Background:  The most relevant toxicity associated with tenofovir disoproxil fumarate (TDF) use is renal dysfunction, which usually presents as asymptomatic kidney tubular dysfunction (KTD). Risk factors associated with KTD in patients on TDF are older age, use of protease inhibitors (PI), and low body weight, all of which could be proxies of increased plasma concentrations of TDF. We investigated the relation between KTD and TDF plasma concentrations in our cohort of HIV⁺ patients treated with TDF for at least 1 year.

Methods:  Blood was drawn to determine mid-dose (C_12h) TDF plasma concentration and creatinine, glucose, phosphate, and uric acid. Urine was collected to measure α1-microglobulin, creatinine, glucose, phosphate, and uric acid. KTD was defined as the presence of at least 2 of the following criteria:  urinary α1-microglobulin/creatinine ratio >15 mg/10 mmol; fractional excretion (FE) of phosphate >20%  in the presence of hypophosphatemia; FE of uric acid  >10% in the presence of decreased plasma uric acid levels; or glucosuria. Multivariate logistic regression was used to study which variable was associated with KTD. The study received a waiver from the local medical ethics committee.  

Results:  A total of 161 HIV⁺ patients were evaluable for analysis. Abnormalities in individual parameters of tubular dysfunction were seen in 101 patients (63%). A total of 17 patients (10.6%) fulfilled the definition of KTD. Patients with KTD had a significantly higher plasma TDF concentration and a higher serum creatinine than patients without KTD. In addition, patients with KTD appeared to have longer TDF exposure, although this difference was statistically not significant. Other variables were comparable between the 2 groups. Multivariate logistic regression showed TDF plasma concentration to be the only variable associated with KTD. Mean C_12h was 0.143 (95%CI 0.102 to 0.183) mg/L in patients with KTD vs 0.108 mg/L (95%CI 0.101 to 0.115) mg/L in patients without KTD (p = 0.04). An ROC analysis, however, did not result in a clear threshold for TDF C_12h to discriminate for KTD with acceptable sensitivity and specificity. A post-hoc analysis was done to associate KTD with the product of C_12h and the duration of TDF exposure (in months). Indeed, this relation was stronger than that of C_12h and KTD (p = 0.02).

Conclusions:  Prolonged presence of elevated TDF plasma concentrations may be an important risk factor for developing KTD and should be prospectively evaluated.