 |
|
|
| |
Renal Report by Wyatt/Klotman Mt Sinai
|
| |
| |
CROI Feb 3-6, 2008, Boston
Christina M. Wyatt, MD
Paul E. Klotman, MD
Mount Sinai School of Medicine, New York, NY
With prolonged survival and aging of the HIV-infected population in the United States and Western Europe, there is increasing interest in non-AIDS conditions that are contributing to morbidity and mortality in ART-treated patients. An early plenary session presented by Andrew Phillips and a symposium on "Aging and AIDS" helped to frame the discussion of these important non-AIDS comorbidities, including kidney disease.
In the scientific sessions, investigators from four large prospective cohorts reported on the incidence and risk factors for progression of chronic kidney disease among patients with HIV. The EuroSIDA investigators have previously described a relatively low prevalence of chronic kidney disease in this largely Caucasian cohort [Mocroft et al. AIDS 2007]. In the current abstract [Abstract 971], Ole Kirk and colleagues describe the incidence of kidney disease progression, defined as a confirmed glomerular filtration rate (GFR) below 60 mL/min/1.73m2 (2.4% of subjects) or a confirmed 25% decline in GFR from baseline (3.2% of subjects). Clinical factors associated with progression of kidney disease included advanced HIV disease (lower current CD4 cell count, higher viral load, and previous AIDS diagnosis), hepatitis C virus co-infection, prior cardiovascular event, and lower baseline GFR. These data are consistent with other reported incidence data for Caucasian populations. They also suggest that advanced HIV disease is associated with renal complications, underscoring the current recommendations that kidney disease is an indication for ART regardless of CD4.
HIV disease-specific characteristics did not appear to predict annual GFR decline in the all-male Multicenter AIDS Cohort Study [Abstract 973]. At baseline, nearly 6% of HIV-positive antiretroviral-treated men had an estimated GFR < 60 mL/min/1.73m2, compared to only 3.6% of HIV-negative controls (p<0.001). The authors considered a GFR decline > 3% per year to be clinically significant. In addition to two traditional kidney disease risk factors, hypertension and smoking, an annual GFR decline of this magnitude was also associated with white race and higher baseline GFR. The latter finding may reflect the poor precision of GFR estimates in the normal range, and brings into question the clinical significance of this endpoint. The association with white race is also unexpected, and contrasts sharply with the strong association between African-American race and GFR decline observed in the Johns Hopkins HIV Cohort, described below, and in the largely male Veterans Affairs cohort [Choi et al. J Am Soc Nephrol 2007]. Proteinuria was observed in 23.4% of antiretroviral-treated participants, compared to only 5.3% of HIV-negative control. The presence of proteinuria was associated with traditional risk factors for chronic kidney disease, including black race, diabetes, hypertension, and smoking. HIV infection and hepatitis C co-infection were also associated with proteinuria, highlighting the potential of the MACS to help define the independent impact of HIV and HCV on kidney disease epidemiology.
In contrast to the limited representation of minority patients in EuroSIDA and MACS, African-American participants account for more than three-quarters of the Johns Hopkins HIV Cohort [Abstract 972]. Gregory Lucas and colleagues dissected the epidemiology of chronic kidney disease among cohort participants, in order to explain the dramatic increase in end-stage renal disease observed among HIV-positive African-Americans. The incidence of chronic kidney disease was approximately 2-fold higher among African-American participants, while the rate of progression to end-stage renal disease was nearly 18-fold higher in African-American compared to white participants. This racial disparity in progression was most prominent among blacks with biopsy-proven HIV-associated nephropathy, but was also observed with other kidney diseases. In a separate report from the Hopkins cohort, kidney disease was the primary admission diagnosis for more than 9% of hospitalizations in cohort participants [Abstract 963]. The incidence rate of hospitalization for non-AIDS comorbidity was significantly decreased among ART-treated patients, although the analysis did not consider individual conditions.
Based on the previously reported association between proteinuria and progressive kidney disease, Samir Gupta and colleagues described the prevalence and slope of proteinuria in the ACTG Longitudinal Linked Randomized Trials (ALLRT) cohort [Abstract 974]. Overall, 20% of cohort participants had a urine protein: creatinine ratio > 0.2 mg/g (approximately 200 mg/day of urinary protein excretion) at the first measurement. Predictors of proteinuria included traditional kidney disease risk factors such as older age, black race, diabetes, and hypertension, in addition to higher baseline HIV viral load and hepatitis C co-infection. An association with prior indinavir use may reflect longer cumulative ART exposure, and will be the subject of further analysis in this cohort, as will a trend towards increased proteinuria with current tenofovir use. Viral load suppression was associated with lower odds of proteinuria, and there was a small but significant decline in the slope of proteinuria during followup. This decline in proteinuria was observed almost exclusively among participants with significant proteinuria at baseline. Further analyses are planned to characterize the changes in protein:creatinine ratio among participants without significant baseline proteinuria, and to determine the impact of cumulative exposure to ART and to different antiretroviral agents.
Several abstracts addressed the potential for renal toxicity associated with nucleotide reverse transcriptase inhibitors. Michael Horberg and colleagues at Kaiser Permanente described the frequency of unconfirmed renal laboratory abnormalities in a retrospective cohort of 1742 enrollees initiating tenofovir-containing regimens, compared to 623 patients starting a new regimen that did not include tenofovir [Abstract 975]. Tenofovir-treated patients had higher baseline CD4 cell counts, but were more likely to be ART-experienced and to have comorbid diabetes. During a median of 2.7 years of followup, tenofovir-treated patients were more likely to experience a single episode of increased serum creatinine, proteinuria, glycosuria, or mildly reduced serum phosphorus (< 2.7 mg/dL) or bicarbonate (<23meq/L). Patients with at least 3 of these abnormalities within a 30-day period were considered to have Fanconi syndrome. This constellation of abnormalities also occurred more frequently in tenofovir-treated patients (10% versus 3.4%). Assessment of at least 3 laboratory values required for the diagnosis of Fanconi syndrome was more common among tenofovir-treated patients, suggesting an appropriate increase in clinical monitoring but also raising the possibility of assessment bias. In addition, observed laboratory abnormalities were based on a single unconfirmed measurement. In the pivotal GS903 trial for tenofovir, similar unconfirmed elevations of creatinine occurred in 4.1% of subjects in the tenofovir arm, but were only confirmed in one-third of those subjects (1.4% overall) on repeat testing. Similar trends were observed for serum creatinine over 2 years of followup (GS 910). The reproducibility of serum phosphorus abnormalities was even lower, ranging from less than 6% in 903 to less than 25% in 910. The current study likely overestimates the true incidence of clinically significant laboratory abnormalities among tenfovir-treated patients; nonetheless, the frequency of observed laboratory abnormalities does reinforce the importance and complexity of toxicity monitoring in ART-treated patients.
In the Swiss HIV Cohort Study, tenfovir use was associated with a small but significant increase in serum alkaline phosphatase, which may be an indicator of increased bone turnover [Abstract 968]. Changes in alkaline phosphatase were hypothesized to result from increased urinary phosphorus losses, although supporting data were not available and changes in alkaline phosphorus were independent of changes in GFR. Future studies are needed to more fully characterize the relationship between tenofovir use, urinary phosphorus/ calcium excretion, and bone mineral density. (from Jules: I think it is unsure if alkaline phosphatase is a reliable marker for changes in bone).
Consistent with the findings of previous randomized clinical trials, the HEAT trial did not demonstrate any difference in incidence of proximal tubular dysfunction between ART-na•ve patients randomized to abacavir/ 3TC or tenofovir/ FTC in combination with boosted lopinavir [Abstract 774]. As with all randomized clinical trials, the results may not be generalizable to sicker patients in clinical practice, but this finding may be particularly important in light of concerns that boosted protease inhibitors are associated with increased risk for tenofovir nephrotoxicity. This was also the first head to head assessment prospectively addressing the question of whether tenofovir is associated with more renal events.
(from Jules: in the 48-weeks results from the CASTLE study [ATV/r vs LPV/r in naives] investigators reported: renal all grade AEs, 2% in both arms; change from baseline at 48 weeks in renal function: mean serum creatinine: +0.05 mg/dL ATV/r, +0.02 mg/dL LPV/r; median calculated creatinine clearance: 1% decrease in both arms. Abbott reported 48-weeks results from a comparison of once vs twice daily Kaletra (LV/r) tablets co-administered with tenofovir/FTC: creatinine clearance <50 ml/min was 2% in both arms. In the TDF/FTC arm of the HEAT Study after 48 weeks: there was no change in median numbers reported for MDRD in Blacks (94 at baseline vs 95 at wk 48) and non-Blacks (84 at baseline, 88 at wk 48); no change in median creatinine clearance (100 at baseline, 106 at wk 48); no change in median Protein:Cr Ratio (0.09 at baseline, 0.09 at wk 48); median change from baseline in serum phosphate was -0.4 (median 3.70 at baseline and 3.20 at wk 48); urine glucose (mg/dL): was median 7.0 at baseline, and 7.0 and at wk 48; mean change in urine glucose was +26: 26.2 mg/dL at baseline and 58.4 mg/dL at wk 48. The study reported 2 (<1%) renal failures on TDF/FTC. Three patients (1%) switched off TDF/FTC due to Proximal Renal Tubule Dysfunction (PRTD); PRTD definition: serum creatinine rise of ≥0.5 mg/dL from BL and serum phosphate < 2mg/dL or either of the former plus any 2 of the following: proteinuria ≥100 mg/dL, glycosuria ≥250 g/dL, low serum potassium <3 meQ/L, or low serum bicarbonate <19 meQ/L.)
The renal toxicity of nucleotides is associated with intracellular accumulation in proximal tubular epithelial cells. Tomas Cihlar and colleagues at Gilead Sciences described the pattern of tissue accumulation and renal tubular transport of a novel nucleotide, GS-9148, which is expected to have lower nephrotoxic potential [Abstract 977a]. The monoamidate prodrug GS-9131 is preferentially hydrolyzed to the active drug in lymphocytes, theoretically targeting drug delivery to peripheral lymphocytes and lymphoid tissue. In vitro studies demonstrated low affinity of GS-9148 for the organic anion transporters hOAT1 and hOAT3, which are responsible for proximal tubular cell uptake of nucleotides. Both the novel nucleotide and tenofovir were shown to be less toxic to proximal tubular epithelial cells in vitro, compared to cidofovir and adefovir. Oral administration of the prodrug GS-9131 to dogs also resulted in minimal accumulation in kidney tissue. The low efficiency of renal tubular cell uptake predicts low nephrotoxic potential compared to previously developed nucleotides, although early clinical studies will focus significant attention on nephrotoxicity in addition to efficacy. Of course, the safety of targeting drug uptake to other tissues will need to be confirmed as well.
Nephrotoxic drugs were a common contributor to acute renal failure among HIV patients at King's College Hospital in London [Abstract 976]. Frank Post and colleagues described the incidence of acute renal failure among 2274 patients evaluated between 1998 and 2005. There were 144 episodes of acute renal failure over 7394 patient-years of followup (2.7 events/ 100 patient-years). Approximately half of the cases occurred within the first 3 months of HIV care, and it is possible that kidney disease was the primary presentation of HIV infection in some patients. Nearly 6% of patients experienced at least one episode of acute renal failure, despite a conservative case definition. The most common factors contributing to acute renal failure were nephrotoxic drugs, infection (both opportunistic and non-opportunistic infections), malignancy, and liver disease. Mortality was 31% among patients with acute renal failure, consistent with a strong association between acute renal failure and poor outcomes in the general population.
(from Jules: In multivariate analysis, nadir CD4 T-cell count < 100 x 109 cells/L (OR 6.7, 2.5-18.3) and AIDS (OR 6.7, 3.4-13.3) were associated with early-onset ARF, while IV drug use (OR 4.8, 1.3-17.7), hepatitis C co-infection (OR 3.4, 1.3-8.6) and nadir CD4 T-cell count < 100 x 109 cells/L (OR 5.8, 2.5-13.4) were associated with late-onset ARF. Author Conclusions: ARF was common and associated with advanced immunodeficiency. The incidence of ARF declined more than 15-fold in patients who had received HIV care for > 3 months. The ARF incidence rate was particularly high during the first 3 months following HIV diagnosis / initiation of HIV care. Beyond the initial 3 months of HIV care, the incidence of ARF was approximately 1 episode per 100 person years. Our results suggest that HIV care, including targeted HAART and OI prophylaxis, is associated with a marked reduction in the risk of developing ARF. As nearly all patients with ARF were hospitalised, nephrotoxic antimicrobials should be used judiciously and dehydration and NSAIDs avoided to reduce the risk of ARF in this setting.)
All six of the clinical cohort studies presented at CROI relied on GFR estimates to define clinical endpoints. Although GFR estimates are more sensitive markers of decreased kidney function, current methods to estimate GFR have not been validated in patients with HIV. The PREPARE investigators presented the results of a small GFR validation study performed in clinical trial participants with well-controlled HIV infection and preserved kidney function [Abstract 977b]. The performance of 24-hour urine creatinine clearance, Cockcroft-Gault creatinine clearance, MDRD GFR, and cystatin C were compared to direct GFR measurement by iothalamate clearance. The mean measured GFR was 134mL/min, and several patients had GFR measurements above 150mL/min. The clinical significance of these "high" GFR measurements is not known, but could suggest pathologic hyperfiltration similar to that observed in early diabetic nephropathy. Correlation with measured GFR was strongest for creatinine clearance estimated by 24-hour urine collection or Cockcroft-Gault, and both MDRD estimates and cystatin C underestimated GFR in this population with normal to high GFR. The performance of different GFR estimates for predicting clinically significant reductions in kidney function could not be determined in this population, since no patients had a measured GFR below 60 mL/min. In addition, the study population included a single woman and only 3 patients of African descent, limiting our ability to extrapolate these results to other populations. The provocative results of the current study suggest a need for larger studies with more diverse patient populations and varying levels of kidney function.
|
| |
|
|
|
|
|
