CROI: Tenofovir and kidney toxicity - low rates seen, but some groups may be more vulnerable

The Thirteenth Conference on Retroviruses and Opportinsitic Infections gathered together five varied studies of kidney toxicity in patients taking tenofovir (Viread) to better establish the frequency of this adverse event.

Physicians have been concerned about tenofovir possibly causing long-term renal toxicity because this is a severe side-effect of the two chemically similar drugs adefovir and cidofovir. In preclinical studies in animals, dose-limiting toxicity was seen in dogs at levels exceeding five times the dose given to humans.

The largest database (Nelson) of patients taking tenofovir, however, offered news that the presenter of the poster, Professor David Cooper, called “reassuring.”

Safety data were combined from 10,343 patients taking tenofovir during its pre-licensing Expanded Access Programmes (EAPs) and from post-marketing safety reports received up to April last year. Between them these added up to 455,392 patient-years of exposure to the drug.

A total of 6% of patients in the EAP experienced serious adverse events (SAEs) but no one SAE was experienced by more than 0.67% of patients.

Kidney problems were second on the list, experienced by 0.57% of patients but less than half of these (0.22% or one in 454 patients) had their renal problems ascribed to tenofovir.

In the post-marketing safety reports (which, being individual safety alerts, will tend to over-represent the frequency of SAEs), overt renal failure occurred in 0.3% of patients, a rate of one case per 2000 patients a year.

Another large study (Heffelfinger) used a different methodology. The Centers for Disease Control (CDC)’s Adult/Adolescent Spectrum of Disease Reports complied safety data on 11,362 patients in ten cities in the United States between 2000 and 2003.

It found rates of ‘renal impairment’ considerably higher than in the tenofovir EAPs and post-marketing reports. The proportion of patients (taking any HAART regimen) experiencing mild, moderate and severe renal impairment comprised 35.1%, 6.4% and 2.6% respectively.

Taking a regimen containing tenofovir was found to increase the risk of renal impairment by 60% compared with patients not taking tenofovir. However there was no statistically significant association between tenofovir and severe impairment.

Several factors were strongly associated with renal impairment. Age was important: patients over 50 were 3.5 times more likely to have severe impairment. Hypertension and diabetes raised the chance of severe impairment by a similar amount.

And having anaemia hugely increased the chances of also having severe impairment (75-fold in fact), but, as was explained in the poster discussion session, anaemia is a consequence rather than a cause of severe kidney insufficiency.

A third study (Guest) found even higher rates of what it called nephrotoxicity (damage to the kidney tubules) in patients taking tenofovir. The Atlanta Veterans Authority Medical Centre found rates of nephrotoxicity of 17.1% in the first year on tenofovir, hypophosphatemia (low blood phosphate) in 13% and what it called renal insufficiency, as measured by creatinine clearance, in 4%.

However this patient group was very treatment-experienced, with 85% having had an AIDS diagnosis, and also quite old (mean age 47.8 years). The study also did not control for other nephrotoxic drugs, though amphotericin B (used for cryptococcal meningitis) increased the renal problem rate threefold.

A fourth study (Crane) looked at the patient cohort from the University of Washington Medical School HIV Clinic in Seattle.

Out of 2236 patients between November 2001 and September 2005, 320 had taken tenofovir. The study found that 245 tenofovir patients had no evidence of renal decline and 79 (24.7%) did. It found that patients over 50 (Odds Ratio 5.6 to 6.1) and patients who had also taken ddI (OR 2.0-2.5) had an increased chance of renal decline. So did patients taking lopinavir/ritonavir, according to one measurement method used. And so did patients with low body weight, with a 5% greater chance of renal decline for every kilogram less than average body weight for height and gender.

A final study (Thompson) was not set up to be a tenofovir study as such. The ESS40006 study assessed two doses of boosted amprenavir (Agenerase) in patients failing their current regimen. In addition all patients received abacavir (Ziagen) plus another NRTI (usually 3TC). The only variable in this unusual regimen was that NNRTI-naïve patients received efavirenz (Sustiva) while NNRTI-experienced patients received tenofovir (Viread) as a third NRTI instead, so Thompson could distinguish kidney function between patients taking tenofovir and ones taking efavirenz.

She found a statistically significant difference in GFR (see below) after 24 weeks on tenofovir, with patients on efavirenz actually having a GFR that improved by 12.66mg/dl compared with one that declined by 9.85mg/dl in patients taking tenofovir.

However the GFR rates did not appear to decline significantly further at 48 weeks, except in the most severe type of impairment.

One thing that became clear from these various studies is that there is no standard clinical marker for what constitutes renal impairment. The three most common measures are:

• Glomerular filtration rate (GFR), a measure of the kidneys’ ability to process fluid. A GFR of 109ml/minute is regarded as normal and under 90ml/minute is regarded as evidence of some renal insufficiency, but ‘severe’ impairment is measured as under 30ml/minute. Two different ways of measuring GFR, called the Cockroft-Gault Method and the Modified Diet in Renal Disease method, give different estimates of GFR.
• Serum creatinine; 0.99mg/dl is regarded as a normal measure, as is a clearance rate of this amount per minute. The Guest study regarded a 50% drop in the creatinine clearance rate as being evidence of renal insufficiency.
• Serum phosphate: the Guest study called at least one measure of less than 20mg/dl as hypophosphatemia. However it was criticised as taking just one baseline and one later measurement as evidence of this, rather than a series of measurements.

Conclusion

What one can say from all these various studies is that the rate of severe kidney problems in people taking tenofovir is low, though somewhat higher than in patients taking other drugs, and that the evidence that it is progressive is sketchy.

However a lack of agreed definitions as to what constitutes mild impairment and failure to control for the use of other drugs that may cause renal problems continues to make it difficult to assess whether it is something to worry about in a larger proportion of patients.