Infect Dis Ther (2018) 7:293–308 https://doi.org/10.1007/s40121-018-0194-1 ORIGINAL RESEARCH Tenofovir-Associated Bone Adverse Outcomes among a US National Historical Cohort of HIV- Infected Veterans: Risk Modiﬁcation by Concomitant Antiretrovirals . . . . Joanne LaFleur Adam P. Bress Joel Myers Lisa Rosenblatt . . . . Jacob Crook Kristin Knippenberg Roger Bedimo Pablo Tebas Heather Nyman Stephen Esker Received: August 25, 2017 / Published online: February 28, 2018 The Author(s) 2018. This article is an open access publication versus those initiating non-EFV-containing ABSTRACT TDF/FTC regimens. Methods: Using national Veterans Health Introduction: Tenofovir disoproxil fumarate Administration clinical and administrative data (TDF) has been associated with greater inci- sets, we identiﬁed a cohort of treatment-naı¨ve dences of bone complications, which might be HIV-infected veterans without bone disease modiﬁed by some concomitantly administered who initiated therapy with TDF/FTC plus EFV, antiretrovirals, possibly by their effect on teno- rilpivirine, elvitegravir/cobicistat, or ritonavir- fovir concentrations. We compared bone boosted protease inhibitors in 2003–2015. The adverse outcomes among treatment-naıve HIV- primary composite adverse bone outcome was infected US veterans initiating efavirenz (EFV)- the unadjusted incidence rate (IR) of osteo- containing TDF/emtricitabine (FTC) regimens porosis, osteopenia, or fragility fracture (any hip, wrist, or spine fracture). To account for selection bias and confounding, we used inverse Enhanced Content To view enhanced content for this article go to https://doi.org/10.6084/m9.ﬁgshare. probability of treatment-weighted Cox propor- tional hazards regression models to calculate adjusted hazard ratios (HRs) for each outcome Electronic supplementary material The online associated with EFV ? TDF/FTC versus each version of this article (https://doi.org/10.1007/s40121- 018-0194-1) contains supplementary material, which is non-EFV-containing TDF/FTC regimen. available to authorized users. J. LaFleur (&) K. Knippenberg H. Nyman J. Crook Department of Pharmacotherapy, University of Division of Epidemiology, University of Utah, Salt Utah, Salt Lake City, UT, USA Lake City, UT, USA e-mail: Joanne.Laﬂeur@pharm.utah.edu R. Bedimo J. LaFleur A. P. Bress J. Crook K. Knippenberg VA North Texas Health Care System, Dallas, TX, Salt Lake City VA Health Care System, Salt Lake USA City, UT, USA R. Bedimo A. P. Bress University of Texas Southwestern Medical Center, Department of Population Health Sciences, Dallas, TX, USA University of Utah School of Medicine, Salt Lake P. Tebas City, UT, USA Perelman School of Medicine, University of J. Myers L. Rosenblatt S. Esker Pennsylvania, Philadelphia, PA, USA Bristol-Myers Squibb, Lawrenceville, NJ, USA 294 Infect Dis Ther (2018) 7:293–308 Results: Of 33,048 HIV-positive veterans, 7161 cumulative exposure to TDF was independently initiated a TDF/FTC-containing regimen (mean predictive of fragility fracture ; however, no age, 50 years; baseline CD4 \200 cells/mm , evaluation of TDF-associated fracture risk based 33.3%; HIV-1 RNA[100,000 copies/ml, 22.3%; on concomitantly administered ART was mean follow-up, 13.0 months). Of these, 4137 undertaken. Very few epidemiologic studies in initiated EFV- and 3024 non-EFV-containing US populations have compared risks of bone regimens. Veterans initiating EFV- versus non- adverse outcomes with TDF across differing EFV-containing TDF/FTC regimens had a lower TDF-containing antiretroviral regimens . IR of the composite bone outcome (29.3 vs. 41.4 ART initiation is associated with increases in per 1000 patient-years), with signiﬁcant risk bone turnover and modest decreases in bone reductions for this outcome [HR, 0.69; 95% mineral density (BMD) , which are greater conﬁdence interval (CI), 0.58–0.83] and fragility with TDF-containing regimens. BMD loss with fracture (HR, 0.59; 95% CI, 0.44–0.78). TDF is ampliﬁed when coadministered with Conclusion: EFV ? TDF/FTC is associated with boosted protease inhibitors (PIs), but effects are a lower risk of adverse bone outcomes compared less certain when coadministered with other with other TDF-containing regimens in the antiretroviral agents [7, 10, 14–25]. For exam- VHA. ple, in the AIDS Clinical Trials Group (ACTG) Funding: Bristol-Myers Squibb. A5224s substudy of ACTG A5202, spine BMD loss was signiﬁcantly greater in those receiving ritonavir-boosted atazanavir (ATV) ? TDF/FTC Keywords: Efavirenz; Fracture; Osteoporosis; (-3.1%) than in those receiving efavirenz Tenofovir disoproxil fumarate; Veterans (EFV) ? TDF/FTC (-1.7%; P = 0.035) . How- ever, the clinical relevance of these observations INTRODUCTION (increased fracture risk) remains uncertain. Regarding fracture risk, available data from Patients with HIV infection have higher rates of randomized controlled trials are limited by the osteoporosis and fragility fractures than unin- small sample sizes and short observation peri- ods of individual trials and the lack of speciﬁc fected individuals, which are not completely explained by differences in traditional risk fac- reporting of fragility fractures . Thus, despite differences in BMD across regimens in tors such as age and body mass index . Ran- domized controlled trials and observational ACTG A5224s, fractures (all traumatic) were studies suggest that HIV infection itself and the uncommon, and rates were similar across regi- initiation of some antiretroviral therapies (ART) mens . In contrast, one recent large cohort may independently increase the risk of bone study of commercial claims data showed lower adverse outcomes [1–8]. For these reasons, the fracture incidence rates (IRs) for EFV ? TDF/FTC safety proﬁles of antiretroviral agents are subject compared with elvitegravir/cobicistat (EVG/c)/ TDF/FTC, rilpivirine (RPV) ? TDF/FTC, and the to increased scrutiny in an effort to reduce treatment-related side effects in the aging HIV overall HIV population in the database . The lower BMD loss and fracture risk with population. Tenofovir disoproxil fumarate (TDF)/emtric- EFV-containing TDF/FTC regimens compared with other third agents combined with TDF/ itabine (FTC) has been the main nucleos(t)ide reverse transcriptase inhibitor backbone of FTC may relate to differences in drug-drug and combination ART for more than a decade; TDF/ drug-food interactions across these regimens. FTC continues to be included in four of the six Tenofovir plasma concentrations are increased recommended regimens for treatment-naı¨ve by 22–37% when TDF is taken with PIs [27, 28], patients with HIV infection in the current cobicistat , or RPV . The bioavailability Department of Health and Human Services of tenofovir is also increased by up to 40% with concomitant food intake [28, 31]. Certain TDF guidelines . TDF has been noted to increase the risk of fragility fractures [7, 10]. In a Veter- regimens, including RPV, EVG/c, and boosted PIs, are taken with food [32–34]. Conversely, ans Health Administration (VHA) population, Infect Dis Ther (2018) 7:293–308 295 EFV ? TDF/FTC is taken in the fasted state, and  with a sensitivity of 86% and positive pre- no clinically relevant drug-drug interactions dictive value of 87%  was used to exclude have been reported between TDF and either EFV patients with evidence of prior ART received or FTC. Therefore, we hypothesized that outside of the VHA, which included the fol- EFV ? TDF/FTC could result in a lower inci- lowing criteria: exposure to any antiretroviral dence of bone adverse outcomes compared with agents during a 1-year period before the index these other TDF-containing regimens. date (the pre-index period), patients whose The purpose of this analysis was to compare index ART regimen was a ‘‘salvage’’ regimen the incidence of bone adverse outcomes among (i.e., composed of both a PI and an NNRTI or treatment-naı¨ve HIV-infected US veterans composed of 5 or more agents), and patients without bone disease who initiated different whose HIV RNA levels before the index date TDF-containing regimens. were low enough (\500 copies/ml) to suggest prior antiretroviral exposure. Veterans were identiﬁed for inclusion by an available index METHODS date for the ﬁrst pharmacy ﬁll for one of the third agents of interest and if they fulﬁlled the Study Design and Data Sets following criteria at the index date: (1) aged C 18 years; (2) at least 6 months of pre-index Using a national cohort of US veterans, we date VHA activity including in- or outpatient conducted a population-based historical cohort services; (3) no evidence of prior treatment study using VHA databases containing clinical, including ﬁlls for antiretrovirals in the pharmacy, and administrative data from more 6 months before the index date; (4) no evidence than 150 VHA hospitals and 850 outpatient of prior bone disease (deﬁned as diagnosis of clinics nationwide . We obtained demo- osteoporosis/osteopenia by International Clas- graphic, laboratory, diagnosis, and utilization siﬁcation of Disease diagnosis codes, Current data from the Veterans Affairs (VA) Corporate Procedural Terminology codes, or classiﬁcation Data Warehouse (CDW), including medical SAS by bone mineral density test result). Codes used data sets for in- and outpatient encounter data, to identify patients with HIV infection are CDW’s raw pharmacy data, and Decision Sup- shown in Supplemental Digital Content, port Systems and CDW’s laboratory data. To Table 1. conduct the analyses, data sets were housed in the VA Informatics and Computing Infrastruc- Exposures ture environment, which enables access to data and tools for reporting and analysis in a secure Exposures of interest included TDF/FTC (either workspace to ensure veterans’ privacy and data as a ﬁxed-dose combination or as separate security. This article does not contain any agents) plus one of the following agents: EFV, studies with human participants or animals EVG/c, RPV, or any one of three ritonavir- performed by any of the authors. The University boosted PIs (i.e., ATV, lopinavir, or darunavir). of Utah Institutional Review Board and the Salt For regimens with separate dosage forms, the Lake City VA Health Care System Ofﬁce of third agent must have overlapped with the Research and Development approved this study. backbone within 30 days. For boosted or enhanced regimens (EVG/c and RTV-boosted Patients PIs), the third agent must have also overlapped with the booster/enhancer for the patient to be The cohort included all HIV-infected antiretro- classiﬁed as taking the regimen. Discontinua- viral-naı¨ve veterans without bone disease who tion of the regimen was deﬁned as having a gap initiated TDF/FTC plus a third antiretroviral of at least 30 days for either the third agent or agent of interest (see Exposures section) during the backbone; patients who discontinued their the period 2003–2015. A validated algorithm regimen were censored on the ﬁrst day of the 296 Infect Dis Ther (2018) 7:293–308 Table 1 Unadjusted baseline characteristics among HIV-infected veterans receiving initial ART with differing TDF/FTC- containing regimens Initial ART Containing TDF/FTC Plus EFV Non-EFV EVG/c RPV RTV-boosted (n 5 4137) (n 5 3024) (n 5 232) (n 5 171) PI (n 5 2621) Demographics and physical Age, years, mean ± SD 50 ± 10 49 ± 9.8 49 ± 13 47 ± 13 50 ± 9.3 Male 4002 (96.7) 2903 (96.0) 221 (95.3) 161 (94.2) 2521 (96.2) Married 337 (8.1) 221 (7.3) 26 (11.2) 20 (11.7) 175 (6.7) Race White 1240 (30.0) 911 (30.1) 73 (31.5) 48 (28.1) 790 (30.1) Black 2492 (60.2) 1762 (58.3) 124 (53.4) 104 (60.8) 1534 (58.5) Hispanic 260 (6.3) 207 (6.8) 19 (8.2) 10 (5.8) 178 (6.8) Asian 34 (0.8) 28 (0.9) 5 (2.2) 2 (1.2) 21 (0.8) Other 22 (0.5) 28 (0.9) 2 (0.9) 2 (1.2) 24 (0.9) Missing 89 (2.2) 88 (2.9) 9 (3.9) 5 (2.9) 74 (2.8) BMI (kg/m , mean ± SD) 26 ± 5.0 25 ± 5.1 26 ± 5.6 27 ± 5.3 25 ± 5.1 Pre-index prognostic indices ? 3 CD4 count, cells/mm \200 1286 (31.1) 1103 (36.5) 60 (25.9) 23 (13.5) 1020 (38.9) 200–299 698 (16.9) 477 (15.8) 33 (14.2) 30 (17.5) 414 (15.8) 300–399 634 (15.3) 342 (11.3) 34 (14.7) 11 (6.4) 297 (11.3) 400–499 401 (9.7) 222 (7.3) 24 (10.3) 33 (19.3) 165 (6.3) C 500 588 (14.2) 426 (14.1) 61 (26.3) 63 (36.8) 302 (11.5) Missing 530 (12.8) 454 (15.0) 20 (8.6) 11 (6.4) 423 (16.1) HIV viral load, copies/ml \10,000 1184 (28.6) 950 (31.4) 65 (28.0) 68 (39.8) 817 (31.2) 10,000–100,000 1434 (34.7) 940 (31.1) 78 (33.6) 67 (39.2) 795 (30.3) [100,000 913 (22.1) 686 (22.7) 61 (26.3) 15 (8.8) 610 (23.3) Missing 606 (14.6) 448 (14.8) 28 (12.1) 21 (12.3) 399 (15.2) Pre-index renal function eGFR, ml/min/1.73 m C 90 2193 (53.0) 1552 (51.3) 143 (61.6) 108 (63.2) 1301 (49.6) 60–89 1360 (32.9) 985 (32.6) 73 (31.5) 54 (31.6) 858 (32.7) 30–59 164 (4.0) 123 (4.1) 3 (1.3) 4 (2.3) 116 (4.4) 15–29 5 (0.1) 7 (0.2) 0 (0.0) 0 (0.0) 7 (0.3) \15 39 (0.9) 53 (1.8) 0 (0.0) 0 (0.0) 53 (2.0) Missing 376 (9.1) 304 (10.1) 13 (5.6) 5 (2.9) 286 (10.9) Infect Dis Ther (2018) 7:293–308 297 Table 1 continued Initial ART Containing TDF/FTC Plus EFV Non-EFV EVG/c RPV RTV-boosted (n 5 4137) (n 5 3024) (n 5 232) (n 5 171) PI (n 5 2621) Pre-index comorbid diagnoses CAD/CVD 447 (10.8) 280 (9.3) 28 (12.1) 13 (7.6) 239 (9.1) Heart failure 132 (3.2) 62 (2.1) 5 (2.2) 2 (1.2) 55 (2.1) Dyslipidemia 659 (15.9) 412 (13.6) 50 (21.6) 29 (17.0) 333 (12.7) Hypertension 1521 (36.8) 910 (30.1) 83 (35.8) 64 (37.4) 763 (29.1) Diabetes mellitus 561 (13.6) 358 (11.8) 30 (12.9) 19 (11.1) 309 (11.8) Chronic kidney disease 281 (6.8) 169 (5.6) 16 (6.9) 9 (5.3) 144 (5.5) End-stage renal disease 15 (0.4) 12 (0.4) 1 (0.4) 0 (0.0) 11 (0.4) Fracture 30 (0.7) 20 (0.7) 4 (1.7) 1 (0.6) 15 (0.6) Viral hepatitis 1122 (27.1) 902 (29.8) 59 (25.4) 42 (24.6) 801 (30.6) Tuberculosis 52 (1.3) 46 (1.5) 1 (0.4) 2 (1.2) 43 (1.6) Psychiatric disorder 1495 (36.1) 1384 (45.8) 119 (51.3) 104 (60.8) 1161 (44.3) Depression 962 (23.3) 878 (29.0) 83 (35.8) 68 (39.8) 727 (27.7) Schizophrenia 140 (3.4) 194 (6.4) 12 (5.2) 9 (5.3) 173 (6.6) Bipolar disorder 635 (15.3) 648 (21.4) 61 (26.3) 63 (36.8) 524 (20.0) Psychosis 231 (5.6) 308 (10.2) 26 (11.2) 18 (10.5) 264 (10.1) Posttraumatic stress disorder 366 (8.8) 379 (12.5) 39 (16.8) 52 (30.4) 288 (11.0) Tobacco use 1279 (30.9) 918 (30.4) 70 (30.2) 77 (45.0) 771 (29.4) Alcohol abuse 946 (22.9) 725 (24.0) 59 (25.4) 51 (29.8) 615 (23.5) Medications Methadone 49 (1.2) 70 (2.3) 2 (0.9) 2 (1.2) 66 (2.5) Proton pump inhibitors 1179 (28.5) 713 (23.6) 56 (24.1) 23 (13.5) 634 (24.2) Bisphosphonates 8 (0.2) 9 (0.3) 0 (0.0) 0 (0.0) 9 (0.3) Testosterone 66 (1.6) 51 (1.7) 2 (0.9) 3 (1.8) 46 (1.8) Data are n (%) unless otherwise indicated Non-EFV includes the EVG/c, RPV, and RTV-boosted PI groups Pre-index comorbidities and clinical characteristics were identiﬁed in the 6–12-month pre-index period c 2 Deﬁned as either a chronic kidney disease diagnosis or two consecutive measures of eGFR\60 ml/min/1.73 m occurring at least 30 days apart Deﬁned as either a diagnosis of end-stage renal disease, kidney transplant, or dialysis Includes abuse, dependence, rehabilitation, and toxicity related to tobacco ART antiretroviral therapy, BMI body mass index, CAD coronary artery disease, CVD cerebrovascular disease, EFV efavirenz, eGFR estimated glomerular ﬁltration rate, EVG/c elvitegravir/cobicistat, FTC emtricitabine, PI protease inhibitor (atazanavir, darunavir, or lopinavir), RPV rilpivirine, RTV ritonavir, SD standard deviation, TDF tenofovir disoproxil fumarate 298 Infect Dis Ther (2018) 7:293–308 ﬁrst 30-day gap following the end of the prior in the unweighted cohort assuming a Poisson days’ supply received by the patient. distribution. To control for confounding by indication and selection bias, we used inverse probability of treatment weighting (IPTW) for Outcomes each patient . Weighted Cox proportional hazards regression models were used to estimate The primary composite outcome was a bone adjusted hazard ratios (HRs) for bone outcomes adverse event deﬁned as a diagnosis of osteo- associated with EFV ? TDF/FTC compared with porosis; a BMD T-score in the osteoporotic or EVG/c ? TDF/FTC, RPV ? TDF/FTC, ritonavir- osteopenic ranges for the femoral neck, total boosted PIs ? TDF/FTC, and all non-EFV regi- spine, distal radius, or total hip; or a diagnosis mens combined. To reduce risk of bias, model or procedure code for likely fragility fracture variability, and unreliable CIs, we excluded any (any hip, wrist/forearm, or spine fracture). BMD group with \5 events from analysis compar- T-scores were extracted from patient radiology isons. We used IPTW in the primary analysis; dual-energy X-ray absorptiometry (DEXA) however, if covariate balance was not achieved reports and clinical notes using a previously with IPTW, we conducted a sensitivity analysis developed natural language processing tool, using matching weights, which are less sensitive with accuracy in the range of 90.4–92.8% to residual and unmeasured confounding. All [38, 39]. All codes used to identify outcomes of analyses were done in SAS version 9.2 (SAS interest are provided in Supplemental digital Institute, Cary, NC). content, Table 2. Covariates RESULTS To control for confounding and selection bias, Patient Characteristics we measured baseline covariates that were selected on the basis of potential associations A total of 7161 patients met all eligibility crite- with treatment and/or outcomes, as found in ria (Fig. 1). Of these, 4137 initiated EFV and published literature and based on prior clinical 3024 initiated non-EFV regimens, including 232 knowledge of ART and HIV. These included with EVG/c, 171 with RPV, and 2621 with baseline demographics, baseline HIV laboratory ritonavir-boosted PIs. At baseline, 26 (0.4%) had measures, baseline BMD measures and related received a bisphosphonate (Table 1). For all diagnoses, lifestyle exposures, other comor- unadjusted comparisons, psychiatric disorders bidities, medication exposures, and calendar were overrepresented in those receiving non- year of the index regimen. All covariates were EFV-containing regimens versus EFV-contain- identiﬁed over a 12-month look-back period ing regimens (Fig. 2 and Supplemental digital preceding the index date. Speciﬁc deﬁnitions content Table 4). After IPTW, covariate balance for all covariates are provided in the Supple- was achieved among all variables for EFV versus mental digital content, Table 3. all non-EFV combined (Fig. 2), versus EVG/c (Supplemental digital content, Fig. 1A), and Statistical Analysis versus ritonavir-boosted PIs (Supplemental dig- ital content, Fig. 1B). For EFV versus RPV, the covariate balance was achieved for 83.0% of We calculated baseline characteristics overall variables (Supplemental digital content, and by treatment group and used standardized Fig. 1C). Mean (standard deviation) follow-up mean differences (SMDs) to compare differences times were 13.0 (19.2) months overall and 15.0 between groups, with SMDs outside the bounds (21.2), 7.6 (7.6), 10.0 (10.5), and 10.5 (16.7) of ± 0.1 indicating meaningful differences . months for EFV, EVG/c, RPV, and ritonavir- We calculated crude IRs of bone adverse out- boosted PIs, respectively. Corresponding med- comes per 1000 patient-years of exposure and ian (interquartile range; min–max) values were associated exact 95% conﬁdence intervals (CIs) Infect Dis Ther (2018) 7:293–308 299 Received a diagnosis for HIV infection between January 1, 2003 through December 31, 2015 N = 33,048 Excluded those who did not receive an ART regimen of interest during the study period N = 19,682 Treated with a fixed-dose combination of TDF/FTC plus EFV, EVG/c, RPV, or RTV-boosted PIs during study period N = 13,366 Excluded those not showing a 6-month pattern of receiving regular VHA care N = 451 Showed a pattern of receiving regular VHA care N = 12,915 Excluded those with evidence of prior ART (N = 5679) or with prior bone disease* (N = 61) or with no age or BMI measurements (N = 14) Incident-treated patients N = 7161 FINAL COHORT EFV+TDF/FTC EVG/c/TDF/FTC RPV/TDF/FTC RTV-boosted PIs n = 4137 (57.8%) n = 232 (3.2%) n = 171 (2.4%) +TDF/FTC n = 2621 (36.6%) EFV/TDF/FTC EFV+TDF/FTC ATV LPV DRV n = 3325 n = 812 n = 1489 n = 642 n = 490 (80.4%) (19.6%) (56.8%) (24.5%) (18.7%) Fig. 1 Patient selection according to eligibility criteria. fumarate, VHA Veterans Health Administration. *Bone ART antiretroviral therapy, ATV atazanavir, BMI body disease deﬁned as diagnosis of osteoporosis/osteopenia by mass index, DRV darunavir, EFV efavirenz, EVG/c international classiﬁcation of disease diagnosis codes, elvitegravir/cobicistat, FTC emtricitabine, LPV lopinavir, current procedural terminology codes, or classiﬁcation by PI protease inhibitor (atazanavir, darunavir, or lopinavir), bone mineral density test result RPV rilpivirine, RTV ritonavir, TDF tenofovir disoproxil 5.0 (2.0–14.2; 0.1–127.7), 5.7 (2.1–17.9; analyses, EFV was associated with a statistically 0.1–110.3), 4.6 (2.1–10.4; 0.2–34.6), 5.9 signiﬁcant 31% lower risk of the composite (2.8–13.4; 0.5–48.3), and 4.2 (1.8–10.7; bone outcome than non-EFV groups combined, 0.1–127.7) months. 25% lower than RPV, and 30% lower than the ritonavir-boosted PI group. Incidence and Risk of Any Bone Outcome Incidence and Risk of Osteoporosis or Osteopenia The unadjusted (crude) IRs and adjusted HRs for all comparisons are summarized in Fig. 3. The crude IR of the composite bone outcome was The crude IR for osteoporosis was lower for EFV lower in the EFV group than in the EVG/c, RPV, compared with all other regimens. In adjusted and ritonavir-boosted PI groups. In adjusted analyses, EFV was associated with a statistically 300 Infect Dis Ther (2018) 7:293–308 Hypertension Index year 2007 Index year 2010 Index year 2009 CD4 300-399 cells/mm Proton pump inhibitor Age, years Index year 2008 CD4 400-499 cells/mm HIV viral load 10,000-100,000 c/mL Heart failure Index year 2011 Dyslipidemia Diabetes mellitus CAD/CVD Chronic kidney disease Body mass index, kg/m African American Male eGFR =90 mL/min/1.73 m Married CD4 200-299 cells/mm Tobacco use Fracture eGFR 60-89 mL/min/1.73 m CD4 =500 cells/mm White race EFV vs Non-EFV eGFR 30-59 mL/min/1.73 m HIV viral load Missing c/mL Unweighted End-stage renal disease Testostorone IPT weighted Index year 2012 Asian race HIV viral load >100,000 c/mL Index year 2004 Bisphosphonate Tuberculosis Hispanic race Alcohol abuse eGFR 15-29 mL/min/1.73 m eGFR missing Other race Missing race Index year 2006 Viral hepatitis HIV viral load <10,000 c/mL CD4 missing eGFR <15 mL/min/1.73 m Index year 2013 Methadone CD4 <200 cells/mm Posttraumatic stress disorder Depression Schizophrenia Bipolar Psychosis Index year 2005 Index year 2014 Psychiatric disorder Index year 2015 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 Overrepresented in Non-EFV ← Balanced→ Overrepresented in EFV Standardized Mean Difference (Cohen's D) Fig. 2 Baseline demographics: veriﬁcation that IPT disease, CKD chronic kidney disease, EFV efavirenz, eGFR weighting achieves baseline covariate balance between estimated glomerular ﬁltration rate, FTC emtricitabine, TDF/FTC-containing regimens with EFV versus all non- IPT inverse probability of treatment, TDF tenofovir EFV. CAD/CVD coronary artery disease/cerebrovascular disoproxil fumarate Infect Dis Ther (2018) 7:293–308 301 signiﬁcant 36% lower risk of osteoporosis than Sensitivity Analysis non-EFV groups combined, 53% lower than EVG/c, and 35% lower than the ritonavir- In matching weight analyses, the risks of any boosted PI group. Too few events were observed bone outcome or fracture for RPV versus EFV with RPV to make comparisons. The crude IR of showed a similar magnitude, suggesting the osteopenia was lower for EFV compared with IPTW ﬁndings were robust regarding residual EVG/c, but for all other comparisons IRs were confounding present for 7/47 (14.9%) of vari- similar. In adjusted analyses, the risk of able comparisons for EFV versus RPV. osteopenia was similar between EFV and the ritonavir-boosted PI group (too few events were DISCUSSION observed to report comparisons for EVG/c or RPV). When combining osteoporosis and EFV ? TDF/FTC was generally associated with osteopenia events, EFV was associated with a lower risks of bone adverse outcomes compared 21% lower risk compared with the ritonavir- with TDF-containing regimens with EVG/c, boosted PI group. Frequencies of available pre- RPV, or ritonavir-boosted PIs in the VHA cohort and post-index DEXA scans as well as baseline characteristics by DEXA scan status are pre- over an average follow-up time of 13.0 months. These ﬁndings suggest that patients with HIV sented in the Supplemental digital content, Tables 6 and 7, respectively. infection receiving EFV ? TDF/FTC, which has no known drug-drug interactions with TDF or FTC, or drug-food interactions, may be at lower Incidence and Risk of Fracture overall risk of bone adverse outcomes compared with those receiving TDF-containing regimens The crude IR of fragility fractures was lower for with EVG/c, RPV, or ritonavir-boosted PIs. EFV compared with EVG/c, RPV, and ritonavir- This study is the ﬁrst to identify an associa- boosted PIs. In adjusted analyses, risk differ- tion between EFV ? TDF/FTC use among veter- ences were statistically signiﬁcant for EFV versus ans and a reduced risk of bone adverse all non-EFV combined (41% lower), RPV (57% outcomes versus other TDF-containing regi- lower), and ritonavir-boosted PIs (40% lower). mens. These ﬁndings conﬁrm those of Nkhoma Of the individual fracture outcomes, EFV had et al. , who conducted a large claims data- a lower incidence compared with each non-EFV base analysis of bone fractures associated with regimen for all fracture sites in all comparisons TDF/FTC and various third agents (although except for hip fractures with ritonavir-boosted adjusted analyses were not possible because of PIs, which had too few events to make com- insufﬁcient numbers of fracture events). parisons. In adjusted analyses, this corre- Nkhoma et al.  found a lower fracture IR sponded to statistically signiﬁcant reduced risks with EFV ? TDF/FTC (3.4 per 1000 patient- for EFV compared with all non-EFV regimens years) compared with RPV ? TDF/FTC (3.6 per combined for vertebral fracture (51% lower) and 1000 patient-years) or with EVG/c ? TDF/FTC wrist/forearm fracture (60% lower). Signiﬁ- (4.4 per 1000 patient-years). The IRs in this cantly lower risks were also observed versus study followed the same general trend as in ritonavir-boosted PIs for vertebral fracture (48% Nkhoma et al. , but were higher in magni- lower) and wrist/forearm fracture (60% lower). tude, possibly because the current study was Too few events were observed for individual conducted in a higher-risk population in a rel- fracture outcomes to make comparisons with atively closed, integrated care network. EVG/c or RPV. Cobicistat, rilpivirine, and ritonavir increase For a summary of event numbers, time to serum creatinine and decrease the estimated events and follow-up times for all bone adverse glomerular ﬁltration rate (GFR) via inhibition of outcomes, see Supplemental digital content, renal tubular transporters [42, 43], but this does Table 5. not appear to affect actual GFR as measured by iohexol clearance . Therefore, the bone 302 Infect Dis Ther (2018) 7:293–308 Crude Incidence (95% CI) per 1000 PY Bone outcome EFV+TDF/FTC Comparator IPTW Adjusted Hazard Ratio (95% CI) EFV+TDF/FTC (n = 4137) vs all other TDF-containing (n = 3024) Any bone adverse outcome 29.3 (24.7-34.5) 41.4 (33.7-50.2) 0.69 (0.58-0.83) Osteoporosis by diag./T-score 8.4 (6.1-11.4) 13.0 (8.9-18.2) 0.64 (0.46-0.89) Osteopenia by T-score 16.2 (12.9-20.1) 17.0 (12.3-22.9) 0.90 (0.69-1.17) Osteoporosis or osteopenia 20.5 (16.7-24.9) 24.7 (19.0-31.7) 0.79 (0.63-0.99) Any fracture (spine/hip/wrist) 10.2 (7.6-13.4) 18.3 (13.4-24.4) 0.59 (0.44-0.78) Spine 2.5 (1.3-4.3) 4.7 (2.4-8.1) 0.49 (0.28-0.85) Hip 4.3 (2.7-6.5) 3.1 (1.3-6.1) 1.73 (0.96-3.11) Wrist/forearm 4.1 (2.5-6.3) 11.1 (7.3-16.0) 0.40 (0.27-0.60) EFV+TDF/FTC (n = 4137) vs EVG/c/TDF/FTC (n = 232) Any bone adverse outcome 29.3 (24.7-34.5) 49.5 (19.9-102.0) 0.92 (0.69-1.22) Osteoporosis by diag./T-score 8.4 (6.1-11.4) 20.8 (4.3-60.7) 0.47 (0.29-0.75) Osteopenia by T-score 16.2 (12.9-20.1) 35.2 (11.4-82.1) — Osteoporosis or osteopenia 20.5 (16.7-24.9) 35.2 (11.4-82.1) 0.74 (0.53-1.03) Any fracture (spine/hip/wrist) 10.2 (7.6-13.4) 13.7 (1.7-49.5) — Spine 2.5 (1.3-4.3) 6.8 (0.2-38.1) — Hip 4.3 (2.7-6.5) 6.8 (0.2-38.1) — Wrist/forearm 4.1 (2.5-6.3) 6.8 (0.2-38.1) — EFV+TDF/FTC (n = 4137) vs RPV/TDF/FTC (n = 171) Any bone adverse outcome 29.3 (24.7-34.5) 43.7 (16.1-95.2) 0.75 (0.59-0.95) Osteoporosis by diag./T-score 8.4 (6.1-11.4) 7.1 (0.2-39.5) — Osteopenia by T-score 16.2 (12.9-20.1) 14.2 (1.7-51.4) — Osteoporosis or osteopenia 20.5 (16.7-24.9) 21.6 (4.5-63.1) — Any fracture (spine/hip/wrist) 10.2 (7.6-13.4) 28.8 (7.8-73.6) 0.43 (0.30-0.61) Spine 2.5 (1.3-4.3) 7.1 (0.2-39.5) — Hip — 4.3 (2.7-6.5) 14.1 (1.7-50.8) Wrist/forearm — 4.1 (2.5-6.3) 7.1 (0.2-39.5) EFV+TDF/FTC (n = 4137) vs RTV-boosted PI+TDF/FTC (n = 2621) Any bone adverse outcome 0.70 (0.58-0.84) 29.3 (24.7-34.5) 40.7 (32.6-50.1) Osteoporosis by diag./T-score 8.4 (6.1-11.4) 12.9 (8.6-18.5) 0.65 (0.46-0.90) Osteopenia by T-score 16.2 (12.9-20.1) 16.0 (11.2-22.2) 0.92 (0.70-1.20) Osteoporosis or osteopenia 20.5 (16.7-24.9) 24.3 (18.2-31.7) 0.79 (0.63-0.99) Any fracture (spine/hip/wrist) 10.2 (7.6-13.4) 17.9 (12.8-24.4) 0.60 (0.45-0.80) Spine 2.5 (1.3-4.3) 4.4 (2.1-8.0) 0.52 (0.30-0.91) Hip 4.3 (2.7-6.5) 2.2 (0.7-5.1) — Wrist/forearm 4.1 (2.5-6.3) 11.6 (7.6-17.0) 0.40 (0.26-0.61) 0.25 2.5 Favors EFV+TDF/FTC Favors Comparator Fig. 3 Unadjusted incidence rates and IPTW adjusted IPTW inverse probability of treatment weighting, PI hazard ratios of bone adverse outcomes among patients protease inhibitor (atazanavir, darunavir, or lopinavir), PY treated with different TDF/FTC-containing regimens patient-years, RPV rilpivirine, RTV ritonavir, TDF teno- with at least ﬁve events per group. CI conﬁdence interval, fovir disoproxil fumarate, – insufﬁcient events to calculate CVD cerebrovascular disease, diag. diagnosis, EFV efavir- hazard ratio. Measured at the femoral neck, total spine, enz, EVG/c elvitegravir/cobicistat, FTC emtricitabine, distal radius, or total hip Infect Dis Ther (2018) 7:293–308 303 effects we observed were more likely due to However, while it appears that EFV ? TDF/FTC relative differences in tenofovir exposure across has a lower risk for bone adverse outcomes than the evaluated regimens. Both RPV/FTC/TDF and PIs and other boosted regimens in combination EVG/COBI/FTC/TDF are recommended to be with TDF/FTC, our study is not generalizable to administered with food, which increases teno- the use of unboosted integrase inhibitors, for fovir exposure compared with fasting adminis- which additional research is needed. tration [33, 34]. In addition, both protease Other options to counter antiretroviral-asso- inhibitors (including RTV)  and cobicistat ciated BMD loss include vitamin D/calcium and/  increase tenofovir exposure by inhibition or zoledronic acid supplementation. Vitamin of intestinal P-gp-mediated efﬂux of TDF. Taken D/calcium supplementation lessened BMD loss together, these effects on tenofovir exposure among patients receiving EFV ? TDF/FTC over may have contributed to the increased risk of 48 weeks [21, 23], and single-dose zoledronic adverse bone outcomes that we observed with acid administered at the initiation of a PI-con- EVG/c/FTC/TDF, RPV/TDF/FTC, and RTV-boos- taining TDF/FTC regimen can prevent the initial ted PIs in this analysis. BMD loss . These options, together with For patients with a high risk of bone adverse careful antiretroviral choice, should be consid- outcomes, the use of the novel formulation ered between patients and health care providers, tenofovir alafenamide (TAF) or abacavir has taking into account the drug resistance proﬁle, been associated with lower BMD losses at the treatment history, other comorbidities, and the time of therapy initiation compared with TDF risk or tolerability of side effects. . TAF, a prodrug of tenofovir, is associated Strengths and limitations of this study are with a 91% lower plasma tenofovir concentra- those common to large epidemiologic studies. tion than that following TDF administration The main strengths of our study were its large while maintaining higher intracellular teno- sample size and detailed data from VHA data fovir concentrations in peripheral blood sets, as well as the use of a natural language- mononuclear cells for HIV suppression . processing tool, which allowed us to extract Biomarkers of bone turnover appear to be less BMD results from radiology and clinical notes. affected by TAF-containing regimens compared Our study also has limitations. The VHA popu- with TDF-containing regimens. TAF is now lation was more than 95% male, which limits recommended in HIV treatment guidelines , generalizability to female populations that and its use is being preferred to TDF by HIV-care could be affected differently by the predictors providers concerned about bone toxicity of TDF identiﬁed in our study. Follow-up times were in the aging HIV population. short, which may have led to underascertain- Given the ongoing widespread use of ment of relevant bone adverse outcomes; how- EFV ? TDF/FTC and the absence of an EFV co- ever, statistically signiﬁcant between-regimen formulation with TAF, the results of the current differences for relevant bone adverse outcomes study may reassure physicians and their were noted despite these follow-up times. patients about the bone safety of this combi- Moreover, other data have demonstrated that nation. Moreover, where alternatives to TDF are the incidence of fracture is highest during the limited (such as in resource-limited settings), or ﬁrst and second years after ART initiation, tail- where use of generic EFV or TDF as a cost ing off thereafter ; fracture risk among men reduction strategy is available, or where with HIV infection is higher among older indi- EFV ? TDF/FTC is available through the viduals  as was the case in this VA popula- Medicines Patent Pool, the choice of a third tion. Thus, we consider that follow-up times in agent remains critical to long-term safety. The the current study were sufﬁcient to detect dif- reduced risk of bone adverse outcomes for ferences in BMD and fracture risk. Although EFV ? TDF/FTC found in this study is of high IPTW was successfully used to adjust for selec- relevance, especially in resource-limited settings tion bias and measured confounders for all where the cost effectiveness of the ﬁxed-dose comparisons with the exception of EFV versus combination has achieved widespread use . RPV, the potential for unmeasured confounders 304 Infect Dis Ther (2018) 7:293–308 and incomplete adjustment for measured con- Utah. The article processing charges were also founders cannot be ruled out. For example, funded by Bristol-Myers Squibb. patients with a range of psychiatric disorders Authorship. All authors had full access to all were less likely to be prescribed EFV-containing of the data in this study and take complete regimens across all comparisons (Fig. 2 and responsibility for the integrity of the data and Supplemental digital content, Fig. 1), likely accuracy of the data analysis. All named authors because of channeling bias consequent to the meet the International Committee of Medical side effect proﬁle of EFV in patients with severe Journal Editors (ICMJE) criteria for authorship psychiatric symptoms . Various psychiatric for this manuscript, take responsibility for the disorders, such as schizophrenia and depres- integrity of the work as a whole, and have given sion, and psychotropic medications are associ- ﬁnal approval to the version to be published. ated with BMD loss and an increased risk of fracture . However, our method of control- Author Contribution. Joanne LaFleur, Adam ling for confounding effectively balanced the P. Bress, Joel Myers, Lisa Rosenblatt, Roger observed differences in these measured charac- Bedimo, Pablo Tebas, Heather Nyman, and teristics. Speciﬁcally, the six psychiatric covari- Stephen Esker contributed to the conception ates examined were balanced after IPTW for the and design of the study. Joanne LaFleur, Adam EFV versus EVG/c and EFV versus ritonavir- P. Bress, Kristin Knippenberg, and Jacob Crook boosted PI comparisons. For the EFV versus RPV contributed to data collection. Joanne LaFleur, comparison, only one remained imbalanced Adam P. Bress, Kristin Knippenberg, and Jacob after IPTW, and this was by a small margin Crook contributed to data analysis. Joanne (SMD of 0.102). In addition, the more conser- LaFleur, Adam P. Bress, Joel Myers, Lisa Rosen- vative matching weights analyses produced blatt, Roger Bedimo, Pablo Tebas, Heather qualitatively similar results to those using Nyman, and Stephen Esker contributed to data IPTW, making it highly unlikely that channel- interpretation. All authors contributed to writ- ing bias affected our results. Finally, as in any ing and editing the manuscript. retrospective observational study, causal asso- ciations cannot be proven; thus, these ﬁndings Editorial Assistance. Editorial Assistance require conﬁrmation in further prospective was provided by Julian Martins of inScience studies. Communications, Springer Healthcare, which was funded by Bristol-Myers Squibb. CONCLUSION Disclosures. This work was supported by Bristol-Myers Squibb by a grant to the University In conclusion, EFV ? TDF/FTC was generally of Utah. Joanne LaFleur received some salary associated with a lower incidence of bone support from this grant and declares no intel- adverse outcomes, including osteoporosis, any lectual property rights related to this research. major fracture, vertebral fracture, and Outside of the funded work, over the last 3 years, wrist/forearm fracture, compared with other the following organizations provided research TDF/FTC-containing regimens in the VHA. The grants to the University of Utah and Joanne third agent in antiretroviral regimens may have LaFleur worked on those projects and/or a signiﬁcant effect on the risk of bone adverse received salary or other types of support that events associated with TDF. were funded by those grants: Gilead Sciences, Inc., Anolinx LLC, Skaggs Foundation, and Agency for Healthcare Research and Quality. ACKNOWLEDGMENTS Adam P. Bress received some salary support from this grant and declares no intellectual property rights related to this research. Outside of the Funding. This work was supported by Bris- funded work, over the last 3 years, the following tol-Myers Squibb by a grant to the University of Infect Dis Ther (2018) 7:293–308 305 organizations provided research grants to the human participants or animals performed by University of Utah, and Adam P. Bress worked any of the authors. The University of Utah on those projects and/or received salary or other Institutional Review Board and the Salt Lake types of support that were funded by those City VA Health Care System Ofﬁce of Research grants: Gilead Sciences, and National Heart, and Development approved this study. Lung, and Blood Institute (NHLBI). Joel Myers is Data Availability. The data sets generated an employee of, and owns stock in, Bristol-My- during and/or analyzed during the current ers Squibb. Lisa Rosenblatt is an employee of, study are not publicly available because of and owns stock in, Bristol-Myers Squibb. Jacob compliance with Veteran Healthcare Adminis- Crook received some salary support from this tration restrictions on data sharing. grant and declares no intellectual property rights related to this research. Outside of the Previous Presentation. These data were funded work, over the last 3 years, the following previously presented in poster form at Infec- organizations provided research grants to the tious Disease Week (IDWeek) October 26–30, University of Utah and Jacob Crook worked on 2016, New Orleans, LA. those projects and/or received salary or other types of support that were funded by those Open Access. This article is distributed grants: Cubist, Gilead Sciences, Inc., Anolinx under the terms of the Creative Commons LLC, Skaggs Foundation, Agency for Healthcare Attribution-NonCommercial 4.0 International Research and Quality, and Utah Department of License (http://creativecommons.org/licenses/ Health. Kristin Knippenberg received some sal- by-nc/4.0/), which permits any non- ary support from this grant and declares no commercial use, distribution, and reproduction intellectual property rights related to this in any medium, provided you give appropriate research. Outside of the funded work, over the credit to the original author(s) and the source, last 3 years, the following organizations pro- provide a link to the Creative Commons license, vided research grants to the University of Utah and indicate if changes were made. and Kristin Knippenberg worked on those pro- jects and/or received salary or other types of support that were funded by those grants: Gilead Sciences, Inc., Anolinx LLC, Skaggs Foundation, Agency for Healthcare Research and Quality, REFERENCES and Utah Department of Health. Roger Bedimo has received grants and research support awar- ded to the Veterans Affairs North Texas Health- 1. Cotter AG, Sabin CA, Simelane S, et al. Relative contribution of HIV infection, demographics and care System from Merck & Co; he has served as a body mass index to bone mineral density. AIDS. scientiﬁc advisor for Bristol-Myers Squibb, 2014;28(14):2051–60. Merck & Co, Inc., and Theratechnologies, Inc. 2. Brown TT, Qaqish RB. Antiretroviral therapy and Pablo Tebas has served as a scientiﬁc advisor to the prevalence of osteopenia and osteoporosis: a Merck & Co, Inc., and is a member of an adju- meta-analytic review. AIDS. 2006;20(17):2165–74. dication committee in a vaccine trial sponsored by Glaxo. Heather Nyman received some salary 3. Hileman CO, Labbato DE, Storer NJ, Tangpricha V, McComsey GA. Is bone loss linked to chronic support and travel expenses from this grant and inﬂammation in antiretroviral-naive HIV-infected declares no intellectual property rights related to adults? A 48-week matched cohort study. AIDS. this research; she has also received consultancy 2014;28(12):1759–67. honoraria from Otsuka and Fresenius. Stephen 4. Young B, Dao CN, Buchacz K, Baker R, Brooks JT. Esker is an employee of, and owns stock in, Investigators HIVOS. Increased rates of bone frac- Bristol-Myers Squibb. ture among HIV-infected persons in the HIV Out- patient Study (HOPS) compared with the US general Compliance with Ethics Guidelines. This population, 2000-2006. Clin Infect Dis. article does not contain any studies with 2011;52(8):1061–8. 306 Infect Dis Ther (2018) 7:293–308 5. Yong MK, Elliott JH, Woolley IJ, Hoy JF. Low CD4 16. Erlandson KM, Kitch D, Tierney C, et al. Weight count is associated with an increased risk of fragility and lean body mass change with antiretroviral ini- fracture in HIV-infected patients. J Acquir Immune tiation and impact on bone mineral density. AIDS. Deﬁc Syndr. 2011;57(3):205–10. 2013;27(13):2069–79. 6. Grund B, Peng G, Gibert CL, et al. Continuous 17. Grant PM, Kitch D, McComsey GA, Tierney C, Ha B, antiretroviral therapy decreases bone mineral den- Brown TT. Differential skeletal impact of tenofovir sity. AIDS. 2009;23(12):1519–29. disoproxil fumarate in young versus old HIV-in- fected adults. HIV Clin Trials. 2015;16(2):66–71. 7. McComsey GA, Kitch D, Daar ES, et al. Bone min- eral density and fractures in antiretroviral-naive 18. Gupta SK, Mi D, Moe SM, Dube MP, Liu Z. Effects of persons randomized to receive abacavir-lamivudine switching from efavirenz to raltegravir on or tenofovir disoproxil fumarate-emtricitabine endothelial function, bone mineral metabolism, along with efavirenz or atazanavir-ritonavir: Aids inﬂammation, and renal function: a randomized, Clinical Trials Group A5224s, a substudy of ACTG controlled trial. J Acquir Immune Deﬁc Syndr. A5202. J Infect Dis. 2011;203(12):1791–801. 2013;64(3):279–83. 8. van Vonderen MG, Lips P, van Agtmael MA, et al. 19. Moyle GJ, Stellbrink HJ, Compston J, et al. 96-Week First line zidovudine/lamivudine/lopinavir/riton- results of abacavir/lamivudine versus tenofovir/ avir leads to greater bone loss compared to nevirap- emtricitabine, plus efavirenz, in antiretroviral- ine/lopinavir/ritonavir. AIDS. 2009;23(11):1367–76. naive, HIV-1-infected adults: ASSERT study. Antivir Ther. 2013;18(7):905–13. 9. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the use of antiretroviral 20. Nelson MR, Elion RA, Cohen CJ, et al. Rilpivirine agents in HIV-1-infected adults and adolescents. versus efavirenz in HIV-1-infected subjects receiv- https://aidsinfo.nih.gov/guidelines/html/1/adult- ing emtricitabine/tenofovir DF: pooled 96-week and-adolescent-arv-guidelines/11/what-to-start. data from ECHO and THRIVE Studies. HIV Clin Accessed 27 Feb 2017. Trials. 2013;14(3):81–91. 10. Stellbrink HJ, Orkin C, Arribas JR, et al. Comparison 21. Overton ET, Chan ES, Brown TT, et al. Vitamin D of changes in bone density and turnover with aba- and calcium attenuate bone loss with antiretroviral cavir-lamivudine versus tenofovir-emtricitabine in therapy initiation: a randomized trial. Ann Intern HIV-infected adults: 48-week results from the Med. 2015;162(12):815–24. ASSERT study. Clin Infect Dis. 2010;51(8):963–72. 22. Tebas P, Kumar P, Hicks C, et al. Greater change in 11. Bedimo R, Maalouf NM, Zhang S, Drechsler H, bone turnover markers for efavirenz/emtricitabine/ Tebas P. Osteoporotic fracture risk associated with tenofovir disoproxil fumarate versus dolute- cumulative exposure to tenofovir and other gravir ? abacavir/lamivudine in antiretroviral ther- antiretroviral agents. AIDS. 2012;26(7):825–31. apy-naive adults over 144 weeks. AIDS. 2015;29(18):2459–64. 12. Nkhoma ET, Rosenblatt L, Myers J, Villasis-Keever A, Coumbis J. Real-world assessment of renal and bone 23. Wohl DA, Orkin C, Doroana M, et al. Change in safety among patients with hiv infection exposed to vitamin D levels and risk of severe vitamin D deﬁ- tenofovir disoproxil fumarate-containing single- ciency over 48 weeks among HIV-1-infected, treat- tablet regimens. PLoS ONE. 2016;11(12):e0166982. ment-naive adults receiving rilpivirine or efavirenz in a Phase III trial (ECHO). Antivir Ther. 13. Cazanave C, Dupon M, Lavignolle-Aurillac V, et al. 2014;19(2):191–200. Reduced bone mineral density in HIV-infected patients: prevalence and associated factors. AIDS. 24. Sax PE, DeJesus E, Mills A, et al. Co-formulated elvite- 2008;22(3):395–402. gravir, cobicistat, emtricitabine, and tenofovir versus co-formulated efavirenz, emtricitabine, and tenofovir 14. Behrens G, Rijnders B, Nelson M, et al. Rilpivirine for initial treatment of HIV-1 infection: a randomised, versus efavirenz with emtricitabine/tenofovir diso- double-blind, phase 3 trial, analysis of results after proxil fumarate in treatment-naive HIV-1-infected 48 weeks. Lancet. 2012;379(9835):2439–48. patients with HIV-1 RNAB 100,000 copies/mL: week 96 pooled ECHO/THRIVE subanalysis. AIDS 25. Zolopa A, Sax PE, DeJesus E, et al. A randomized Patient Care STDS. 2014;28(4):168–75. double-blind comparison of coformulated elvite- gravir/cobicistat/emtricitabine/tenofovir disoproxil 15. Daar ES, Tierney C, Fischl MA, et al. Atazanavir plus fumarate versus efavirenz/emtricitabine/tenofovir ritonavir or efavirenz as part of a 3-drug regimen for disoproxil fumarate for initial treatment of HIV-1 initial treatment of HIV-1. Ann Intern Med. infection: analysis of week 96 results. J Acquir 2011;154(7):445–56. Immune Deﬁc Syndr. 2013;63(1):96–100. Infect Dis Ther (2018) 7:293–308 307 26. Bedimo R, Rosenblatt L, Myers J. Systematic review HIV-infected veterans starting combination of renal and bone safety of the antiretroviral regi- antiretroviral therapies. AIDS. men efavirenz, emtricitabine, and tenofovir diso- 2007;21(12):1579–89. proxil fumarate in patients with HIV infection. HIV clinical trials. 2016;17(6):246–66. 37. Gandhi NR, Tate JP, Rodriguez-Barradas MC, et al. Validation of an algorithm to identify antiretrovi- 27. Jullien V, Treluyer JM, Rey E, et al. Population ral-naive status at time of entry into a large, obser- pharmacokinetics of tenofovir in human immun- vational cohort of HIV-infected patients. odeﬁciency virus-infected patients taking highly Pharmacoepidemiol Drug Saf. 2013;22(9):1019–25. active antiretroviral therapy. Antimicrob Agents Chemother. 2005;49(8):3361–6. 38. LaFleur J, DuVall SL, Willson T, et al. Analysis of osteoporosis treatment patterns with bisphospho- 28. Gilead Sciences. Prescribing information for nates and outcomes among postmenopausal veter- VIREAD (tenofovir disoproxil fumarate). http:// ans. Bone. 2015;78:174–85. www.gilead.com/*/media/ﬁles/pdfs/medicines/ hiv/viread/viread_pi.pdf?la=en. Accessed 19 Oct 39. LaFleur J, Ginter T, Curtis JR, et al. A novel method for obtaining bone mineral densities from a dataset of radiology reports and clinic notes: natural lan- 29. European Commission Health and Food Safety. guage processing in a national cohort of post- Tybost Product Characteristics. Available at http:// menopausal veterans. Presented at the 2013 annual ec.europa.eu/health/documents/community- meeting of the American Society for Bone and register/2015/20150210131020/anx_131020_en. Mineral Research (ASBMR), Baltimore Convention pdf. Accessed 2 May 2016. Center, Baltimore, Maryland; 2013, October 4–7. http://www.asbmr.org/education/ 30. Hoetelmans R, Van Heeswijk R, Kestens Dea. Phar- AbstractDetail?aid=99cbf7b6-f3c3-49c4-9984- macokinetic interaction between the novel non- 0ba2962f1812. Accessed 13 Oct 2017. nucleoside reverse transcriptase inhibitor (NNRTI) TMC278 and tenofovir disoproxil fumarate (TDF) in 40. Austin PC. Balance diagnostics for comparing the healthy volunteers. In: 3rd IAS conference on HIV distribution of baseline covariates between treat- pathogenesis, treatment and prevention, Brazil, ment groups in propensity-score matched samples. July 2005, abstract WePe3.3C15. Stat Med. 2009;28(25):3083–107. 31. Zimmermann AE, Pizzoferrato T, Bedford J, Morris 41. Robins JM, Hernan MA, Brumback B. Marginal A, Hoffman R, Braden G. Tenofovir-associated acute structural models and causal inference in epidemi- and chronic kidney disease: a case of multiple drug ology. Epidemiology. 2000;11(5):550–60. interactions. Clin Infect Dis. 2006;42(2):283–90. 42. Maggi P, Montinaro V, Mussini C, et al. Novel 32. Bristol Myers Squibb, Gilead Sciences. Prescribing antiretroviral drugs and renal function monitoring information for ATRIPLA (efavirenz/tenofovir of HIV patients. AIDS Rev. 2014;16(3):144–51. disoproxil fumarate/emtricitabine). http:// packageinserts.bms.com/pi/pi_atripla.pdf. Accessed 43. Lepist EI, Zhang X, Hao J, et al. Contribution of the 19 Oct 2017. organic anion transporter OAT2 to the renal active tubular secretion of creatinine and mechanism for 33. Gilead Sciences. Prescribing information for COM- serum creatinine elevations caused by cobicistat. PLERA (rilpivirine/tenofovir disoproxil fumarate/ Kidney Int. 2014;86(2):350–7. emtricitabine). https://gilead.com/*/media/ﬁles/ pdfs/medicines/hiv/complera/complera_pi.pdf. 44. German P, Liu HC, Szwarcberg J, et al. Effect of Accessed 19 Oct 2017. cobicistat on glomerular ﬁltration rate in subjects with normal and impaired renal function. J Acquir 34. Gilead Sciences. Prescribing information for STRI- Immune Deﬁc Syndr. 2012;61(1):32–40. BILD (elvitegravir/cobicistat/tenofovir disoproxil fumarate/emtricitabine). http://www.gilead.com/ 45. Tong L, Phan TK, Robinson KL, et al. Effects of */media/ﬁles/pdfs/medicines/hiv/stribild/stribild_ human immunodeﬁciency virus protease inhibitors pi.pdf?la=en. Accessed 19 Oct 2017. on the intestinal absorption of tenofovir disoproxil fumarate in vitro. Antimicrob Agents Chemother. 35. Veterans Health Administration Ofﬁce of Research 2007;51(10):3498–504. and Development. Ofﬁce of Research & Develop- ment. Available from http://www.research.va.gov/ 46. Lepist EI, Phan TK, Roy A, et al. Cobicistat boosts default.cfm. Accessed 24 Feb 2017. the intestinal absorption of transport substrates, including HIV protease inhibitors and GS-7340, 36. Braithwaite RS, Kozal MJ, Chang CC, et al. Adher- in vitro. Antimicrob Agents Chemother. ence, virological and immunological outcomes for 2012;56(10):5409–13. 308 Infect Dis Ther (2018) 7:293–308 47. Ruane PJ, DeJesus E, Berger D, et al. Antiviral 50. Yin MT, Kendall MA, Wu X, et al. Fractures after activity, safety, and pharmacokinetics/pharmaco- antiretroviral initiation. AIDS. 2012;26(17):2175–84. dynamics of tenofovir alafenamide as 10-day monotherapy in HIV-1-positive adults. J Acquir 51. Triant VA, Brown TT, Lee H, Grinspoon SK. Fracture Immune Deﬁc Syndr. 2013;63(4):449–55. prevalence among human immunodeﬁciency virus (HIV)-infected versus non-HIV-infected patients in 48. Pinheiro Edos S, Antunes OA, Fortunak JM. A sur- a large U.S. healthcare system. J Clin Endocrinol vey of the syntheses of active pharmaceutical Metab. 2008;93(9):3499–504. ingredients for antiretroviral drug combinations critical to access in emerging nations. Antiviral Res. 52. Bristol Myers Squibb. Prescribing information for 2008;79(3):143–65. SUSTIVA (efavirenz). http://packageinserts.bms. com/pi/pi_sustiva.pdf. Accessed 19 Oct 2017. 49. Ofotokun I, Titanji K, Lahiri CD, et al. A single-dose zoledronic acid infusion prevents antiretroviral 53. Brown MJ, Mezuk B. Brains, bones, and aging: psy- therapy-induced bone loss in treatment-naive HIV- chotropic medications and bone health among older infected patients: a phase IIb trial. Clin Infect Dis. adults. Curr Osteoporos Rep. 2012;10(4):303–11. 2016;63(5):663–71.
Infectious Diseases and Therapy – Springer Journals
Published: Feb 28, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera