| |
Renal Dysfunction & Tenofovir
|
| |
| |
"Changes in Renal Function Associated with Tenofovir Disoproxil Fumarate Treatment, Compared with Nucleoside Reverse-Transcriptase Inhibitor Treatment"
Clinical Infectious Diseases, April 15 2005;40:1194-1198
Joel E. Gallant, Michelle A. Parish, Jeanne C. Keruly, and Richard D. Moore
The Johns Hopkins University School of Medicine, Baltimore, Maryland
In our large observational cohort, use of tenofovir disoproxil fumarate (n = 344) was associated with a greater decline in renal function than was use of alternative nucleoside analogues (n = 314). Other associations included a lower CD4 cell count, decreased renal function at baseline, and diabetes. The declines were modest and did not lead to greater rates of discontinuation of therapy.
Tenofovir disoproxil fumarate (TDF) is the first nucleoside reverse-transcriptase inhibitor (NRTI) approved by the US Food and Drug Administration for the treatment of HIV disease. It is renally excreted via a combination of glomerular filtration and active tubular secretion. Adefovir dipivoxil, a related NRTI, caused proximal renal tubular dysfunction at a dosage of 60-120 mg/day, which is required to inhibit HIV replication [1]. However, in clinical trials, TDF has demonstrated an excellent renal safety profile. In the Gilead Sciences 903 trial, in which 299 treatment-naive patients received TDF in combination with lamivudine and efavirenz, there were no instances of renal failure or grade 3/4 elevations in the serum creatinine level through 144 weeks, and there were no significant differences in renal function, compared with that of patients taking stavudine [2, 3].
Nevertheless, renal impairment, including but not limited to proximal renal tubular dysfunction, has been reported to be associated with TDF use [4-11]. The majority of cases of renal impairment have occurred in patients with underlying systemic or renal disease or in patients taking nephrotoxic agents. However, some cases have occurred in patients without any identified risk factor. Coadministration of TDF with drugs eliminated by active tubular secretion may increase concentrations of either tenofovir or the coadministered drug because of competition for this pathway, and drugs that decrease renal function may increase the concentration of tenofovir.
The reports of renal dysfunction in TDF-treated patients raise concerns about the potential for nephrotoxicity with use of this drug. We analyzed data from a large clinical database to assess the effect of treatment with TDF on renal function in clinical practice.
Methods. We analyzed data from the Johns Hopkins HIV Clinical Cohort. The methods of data collection for this observational longitudinal cohort have been described [12]. This analysis focuses on all patients who started therapy between 1 January 2001 and 31 December 2003 with either TDF or an alternative NRTI as part of a HAART regimen. All data regarding prescribed antiretroviral use was collected by dates of use and dose. Creatinine clearance (CLCr) was calculated using the Cockcroft-Gault equation, which estimates CLCr on the basis of the serum creatinine level, weight, and sex of the patient [13]. CLCr was calculated using the average of the 2 serum creatinine levels determined closest to initiation of therapy. For each subject, the maximum serum creatinine level of those measured within 1 year after initiation of and during the treatment regimen was determined. The subsequent (or preceding, if missing) measurement of serum creatinine was averaged with the maximum level to obtain the value used to calculate the change in CLCr. The average of the 2 measurements was used to calculate CLCr to minimize regression to the mean.
We calculated the absolute and percentage change in CLCr from baseline for each patient and compared these changes in TDF-treated patients with those in NRTI-treated patients. We also assessed the associations of other variables with change in CLCr, including age, ethnicity, sex, HIV transmission risk factor, diabetes status, hypertension status, CLCr at baseline, CD4 cell count and HIV-1 RNA load, and concurrently administered antiretroviral agents. The Wilcoxon rank sum test was used for bivariate comparisons with the percent change in CLCr. Multivariate least-squares linear regression was used to assess the associations of multiple factors with percent change in CLCr. We plotted the mean values (and standard error) of CLCr at baseline and at 3-month intervals after initiation of treatment for 1 year. We used Student's t test to compare the change between the mean value at baseline with the mean value at each time interval for the 2 groups.
RESULTS
A total of 344 patients received TDF, and 314 patients received an alternative NRTI. The characteristics of the sample are shown in table 1. There were no substantive differences between the groups, except for the expected greater use of alternative NRTIs in the NRTI group, the greater use of lopinavir-ritonavir in the TDF group, and the greater use of efavirenz among patients receiving NRTIs. It is now recommended that the dosing interval of TDF be modified for patients with a CLCr of <50 mL/min [14, 15]. Only 2 patients, who had a serum creatinine level of >2.0 mg/dL at baseline, received <300 mg of TDF per day.
Table 2 shows serum creatinine levels and CLCr at baseline and changes in serum creatinine levels and CLCr over the period of treatment. The TDF group had significantly greater increases in serum creatinine levels and decreases in absolute and percentage CLCr, compared with the NRTI group. There was no difference in the rate of discontinuation of treatment coincident with maximum decline in CLCr: only 19 (5.5%) of TDF-treated patients and 21 (6.7%) of NRTI-treated patients discontinued therapy at the time of maximum decline in renal function.
Table 2. Comparison of change in renal function between patients who received tenofovir disoproxil fumarate (TDF) and patients who received nucleoside reverse-transcriptase inhibitors.
|
|
 |
| |
Figure 1 shows the change in CLCr from the start of treatment through day 365 of therapy (in 90-day increments). It can be seen that the change in CLCr was apparent after 90 days of therapy and persisted for the entire year.
Figure 1. Plot of mean (with standard error) CLCr over time for patients who received TDF and patients who received other NRTIs (last CLCr on treatment carried forward if treatment stopped).
|
|
 |
| |
In addition, regarding the comparison of TDF use with use of other NRTIs, diabetes was associated with a -13% change in CLCr (vs. a change of -8% in patients without diabetes [P < .04]); a baseline CD4 cell count of <50 cells/mm3 was associated with a change of -14% (vs. a change of -8% for patients with a baseline count of >50 cells/mm3 [P < .01]); a baseline HIV-1 RNA load of >20,000 copies/mL was associated with a change of -15% (vs. a change of -8% for patients with a baseline load of >20,000 copies/mL [P < .02]); and receipt of initial HAART was associated with a change of -11% (vs. a change of -8% for patients who received HAART subsequently [P < .03]). Hypertension status, age, sex, ethnicity, the presence of concomitant hepatitis C or B, use of lopinavir/ritonavir, or use of any other specific protease inhibitor or NRTI were not associated with the percent change in CLCr (P > .05).
Adjusting for these variables in a multivariate analysis, only TDF use (P = .006) and a CD4 cell count of <50 cells/mm3 (P < .001) were associated with CLCr decline. A baseline CLCr value of <50 mL/min and diabetes were less strongly associated (P = .10). Because the dosage of TDF was only adjusted for poorer renal function (defined as a CLCr of <50 mL/min) in 2 patients, the multivariate analysis was repeated for those patients with a baseline CLCr of >50 mL/min. The results were similar; TDF use (P = .004), a low CD4 cell count (P < .001), and diabetes (P = .06) were associated with CLCr decline. In both analyses, hypertension status, use of lopinavir-ritonavir or other antiretroviral agents, HIV-1 RNA load, previous use of adefovir dipivoxil, age, sex, ethnicity, and HIV transmission risk factor were not associated with CLCr decline, after adjustment for the variables in table 2.
AUTHOR DISCUSSION
TDF use has not been associated with nephrotoxicity in clinical trials [2, 3, 16-18 However, it has been suggested that clinical trials may not represent "real world" scenarios, because patients with renal insufficiency or risk factors for renal insufficiency are often excluded. Data from clinical cohorts are somewhat conflicting. Although several studies have shown little evidence that nephrotoxicity is associated with use of TDF [19-23], other studies have shown evidence of an association with a modest decline in renal function [24-29].
In our study, TDF use was associated with a significantly greater decline in CLCr, compared with the decline associated with the use of alternative NRTIs. The relative decline in CLCr associated with the use of TDF, compared with that associated with the use of alternative NRTIs, was 4%. To minimize temporal bias, the period of treatment with NRTI was kept within the same study period as treatment with TDF. Because this was an observational study, and because urinalysis and serum phosphate analysis are not performed routinely in our clinic, we could not assess the frequency of proteinuria or Fanconi syndrome.
In addition to TDF use, the other variable strongly associated with CLCr decline was advanced immunosuppression (defined as a CD4 cell count of <50 cells/mm3). There were trends toward associations with diabetes and with a baseline CLCr of <50 mL/min. With the exception of only 2 patients, the dosage of TDF was 300 mg/day for all patients. Whether a lower dose of TDF for patients with a lower CLCr at initiation of therapy would have modified the decrease in CLCr is unknown. However, when restricting our analysis to patients with a baseline CLCr of >50 mL/min-patients who should receive full doses of TDF, according to current dosing recommendations [15]-we found the same modest decline in renal function. Patients who received treatment with TDF were less likely to be receiving initial HAART than were patients who received treatment with other NRTIs. However, receipt of initial HAART was not associated with CLCr decline, compared with receipt of subsequent HAART.
It is possible that the emerging difference between CLCr decline associated with TDF use and that associated with the use of other NRTIs is a result of a longer duration of treatment. In an earlier analysis of these data [30], which found no significant difference in renal function, the median treatment duration was 246 days, compared to 322 days for the current analysis. More patients who received TDF were also taking lopinavir-ritonavir, which has been shown to increase tenofovir levels [31]. However, concomitant use of lopinavir-ritonavir or other antiretroviral drugs was not associated with CLCr decline.
Although statistically significant, the decline in CLCr associated with TDF use was small and was not associated with a greater rate of discontinuation, probably because the changes in creatinine levels appeared trivial to the clinician. Because the majority of these patients continue to receive treatment, we will continue to observe this cohort to assess renal function changes as the duration of treatment increases.
There may be a modest decline in renal function in TDF-treated patients with prolonged use, especially in patients with advanced HIV disease, diabetes, or a decreased renal function at baseline. The clinical significance of these findings is unclear. Because the decrease in CLCr was gradual and did not lead to clinically significant renal failure, there is no reason to withhold treatment with TDF from patients who would benefit from it. However, the data from our large cohort emphasize the importance of assessing renal function prior to initiation of such treatment, making appropriate dose adjustments for patients with impaired renal function at baseline, and monitoring changes in renal function in TDF-treated patients.
|
|
| |
| |
|
|
|
