icon-    folder.gif   Conference Reports for NATAP  
  19th Conference on Retroviruses and
Opportunistic Infections
Seattle, WA March 5 - 8, 2012
Back grey_arrow_rt.gif
Renal Update CROI 2012
  Samir K. Gupta, MD, MS
Division of Infectious Diseases
Indiana University School of Medicine

Several key issues related to renal disease and toxicities were highlighted at CROI this year. These primarily centered on ART-related renal effects, differing estimates of renal function, and the outcomes related to renal dysfunction.

Dolutegravir and Cobicistat - Effects on the Kidney

Perhaps the most intriguing and concerning ART-related renal toxicities discussed this year were perhaps not really toxicities at all. Two new integrase inhibitors, dolutegravir and cobicistat-boosted elvitegravir, were shown to be overall quite safe and highly effective in achieving virologic suppression when compared to efavirenz and atazanavir-ritonavir in ART-naïve patients. However, there were appreciable increases in serum creatinine values with the use of these two new antiretrovirals. Both dolutegravir and cobicistat are thought to increase serum creatinine values by inhibiting the elimination pathway of muscle-derived creatinine through the proximal renal tubules. By doing so, more creatinine naturally builds up in the circulation, giving the appearance of worsening renal function. However, this inhibition, which is similar to what is found with high dose trimethoprim or cimetidine, is not a real toxicity but simply a manifestation of these drugs' effects on normal renal physiology.

Sax et al (Abstract 101) presented the 48 week findings on an ART-naïve, double-blind, randomized trial comparing efavirenz(EFV)/emtricitabine(FTC)/tenofovir(TDF) vs. the new 'Quad' single tablet dose of elvitegravir(EVG)/cobicistat/FTC/TDF; there were approximately 350 participants in each study arm. Serum creatinine increased eventually by 0.14mg/dL more in the Quad arm compared to the EFV/FTC/TDF arm by week 48; this increase was seen as early as week 2 of the study. This increase in serum creatinine translated to a creatinine clearance reduction of 14.3mL/min with the Quad compared to 3.0mL/min with EFV/FTC/TDF by week 48. Drug discontinuations due to renal events were 1.4% vs. 0% in the Quad vs. EFV/FTC/TDF arm, respectively.

DeJesus et al (Abstract 627) presented a companion trial assessing the Quad pill in ART-naïve patients. In this other 48 week, double-blind, randomized trial, the Quad pill was compared to FTC/TDF+atazanavir(ATV)/ritonavir(r). The Quad treatment resulted in an average additional increase in serum creatinine of 0.04mg/dL over what was found with FTC/TDF+ATV/r. This additional increase appeared by week 2, similar to what was found by Sax et al, and persisted through week 48. However, only 0.3% of the roughly 350 participants in each arm had treatment discontinued due to renal adverse events. Thus, the results of these two trials demonstrate that cobicistat leads to an early physiologic increase in serum creatinine (and a reduction in creatinine clearance) by week 2 of use which persists, but does not worsen, through one year of treatment.

The effects of dolutegravir on renal function were demonstrated by Stellbrink et al (Abstract 102LB) in the ART-naïve, double-blind, phase 2b, dose-finding, randomized trial of 3 different doses of dolutegravir vs. efavirenz (with either FTC/TDF or 3TC/ABC as the nucleoside backbone). Increases of 0.10-0.15mg/dL of serum creatinine were seen across all arms and did not differ by FTC/TDF or 3TC/ABC use. Increases in urine albumin-creatinine ratios were small and similar across arms. However, this study was relatively small, and the larger phase 3 studies will likely shed more light on the effects of dolutegravir on the kidney.

Although these increases in serum creatinine with dolutegravir or cobicistat likely do not reflect true nephrotoxicity, interpreting these changes in serum creatinine will be difficult and poses several questions. If a patient develops an increase in serum creatinine while on one of these drugs, does this truly just reflect a benign physiologic effect? If the change in serum creatinine is small and appears early, is this sufficient to satisfy the HIV caregiver and the patient that there is nothing to worry about? How sure is it that a change in serum creatinine in the general HIV population with other comorbidities is not perhaps due to tenofovir, atazanavir, or another etiology of renal injury? Perhaps alternative markers of renal function will help settle this issue. Certainly additional research is warranted.

More on Renal Dysfunction Associated with Tenofovir and Atazanavir

A few papers presented at this year's CROI once again focused on tenofovir's and atazanavir's effects on kidney function. Ryom et al presented data from the D:A:D study (Abstract 865), which also included participants of the EuroSIDA cohort, that assessed the effects of various antiretrovirals on the development of renal dysfunction. In patients with normal renal function, defined as an estimated creatinine clearance of 90mL/min using the Cockcroft-Gault equation, were assessed for development of either a confirmed creatinine clearance of ≤70mL/min or ≤60mL/min. Creatinine clearance rates after drug discontinuation were also determined. In this analysis, cumulative use of TDF, ATV/r, and lopinavir/r were each associated with ensuing creatinine clearances ≤70mL/min; however, only lopinavir/r and abacavir use were associated with creatinine clearances ≤60mL/min. Overall though, the increased risks of renal dysfunction associated with the use of these drugs were quite small. Interestingly, and perhaps reassuringly, renal function appeared to improve significantly once these drugs were discontinued.

In abstract 870, Bonjoch et al performed an analysis of their clinical cohort from Spain. In this study, a majority of patients who developed nephrotoxicity with the use of tenofovir recovered quickly once this drug was stopped. However, the length of time to see this recovery may be over one year in some patients. In addition, nearly a third of their cases did not fully recover. Interestingly, having a higher nadir CD4 count at time of TDF initiation and better CD4 count at time of drug discontinuation were both associated with renal function recovery.

Although data from randomized trials are more informative in demonstrating causality for drug-associated benefits and risks, these large observational studies do corroborate other previously published cohort studies. Certainly, more study is certainly needed to determine which patients are most likely to improve after discontinuation of an offending antiretroviral and which patients perhaps should not be placed on these drugs in the first place in order to prevent irreversible renal injury.

Which Renal Function Equation Should We Use?

Because directly measuring renal function by determining the renal clearance of an IV or SQ exogenous agent is not practical in clinical practice, practitioners use estimating equations based on endogenous markers, such as serum creatinine or cystat and which include other clinical factors (e.g. age, race, sex, and weight). There are several such renal estimating equations in clinical use. But it is unclear which one is most accurate in the HIV-infected population. In attempts to answer this important question, several studies compared these renal estimating equations to 'gold standard' exogenous renal clearance measurements. In Abstract 862, Praditpornsilpa et al found that the MDRD equation, when using a Thai-specific adjustment factor, was the most accurate equation assessed. In Abstract 863, Lucas et al found that in a clinical cohort in Baltimore, that the MDRD equation, the Cockcroft-Gault (CG) equation, and the CKD-EPI equation all performed relatively well, although CG was heavily biased by the body mass index of the patient studied. This is not unanticipated given that CG, but not the other equations, includes a weight factor in its equation. Margolick et al in Abstract 864 examined factors that affected glomerular filtration measurements using iohexol clearance methods in the MACS cohort. These investigators found that race, history of AIDS, and hepatitis C infection, but not HIV infection status nor use of any antiretroviral (including tenofovir), influenced these measurement results. Overall, these important studies are increasing our knowledge of how best to estimate renal function in the clinic and how different these estimates are based on the patient population actually studied.

However, the 'holy grail' of estimating equations for renal function in the HIV-infected population has not yet been established. This equation should be accurate in patients both on and not on ART as well account for differences in age, sex, body size, and race/ethnicity. Hopefully such a 'one size fits all' equation will be developed soon.

Consequences and New Predictors of HIV-Related Renal Disease

Several published studies have demonstrated the reduced renal function and markers of renal injury (proteinuria and albuminuria) variably predict progression to end stage renal disease, cardiovascular disease, progression to AIDS or new AIDS-defining event, and all-cause mortality. Several CROI abstracts revolved around these long-term consequences of renal impairment.

Lucas et al in Abstract 866 performed a unique study by comparing the outcomes of patients with renal dysfunction as estimated using a creatinine-based estimate (CKD-EPI equation) with serum cystatin C. Cystatin C as a measure of GFR (higher values indicate worse glomerular function) has been shown in the general population to be a better prognostic indicator of long-term outcomes than creatinine-based estimates. Using the SMART trial population, cystatin C was a better predictor of overall mortality, cardiovascular events, and opportunistic infections than CKD-EPI using serum creatinine. So perhaps cystatin C is a better measure of overall renal dysfunction and its influence on important clinical outcomes.

In Abstract 869, Lima et al corroborated several studies by showing once again that albuminuria, defined as a spot urine albumin to creatinine ratio of 30mg/g, was independently associated with all cause mortality in a long-term study of patients from British Columbia. The unique aspect of this study was that the risk of death was much stronger in those patients followed for more than 10 years compared to those followed for less than 10 years. With the HIV patient population fortunately living much longer, assessing long-term outcomes, like that done in this study, will be increasingly important.

Polymorphisms in both the APOL1 and MYH9 genes have been found to explain most of the risk of HIV-associated nephropathy (HIVAN) in those of African ancestry. Limou et al examined these polymorphisms and the risk of eventual renal impairment in SMART trial participants of African ancestry in Abstract 873. These investigators found that specific polymorphisms in these genes are indeed associated with lower renal function at baseline and with worse renal function over time. Thus, genetic influences may play a role in renal disease in the HIV-infected population.

Finally, Neuhaus et al in Abstract 830/O-271 examined the influence of the inflammatory markers IL-6 and hsCRP and the coagulation marker D-dimer on the eventual development of chronic kidney disease (CKD) in patients participating in SMART and the ESPRIT studies. Using a case-control design, baseline levels of these markers were not associated with development of CKD. This was an unexpected result given that these markers were associated with cardiovascular disease and overall mortality in SMART. Thus, it cannot be assumed that inflammation and coagulation definitely influence all outcomes in those with HIV.

These studies further support the need and importance to include renal health measures in all new studies of long-term clinical outcomes in those with HIV.