Systematic review of antihypertensive adherence interventions and investigation of antihypertensive drug discontinuation in the Glasgow Blood Pressure Clinic cohort

Bazuhair, Mohammed Abdulghaffar (2022) Systematic review of antihypertensive adherence interventions and investigation of antihypertensive drug discontinuation in the Glasgow Blood Pressure Clinic cohort. PhD thesis, University of Glasgow.

Abstract

Hypertension (HTN) is a major risk factor for cardiovascular diseases. Suboptimal medication adherence is a well-recognized factor contributing to the poor control of blood pressure (BP) in HTN. Estimated HTN prevalence in Scotland in the adult population from 2014 to 2017 for all age groups in both sexes was 58.7%. While systematic reviews have demonstrated improvements in adherence and clinical outcomes for atherosclerotic cardiovascular disease secondary prevention, the evidence for BP response is inconsistent. This systematic review evaluates the effectiveness of medication adherence interventions on antihypertensive related BP response.

We searched bibliographic databases between 2000 and 2019 using EMBASE, Medline, and Cochrane without language restrictions. Eligible studies were required to be randomised controlled trials, incorporating participants identified as having hypertension, and utilizing interventions aimed at improving adherence antihypertensive therapy with measurement of both BP response and adherence. Two reviewers independently determined the eligibility of studies and extracted data. Standardized mean difference effect sizes were calculated in random effect models. We identified 52 trials (n= 19673 participants).

The majority of trials (90%) were from high-income countries. Most had unclear risk of bias because of poor methodological descriptions. Studies which provided mean difference in SBP (N=13740 from 31 trials) or DBP (N=10665 from 29 trials) were included in the metaanalysis. The interventions resulted in significant reductions in both SBP (MD −2.55 mmHg [95% CI −3.87 to −1.23]) and DBP (-1.47 mmHg [CI −2.23 to -0.71]). Seven subgroups were analysed because of significant heterogeneity (SBP: I2=82%; DBP I2=73): intervention type, age, use of home BP, study duration, geographical factors, effect on Medication Adherence, health provider, care setting and geographical. Education intervention, studies of 6-month duration, and interventions that concomitantly increased MA showed significant reductions in SBP and DBP. Funnel plot demonstrated publication bias. Egger's Regression test was statistically significant with both SBP and DBP (p< 0.001). In term of medication adherence (MA), 12 trials (n= 6450) reported MD in MA were included in the quantitative analysis and divided into 6 groups based on the MA assessment method (MMAS-4, MMAS-8, MARS-5, Hill-bone, Medication refill and Electronic Drug Monitoring. Better effect on MA in studies that used MMAS-4, MMAS-8, and Hill-Bone self-report assessment methods. The studies that used the MARS-5 self-report assessment method and the studies that used Medication refill showed results favouring control. The Electronic Drug Monitoring group results showed extremely wide confidence intervals. Electronic drug monitoring presented as outliers in funnel plot. The variation among studies in terms of intervention, duration, sample size and population precluded any formal comparison between the included studies. Longer trials with clearly defined methodologies are recommended to establish whether adherence and BP control can be maintained with reduction in clinical events.

In chapter four, Glasgow Blood Pressure Clinic data were used to investigate the patterns of antihypertensive drug discontinuation in a large cohort of treated hypertensive patients. The study start date was Jan 1st 2006 and the study end date was March 31st 2013. The prescription data was classified into the following antihypertensive drug class groups – ACEI, ALPHABLOCKER, ARB, BB, DHP-CCB, NONDHP-CCB, DIURETIC, MRA, OTHER. The definition of a new antihypertensive drug class prescription was based on the patient showing no receipt of the drug class in the prescription data available from 2004 onwards. Discontinuation of a drug class was defined as the date of last prescription of the drug class with no receipt for at least 180 consecutive days. As a result, there were 4961 patients included in this study with average age 60.2 (14.5) years, greater proportion of females (53%). Alpha-blockers and ARB showed the highest discontinuation rates as first prescriptions. ACEI when prescribed as the first antihypertensive agent was 33.5% and this decreased to 27.5% when it was prescribed second. Alphablochers had a discontinuation rate of 44.1% when prescribed as the first antihypertensive agent and this declined to 20% when it was prescribed as the fourth agent. The significant predictors of drug discontinuation are age, Charlson comorbidity score and SIMD. Using Kaplan-Meier, the median time to discontinuation of any antihypertensive drug class was 4.1 years (95% C.I. 3.25-4.25 years) and the median time to discontinuation for all major drug classes were roughly similar except for ACEI which showed the lowest discontinuation rate. Additional analysis integrating other covariates and outcomes will help understand whether drug discontinuation or persistence have long term impact on morbidity and mortality.

In conclusion, the results of this thesis have shown that the current published literature on drug adherence interventions for hypertension is very heterogeneous limited by lack of consistency in definition and measurement of adherence or outcomes which may have affected the conclusions. Studies with significant results could have a greater likelihood of publication. Confounding and moderating variables are different in different studies. The clinical implications of this project require the generation of more generalisable results and this project has provided some direction for future research efforts. These include establishment of a uniform method of measuring adherence that is appropriate for hypertension, well-powered randomised and blinded studies. For retrospective studies, unselected population data representative of the hypertensive population across all age range and severity is crucial. Consideration for polypharmacy and multimorbidity is essential in any analysis plan.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Additional Information: Supported by funding from King Abdulaziz University.
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Medicine, Dentistry & Nursing
• Supervisor's Name: Padmanabhan, Professor Sandosh
• Date of Award: 2022
• Depositing User: Theses Team
• Unique ID: glathesis:2022-82706
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 17 Feb 2022 10:20
• Last Modified: 08 Apr 2022 16:50
• Thesis DOI: 10.5525/gla.thesis.82706
• URI: https://theses.gla.ac.uk/id/eprint/82706