Therapeutic drug monitoring: may predict nelfinavir failure, even when viral load controlled

Therapeutic drug monitoring (TDM) of nelfinavir levels may predict less durable response, even in patients with early virologic suppression, according to findings from a French cohort published in the Journal of Acquired Immune Deficiency Syndromes this month.

The study reported on the relationship between durability of virologic response and drug blood levels in 407 patients starting their first protease inhibitor regimen with either indinavir or nelfinavir. All subjects included had successful viral load suppression below 500 copies four months after starting PI-containing treatment.

Patients from a larger cohort were selected for inclusion if the time between dosing and the blood level determination had been recorded at the month 4 visit. Measurements of indinavir and nelfinavir levels were performed by high-performance liquid phase chromatography on frozen samples taken at that visit. Interpatient variability in the time between dosing and blood draw was adjusted using a model that predicted expected blood drug concentrations over time for each drug, then expressed each individual’s result as a ratio of the observed to the predicted concentration (O/P). A lower ratio indicates a possibly sub-optimal drug concentration and a hypothetically greater risk of subsequent loss of suppression. The O/P ratios were grouped by quartile and patient outcomes analysed within these groupings over a median of 20 months follow-up.

During follow-up, equal proportions (29%) of patient within each drug group had experienced loss of suppression (49 of 167 patients treated with nelfinavir; 69 of 240 treated with indinavir). Among those on indinavir, there was no association between the O/P ratio and risk of treatment failure by either univariate or multivariate analysis. In nelfinavir-treated patients, using drug concentrations alone, there was a non-significant trend to having a greater risk of virologic rebound among those with the lowest O/P ratio. By multivariate analysis, with adjustments for prior antiretroviral therapy, adherence, baseline CD4 count and viral load, a significant risk was observed between being in the lowest quartile of O/P ratios and having loss of viral suppression (RH = 2.20; 95% CI = 1.19-4.04, P=0.01).

No strong signals were detected in this observational study, and the authors caution that their simple modeling of the pharmacologic data may be a limitation. The cutoff of 500 copies/mL for HIV RNA does not adequately report the depth of the response at four months and it would be helpful to know the proportion of patients with virologic response below 50 copies at that time point; perhaps that would be a more sensitive predictor of durability. Yet the authors conclude that it may be useful to monitor drug levels in every nelfinavir-treated patient, even those with successful virologic response, and to adjust dosages to achieve optimal concentrations. However, they also note that nelfinavir concentration variability may be due to lack of adherence to recommendations that the drug be taken with food, and advise that nelfinavir food restrictions be more strictly emphasised.