Abstract

MOPE087 - Poster Exhibition

Proton pump inhibitor and protease inhibitor use are associated with fragility fracture risk in HIV-positive male veterans

J. Womack^1,2, J. Goulet^2,3, C. Gibert^4,5, C. Brandt^2,6, B. Gulanski^2,7, D. Rimland^8,9, M. Rodriguez-Barradas^10,11, J. Tate², M. Yin¹², A. Justice^2,7, Veterans Aging Cohort Study Project Team

¹Yale School of Nursing, New Haven, United States, ²VA Connecticut Healthcare System, West Haven, United States, ³Yale School of Medicine, Psychiatry, New Haven, United States, ⁴Veterans Affairs Medical Center, Washington, United States, ⁵George Washington University, Medicine, Washington, United States, ⁶Yale Center for Medical Informatics, New Haven, United States, ⁷Yale School of Medicine, Internal Medicine, New Haven, United States, ⁸Veterans Affairs Medical Center, Atlanta, United States, ⁹Emory University School of Medicine, Atlanta, United States, ¹⁰Michael E. DeBakey VA Medical Center, Medical Service, Houston, United States, ¹¹Baylor College of Medicine, Medicine, Houston, United States, ¹²Columbia University Medical Center, Medicine, New York, United States

Background: Prior research suggests that severity of illness plays a role in fracture risk among HIV infected (HIV+) individuals. The VACS Index has been shown to be a novel predictor of 5 year mortality and ICU admissions for HIV+ individuals and thus may capture illness severity. We explored the association of traditional and HIV-specific risk factors for fragility fracture when controlling for the VACS Index as a time-updated covariate.
Methods: Cox regression was used to assess risk factors for incident hip, vertebral, or upper arm fracture as defined by ICD-9-CM codes in HIV+ men enrolled in the Veterans Aging Cohort Study (VACS). The VACS Index, protease inhibitor use, tenofovir use, steroid use, and proton pump inhibitor (PPI) use were modeled as time-varying covariates. All other variables were evaluated at study enrollment. Missing data was imputed using SAS PROCs MI and MIANALYZE.
Results: [40,115 HIV+ male Veterans experienced 588 fragility fractures over a mean follow-up of 6.0±3.9 years (Tables 1-3). The VACS Index (HR: 1.08; 95% CI: 1.06, 1.11), white race (1.94; 1.64, 2.30), body mass index (BMI) (0.88;0.78, 0.98), alcohol-related diagnoses (1.58; 1.21, 2.08), cerebrovascular disease (CVD) (1.87; 1.09, 3.20), PPI use (1.81; 1.48, 2.20), and protease inhibitor use (1.28; 1.07, 1.52) were associated with fracture risk (Tables 2 and 3).

Variable	Means, HIV+ Male Veterans N=40,115
Follow-up time	6.0±3.9 years
Age at baseline	46±10 years
Age at time of fracture	53±10 years

[Table 1. Sample Characteristics (Means)]

Variable	Median (IQR), HIV+ Male Veterans N=40,115	Hazard Ratio (95% CI)
CD4 (cells/mm3) at baseline	278 (112, 472)	--
HIV RNA (copies/mL) at baseline	10,408 (480, 77987)	--
VACS Index Score	33 (18-53)	1.08 (1.06, 1.11)
BMI (m/kg2) at baseline	25 (22, 28)	0.88 (0.78, 0.98)

[Table 2. Sample Characteristics(Median)/Regression]

Variable	Percent, HIV+ Male Veterans, N=40,115	Hazard Ratio (95% CI)
White race	37	1.94 (1.64, 2.30)
Alcohol-related diagnoses	16	1.58 (1.21, 2.08)
Cerebrovascular disease/stroke	2	1.87 (1.09, 3.20)
Any proton pump inhibitor use	36	1.81 (1.48, 2.20)
Any protease inhibitor use	64	1.28 (1.07, 1.52)
Smoking (past or current)	75	1.11 (0.90, 1.36)
Any steroid use	21	1.36 (0.97, 1.91)
Tenofovir use	22	1.16 (0.89, 1.52)
Drug use/abuse	19	1.08 (0.82, 1.42)

[Table 3. Sample Characteristics (%) /Regression]


Conclusions: Even when adjusting for the VACS Index, a novel marker of HIV-disease severity, PPI and protease inhibitor use were associated with fragility fractures among HIV+ male Veterans. CVD, an important predictor of fragility fractures in the general population likely mediated through falls, was also an important predictor of fractures. Steroid use and smoking were not significant risk factors in this model. The low level of exposure to steroids and the fact that we accounted for BMI, alcohol use, and disease severity may account for the weak associations.
Download the e-Poster