | Boston, Massachusetts
Abstract Number: 



Andrew Carr1, Birgit Grund2, Ann V. Schwartz3, Anchalee Avihingsanon4, Sharlaa Badal-Faesen5, Jose I. Bernadino6, Patrick W. Mallon7, Vicente Estrada8, Alberto M. La Rosa9, Sanjay Pujari10, David White11, Nicole Wyman Engen2, Kristine E. Ensrud2, Jennifer Hoy12

1St. Vincent's Hosp, Sydney, NSW, Australia,2University of Minnesota, Minneapolis, MN, USA,3University of California San Francisco, San Francisco, CA, USA,4Chulalongkorn University, Bangkok, Thailand,5University of the Witwatersrand, Johannesburg, South Africa,6Hospital La Paz Institute for Health Research, Madrid, Spain,7University College Dublin, Dublin, Ireland,8University Hospital Clinic of San Carlos, Madrid, Spain,9Asociacion Civil Impacta Salud y Educacion, Lima, Peru,10Institute of Infectious Diseases, Pune, India,11Heart of England NHS Foundation Trust, Birmingham, UK,12Monash University, Melbourne, VIC, Australia

Abstract Body: 

Initial antiretroviral therapy (ART) in adults with normal CD4 counts accelerates loss of bone mineral density (BMD) over the first 1-2 years. Whether this loss continues with longer therapy is unclear.

We compared the effects of immediate and deferred ART on BMD change in adults in the BMD substudy of START, which randomized ART-naïve adults with CD4>500 cells/µL to immediate or deferred (CD4<350) ART. Deferred group offered ART after May 2015. BMD was measured annually for up to 5 years at the total hip and lumbar spine (L1-L4) by dual-energy X-ray absorptiometry. Mean percent changes in BMD from baseline and annual percent changes were estimated and compared between treatment groups using longitudinal mixed models, by intention to treat (ITT) and by ART use (Immediate vs. no ART [Deferred group censored at ART start]). We also assessed predictors of BMD change within group.

411 participants were included (Immediate=201; Deferred=210). Median baseline age was 32 years, with 80% non-white, 24% women and median CD4 count 643 cells/µL. Groups were well balanced at baseline. The most common initial drugs in the Immediate group were tenofovir disoproxil fumarate (TDF; 83%) and efavirenz (79%); a protease inhibitor was used by 13%. Mean follow-up was 4.5 years. In the Immediate group, 96%-97% of participants used ART throughout Years 1-5. In the Deferred group, 16%, 28%, 58%, and 85% used ART at the Year 1, 2, 3, and 4 visits, respectively. BMD changes by ITT and by ART use are shown in the Table. Averaged through follow-up, BMD decreased more in the Immediate versus the Deferred group (Table, first 2 rows), but groups converged by Year 3 at the spine (diff=-0.5, p=0.26) and Year 4 at the hip (diff=-0.2, p=0.68) as most Deferred group participants started ART. In the Immediate group, BMD declined by 2.1% at the spine and 2.0% at the hip during Year 1; afterwards, BMD was stable at the spine and continued to decline at the hip by 0.5% per year. The annual rates of BMD change after Year 1 were similar in the Immediate group and those who remained ART-naïve in the Deferred group. No clinical, HIV-related or ART characteristic consistently predicted greater BMD loss with immediate ART (including use of TDF), or while ART-naïve in the Deferred group.

BMD declined at both hip and spine after ART initiation. After Year 1, BMD change was comparable between the Immediate group and those who remained ART-naive, suggesting that bone loss slows after the first year of ART.

Session Number: 
Session Title: 
Presenting Author: 
Andrew Carr
Presenter Institution: 
St Vincent's Hospital