September 16, 2003
Dr. Peter Reiss from the University of Amsterdam presented a review of HIV resistance testing and therapeutic drug monitoring (TDM) in clinical practice.
He began his review with a discussion of the factors associated with the development of resistance. They include patient factors such as adherence levels and drug disposition; drug properties such as pharmacokinetics (half life), potency and interactions with other meds; and viral factors such as transmitted resistance virus and the genetic barrier required to confer resistance (i.e., the number of mutations required to confer resistance).
First he focused on how these tests should be used in treatment-naive patients. He emphasized the importance of the initial treatment regimen -- this is typically when we see the highest level of success. So, all available tests at our disposal should be used to maximize the treatment response at this crucial point.
Resistance and drug exposure are linked. There is usually a minimal concentration required to fully suppress virus replication. However, most HIV drugs have a narrow therapeutic window, meaning that the concentration required for viral suppression is extremely close to the concentration that results in significant toxicities. Resistance development is thought to result in part from sustained inadequate drug concentrations. Preventing inadequate drug levels, perhaps through the use of TDM, can optimize drug concentration and hence treatment effect.
Dr. Reiss then illustrated the magnitude of the resistance problem. Several studies recently published or presented show the prevalence of drug resistance in newly infected patients in the U.S. to be around 5-15 percent and up to 50 percent in the HIV-treated population. Much of that level of resistance is thought to result from prior suboptimal treatment regimens.
However, he felt that there is a lower risk of resistance development with contemporary regimens because of HAART's increased potency. In most patients infected with resistant virus, there is also some wild type virus without mutations. In the absence of HIV medications, the resistant viral strains recede into the background below the level of detection by current resistance testing assays, but they still continue to exist.
Recent studies have also shown that some resistance mutations disappear within a year of infection, while some are still detectable even after two years of infection. This has resulted in a modification in the resistance testing guidelines that now calls for the use of testing in drug naive patients, up to two years after acute infection.
Dr. Reiss discussed the case for TDM in clinical practice. A recent review indicated that: there is extensive interpatient variation in HIV drug concentration; the range of drug concentration that is tolerated and effective is narrow; and accurate assays to measure the concentration of HIV drugs are now available, and in most cases validated. He stated that there may be some limitations when using TDM, namely that: inadequate levels may not reflect insufficient dosing; intrapatient variation may require multiple samples to be taken from a given patient; most assays measure total blood plasma concentration and not the protein-unbound fraction, which is the active amount of the compound; there is uncertainty regarding whether pharmacokinetic sampling can predict response or target concentration; and that these issues may be accentuated in the setting of drug resistant virus.
Dr. Reiss presented data from the ATHENA study (click here for a summary of this study), which was a TDM study conducted in the Netherlands. In this study, patients received either indinavir (IDV, Crixivan) or nelfinavir (NFV, Viracept) as the protease inhibitor (PI). PI levels from random samples were determined using high pressure liquid chromatography (HPLC). Because these were random levels taken from patients who could have ingested their PI dose at any time, a concentration ratio (CR) consisting of the individual concentration divided by the time-adjusted population value was used to correct for the possibility of variability in PI levels. Practitioners either got a TDM report or no TDM report. Results showed that the rate of discontinuation of the regimen was lower and the treatment response better in the arm in which practitioners received the TDM report and could make dose modifications of the PI as necessary.
As an example of how TDM can be useful, Dr. Reiss mentioned that there may be distinctions between men and women when it comes to efavirenz (EFV, Sustiva). In general, women had higher trough levels of efavirenz, which were closer to what would be considered the toxic range.
But resistance and TDM testing is not only important for naive patients. Dr. Reiss briefly reviewed clinical trials which studied resistance testing in treatment-experienced patients. Several studies support the use of testing, irrespective of whether one is using genotypic or phenotypic assays. However, several studies did not show a clinical benefit. These studies all had different study designs and Dr. Reiss pointed out that the clinical benefit might have been compromised by significant cross resistance among drugs in the same class, making the interpretation of the resistance test information difficult. He emphasized that resistance test information for the current regimen is the most accurate; that the absence of resistance does not rule out resistance, particularly for drugs that the patient has had prior known exposure; and that pharmacokinetic enhancement (i.e., increasing the PI dose) may compensate for some PI-associated resistance. He also said that if regimen failure or resistance is suspected because of an increasing viral load, confirmation should always be done by repeating the viral load test. Plus, of course, he pointed out that there is no substitute for a good antiretroviral history. In the absence of resistance, particularly with a failing regimen containing a PI, one should suspect inadequate exposure and consider a dosage adjustment.
Dr. Reiss closed his presentation saying that perhaps using an inhibitory quotient (IQ) -- which is defined as the trough concentration of an HIV drug divided by the effective or inhibitory concentration required to inhibit 50 percent of the patients virus strain -- might be a more accurate marker for predicting a drug's efficacy. There is actually a Web site, www.HIVpharmacology.com, that describes the current accepted trough concentrations for all HIV drugs that might be useful. Plus, there are draft TDM guidelines coming. And finally, Dr. Reiss reiterated that both resistance testing and TDM are important tools that should be used as adjuncts to good clinical judgment, and still require further study.
Dr. Joseph Eron, Director of Clinical Core UNC Center for AIDS Research, Dr. Ruth Tuomala from Brigham and Women's Hospital in Boston and Dr. Andrew Carr of St. Vincent's Hospital, Sydney were part of a roundtable discussion on the topics discussed by the speakers during this symposium. Several questions were asked by audience members, for which one of the speakers then answered the question. Below you'll find a brief description of some of the questions asked.
One person asked what would be a sufficient time period after starting T-20 (enfuvirtide, Fuzeon) to determine whether the regimen was effective or not. Dr. Eron felt that four weeks might be too short of a treatment duration to see a response. He thought that eight to 12 weeks might be more acceptable. The response would be dependent on things like the genotypic susceptibility score (GSS), which is a number that predicts how many other HIV drugs still have activity against the patient's virus; how many new or recycled drugs were included in the regimen along with T-20; and the patient's prior antiretroviral history. Dr. Eron went on to say that perhaps a practitioner should wait, if clinically possible, until there were more additional treatment options if T-20 could not be given with other currently available drugs that still had activity.
Someone else asked about the relative role of genotyping versus phenotyping versus virtual phenotyping. Dr. Reiss felt that phenotyping did not add significantly to genotyping given the increased cost of phenotyping. Dr. Eron added that there is evidence that virtual phenotyping might have some advantage over true phenotyping. This is probably the result of true phenotyping only having the results of a single patient's viral strain, whereas virtual phenotyping is essentially a genotype that then utilizes a rather large database to match many phenotypes to the patient's specific genotype resulting in the increased predictive power of phenotypic information.
Someone else asked about why there would be increased maternal risk of death in HIV-infected mothers who were breast-feeding. Dr. Tuomala felt that this was probably primarily related to malnutrition in the mothers. Someone then wondered about the use of TDM in HIV-infected pregnant women. Dr. Tuomala and Dr. Reiss stated that there was a serious lack of information regarding HIV medication drug levels during pregnancy. They briefly mentioned that nelfinavir levels are known to be affected.
Finally a question was asked about the mechanisms causing lipodystrophy. Dr. Carr briefly stated that there is no good animal model, and that there have been insufficient in vitro studies to determine the exact mechanisms. Some in vitro results have been different than what has been seen in patients. Some cytokines appear to play a role in the development of lipodystrophy, but what role tumor necrosis factor (TNF) may play is unclear.