icon-    folder.gif   Conference Reports for NATAP  
  Conference on Retroviruses
and Opportunistic Infections (CROI)
February 22-25, 2016, Boston MA
Back grey_arrow_rt.gif
The Kidney Report - CROI 2016
  Feb 22-24, Boston
Christina Wyatt MD
Associate Professor, Medicine/ Nephrology
Icahn School of Medicine at Mount Sinai New York, NY
Renal safety of tenofovir alafenamide (TAF)
The results of ongoing and completed trials of TAF continue to suggest a potential for improved renal safety, including the 48-week results of a randomized switch study comparing FTC/TAF versus FTC/TDF in adults previously suppressed on FTC/TDF (Oral abstract 29). Similar to prior studies comparing E/C/F/TAF to E/C/F/TDF, there were statistically significant improvements in CrCl and bone mineral density among participants randomized to FTC/TAF, along with a decline in urinary excretion of the low molecular weight proteins retinol binding protein (RBP) and beta2-microglobulin (B2M). Study drug was discontinued in one participant with pre-existing proteinuria in the FTC/TDF arm and no participants in the FTC/TAF arm.
In 72-week follow-up of an open label study of E/C/F/TAF in virologically suppressed adults with CrCl 30-69 at enrollment, study drug was discontinued in 2% of participants because of decreased CrCl (Abstract 680). All five participants had baseline CrCl<50 ml/min and at least one traditional risk factor for chronic kidney disease (CKD). In participants who switched from a TDF-containing regimen, albuminuria and proteinuria declined significantly, while bone mineral density increased. More importantly, the prevalence of clinically relevant albuminuria and proteinuria also decreased significantly following the switch to E/C/F/TAF.
The top-line 48-week results of two randomized trials comparing E/C/F/TAF to E/C/F/TDF in participants with CrCl > 60 ml/min were previously published in Lancet and presented in support of regulatory applications. Two post hoc analyses used the data from this combined study cohort to provide additional information on the renal safety of TAF. The first analysis evaluated the safety of TAF in participants at increased risk of CKD as assessed by the D:A:D CKD risk score (Abstract 681). The stratified analysis suggests that the favorable effect of TAF on CrCl and urinary excretion of RBP and B2M is maintained in participants with risk factors for CKD. Fewer participants in the TAF arm experienced a decline in CrCl below 60ml/min, although the difference was of marginal statistical significance. There were also significantly greater declines in proteinuria and urinary RBP and B2M in participants randomized to TAF versus TDF.
A second analysis evaluated the renal safety of TAF at 96 weeks (Abstract 682). Consistent with published data and the other studies presented at CROI, this analysis demonstrated a favorable effect of TAF on CrCl and low molecular weight protein excretion. There was also a significant difference in study drug discontinuation for renal adverse events, with 6 events occurring in participants randomized to E/C/F/TDF and no events in those randomized to E/C/F/TAF (p=0.03).
Taken together with the switch study in participants with decreased CrCl, these data provide the first evidence that the improvement in renal biomarkers observed with TAF versus TDF may ultimately translate into an improvement in clinically relevant outcomes with cumulative use. Although urinary RBP and B2M are sometimes considered biomarkers of tubular injury, they are more accurately classified as biomarkers of proximal tubular dysfunction. The rapid decline in these biomarkers previously reported in participants switching from TDF to TAF are more consistent with tubular dysfunction than true injury, raising the question of whether these improvements predict an improvement in more relevant clinical outcomes such as overt proximal tubular injury, proteinuria, or decline in GFR. Longer follow-up in real world populations will determine the true clinical significance of the observed improvements in surrogate markers with TAF.
Renal Safety of Tenofovir Disoproxil Fumarate (TDF) in HIV treatment
Several observational studies provided additional insight into kidney toxicity with TDF. A retrospective analysis in the UK-CHIC cohort evaluated the incidence and risk factors for TDF-associated proximal tubulopathy, as defined by at least 2 markers of proximal tubular dysfunction (n=52) or biopsy-proven acute tubular necrosis (n=17). The definition did not require decline in CrCl, which accompanied approximately half of the cases (Abstract 683). In adjusted analysis, factors independently associated with proximal tubulopathy included older age, white race, longer TDF exposure, lower CD4 nadir, and PI use.
Analysis of data from 18,596 HIV-positive adults in NA ACCORD demonstrated no significant decline in CKD-EPI GFR over the first 6 months of TDF versus other NRTIs in participants with baseline eGFR ≥ 90ml/min/1.73m2 (Abstract 684). In participants with lower baseline eGFR, there was a significantly greater decline in eGFR in those on TDF. Although there was some recovery in eGFR after the first 6 months in participants on TDF, additional studies are needed to determine the clinical relevance of the early decline.
In a cross-sectional study of 884 HIV-positive men in MACS, cumulative exposure to TDF was associated with small but statistically significant increases in urine biomarkers of kidney injury (IL-18 and KIM-1) and a proposed marker of renal fibrosis (Abstract 685). There was no significant association between TDF exposure and albuminuria. Future studies should consider whether these biomarkers can predict clinically relevant kidney injury in individuals exposed to TDF, or whether they can provide insights into the mechanism of injury. These biomarkers are not currently recommended for clinical use in the diagnosis of TDF toxicity.
A prospective cohort study of 310 HIV-positive adults with baseline CrCl >60ml/min and without comorbid diabetes or hypertension evaluated risk factors for TDF discontinuation (Abstract 426). In addition to demographic factors and comedications, the authors considered the predictive ability of proximal tubular dysfunction (as defined by elevations in urinary RBP), plasma and urine tenofovir concentrations, and genetic polymorphisms in transport proteins that have been linked to TDF toxicity or tenofovir concentrations in prior studies. TDF was discontinued in 32 participants over median followup of 21 months, including 24 discontinuations for suspected kidney injury. In adjusted analyses, TDF discontinuation for kidney injury was associated with older age, male sex, PI use, and ABCB1 TT genotype (encoding multidrug resistance protein 1, or MRP1). Lower urine: plasma tenofovir concentration was associated with higher levels of urinary RBP and urinary protein, but was not independently associated with TDF discontinuation in this small study. As with prior genetic studies, the role of MRP1 in promoting TDF toxicity is not clear, as this transporter is not thought to play a major role in tenofovir transport.
Renal safety of TDF-based PrEP
Three abstracts presented in a themed discussion evaluated the renal safety of TDF-based PrEP. In a substudy of the iPrEx open-label extension study (iPrEx OLE, n=202), hair levels of tenofovir were shown to correlate with eGFR decline and with risk of MDRD eGFR<70 ml/min/1.73m2 in men (Abstract 866). Other factors associated with eGFR decline included baseline eGFR < 90 and age > 40 years.
In a substudy of the US Demo Project, dried blood spots were used to evaluate the correlation between tenofovir plasma concentrations and CrCl decline in 557 men (Abstract 867). There was a mean decline in CrCl of 6ml/min from baseline to week 12, with no further decline through 48 weeks. Tenofovir concentrations consistent with ≥ 2 doses/week were associated with greater decline in CrCl, with a dose response relationship noted for higher levels of dosing. In adjusted analysis, tenofovir concentrations indicative of ≥ 2 dose/ week and concurrent use of medications for diabetes or hypertension were associated with greater decline in CrCl. In contrast to the majority of studies, which have associated older age with GFR decline, age < 25 years was associated with greater CrCl decline in this relatively young population. Of note, there were no cases of CrCl<60 ml/min and no confirmed creatinine events through 48 weeks.
In a random substudy of Partners PrEP (n=1549), there was no statistically significant difference in the prevalence of proximal tubulopathy at 24 months between participants randomized to FTC/TDF versus placebo (Abstract 868). There was a higher prevalence of isolated hyperuricosuria and isolated non-albumin proteinuria in the active PrEP arm. In a nested case-control study, proximal tubopathy and individual markers of proximal tubular dysfunction did not predict clinically relevant decline in eGFR, defined as ≥ 25% decline from baseline.
These three studies suggest a low incidence of clinically relevant kidney injury among healthy HIV-negative adults taking PrEP for HIV prevention, although there is evidence of a dose-response relationship with mild decline in CrCl or eGFR. Individuals with lower baseline eGFR also appear to at greater risk for kidney function decline. Future studies should focus on identifying individuals at highest risk of kidney injury with longer exposure to TDF-based PrEP, to guide toxicity monitoring and/ or the use of alternative preventive measures.
Chronic kidney disease in HIV-positive adults
Several observational studies focused more broadly on CKD in HIV-positive adults. In a prospective cohort study, HIV-positive adults (n=191) and HIV-negative controls (n=100) underwent direct measurement of GFR by iohexol clearance (iGFR) at baseline and annually (Abstract 687). At baseline, measured GFR was lower in HIV-positive adults, although there was no significant difference in estimated GFR at baseline or in iGFR slope over time between HIV-positive and HIV-negative participants overall. Among HIV-positive participants, non-suppressed HIV-RNA, albuminuria, and higher hemoglobin A1c were associated with more rapid iGFR decline. The investigators also measured plasma levels of fibroblast growth factor 23 (FGF23), a regulator of phosphorus metabolism that has been associated with CKD and mortality (Abstract 651). Among HIV-positive participants, higher baseline FGF23 levels were associated with worsening of albuminuria and carotid intima media thickness, a noninvasive surrogate marker of atherosclerosis.
Three studies evaluated the association between markers of immunosuppression and risk of CKD. In the D:A:D cohort, several measures of immunosuppression based on CD4 cell count were associated with increased risk of CKD, as defined by a confirmed CrCl < 60 ml/min (Abstract 687). In adjusted analysis, the percentage of follow-up time with CD4 < 200 was independently associated with CKD. The impact of immunosuppression was more pronounced in individuals who were otherwise at low risk of CKD as determined by the D:A:D risk score. In a cross-sectional substudy of the Aquitaine cohort (n=849), investigators evaluated the correlation between a recently published index of immune activation/ immune senescence and prevalent CKD, as assessed by ICD-9 diagnostic codes and registry linkage (Abstract 689).
In adjusted analysis, a higher index of immune activation/ immune senescence was independently associated with eGFR < 60ml/min/1.73m2. In a single-center cohort study from Denmark, baseline levels of plasma soluble CD163, a marker of monocyte/ macrophage activation, were associated with incident CKD over a median of 10.5 years of follow-up (Abstract 690). In adjusted analysis, the highest quartile of soluble CD163 levels was associated with a > 8-fold increase in the risk of CKD. Taken together, these studies support a role for systemic inflammation and immune dysregulation in promoting CKD in the setting of HIV infection.
A biomarker substudy of the WIHS cohort (n=902) was recently published in Nephrology Dialysis Transplantation (Abstract 691). A panel of 3 urine biomarkers modestly improved the prediction of CKD when added to a multivariable model including relevant clinical variables. Future studies are needed to develop simple cut-offs and standardize biomarker assays prior to adoption in clinical practice.
In a cohort of 540 HIV-positive adults with documented ESRD in NA ACCORD, there were 255 deaths over a median follow-up of 2.5 years (Abstract 688). In adjusted analysis, baseline factors associated with increased mortality included age > 60 years, hypercholesterolemia, AIDS-defining illness, lack of ART prescription (at time of ESRD diagnosis), use of TDF prior to ESRD diagnosis, and lower CD4. As in the HIV-negative ESRD population, older black patients had a survival advantage over older white patients, while this pattern was reversed in those younger than 60.