Pharmaceutical companies have a new buzzword this season: Cmin. This is the minimum concentration of a drug measured between one dose and the next, and it's become a focus of maximum competition between the manufacturers of protease inhibitors as they struggle to sell the merits of their products.
"Each year there's a new controversial topic in HIV treatment" says Andrew Hill, International Medical Manager of Roche. "It's like the Spring and Autumn collections. One year it's phosphorylation, then HIV RNA analyses. This season it's pharmacology".
At last month's International Workshop on HIV Drug Resistance in Sitges, Spain, Andrew Hill challenged competitors to come up with a common standard for talking about drug levels, and called for an end to the `mine's bigger than yours' spinning which has characterised much of the discussion over how to boost blood levels of protease inhibitors.
Why boost?
Boosting the levels of a protease inhibitor by using ritonavir has become a widely accepted strategy for improving the durability of protease inhibitor therapy. The rationale is that achieving higher drug levels can reduce the risk of drug resistance, or overcome low level drug resistance, and reduce the number of pills to be taken and the frequency of dosing.
However, not all drug boosting works in the same way, due to the differing ways in which the protease inhibitors are metabolised.
While ritonavir greatly increases blood levels of saquinavir or ABT-378, it has little effect on the peak levels of indinavir, amprenavir or nelfinavir, because these drugs are already well absorbed. Instead, dosing ritonavir slows the elimination of these drugs, allowing them to be dosed less frequently. Nevertheless, even a slight boost to the peak levels of indinavir can result in more cases of nephrolithiasis, the kidney complication caused by a build-up of razor-sharp indinavir crystals. The BEST study of ritonavir/indinavir versus indinavir three times daily found a significantly greater risk of this complication amongst the ritonavir group, although study investigator Jose Gatell argues that it can be managed by increasing fluid intake.
Another problem is that larger amounts of ritonavir are needed to boost these drugs, resulting in more frequent side effects. In contrast, Roche say they are confident that 100mg of ritonavir is enough to increase Fortovase levels with minimal side effects.
One of the chief benefits of drug `boosting', as Roche like to call it, is the ability to overcome some loss of sensitivity by swamping the virus with a massive increase in drug concentrations. However, Andrew Hill argues that we don't know the real Cmin or IC50s for protease inhibitors in drug naïve people, let alone in those with some degree of resistance, which can make it hard to interpret resistance tests and the claims of manufacturers regarding the impact of their products on drug-resistant virus.
Investigators are particularly interested in the ratio of the Cmin to the IC50 or the IC95 (the drug concentration needed to reduce viral replication by 50% or 95%), which indicates the degree to which drug levels remain above the threshold needed to maintain viral suppression.
Abbott Laboratories investigator Dale Kempf says that ABT-378 combined with ritonavir has a Cmin/IC50 ratio of 30, compared to ratios in single figures for other PIs when used alone. Similarly, Jon Condra of Merck has calculated that the Cmin/IC95 ratio of indinavir and ritonavir varies from 24.2 to 68.5 depending on the dose of ritonavir used. In comparison, the saquinavir/ritonavir combination had a ratio of just 1.7.
These comparisons are already surfacing in talks by scientists, but Andrew Hill and others I spoke to are sceptical about the validity of such comparisons.
"It depends on how you measure the Cmin, the IC50 or the IC95 and how you make the adjustment for protein binding", argues Andrew Hill.
For example, some people with very high Cmin values in a study group will skew the mean, so the median value is a more accurate measure. In addition, a study comparing HIV-positive individuals with healthy volunteers has shown that people with HIV absorbed much less indinavir, and had a Cmin nearly tenfold lower than their uninfected counterparts.
"The Cmin is a range of values in different patients so we need to know the variation in a population when we are looking at data", says Professor David Back. He agrees that comparing drugs on the basis of pharmacokinetic data is a minefield.
"Companies have got to be very careful. It's an issue that's got to be resolved, and we should aim for a uniformity in the way that this data is presented, or at least some transparency so that we know how companies made their adjustments for protein binding, for example. The IC50 can differ from one laboratory to another and from one cell type to another, regardless of protein binding, so we need to know these numbers too" says David Back.
HIV-positive individuals have much higher rates of protein binding of drugs than the uninfected medical students who typically volunteer for pharmacokinetic studies.
When Hill and colleagues calculated the IC50 for protease inhibitors using a variety of protein binding estimates, they came up with vast variations, suggesting that companies, clinicians and people with HIV will be unable to compare like with like until researchers standardise the methods they use for measuring drug concentrations.
For example, by fiddling around with the Cmin, IC95 and protein binding figures for the saquinavir/ritonavir combination, Andrew Hill was able to demonstrate Cmin/IC95 ratios ranging from 0.3 to 53.8, with a similar spread for other ritonavir-boosted PIs.
Mike Robertson of Merck Sharp and Dohme points out that Merck were the first to highlight the difficulties of making comparisons, and the need for standardisation.
"I think it is important to get the same labs doing the assay, and to standardise the methodology. To measure phenotypic resistance, we used highly resistant virus isolates gathered from participants in the early indinavir monotherapy trials."
He also questions the relevance of measuring or reporting the IC50.
"The IC50 is a sharper cut-off. If you look at the graph of plasma drugs levels, it occurs on the steepest part of the curve so it's more reproducible, whereas the range of values for the IC95 could be much greater because the curve has flattened out by that point. " Having said that, Mike Robertson is nevertheless satisfied with the accuracy and reproducability of the IC95 results produced from the Virologic phenotypic assay, and says "We think the IC95 is more clinically relevant. With the IC50 you are halfway on the way to failure."
"This is all virtual medicine and what we need are clinical trials which measure plasma levels of drug in the same way as phenotypic resistance. For example, I don't think we know what the trough level should be in order to suppress viral replication - we don't know the magic number so that's why we need to do the trials".