Renal effects of novel antiretroviral drugs

Abstract

Chronic kidney disease (CKD) is a critical comorbidity for patients living with HIV, with an estimated prevalence between 2.4 and 17%. Such patients are increasingly affected by diseases associated with ageing, including cardiovascular disease and CKD, and the prevalence of risk factors such as smoking and dyslipidaemia is increased in this population. Proteinuria is also now recognized as a common finding in individuals living with HIV. While combination antiretroviral (ARV) treatments reduce CKD in the HIV-infected population overall, some ARV drugs have been shown to be nephrotoxic and associated with worsening renal function. Over the last few years, several highly efficacious new ARV agents have been introduced. This brief review will look at the novel agents dolutegravir, raltegravir, elvitegravir, cobicistat, tenofovir alafenamide fumarate and atazanavir, all of which have been licensed relatively recently, and describe issues relevant to renal function, creatinine handling and potential nephrotoxicity. Given the prevalence of CKD, the wide range of possible interactions between HIV, ARV therapy, CKD and its treatments, nephrologists need to be aware of these newer agents and their possible effect on kidneys.

INTRODUCTION

Chronic kidney disease (CKD) is increasingly important as a critical comorbidity for patients living with HIV: the life expectancy of appropriately treated individuals living with HIV is now similar to that of the general population [1]; the prevalence of patients with fully suppressed HIV on combined antiretroviral (ARV) treatment is increasing and the HIV population is ageing. Such patients are increasingly exposed to, and often more affected by, diseases associated with ageing, including cardiovascular disease and CKD, and the prevalence of risk factors such as smoking and dyslipidaemia is increased in this population [2, 3]. There is also growing evidence for an ageing phenotype that may be driven by a pro-inflammatory state in patients infected with HIV, and of accelerated ‘immunosenescence’, both of which may contribute to this burden of comorbidity. Proteinuria is also increasingly recognized as a common finding in individuals living with HIV in the absence of known kidney disease, although first described >20 years ago, and overall the estimates of the prevalence of CKD in this population are between 2.4 and 17% [4–6]. Not only are individuals living with HIV subject to the traditional risk factors for CKD, but high HIV viraemia and low CD4 counts (usually in untreated or poorly adherent patients) are associated with acute kidney injury (AKI), CKD and progression to end-stage renal failure [7]. While cART has been shown to reduce CKD in the HIV-infected population overall (mostly by rapidly controlling viral infection and associated AKI from HIV associated nephropathy, opportunistic infection or immune complex glomerular damage), some ARV drugs have been shown to be nephrotoxic and associated with worsening renal function.

The protease inhibitors indinavir, lopinavir and atazanavir have all been associated with renal stone formation and while tenofovir disoproxil fumarate (TDF) and atazanavir have been shown to cause acute tubular injury and tubulo-interstitial nephritis [8]. Renal complications following TDF exposure range from mild proximal tubulopathy to a fulminant Fanconi syndrome; TDF has also been associated with a progressive decline in renal function, although the mechanisms are unclear. Additionally, many ARV drugs have been shown to interfere with creatinine handling in the kidney, classically reducing tubular secretion, leading to misdiagnosis of renal dysfunction from measures of serum creatinine and estimated glomerular filtration rate (eGFR) alone [9, 10].

Over the last 3–4 years there has been a further increase in the number of available antiretroviral (ARV) drugs. Given the prevalence of CKD, the wide range of possible interactions between HIV, ARV therapy, CKD and its treatments, nephrologists need to be aware of these newer agents and their possible effect on the kidneys. This brief review will look at the novel ARV agents dolutegravir (DTG), raltegravir (RTG), elvitegravir (EVG), cobicistat (COBI), tenofovir alafenamide fumarate (TAF) and atazanavir (ATZ), all of which have been licensed relatively recently, and describe issues relevant to renal function, creatinine handling and potential nephrotoxicity (Table 1). The renal effects of older ARV drugs have been reviewed in detail elsewhere [11].

Table 1.

Possible renal effects of novel ARV drugs

ARV drug	Renal effect
Raltegravir	Inhibition of tubular secretion of serum creatinine by organic cation transporters Rhabdomyolysis Possible reduction in true GFR
Elvitegravir	None reported
Dolutegravir	Inhibition of tubular secretion of serum creatinine by organic cation transporters
Cobicistat	Inhibition of tubular secretion of serum creatinine by MATE 1 Increase in serum concentration of TDF with potential risk for enhancement of TDF tubular toxicity if co-administered (not proven)
Tenofovir alafenamide fumarate	No renal toxicity reported
Atazanavir	Renal stones Acute tubular injury (interstitial nephritis)

MATE1, multidrug and toxin extruder protein 1; TDF, tenofovir disopoxil fumarate.

Ritonavir (often included in combination therapy as a booster) partially inhibits MATE1 and may therefore increase serum TDF.

RALTEGRAVIR

Raltegravir (RTG) is one of several new integrase inhibitors, drugs that block the integration of HIV genetic material into human DNA, an essential step for viral replication. It has demonstrated efficacy in both treatment-naive and -experienced patients in combination with other ARV drugs. RTG is rapidly absorbed, with a half-life of 7–12 h, and a low propensity for drug–drug interactions [12]. This is due to its metabolism by glucuronidation, primarily by uridine glucuronosyl transferase 1A1 (UGT1A1), a low-affinity, high-capacity pathway, and also due to a lack of involvement with CYP450 [13]. RTG is 83% plasma-protein bound and excreted in both urine and faeces (32 and 51%, respectively), but dose adjustment is not needed in renal impairment [13]. RTG is only minimally removed during haemodialysis [14].

Trial data suggest RTG may cause an increase in serum creatinine without affecting renal excretory or tubular function. The SAILING study evaluated renal adverse effects in both RTG- and DTG-exposed individuals as part of a non-inferiority study [15]. Both groups exhibited similar small increases in creatinine at Week 2 (approximately 4 and 10 µmol/L in the RTG and DTG treatment arms, respectively) that remained stable for the duration of the 48-week trial. This finding was echoed in the SPRING-2 study (patients randomized to RTG or DTG with either TDF/emtricitabine or abacavir/lamivudine for 96 weeks), which also showed a similar small, stable increase in the mean serum creatinine of 8.2 µmol [with a decrease in mean creatinine clearance (CrCl) of 9.3 mL/min] but no evidence of the development of proteinuria in RTG-treated patients after 48 weeks of therapy [16]. Given the similarities between their effects on CrCl, it has been suggested that RTG may have an effect on tubular function similar to that of DTG. However, in a smaller study of 30 patients, individuals switched from efavirenz to RTG had increases in both serum creatinine and serum cystatin C, with a decrease in GFR (cystatin) of 8.50 mL/min/1.73² (SD 11.04) in the switch group when measured by cystatin C clearance [17]. This suggests a possible genuine reduction in renal function rather than an effect on creatinine handling by renal tubular cells. Furthermore, RTG has no effect on organic cation transporter 2 (OCT2), unlike DTG.

A significant problem with many of these studies is the failure to measure true GFR (e.g. using iothalamate or ethylenediamine tetraacetic acid clearance). Overall, however, it seems likely that RTG may cause a very small increase in serum creatinine within the first few weeks of treatment without a subsequent decline in renal function in the medium term.

There are several case reports of RTG associated with episodes of rhabdomyolysis, in one patient causing significant AKI [18]. However, it is not at all clear that RTG was the cause in any of these patients. In the latter case, the patient developed dialysis-requiring AKI, but with a peak creatinine kinase of only 8947 u/L, while taking RTG, TDF and emtricitabine (which he had been on for 2 years) and non-steroidal anti-inflammatory agents, clorazepate and trazadone. He made a full recovery. Therefore, it is unclear if this was a causal link, but it may be prudent to be aware in patients who might be at particular risk and ensure patients seek medical advice if they develop muscle pain or dark urine.

ELVITEGRAVIR

Elvitegravir (EVG) is an integrase inhibitor that undergoes extensive primary metabolism by hepatic and intestinal cytochrome CYP3A and subsequent glucuronidation by UGT1A1/3. As such, it is always administered with a CYP3A inhibitor, either COBI or ritonavir (RTV, which act as pharmacokinetic boosters to facilitate once-daily dosing [19]. It is included in the fixed-dose combination tablet Stribild, alongside COBI, emtricitabine and TDF. The unboosted half-life of EVG is approximately 3 h, but this is increased to 9 h with the co-administration of RTV or COBI. It is >99% protein bound in plasma and predominantly excreted in faeces (94%), with the remainder (6%) excreted in urine [13, 19]. As its metabolism is primarily hepatic, EVG does not require dose adjustment in renal impairment [20].

Phase 3 trials comparing EVG to RTG showed no difference in renal function between the two drugs, but a median decrease in eGFR from baseline to 96 weeks of 10.8 mL/min/1.73 m² in the EVG arm compared with 11.7 mL/min/1.73 m² for RTG, with no documented formal renal adverse effects [21]. The booster in this study was RTV, not COBI. Given that all these patients had normal renal function at baseline, and that eGFR is significantly variable in patients with normal renal function, the meaning of these data is unclear. There remains a possibility that as with RTG, EVG has a minor effect on physiological creatinine handling. No other renal effects of EVG have been reported.

DOLUTEGRAVIR

DTG is the third integrase inhibitor to be approved (in January 2014 in Europe) and is available as an individual agent for once-daily dosing or in a fixed drug combination with abacavir and lamivudine, branded under the trade name Triumeq [22]. Similar to RTG, it is predominantly metabolized by UGT-1A1. It is extensively protein bound (>99%) and excreted primarily in faeces, with <1% excreted unchanged in the urine [23]. Therefore, DTG does not need to be dose adjusted, even in severe renal impairment (CrCl < 30 mL/min) [24].

In vitro data have shown DTG to inhibit OCT2 on the basolateral side of proximal tubule cells [10]. OCT2 substrates include endogenous substances such as creatinine and serotonin alongside drugs including metformin, ranitidine and propranolol [25]. Therefore, in a similar fashion to cimetidine and other OCT2 inhibitors, DTG can block the tubular uptake of creatinine from the blood, leading to increases in serum creatinine and decreased eGFR or CrCl, without changing true GFR [26]. In the SPRING-2 study, the mean estimated CrCl decreased 16.5 mL/min in the DTG group compared with 5.4 mL/min in the RTG group at the end of 48 weeks. This change occurred during the first 2–3 weeks of treatment and was non-progressive. There was no documented increase in the urinary protein:creatinine ratio in either the DTG- or RTG-treated patients and there were no patient withdrawals due to renal adverse events [16]. The VIKING study of additional DTG in patients failing to suppress HIV on another regimen showed a similar pattern of modest, non-progressive creatinine increase in patients (12.38 µmol/L for both cohorts) treated with DTG, occurring within 2 weeks after initiation of therapy and plateauing at 4 weeks [27]. It has been demonstrated that these increases in serum creatinine, seen soon after starting DTG, do not correspond with a decrease in true GFR when measured using iohexol [28].

Overall therefore, DTG induces a noticeable increase in serum creatinine following initiation of therapy due to non-pathologic tubular blockade of creatinine secretion through the inhibition of OCT2. No true renal adverse actions of DTG have been reported.

COBICISTAT

COBI is a pharmaco-enhancer with no ARV activity that acts to increase the concentration of certain drugs through potent inhibition of CYP3A. COBI has greater CYP450 enzyme specificity than RTV, another pharmaco-enhancer, with no effect on CPY1A2, CYP2C9 or CYP2C19, suggesting that there should be fewer drug–drug interactions than observed with RTV [29, 30]. It is licensed for use as a booster with ATZ and darunavir, as well as in combination with EVG/TDF/emtricitabine (FTC) in the once-daily, fixed combination drug Stribild. COBI undergoes predominantly hepatic metabolism and does not require dose adjustment in renal impairment.

In phase 3 trials of Stribild compared with Atripla (efavirenz/TDF/FTC), five patients had renal adverse effects that required treatment discontinuation prior to Week 48 and all were in the treatment arm containing COBI. Two patients were reported to have an increased serum creatinine with no further data available in the literature, two developed renal failure and the fifth patient had Fanconi syndrome [31]. Four of these patients had evidence of proximal tubulopathy, defined as a combination of hypophosphataemia, glycosuria or proteinuria. A median eGFR decrease of 14.3 mL/min/1.73 m² (interquartile range 4.3–24.2) was noted in the COBI/EVG/TDF/FTC treatment arm at 4 weeks and remained stable thereafter. In weeks 48–144 of the same trial, three further patients were withdrawn from the trial due to renal adverse events: all had isolated creatinine increases with no evidence of proximal tubulopathy. In non-inferiority studies comparing Stribild with RTV/ATZ/TDF/FTC there were four withdrawals from the study due to renal adverse events at 144 weeks, with no cases of proximal tubulopathy observed and a mean eGFR decrease of 14.1 mL/min/1.73 m² at 48 weeks [32, 33]. Renal safety was assessed in patients with an eGFR ≥50 mL/min/1.73 m² and the same non-progressive creatinine increase was noted but was not associated with a decrease in cystatin C-based eGFR [20].

COBI is a known inhibitor of MATE1 (multidrug and toxin extruder protein 1), a transporter on the apical side of the proximal tubule cells responsible for the efflux of creatinine into the urine [34]. The effect of COBI on true GRF and CrCl has been investigated using iohexol-measured GFR and showed that COBI decreased CrCl but had no effect on measured GFR. In the same study, serum creatinine levels returned to baseline after stopping COBI [35]. The possible renal effects of Stribild may be related to the actions of COBI on TDF handling. Co-administration of TDF and COBI results in a 25–30% increase in the area under the curve (AUC) and peak serum concentration (C_max) of TDF [10, 36]. Since the inhibition of MATE1 should not affect TDF [37], this increase could potentially be explained by previous in vitro studies showing inhibition of intestinal transporter P-glycoprotein by COBI [38]. TDF is a substrate for P-glycoprotein-mediated efflux and its inhibition increases the intestinal absorption of TDF [39]. This increased exposure to tenofovir, particularly in combination with decreased GFR from anther cause, may explain why in phase 3 trials the rates of patient withdrawal from renal adverse events were higher with Stribild than comparators.

Overall, patients experiencing an increase in serum creatinine or a decrease in eGFR when started on Stribild require careful monitoring. An early, small, non-sustained increase in serum creatinine is likely to be an effect of COBI on creatinine secretion alone, and should not require drug cessation or further investigation, but must be distinguished from TDF proximal tubulopathy or AKI (see Table 2).

Table 2.

Distinguishing ARV drug induced nephrotoxicity from effects on tubular cell drug transporters

Inhibition of tubular creatinine secretion	Tubular cell damage or other causes of AKI
Increase in serum creatinine of up to 20 µmol/L or decrease in eGFR of <24 mL/min	No limit to increase in serum creatinine
Change in serum creatinine or eGFR always occurs within 4 weeks of drug therapy	Change in serum creatinine or eGFR can occur at any time, often months, after drug initiation
Change in serum creatinine or eGFR is non-progressive (no further decline after initial change)	Change may be progressive (but not always)
No proteinuria or haematuria	Low-level proteinuria is common (often in the absence of albuminuria) and may have glycosuria or aminoaciduria
No change in blood pressure	No change in blood pressure usually, but may increase
No change in serum electrolytes	May have full or partial Fanconi syndrome with phosphaturia, low serum phosphate, mild acidosis and orthoglycaemic glycosuria

TENOFOVIR ALAFENAMIDE FUMARATE

Tenofovir, in the form of TDF, has been in use for 15 years and remains a widely use, highly efficacious, first-line agent in treating patients with HIV, with >7.5 million person-years of use in various combinations. From a renal perspective, the major potential toxicity is tubular dysfunction, ranging from low-level proteinuria to a full-blown Fanconi syndrome (rare), but small, non-progressive increases in serum creatinine are also observed after initiating therapy and increasingly TDF has been associated with a progressive sustained slow decline in renal function although the mechanisms are unclear [6]. Pathological findings are usually of acute tubular damage, often with mitochondrial abnormalities in tubular cells. Most patients recover, usually fully, on TDF withdrawal.

TAF is a novel prodrug of tenofovir currently undergoing phase 3 trials that may offer improved renal safety over TDF [40]. TAF is more stable in plasma and is metabolized intracellularly by cathepsin A, an enzyme that is highly expressed in lymphoid tissues. Therefore, TAF can achieve higher intracellular levels of the active moiety tenofovir diphosphate, with lower levels of circulating tenofovir when compared with TDF. This more targeted treatment could potentially result in fewer renal complications [41, 42].

In comparisons between two fixed-dose combination drugs, EVG/COBI/FTC/TAF versus EVG/COBI/FTC/TDF (Stribild), the initial results suggest a more favourable renal side effect profile for TAF than TDF. There were four renal discontinuations in the TDF arm compared with none in the TAF group and there was a lesser reduction in eGFR in the TAF group when compared with the TDF group, 5.5 and 10.1 mL/min/1.73 m², respectively (P = 0.041), with no cases of proximal tubulopathy in either treatment group [40]. TAF has been shown to be safe in mild to moderate renal impairment (30–69 mL/min), and a switch from a mixture of TDF- and non-TDF-containing regimens to fixed-dose EVG/COBI/FTC/TAF in a further study resulted in no change in eGFR but significant improvements (reduction) in proteinuria [40]. These studies were among the first to directly randomize patients with documented impaired renal function and to measure true GFR in addition to eGFR.

If these results are confirmed in ongoing studies, it is likely that TAF would replace TDF, especially given the potential improvement in mild tubular dysfunction seen on switching patients from TDF- to TAF-containing regimens, although the true clinical significance of these changes remains to be defined.

ATAZANAVIR

ATZ is a widely used HIV protease inhibitor that is usually taken with a boosting dose of RTV. Classically, ATZ is prescribed in combination with a tenofovir- or abacavir-containing nucleoside/nucleotide ‘backbone’. ATZ is metabolized mainly by cytochrome P450 3A4 and is primarily excreted in faeces (79%), with only 7% of ATZ excreted unchanged in the urine, although the solubility of ATZ decreases with an increase in the alkalinity of urine [43]. Common combinations include ATZ/RTV and Truvada (TDF/FTC) or ATZ/RTV and Kivexa (abacavir/lamivudine). The association of the older protease inhibitor indinavir and renal stone formation has been well described, with up to 30% of exposed individuals developing indinavir crystalluria. More recently, RTV boosted ATZ (ATZ/r) has also been shown to be associated with an increased risk of nephrolithiasis, with a single-centre study suggesting a hazard ratio of 10.44 (P ≤ 0.001) when compared with other protease inhibitors [44]. High concentrations of ATZ have been detected in urine [45], and in some reports, stones have been retrieved that have been shown to be composed entirely of ATZ. No clear risk factors have been established for nephrolithiasis in patients treated with ATZ. In one study, the duration of ATZ/r therapy was longer in patients with crystalluria, and this is consistent with several previous reports showing an association between the duration of ATZ exposure and nephrolithiasis. Other proposed risk factors include CKD [44], hyperbilirubinaemia and a history of urolithiasis [46]. Given the relatively high frequency of renal stones in the general population, it is always important to characterize stones biochemically, and it should be noted that ATZ stones are radiolucent.

Alongside its association with nephrolithiasis, patients on ATZ have been shown to have significant reductions in eGFR. One study showed a decrease in eGFR of 10.4 mL/min/1.73 m² (range 6.7–13.1) over 12 months in patients treated with ATZ. Higher rates of tubular damage have also been reported, with 14.1% of patients showing evidence of proximal tubular dysfunction at the end of 12 months [47].

CONCLUSION

In summary, the newer, highly effective, ARV medications have a wide range of effects on renal function, without any current evidence for significant nephrotoxicity, although there is a lack of any data on long-term use and of any histopathological findings in patients taking these agents, or even surrogate markers of renal damage. It is important to recognize that many of these novel agents will effect estimates of GFR due to modification of creatinine handling in the tubule, and that different clinical studies often use different estimating formulae for GFR, and sadly very few actually measure true GFR. The agents DTG and COBI have been shown particularly to interfere with tubular creatinine handling, causing increases in serum creatinine with no actual nephrotoxicity or effects on true GFR. TAF has been designed to have improved renal safety due to its more targeted mechanism of action and pharmacokinetic profile, but this will only be confirmed once it is in widespread use. ATZ has been associated with acute tubular injury and nephrolithiasis and its use in patients with known CKD should be closely monitored. The introduction of these newer agents will require a close working relationship between HIV and renal physicians in order to appropriately investigate (but not over-investigate) and manage patients who manifest changes in serum creatinine.