Background:  Tenofovir disoproxil fumarate (TDF) has been associated with renal toxicity in previous reports. Risk factors identified were lower baseline weight, baseline creatinine or creatinine clearance, co-administration with lopinavir/ritonavir or didanosine, exposure to amphotericin B, ART-naive status, and intravenous drug use. Our research aim was to determine whether low-dose ritonavir (RTV) is an independent risk factor for nephrotoxicity in patients taking TDF.

Methods:  Our retrospective cohort included adults under care for HIV infection without pre-existing renal disease (creatinine clearance ≥75 cc/minute) and on stable combination ART from 2000 to 2005 with regular renal function assessments (n = 635). We formed 3 study groups:  no TDF (N_(–)TDF = 380), TDF without RTV (N_TDF&(–)RTV = 71), and TDF with RTV (N_TDF&RTV = 184). We examined 5 confirmed renal function metrics:  serum creatinine ↑ 0.5 mg%, ↓ Cockcroft-Gault (C-G) CrCl  30 cc/minute, ↓ C-G CrCl 25%,  ↓ MDRD clearance by 30 cc/minute, and ↓ modification of diet in renal disease (MDRD) clearance by 25%. Time to nephrotoxicity was analyzed with Kaplan-Meier analysis and Cox-proportional hazards models were fit to identify independent risk factors for nephrotoxicity. Co-variates included demographic characteristics, baseline body mass index, CD4, viral load, MDRD clearance, didanosine (ddI) exposure, prior ART experience, and concomittant nephrotoxins. Co-variates significant (p <0.05) in unadjusted analyses were entered into multipredictor Cox models.

Results:  Event rates varied from 6.9% to 28.7% for different metrics. Of 635 patients at risk, 33% met at least 1 nephrotoxicity endpoint. In unadjusted analyses, relative to no TDF, TDF+RTV was associated with increased risk of MDRD ¯ 30 cc/minute (HR 1.6, p = 0.01) and MDRD ¯ 25% (HR 1.6, p = 0.003). TDF without RTV was not associated with nephrotoxicity by any endpoint metric. In models adjusted for body mass index, baseline CD4, baseline viral load, and baseline MDRD clearance, TDF+RTV remained a significant predictor of MDRD ¯ 30 cc/minutes (adjusted HR 1.9, p = 0.001), and MDRD ¯ 25% (adjusted HR 1.8, p <0.0001), whereas TDF without RTV was not associated with increased risk.

Conclusions:  In our cohort, low-dose RTV in TDF-containing regimens was associated with nephrotoxicity by 2 MDRD clearance metrics in patients without pre-exisitng renal disease, while TDF alone posed no additional risk. The pathophysiologic basis of the observed drug interaction is unknown.