Adherence Strategies

23-plus years survival with fair and better adherence

Poor adherence has less than half survival time

Investigators intending to develop an economic model for HIV adherence made some surprising discoveries, including the conclusion that the initial antiretroviral therapy is the strongest one, but it doesn't matter which ART regimen is taken first.

"From a modeling perspective it really doesn't matter which one they take first," says Teresa L. Kauf, PhD, MS, an associate professor at the University of Florida at Gainesville.

"From a clinical practice perspective, if the first regimen is the best one, then you should think about what options are available for the patient and make the best choice based on tolerability, side effects, and things like that," Kauf says.

The study found that among HIV patients who were first treated with a non-nucleoside reverse transcriptase inhibitor (NNRTI), there was a mean survival of 25.38 years for those who achieved greater than 95 percent adherence, and there was a mean survival of 25.01 years for those who achieved 77–94 percent adherence, and there was a mean survival of 23.36 years for those who achieved 49–76 percent adherence. Patients who were adherent less than 49 percent of the time had a mean survival of 12.41 years.1

The findings were very similar for patients who started on a boosted protease inhibitor regimen, with mean survivals of 25.36, 25.10, 22.85, and 12.38 years, ranging from excellent adherence to poor adherence.1

"If you can get the patient to adhere, it really doesn't matter if you start with NNRTI or boosted PI," Kauf says. "Just put patients on a regimen that works for them."

It wasn't a surprise to find that patients with poor adherence would have less survival time than those with excellent adherence, but Kauf says she had expected to see a bigger survival difference.

"The data underlying this are based on clinical trials," Kauf says. "Patients come in and start on one regimen in the model, and those rates are based on clinical trials."

A portion of patients will have suppressed HIV even if they are not the greatest adherers, and then it will take a while for the suppression to fail. But once it does fail, the model assumes that the patient is switched to a new regimen where suppression can be re-established, Kauf explains.

"Essentially, we assumed that regardless of whether you started with NNRTI or boosted PI, if you failed that first regimen, you'd be put on the other one, and so patients would still come into contact with a boosted PI regimen," Kauf says.

The model's findings about adherence suggest only subtle differences in survival between excellent, good, and fair adherence.

"From this analysis it seems that even for patients who are not very good adherers, who have fair adherence of 49 to 76 percent, which most physicians would say is bad, it is worth taking the drugs," Kauf says.

The study also addresses quality of life by measuring quality-adjusted life years, finding that with either regimen introduced first, patients who have 95 percent or greater adherence have a mean of 15.13 quality-adjusted life years. Those with poor adherence of less than 49 percent have a quality-adjusted, life-year mean of 7.65 for NNRTI regimens and 7.63 for boosted PI regimens.1

"There are several studies trying to assess HIV patients' quality of life, and we used a utility measure of quality of life," Kauf says. "People will survey patients and classify responses according to CD4 cell levels and those with lower CD4 levels generally have lower quality of life."

Investigators used this weight, multiplied it by the survival years and came up with the quality-adjusted life years, she explains.

"Every time patients in the model spent three months in a particular CD4 category, we weighed that by this utility weight," Kauf says. "Then we add up all the months they spent in different health states, and that gives us the quality-adjusted life years."

The patients in the excellent adherence category still would be expected to have a 25-year survival, but the value of their life in terms of disease burden would make it equivalent to only 15 years of a disease-free life, she notes.

"So if your survival was going to be 25 years in various states of HIV infection, that would be equivalent to 15 years if you were completely healthy," she says. "Some people would trade that extra 10 years if they could be healthy."

For the quality-adjusted life years, the mean years for 77–94 percent adherence were 14.89 years for NNRTI regimens and 14.97 for boosted PI regimens; for those whose adherence was fair at 49–76 percent, the mean quality-adjusted life years was 13.97 for NNRTI regimens and 13.59 for boosted PI regimens.1

The model was not able to examine the impact of adherence and the initial use of nucleoside reverse transcriptase inhibitors (NRTIs), although that's an area investigators are working on, Kauf says.

"So that could be a factor, and it might be one reason why these numbers seem like less of a range than you might expect," Kauf adds.

"NNRTIs are easy to model because once a patient develops resistance, they're not put on NNRTIs again," she says. "With a boosted PI regimen, there still is some efficacy if you could get the patient put on a new NNRTI."

Since the study is based on a model and there are these drawbacks, it would be hard to predict what an individual patient would see in terms of treatment failure, Kauf notes.

"We're in the process of trying to validate the model, using data from a database project called CHORUS," Kauf adds. "This database followed HIV patients for several years and we'll use that data to validate some of the numbers from our model, including efficacy, side effects, and others."


  1. Kauf T, et al. Simulation modeling of adherence and resistance on long-term outcomes in HIV. Presented at the 44th Annual Meeting of the Infectious Diseases Society of America, held Oct. 12–15, 2006, in Toronto, Ontario, Canada. Abstract:978.