icon- folder.gif   Conference Reports for NATAP  
  XVII International AIDS Conference
Mexico City
3-8 August 2008
Back grey_arrow_rt.gif
Cohort Studies Disagree on Tenofovir Role in Kidney Trouble
  XVII International AIDS Conference
August 3-8, 2008
Mexico City
Mark Mascolini
A retrospective 323-person cohort study involving a majority African-American population saw no link between tenofovir therapy and more than a 25% drop in glomerular filtration rate (GFR), a signal of kidney problems [1]. Meanwhile, cross-sectional (one-time) analysis of a 671-person cohort did tie tenofovir to a GFR below 90 mL/min/1.73/m2 [2]. Results of the two studies also disagreed on how race and weight affect risk of renal impairment.
Collaborating with GlaxoSmithKline investigators, M. Keith Rawlings and colleagues at the AIDS Arms Peabody Health Clinic in Dallas scoured their electronic database to find antiretroviral-treated people monitored continuously for at least 1 year who had two or more GFR estimates by the Modification of Diet in Renal Disease (MDRD) method. The Peabody investigators noted that, unlike the Cockroft-Gault formula for reckoning GFR, the MDRD factors in body surface area and black race, which predisposes people to kidney disease.
Of the 323 patients studied, 198 (61%) were black, 167 (52%) took tenofovir, and 156 (48%) did not take tenofovir. Racial distribution varied by tenofovir experience, with 96 African Americans (48%) taking tenofovir versus 41 whites (66%). That imbalance could reflect clinician concern that tenofovir may heighten the risk of kidney toxicity in already high-risk black patients. While 111 cohort members (34%) had hepatitis C virus (HCV) coinfection, 94 (29%) had hypertension.
A proportional hazards model to find correlates of more than a 25% drop in GFR figured by MDRD considered numerous baseline and demographic variables, individual antiretrovirals, hypertension, diabetes, viral hepatitis, time taking antiretrovirals, and duration of monitoring. People with hypertension when first evaluated had a 70% higher risk of renal impairment (hazard ratio [HR] 1.706, P = 0.0158). And every additional day of follow-up lowered the risk 0.1% (hazard ratio 0.999, P = 0.0017).
When Rawlings used the Cockroft-Gault method to estimate GFR, every additional year of age upped the risk of kidney impairment 3.9% (HR 1.039, P < 0.0077), every additional pound of weight lowered the odds 1% (HR 0.987, P = 0.0023), and every additional day of follow-up trimmed the risk 0.1% (HR 0.999, P = 0.0043).
SUN Study investigators used the simplified MDRD calculator (sMDRD) to estimate GFR in 671 people with an average age of 41 years, including 198 African Americans (30%), 65 Hispanics (10%), and 394 whites (59%) [2]. All study participants were adults with either (1) no antiretroviral experience and a CD4 count of 100 to 500 or (2) antiretroviral experience and more than 100 CD4 cells.
These residents of Denver, Minneapolis, Providence, and St. Louis had an average GFR of 96.8 mL/min/1.73m2. Four hundred cohort members (60%) had a GFR at or above 90, signaling normal kidney function, while 250 people (37%) had stage-2 disease (GFR 60 to 89) and 21 (3%) had chronic kidney disease (GFR 30 to 59). Cystatin C, a kidney function marker that goes up as GFR goes down, averaged 0.90 mg/L, and 127 people (19%) had a cystatin C level above 1.0 mg/L. Three quarters of cohort members were taking antiretrovirals, and almost the same proportion had a viral load below 400 copies.
E. Turner Overton and coworkers found that people with a GFR under 90 were significantly older (43.6 versus 39.4 years, P < 0.001), weighed significantly more (82.3 versus 78.2 kg, P = 0.006), had a higher body mass index (27.1 versus 26.8 kg/m2, P = 0.033), had higher cystatin C (0.95 versus 0,87 mg/L, P < 0.001), were more likely to have hypertension (27% versus 17%, P = 0.004), were more likely to be taking tenofovir (44% versus 35%, P = 0.026), and were more likely to be taking ritonavir (39% versus 25%, P < 0.001). No other antiretrovirals correlated with a sub-90 GFR.
Multivariate analysis considering these and other variables identified six independent predictors of a GFR below 90 mL/min/1.73m2:
- Weight above 148 kg: odds ratio [OR] 2.34, 95% confidence interval [CI] 1.50 to 3.67, P < 0.001
- Urine albumin above 30 mg/dL: OR 2.01, 95% CI 1.0 to 3.75, P = 0.029
- Cystatin C above 1.0 mg/L: OR 1.81, 95% CI 1.15 to 2.84, P = 0.010
- Hypertension: OR 1.76, 95% CI 1.13 to 2.73, P = 0.013
- Current tenofovir or ritonavir: OR 1.68, 95% CI 1.18 to 2.39, P = 0.004
- White race: OR 1.66, 95% CI 1.15 to 2.40, P = 0.007
Overton and colleagues did not speculate on why white race inflated the risk of kidney impairment in their cohort. Nor did Rawlings and coworkers ponder why higher weight lowered the risk of compromised renal function in their study group [1]. The Peabody-GSK team did note that the relatively high initial average GFR (above 113 mL/min) in their population may have eased the risk of tenofovir-related kidney toxicity in this group. Overton and colleagues stressed that the cross-sectional nature of their analysis limits its predictive power.
One possibly critical difference between these studies is their different definition of renal insufficiency--GFR under 90 mL/min in the Overton SUN study [2] and more than a 25% GFR drop in the Rawlings analysis [1]. By MDRD, only 48 of 280 people (17%) in Rawlings' cohort had an initial GFR under 90, compared with 271 of 671 (40%) in the SUN cohort. While 31 of 113 tenofovir takers (27%) had an initial sub-90 GFR in the Rawlings cohort, 17 of 136 tenofovir-naive people (12.5%) had a starting GFR under 90. Rawlings did not report absolute GFRs at the follow-up point used to determine GFR change.
From Jules: Overton concluded: "...chronic kidney disease is likely to become more prominent as patients age and are on long-term HIV therapy (aging is a major risk factor for kidney disease).....given that this evaluation was of only the baseline parameters, we may be limited by the cross-sectional design of this analysis. Evaluation of longitudinal data will clarify whether these findings represent a possible trend towards decreasing GFR with longer use of HAART, or if other factors (i.e., advancing age, further immune reconstitution) may outweigh any HAART-aasociated effects on GFR over time.....the sMDRD equation has not been properly validated in the HIV-infected population. The SUN Study group is looking at the validity of this parameter in other analyses.....Additional research is warranted to evaluate risk factors associated with kidney dysfunction and determine other markers such as cystatin C to detect early renal dysfunction."
Impact of Drug Therapy and Co-Morbidities on the Development of Renal Impairment in HIV-Infected Patients: Results of a Large Retrospective Database Study - (08/20/08) -
1. Rawlings MK, Klein J, Klingler ET al. Impact of drug therapy and co-morbidities on the development of renal impairment in HIV-infected patients. Results of a large retrospective database study. XVII International AIDS Conference. August 3-8, 2008. Mexico City. Abstract THPE0182.
2. Overton ET, Mondy K, Conley L, et al. Prevalence and predictors of renal insufficiency among HIV-infected patients in the study to understand the natural history of HIV/AIDS in the era of effective therapy. XVII International AIDS Conference. August 3-8, 2008. Mexico City. Abstract THPE0231.