Direct Oral Anticoagulants (DOACs) Use in Patients with Renal Insufficiency and Obesity

  • Ezekwesiri Nwanosike

Student thesis: Doctoral Thesis

Abstract

The popularity of Direct Oral AntiCoagulants (DOACs) for approved indications has risen dramatically following their introduction to the UK’s National Health Service (NHS) due to their convenience of dosing surpassing warfarin. However, prescribing these medications to high-risk patients has been challenging since mainly due to uncertainties around limited clinical trial data. Patients with chronic kidney disease and obesity pose a risk in particular as DOAC dosing was significantly affected by the variables such as, body weight and renal function. Due to the increased prevalence of CKD and obesity among the NHS patient population, the cost savings of preferring DOACs over warfarin was no longer
beneficial due to higher costs of mortalities and consequential morbidities (e.g., strokes and bleeding events). There are very limited interventional studies to rationalise the sample sizes to generalise findings. Therefore, a retrospective real-world data-driven approach was used in this thesis in an attempt to optimise the DOACs dosing regimen for patients with renal impairment and obesity.

The main data-driven techniques used in the thesis employed machine learning and multivariate logistic regression (The systematic review in Chapter 6 describes the potential of in-silico modelling). These were applied to a pre-processed dataset, carefully collected from Calderdale and Huddersfield NHS Foundation Trust Hospitals, and profiled accordingly. The methodology was executed in three phases: overall analysis of the full dataset, comprising different BMI categories (Chapter 3), the data analyses comprising patients with morbid obesity only (Chapter 4), and the analyses of the overall dataset comprising patients with different categories of renal impairment (Chapter 5).

The factors that influenced the clinical outcomes (such as mortality, ischaemic stroke, clinically relevant non-major bleeding (CRNMB), thromboembolism, length of stay, and emergency visits) in renal impairment and obesity were then determined following data analysis. Some of these factors, which included the individual DOACs administered, exerted a protective effect, while others worsened the safety and, or efficacy indicators. Also, it was found that some of the machine learning models employed in the thesis predicted the target (i.e., DOAC dose regimen) more accurately than others. Chapter 7 provides a discussion of the findings and makes reference and comparison with the existing evidence in the literature. More importantly, the results from patients with renal impairment and obesity were compared. Overall, the aim of generating real-world evidence for optimising DOACs safety and effectiveness in obesity and renal impairment was achieved. Our findings would support clinicians’ decision-making by reducing the uncertainty in DOACs prescribing.

There is a need to validate the thesis findings with well-designed prospective studies. There is also a need to explore pharmacometrics analyses and advanced data-driven techniques such as reinforcement learning to arrive at more precise DOAC dosing estimates for patients with renal impairment and obesity.
Date of Award23 Feb 2024
Original languageEnglish
SupervisorShahzad Hasan (Main Supervisor) & Barbara Conway (Co-Supervisor)

Cite this

'