Switching from Abacavir/Lamivudine to Tenofovir DF/Emtricitabine Reduces Biomarkers of Inflammation: A Randomized Proof of Concept Study

The spectrum of morbsidity and mortality among HIV positive persons is now more commonly characterized by long-term complications such as atherosclerotic cardiovascular disease (CVD) [1]. As a result, HIV clinical management now emphasizes modifying traditional and HIV-specific risk factors that contribute to CVD risk. HIV factors, in part, include toxicities from individual components of antiretroviral therapy (ART) [2]. Exposure to the nucleoside reverse transcriptase inhibitor (NRTI) abacavir has been associated with increased CVD event risk in some studies, but not in others [3-10]. Few studies have prospectively explored the potential adverse effects of abacavir among virally suppressed patients, data are needed to determine whether tailoring individual components of ART (e.g., stopping abacavir) is a viable strategy to reduce CVD risk.


Introduction
The spectrum of morbsidity and mortality among HIV positive persons is now more commonly characterized by long-term complications such as atherosclerotic cardiovascular disease (CVD) [1]. As a result, HIV clinical management now emphasizes modifying traditional and HIV-specific risk factors that contribute to CVD risk. HIV factors, in part, include toxicities from individual components of antiretroviral therapy (ART) [2]. Exposure to the nucleoside reverse transcriptase inhibitor (NRTI) abacavir has been associated with increased CVD event risk in some studies, but not in others [3][4][5][6][7][8][9][10]. Few studies have prospectively explored the potential adverse effects of abacavir among virally suppressed patients, data are needed to determine whether tailoring individual components of ART (e.g., stopping abacavir) is a viable strategy to reduce CVD risk.
To explore potential CVD-related toxicity from abacavir exposure, we measured plasma biomarkers from stored specimens as part of a proof of concept pilot study assessing the effects of switching from an abacavir-to a tenofovir disoproxil fumarate (DF)-based ART regimen.

Study design
Participants were recruited from Hennepin County Medical Center (HCMC) and Abbott Northwestern (ANW) HIV Clinics in Minneapolis, Minnesota, USA.We enrolled medically stable HIVpositive participants with viral suppression (HIV RNA <48 copies/ mL) for ≥3 months, taking fixed-dose abacavir/lamivudine and one additional antiretroviral medication. Exclusion criteria included: a creatinine clearance <50 mL/min (estimated by Cockcroft-Gault formula), a contraindication to taking tenofovir DF, and the presence of resistance mutations to abacavir, tenofovir DF, lamivudine or emtricitabine. The randomized intervention consisted of: a) maintaining their current regimen ('ABC group'), or b) switching to tenofovirDF/emtricitabine-based ART ('TDF group'), while maintaining the third component of ART in place. The study goal was to explore and quantify any treatment effect on vascular inflammation and endothelial dysfunction, with additional analysis of inflammatory biomarkers planned a priori, that might inform additional analyses of stored specimens from larger clinical trials [11]. Planned sample size for this pilot phase was based on ability to recruit. All participants underwent informed consent. The study protocol was approved by ethics committees at both sites and registered at ClinicalTrials.gov (NCT#00998582).

Laboratory methods
Biomarker levels, from cryo preserved plasma specimens collected at month 1 and 6, were measured at the Laboratory for Clinical Biochemistry Research at the University of Vermont using stored plasma. Inflammation and coagulation was assessed via levels of high sensitivity C-reactive Protein (hsCRP) by NB™II nephelometer N Antiserum to Human CRP (Siemens Diagnostics), interleukin-6 (IL-6) by Chemiluminescent Sandwich ELISA (R&D Systems), and D-dimer by immunoturbidometric methods (Diagnostica Stago). Vascular inflammation and endothelial dysfunction was indirectly assessed via the following biomarkers: soluble intercellular adhesion moledule-1 (sICAM) with ELISA from Parameter Human Immunoassay (R&D Systems), von Willebrand Factor (vWF) by Stago STA-R analyzer (Stago Diagnostics), serum thrombomodulin (sTM) by Chemiluminescent Sandwich ELISA (R&D Systems), and asymmetric dimethylarginine (ADMA) by competitive enzyme-linked immunosorbent assay (Euroimmun).All samples were analyzed blinded to treatment group.

Statistical methods
Baseline medians were compared utilizing Wilcoxon tests; proportions were compared with chi-squared or Fisher's exact tests. Differences between randomized groups were estimated with generalized linear models adjusted for baseline biomarker level, with the primary comparison at month 6. Biomarker levels were log etransformed before analysis; results are reported as percent differences between treatment groups by exponentiating log e -transformed values. Additional analyses adjusted for baseline 10-year Framingham Risk Score (FRS) for coronary heart disease event and estimated glomerular filtration rate (eGFR). A global assessment composite measure [12] was considered by grouping biomarkers together based on: inflammation/ coagulation (hsCRP, IL-6, and D-dimer) and vascular inflammation (sICAM, ADMA, vWF, sTM) assessments. We expected a similar treatment effect on the individual markers within each composite, so the intent in generating composite scores was to reduce measurement variability since the biomarkers within each panel represent overlapping biology. With the "rank-sum" method biomarker level values in each composite were ranked and then summed by treatment group. Models were used to compare the sum of the ranks between treatment groups. Two-sided p-values and 95% confidence intervals (CIs) are cited, with no adjustments for multiple comparisons. Analyses were performed using SAS version 9.2 (SAS Institute) and R version 2.9.

Results
Twenty-seven participants were screened and randomized to either stay on abacavir/lamivudine-(n=13) or switch to tenofovir DF/ emtricitabine-(n=14) based ART. Among randomized participants, 27 were seen at month one and 26 at month six visits. Baseline characteristics and biomarker levels are presented in Table 1. Individual CVD risk factors did not differ by randomized allocation, but the Framingham Risk Score (FRS) for 10-year risk of a coronary heart disease event tended to be higher among the abacavir (8.5%) versus tenofovir (4.7%) group (p=0.052). The third antiretroviral medication was typically a PI (n=18) or NNRTI (n=7), but two participants were taking an integrase inhibitor (raltegravir).

Discussion
In this proof of concept study, we randomized HIV-positive participants with viral suppression taking fixed-dose abacavir/ lamivudine-based ART to remain on their current regimen or switch the nucleoside component to tenofovir DF/emtricitabine. Participants switching to tenofovir DF had significantly lower hsCRP levels and a lower inflammation/coagulation rank score after 6 months, when compared to those that stayed on abacavir. These pilot data raise the possibility that there may be greater systemic inflammation among patients on stable abacavir regimens. Though the clinical implications of these findings are unclear, the plasma biomarkers and rank score were chosen based on epidemiologic data demonstrating associations between these biomarkers and risk for CVD and all-cause mortality among HIV positive patients [13][14][15][16].
Our study provides additional randomized data among virally suppressed patients and are consistent with findings from the larger AIDS Clinical Trials Group (ACTG) study A5224s of starting ART [17]. In this study of 244 participants, hsCRP levels were higher 24 and 96 weeks after starting abacavir/lamivudine compared with those starting tenofovir DF/emtricitabine. A similar effect was reported in a nonrandomized comparison within the Strategies for Management of Anti-Retroviral Therapy (SMART) study, where patients receiving abacavir had higher baseline hsCRP and IL6 (interleukin-6), compared with patients receiving other study NRTIs [18]. In another non-randomized comparison of 57 ART-treated participants with suppressed viral load, hsCRP levels increased 3 months after switching to an abacavir containing regimen but no increase was seen after 12 months or among those who switched to a tenofovir DF-based regimen [19].
Conversely, several other studies have failed to show any differential effect on inflammatory biomarker outcomes between abacavir and other NRTIs [20][21][22][23]. In the Multicenter AIDS Cohort Study (MACS) and Women's Interagency HIV Study (WIHS) cohort, none of the inflammatory biomarkers studied (hsCRP, IL-6 and D-dimer) were higher among participants starting abacavir versus other NRTIs [20]. The BICOMBO study compared 80 ART-treated virally suppressed participants randomized to switch their NRTI backbone to either abacavir/lamivudine or tenofovir/emtricitabine, and found no difference in inflammatory (including hsCRP levels) and vascular inflammation biomarkers [21]. Null biomarker findings were also reported in the Simplification of antiretroviral therapy with Tenofovir-Emtricitabine or Abacavir-Lamivudine (STEAL) study, a randomized study of switching to abacavir-versus tenofovir DFbased ART regimens [22]. Importantly, the STEAL study also included a virally suppressed target population and the clinical endpoint analyses did find an association between abacavir use and CVD risk. This discrepancy in the STEAL study may be explained by alternative mechanisms for the potential abacavir-mediated CVD risk, other than inflammation, include inducing platelet hyper-reactivity [24,25]. Ultimately, disparate findings from multiple studies to date highlight the need for clinical endpoint studies (e.g., of CVD events) to determine if there is a differential effect of abacavir on CVD risk.
Our study was limited by its small size, exploratory nature, and use of plasma biomarker outcomes. The risk for both type 1 and type II statistical error mean findings should be interpreted with caution. The study design is also unable to control for any carry-over effect (e.g., abacavir-related abnormalities that persist after the drug is stopped). Despite these issues, our findings highlight a potential proinflammatory effect of abacavir exposure. Larger randomized studies are needed to verify whether switching from abacavir-based ART reduces inflammation and corresponding changes in clinical risk for CVD.

Funding
Gilead Pharmaceuticals, investigator initiated and sponsored project. Study investigators were not employed or contracted by Gilead to conduct this study, and had full control over conduct of the study, including study design, data analysis, and the content of this manuscript.