| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Divisions of Endocrinology (T.K.), Infectious Diseases (P.K., J.T.), Department of Medicine (T.T.B., M.D.R.), General Clinical Research Center (R.K.), Georgetown University Medical Center, Washington, D.C. 20007; and Division of Endocrinology and Metabolism (T.T.B., M.D.R., A.S.D.), Johns Hopkins University, Baltimore, Maryland 21287
Address all correspondence and requests for reprints to: Todd T. Brown, M.D., Division of Endocrinology and Metabolism, Johns Hopkins University, 1830 East Monument Street, Suite 333, Baltimore, Maryland 21287. E-mail: tbrown27{at}jhmi.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
More recently, disturbances in bone metabolism, particularly osteopenia and osteoporosis, have been described in young HIV-infected patients, with prevalence estimates in cross-sectional studies varying from 55% (1) to 89% (2). In addition, disordered bone remodeling, with increased bone resorption and changes in bone formation, has also been reported (3).
The mechanisms underlying these changes in bone metabolism are unclear and are likely multifactorial in nature. Protease inhibitors (4, 5, 6) and nucleoside analogs (7) have both been implicated in the pathogenesis of HIV-associated bone disease, but these results are controversial (8). Other nonmedication factors such as cytokine elaboration induced by HIV infection (9) and low pretreatment body weight (7) have also been associated with bone disease.
Also controversial is the relationship between the metabolic and morphologic complications and the bone disorders. Huang et al. (10) demonstrated that reduced bone mineral density (BMD) was independently correlated with abdominal visceral fat measured by quantitative computed tomography in HIV-infected patients. Carr et al. (7) found no association between reduced BMD and dyslipidemia or insulin resistance in cross-sectional study of 221 HIV-infected men. The demonstration of an association between the metabolic and bone complications would provide further evidence for a shared mechanism. In addition, such an association would help to identify patients who are at increased risk of osteopenia and osteoporosis and who, therefore, may benefit from screening for this largely asymptomatic condition.
We conducted a cross-sectional study of HIV-infected men and women and healthy control subjects to determine: 1) the prevalence of reduced BMD in HIV-infected subjects and 2) whether reduced BMD or markers of bone turnover are associated with impaired glucose tolerance, dyslipidemia, central fat accumulation, or peripheral fat wasting in HIV-infected patients.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Between October 2000 and June 2001, 51 HIV-infected patients at the Georgetown University HIV clinic on HAART who met the inclusion and exclusion criteria were successively enrolled in the study. In addition, 22 healthy subjects meeting the inclusion and exclusion criteria were enrolled as control subjects. Control subjects were recruited by advertisements and word of mouth and were frequency matched to the HIV-infected patients by age and gender. Inclusion criteria included an age of 30–55 yr, CD4 count greater than 200 cells/mm3 and treatment with HAART [either protease inhibitor (PI) based or nonnucleoside reverse transcriptase inhibitor (NNRTI) based] greater than 6 months. Exclusion criteria were an opportunistic infection in the past 3 months; treatment with steroids in the past 12 months or for more than 3 months; amenorrhea for more than 6 months or postmenopausal state; treatment with hormonal agents including GH, thyroid hormone, testosterone, megestrol, or oxandrolone in the past year (estrogen/progesterone used for contraception was permitted); or an underlying medical condition predisposing to osteoporosis. In addition, patients with current or past treatment with both PIs and NNRTIs were excluded to be able to assess more accurately the independent contribution of these classes of antiretrovirals to the metabolic and skeletal outcome measures. All volunteers gave written informed consent before participating. The study was approved by the Georgetown University Institutional Review Board.
Materials and methods
Age, duration of HAART, type of HIV therapy, history of AIDS (past CD4 count < 200 cells/mm3 or AIDS-defining condition), and body mass index (BMI) were recorded at initial assessment. Waist/hip ratio was defined as the circumference of the waist at the umbilicus divided by the circumference of the hips across the greater trochanter.
After an overnight fast, a standard 2-h, 75-g oral glucose tolerance test (OGTT) was performed in addition to a fasting lipid profile. Glucose measurements were done at baseline and 30, 60, 90, and 120 min after the glucose load. The plasma glucose concentration 2 h after the glucose load was used as a primary measure of hyperglycemia. Glucose area under the curve (AUC) was calculated by the trapezoidal method. Non-high-density lipoprotein (HDL) cholesterol, which is the most sensitive cholesterol marker of cardiovascular risk in patients with the metabolic syndrome (11) and type 2 diabetes mellitus (12), was used as a measure of dyslipidemia. Serum osteocalcin (a marker of bone formation) and 2-h urinary deoxypyridinoline (a marker of bone resorption) were also measured. Urinary deoxypyridinoline has been shown not to vary significantly during the menstrual cycle (13). Serum calcium, phosphorous, intact PTH, thyroid stimulating hormone, testosterone (males), estradiol (females), IGF-1, 1,25 dihydroxy vitamin D, 24-h urinary-free cortisol, and 24-h urinary calcium were measured to assess for secondary causes of reduced BMD. All laboratory tests were processed by Quest Diagnostics, Inc. (Baltimore, MD) using standard techniques. All testing took place at the Georgetown University General Clinical Research Center.
Within 2 wk of the laboratory assessment, regional body composition, and BMD testing was determined by dual x-ray absorptiometry (DXA) using a single machine after daily calibration (Hologic-4500, Hologic Co., Waltham, MA, intrasubject coefficient of variation, 2%). Central adiposity was determined by the ratio of truncal fat mass (in grams) to total fat mass. Extremity fat was calculated as the ratio of the extremity fat mass to total fat mass. Ratios were multiplied by 100 to allow better interpretation of the resulting measure. These calculations have been previously used to measure adiposity in HIV-infected patients (14). All anthropomorphic and DXA measurements were done by the same two investigators (T.B and M.R.).
BMD was determined at each of the following sites: the spine, total hip, hip neck, and total forearm. Subjects were classified as having reduced BMD if the BMD at any of these sites was greater than 1 SD less than that of young volunteers in the Hologic database (t-score < -1.0), in accordance with World Health Organization t-score criteria (15). Normal BMD was defined as a t-score greater than -1.0. Osteopenia is defined as a t-score between -1.0 and -2.5 at any of the four sites examined, and osteoporosis is defined as t-score < -2.5.
Statistical analysis
Comparisons of the demographic and metabolic data between the HIV-infected and healthy control groups were using Student’s t test or 
2 analysis, depending on the type of variable used. A similar analysis was performed in the subset of the HIV-infected subjects, stratifying by the presence or absence of reduced BMD. In this subset, bivariate associations between reduced BMD and demographic and metabolic variables were determined by logistic regression. A multivariate model was constructed using independent variables which were significantly associated (P < 0.05) with the presence of reduced BMD on bivariate analysis. Likelihood ratio testing was used to test the independent significance of each covariate in the model, and results were confirmed with forward stepwise regression. Age and BMI were considered to be required covariates.
Predictors of bone formation (serum osteocalcin) and bone resorption (urine deoxypyridinoline) were determined by bivariate and multivariate linear regression. Forward, stepwise regression was performed to identify important covariates. All analyses were performed using STATA 7.0 (College Station, TX) and P < 0.05 were considered significant.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Comparisons between the 51 HIV-infected subjects and the 22 healthy control subjects are presented in Table 1
. The two groups were comparable on all demographic variables, including age, gender, and race. Among the 18 HIV-infected subjects classified as non-White, 15 were African American, one was Hispanic, and two were Asian. The non-White HIV-negative subjects included two African-Americans and two Asians. Sixty-three percent (32 of 51) of the HIV-infected patients had reduced BMD (28 osteopenia, four osteoporosis), compared with 32% (7 of 22) of the control subjects (seven osteopenia, 0 osteoporosis) (P = 0.02). T-scores by DXA in the HIV-infected group were significantly lower in the spine and hip but similar in the forearm. None of the subjects had a history of atraumatic fracture. Markers of bone resorption (urine deoxypyridinoline) were higher in the HIV group, whereas bone formation markers were similar between the two groups, suggesting uncoupling between bone formation and bone resorption in HIV-infected subjects. HIV-infected subjects also had higher glucose concentrations during the OGTT (glucose AUC, 16740 ± 3759 vs. 14805 ± 3472 mg/dl/min, P = 0.043), increased central adiposity (trunk fat/total fat, 0.575 ± 0.091 vs. 0.496 ± 0.068, P < 0.001), decreased fat in the extremities (extremity fat/total fat, 0.342 ± 0.013 vs. 0.437 ± 0.012, P < 0.001), and higher total and non-HDL cholesterol (total cholesterol, 194 ± 7.2 vs.172.3 ± 5.9 mg/dl, P = 0.065; non-HDL cholesterol, 152.8 ± 52.5 vs.121.9 ± 33.7 mg/dl, P = 0.013), compared with the healthy control subjects. One HIV-infected subject was unable to complete the OGTT.
|
Table 2 presents the demographic, clinical, and metabolic characteristics of the 51 HIV-infected subjects, grouped by the presence or absence of reduced BMD (osteopenia or osteoporosis). No significant differences in demographic or clinical variables were found between the two groups, including type of HAART regimen used (PI sparing vs. PI containing) or duration of HAART treatment. Nineteen subjects were taking one PI, five subjects were on dual-PI therapy, and one subject was on three PIs (overall, six on saquinavir, seven on ritonavir, 13 on indinavir, and six on nelfinavir). Among the 26 subjects on a PI-sparing regimen (NNRTI based), there were 14 on nevirapine, 11 on efavirenz, and one on delavirdine. All HIV-infected subjects were taking two nucleoside analogs (17 on zidovudine, 33 on stavudine, 42 on lamivudine, four didanosine, four on abacavir, and one on zalcitabine). One subject was taking three nucleoside analogs (zidovudine, lamivudine, abacavir) in addition to efavirenz. No specific antiretroviral medication was associated with the presence of reduced BMD (data not shown). The mean CD4 count and percentage of subjects with an undetectable viral load did not differ between the groups. The range of viral load in those 10 subjects with detectable viral load was 132–13,923 copies/ml. Markers of bone formation and bone resorption were also similar between the two groups.
|
Figure 1 shows the mean glucose concentrations for both groups during the OGTT at each time point. Both fasting plasma glucose and 2-h postload glucose were significantly higher in those HIV-infected patients with reduced BMD. The 90-min postload concentrations tended to be higher in the reduced BMD group (148 ± 48 vs. 126 ± 35 mg/dl, P = 0.089), whereas the 30- and 60-min glucose concentrations were similar between the two groups. The glucose AUC tended to be higher in those HIV-infected subjects with reduced BMD (17,465 ± 4000 vs. 15,557 ± 3068 mg/dl·min, P = 0.081).
|
). The addition of age and BMI into the model did not affect the associations. In the multivariate analysis, the risk of reduced BMD increased by 2% for each milligram per deciliter increase in the 2-h postload plasma glucose (odds ratio 1.02, 95% confidence interval 1.00–1.05, P = 0.048). Independent of postload glucose concentrations, the risk of reduced BMD increased by 10% for each 1% increase in the ratio between the central fat and total fat (odds ratio 1.10, 95% confidence interval 1.00–1.20, P = 0.043). No effect modification was observed between the type of HAART therapy (PI sparing vs. PI containing) and the metabolic variables (data not shown).
|
). These findings did not change when extremity fat, type of HAART therapy, and duration of HAART therapy were included as required covariates. No effect modification was found between these independent variables and the type of HAART therapy (data not shown).
|
Table 5 presents the results for the laboratory evaluation of secondary causes of reduced BMD. Table 5A compares the HIV-infected subjects with the HIV-negative subjects. With the exception of urinary-free cortisol, there were no significant differences between the groups. Whereas the HIV-infected subjects had significantly lower urinary-free cortisol values than the HIV-negative subjects, both mean values were well within the normal range. Table 5B shows that there were no differences between the HIV-infected subjects with reduced BMD and those with normal BMD in any of the variables examined. IGF-1 tended to be lower in HIV-infected subjects with reduced BMD, compared with HIV-infected subjects with normal BMD (162.9 ± 56.0 vs. 187.8 ± 39.3 ng/ml, P = 0.094). This apparent association with reduced BMD was not seen when IGF-1 was included in a multivariate logistic regression model with age, BMI, plasma glucose at 2 h, and trunk fat/total fat, although the magnitude and precision of the other estimates in the model did not change (data not shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The association between postload hyperglycemia and reduced BMD has not been previously reported. Carr et al. (7) showed no difference between 211 HIV-infected patients with normal BMD and those with reduced BMD in fasting glucose concentration, fasting insulin concentration, or insulin resistance as measured by the homeostasis model. However, glucose was not measured after an OGTT, which is more sensitive than fasting glucose measurements and correlates more closely with both microvascular and macrovascular complications of hyperglycemia (16).
The mechanisms underlying the observed associations deserve closer scrutiny. Although some reports have demonstrated an increased prevalence of osteopenia and osteoporosis in patients with type 2 diabetes mellitus (17), perhaps as a result of osteoblast dysfunction in the setting of hypergylcemia (18), a direct causal pathway between hyperglycemia and reduced BMD seems unlikely in this population given the mild hyperglycemia observed. A more likely explanation is that these two disorders, along with central adiposity, share a common mechanism.
The use of protease inhibitors would seem to be the most likely common causative factor. Several protease inhibitors, including ritonavir, indinavir, and saquinavir, have been shown to increase osteoclast activity in vitro (19), perhaps leading to reduced BMD. Similarly, the protease inhibitor, indinavir, has been shown to directly cause insulin resistance by inhibiting glucose movement through the glut 4 transporter (20). Whereas this common mechanism seems plausible, we found no difference between the reduced and normal BMD groups in the use of protease inhibitors. Huang et al. (10) also failed to find an association with the use of protease inhibitors. However, both studies did not have sufficient power to examine the relative contribution of individual protease inhibitors.
Nucleoside analogs are used in all HAART regimens and have been indirectly implicated in the pathogenesis of reduced BMD because high plasma lactate levels, a known consequence of nucleoside analog therapy, have been shown to be independently associated with reduced BMD (7). However, abnormalities in glucose homeostasis and central adiposity have not been directly linked to the use of nucleoside analogs, and therefore it is less likely that this type of therapy is a common mechanism. This hypothesis was not specifically evaluated in the present study.
The link between the metabolic and bone abnormalities is likely complex, resulting from an interaction of medications, disease factors, and patient factors, such as a genetically based vulnerability. Susceptibility genotypes are beginning to emerge in the metabolic complications of HIV. For example, certain apolipoprotein C-III polymorphisms were found to account for approximately 43% of the variability-associated hypertriglyceridemia and low HDL in HIV patients on HAART, suggesting an interaction between medication and host factors (21). Common genotypes may also underlie the development of hyperglycemia, central adiposity, and reduced BMD and deserve further investigation.
Bone markers have been shown to predict fracture risk independently of BMD in some populations, but their use remains controversial (22). Markers have also been used to assess the relationship between bone formation and bone resorption, a process that is normally coupled. In HIV patients, uncoupling of bone formation and bone resorption has been shown to relate to disease activity and cytokine activation (23), which appear to suppress bone formation, as measured by serum osteocalcin. Interestingly, in the setting of highly active antiretroviral therapy, markers of bone formation increase, whereas markers of bone resorption remain constant, suggesting partial recoupling of these two processes. Enhanced bone resorption, however, has been demonstrated during a longitudinal analysis of HIV patients on therapy, consistent with a continuous high bone turnover state (3). Other studies have shown normal bone formation and resorption markers in HIV patients with and without clinical lipodystrophy, compared with HIV-negative patients (10).
In the present study, we found similar markers of bone formation (serum osteocalcin) in HIV-infected patients on HAART, compared with HIV-negative patients, but elevated markers of bone resorption, confirming the previous findings of increased bone turnover in HIV patients (3, 24). In addition, we found that bone resorption was independently related to non-HDL cholesterol, an association that has not been previously reported. A direct, causal relationship between elevated non-HDL cholesterol and bone resorption is unlikely, but the association probably represents and interaction between patient and medication factors. Because patients with increased bone resorption may be at an increased risk of fracture, the finding of an association between urine deoxypyridinoline and dyslipidemia may help to identify high-risk patients.
Our study has several limitations. Given our small sample size, we were unable to fully explore potential common mechanisms, such as the effects of individual protease inhibitors or nucleoside analogs. These hypotheses will need to be tested in larger populations. Second, the cross-sectional design makes the evaluation of causal relationships difficult. Third, DXA was used to assess fat content. Although this methodology is considered accurate, both sc and visceral fat is measured, which may be metabolically distinct compartments. Finally, the small proportion of women reduces the external validity of the results when applied to women. Because of important gender distinctions in fat composition and bone metabolism, further work is required to confirm these findings in a population of women. Despite these limitations, these findings should raise new interest in searching for a common mechanism between the metabolic and skeletal complications of HIV.
This association may also have implications for bone density screening. Despite the widespread reports of increased prevalence of reduced BMD in HIV patients, their risk of fracture has not yet been determined. As a result, current management guidelines recommend against population screening with DXA (25). With confirmation of the current findings in larger populations, patients with central adiposity and hyperglycemia may represent a group of patients who are an increased risk of osteopenia and osteoporosis and may benefit from screening.
|
Footnotes |
|---|
This work was presented in part at the 9th Conference on Retroviruses and Opportunistic Infections, Seattle, Washington, 2002.
Abbreviations: AUC, Area under the curve; BMD, bone mineral density; BMI, body mass index; DXA, dual x-ray absorptiometry; HAART, highly active antiretroviral therapy; HDL, high-density lipoprotein; NNRTI, nonnucleoside reverse transcriptase inhibitor; OGTT, oral glucose tolerance test; PI, protease inhibitor.
Received August 29, 2003.
Accepted November 12, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  V. A. Triant, T. T. Brown, H. Lee, and S. K. Grinspoon Fracture Prevalence among Human Immunodeficiency Virus (HIV)-Infected Versus Non-HIV-Infected Patients in a Large U.S. Healthcare System J. Clin. Endocrinol. Metab., September 1, 2008; 93(9): 3499 - 3504. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Bolland, A. B. Grey, G. D. Gamble, and I. R. Reid Low Body Weight Mediates the Relationship between HIV Infection and Low Bone Mineral Density: A Meta-Analysis J. Clin. Endocrinol. Metab., December 1, 2007; 92(12): 4522 - 4528. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. E. Dolan, J. R. Kanter, and S. Grinspoon Longitudinal Analysis of Bone Density in Human Immunodeficiency Virus-Infected Women J. Clin. Endocrinol. Metab., August 1, 2006; 91(8): 2938 - 2945. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Koutkia, B. Canavan, J. Breu, and S. Grinspoon Effects of Growth Hormone-Releasing Hormone on Bone Turnover in Human Immunodeficiency Virus-Infected Men with Fat Accumulation J. Clin. Endocrinol. Metab., April 1, 2005; 90(4): 2154 - 2160. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, P = 0.089.