Subtherapeutic Acetazolamide Doses as a Noninvasive Method for Assessing Medication Adherence

Abstract

Adherence monitoring is a vital component of clinical efficacy trials, as the regularity of medication consumption affects both efficacy and adverse effect profiles. Pill-counts do not confirm consumption, and invasive plasma assessments can only assist post hoc assessments. We previously reported on the pharmacokinetics of a potential adherence marker to noninvasively monitor dosage consumption during a trial without breaking a blind. We reported that consumption cessation of subtherapeutic 15 mg acetazolamide (ACZ) doses showed a predictable urinary excretion decay that was quantifiable for an extended period. The current study describes the clinical implementation of 15 mg ACZ doses as an adherence marker excipient in distinct cohorts taking ACZ for different "adherence" durations. We confirm that ACZ output did not change (accumulate) during 18-20 days of adherence, and developed and assessed urinary cutoffs as nonadherence indicators. We demonstrate that whereas an absolute concentration cutoff (989 ng/mL) lacked sensitivity, a creatinine normalized equivalent (1,376 ng/mg ACZ) was highly accurate at detecting nonadherence. We also demonstrate that during nonadherent phases of three trials, creatinine-normalized urinary ACZ elimination was reproducible within and across trials with low variability. Excretion was first order, with a decay half-life averaging ~ 2.0 days. Further, excretion remained quantifiable for 14 days, providing a long period during which the date of last consumption might be determined. We conclude that inclusion of 15 mg ACZ as a dosage form adherence marker excipient, provides a reliable and sensitive mechanism to confirm medication consumption and detect nonadherence during clinical efficacy trials.

Figure 1. NIDA trial Creatinine-normalized and Absolute Urinary Acetazolamide Concentration Data

Creatinine normalized (Panel A) and absolute concentration (Panel B) acetazolamide (ACZ) values for all participants in the NIDA trial (n=10). 15 mg ACZ/day was administered on Days −4 to −1. No drug administered during the nonadherence phase. Grey Overlay curves: represent the mean best fit predicted curves from two-compartmental pharmacokinetic analysis of each individual’s data collected after dose administration on day −1 (4^th dose day) ± 95% CI. Insets: Mean half-lives ± 95%CI calculated from output of two-compartmental pharmacokinetic analysis of each individual’s data collected after final dose administration on Day −1.

Figure 2:. VAMC Creatinine-normalized and Absolute Urinary Acetazolamide Concentration data sets.

Creatinine normalized (Panel A) and absolute concentration (Panel) acetazolamide (ACZ) values for all participants in the VAMC trial (n=10). Participants received 15mg ACZ/day during the adherence phase and no ACZ during the nonadherence phase Curve through Nonadherence phase: Non-linear regression analysis (bold line) ± 95%CI (dotted lines) of all pooled participant data based on samples collected after final dose administration on Day −1. Insets: Mean half-lives ± 95%CI calculated by of pooled participant data (day 1 to day 14) based on samples collected after final dose administration on Day −1.

Figure 3:. Normalized and Absolute Concentration Values for all Trials Overlaid

All participant acetazolamide (ACZ) data from different trials are overlaid to demonstrate reproducibility different populations. A: Creatinine-normalized ACZ data from the NIDA and VAMC trials B: Absolute concentration data from the NIDA, VAMC and UKY trials Solid horizontal line: denotes 5% cutoff used in statistical analyses. Dotted line marks day 0 (1^st day with no dose) Final two “adherence” days and “nonadherence” days 0-3 are expanded for ease of viewing. Subject AB12 data shown as chained filled diamonds, to illustrate intermittent adherence.

Figure 4:. Graphic Definitions of the Logistic Regression Performance Parameters reported in Table 2.

Graphic Definitions of the logistic regression performance parameters reported in Table 2. The parameter definitions highlighted as grey / dashed boxes overlaid on the test data sets, with the 5^th percentile Cr-normalized cutoff shown for illustrative purposes The data included in the test sets include the entire adherence phase (left of vertical dotted line) and days 1-3 of the nonadherence phase (Right of the vertical dotted line). Participants considered “nonadherent” only when 24h have passed after the time medication should be taken ie, day 1.