Therapeutic drug monitoring of HIV-protease inhibitors to assess noncompliance

Objective: To determine plasma concentration ratio limits (CORALS) for HIV-protease inhibitors outside of which random plasma concentrations reflect partial compliance or noncompliance. In the absence of a gold standard for measuring compliance and to avoid complex techniques, measuring plasma concentrations may be an objective and easy way to check noncompliance. Methods: Pharmacokinetic curves after observed ingestion were recorded in patients on steady-state indinavir 800 mg TID (n = 22), ritonavir 400 mg/saquinavir 400 mg BID (n = 22, ritonavir; n = 17, saquinavir hard-gel capsules), or nelfinavir 750 mg TID (n = 4) or 1250 mg BID (n = 4). Concentration ratios were calculated by dividing the measured concentration by the median population value at the corresponding sampling time. Limits were based on the minimum P-05 (5(th) percentile) and maximum P-95 of these ratios found during the sampling interval. With these limits the authors determined (1) the proportion of patients falsely judged as noncompliers after observed ingestion, (2) discrimination between compliers and noncompliers, and (3) the absolute percentage of noncompliers. To judge the last two elements, two sets of random plasma samples were included: (1) samples sent in by the physician on suspicion of noncompliance (indinavir, n = 42; nelfinavir, n = 22;) or from a study population stating imperfect compliance in a questionnaire (ritonavir/saquinavir, n = 11); (2) control samples sent in routinely for monitoring therapeutic levels (indinavir, n = 41; nelfinavir, n = 201) or drawn from patients who stated perfect compliance in the questionnaire (ritonavir/saquinavir, n = 35). Results: The following CORALS were found: indinavir <0.23 or >3.3; nelfinavir <0.36 or >2.1; ritonavir <0.18 or >1.9; saquinavir <0.28 or >2.3. In 31% to 55% of the patients suspected of noncompliance a plasma concentration ratio outside these limits was found. If a ratio was outside the limits, there was a 68% to 88% chance that that plasma sample belonged to a patient suspected of noncompliance, compared with the controls (all Chi-squared tests P < 0.05). Compared with observed ingestion, these chances ranged from 87% to 92%. Conclusion: By using concentration ratio limits (CORALS), plasma samples of protease inhibitors with values outside these limits are highly indicative of partial or noncompliance.