| |
Increased Rates of Bone Fracture among HIV-Infected Persons in the HIV Outpatient Study (HOPS) Compared with the US General Population, 2000-2006 - pdf attached
|
| |
| |
Download the PDF here
Clin Infect Dis. ciq242 first published online March 10, 2011
Benjamin Young,1,4 Christine N. Dao,2 Kate Buchacz,2 Rose Baker,3 John T. Brooks,2 and the HIV Outpatient Study (HOPS) Investigatorsa
1Rocky Mountain CARES/DIDC, Denver, Colorado; 2Division of HIV/AIDS Prevention, Centers for Disease Control and Prevention, Atlanta, Georgia; 3Cerner, Vienna, Virginia; and 4Health Connections International, Amsterdam, the Netherlands
"Age-adjusted fracture rates among HOPS patients were higher than rates in the general US population during the period 2000-2006. .......In summary, our study suggests that HIV-infected adults, particularly adults age 25-54 years of age, are at an increased risk of bone fracture, compared with the general population. In light of this finding and the established relationship between low BMD and increased fracture risk in the general adult population, we recommend that screening for and correcting reversible causes of low BMD and fall risk be incorporated into routine clinical care of HIV-infected patients, as has been recently argued [20].......Screening HIV-infected adults for fracture risk has been recommended for individuals aged ≥40 years by the European AIDS Clinical Society and ≥50 years by McComsey and colleagues [19, 20]. Our findings suggest that younger HIV-infected adults are also at significant risk for fragility fractures and should be considered for similar screening interventions."
"Age-adjusted fracture rates among HIV-infected men and women ages 25-54 years in the HOPS were consistently higher than those in adults in the US general population captured by the NHAMCS-OPD during the study period. Moreover, the relative proportion of fractures at fragility sites (spine, hip, and wrist) was higher among HOPS patients than among participants surveyed in the NHAMCS-OPD. Among HOPS patients, increasing age, HCV co-infection, and BMI <18.5 were independently associated with fragility fractures.......HOPS men aged 25-54 years experienced proportionally more fractures at the wrist and vertebra (I had a wrist fracture & suspected that wirst fractures for some reason may be more likely than other parts of the body in HIV+)......HOPS women aged 25-54 years experienced proportionally more fractures at vertebra and femoral neck sites (P < .01 and P = .04, respectively).......we believe that our analysis is the first to highlight a possible association of low nadir CD4+ cell count with incident fracture rates. The causal mechanism by which low nadir CD4+ cell count is associated with low BMD and fracture risk is unclear and warrants further investigation........In univariate Cox proportional hazards analyses, age >47 years, CD4+ cell count nadir <200 cells/mm3, baseline AIDS diagnosis, current tobacco use, HCV infection, diabetes, substance abuse, and peripheral neuropathy were associated (P < .05) with increased fracture risk in the contemporary HAART era.....In multivariable analyses, factors associated with increased fracture risk included age >47 years, nadir CD4+ cell count <200 cells/mm3, HCV infection, diabetes, and substance abuse.....Risk factors associated specifically with fragility fractures during the period 2002-2008 (n = 42) in univariate analysis included increasing age (hazard ratio [HR], 1.44 per 10 years; 95% CI, 1.05-1.98), HCV infection (HR, 2.18; 95% CI, 1.12-4.26), BMI <18.5 compared with all other BMI categories (HR, 3.24; 95% CI, 1.00-10.48), peripheral neuropathy (HR, 2.34; 95% CI, 0.98-5.55), and use of SSRIs (HR, 2.01; 95% CI, 0.96-4.20). In multivariable analyses, 3 factors remained independently associated with experiencing a fragility fracture: increasing age (adjusted HR [aHR], 1.43 per 10 years; 95% CI, 1.03-1.98), HCV infection (aHR, 1.99; 95% CI, 1.01-3.90), and BMI <18.5 (aHR, 3.72; 95% CI, 1.14-12.09)."
Figure 1. Rates of bone fracture among patients in the HIV Outpatient Study (HOPS), compared with rates among patients in the National Hospital Ambulatory Medical Care Survey of Outpatient Departments (NHAMCS-OPD) for adults aged 25-54 years. HOPS rates were indirectly standardized by age and sex using rates from NHAMCS-OPD. P values represent statistical significance of trends in annual values 2000-2008 for HOPS and 2000-2006 for NHAMCS-OPD.
Abstract
Background. Among persons with HIV infection, low bone mineral density is common and has raised concerns about increased risk of fracture.
Methods. We analyzed data from the HIV Outpatient Study (HOPS), an open prospective cohort study of HIV-infected adults who were followed up at 10 US HIV clinics. We assessed rates of first fractures at any anatomic site during the period 2000-2008. We indirectly standardized the rates of fracture in the HOPS to the general population by age and sex, using data from outpatients in the National Hospital Ambulatory Medical Care Survey (NHAMCS-OPD). We examined factors associated with fractures using Cox proportional hazards modeling.
Results. Among 5826 active HOPS patients whose data were analyzed (median baseline age, 40 years; male sex, 79%; white race, 52%; exposure to antiretroviral therapy, 73%), 233 patients had incident fractures (crude annual rates, 59.6-93.5 fractures per 10,000 persons). Age-standardized fracture rates increased from 2000 to 2002 (P = .01) and stabilized thereafter. Among persons aged 25-54 years, both fracture rates and relative proportion of fragility fractures were higher among HOPS patients than among patients in the NHAMCS-OPD. In addition to older age and substance abuse, nadir CD4+ cell count <200 cells/mm3 (adjusted hazard ratio [aHR], 1.60; 95% confidence interval [CI], 1.11-2.31), hepatitis C infection (aHR, 1.61; 95% CI, 1.13-2.29) and diabetes (aHR, 1.62; 95% CI, 1.00-2.64) were associated with incident fractures.
Conclusions. Age-adjusted fracture rates among HOPS patients were higher than rates in the general US population during the period 2000-2006. Clinicians should regularly assess HIV-infected persons for fracture risk, especially those with low nadir CD4+ cell counts or other established risk factors for fracture.
Low bone mineral density (BMD) is common among persons with human immunodeficiency virus (HIV) infection [1-4]. A recent meta-analysis estimated that 67% of HIV-infected persons exhibited reduced BMD, and 15% had osteoporosis [5]. The increased prevalence of low BMD among HIV-infected patients has raised concerns that as cohorts of HIV-infected patients age, bone fracture rates will increase [6]. Although multiple analyses have reported decreased BMD in HIV-infected persons, there are few data regarding incidence rates of fractures [7-9]. One US report found that fracture prevalence was increased significantly in both HIV-infected men and HIV-infected women, compared with persons whose HIV status was negative or unknown [7]. A study of antiretroviral (ART)-treated HIV-infected persons in France from 1997 through 2007 found an increasing incidence of fractures, but this study did not include a general population comparison group [8].
In this report, we compared rates of bone fracture among a large and demographically diverse cohort of ambulatory HIV-infected adults with rates of bone fracture among adult outpatients in the general US population and explored risk factors for incident fractures among HIV-infected adults.
RESULTS
Trends in Fracture Incidence
Of 5826 HOPS patients with data analyzed (median age, 40 years), 79.0% were male, 51.8% were non-Hispanic white, and 72.7% were exposed to ART (Table 1). During a median follow-up period of 3.8 years (interquartile range [IQR], 1.5-7.4 years), 233 patients experienced a first fracture, of whom 32 (13.7%) were hospitalized. In the HOPS, sex- and age-standardized fracture rates per 10,000 persons increased from 57.7 during 2000 to 84.8 during 2002 (P = .01) and stabilized thereafter, measuring 89.9 during 2008; among NHAMCS-OPD patients, rates of fracture during 2000-2006 did not change significantly (Figure 1).
Crude fracture rates, as well as age- and sex-adjusted fracture rates, among HOPS patients and in the general US population surveyed by the NHAMCS-OPD are shown in Table 2. Differences in standardized rates were statistically significant (ie, there was no overlap between 95% CIs) in most years overall and for men, although not for women (Table 2). Among all HOPS patients and among those ages 25-54 years, who constituted 87% of all the HOPS patients included in this analysis, standardized rates of fracture increased significantly during the period 2000-2008 (P = .01, Figure 1) and were consistently higher than rates in NHAMCS-OPD during 2000-2006 (95% CIs not shown).
HOPS men aged 25-54 years experienced proportionally more fractures at the wrist and vertebra (P < .01, Table 3). HOPS women aged 25-54 years experienced proportionally more fractures at vertebra and femoral neck sites (P < .01 and P = .04, respectively). Both HOPS men and women experienced relatively fewer fractures at nonfragility sites, compared with NHAMCS-OPD patients (Table 3).
Risk Factors Associated with Fractures in HOPS Patients
In univariate Cox proportional hazards analyses, age >47 years, CD4+ cell count nadir <200 cells/mm3, baseline AIDS diagnosis, current tobacco use, HCV infection, diabetes, substance abuse, and peripheral neuropathy were associated (P < .05) with increased fracture risk in the contemporary HAART era (Table 4). There were no observed associations between risk of fracture and sex, race/ethnicity, age, BMI, ART exposure, and use of selective serotonin reuptake inhibitors (SSRIs), proton pump inhibitors, or thiazolidinediones. In multivariable analyses, factors associated with increased fracture risk included age >47 years, nadir CD4+ cell count <200 cells/mm3, HCV infection, diabetes, and substance abuse (Table 4).
Risk factors associated specifically with fragility fractures during the period 2002-2008 (n = 42) in univariate analysis included increasing age (hazard ratio [HR], 1.44 per 10 years; 95% CI, 1.05-1.98), HCV infection (HR, 2.18; 95% CI, 1.12-4.26), BMI <18.5 compared with all other BMI categories (HR, 3.24; 95% CI, 1.00-10.48), peripheral neuropathy (HR, 2.34; 95% CI, 0.98-5.55), and use of SSRIs (HR, 2.01; 95% CI, 0.96-4.20). In multivariable analyses, 3 factors remained independently associated with experiencing a fragility fracture: increasing age (adjusted HR [aHR], 1.43 per 10 years; 95% CI, 1.03-1.98), HCV infection (aHR, 1.99; 95% CI, 1.01-3.90), and BMI <18.5 (aHR, 3.72; 95% CI, 1.14-12.09).
DISCUSSION
In this large and diverse cohort of HIV-infected US adults, incident first fracture rates increased slightly during the period 2000-2008, with the increase occurring mostly during 2000-2002. Age-adjusted fracture rates among HIV-infected men and women ages 25-54 years in the HOPS were consistently higher than those in adults in the US general population captured by the NHAMCS-OPD during the study period. Moreover, the relative proportion of fractures at fragility sites (spine, hip, and wrist) was higher among HOPS patients than among participants surveyed in the NHAMCS-OPD. Among HOPS patients, increasing age, HCV co-infection, and BMI <18.5 were independently associated with fragility fractures.
Our finding that incident fracture rates among HIV-infected HOPS patients were greater than rates among predominantly HIV-uninfected counterparts in the general population confirms similar findings from prior reports [7]. The slight annual increase in fracture rates among the HOPS patients might reflect a true increase as HIV-infected patients experience improved survival or an improved capture of fracture data within the HOPS because of increasing awareness of bone health issues, or both.
We confirmed that several established risk factors for fracture (ie, older age, substance abuse, HCV co-infection and diabetes) were associated with incident fractures among HOPS patients. The Veterans Aging Cohort Study documented that incident fragility fractures were independently associated with HIV infection, as well as with cachexia, white race/ethnicity, alcohol abuse, and increasing age [17]. The Women's Interagency HIV Study (WIHS) found that fractures were associated with white race/ethnicity, increasing age, a history of AIDS-defining illness, and a history of prior fracture [18] but found no association with HIV infection in this generally young cohort of mostly pre-menopausal women [18]. Although previous studies have linked advanced HIV disease and its markers to low BMD, we believe that our analysis is the first to highlight a possible association of low nadir CD4+ cell count with incident fracture rates. The causal mechanism by which low nadir CD4+ cell count is associated with low BMD and fracture risk is unclear and warrants further investigation.
The optimal clinical management of bone health in HIV-infected individuals is not well defined and remains controversial. Screening HIV-infected adults for fracture risk has been recommended for individuals aged ≥40 years by the European AIDS Clinical Society and ≥50 years by McComsey and colleagues [19, 20]. Our findings suggest that younger HIV-infected adults are also at significant risk for fragility fractures and should be considered for similar screening interventions.
This analysis has several important limitations. First, bone fractures were ascertained from diagnoses of clinical events charted by HIV care providers and were not confirmed by review of radiographic findings or central adjudication. Second, subjects were not routinely queried about bone fractures that occurred outside of the HOPS clinic, which could have led to under-reporting of events that biased our analysis towards underestimation of fracture incidence. Third, for a majority of HOPS patients, we were not able to systematically determine the clinical venue where they initially presented for treatment of their fracture and thus could not restrict comparisons of HOPS data with national datasets by place of initial treatment (ie, hospital, outpatient clinic, or emergency department). Nevertheless, we believe that, despite methodological differences in fracture ascertainment, NHAMCS-OPD was our best available source of comparative data on fracture rates in the general population. Like the HOPS, the NHAMCS-OPD captures fractures among adults who initially present to outpatient departments for treatment of a fracture, who present to outpatient departments for follow-up treatment of a fracture, or who report a fracture to their health care provider, at which point this event is captured in their medical record and abstracted. Fourth, in both the HOPS and NHAMCS-OPD, we were unable to assess the completeness with which fracture events were captured; however, we have no a priori reason to believe that completeness would differ meaningfully between these datasets in a direction that would have biased against the null hypothesis. Fifth, although our analysis of HOPS data included >5800 patients, fractures are rare events, and rates in the HOPS are thus subject to considerable variability, which limited our ability to stratify by age, sex, and race/ethnicity. Sixth, we had no data on BMD for HOPS patients and were thus unable to assess its contribution to the risk of fracture. Finally, we did not have adequate data on use of tobacco, alcohol, or other substances of abuse among HOPS patients to explore a dose-response association of exposure to these agents with the risk of fracture.
In summary, our study suggests that HIV-infected adults, particularly adults age 25-54 years of age, are at an increased risk of bone fracture, compared with the general population. In light of this finding and the established relationship between low BMD and increased fracture risk in the general adult population, we recommend that screening for and correcting reversible causes of low BMD and fall risk be incorporated into routine clinical care of HIV-infected patients, as has been recently argued [20].
| |
| |
| |
|
|
|
