Print this page    •   Back to Web version of article

Treatment Interruptions: Reviewing What We've Learned So Far

By Kaleo Staszkow, M.D.

May/June 2006

Treatment of HIV is interrupted often and for many reasons. Technically, skipping a dose of your meds or even taking them an hour late is a treatment interruption of sorts. Not infrequently, patients have to weigh the advice of their physician to continue treatment against difficult personal or economic realities. Or a physician may recognize that a patient may temporarily be too distracted by medical illness, drug addiction, financial hardship, or another personal issue to make a commitment to taking HIV medicines as directed. Occasionally, meds have to be stopped due to a severe side effect or toxicity such as hyperlactatemia caused by NRTIs. In these sorts of situations, patients derive some clearly defined benefit from stopping treatment that they have decided outweighs the risks posed by untreated HIV. It is understood, however, that the price one pays for the benefits from interrupting treatment in these settings may be diminished control of HIV and progressive impairment of the immune system.

In contrast, the concept of strategic treatment interruption (or structured treatment interruption) is that treatment can be interrupted in some specific manner for the express purpose of improving outcomes such as restoration of control of HIV in patients experiencing virologic failure, reducing exposure to HIV medications, minimizing complications of treatment, or improving patients' well being through structured "drug holidays" and easier-to-take regimens that are as or more effective than standard therapy. Another compelling feature of any treatment strategy that is found to be as effective as standard, continuous HAART (Highly Active Antiretroviral Therapy), but uses less medicine over time, would be a substantial reduction in cost. For example, a fixed-length treatment interruption regimen with patients spending two months on and four months off meds would allow an AIDS Drug Assistance Program with a limited budget to provide three times as many people with antiretroviral medications each year. Note, however, that any treatment strategy that seems to save money initially but turns out to be less effective than standard HAART, will likely prove far more expensive to the state in the long run due to the cost of treating complications of HIV and AIDS which could have been prevented.

Treatment interruption as a strategy to improve clinical outcomes in HIV management was first explored as a potential means of combating multi-drug resistant virus in heavily treatment experienced patients. HIV becomes resistant to a drug when it mutates in a way that renders the drug ineffective. Resistance mutations, however, often come at a cost to the virus in terms of fitness. That is to say, resistance mutations can decrease the ability of HIV to multiply, but that without resistance mutations, HIV could not multiply at all in the presence of drugs. Remove the drugs from the equation, however, and the formerly crucial resistance mutation is now a liability to the virus. In this setting, wild-type (meaning without drug resistance mutations) virus, because it is more fit, rapidly becomes the predominant virus and resistant virus rapidly becomes undetectable.

It was therefore hoped that during a treatment interruption, competition with wild-type virus would rapidly eliminate drug resistant virus. Unfortunately, it has been demonstrated repeatedly and convincingly that treatment interruption, when undertaken in highly drug resistant patients, does not eliminate drug-resistance, which returns promptly upon restarting meds; and can, in fact, lead to additional resistance and more rapid progression of HIV. This is not to say that treatment should never be interrupted in patients with resistance to multiple drugs, but that treatment interruption in this setting does not work as a strategy for improving control of HIV and is likely to cause more harm than good.

Just because treatment interruption is not a viable strategy for achieving better virologic control in heavily treated patients with lots of drug resistance doesn't mean it may not be advantageous for other patients in other settings. Interrupting treatment when a patient's regimen is failing might produce entirely different results than when treatment is working very well. Accordingly, patients responding to HAART have participated in many recent clinical trials involving treatment interruptions. Among the key issues addressed by recent trials are the following.

Paramount in the design of any clinical trial is the safety of its participants. In a randomized, controlled trial of non-continuous treatment of HIV, the primary safety concern is that the experimental treatment strategy will prove to be inferior to standard therapy, posing the dangers of more rapid progression of HIV, increased risk for complications of HIV and AIDS, and drug resistance leading to increased difficulty controlling HIV when therapy is re-started. However, because long-term HAART, while life-saving, can be dangerous itself, investigation of potentially safer treatment strategies is warranted. Logically, any treatment strategy that involves reduced patient exposure to anti-HIV medications is likely to pose a significantly lower risk than standard therapy of treatment-related illnesses. Before randomized clinical trials with large numbers of individuals could be undertaken, a body of data from smaller studies and retrospective analyses was gathered that supported the idea that patients who have just started HAART and patients whose meds are working very well would not be subject to the safety concerns that pertain to the failed treatment strategy of interrupting therapy in order to restore drug efficacy in heavily treatment experienced patients.

In 2005, the International Study Group on CD4-monitored Treatment Interruptions published a study in the journal AIDS that prospectively evaluated 139 patients with undetectable viral loads and CD4 counts greater than 500 who agreed to interrupt treatment until their CD4 counts dropped to below 350. All patients had been on HAART for a minimum of 12 months and were initially started on treatment when their CD4 counts had dropped below 350 but not below 250. Patients were able to stay off therapy with minimal risk of progression to AIDS for a median time of 14 months. The study looked at only a small number of patients and larger trials are needed to confirm the results. Also, the study was not designed to identify strategic advantages to interrupting treatment. A particular strength of this study is that the key features of treatment history that define the cohort (viral load below 50 and CD4 above 500 after 12 or more months of treatment and CD4 nadir -- lowest point ever -- above 250) will be fairly common in large numbers of patients receiving standard therapy for HIV in "real world" settings outside of clinical trials. Most patients, for example, who are being treated according to current CDC guidelines would share these characteristics. The results would seem to suggest, therefore, that for a very large number of patients, interrupting treatment might at least be done safely. However, whether or not this kind of CD4-guided treatment interruption offers any strategic advantage in patients' lifelong fight against HIV remained unclear.

One important feature of any non-continuous treatment strategy should be a decrease in the adverse effects of anti-HIV medications. Standard HAART, given continuously, can keep HIV under control and restore immune system health indefinitely. However, a growing number of illnesses are now known or suspected to occur more frequently in patients on HAART. Among these are cardiovascular disease, including heart attack and stroke, liver disease, kidney disease, blood cell disorders, and lipodystrophy and lipoatrophy. The latest news on treatment interruptions, from this year's Conference on Retroviruses and Opportunistic Infections (CROI), presented a setback to side effect management.

The SMART trial was designed to prove that even if patients who interrupted treatment experienced slightly worse outcomes in terms of progression of HIV, that this would be offset by significant reductions in the rates of treatment associated adverse events. A total of 6,000 patients were to have been recruited and randomized to receive either continuous HAART or to stop meds at the beginning of the study. The patients in the treatment interruption arm would then resume treatment when the CD4 count fell below 250 and then interrupt therapy again when the CD4 count rose above 350 and so on. The study was halted in January, however, when preliminary review of the first two years of data confirmed that patients who interrupted treatment were not only more likely to experience HIV disease progression, but were also more likely to experience serious treatment-related adverse events, namely heart attack, stroke, coronary artery disease requiring surgery, and kidney or liver disease. So in this large randomized trial (5,472 patients were enrolled at the time the study was stopped), designed specifically to evaluate CD4-guided treatment interruptions as a treatment strategy, the data strongly suggest that not only is interrupting treatment in this way unsafe because HIV is less well controlled, but that it is also futile. In other words, treatment interruption did not reduce the risk of treatment-related adverse events as hoped. Instead, patients in the treatment interruption arm had a significantly higher risk of heart, liver, and kidney problems and nearly double the likelihood of death (1.7% vs 0.9%) than those on continuous therapy.

Data from SMART make a very strong case against CD4-guided treatment interruptions, though the possibility remains that results might be better with higher CD4 cut-offs. Also, other treatment strategies are being studied. Several other clinical trials evaluating various non-continuous treatment strategies have presented data this year.

The Staccato study randomized 558 patients on HAART (most on a regimen that included a boosted protease inhibitor drug) to one of three arms: week on-week off (patients in this arm took their medicines every other week), CD4-guided therapy in which patients stopped and started meds when their CD4 count went above or below 350, and standard, continuous therapy. The week on-week off arm was stopped early due to inferior performance. The patients in the CD4-guided treatment interruption arm had more HIV-related diseases such as thrush and lower CD4 counts. On the plus side, patients in this arm also experienced less diarrhea, less neuropathy, and had lower cholesterol. There was no difference in virologic control or the emergence of drug resistance in the two groups.

In the media, a big deal has been made about comparing findings from SMART and Staccato and the two sets of data have even been declared contradictory by some. Taken in context, however, the data from SMART which show that CD4-guided therapy is dangerous because of a greater risk of HIV progression combined with an increased risk of treatment-related medical problems, says something altogether different about interrupted treatment than the data from Staccato. While Staccato does show that treatment interruptions may not cause increased rates of drug resistance (especially if there's a boosted protease inhibitor in the regimen), it failed to show that the patients in the treatment interruption arm had any strategic advantage in terms of safety or efficacy compared with the patients on standard therapy.

The Trivacan trial conducted in Ivory Coast by the French National Agency for AIDS Research was designed to compare a CD4-guided treatment interruption arm with a fixed-length treatment interruption arm. The 840 participants (77% women), who had been on HAART a minimum of six months with a minimum CD4 count of 350 and a maximum viral load of 300, were randomized to CD4-guided treatment using CD4 counts of 350 and 250 to stop and re-start treatment, or to a fixed-length arm of two months on -- four months off, or to standard, continuous HAART. Ninety-one percent of participants were on AZT [Retrovir] + 3TC [Epivir] + Sustiva. In October 2005, the CD4-guided treatment interruption arm was stopped by the data safety monitoring board because serious illness was significantly more frequent among patients in this arm. Recently presented 19 month follow-up data at CROI showed that the over-all incidence of serious illness was 2.6 times greater in the CD4-guided treatment interruption arm than in the continuous treatment arm. Those receiving CD4-guided treatment were 16 times more likely to be diagnosed with invasive bacterial infections and 85% of these infections were caused by antibiotic resistant bacteria. They were also three times more likely to have severe thrush and 1.5 times more likely to be diagnosed with TB (tuberculosis). No data was presented from the fixed length (two months on-four months off) treatment interruption arm which is continuing.

Another trial, however, with a similar fixed-length treatment interruption arm was terminated on March 15 after a review of preliminary data revealed that patients in the treatment interruption arm were significantly more likely to experience HIV related illnesses. DART (Development of Antiretroviral Therapy in Africa) randomized 799 patients with a minimum CD4 count of 300 after 12 to 18 months of HAART to either standard, continuous HAART or to a 12 week on-12 week off fixed length treatment interruption arm. The rate of HIV-related illness (the most common being esophageal Candidiasis) was over four times greater in the treatment interruption arm.


Treatment interruptions will always have a place when certain specific interests of an individual patient conflict with the need to treat that individual's HIV or when individual patients experience serious medication toxicities or intolerable side effects. However, if the goal is treating HIV in order to allow patients to live free from the many dangers of immune system impairment and to prevent the premature disability and death that inevitably occurs with untreated AIDS, then it must be said that no treatment interruption strategy has yet been demonstrated to be superior to continuous treatment with HAART. This is not to say that future clinical trials may not identify more successful treatment strategies. Also, with the development of new classes of drugs such as entry inhibitors and integrase inhibitors, the development of successful non-continuous treatment strategies may be greatly enhanced. A successful therapeutic vaccine or some novel approach to treating HIV that has yet to be conceived might very well set the stage for the ultimate treatment interruption. At present, un-interrupted therapy with HAART seems the best way to ensure that people who need treatment for HIV today will be around when that happens.

Kaleo Staszkow, M.D., specializes in providing comprehensive primary care for adults with HIV at NorthStar Medical Center on Chicago's North Side. He is also an investigator in several clinical trials of novel therapies for HIV and HIV-associated lipodystrophy. He may be reached via e-mail at

Got a comment on this article? Write to us at

This article was provided by Positively Aware. It is a part of the publication Positively Aware. You can find this article online by typing this address into your Web browser:

General Disclaimer: is designed for educational purposes only and is not engaged in rendering medical advice or professional services. The information provided through should not be used for diagnosing or treating a health problem or a disease. It is not a substitute for professional care. If you have or suspect you may have a health problem, consult your health care provider.