Observational Study Medicine ® OPEN Impact of protease inhibitors on the evolution of urinary markers Subanalyses from an observational cross-sectional study ∗ Anna Bonjoch, PhD, MDa, , Jordi Puiga, Nuria Pérez-Alvarez, PhD, MSca,b, Javier Juega, MDc, Patricia Echeverría, PhD, MDa, Bonaventura Clotet, PhD, MDa,d,e, Ramón Romero, PhD, MDc, J. Bonet, PhD, MDc, E. Negredo, PhD, MDa,e Abstract Kidney injury (defined as the presence of albuminuria, proteinuria, glycosuria [without hyperglycemia], hematuria, and/or renal hypophosphatemia) is an emerging problem in human immunodeficiency virus (HIV)-infected patients, although few data are available on the role of protease inhibitors (PIs) in this condition. To determine the time to kidney injury in a cohort of HIV-infected patients receiving a PI-containing regimen. We report the results of a subanalysis of a published cross-sectional study. The subanalysis included only patients receiving PI- containing regimens for more than 6 months (377 of the overall 970 patients). We determined associated factors and constructed receiver operating characteristic curves to estimate time to kidney injury depending on the PI used. The percentage of patients with kidney injury was 27.7% for darunavir, 27.9% for lopinavir, and 30% for atazanavir. Time to kidney injury was as follows: 229 days for atazanavir/ritonavir (area under the curve [AUC], 0.639; sensitivity, 0.89; specificity, 0.41); 332 days for atazanavir/ritonavir plus tenofovir (AUC, 0.603; sensitivity, 0.75; and specificity, 0.29); 318 days for nonboosted atazanavir (AUC, 0.581; sensitivity, 0.89; and specificity, 0.29); 478 days for lopinavir/ritonavir (AUC, 0.566; sensitivity, 0.864; and specificity, 0.44); 1339 days for lopinavir/ritonavir plus tenofovir (AUC, 0.667; sensitivity, 0.86; and specificity, 0.77); 283 days for darunavir/ ritonavir (AUC, 0.523; sensitivity, 0.80; and specificity, 0.261); and 286 days for darunavir/ritonavir plus tenofovir (AUC, 0.446; sensitivity, 0.789; and specificity, 0.245). The use of lopinavir/ritonavir without tenofovir was a protective factor (odds ratio = 1.772; 95%CI, 1.070–2.93; P = 0.026). For all PIs, the percentage of patients with kidney injury exceeded 27%, irrespective of tenofovir use. The longest time to kidney injury was recorded with lopinavir/ritonavir. These results demonstrate the need for renal monitoring, including urine samples, in patients receiving a PI-based regimen, even when tenofovir is not used concomitantly. Abbreviations: AUC = area under the curve, eGFR = estimated glomerular filtration rate, HIV = human immunodeficiency virus, PI = protease inhibitor, ROC = receiver operating characteristic, TDF = tenofovir disoproxil fumarate. Keywords: AIDS, HIV, kidney impairment, nephrotoxicity, protease inhibitors, proteinuria 1. Introduction Editor: Alejandro Vallejo. Funding/support: AB has received fees from ViiV, Jansen Cilag, AbbVie, and Renal abnormalities (defined as alteration in urine and blood Roche; JP has received fees from AbbVie; PE has received fees from ViiV, markers, i.e., presence of microalbuminuria, macroalbuminuria, Jansen Cilag, and AbbVie; EN and BC have received fees from ViiV, Merck, or proteinuria and/or the presence of glycosuria [without Jansen Cilag, Bristol-Meyers Squibb, Gilead, and AbbVie. hyperglycemia], hematuria, and/or hypophosphatemia, or an The remaining authors have no conflicts of interest to disclose. estimated glomerular filtration rate [eGFR] < 60 mL/min/ a Unitat VIH, Fundació Lluita contra la SIDA, Hospital Germans Trias i Pujol, 1.73 m2) are an emerging problem in the human immunodefi- Universitat Autònoma de Barcelona, Badalona, Spain, b Statistics and Operations ciency virus (HIV)-infected population owing to increasing life Research Department, Universitat Politècnica de Catalunya, Spain, c Nefrology department, Hospital Germans Trias i Pujol, Spain, d IrsiCaixa Foundation, Spain, expectancy and prolonged exposure to antiretroviral drugs. e Universitat de Vic-Universitat Central de Catalunya, Barcelona, Spain. Prevalence is as high as 30% in some series.[1] Kidney ∗ Correspondence: Anna Bonjoch, Unitat VIH, Fundació Lluita contra la SIDA, abnormalities usually develop slowly and silently, but can Hospital Germans Trias i Pujol, Universitat Autònoma de Barcelona, Ctra de progress to irreversible chronic disease. However, the most Canyet, s/n, 08916 Badalona, Barcelona, Spain (e-mail: abonjoch@flsida.org). relevant consequence of some alterations such as proteinuria or Copyright © 2016 the Author(s). Published by Wolters Kluwer Health, Inc. All altered GFR is probably the increased risk of cardiovascular rights reserved. events and death.[2,3] This is an open access article distributed under the Creative Commons Attribution-ShareAlike License 4.0, which allows others to remix, tweak, and build The relationship between antiretroviral agents and renal upon the work, even for commercial purposes, as long as the author is credited abnormalities has been established mainly for tenofovir dis- and the new creations are licensed under the identical terms. oproxil fumarate (TDF) and for its association with tubular Medicine (2016) 95:32(e4507) toxicity. TDF-related kidney injury is more likely when patients Received: 12 May 2016 / Received in final form: 13 July 2016 / Accepted: 14 receive concomitant protease inhibitors (PIs) owing to pharma- July 2016 cokinetic interactions.[4,5] One meta-analysis pointed to a decline http://dx.doi.org/10.1097/MD.0000000000004507 in kidney function associated with PIs, independently of 1 Bonjoch et al. Medicine (2016) 95:32 Medicine concomitant use of tenofovir.[6] Nonetheless, few data are All candidate predictor variables were entered in a logistic available on the impact of PIs on proteinuria and other signs of regression analysis, with kidney injury (as defined above) as the kidney injury. dependent variable. The variables included in the logistic The aim of the present study was to determine the role of PIs regression analysis were time on PIs with and without TDF, in kidney injury by evaluating time to appearance of renal time on TDF, time on nucleoside reverse-transcriptase inhibitors, abnormalities in a cohort of HIV-infected patients receiving a PI- current use of each PI, current use of TDF, age, time since containing regimen. diagnosis of HIV infection (it is defined as the date of the 1st positive HIV test result), time on antiretroviral therapy, age, current and nadir CD4 + cell count, undetectable viral load, 2. Methods hepatitis coinfection, use of nephrotoxic drugs, and body mass We performed a subanalysis of data collected from a cross- index. Variables with a P-value <0.20 were entered into the sectional study.[7] Briefly, the main study included 970 HIV- multivariate model. Backward selection (Wald test) was used to infected outpatients who consecutively attended our HIV Care build the multivariate model. Forward selection was subsequent- Unit and agreed to participate (from January 2011 to December ly applied to verify the stability of the results. Multicolinearity 2012). All the patients signed an informed consent document, was avoided by excluding from the multiple regression 2 and the Institutional Ethics Committee of the hospital and the covariates with a correlation coefficient >75%. local health authorities approved the study (code: EO-10-035). All analyses were performed using SPSS (version 15.0; SPSS For the subanalysis, we selected those patients who were Inc, Chicago, IL). receiving a PI-containing regimen for more than 6 months and had an eGFR > 60 mL/min/1.73 m2 (377 out of 970). The 3. Results exclusion criteria included baseline renal disease, any baseline Of the 377 patients included, 104 (29.4%) were receiving renal alterations (baseline defined as the time of the initiation of lopinavir/ritonavir-containing regimens, with and without TDF, in the current antiretroviral regimen), and extra renal causes of 28 and 76 cases, respectively; 177 patients (50%) were receiving glycosuria and hypophosphatemia. The objective was to compare darunavir/ritonavir, with and without TDF, in 68 and 99 cases; the impact of the 3 most commonly used PIs (darunavir, and 73 (20.6%) were receiving atazanavir/ritonavir, with and atazanavir, and lopinavir) on specific renal parameters by without TDF, in 30 and 43 cases, respectively. Only 23 patients estimating the predicted time to abnormality. We also defined were receiving nonboosted atazanavir without TDF in all cases. associated risk factors. The demographic and clinical characteristics of the patients The dose for boosted PIs was standard: atazanavir/ritonavir, and the time on the specific PI are shown in Table 1. Potentially 300/100 mg once daily; lopinavir/ritonavir, 400/100 mg twice nephrotoxic drugs were collected: angiotensin-converting en- daily; darunavir/ritonavir 800/100 mg once daily; and non- zyme inhibitors, angiotensin receptor blockers, cyclosporine, boosted atazanavir 400 mg once daily. TDF was administered at nonsteroidal antiinflammatory drugs, and tacrolimus. After a 300 mg once daily. median (interquartile range) time on the PI of 103 (62–144) The demographic and HIV-related data collected included time months, the percentage of patients fulfilling the criteria for kidney on antiretroviral treatment, time on tenofovir, time on PIs, and injury (as defined above) was 27.7% for darunavir (27.9% with time on each specific current PI. Blood and urine samples were collected under fasting conditions. Kidney injury was defined as the presence of micro- or Table 1 macroalbuminuria or proteinuria, that is, an albumin/creatinine Patient characteristics. ratio of >30 mg/g and/or protein/creatinine ratio of >200 mg/g Overall population N = 377 and/or the presence of glycosuria (without hyperglycemia), hematuria, and/or hypophosphatemia (<2.5 mg/dL).[8] Gender (male), n, % 282 (75%) Concomitant use of TDF was defined as exposure to the drug Ethnicity (white), n, % 358 (95%) Age, years 47.45 (43–51.8) for more than 6 months. Patients were considered not to have Diabetes, n, % 11 (3%) received TDF if they had never received it, had received it for less Hypertension, n, % 32 (8.6%) than 3 months, or if they had discontinued it for more than 4 Smoking, n, % 204 (54%) months considered the mean time for reversibility of TDF-related Hepatitis B/C coinfection, n, % 119 (31.6%) toxicity, as is described in a previous study on the reversibility of Potentially nephrotoxic concomitant therapy, n, % 47 (12.5%) TDF-related toxicity.[9] Time since diagnosis of HIV infection, years 17.3 (12–21.5) Current CD4 + T-cell count, cells/mL 553 (396–779) Nadir CD4 + T-cell count, cells/mL 181 (92–276.5) 2.1. Statistical analyses Suppressed viral load, n, % 323 (85.6%) Continuous variables were expressed as median (interquartile Time on ARV treatment, months 171 (108–209.6) range), and categorical variables were expressed as frequency Time on PIs, months 103.2 (62–144) Time on TDF, months 32.3 (17.8–49.7) (percentage), unless stated otherwise. Time on ATZ (without TDF), months 17.5 (7.3–28.3) Receiver operating characteristic (ROC) curves were used to Time on ATZ + TDF, months 32.5 (8.4–45.2) calculate a cut-off for time on treatment according to the presence Time on LPV (without TDF), months 28.9 (15.6–41.9) or absence of kidney injury. The time for the decision threshold Time on LPV + TDF, months 44.1 (25.7–54.7) was selected by setting a minimum sensitivity of 80% and the Time on DRV (without TDF), months 16.1 (9.6–25.4) largest specificity available. An area under the curve (AUC) of Time on DRV + TDF, months 16.5 (9.9–33.5) 0.5 was considered a random allocation; an AUC greater than Data are expressed as the median (IQR) unless otherwise indicated. ARV = antiretroviral, ATZ = 0.5 was considered acceptable as a sign of possible relationship. atazanavir, DRV = darunavir, IQR = interquartile range, LPV = lopinavir, n = number of patients, PI = Statistical significance was set at 0.05. protease inhibitor, TDF = tenofovir. 2 Bonjoch et al. Medicine (2016) 95:32 www.md-journal.com and 30.3% without TDF), 27.9% for lopinavir (25% with and AUC, 0.523) and 286 days for darunavir/ritonavir plus TDF 29% without TDF), and 30% for atazanavir (20% with and (sensitivity, 0.789; specificity, 0.245; and AUC, 0.446). 45% without TDF). No statistically significant differences were The only factor in the logistic regression analysis that was observed when we compared each PI with and without TDF. associated with a decrease in the risk of renal abnormalities was Similarly, no significant differences were observed when we current use of lopinavir/ritonavir without TDF (odds ratio = compared all the PIs with TDF and without TDF. 1.772; 95%CI, 1.070–2.93; P = 0.026). No other correlations Overall, altered protein/creatinine and/or albumin/creatinine were established. ratios were the most frequent alteration detected (65% of all alterations). The ROC curve analysis (Fig. 1) showed an estimated time to 4. Discussion kidney injury of 229 days for atazanavir/ritonavir when it was used without TDF (sensitivity, 0.89; specificity, 0.41; and AUC, We determined predicted time to kidney injury to analyze the 0.639), 332 days when it was administered with TDF (sensitivity, impact of various PIs on renal alterations in HIV-infected patients 0.75; specificity, 0.29; and AUC, 0.603), and 318 days when with an eGFR > 60 mL/min/1.73 m2. Abundant recent data are nonboosted atazanavir was administered without TDF (sensitiv- available for PIs and renal abnormalities[10,11]; however, few data ity, 0.89; specificity, 0.29; and AUC, 0.581). have been reported on the impact of these agents on kidney Time until kidney injury was 478 days for lopinavir/ritonavir injury, defined as proteinuria, hematuria, glycosuria, and/or without TDF (sensitivity, 0.864; specificity, 0.44; and AUC, hypophosphatemia.[12] 0.566) and 1339 days for lopinavir/ritonavir plus TDF The importance of an early detection of abnormal biomarkers (sensitivity, 0.86; specificity, 0.77; and AUC, 0.667). related with renal damage in the 1st stages, before the advent of Estimated time to kidney injury was 283 days for darunavir/ clinical symptoms, lies in the possible link with the reversibility of ritonavir without TDF (sensitivity, 0.80; specificity, 0.261; and the damage. The safety profile of TDF has been well studied, in Lopinavir/ritonavir Atazanavir/ritonavir Darunavir/ritonavir 1,0 1,0 1,0 0,8 0,8 0,8 Sensitivity 0,6 0,6 Sensitivity 0,6 Sensitivity 0,4 0,4 0,4 0,2 0,2 0,2 0,0 0,0 0,0 0,0 0,2 0,4 0,6 0,8 1,0 0,0 0,2 0,4 0,6 0,8 1,0 0,0 0,2 0,4 0,6 0,8 1,0 1 - Specificity 1 - Specificity 1 -Specificity 1,0 1,0 1,0 0,8 0,8 0,8 0,6 0,6 Sensitivity 0,6 Sensitivity Sensitivity 0,4 0,4 0,4 0,2 0,2 0,2 0,0 0,0 0,0 0,2 0,4 0,6 0,8 1,0 0,0 0,0 0,2 0,4 0,6 0,8 1,0 0,0 0,2 0,4 0,6 0,8 1,0 1 - Specificity 1 -Specificity 1 - Specificity Figure 1. ROC curve is using the top discriminatory features of time on different PIs exposition to classify the individuals as having developed kidney injury or not having developed kidney injury. (A) Boosted PIs without TDF; (B) boosted PIs plus TDF. PI = protease inhibitors, ROC = receiver operating characteristic, TDF = tenofovir). 3 Bonjoch et al. Medicine (2016) 95:32 Medicine contrast, scarce data are available about renal injury with PI outcome, and no data were recorded for previous antiretroviral strategies. history or previous toxicity. Since we included mainly white The 1st remarkable finding of this study is the longer time to men, our results cannot be extrapolated to other populations. kidney injury with lopinavir (478 and 1339 days without and Statistical methods that include time, such as a Cox model, could with TDF, respectively). Time to kidney injury was much shorter not be used owing to the cross-sectional design of the study. for the remaining PIs (229 and 332 days for atazanavir and 283 Finally, the ROC curve analysis did not show correlations in all and 286 days for darunavir, without and with TDF, respectively), cases (e.g., darunavir), probably because the number of patients although the AUC was less accurate in the case of darunavir. included or the duration of the evaluation was insufficient. These results could reflect lower renal toxicity with lopinavir than To conclude, our study identified a considerable proportion of with the other PIs, as corroborated by the results of the logistic patients with signs of kidney injury while receiving PI-based regression. regimens. The longest time to kidney injury was recorded for Also interesting was the considerably high proportion of lopinavir/ritonavir. Although further longitudinal studies are patients (30%, similar for all 3 agents) who developed kidney needed to confirm the long-term consequences of these findings, injury, even when TDF was not used concomitantly. This result our results have a key clinical implication: give the high seems to contradict the results of the logistic regression analysis prevalence of renal abnormalities and the recognized silent with respect to lopinavir and the ROC curves with respect to the nature of the initial phases of kidney injury it is important to time taken to develop kidney injury in the case of lopinavir. stress the need to monitor renal parameters, including urine Nevertheless, compared with the other PIs, exposure to lopinavir values, even in patients not receiving concomitant tenofovir. was longer, although the number of renal abnormalities was similar to that of the other drugs. Our results agree with those of previous studies in that the Acknowledgements worst result (defined as the shortest time to development of The authors thank Thomas O’Boyle for editorial assistance. alterations in urinary markers) was with atazanavir/ritonavir, followed by lopinavir/ritonavir.[10,11,13] Few data are available on the use of darunavir/ritonavir.[14] Nevertheless, these studies References assessed changes in eGFR, incidence of chronic kidney disease, or [1] Winston J, Deray G, Hawkins T, et al. Kidney disease in patients with both, but not the presence of proteinuria or other signs of kidney HIV infection and AIDS. Clin Infect Dis 2008;47:1449–57. [2] French AL, Gawel SH, Hershow R, et al. Trends in mortality and causes injury. Atazanavir in particular has the potential to yield of death among women with HIV in the United States: a 10-year study. crystalline precipitate in urine, leading to crystalluria, tubuloin- J Acquir Immune Defic Syndr 2009;51:399–406. terstitial nephritis, and acute or chronic kidney disease.[15] In the [3] Wyatt CM, Hoover DR, Shi Q, et al. Pre-existing albuminuria predicts case of lopinavir/ritonavir, the mechanism that leads to a decline AIDS and non-AIDS mortality in women initiating antiretroviral in the eGFR remains unclear, since the results vary depending on therapy. Antivir Ther 2011;16:591–6. [4] Goicoechea M, Liu S, Best B, et al. Greater tenofovir-associated renal the equation used, although ritonavir could play a role in the function decline with protease inhibitor-based versus nonnucleoside inhibition of the transport protein of tenofovir.[10] In addition, reverse-transcriptase inhibitor-based therapy. J Infect Dis 2008;197: lopinavir-related acute interstitial nephritis was recently 102–8. reported.[16] Darunavir has only been associated with asymp- [5] Young J, Schafer J, Fux CA, et al. Renal function in patients with HIV starting therapy with tenofovir and either efavirenz, lopinavir or tomatic nephrolithiasis,[17] although it is the drug with shortest atazanavir. AIDS 2012;26:567–75. follow-up. [6] Bagnis CI, Stellbrink HJ. Protease inhibitors and renal function in Previous studies confirm increased toxicity of TDF when patients with HIV infection: a systematic review. Infect Dis Ther administered concomitantly with boosted PIs.[4,18] This effect is 2015;4:15–50. related to inhibition of multidrug resistance-associated protein 4 [7] Bonjoch A, Juega J, Puig J, et al. High prevalence of signs of renal damage despite normal renal function in a cohort of HIV-infected patients: by ritonavir, the subsequent increase in intracellular tenofovir evaluation of associated factors. AIDS Patient Care STDS 2014; concentrations, and the greater inhibition of DNA-Y polymerase 28:524–9. and depletion of mitochondrial DNA.[19,20] Nevertheless, in our [8] Montanes Bermudez R, Gracia Garcia S, Perez Surribas D, et al. study, time to kidney injury was longer in patients receiving Consensus document. Recommendations on assessing proteinuria during the diagnosis and follow-up of chronic kidney disease. Nefrologia lopinavir or atazanavir concomitantly with TDF. These findings 2011;31:331–45. should be interpreted with caution owing to the small sample size [9] Bonjoch A, Echeverria P, Perez-Alvarez N, et al. High rate of reversibility in both groups (only 27 and 28 patients with atazanavir and of renal damage in a cohort of HIV-infected patients receiving tenofovir- lopinavir, respectively, were also receiving TDF) and to the containing antiretroviral therapy. Antiviral Res 2012;96:65–9. process for selection of patients, which focused on PIs. The [10] Mocroft A, Kirk O, Reiss P, et al. Estimated glomerular filtration rate, chronic kidney disease and antiretroviral drug use in HIV-positive selection criteria excluded patients receiving TDF but not PIs, patients. AIDS 2010;24:1667–78. patients with less than 6 months’ exposure to TDF, patients who [11] Ryom L, Mocroft A, Kirk O, et al. Association between antiretroviral discontinued the drug early (3 months), and patients who had exposure and renal impairment among HIV-positive persons with interrupted therapy for more than 4 months. The selection normal baseline renal function: the D:A:D study. J Infect Dis 2013; 207:1359–69. process excluded subjects with early toxicity associated with [12] Calza L, Vanino E, Magistrelli E, et al. Prevalence of renal disease within TDF, although it could lead us to underestimate the impact of the an urban HIV-infected cohort in northern Italy. Clin Exp Nephrol drug on kidney markers and thus restrict conclusions about TDF 2014;18:104–12. as a separate agent. [13] Wyatt CM, Winston JA, Malvestutto CD, et al. Chronic kidney disease in Our study has several limitations. Its cross-sectional design HIV infection: an urban epidemic. AIDS 2007;21:2101–3. [14] Rockwood N, Mandalia S, Sirokosta J, et al. A comparative analysis could lead us to overestimate the prevalence of kidney injury, of risk factors associated with efavirenz, atazanavir/r, darunavir/r, because only 1 blood parameter and 1 urine parameter were lopinavir/r, and renal impairment. In: 17th Annual Conference of the determined. In addition, no follow-up was available to assess British HIV Association (BHIVA). Bournemouth, 2011. 4 Bonjoch et al. Medicine (2016) 95:32 www.md-journal.com [15] Izzedine H, M’Rad MB, Bardier A, et al. Atazanavir crystal nephropathy. [18] Reynes J, Cournil A, Peyriere H, et al. Tubular and glomerular AIDS 2007;21:2357–8. proteinuria in HIV-infected adults with estimated glomerular filtration [16] Chughlay MF, Njuguna C, Cohen K, et al. Acute interstitial nephritis rate>/ = 60 ml/min per 1.73 m2. AIDS 2013;27:1295–302. caused by lopinavir/ritonavir in a surgeon receiving antiretroviral [19] Hall AM, Hendry BM, Nitsch D, et al. Tenofovir-associated kidney postexposure prophylaxis. AIDS 2015;29:503–4. toxicity in HIV-infected patients: a review of the evidence. Am J Kidney [17] de Lastours V, Ferrari Rafael De Silva E, Daudon M, et al. High levels of Dis 2011;57:773–80. atazanavir and darunavir in urine and crystalluria in asymptomatic [20] Izzedine H, Harris M, Perazella MA. The nephrotoxic effects of HAART. patients. J Antimicrob Chemother 2013;68:1850–6. Nat Rev Nephrol 2009;5:563–73. 5