NNRTIs more forgiving for occasionally forgetful patients; but are they the best option for patients with major adherence issues?

Individuals who first obtain an undetectable viral load on highly active antiretroviral therapy (HAART) and then adhere to the regimen between 76% and 99% of the time have less risk of ‘failing’ on a non-nucleoside (NNRTI)-based HAART combination, compared with a PI-based regime. These findings are the conclusion of a prospective Italian study and are published in the January 1^st online issue of Clinical Infectious Diseases.

However, an accompanying editorial warns that NNRTI-based therapy is not always the best choice for patients who have adherence issues, suggesting that since NNRTI-based HAART is more forgiving of the occasional late dose, but less forgiving of unplanned multiple drug holidays, individualised assessment is needed at the clinic level due to the “multidimensional problem of adherence.”

This study of 543 patients, set in an HIV outpatient clinic near Milan, Italy, sought to ascertain the effect of adherence to various HAART regimens on the risk of virological failure (which was defined here as two consecutive viral load tests above 500 copies/ml). Adherence was determined at baseline using a one-off self-reported questionnaire that asked about missed doses and timing in the previous four days.

All patients were on a stable HAART regimen for six months or longer, and had a viral load of less than 50 copies/ml at enrolment into the study. The majority (314) of patients were taking NNRTI-based HAART: 178 on efavirenz (Sustiva) and 136 on nevirapine (Viramune). Of the 153 patients on HAART that contained a protease inhibitor (PI) only about one-third were taking a boosted PI regimen: 33 on ritonavir-boosted lopinavir (Kaletra), twelve on ritonavir-boosted saquinavir (Invirase, Fortovase) and twelve on ritonavir-boosted amprenavir (Agenerase). Sixty-six patients took nelfinavir (Viracept) and 33 took indinavir (Crixivan) alongside two nucleoside reverse transcriptase inhibitors (NRTIs). A further 76 patients took triple NRTI therapy.

Although the patient population included individuals who were taking HAART for the first and second time, as well as patients who were on salvage regimens, and there was also a mix of patients who had previously experience virological failure alongside those who had no previous virological failure, these variables were not found to be statistically significant when examining their correlation to the study’s virological success or failure rate.

The only two factors that were significantly associated with virological success or failure were the type of HAART received (p=.037) and the rate of adherence reported at baseline (p<.0001 analysis.="" and="" both="" in="" multivariate="" this="" true="" univariate="" was="">

The study found a linear association between the reported adherence rate and the rate of virological failure in the six months after the baseline visit. These were:

• Adherence 75% or less: rate of virological failure = 17.4%
• Adherence 76-85%: rate of virological failure = 12.2%
• Adherence 86-95%: rate of virological failure = 4.3%
• Adherence 96% or more: rate of virological failure = 2.4%

However, the risk of virological failure differed depending on the type of HAART regimen taken. For example, adherence of up to 85% to a PI-based regimen resulted in a virological failure rate greater than 20%; whereas adherence of less than 75% to an NNRTI-based regimen resulted in a virological failure rate greater than 10%.

When they represented these results in terms of risk (using odds ratios, OR), the investigators found that NNRTI-based HAART had a reduced risk of virological failure (OR 0.271: 95% CI, 0.116-0.635) compared with PI-based HAART.

For those with a low adherence rate (i.e. less than 75%) the OR was 0.379, but with a wide confidence interval from 0.088-1.639, and therefore not significant. However, for patients with an adherence rate of greater than 75% the OR was 0.309 (95% CI, 0.105-0.911). The results were confirmed even when the one-third of patients with self-reported 100% adherence rates were excluded from analysis.

The investigators also observed a statistically significantly different (p=.018) adherence rate between patients on PI- and NNRTI-based regimens (mean adherence rate 89.8% vs. 93.6%, respectively). As with previous studies they found that the number of pills in the regimen (p=.021) and the number of daily doses were the two significant factors of adherence rates; once-daily regimens (mean adherence rate 94.8%) were adhered to greater than twice-daily regimens (mean adherence rate 92%; p=.042) and greater than twice-daily regimens (mean adherence rate 91.9%; p=.009).

In an accompanying editorial, Dr Hernando Knobel argues that these findings suggest a “possible benefit of the use of an NNRTI as therapy for ‘forgetful’ patients and, consequently, as the first choice of therapy for previously nonadherent patients or patients suspected of being nonadherent [and] should be discussed in more depth.”

However, he goes on to cite other studies that found a high rate of virological failure in patients on NNRTI-based HAART who took frequent drug holidays. This is due to the long half-life of NNRTIs, which is forgiving of occasional missed or late doses, but which can result in drug exposure to replicating virus, and therefore the emergence of resistance, when NNRTIs are frequently stopped - and the restarted - for longer than several days at a time. “From a clinical point of view,” he writes, “a paradoxical situation exists. A very potent, convenient, and easy-to-take class of drugs (i.e. NNRTIs) is very safe and effective for adherent patients; however, for patients with a high risk of low adherence – in particular, patients who take intermittent therapy – treatment involving this class of drugs is not the best option.”

He concludes, therefore, that, “nowadays, great emphasis should be placed on achieving full adherence in the long term by using different strategies according to the clinical scenario. Appropriate selection of drugs for treatment, which is only one aspect of the multidimensional problem of adherence, must be a carefully made and individualized (sic) decision base on sound clinical judgment (sic).”