CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

Methods: We have measured BMD and TBS in our Dallas cohort of 450 veterans: 45 HIV/HCV, 151 HIV, 103 HCV and 151 uninfected. We conducted analysis of covariance comparing TBS between groups, controlling for age, race, BMI and smoking. We then calculated FRAX® scores in all participants, with BMD alone (FRAX-BMD) and with TBS included (FRAX- BMD-TBS). Results: Both HIV and HCV were associated with lower total hip and femoral neck BMD (Table). Compared to controls, HCV and HIV/HCV subjects had significantly lower meanTBS scores (p=0.048 and 0.009, respectively). HIV/HCV also had lower TBS than HIV (p=0.02). Mean TBS scores were similar between HIV and controls (p=0.65) and between HCV and HIV/HCV (p=0.27) (Table). Compared to controls, HCV was associated with higher covariate-adjusted FRAX® scores (p=0.01), but HIV was not (p=0.36). The inclusion of TBS in the FRAX® calculator resulted in a significantly higher estimated absolute risk of major OF in HCV mono-infected and HIV/HCV co-infected, but not in HIV mono-infected (Table). The calculated absolute 10-year probability of fracture increased significantly when TBS was included in the HIV/HCV (+0.4; p=0.006) and HCV group (+0.3; p<0.0001) in HCV. It did not significantly change in the HIV group (-0.1; p=0.65). Conclusion: Our results suggest that the increased OF risk among HCV-infected individuals may be mediated by altered bone micro-architecture as assessed by TBS. Using FRAX- BMD-TBS may provide a more accurate risk assessment of the real fracture risk in this population.

678 FRAX-TOOL FRACTURE-RISK COMPARISON: HAVE A LITTLE BACKBONE! Linda Battalora 1 , Carl Armon 2 , Kate Buchacz 3 , John Hammer 4 , John Spear 1 , John T. Brooks 3 , Benjamin Young 5 , Edgar T. Overton 6 1 Colorado Sch of Mines, Golden, CO, USA, 2 Cerner Corporation, Kansas City, MO, USA, 3 CDC, Atlanta, GA, USA, 4 Denver Infectious Disease Consultants, Denver, CO, USA, 5 APEX Family Med, Denver, CO, USA, 6 Univ of Alabama at Birmingham, Birmingham, AL, USA Background: While osteoporosis, determined by dual energy X-ray absorptiometry (DXA), remains a key determinant for fragility fractures, the Fracture Risk Assessment (FRAX) tool incorporates several risk factors, including femoral neck (hip) BMD, to provide 10-year fracture probability (10YFP). Lumbar spine (L1-L4) BMD measurements often differ from hip BMD and are not included in the FRAX tool. A method exists that uses L1-L4 T-score to adjust FRAX tool results. We compared FRAX tool results using hip BMD “unadjusted FRAX” (uFRAX) and modified with L1-L4 T-scores “adjusted FRAX” (aFRAX). Methods: We analyzed available DXA values of the left hip, L1-L4 T-scores and clinical data collected prospectively during 2004-2012 from two CDC-sponsored HIV cohorts. Osteoporosis of either site was defined as a T-score<-2.5. We calculated FRAX 10YFP of a major osteoporotic fracture (hip, spine, forearm, or shoulder) with both uFRAX and aFRAX tool results. Harrell’s C-statistics and Cox proportional hazards analyses of factors associated with incident fracture were performed using both the uFRAX and aFRAX scores. Results: Characteristics of 1000 persons included in the analysis were: median age 43 years (interquartile range [IQR] 36-49), 83%male, 67% non-Hispanic white, median CD4+ cell count [CD4] 461 cells/mm³ (IQR 312-659). Among 86 patients with osteoporosis at either L1-L4 spine or hip, only 24 had osteoporosis at both. During 4056 person-years (py) of observation after DXA, we identified 84 incident fractures (20.7/1000py) including 22 major osteoporotic fractures (5.4/1000py). Fracture incidence increased when either the uFRAX or aFRAX score was≥3% (Figure). Of 84 fractures, 41 occurred among 291 persons (29%) with aFRAX score≥3%. Using the L1-L4 aFRAX tool had minimal effect on the estimate of fracture risk for the population. C-statistic values increased from 0.63 to 0.64 by including L1-L4 data. In multivariable Cox proportional hazards analyses with uFRAX, HCV co-infection (HCV) (Hazard Ratio [HR] 1.91, 95% confidence interval [CI] 1.15-3.19) and FRAX 10YFP≥3% (HR 2.32, CI 1.51-3.56); with aFRAX, HCV (HR 1.84, CI 1.10-3.07) and FRAX 10YFP≥3% (HR 2.59, CI 1.68-3.97); were associated with incident fracture. Conclusion: In a large convenience sample of relatively young HIV-infected US adults, a FRAX 10YFP≥3%, and HCV were significantly associated with elevated risk of incident fracture. uFRAX and aFRAX tools had similar associations with incident fracture rates. 679 FRACTURE-RISK ESTIMATES IN HIV-POSITIVE AND -NEGATIVE SUBJECTS: DATA FROM HIV UPBEAT Aoife G. Cotter 1 , Dylan Shannon 1 , Tara McGinty 1 , Alan Macken 1 , Eoin Kavanagh 2 , Geraldine McCarthy 2 , Jennifer J. Brady 2 , Juliet Compston 3 , Caroline Sabin 4 , Patrick W. Mallon 1 1 Univ Coll Dublin, Dublin, Ireland, 2 Mater Misericordiae Univ Hosp, Dublin, Ireland, 3 Cambridge, Cambridge, UK, 5 Univ Coll London, London, UK Background: Although low bone mineral density (BMD) and fractures are common in HIV it is unclear if currently available, general population-derived fracture risk prediction tools are valid in HIV. We aimed to determine if adding HIV as a secondary risk factor and/or markers of bone quality (BMD or trabecular bone score (TBS)) impacted fracture risk estimates using commonly used risk prediction tools. Methods: Fracture risk (10-year major osteoporotic and hip fracture) was calculated using the Fracture Risk Assessment Tool (FRAX), Garvan and QFracture algorithms from baseline data, including falls history, in a prospective cohort study of HIV+ and HIV- subjects. From femoral neck (FN) dual xray absorptiometry (DXA) we derived TBS using iNsight software v2.2.1. FRAX and Garvan algorithms include BMD, Qfracture and Garvan incorporate falls history but Qfracture does not include BMD. FRAX permits inclusion of TBS and non-specific secondary risks (including HIV), while Garvan and Qfracture include disease-specific secondary risks (but not HIV). Using non-parametric analyses, we compared FRAX risk estimates between groups based on clinical risk alone, recalculated with HIV as a secondary risk and again after adding FN BMD and TBS. Results are median [IQR] unless specified. Results: In 202 HIV+ (age 39 [33,46] yrs, 58%male, 40% African) compared to 263 HIV- subjects (age 42 [34,49] yrs, 44%male and 25% African), falls prevalence (past month) was similar (HIV+ 5.0% vs HIV- 6.5%, P=0.50) but BMD significantly lower in the HIV+ versus HIV- group (FN z-scores -0.3[-1.3, 0.6] vs 0.2[-0.5, 1.3] (P<0.0001)). Although there was no between group difference in FRAX risk derived from clinical factors, incorporating HIV as a secondary risk significantly increased risk estimates in the HIV+ group; major osteoporotic 0.9 [0.7, 1.9]% and hip fracture 0.1 [0.0, 0.5]%. After inclusion of BMD or BMD+TBS, FRAX risk estimates remained increased for major osteoporotic fracture but reduced for hip fracture (table 1). There were no significant between-group differences in major osteoporotic fracture risk using Garvan or Qfracture, although the HIV+ group had significantly greater Garvan hip fracture risk (table 1). Conclusion: Although fracture risk using Garvan and Qfracture was similar between groups, adding HIV to FRAX significantly amended fracture risk, with the effect attenuated for hip fracture with addition of measures of bone quality. Further research is needed to validate these tools in HIV+ populations.

Poster and Themed Discussion Abstracts

CROI 2017 294