icon-    folder.gif   Conference Reports for NATAP  
  17th CROI
Conference on Retroviruses
and Opportunistic Infections
San Francisco CA
February 16-19, 2010
Back grey_arrow_rt.gif
The Kidney at CROI
  Christina M. Wyatt, MD
Assistant Professor, Medicine/ Nephrology
Mount Sinai School of Medicine
New York, NY
Kidney disease and other non-AIDS complications in the aging HIV population were highlighted at this year's meeting. Two oral abstracts framed much of the discussion of kidney disease. Phase 2 study results were reported for Gilead's new fixed-dose single-tablet regimen "Quad" [#58; see report on 2/16/10]. There was a significant decline in estimated glomerular filtration (eGFR) over 24 months among participants randomized to the study drug, which appeared to be primarily related to the novel boosting agent GS-9350. In a separate analysis of healthy volunteers randomized to GS-9350 or placebo, there was a rapid and reversible decline in eGFR among those randomized to GS-9350, but no change in true GFR as directly measured by iohexol clearance. This result suggests that GS-9350 may interfere with the tubular secretion of creatinine, similar to trimethoprim and cimetidine. In vitro studies are planned to confirm this interaction. Although this appears to be a benign effect, these laboratory abnormalities may complicate the management of patients with marginal kidney function.
The EuroSIDA investigators reported an association between exposure to several individual ART agents and the risk of chronic kidney disease (CKD) in this large, predominantly Caucasian cohort [#107]. CKD was defined as a confirmed eGFR < 60, or a 25% decline in eGFR for participants with baseline eGFR < 60. In adjusted analysis, the incidence of CKD was significantly increased with longer cumulative exposure to tenofovir (incidence rate ratio, IRR 1.16), indinavir (IRR 1.12) or atazanavir (IRR 1.21), and marginally increased with longer exposure to lopinavir (IRR 1.08). This increase in risk was observed in participants who were actively using the agent in question, and returned to reference levels immediately after discontinuation of the protease inhibitors. Compared to patients who were never exposed to tenofovir, patients who discontinued tenofovir still had a significant increase in CKD events in the 12 months following discontinuation. Rather than an ongoing adverse effect of tenofovir, the authors hypothesized that this residual risk may reflect increased awareness of tenofovir nephrotoxicity and a low clinical threshold for discontinuing tenofovir in patients with a decline in eGFR, regardless of etiology. (etiology or cause for decline I eGFR could be due to other reasons like diabetes).
The observed association between atazanavir and CKD has not been previously reported, and raises the possibility that atazanavir may have a nephrotoxic effect similar to the well-known effect of indinavir, which is known to cause acute and chronic interstitial nephritis in addition to stones. Although this is biologically plausible, it is also possible that the observed association merely reflects residual confounding. Because atazanavir may be prescribed preferentially to patients with cardiovascular risk factors, who are also at increased risk for CKD, atazanavir could be a marker of increased CKD risk rather than a causative agent. The observed associations should be validated in other populations with particular attention to potential confounders, and future studies are needed to explore possible mechanisms. Nonetheless, this study is the first to explore the epidemiologic association of multiple individual ART agents with CKD. Unlike previous studies, which have singled out a small number of suspected nephrotoxins for study, the EuroSIDA investigators explored the impact of all available ART agents. Of note, they did not have adequate power to exclude an effect of newer ART agents on CKD risk. (note from Jules: in the oral talk the presenter in his summary said: AIDS, non-AIDS malignancies and coinfection with HCV were also independently associated with CKD).
CROI:Chronic kidney disease and exposure to ARTs in a large cohrt with long-term follow-up: EuroSida - (02/20/10)
Epidemiology of CKD and Acute Kidney Injury in HIV-infected Cohorts
In addition to EuroSIDA, several other large cohorts reported data on the incidence and predictors of CKD. Among ART-naïve participants in the Swiss HIV Cohort Study, cumulative ART exposure was associated with a small increase in eGFR over a median follow-up of 2.5 years, while tenofovir exposure was associated with a decline in eGFR [#740]. The Aquitaine Cohort investigators also identified an increased risk of CKD (eGFR<60 for at least 3 months) associated with cumulative tenofovir exposure [#741]. Of note, these cohorts did not report data on the association of other individual ART agents with eGFR decline or CKD. In the US Military HIV Natural History Study, HIV diagnosis prior to the ART era was associated with a 6-fold increase in CKD incidence (eGFR < 60 for at least 3 months), suggesting a benefit of ART in this relatively young cohort with free access to care [#738]. These investigators did not explore the association with individual ART agents or regimens. In contrast to previous studies demonstrating strong racial disparities, there was no increased risk of CKD among African-American participants. In a competing risks analysis, there was no increase in the risk of ESRD or death, although the number of events was low. The authors hypothesized that equal access to care, low rates of IDU, and selection of participants for excellent health at baseline may have eliminated some of the disparity in risk observed in other cohorts. Other traditional risk factors for CKD were also investigated in each of these cohorts; however, the only risk factors identified in all three cohorts included older age and female gender. While older age is an established CKD risk factor, it is likely that the association with female gender reflects a bias in the performance of GFR estimates.
Investigators from the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) reported a biphasic relationship between ART exposure and eGFR slope in ART-naïve patients [#735, revised title "HAART is associated with improved kidney function despite early GFR reductions"]. Among participants with baseline eGFR > 60, the eGFR slope was stable or improved during the first year of ART among those with no exposure to tenofovir or ritonavir, while there was a suggestion of an early eGFR decline among those exposed to both tenofovir and ritonavir. The eGFR slope was stable after year 1 regardless of regimen. Of note, the CNICS investigators reported a notable difference in eGFR slopes depending on which GFR estimating equation was used.
Although estimating equations are far more sensitive than serum creatinine alone for the detection of CKD, none of the available estimates have been validated in HIV. In a cross-sectional study of ART-treated individuals, Guaraldi and colleagues explored differences between the prevalence of eGFR < 60 as defined using three different GFR estimating equations: Cockcroft-Gault, MDRD, and CKD-EPI [#704]. The new CKD-EPI equation has been shown to perform more accurately in the normal or near-normal GFR range in the general population, and in the current study significantly fewer individuals were categorized as having eGFR<90 by CKD-EPI compared to the other equations (level of agreement, Kappa 0.29 for MDRD and 0.38 for Cockcroft-Gault). The correlation between eGFR<90 and higher coronary artery calcium score, a surrogate marker for cardiovascular disease, was also highest when eGFR was defined by CKD-EPI. Future studies should investigate which GFR estimate is the most accurate in HIV-infected individuals when compared to a gold standard measure of GFR.
Because currently available GFR estimates were derived in North American populations, there is a particularly urgent need to validate these equations in African populations with a high prevalence of HIV infection and high background risk for CKD. Data from ICAP's MTCT-Plus program demonstrated clinically significant variability in the prevalence of decreased eGFR in a multi-national African cohort depending on the GFR estimate used [#733]. The proportion of participants who would require ART dose modification for an eGFR < 50 varied from 3% by Cockcroft-Gault to 2.1% by 4-variable MDRD. The overall prevalence of eGFR < 60 was approximately 8%, similar to published estimates from Kenya and much lower than estimates from Nigeria. In multivariate analysis, the strongest predictor of decreased eGFR was female gender, most likely reflecting a bias in the performance of the GFR estimates in this population.
While the majority of CKD abstracts focused on mild to moderate changes in eGFR, a small retrospective cohort study from Spain evaluated clinical events among HIV-infected patients with established ESRD [#739]. Although the 1-year survival rate of 95% is better than previously reported in US studies, 5-year survival and access to kidney transplant were both lower among HIV-infected dialysis patients compared to matched HIV-negative controls. ART was associated with improved survival among HIV-infected dialysis patients, consistent with published data. In contrast to previous studies, hemodialysis was associated with improved outcomes compared to peritoneal dialysis, although it was not possible to adjust for all potential confounders in this small study.
In the ATHENA cohort, Zhang and colleagues reported an increased risk of non-AIDS events among participants with CD4 < 200 [#503]. The risk of a clinically defined acute or chronic renal event was increased more than 11-fold among participants with a CD4< 200 compared to those with a CD4 > 500. In contrast to SMART, these investigators did not observe any increase in the risk of non-AIDS events with ART interruption, although the viral load was suppressed during 89% of the follow-up time.
A retrospective analysis of clinical data from King's College Hospital in London also identified low CD4 as a risk factor for acute renal failure (ARF), defined as a 40% decline in eGFR to a nadir eGFR below 60 [#734]. Approximately 7% of patients experienced at least one episode of ARF during a median follow-up of 2.5 years, with the majority of events occurring in hospitalized patients. In adjusted analysis, the only factors independently associated with ARF were CD4 <100 and eGFR <75. While non-black race, IDU, and HCV co-infection were associated with ARF in the unadjusted analysis, these factors were not independently associated with ARF in this racially diverse cohort with a low prevalence of HCV co-infection (10%). Because it is easier for a patient with low baseline eGFR to meet the study criteria, the authors plan to test the robustness of their findings using an alternative definition of ARF.
Tenofovir Toxicity
Several small studies investigated the incidence of tenofovir nephrotoxicity in previously understudied patient populations. In a multi-center cohort study including 102 HIV-Hepatitis B co-infected individuals treated with tenofovir-containing regimens, de Vries-Sluijs and colleagues described a relatively small but statistically significant decline in estimated GFR over a median follow-up of 53 months [#631]. Consistent with previous cohort studies in HIV-infected individuals, the GFR decline was most prominent immediately following tenofovir initiation. Three co-infected patients discontinued tenofovir secondary to a decline in kidney function. In two small pediatric studies, there was no evidence of tenofovir toxicity among 27 stable pediatric patients who were switched to a tenofovir-containing ART regimen [#870] or among 20 seronegative children exposed to tenofovir-containing ART in utero [#925].
Two small studies evaluated the relationship between markers of tubular dysfunction and markers of bone turnover. In a small study of patients with overt proximal tubular dysfunction or isolated hyperphosphaturia, evidence of bone turnover was more common and more pronounced among individuals with overt tubular dysfunction as defined by the abnormal urinary excretion of at least 3 of the following: phosphate, protein, uric acid, or glucose [#748]. The authors also observed a high prevalence of vitamin D deficiency and hyperparathyroidism. To further evaluate the impact of vitamin D deficiency and elevated PTH on the relationship between tubular dysfunction and bone turnover, these investigators explored the correlation between renal phosphate handling, vitamin D, and PTH in a cross-sectional study [#749]. Vitamin D deficiency was associated with slightly lower urinary phosphate losses, while PTH was not correlated with renal phosphate handling. While these findings suggest that vitamin D deficiency and hyperparathyroidism do not mediate the proposed relationship between tubular dysfunction and increased bone turnover, they may still modify that relationship.
The field has not yet established a consensus panel of tubular injury markers for use in research or clinical practice. In a convenience sample of 206 HIV-infected patients with simultaneous measurement of random urine protein: creatinine and urine albumin: creatinine, Samarawickrama and colleagues investigated the utility of these readily available measures to distinguish tubular injury from glomerular disease [#737]. Patients with an elevated urine protein:creatinine but with normal urine albumin:creatinine ratio were presumed to have tubular proteinuria, while patients with elevations in both were considered to have glomerular proteinuria. Tubular proteinuria was associated with lower plasma phosphorus and higher fractional excretion of phosphate, as well as with current or previous exposure to tenofovir. In a small subgroup of 8 patients who underwent clinically indicated kidney biopsy, the biopsy findings were consistent with the designation based on proteinuria and albuminuria. If these findings are validated in a prospective study with biopsy-confirmed diagnoses, the combination of urine albumin and protein: creatinine ratios may provide an additional non-invasive approach to distinguish tenofovir toxicity from comorbid CKD.
Pathogenesis of ART- and HIV-related Kidney Disease
The pathogenesis of tenofovir toxicity and other forms of HIV and ART-related kidney disease was the subject of several studies. Pushpakom and colleagues investigated the role of a novel tubular transporter in the same family as MRP4, a known transporter for tenofovir, and MRP2, which has been linked to tenofovir toxicity in small genetic studies but which is not known to be directly involved in tenofovir transport [#742]. In a series of in vitro experiments, these investigators demonstrated that tenofovir is a substrate for MRP7. In subsequent analysis, the single nucleotide polymorphism (SNP) rs9349256 in the gene encoding MRP7 was independently associated with the development of proximal tubular injury. Because of the limitations of small SNP analyses, this association should be validated in larger studies.
In a small proof of concept study, Gupta and colleagues compared the urine cytokine profile from HIV-infected patients with and without proteinuria, suggesting a role for local inflammation mediated by the NFkB pathway [#736]. Urinary MCP-1 and RANTES were higher in the setting of proteinuria, and MCP-1 remained independently associated with proteinuria and albuminuria after adjusting for CD4 and viral load. Proteomic analysis was also performed in a small number of urine specimens to demonstrate feasibility. Because glomerular or tubular injury may allow the non-selective loss of proteins in urine, additional studies are needed to identify proteins that may be involved in the pathogenesis of HIV-related kidney disease.
In vitro studies from investigators at Mount Sinai explored the mechanism and downstream effects of direct HIV infection of the kidney. While local infection of the kidney is required for the development of HIV-associated nephropathy (HIVAN), it is not known how HIV infects the kidney in the absence of CD4 or traditional co-receptors. Using a fluorescently tagged HIV, Chen and colleagues described the efficient cell-cell transfer of HIV from infected T cells to renal tubular cells in co-culture [#743]. The rapid internalization of viral RNA by tubular cells was sufficient for the expression of HIV genes. Snyder and colleagues investigated the downstream effects of HIV Vpr expression in renal tubular cells, building on previous work showing that Vpr induces apoptosis in tubular cells [#744]. In a series of in vitro experiments in a human tubular cell line, these investigators identified the ERK-Caspase 8 pathway as a novel pathway for Vpr-induced apoptosis in this cell type. These studies provide additional insight into the pathogenesis of HIVAN, and future studies are needed to investigate whether local HIV infection and viral gene expression may also contribute to the pathogenesis or progression of other kidney diseases in HIV-infected individuals.