Tenofovir and indinavir associated with increased risk of chronic renal failure in EuroSIDA study
Simon Collins, HIV i-Base
Results from a cross-sectional analysis of chronic renal disease from a large European cohort were reported in the 13 May issue of AIDS. 
EuroSIDA is a prospective international study in patients from 24 European countries (plus Argentina and Israel) that has been running since 1994 and now includes over 14,000 patients and 51,000 years of follow-up. Routine collection of serum creatinine was only included in data collection from January 2004 and this analysis included all patients (n=4474) who had 2 or more serum creatinine measurements from January 2004.
Chronic renal failure (CRF) was defined as two consecutive glomerular filtration rate (GFR) < 60mL/min per 1.73sq using either Cockcroft-Gault (CG) or Modification of Diet in Renal Disease (MDRD), both adjusted for body surface area (BSA).
Using CG, the median GFR at baseline was 94.4 (IQR, 80.5-109.3).
158 patients (3.5%) had CRF using CG and 209 (4.7%) using MDRD. 101 (2.3%) had CRF using both measurements.
The researchers reported a high degree of correlation between CG and MDRD (0.773, p<0.0001) but this dropped to 0.518 if CG was not adjusted for BSA. CG generally gave higher readings, and the difference was greatest in the youngest patients. For patients aged <40 years the mean difference was 8.29 (95CI 7.33-9.25) and in patients aged over 50 years MDRD gave a higher value, mean difference 1.17 (95CI 0.27-2.07).
Patients with CRF were older (median 61.9 vs 43.1 years), had lower CD4 nadirs (median 80 vs 137 cells/mm3) and were more likely to be diagnosed with AIDS (44.3 vs 30.4%), diabetes (16.5 vs 4.3%) or hypertension (53.8 vs 26.4%), all p<0.001.
The relationship between CRF and use of antiretroviral and renal-toxic drugs was then examined.
Of 2118 patients never exposed to indinavir, 38 had CRF (1.8%) compared to 120/2356 (5.1%) of patients who had used indinavir. For tenofovir, these figures were 98/3213 (3.1%) and 60/1261 (4.8%) in the non-exposed and exposed patients respectively.
Details of antiretroviral drug use are included in Table 1.
Table 1: Antiretroviral drug exposure in patients with CRF
|ARV||No. exposed||Approx % exposed with CRF||No. not exposed||Approx % not exposed with CRF|
|RTV (any PI)||2457||2.8||2017||4.1|
Any use of indinavir (OR, 2.49) or tenofovir (OR, 2.18) was associated with increased odds of CRF, and results are detailed below in Table 2. Similar results were found when the analysis was repeated looking at duration of exposure to ARVs. Each additional year of exposure to indinavir increased the risk of CRF at baseline by 15% (OR, 1.15; 95CI 1.05-2.24; p = 0.0013), and to tenofovir by 60% (OR, 1.60; 95CI 1.20-2.15; p=0.0015).
Use of ritonavir, as single or part of boosted PI regimen, including with tenofovir had no increased risk. T-20 use showed a decreased risk.
Table 2: Multivariate analysis of factors associated with confirmed GFR< 60 (CG)
|E Europe vs rest||2.45||1.35-4.45||0.0033|
|Prior AIDS vs none||1.34||0.88-2.02||0.17|
|Age (per 10 yrs older)||5.47||4.45-6.72||<0.0001|
|CD4 nadir (per 50% higher)||0.90||0.82-0.99||0.028|
|Viral load (above 500 vs lower)||1.54||0.98-2.41||0.062|
|Baseline (per 12 mo later)||1.65||1.04-2.62||0.33|
|Any TDF use||2.18||1.15-3.81||0.0061|
|Any IDV use||2.49||1.62-3.83||<0.0001|
|Any T-20 use||0.13||0.03-0.65||0.013|
|Any RTV use (single PI)||0.77||0.39-1.50||0.44|
|Any RTV (2 or mor PI regimen)||0.89||0.56-1.43||0.64|
|Any renal-toxic drug||1.47||0.94-2.30||0.089|
* acyclovir, pentamidine, foscarnet, cidofovir or amphotericin B.
There was a significantly greater proportion of patients with CRF at baseline that had used renal-toxic drugs (pentamadine, cidofovir, foscarnet). These differences are detailed in Table 3.
Table 3: Proportion of patients with baseline CRF by exposure to renally toxic drugs
|% exposed||% not exposed||p|
|Any renal drug exposure||27.2%||15.9%||<0.0001|
The study provided important information on the use of CG and MDRD calculations and the importance of adjusting for BSA. For patients identified as being at higher risk, the use of more sensitive tests including estimating urinary clearance of IV-infused insulin or labeled tracers with timed urine collection is recommended.
The authors noted several limitations to cross-sectional analysis including the relatively recent decision to measure creatinine, the impossibility of showing whether CRF occurred before or after exposure to drugs, and whether it was a consequence of exposure to individual drugs, and the impact of other factors that cannot be adjusted for (such as use of NSAIDs, and that choice of ARV treatment was not randomised).
Future analysis are planned in order to provide longitudinal analyses from this database.
This should also include analyses for patients at highest risk, and with overlapping risk factors (previous AIDS, use of nephrotoxic drugs, age, and use of indinavir and tenofovir etc) in order to consider risk and impact of using nephrotoxic and renally cleared drugs in these patients.
- Moorcroft A, Kirk O, Gatell J et al for EuroSIDA Study Group. Chronic renal failure among HIV-1 infected patients. AIDS 2007, 21:1119-1127.