The development, validation, and evaluation of quantitative assays for determining adherence of heart failure patients to carvedilol, enalapril and perindopril

Doctoral Thesis


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title
Background: Heart failure (HF) is a global pandemic with a rising prevalence rate in low- and middle-income countries (LMICs). Poor medication adherence contributes to the impact of chronic diseases such as HF. However, there are sparse adherence data on HF patients in sub-Saharan Africa (SSA). This is problematic as African HF patients have a high mortality rate, which is poorly understood. Poor medication adherence could contribute to the high mortality rate of African HF patients. Objective adherence measures are better than subjective measures (for example, patient recall) at predicting outcomes. In addition, the adherence method should be applicable to resource-scarce settings. Novel multiplex assays were developed to quantify carvedilol, enalaprilat and perindoprilat in dried blood spots (DBS) and correlated with plasma. Carvedilol, enalapril and perindopril are medications commonly used to treat HF, with enalaprilat and perindoprilat being the active metabolites of enalapril and perindopril, respectively. The developed assays were then evaluated in terms of their ability to discern between non-adherent and adherent patients and their suitability for use in resource-scarce settings. Method: The DBS and plasma assays were validated per the United States Food and Drug Administration (FDA) guidelines. The plasma assay was validated over a calibration range of 0.2–200 ng/mL for carvedilol, enalaprilat and perindoprilat. The DBS assay was validated over a range of 1.00–200 ng/mL for the three analytes. The DBS assays were correlated with plasma concentrations in a pilot intensive pharmacokinetic study of six patients. The correlation was determined using Deming regression, with Bland–Altman analysis used to establish agreement between observed and calculated plasma concentrations. Calculated plasma concentrations were obtained using the Deming regression equations describing the relationship between DBS and plasma concentrations. Results: Accuracy, precision, selectivity and sensitivity were proven with complete and reproducible extraction recovery at all concentrations tested for both assays. Stability of the analytes in the matrix and throughout sample processing was proven for both assays. The full range of plasma pharmacokinetic samples could be quantified for all analytes, with the lower limit of quantification (LLOQ) of 0.2 ng/mL proving to be sufficient. The pharmacokinetic pilot study's full range of DBS concentrations could be quantified for enalaprilat but not for carvedilol and perindoprilat. The LLOQ of 1.00 ng/mL was not sensitive enough to quantify the lowest concentrations of some patients for these two analytes. Good correlations were observed between DBS and plasma pharmacokinetic samples, with the Pearson's correlation coefficient (r) greater than 0.94 for all analytes. The difference between the observed and calculated plasma concentrations was less than 20% of their mean for > 67% of samples for all analytes, indicating good agreement between observed and calculated plasma concentrations for all analytes. Conclusions: The plasma assay is suited for evaluating patient adherence to carvedilol, enalapril and perindopril medication. The assay is robust and sensitive enough to discern between those who are adherent and non-adherent. Due to the wealth of pharmacokinetic data available for the analytes in plasma, through pharmacokinetic modelling, it is possible to determine the most appropriate dose and weight-specific adherence interpretation for that patient rather than relying on a general cut-off value. In other words, adherence interpretation can be individualised based on a patient's own dose and weight. Plasma as a matrix, however, is not very amenable to resource-scarce settings. The matrix requires strict storage and transport conditions, so creating additional logistic difficulties and expenses in resource-scarce and remote locations. These are difficulties that would have to be accommodated to use the assay. It was found that the DBS assay is more suitable as a screening assay for carvedilol and perindoprilat than as an assay to gauge adherence. The assay is suitable as an adherence determining assay for enalaprilat, however. The prolonged terminal half-life of enalaprilat allows sufficient DBS concentrations to track adherence. The DBS assay's higher LLOQ and the higher concentration of the analytes in plasma versus that of whole blood places the assay at a stark disadvantage in terms of sensitivity relative to the plasma assay. DBS samples have a significant advantage over plasma samples in their less stringent storage and transport requirements. As a matrix, DBS is far more conducive to remote and resource-scarce areas when compared to plasma. The robustness of both assays was proven with cross-validation using actual clinical samples. Good agreement between observed and calculated plasma concentrations means that DBS concentrations, once normalised, can be used interchangeably with plasma samples. DBS samples can be collected at the sampling sites, taking advantage of the DBS matrix's less stringent storage and transportation requirements. Once the samples are analysed, the concentrations can be converted to plasma concentrations, which can be interpreted more efficiently in terms of adherence. However, this would only be feasible for enalaprilat, as the DBS assay for the carvedilol and perindoprilat analytes lacked sensitivity to reflect ingestion within the last 24 hours.