iconstar paper   HIV Articles  
Back grey arrow rt.gif
Bone Density, Microarchitecture, and Tissue Quality After Long-Term Treatment With Tenofovir DF/Emtricitabine or Abacavir/Lamivudine
  Download the PDF here
With Tenofovir/Emtricitabine or Abacavir/Lamivudine
JAIDS July 12017 - Robert Guerri-Fernandez, MD, PhD,* Daniel Molina-Morant, MD,* Judit Villar-Garcia, MD,* Sabina Herrera, MD,* Alicia Gonzalez-Mena, MD,* Ana Guelar, MD,* Marta Trenchs-Rodriguez, MD, Adolfo Diez-Perez, MD, PhD,* and Hernando Knobel, MD, PhD* From the *Hospital del Mar Research Institute, Infectious Diseases Department, Barcelona, Spain; Departament de Medicina, Universitat Autonoma de Barcelona, Spain; and Catalan Institute of Health, ABS Passeig de Sant Joan, Barcelona, Spain.
Objectives: HIV infection has been associated with reduced bone mineral density (BMD). Antiretroviral therapy (ART) has a deleterious effect on BMD, but its effect on bone fragility is not clear. The objective of this study is to analyze the BMD, microarchitecture, and tissue quality of bone in patients receiving long-term tenofovir- or abacavir-based ART.
Design: We conducted a cross-sectional study in patients with HIV undergoing tenofovir or abacavir ART for more than 5 years.
Methods: We measured BMD using dual X-ray absorptiometry ,bone michroarchitecture using trabecular bone score (TBS), and bone tissue quality using microindentation. TBS is a dual X-ray absorptiometry–based software that is more highly correlated with bone fragility than BMD. Microindentation (BMSi) directly assesses bone quality at the tissue level. Results: A total of 63 patients were included in this study, with 36 belonging to the TDF-FTC group and 27 to the ABC-3TC group. Patients receiving TDF-FTC treatment showed lower BMD values than those in the ABC-3TC group. We found no differences in TBS or microindentation between the 2 groups. However, after adjusting for sex, age, body mass index, and 25[OH]vitD we found lower BMSi and thus poorer bone properties in the TDF-FTC group than in the ABC-3TC group [beta coefficient -3.594 (confidence interval: 95% -0.12 to -7.61); P = 0.043].
Conclusions: Long-term treatment with TDF-FTC leads to impaired bone health, not only in terms of BMD but also in terms of bone quality, another determinant of overall bone strength. To complement BMD-based predictions, these other techniques may also be used to identify patients with excess fracture risk.
Antiretroviral therapy (ART) and HIV infection have been associated with reduced bone mineral density (BMD) as measured by dual X-ray absorptiometry (DXA). Despite the potentially toxic effects of ART on bones, bone disease in HIV-treated patients is not well characterized. More specifically, the antiretroviral drug tenofovir disoproxil fumarate (TDF) has been associated with reduced BMD and higher levels of bone turnover markers.1 Several mineral alterations can already be detected soon after ART initiation, and as treatment continues, there is a progressive decrease in BMD and an increase in bone turnover markers. In fact, these changes can even persist for much longer: in longitudinal follow-up studies of up to 96 weeks, ART-treated patients continued to exhibit progressive BMD loss.2 However, no clear association between fracture risk and antiretroviral drug use has been discovered. Nonetheless, the clinical impact of bone disease among patients undergoing long-term ART is an important issue. In patients with HIV, aside from ART, a combination of several other factors such as harmful lifestyles and altered hormone levels may also have an impact on bone health.
Bone strength is an integration of both bone density and bone quality.3 As such, measuring DXA only provides limited information about overall bone health. Unfortunately, in many situations, this DXA-measured bone density is not sufficient to thoroughly assess fracture risk.4–6A more comprehensive assessment including other key aspects of bone strength, such as bone microarchitecture and mechanical properties (tissue quality), is needed to better evaluate the propensity of bone to fracture. This is especially true in situations where BMD does not fully account for the decrease in bone strength. To assess bone microarchitecture, new DXA-based software has been developed. This enables a trabecular bone score (TBS), an index of bone microarchitecture, to be derived from lumbar spine DXA images.7 The TBS is a BMD-independent predictor of fracture risk. A recent meta-analysis found that the TBS can significantly predict fracture risk independently of the FRAX diagnostic tool.8 With respect to the quality component of bone strength, microindentation techniques can provide direct assessments of the mechanical properties of bone at the tissue level.5,6,9 Through single-impact microindentation, a direct measurement of the resistance of cortical bone tissue to the appearance of microcracks is obtained. This phenomenon closely mimics the initial crack of a fracture9,10 and is correlated with the propensity of bone to fracture.4 Furthermore, as measured by microindentation, patients with HIV show worse mechanical properties than noninfected individuals irrespective of BMD.11
The aim of this study was to determine the consequences of long-term tenofovir/emtricitabine (TDF-FTC) or abacavir/lamivudine (ABC-3TC) treatment on bone health. The effects of these 2 common ART regimens were analyzed by measuring BMD, microindentation, TBS-assessed bone microarchitecture, and bone turnover markers.

We conducted a cross-sectional study in the outpatient clinic of the Infectious Diseases Department at Hospital del Mar in Barcelona, Spain. The study was approved by the local Clinical Research Ethics Committee, and all participants provided written informed consent (Ethical Committee 2013/5250/I).
For each patient, we obtained information regarding general clinical history, specific HIV history, and past treatments. All patients in the study were HIV infected, older than 50 years, and had been subjected to ART (either TDF-FTC or ABC-3TC treatment) for more than 5 years. All patients displayed good treatment adherence, undetectable viral loads (<50 copies/mL), and CD4 T-lymphocyte counts of >250/mL. All individuals who had undergone treatment with bone-active drugs or who had a condition that could interfere with bone metabolism were excluded from the study. This included patients with liver disease, severe chronic active alcoholism (defined as >40 g/d), malignancy, Cushing syndrome, hypogonadism, hyperthyroidism, hypopituitarism, hyperparathyroidism, chronic kidney disease (defined as glomerular filtration rate <60 mL•min-1•1.73 m-2 for at least 3 months), and specifically we excluded patients who presented in any previous control abnormal creatinine levels, and also those cases where therapy was changed because of abnormal kidney function, chronic obstructive pulmonary disease, chronic hepatitis C or B, diabetes mellitus type 1 or type 2, or neuropathic disease. Furthermore, patients who used glucocorticoids or opioids, who were active intravenous drug consumers, or who had a history of switching between TDF-FTC and ABC-3TC treatments were also excluded.
Bone Measurements
For each patient, we recorded risk factors for bone fracture and performed spinal x-rays to detect vertebral fractures. X-rays were assessed by 2 independent observers. We accepted grade I or above (a loss of >20% of vertebral height) as a true fracture. We used chemiluminescent immunoassays (CLIAs) to determine several fasting routine bone-specific laboratory values. Each immunoassay had an interassay coefficient of variation of 10%. Specifically, we measured levels of intact parathormone (Siemens); bone alkaline phosphatase (Roche Diagnostics, Switzerland); the amino pro-peptide of type I collagen (PINP), cross-linked collagen type I, C-telopeptide (CTX), and serum 25-hydroxyvitamin D (Roche Diagnostics); bone-specific alkaline phosphatase (Roche Diagnostics) and high-sensitivity C-reactive protein [CLIA (Immulite 2000; Siemens)], erythrocyte sedimentation rate, beta-2 microglobulin (CLIA) (Immulite 2000; Siemens); D-dimer [immunoturbidimetry (ACL TOP300)] and fibrinogen [Clauss method (ACL TOP300)]. BMD was measured at the lumbar spine and the hip by DXA (DXA-Hologic QDR4500SL; S/N 45,329). The coefficient of variation for the DXA measurements was 1% in the spine and 1.7% for the femoral neck. The TBS, a gray-level textural metric that can be extracted from two-dimensional lumbar spine DXA images, is related to bone microarchitecture and provides skeletal information that is not captured in standard BMD values. TBS values are considered to be normal when above 1.35. Values between 1.20 and 1.35 imply a partially impaired microarchitecture, whereas values below 1.20 refer to a degraded microarchitecture. TBS values were evaluated in the same regions that were used for lumbar spine BMD measurements using iNsight v 2.1 (Med.Imaps, Merignac, France). Microindentation tests were performed using a handheld OsteoProbe Reference Point Indenter (Active-Life-Scientific, Santa Barbara, CA) according to a recently published protocol.12 Briefly, under local anesthesia, software provides the average value of eight 30 N indentations performed on the anterior face of the tibia using a test probe with a conic edge of 4 μm. This average is then normalized to the average of 5 indentations made on a polymethylmethacrylate block. The ratio between the bone and polymethilmethacrylate measurements provides a bone mineral strength index (BMSi) parameter.12 This technique has previously been validated in humans, specifically within the HIV-infected population.9,11 Importantly, this procedure takes less than 5 minutes, causes minimal discomfort to the patient, and no complications have been described in published studies. The coefficient of variation for microindentation in our laboratory is 3% (Fig. 1).
Statistical Analysis
Sample size number was calculated based on previous publications.11 Assuming a type I and II error of 0.05 in a 2-sided test, 27 HIV-infected patients treated with TDF-FTC and 27 treated with ABC-3TC were required to detect a statistically significant difference of ≥5 units of BMSi. The common SD was assumed to be 5.
We used 2-sample t tests and χ2 tests to compared quantitative and categorical variables, respectively, between HIV-infected individuals treated with TDF-FTC and those treated with ABC-3TC. We used multivariable linear regression modeling with backward stepwise variable selection to identify factors associated with BMSi. The variables included in the model at the first step of variable selection were, after locking age, sex and body mass index (BMI), treatment group, CD4 count, spine BMD, 25[OH]vitD, CTX, spine BMD, and total hip BMD. Any variable with P > 0.1 was eliminated from the model, and results with P < 0.05 (2-tailed) were considered statistically significant. Analyses were performed using Stata/IC 13.1.

This study included a total of 63 patients between May 2015 and May 2016, of which 36 belonged to the TDF-FTC group and 27 to the ABC-3TC group. The mean ages of the TDF-FTC and ABC-3TC groups were 56.3 (SD = 6.2) and 63 (SD = 10), respectively (P = 0.0001). The median time after initiation of ART was 7.8 years (interquartile range, IQR = 3.2) in the TDF-FTC group versus 8.9 years (IQR = 3.4) in the ABC-3TC group (P = 0.086). No current or past differences in renal function were detected among groups. There were no differences in the use of protease inhibitor as a third drug in the TDF versus ABC groups [3 (8%) versus 2(7%); P = 0.893], respectively. Most demographic and clinical parameters were well balanced between the 2 groups (Table 1).
The median level of the bone formation marker PINP was higher in patients with HIV-1 receiving TDF-FTC treatment (54.3 ng/mL, IQR = 36.5) than in those receiving ABC-3TC (38.1 ng/mL, IQR = 18.5) (P  = 0.0171). We observed a similar effect with respect to the median level of bone resorption marker CTX [0.432 ng/mL (IQR = 0.303) and 0.310 (IQR = 0.12) ng/mL, respectively; P = 0.05]. However, after adjusting by age, sex, BMI, and vitamin D levels, the differences in PINP, CTX, and bone disappeared between TDF and ABC groups. In contrast, we found no differences in vitamin D levels between groups [27.5 ng/mL (IQR = 21.2) versus 23.1 ng/mL (IQR = 17), respectively; P = 0.339].
The median BMD in the lumbar spine of HIV-1–infected patients under TDF-FTC treatment was significantly lower than that of those under ABC-3TC treatment [0.926 g/cm2 (SD = 0.2) versus 1.015 g/cm2 (SD = 0.2); P = 0.04]. This difference was also reflected in the Z-score of the BMD (table 2). In contrast, there were no significant differences between the groups with respect to the BMD (and Z-scores) in either the femoral neck or the total hip (Table 2).
Furthermore, we found no difference in TBS between the TDF-FTC and ABC-3TC groups, even after adjusting for age and BMI [1.254 (IQR = 0.015) versus 1.221 (IQR = 0.026); P = 0.14]. We also found no differences in BMSi between the TDF-FTC and ABC-3TC groups in the unadjusted analysis [81.02 (IQR = 0.819) versus 82.68 (IQR = 1.3); P = 0.269]. The results of the multivariate analysis are shown in Table 3. However, after adjusting for patients' age, sex, BMI, and 25[OH] vitD levels, we found a lower BMSi (ie, poorer bone material properties) in the TDF-FTC group compared with the ABC-3TC group [beta coefficient -3.594 (confidence interval: 95% --0.12 to -7.61); P = 0.043].
In an alternative multivariable analysis, we also locked in the analysis of total hip BMD for an additional adjustment. We used total hip as long as tibial BMD was not available. After this adjustment, differences between groups of treatment remained similar to the previous analysis and statistically different.
For BMSi, we found a strong, significant correlation with 25[OH]vitD levels (r = 0.257; P = 0.04), but no correlation with the BMD or TBS. Similarly, we did not observe any correlation for BMD and TBS values with turnover markers, hsCRP, or 25[OH]vitD levels. No fractures were recorded in any of the patients recruited.

  iconpaperstack View Older Articles   Back to Top   www.natap.org