Antiretroviral therapy (ART) has revolutionized the management of HIV infection. A recent analysis estimates that currently available combination ART regimens have increased the life expectancy of HIV-infected individuals by approximately 24 years.1 The limitations of ART are plain to HIV clinicians and their patients. Although generally well tolerated, ART can be complicated by immediate and chronic adverse effects. Further, therapy is expensive and must be taken for years, if not for life. The astoundingly rapid ability of HIV to replicate and produce functional but mutated virus has presented the greatest challenge to the long-term control of the infection. In corrections, where HIV-infected inmates may pass in and out of prison and jails and have intermittent exposure to ART, HIV drug resistance is not uncommon. For correctional health providers managing HIV infection, an understanding of HIV drug resistance is essential. Below are some of the most commonly asked questions regarding ART resistance.
The prevalence of ART drug resistance has changed over time. Early in the epidemic, patients treated with zidovudine (AZT) or stavudine (d4T) mono-therapy quickly developed resistance to these drugs. Likewise, dual nucleoside reverse transcriptase inhibitor (NRTI) regimens used in the early 1990s also led to NRTI drug resistance, albeit at a slightly slower rate compared with mono-therapy. With the advent of protease inhibitors and use of triple combination therapy, profound reductions in HIV viremia were achieved. However, treatment failure rates, typically a consequence of suboptimal adherence to regimens requiring three times a day administration and/or large numbers of pills, were common. In a study from the Johns Hopkins HIV clinic in Baltimore, only 37% of their patients starting their first protease inhibitor (PI) based regimen between 1996 and 1998 had HIV viral load levels below the limit of detection (500 copies/mL).2 An analysis of drug resistance from this early era of potent HIV therapy found that two thirds of individuals in a representative sample of patients receiving HIV therapy in the US had HIV viremia of at least 500 copies/mL; of these 76% had evidence of drug resistant virus on testing.3
Of course, many of the patients developing resistant HIV had previously been exposed to suboptimal regimens, leading to the cultivation of drug resistance mutations of the virus that hamstrung their subsequent combination therapies. A recent study of drug resistance conducted at the HIV clinic at The University of North Carolina, found that starting therapy with a regimen that did not contain a PI or a non-nucleoside reverse transcriptase inhibitor (NNRTI) as part of a three-drug combination was a strong predictor of triple class (i.e. NRTI, NNRTI and PI) drug resistance.4 Interestingly, of those few patients who started on a triple drug regimen containing a PI or NNRTI but developed triple class resistance, almost all were on a PI that was not co-administered with ritonavir (i.e. un-boosted PIs).
Heavy reliance on highly potent ritonavir-boosted PIs and NNRTIs as the anchors of HIV therapy has led to a profound suppression of HIV replication, hampering the development of drug resistance. Studies of ritonavir-boosted PI regimens have consistently demonstrated that virologic failure to these agents is rarely associated with resistance to the PI, but rather to the companion drugs in the regimen.5-7 In contrast, virologic failure to NNRTI-based regimens is more likely to be accompanied by detectable resistance to the NNRTI, often in tandem with the M184V (see HIV 101) mutation that confers severely reduced susceptibility to lamivudine (3TC) and emtricitabine (FTC)8,9 (See Resources for link to a guide to reading HIV genotype resistance test).
Recently presented results from an AIDS Clinical Trials Group (ACTG) study in which treatment-naive patients were assigned to 2 NRTIs plus the NNRTI efavirenz (EFV) versus 2 NRTIs plus the boosted PI lopinavir/ritonavir (LPV/r) provides some insights into the frequency of drug resistance with current ART.8 After 96 weeks, both of these study groups experienced high levels of virologic suppression below 50 copies/mL (89% for EFV versus 77% for LPV/r; p = 0.003). Study defined virologic failure (a lack of 10 fold or greater drop in viral load, virologic rebound before week 32, failure to suppress to less than 200 copies/mL after 32 weeks or rebound after week 32) was observed in 94 of the 253 subjects randomized to LPV/r + 2 NRTIs and 60 of the 250 assigned to EFV + 2 NRTIs. For the LPV/r + 2 NRTI arm, 52 of these 94 had resistance testing results available and 8 had NRTI mutations detected but none had any major PI mutations evident. In contrast, 33 of the 60 subjects on EFV + 2 NRTIs with genotype results available had one or more NRTI mutations detected and 16 (48% of those with genotypes in this arm) had NNRTI resistance.
The take-home lesson from this trial and other studies of resistance to current ART regimens is that the great majority of patients treated with ART will achieve virologic suppression. Of those who do experience virologic failure, a substantial proportion does not have drug resistance evident at the time of failure. For some of these patients, a total lack of adherence could explain this observation as the absence of drug removes the selective pressure applied by ART and permits non-resistant wild-type virus to rebound. In other cases, the presence of drug resistance appears to be influenced by the composition of the regimen, with resistance rarely detected when regimens contain a ritonavir-boosted PI.
The dogma for several years has been that HIV-infected individuals receiving ART should take, at a minimum, 95% of the doses of medication prescribed. This canon of HIV management was rooted in the findings of an important study conducted in the late 1990s in a Veteran's Administration hospital in Nebraska in which adherence to HIV therapies was monitored using electronic medication bottle caps that recorded the opening of the medication bottle.10 In this study, adherence at 95% or better was associated with the least risk of uncontrolled viremia with only 22% of patients at this level of adherence having a viral load >400 copies/mL compared to 61% with adherence between 80% and 95%. Lower levels of adherence were associated with even greater rates of detectable virus. It should be noted that patients in this study had varying degrees of treatment experience; for some this was the first regimen and others were highly treatment experienced. Further, in this cohort, PIs were commonly used and were not boosted with ritonavir.
The advent of boosted-PIs and the emergence of NNRTIs since this study was conducted have likely shifted the required adherence level downward as these medications achieve high concentrations in the blood plasma, have relatively long half lives, and, in the case of boosted-PIs, are relatively 'resistant' to resistance. While it remains unclear just how adherent patients need to be to the potent therapies now available, there is some evidence that rates of virologic failure to these therapies can be low even when adherence falls below 90-95%. Provocative work by David Bangsberg and colleagues at UCSF suggests that adherence to our current first line therapies may not need to be absolutely in the range of 90%-100% for these regimens to achieve and maintain viral suppression. In his studies of a cohort of marginally housed HIV-infected men and women in San Francisco, he found that the risk of resistance to NNRTI-based regimens increased only when adherence dropped to below 54%.11 Above this level of adherence, the overwhelming majority of patients had viral loads of 12
This does not mean that patients should not be encouraged to take all their medications as directed. Patients should continue to be encouraged to take their medications exactly as directed. The San Francisco study results have yet to be replicated by others. Further, the study used 400 copies/mL as a threshold for undetectable rather than 50 copies/mL and low level virus may increase risk of resistance development over time. Individual variability in drug metabolism and other factors may make it dangerous to apply the aggregate data from this particular investigation to a patient. However, at the same time clinicians should appreciate that the 90%-95% adherence threshold we have held our patients to may no longer be justified when NNRTI- or ritonavir-boosted PI-based therapies are used in patients with susceptible virus. This is important when considering the administration of punitive responses to ART adherence levels that are below 90%. In some prisons and jails, adherence below a certain level may lead to treatment discontinuation. While the rationale for such an approach is sound, the adherence threshold chosen may need to be less rigid and reflect the available data on current potent ART regimens. These data suggest a more accurate level may be lower than 95% for NNRTIs and probably for boosted PI-containing regimens.
HIV drug resistance tests have been incredibly useful to clinicians managing HIV infection. However, as with any tool or test, it is important for the user to understand their limitations (See Resources for link to a comparison of HIV resistance test).
HIV resistance tests can detect a drug resistant virus that is present in sufficient numbers. If a resistant viral strain exists in very low numbers, the tests will likely miss it. Therefore, resistance can be present at low levels and not be picked up by the resistance tests. This is more likely to happen in several situations. For example, sometimes a resistance test is performed when there is very little virus around (i.e. a viral load of 1,000 copies/mL or less). As the overall amount of virus is low, the HIV drawn into the sample for testing may not be very representative of the viral population present in the circulation and resistant mutants may have been missed. Similarly, when a patient has developed drug resistance and then stops their medication, the wild-type (not drug resistant) virus re-emerges as it can now replicate freely with the removal of the HIV drugs. As the wild-type grows, it dilutes the population of the resistant virus, making it harder to detect. Importantly, even though the wild-type increases to great numbers, the resistant virus is not gone but continues to be present in low numbers and, under the right conditions, such as re-application of the HIV medications, can be selected to grow preferentially. It is currently believed that once a resistant virus is cultivated, it never disappears, but poses a lingering threat to the responses to future ART.
The answer to this question depends on where the person with HIV is infected. In major cities of the US and Europe, anywhere between 10-25% of people acquiring HIV are infected with HIV that is resistant to at least one HIV drug.13-18 In all studies, resistance is highest among the ART drug classes to NRTIs and increasingly to NNRTIs, followed by PIs. Data from New York, which may well represent a worst-case scenario, found an increase over time in acquired ART resistance among 361 individuals with acute or recent HIV infection diagnosed between 1995-2003.14 Comparing the periods of 1995-1998 and 2003-2004, rates of ART resistance increased from 13% to over 24%. NRTI resistance was the most prevalent but was relatively stable. PI resistance was fairly uncommon (1.3%) in the earlier period but rose to 7.1% of patients in the more recent period -- a trend that did not reach statistical significance. For NNRTIs, resistance paralleled their popularity with a significant rise from 2.6% to more than 13%. Other studies of recently infected patients in other locations generally report a slightly lower prevalence of drug resistance than found in this study.
Studies of drug resistance in chronically infected but treatment-naive patients have also been reported.13,17 In one study of 491 patients under care in 25 US cities from 1999-2001, 11% had at least one ART resistance mutation detected.13 Most of the mutations detected were those that reduce susceptibility to the NRTIs, especially the thymidine analoges and 3TC/FTC.
The rate of transmitted drug-resistance in non-metropolitan areas such as rural regions of the US South -- where most people with HIV infection in this country live -- is not clear. What has become evident is that acquired drug resistance has ramifications for subsequent treatment success. Several studies have observed increased rates of treatment failure in patients with pre-therapy resistance.19-22 Although wild-type virus may be able to out-compete resistant mutants, some resistance mutations can persist at levels that permit long-term detection by resistance assays; mutations conferring resistance to all three of the initial ART classes have been described in chronically infected, treatment-naive patients. For these reasons, the US Department of Health and Human Services (DHHS) guidelines on initial therapy of HIV infection, updated in October 2006, recommend genotype resistance testing be performed in all treatment naive patients prior to initiation of HIV therapy.23
In vitro studies and some clinical trials suggest that certain mutations of HIV in response to drug therapy may reduce the pathogenicity of the virus.24-26 To understand how resistance can impact the ability of the virus to replicate it is helpful to consider the dynamics of HIV replication in terms of basic Darwinian evolution. Within the body of an HIV-infected individual ART selects for mutated virus that can survive in a milieu includes these drugs. Continued pressure by the ART favors the persistence of such drug resistant virus. However, in many types of drug resistant mutations the virus evolves to evade the effects of drug therapy but comes with a cost to their ability to replicate relative to wild-type virus. That is, drug resistant virus, in becoming mutated, may not work as well as virus that does not contain drug resistance mutations.
In some cases, the reduced fitness of resistant virus can be exploited during therapy to slow the pace of disease progression. Specifically, in cases where there are few antiretrovirals to which the virus remains susceptible the continued use of certain agents to which the patient's virus is known to be resistant may be used to maintain levels of the relatively less 'fit' resistant virus. The best example of this effect is found in the M184V resistance mutation. This mutation essentially neutralizes any antiviral activity of 3TC and FTC. However, this mutation has been reported to inhibit the ability of the virus to replicate and may sensitize resistance virus to the effects of AZT and tenofovir. In one study, patients failing 3TC-containing therapy with a M184V mutation were randomized to continue their 3TC alone as mono-therapy or stop all ART.27 Those that continued on 3TC alone had a truncated rise in HIV viral load and reduced declines in CD4 cell counts compared to those who were no longer taking ART. Thus, continued 3TC had some effect on viral replication even though the M184V mutation was evident. Some clinicians take advantage of this phenomenon and maintain 3TC or FTC as part of salvage regimens in patients with drug resistance that includes the M184V mutant.
Whether resistance to other antiretrovirals can yield similar effects on fitness is being studied. There is some evidence that resistance to the thymidine analogues (AZT and stavudine), as well as the K65R mutation that can be selected for during therapy with tenofovir as well as some NRTIs, have a detrimental effect on viral fitness. What is clear though, is that it is probably not wise to maintain an NNRTI when resistance to this class of drugs is detected. Unlike the major NRTI mutations and the constellation of mutations that are usually needed to reduce susceptibility to the PIs, the primary NNRTI mutations seem to have less of an effect on the replication ability of HIV. This may explain why the K103N NNRTI mutation can remain detectable by genotype resistance testing long after NNRTI exposure. Further, early clinical trial data of a second generation NNRTI, etravirine (TMC-125), indicate reduced response to this new agent with an accumulation of NNRTI resistance mutations.28 The bottom line is that, in general, maintaining NNRTI therapy in the face of documented NNRTI resistance mutations should not be done. This is not to say that continued therapy with a PI or NRTI could not also lead to the accumulation of additional mutations once resistance has developed. It can, and continuation of any ART to which the virus is resistant has to be considered very carefully when effective treatment remains available.
The best way to manage resistance is introduction of drugs to which additional resistance has not developed. A number of new drugs in new classes are being developed and are expected to become available to patients within the next two to five years. These include inhibitors of the HIV integrase enzyme and the processes that lead to viral maturation within a CD4 cell, blockers of the CCR5 co-receptor and the CD4 receptor, as well as new drugs in existing classes. In addition, there are recently approved antiretrovirals that have activity against certain types of resistant virus that can be employed to craft a new regimen when a prior combination fails.
Whether using new novel agents, existing drugs, or recycling previously used regimens to which the virus is not considered to have resistance, the principle remains the same: use as many drugs in the new regimen as possible that have activity against the virus that exists in that patient. Studies of new antiretrovirals in treatment-experienced patients have consistently demonstrated that when these drugs are coupled with agents to which the virus remains susceptible, response to therapy is better.29-31
We are approaching a critical mass of potent therapies that can be used in such salvage regimens. With the approvals of tipranavir and darunavir -- PIs that are boosted with ritonavir and have activity against many strains of virus resistant to other PIs -- plus the availability of enfuvirtide (T-20), new regimens that offer a reasonable chance of success can be devised. Studies of both of these PIs demonstrate unprecedented responses in treatment-experienced patients, especially when used with other active drugs.29-31
Etravirine, the next generation NNRTI; MK-0518, the first HIV integrase inhibitor and maraviroc, a CCR5 inhibitor are all offered via expanded access programs. While these therapies will remain out of reach for most correctional facilities until FDA approval, they offer the promise of effective agents that, when available, can complete an attractive salvage regimen. For this reason, patients with multiple drug resistant virus may be best managed, when possible, by a delay in a change of therapy until one or more of these new agents becomes available.
While it is ultimately the patient's responsibility to take his or her medication, the clinician must chose regimens that are sound and are most likely to provide long-term viral suppression. Initial therapy should be prescribed with the goal of long-term viral suppression. There are antiretroviral regimens that have been found to be less effective than the currently recommended first-line therapies. Some, like the fixed dose combination of abacavir, 3TC and AZT have been found to be suboptimal compared to preferred regimens. Other combinations have been found to be outright dangerous (e.g. all other triple NRTI regimens except, perhaps, tenofovir + 3TC or FTC + AZT) and should never be prescribed. The US DHHS recommendation for ART for adults and adolescents is a user-friendly guide to initial therapy and lists clearly the preferred regimens, alternatives for special cases and completely contraindicated combinations.23 Unless there is an extremely compelling rationale, the drugs listed in the preferred regimens should be used in all cases of initial HIV therapy.
Baseline genotypic resistance testing prior to the start of HIV therapy is becoming less and less optional. The guidelines are clear on the utility of this test in reducing downstream problems for patients and a recent analysis suggests such testing is cost-effective.23 Detecting transmitted resistance or mutations that may linger from prior ART exposure can help guide proper treatment during the incarceration and after release. The price of the genotype resistance assays has come down, making cost less of a justification for ignoring the DHHS recommendations.
HIV clinicians have become accustomed to evaluating HIV-infected patients every three to four months. However, clinicians must be attuned to the development of changes in the viral load that may signal the emergence of drug resistance and act on these data prior to the next patient visit. Unexpected changes in viral load should prompt immediate reevaluation of the patient and the drawing of a genotype resistance test. At the North Carolina Department of Corrections, we are able to order a genotype and if the viral load is undetectable, the genotype is not run by the commercial laboratory -- thus, avoiding unnecessary billing. Prompt action can prevent further cultivation of resistance that can handicap future treatment options.
For patients new to the system, the greatest challenge can be determining what therapies they have been exposed to in the past. As tiresome as it is, obtaining a release of information and old records from outside providers can ultimately be time- and money-saving. In cases where little can be learned about the prior ART history, restart of the last regimen (if it is not some bizarre combination) can be attempted with a genotype obtained after two to four weeks to detect major resistance mutations that may lead to the overhaul of this regimen. So as not to perpetuate the 'black box' of HIV treatment history, inmates should be given a record of their medications prior to release. Wallet-sized cards that can list HIV medications and other essential clinical data are available from at least two pharmaceutical companies. 'Home-made' versions created by corrections staff can work equally as well. When possible, a listing of the major ART resistance mutations should be added to these cards for the benefit of the patient's future providers.
Lastly, salvage regimens should be created with considerable thought. Salvage HIV therapy generally yields diminishing returns with each subsequent combination less likely to be effective compared to the previous. New therapies may help increase the odds of treatment success beyond initial therapy but, it continues to be imperative that active agents not be wasted by being included in regimens that are predicted to be impotent based on resistance or patient history (i.e. if the patient was on AZT mono-therapy for six years in the 1980s, it is safe to assume they are resistant to this medication even if the resistance test does not detect AZT associated mutations). There should be a low threshold for consultation with an HIV expert when considering the management of the treatment-experienced patient. Outreach to such experts in the community, at academic medical centers or other correctional facilities should be sought and lines of communication established.
Resistance happens. However, resistance to HIV medications need not be inevitable. Potent therapies are now available in extremely convenient formulations and dosing schedules. Adherence remains a cornerstone of drug resistance prevention and correctional facilities have unique advantages in the monitoring and encouragement of treatment adherence. In addition, close surveillance of response to HIV therapy and quick action when viral load increases are detected can forestall further damage from evolving mutations. New drugs in existing classes that are already FDA approved and those expected to be shortly, hold the promise of a new chance for many patients who have developed HIV drug resistance. Wise use of these medications based on clinical trials data, patient history and detected and suspected ART resistance will increase the odds for treatment success.
David Alain Wohl, M.D., is Associate Professor of Medicine, Division of Infectious Diseases, AIDS Clinical Trials Unit at the University of North Carolina-Chapel Hill.
Disclosures: Grant Support: Abbott Laboratories, Gilead Sciences, Inc., Roche Pharmaceuticals, National Institutes of Health; Speakers Bureau: Gilead Sciences, Inc., Abbott Laboratories, Bristol-Myers Squibb, Roche Pharmaceuticals, Boehringer Ingelheim.