Prevalence and Predictors of Low Bone Mineral Density and Fragility Fractures Among HIV-Infected Patients at One Italian Center After Universal DXA Screening: Sensitivity and Specificity of Current Guidelines on Bone Mineral Density Management

Abstract

Low bone mineral density (BMD) is frequent in HIV infection regardless of the use of antiretroviral therapy (ART). Uncertainties remain, however, as to when in HIV infection BMD screening should be performed. We designed a prospective study to estimate the efficacy of universal BMD screening by dual-energy X-ray absorptiometry (DXA). Since April 2009 through March 2011, HIV patients attending our Center were offered femoral/lumbar DXA to screen BMD. Low BMD for chronological age, that is significant osteopenia, was defined as a Z-score ≤ −2.0 at femur and lumbar spine. Nontraumatic bone fractures (NTBFs) were evaluated. The final sample included 163 patients. A Z-score ≤ −2.0 at any site was observed in 19.6% of cases: among these, 18.8% had no indication to DXA using current Italian HIV guidelines for BMD screening. A lower femoral Z-score was independently associated with lower BMI, AIDS diagnosis, HCV co-infection, antiretroviral treatment, and NTBFs; a lower lumbar Z-score with age, BMI, Nadir CD4 T-cell counts, and NTBFs. Prevalence of NTBFs was 27.0%, predictors being male gender, HCV co-infection, and lower femoral Z-scores. Our results suggest that measuring BMD by DXA in all HIV patients regardless of any further specification may help retrieving one-fifth of patients with early BMD disorders not identified using current criteria for selective screening of BMD.

Introduction

Several clinical and basic science studies in recent years documented that bone mineral density (BMD) disorders, mainly osteopenia (OP) and osteoporosis (OT), are more frequent in human immunodeficiency virus (HIV) infected subjects than in the general population.^1

–6 Best estimates gauge that the overall frequency of BMD is threefold higher in HIV-infected patients, despite a preponderant proportion of males.^3,4

A persistent inflammatory status, drug addiction, vitamin D and other nutritional deficiencies, low Body Mass Index (BMI), hypogonadism, hyperthyroidism, hyperparathyroidism, as well as renal dysfunction, have been indicated as independent predictors of OP in this setting.^1

–6 An association between OP, duration of HIV infection, and length of highly active antiretroviral therapy (HAART) was also repeatedly reported.^1

–6 In particular, among antiretrovirals, tenofovir (TDF), protease inhibitors (PIs), and efavirenz (EFV) were found to be related with BMD disorders through various mechanisms.^3

–6 Finally, hepatitis C virus (HCV) co-infection was recently associated with an increased risk of BMD reduction and nontraumatic bone fractures (NTBFs) in HIV infected patients.^7
–9

Despite the above recalled evidence, it is still uncertain at what stage of HIV infection bone turnover and BMD should be measured in clinical practice. Current Italian guidelines for the management of HIV infection emphasize the importance of assessing bone metabolism and BMD in order to support the early management of metabolic abnormalities and the prevention of bone fragility;^10,11 dual-energy X-ray absorptiometry (DXA) is strongly recommended for HIV patients with at least two risk factors other than HIV.^10,11 Despite that, a recent Italian multicentric survey revealed that BMD screening was judged as a relevant priority only by 55% of the responding centers, and even in the remaining units only selected patients were screened, mainly those with low energy fractures, current or former use of steroids, and low BMI.¹² Interestingly, other HIV-related risk factors, such as AIDS diagnosis, long-term HAART, low CD4 T-cell Nadirs, high viral loads, and wasting were rarely taken into consideration.¹²

From 2009 to 2011 we carried out a prospective study to evaluate the potential efficacy of a universal BMD screening policy at our HIV Clinic, systematically searching for BMD disorders by DXA, and investigating potential predictors of bone fractures in a consecutive, unselected sample of HIV-infected patients, including all males and asymptomatic attendees. Here we are reporting the results of the baseline evaluation of all included patients.

Materials and Methods

Study design

The study protocol was approved by the local Ethical Committee and Health District Authorities in 2008 (Authorization n° 943; August 28, 2008). Since April 2009 through March 2011, all HIV patients attending our Center at any stage of the disease (during either first visit or follow-up) were offered participation. Patients providing a written informed consent had to compile a clinical questionnaire, undergo laboratory screening, and a femoral and lumbar DXA. According to the approved study protocol, all participants will be re-evaluated 36 months after enrollment.

History of NTBFs in adulthood,¹³ BMI, presence of any type of lipodystrophy,¹⁴ blood pressure,¹⁵ HCV co-infection, co-morbidities, smoking, alcohol and drug abuse, regular physical activity, working, and family status were evaluated at enrollment, together with the type and line of HAART, Nadir, and current CD4 T-cell counts and current HIV viremia. Among laboratory parameters, serum 25-OH D3, calcemia, osteocalcin, bone alkaline phosphatase, ACTH, TSH, FSH, and PTH were assayed.^6,16,17 BMD was measured by femoral and lumbar spine Hologic QDR 4500A DXA and was expressed in grams of mineral per square centimeter scanned (Hologic Inc., Bedford, MA, USA).¹⁸ The evaluation of normal or low BMD was performed comparing expected BMD for the patient's age and sex (Z-score) or “young normal” adults (T-score): differences were expressed in standard deviation (SD). According to WHO classification,¹⁹ OP is diagnosed when T-scores range between −1 and −2.5 standard deviations; OT when T-scores are below −2.5 SD.¹⁹ However, in consideration of the fact that our universal sample was prevalently constituted by men <50 years old and premenopausal women, low BMD for chronological age was defined as a Z-score ≤ −2.0, in line with current guidelines and previously reports on the HIV population.^7,20,21

Study outcomes

According to current guidelines, given the relatively young age of our population, the outcome measures were femur and lumbar Z-scores (ZSF and ZSL, respectively) and low BMD for chronological age, that is significant OP, was dichotomically defined as a Z-score ≤ −2.0 SD.²⁰

However, additional analyses were performed considering femur and lumbar T-Score (TSF and TSL, respectively) values and two additional dichotomizations were introduced for T-scores < −1.0 versus ≥ −1.0 and for T-scores < −2.5 versus ≥ −2.5.

Data analysis

Initial descriptive statistics were used to analyze the distribution of low BMD for chronological age in the sample and according to all variables investigated. Both multivariable logistic and linear regression analyses were used to evaluate potential independent predictors of low BMD. In logistic analyses (three separate models), the dependent variables were femoral low BMD, lumbar low BMD, and femoral or lumbar low BMD (identified by a Z-score ≤ −2.0). In linear regression analyses (two models), the dependent variables were ZSF and ZSL. We defined the regression models a priori including potential confounders (age, gender, and duration of HIV infection) and used a forward selection process for the other variables, adopting the following criteria for inclusion: p<0.1; change in significant odds ratios (ORs) or coefficient >30%. Each covariate was tested in its original form or transformed if needed (Shapiro-Wilk test). In addition, each variable included was tested for multicollinearity, for potential interaction, and/or quadratic/cubic terms.²²

In logistic regression analysis, the outlier analysis was based upon the calculation of Pearson and standardized residuals, the change in Pearson chi-square and deviance chi-square, Dbeta influence statistic and leverage (hat diagonal matrix). The validity of the final linear regression model was assessed as follows. The assumption of constant error variance was checked graphically, plotting Pearson residuals vs. fitted values, and formally, using the Cook-Weisberg test for heteroskedasticity. High leverage observations were identified by computing Pearson, standardized and studentized residuals, Cook's D influence, and Welsch distance.²² We found seven influential observations in the logistic model and eleven high-leverage observations in the linear model. In both cases, we repeated the analysis excluding these observations, with no substantial changes. We thus reported the results of the analyses including the entire sample. Missing items were <5% in all models, and no missing data imputation technique was adopted.

An additional logistic regression model was fit—following the same above criteria—to investigate potential predictors of NTBFs, and confidence Intervals (CI) for specificity and sensitivity were computed according to the efficient-score method (corrected for continuity) described by Newcombe.²³

Statistical significance was defined as a two-sided p-value<0.05 for all analyses, which were performed using STATA 10.1 (Stata Corp., College Station, TX, 2007).

Results

Characteristics of the sample

The study was proposed to 241 consecutive patients; 186 (77.2%) agreed to participate. Of these, however, only 163 (88.1%) underwent BMD measurement by DXA, entering the final sample. The main characteristics of the final sample are reported in Table 1: 70.1% males; mean age 44.2±10.0 years; mean BMI 23.7±4.1; mean duration of HIV infection 13.5±7.1 years. Mean Nadir CD4 T-cell counts were 251±192 cells/mmc, mean CD4 T-cell counts at enrollment 538±285 cells/mmc. HAART-naïve patients were 33 (20.3%); among 130 patients treated with HAART, 77 (59.2%) were taking PIs and 91 (70.0%) had TDF in their HAART backbone.

Table 1.

Characteristics of the Sample

	Screened with DXA (n=163)
Male gender, %	70.6
Mean age in years (SD)	44.2 (10.0)
Risk factors for HIV infection, %
Drug abusers	22.7
Heterosexuals	48.5
Homosexuals	26.4
Transfusions	2.5
AIDS diagnosis, %	35.0
Mean duration of HIV infection in years (SD)	13.4 (7.1)
HCV co-infection, %	22.7
HAART naïve, %	20.3
Third drug in HAART^#, %
Lopinavir/r	19.6
Atazanavir/r	12.9
Other PIs	13.5
Efavirenz	15.3
Nevirapine	16.0
Tenofovir in HAART^# backbone, % (n)	70.0 (130)
Mean Nadir CD4 T-cell counts (SD), cell/mmc	251 (192)
Mean CD4 T-cell at enrollment (SD), cell/mmc	538 (285)
Mean body mass index (SD^*)	23.7 (4.1)
Any type of lipodystrophy, %	43.0
Diagnosis of hypertension, %	22.1
Gainful employment, %	68.7
Married, %	41.7
Habitual physical practice, %	37.0
Cigarette smokers, %	49.7
Alcohol abusers, %	8.6
Mean osteocalcina (SD), ng/mL	26.9 (22.1)
Mean bone alkaline phosphatase (SD), μg/L	19.7 (12.2)
Mean vitamin D (SD), ng/mL	15.2 (13.2)

AIDS, acquired immunodeficiency syndrome; DXA, dual-energy X-ray absorptiometry; HAART, highly active antiretroviral therapy; HCV, hepatitis C virus; HIV, human immunodeficiency virus; PIs, protease inhibitors; SD, standard deviation.

Bone mass density measurements

As shown in Table 2, either femoral or lumbar Z-scores ≤ −2.0 revealed low BMD for chronological age in 19.6% of the patients, whereas either femoral or lumbar T-scores revealed OP in 49.7% of the patients. A frank osteoporotic score (T-score ≤ −2.5) was observed in 13.5% of the sample, and a T-score < −1 at any site, including both osteopenic and osteoporotic patients, was recorded in 63.2% of all patients (46.3% at femoral and 55.8% at spinal sites).

Table 2.

Prevalence of Osteopenia and Osteoporosis in the Total Sample (n=163)

Site	Osteopenia criterion 1 (Z-score ≤ −2) %	Osteoporosis (T-score ≤ −2.5) %	Osteopenia criterion 2 (−2.5<T-score<−1) %	Osteopenia or osteoporosis (T-score<−1) %
Either femoral or lumbar	19.6	13.5	49.7	63.2
Femoral	9.3	5.6	40.7	46.3
Lumbar	16.6	11.0	44.8	55.8

Based on the assays indicated in the methods, no cases of secondary osteoporosis were diagnosed. Furthermore, serum 25-OH D3 levels were found to be low or very low in as many as 61.2% of tested patients; replacement therapy was instituted in all such cases and the effects of such a strategy will be re-evaluated.

In accordance with current Italian HIV guidelines for BMD screening, DXA would have been recommended only to 81.2% of patients with an osteopenic Z-score and to as few as 68.9% of patients with an osteopenic T-score at any site (Table 3).According with such proportions, the sensitivity of current Italian guidelines proved relatively low: 32.1% or 26.7% according to Z- or T-scores, respectively (Table 3). Notably, adding HCV co-infection among the indications for prescribing DXA, only two additional patients with an osteopenic T-score would have been caught; none considering Z-scores.

Table 3.

Comparison Between Italian Guidelines Indication to Perform DXA and Result Universal BMD Screening (163 Patients)

	Indication to screen for osteopenia ^a % (n)	No indication to screen for osteopenia % (n)	Total % (n)	Guidelines criteria specificity % (95% CI)	Guidelines criteria sensitivity % (95% CI)
Diagnosis based upon Z-score ^b				81.2 (63.0–92.1)	32.1 (24.3–40.9)
Osteopenia at any site	81.2 (26)	18.8 (6)	19.6 (32)
No osteopenia	67.9 (89)	32.1 (42)	80.4 (131)
Total	70.6 (115)	29.4 (48)	100 (163)
Diagnosis based upon T-score ^c				68.9 (58.9–77.5)	26.7 (16.5–39.9)
Osteopenia at any site	68.9 (71)	31.1 (32)	63.2 (103)
No osteopenia	73.3 (44)	26.7 (16)	36.8 (60)
Total	70.5 (115)	29.5 (48)	100 (163)

According to current Italian guidelines.^10,11 ^b Z-score ≤ −2, ^c T-score between −2.5 and −1.

BMD, bone mineral density; DXA, dual-energy X-ray absorptiometry.

Predictors of osteopenia

The multivariate linear regression model, adjusted for age, gender, duration of HIV infection, and BMI, showed an independent association between lower ZSF and lower BMI, AIDS diagnosis, HCV co-infection, treatment with HAART and NTBFs (Table 4). Regression for TSF yielded overlapping results (data not shown). As to lumbar scores, regression models for ZSL revealed the additional association of lower ZSL with age and CD4 T-cell Nadirs. Regression models for TSL confirmed the same predictors with the exception of age (data not shown). According to a dichotomic partition of Z-scores, the same factors were associated with a higher likelihood of OP at any site in multivariate logistic models (Table 5); similar results were found for dichotomic partitions of T-scores (data not shown).

Table 4.

Relationship Between Selected Variables and Femoral or Lumbar Z-Scores ^a

	Femoral Z-score				Lumbar Z-score
Categorical variables	Mean	p	Adjusted regression coefficient (95% CI)	p	Mean	p	Adjusted regression coefficient (95% CI)	p
Gender
Female	−0.36		(Ref.)		−0.36		(Ref.)
Male	−0.69	0.045	−0.22 (−0.5; 0.05)	0.1	−0.89	0.02	−0.4 (−0.75; 0.03)	0.1
AIDS
No	−0.43		(Ref.)		−0.59
Yes	−0.89	0.006	−0.44 (−0.7; −0.17)	0.001	−1.0	0.05
HCV co-infection
No	−0.39		(Ref.)		−0.53		(Ref.)
Yes	−1.31	<0.001	−0.55 (−0.9; −0.2)	0.002	−1.42	<0.001	−0.4 (−0.86; 0.12)	0.1
Lipodystrophy
No	−0.48				−0.65
Yes	−0.74	0.1			−0.84	0.3
Co-morbidities
No	−0.50				−0.60
Yes	−0.63	0.4			−0.78	0.4
HAART
No	−0.14		(Ref.)		−0.46
Yes	−0.71	0.003	−0.38 (−0.72; −0.34)	0.031	−0.80	0.2
TDF-including regimens
No	−0.37				−0.51
Yes	−0.76	0.01			−0.90	0.06
PIs-including regimens
No	−0.38				−0.51
Yes	−0.83	0.004			−0.97	0.02
EFV-including regimens
No	−0.58				−0.73
Yes	−0.63	0.8			−0.74	0.9
TDF-PIs-including regimens
No	−0.50				−0.60
Yes	−0.76	0.1			−0.97	0.08
Other drugs
No	−0.55				−0.84
Yes	−0.62	0.7			−0.66	0.4
Nontraumatic fracture
No	−0.37		(Ref.)		−0.47		(Ref.)
Yes	−1.20	<0.001	−0.59 (−0.88; −0.29)	<0.001	−1.44	<0.001	−0.81 (−1.2; −0.4)	<0.001
Cigarette smoking
No	−0.38				−0.54
Yes	−0.81	0.006			−0.92	0.06
Alcohol abuse
No	−0.44				−0.74
Yes	−0.61	0.6			−0.57	0.6
Drug addiction
No	−0.40				−0.57
Yes	−1.22	<0.001			−1.28	0.004
Gainful employment
No	−0.53				−0.84
Yes	−0.73	0.3			−0.68	0.5
Physical activity
No	−0.55				−0.67
Yes	−0.65	0.5			−0.81	0.5

Continuous variables	Spearman Rho	p	Adjusted regression coefficient (95% CI)	p	Spearman Rho	p	Adjusted regression coefficient (95% CI)	p
Age, 1-year increase	0.06	0.4	0.005 (−0.01; 0.01)	0.7	0.19	0.01	0.04 (0.02; 0.05)	<0.001
Body mass index, 1-unit increase	0.42	<0.001	0.1 (0.07; 0.13)	<0.001	0.29	<0.001	0.07 (0.02; 0.01)	0.04
Duration of HIV infection, 1-year increase	−0.21	0.006	0.02 (−0.007; 0.04)	0.2	−0.09	0.2	−0.01 (−0.04; 0.02)	0.7
Nadir CD4 T-cell, 10-cell/mmc increase	0.17	0.012			0.17	0.03	0.01 (0.01; 0.02)	0.004
Current CD4 T-cell, 10-cell/mmc increase	0.08	0.3			0.23	0.003
Duration of HAART^#, 1-year increase	−0.23	0.003			−0.08	0.3
Serum level vitamin D, 1-ng/ml increase	0.02	0.8			−0.05	0.5

Indices of low BMD for chronological age.

None of the variables that have not been included in the linear regression models were significant in multivariate analysis.

AIDS, acquired Immunodeficiency syndrome diagnosis; HAART, highly active antiretroviral therapy; HCV, hepatitis C virus; HIV, human immunodeficiency virus.

Table 5.

Relationship Between Selected Variables and Femoral or Lumbar Osteopenia (Defined as Z-score ≤ −2.0)

	Femoral/lumbar osteopenia		Logistic regression		Femoral osteopenia		Logistic regression		Lumbar osteopenia		Logistic regression
Categorical variables	%	p	Adj. OR (95% CI)	p	%	p	Adj. OR (95% CI)	p	%	p	Adj. OR (95% CI)	p
Gender
Female	14.6		1 (Ref.)		8.3		1 (Ref.)		10.4		1 (Ref.)
Male	21.7	0.3	1.4 (0.5–4.2)	0.5	9.6	0.8	1.0 (0.2–4.7)	0.9	19.1	0.2	1.9 (0.6–6.1)	0.3
AIDS
No	14.2		1 (Ref.)		6.7				12.3		1 (Ref.)
Yes	28.8	0.02	3.1 (1.2–8.0)	0.02	14.0	0.1			24.6	0.04	2.4 (0.9–6.4)	0.07
HCV co-infection
No	12.7				4.0		1 (Ref.)		11.9
Yes	43.2	<0.001	2.7 (0.9–8.7)	0.06	27.8	<0.001	11.2 (2.0–63)	0.006	32.4	0.003
Lipodystrophy
No	17.2				8.6				14.0
Yes	22.9	0.4			10.1	0.7			20.0	0.3
Co-morbidities
No	12.0				6.0				12.0
Yes	23.6	0.08			11.0	0.3			19.1	0.3
HAART
No	9.1				6.1				9.1
Yes	22.3	0.08			10.1	0.5			18.5	0.2
TDF-including regimens
No	14.1				8.5				12.7
Yes	24.0	0.1			9.9	0.8			19.6	0.2
PIs-including regimens
No	14.0				7.0				11.6
Yes	26.0	0.05			12.0	0.2			22.1	0.07
EFV-including regimens
No	19.5				10.2				16.7
Yes	20.0	0.9			4.0	0.3			16.0	0.9
TDF-PIs-including regimens
No	16.9				9.4				14.1
Yes	24.6	0.2			8.9	0.9			21.1	0.2
Other drugs
No	23.4				7.8				21.9
Yes	17.7	0.4			10.5	0.6			13.5	0.2
Non-traumatic fracture
No	11.8		1 (Ref.)		4.2		1 (Ref.)		10.1		1 (Ref.)
Yes	40.9	<0.001	4.6 (1.7–12)	0.002	23.3	<0.001	4.2 (1.4–15)	0.03	34.1	<0.001	5.0 (1.9–13)	0.001
Cigarette smoking
No	14.6				4.9				14.6
Yes	24.7	0.1			13.8	0.05			18.5	0.5
Alcohol abuse
No	20.1				8.7				16.8
Yes	14.3	0.6			15.4	0.4			14.3	0.8
Drug addiction
No	13.5				4.8				13.5
Yes	40.5	<0.001			24.3	<0.001			27.0	0.05
Gainful employment
No	27.5				16.0				21.6
Yes	16.1	0.09			6.3	0.05			14.3	0.2
Physical activity
No	19.8				9.0				16.8
Yes	19.7	0.9			9.8	0.8			16.4	0.9

Continuous variables
Age, 1-year increase	–	1.0 (0.9–1.0)	0.5	–	1.0 (0.9–1.0)	0.2	–	1.0 (0.9–1.0)	0.2
Body mass index, 1-unit increase	–	0.8 (0.7–0.9)	0.02	–	1.0 (0.8–1.1)	0.6	–	0.8 (0.7–0.9)	0.02
Duration of HIV infection, 1-year increase	–	1.0 (0.9–1.0)	0.9	–	1.0 (0.9–1.1)	0.4	–	1.0 (1.0–1.1)	0.2
Nadir CD4 T-cell, 10-cell/mmc increase	–		NI	–		NI	–		NI
Current CD4 T-cell, 10-cell/mmc increase	–		NI	–		NI	–		NI
Duration of HAART, 1-year increase	–		NI	–		NI	–		NI
Serum level vitamin D, 1 ng/ml increase	–		NI	–		NI	–		NI

AIDS, acquired immunodeficiency syndrome diagnosis; HAART, highly active antiretroviral therapy; HCV, hepatitis C virus; HIV, human immunodeficiency virus; NI, not included; none of the variables that have not been included in the logistic regression models were significant in multivariate analysis; OR, odds ratio.

Fractures

Data about NTBFs in adulthood were available for 184 of the 186 patients initially consenting to the study. Prevalence of NTBFs in the total sample was 27.0% (49 cases). Wrist was the most frequent site, as it was reported by 26 (53.1%) patients, followed by vertebral fractures (10, 20.4%) and femoral fractures (4, 8.2%). Other less frequently reported sites were hand (3 cases), shoulder (2 cases), elbow (2 cases), and ankle (2 cases). At multivariate analyses, the only significant predictors of NTBFs were male gender, HCV coinfection and lower ZSF (Table 6).

Table 6.

Predictors of Nontraumatic Bone Fractures

Categorical variables	Nontraumatic fracture In the total sample %	p	Logistic regression Adj. OR (95% CI)	p
Gender
Female	9.8		1 (Ref.)
Male	33.1	0.001	4.4 (1.4–14.5)	0.014
AIDS
No	23.5
Yes	32.0	0.2
HCV co-infection
No	20.4
Yes	47.6	<0.001	3.9 (1.2–12.2)	0.021
Lipodystrophy
No	22.1
Yes	32.1	0.1
Co-morbidities
No	17.3
Yes	31.2	0.06
HAART
No	20.0
Yes	28.2	0.3
Other drugs
No	23.1
Yes	30.4	0.3
Cigarette smoking
No	20.0
Yes	33.0	0.04
Alcohol abuse
No	25.0
Yes	39.0	0.2
Drug addiction
No	22.4
Yes	41.5	0.02
Gainful employment
No	36.4
Yes	21.4	0.04
Physical activity
No	27.0
Yes	22.2	0.6

Continuous variables
Z-score (femur), 1-unit increase	–	0.33 (0.19–0.59)	<0.001
Z-score (lumbar), 1-unit increase	–	NI ^a
Age, 1-year increase	–	1.0 (0.9–1.0)	0.3
Body mass index, 1-unit increase	–	1.1 (0.9–1.2)	0.1
Duration of HIV infection, 1-year increase	–	NI
Nadir CD4 T-cell, 10-cell/mmc increase	–	NI
Current CD4 T-cell, 10-cell/mmc increase	–	NI
Duration of HAART, 1-year increase	–	NI
Serum level vitamin D, 1-ng/ml increase	–	NI

Discussion

Our investigation was aimed at estimating the prevalence and predictors of OP in HIV-infected patients at our site, as well as the sensitivity and specificity of current guidelines to identify patients requiring a BMD screening.

Our findings provide further evidence that potentially relevant reductions in BMD are very frequent in HIV-infected patients.^1

–6 Indeed, 19.6% of patients in our sample had either a ZSF or ZSL ≤ −2.0, a rate in the same range as in many other reports, whereby DXA screening was performed for at risk patients only.^16,24
–26 When we considered milder levels of BMD reduction, 63.2% of patients had a femoral or lumbar T-score < −1.0 SD. Remarkably, the mean age of our sample was 44.2±10.0 years and males were 70.1%. In the ESOPO study, HIV-negative Italian women aged 40–79 years were osteopenic or osteoporotic (T-score < −1 SD) in 44.7% of investigated cases, and HIV-negative Italian men aged 60–79 only in 36% of cases.²⁷ None of the other recent datasets of HIV infected patients reporting a high prevalence of OP and progression to lower BMD was directly comparable to ours, as they did not share the same inclusion criteria and/or study design, none of them providing universal screening of BMD as the only criterion.^{16,21,24
–26,28,29} Bonjoch et al.,²⁶ for instance, found 47.5% of osteopenic and 23% of frankly osteoporotic patients among 671 selected HIV patients at increased risk of BMD loss, investigated by retrospective retrieval of DXA scan data. In a longitudinal assessment of BMD in HIV-infected women compared with uninfected controls, Dolan et al.¹⁶ found lower levels of BMD among HIV-infected patients and higher levels of osteocalcin and urine N-telopeptide of type 1 collagen, an observation confirmed at follow-up. In a cross-sectional study, Bruera et al.²¹ compared BMD in a sample of 142 subject including four groups: antiretroviral-naïve HIV patients, HIV patients on HAART without PIs, HIV patients on a PI-based HAART, and a control groups including healthy subjects.²¹ BMD was lower in all groups of HIV patients in comparison with healthy controls; differences among HIV groups were not statistically significant.²¹ It is of relevance, however, that most such studies reported prevalences of BMD disorders in the same range as ours, despite considering only selected populations at increased risk: in particular, reported prevalences ranged between 21% and 47.5% for OP and between 3% and 23% for OT.^{16,21,24,28,29} Our inclusion of all consenting patients, many of whom at low risk for BMD disorders in accordance to current guidelines, yielded similarly high proportions of low BMD. Therefore, the lack of a “diluting effect” in our sample suggests that a significant proportion of HIV patients without a validated indication to DXA may yet present acknowledgeable levels of BMD reduction.

Similar prevalences of low BMD for chronological age were found (19.8% and 21.7%, respectively), considering two subsets in the final sample: 116 patients under 50 years and 115 male patients; furthermore, multivariate analyses including only the 116 younger patients or the 115 male patients yielded identical results (data not shown). These additional findings once more suggest that in subjects with HIV infection BMD may be lower due to mechanisms relatively independent of age and gender as in the general population.

We therefore tried to evaluate the sensitivity and specificity of current HIV management guidelines in diagnosing patients with OP: in accordance with current Italian criteria,^10,11 18.8% and 31.1% of our sample would not have been included in the screening process for BMD, considering osteopenic Z-scores and T-scores, respectively. Although other guidelines include slightly different lists of risk factors for BMD loss, similar results would have been obtained following EACS or IDSA guidelines.^30,31 These findings highlight a potential lack of sensitivity of current HIV guidelines for early identification and prevention of BMD disorders in this setting, HIV infection emerging as a cause of OP per se.^10,11,30,31 Interestingly, as overt OT is a condition impacting on overall mortality both in the general and special populations,^32,33 current guidelines for OP/OT recommend screening in the general population when at list one major or two minor risk factors for primary OT are present, and whenever OT may be secondary to other conditions.^34,35 It would be therefore reasonable, at this stage, to acknowledge OP as secondary to HIV infection per se, as well as to all co-morbidities in its wake.

Our investigation provided other possibly relevant data to the management of BMD disorders in HIV-infected patients. First, HCV co-infected patients had higher levels of BMD loss in all our univariate and multivariate analyses. This is in line with several recent reports,^7
–9,36 indicating a prevalence of OP as high as 28% in HIV-HCV co-infected patients with controlled HIV replication³⁶ and lower Z-scores in co-infected than in mono-infected women.⁷ In HIV/HCV co-infected patients, HCV may speed BMD loss through multiple mechanisms, so that an amplification of the effects of HIV on bone metabolism and BMD may well ensue in such patients.⁷

Second, a high percentage (27.0%) of patients reported at least one fragility fracture, correlated with male gender, HCV co-infection and lower femoral Z-scores, as in other reports.^8,9,33 Several lines of evidence underline the relationship between BMD reduction and a higher risk of fragility fractures in elderly people in the general population.^37,38 However, the comparative prevalence of fractures was 2.87% and 1.77% in HIV-positive versus HIV-negative patients in the US Healthcare Registry study, with a significantly higher prevalence of vertebral, hip, and wrist fractures in HIV-infected patients.⁵ Two additional recent reports confirmed this observation, yielding similar results.^7,39 Furthermore, both studies suggested that the relative risk of low-energy fractures was higher in HIV/HCV co-infected patients.^7,39 In line with a recent report documenting a prevalence of vertebral fractures as high as 23.3% among asymptomatic HIV-infected patients using X-ray evaluation as a screening procedure,⁴⁰ our data may yet represent an underestimate of the true prevalence of NTBFs, most of them being asymptomatic.⁴¹

Third, lower levels of ZSF were consistently found in patients treated with any HAART regimen in comparison with untreated patients, no association being found for any specific drug or drug regimen, possibly due to the small size of our sample and the even smaller proportion of untreated patients. Notably, several studies investigated the role of single drugs or specific drug combinations in ensuing BMD disorders in HIV-infected patients.^{21,28,29,42

–45} In a sub-study of the SMART trial,⁴⁵ BMD reductions were higher in patients in the group under continuous HAART than in those undergoing HAART discontinuations. Evidence collected with different experimental designs suggested that TDF, protease inhibitors, and EFV were more frequently related with BMD reductions in HIV treated patients.^3,6,26,46 Some reports documented higher levels of markers of bone turnover in HIV-infected patients on HAART, possibly related to increased osteoclast activation in ritonavir boosted PIs regimens;^47
–49 other reports, however, failed to establish such a clear-cut association between BMD reductions and specific HAART regimens.^42,44

Finally, BMI was tightly related with BMD in our series, in line with several other reports, suggesting that low BMIs should be considered a priori in HIV patients as a possible warning for BMD disorders.^{6,21,37,42,50
–52} In addition, the fact that some variables were associated with lower BMD at the lumbar level, whereas others with femoral BMD only, may be related to a differential impact of different factors on trabecular vs. cortical bones, as already reported in the literature;⁵³ these aspects may be the object of further investigations.

Our baseline study may be limited by a major bias: we do not know what the peak of bone mass density was in our patients in their 20s. Peak of bone mass density at 20s is, absolutely, the most important predictor of later bone loss, surely more relevant than viral or HAART-related damage: a patient with a high peak density will rarely suffer an osteoporotic fracture, even if on an osteotoxic regimen; conversely, a patient with a low peak BMD will risk an osteoporotic fracture even if in therapy with a friendly regimen and free of other risk factors.^3,24,54 In our multivariate model we could adjust for many variables, but not the most important at all. It is likely, however, that different peak densities may be evenly distributed among different risk groups in our sample, due to its relatively large size and unselected nature.

Considering the recommendation of current Italian HIV guidelines^10,11 and NICE guidelines for Osteoporosis⁵⁵ to assess BMD and fragility fracture risk at least 18 months after their first evaluation, all participants will be offered re-evaluation 36 months after enrollment. We plan to re-evaluate all the involved factors already considered at baseline, and in particular: HAART therapy and other drugs in addition to HAART (including vitamin D replacement); clinical, laboratory, radiological, and behavioral aspects; and BMD measurements by DXA. The time span and the complete re-assessment should be enough to identify new upcoming fractures.

In conclusion, HIV-infected patients were once more found at increased risk for OP and fragility fractures. Our results suggest that measuring BMD in HIV patients regardless of any further specification may help retrieving one-fifth to one-third of patients with BMD disorders not identified using current criteria for selective screening in the setting of HIV infection. Fragility fractures were frequent and correlated with male gender, HCV co-infection, and lower femoral Z-scores, suggesting once more that these findings may have clinical significance for long-term management of HIV infection.

Footnotes

Acknowledgments

EM, TU, EP, FV, and FT were partly funded by grants by the Fondazione onlus Camillo De Lellis Per l'Innovazione e la Ricerca in Medicina, Pescara, Italy. We are heavily indebted with Mrs. Loredana Puglielli, Mrs. Angela Pisciella, and Mrs. Teresa Moschiano for their collaboration and their professional ability to deal with patients. The authors also thank Mr. Galileo Camplone, who acted very collaboratively as specialized technician in Medicine Nuclear Unit.

Author Disclosure Statement

The authors declare no conflict of interest.

References

Mondy

, Tebas

. Emerging bone problems in patients infected with human immunodeficiency virus. Clin Infect Dis, 2003; 36:S101–S105.

Amorosa

, Tebas

. Bone disease and HIV infection. Clin Infect Dis, 2006; 42:108–114.

McComsey

, Tebas

, Shane

, et al. Bone disease in HIV infection: A practical review and recommendations for HIV care providers. Clin Iinfect Dis, 2010; 51:937–946.

Brown

, Qaqish

. Antiretroviral therapy and the prevalence of osteopenia and osteoporosis: A meta-analytic review. AIDS, 2006; 20:2165–2174.

Mallon

PWG

. HIV and bone mineral density. Curr Opin Infect Dis, 2010; 23:1–8.

Borderi

, Gibellini

, Vescini

, et al. Metabolic bone disease in HIV infection. AIDS, 2009; 23:1297–1310.

Lo Re

3rd , Guaraldi

, Leonard

, et al. Viral hepatitis is associated with reduced bone mineral density in HIV-infected women but not men. AIDS, 2009; 23:2191–2198.

Hansen

, Gerstoft

, Kronborg

, et al. Incidence of low and high-energy fractures in persons with and without HIV infection: A Danish population-based cohort study. AIDS, 2012; 26:285–293.

Walker Harris

, Sutcliffe

, Araujo

, et al. Hip bone geometry in HIV/HCV-co-infected men and healthy controls. Osteoporos Int, 2011; 23:1779–1787.

10.

Antinori

, Marcotullio

, Ammassari

, et al. Italian guidelines for the use of antiretroviral agents and the diagnostic-clinical management of HIV-1 infected persons. New Microbiol, 2011; 34:109–146.

11.

http://www.salute.gov.it/imgs/C_17_pubblicazioni_2074_allegato.pdf (Last accessed on July 10, 2014 ).

12.

Vescini

, Di Campli

, Borderi

, and Italian SorBone workgroup. Preliminary result of an Italian survey on the diagnosis of osteoporosis in HIV. IAS Rome, 2011 http://pag.ias2011.org/EPosterHandler.axd?aid=2159 (Last accessed on April 14, 2013 ).

13.

Center

, Bliuc

, Nguyen

, Eisman

. Risk of subsequent fracture after low-trauma fracture in men and women. JAMA, 2007; 297:387–394.

14.

Carr

. HIV lipodystrophy: Risk factors, pathogenesis, diagnosis and management. AIDS, 2003; 17:S141–S148.

15.

O'Brien

, Asmar

, Beilin

, et al. European Society of Hypertension Working Group on Blood Pressure Monitoring. Practice guidelines of the European Society of Hypertension for clinic, ambulatory and self blood pressure measurement. J Hypertens, 2005; 23:697–701.

16.

Dolan

, Kanter

, Grinspoon

. Longitudinal analysis of bone density in human immunodeficiency virus-infected women. J Clin Endocrinol Metab, 2006; 91:2938–2945.

17.

Brockmeyer

, Kreuter

, Bader

, et al. Prevalence of endocrine dysfunction in HIV-infected men. Horm Res, 2000; 54:294–295.

18.

http://www.hologic.com/it/osteoporosis-assessment/skeletal/overview (Last accessed on July 10, 2014 ).

19.

Binkley

, Bilezikian

, Kendler

, et al. Official positions of the International Society for Clinical Densitometry and Executive Summary of the 2005 Position Development Conference. J Clin Densitom, 2006; 9:4–14.

20.

Foundation NO. Clinician's guide to prevention and treatment of osteoporosis. Washington, DC: National Osteoporosis Foundation (2010) http://www.nof.org/files/nof/public/content/file/344/upload/159.pdf (Last accessed on July 10, 2014 ).

21.

Bruera

, Luna

, David

, et al. Decreased bone mineral density in HIV-infected patients is independent of antiretroviral therapy. AIDS, 2003; 17:1917–1923.

22.

Hamilton

. Statistics with Stata (Updated for version 9). Belmont, CA: Thomson–Brooks/Cole, 2006.

23.

Newcombe

. Two-sided confidence intervals for the single proportion: Comparison of seven methods. Stat Med, 1998; 17:857–872.

24.

Arnsten

, Freeman

, Howard

, et al. HIV infection and bone mineral density in middle-aged women. Clin Infect Dis, 2006; 42:1014–1020.

25.

Arnsten

, Freeman

, Howard

, et al. Decreased bone mineral density and increased fracture risk in aging men with or at risk for HIV infection. AIDS, 2007; 21:617–623.

26.

Bonjoch

, Figueras

, Estany

, et al. High prevalence of and progression to low bone mineral density in HIV-infected patients: A longitudinal cohort study. AIDS, 2010; 24:2827–2833.

27.

Maggi

, Noale

, Giannini

, et al. Quantitative heel ultrasound in a population-based study in Italy and its relationship with fracture history: The ESOPO study. Osteoporos Int, 2006; 17:237–244.

28.

Carr

, Miller

, Eisman

, Cooper

. Osteopenia in HIV-infected men: Association with asymptomatic lactic acidemia and lower weight pre-antiretroviral therapy. AIDS, 2001; 15:703–709.

29.

Bolland

, Grey

, Horne

, et al. Bone mineral density is not reduced in HIV-infected Caucasian men treated with highly active antiretroviral therapy. Clin Endocrinol, 2006; 65:191–197.

30.

http://www.eacsociety.org/Portals/0/140601_EACS%20EN7.02.pdf (Last accessed on July 10, 2014 ).

31.

Aberg

, Gallant

, Ghanem

, et al. Primary care guidelines for the management of persons infected with HIV: 2013 update by the HIV medicine Association of the Infectious Diseases Society of America. Clin Infect Dis, 2014; 58:1–10.

32.

Eisman

, Bogoch

, Dell

, et al. Making the first fracture the last fracture: ASBMR task force report on secondary fracture prevention. J Bone Miner Res, 2012; 27:2039–2046.

33.

Hasse

, Ledergerber

, Furrer

, et al. Morbidity and aging in HIV-infected persons: the Swiss HIV cohort study. Clin Infect Dis, 2011; 53:1130–1139.

34.

U.S. Preventive Services Task Force. Screening for osteoporosis: U.S. preventive services task force recommendation statement. Ann Intern Med, 2011; 154:356–364.

35.

Lim

, Hoeksema

, Sherin

. Screening for osteoporosis in the adult U.S. population: ACPM position statement on preventive practice. Am J Prev Med, 2009; 36:366–375.

36.

El-Maouche

, Mehta

, Sutcliffe

, et al. Controlled HIV viral replication, not liver disease severity associated with low bone mineral density in HIV/HCV co-infection. J Hepatol, 2011; 55:770–776.

37.

Coin

, Sergi

, Benincà

, et al. Bone mineral density and body composition in underweight and normal elderly subjects. Osteoporos Int, 2000; 11:1043–1050.

38.

Piscitelli

, Tarantino

, Chitano

, et al. Updated incidence rates of fragility fractures in Italy: extension study 2002–2008. Clin Cases Miner Bone Metab, 2011; 8:54–61.

39.

Collin

, Duval

, Le Moing

, et al. Ten-year incidence and risk factors of bone fractures in a cohort of treated HIV1-infected adults. AIDS, 2009; 23:1021–1024.

40.

Borderi

, Calza

, Colangeli

, et al. Prevalence of subclinical vertebral fractures in HIV-infected patients. New Microbiol, 2014; 37:25–32.

41.

Grigoryan

, Guermazi

, Roemer

, et al. Recognizing and reporting osteoporotic vertebral fractures. Eur Spine J, 2003; 12:S104–S112.

42.

Mondy

, Yarasheski

, Powderly

, et al. Longitudinal evolution of bone mineral density and bone markers in human immunodeficiency virus-infected individuals. Clin Infect Dis, 2003; 36:482–490.

43.

Fux

, Rauch

, Simcock

, et al. Tenofovir use is associated with an increase in serum alkaline phosphatase in the Swiss HIV Cohort Study. Antivir Ther, 2008; 13:1077–1082.

44.

Bolland

, Wang

, Grey

, et al. Stable bone density in HAART-treated individuals with HIV: A meta-analysis. J Clin Endocrinol Metab, 2011; 96:2721–2731.

45.

Grund

, Peng

, Gibert

, et al. Continuous antiretroviral therapy decreases bone mineral density. AIDS, 2009; 23:1519–1529.

46.

Gibellini

, Borderi

, de Crignis

, et al. Analysis of the effects of specific protease inhibitors on OPG/RANKL regulation in an osteoblast-like cell line. New Microbiol, 2010; 33:109–115.

47.

Yin

, McMahon

, Ferris

, et al. Low bone mass and high bone turnover in postmenopausal human immunodeficiency virus-infected women. J Clin Endocrinol Metab, 2010; 95:620–629.

48.

Yin

, Modarresi

, Shane

, et al. Effects of HIV infection and antiretroviral therapy with ritonavir on induction of osteoclast-like cells in postmenopausal women. Osteoporos Int, 2011; 22:1459–1468.

49.

Santiago

, Oguma

, Brown

, Laurence

. Noncanonical Wnt signaling promotes osteoclast differentiation and is facilitated by the human immunodeficiency virus protease inhibitor ritonavir. Biochem Biophys Res Commun, 2012; 417:223–230.

50.

Greco

, Fornari

, Rossi

, et al. Is obesity protective for osteoporosis? Evaluation of bone mineral density in individuals with high body mass index. Int J Clin Pract, 2010; 64:817–820.

51.

Saarelainen

, Kiviniemi

, Kröger

, et al. Body mass index and bone loss among postmenopausal women: The 10-year follow-up of the OSTPRE cohort. J Bone Miner Metab, 2012; 30:208–216.

52.

Bolland

, Grey

, Gamble

, Reid

. Low body weight mediates the relationship between HIV infection and low bone mineral density: A meta-analysis. J Clin Endocrinol Metab, 2007; 92:4522–4528.

53.

Crilly

, Cox

. A comparison of bone density and bone morphology between patients presenting with hip fractures, spinal fractures or a combination of the two. BMC Musculoskelet Disord, 2013; 14:68.

54.

Mulligan

, Harris

, Emmanuel

, et al. Low bone mass in behaviorally HIV-infected young men on antiretroviral therapy: Adolescent Trials Network Study 021B. Clin Infect Dis, 2012; 55:461–468.

55.

National Clinical Guideline Centre (UK). Osteoporosis: Fragility Fracture Risk: Osteoporosis: Assessing the Risk of Fragility Fracture. (2012). http://www.nice.org.uk/guidance/cg146/resources/cg146-osteoporosis-fragility-fracture-full-guideline3 (Last accessed on November 28, 2014 ).