Atema, Kris (2022) Prediction model of clinical outcome after transcatheter aortic valve implantation (TAVI) with incorporated blood-based biomarkers. thesis, Medicine.
Full text available on request.Abstract
Introduction Transcatheter Aortic Valve implantation (TAVI) has become available as an alternative treatment for patients with severe aortic stenosis (AS). However, complications still occur. Also 40 to 50% of the patients report no subjective improvement in quality of life. Identification of predictors, in particular biomarkers, that are predictive of poor clinical response to TAVI, could improve patient selection and counselling in the future. Method 248 patients who received TAVI between 2009 and 2021 were retrospectively analysed. Data of seven pre-procedural potential biomarkers (Sodium, Potassium, Urea, Creatinine, Hemoglobin, NTproBNP & eGFR) were collected in a database. The primary outcome was defined as cardiovascular (CV) death, stroke or readmission with cardiac cause. The secondary outcome was defined as symptoms (NYHA class) at one year (response is defined as NYHA 1 or 2 at one year and non-response is defined as NYHA 3-4 at one year) Univariable and multivariable regression analyses were used to create prediction models. Results 248 patients were included. Mean age was 78 years and 52.4% were female. Within one year, 49 (20%) had reached primary outcome. The strongest prediction model included Atrial fibrillation (AF) /flutter (HR 2.571, p=0.003), right bundle branch block (RBBB) (HR 3.145, p=0.003), Sodium (HR 0.914, p=0.041) and Log-NTproBNP (HR 1.302, p=0.021). Adding biomarkers to the clinical characteristics improved the C-statistic from 0.677 to 0.714. With respect to the secondary outcome, 28 out of 223 (13%) patients had a poor symptomatic response. BMI, COPD and decompensated AS were found to be significant predictors of failure to improve symptomatically, resulting in a final prediction model including BMI (OR 1.099, p=0.010) and COPD (OR 3.307,p=0.005) (C-statistic =0.697). Routine biomarkers were not found to be significantly predictive for poor symptomatic outcome. Conclusion After multivariable analysis the significant predictors of CV death, stroke or cardiac hospitalization were AF/flutter, RBBB, Sodium and Log-NTproBNP, while significant predictors of poor symptomatic response were BMI and COPD. These predictors can support clinical decision making and improve patient counselling.
| Item Type: | Thesis (UNSPECIFIED) |
|---|---|
| Supervisor name: | Voors, Prof. dr. A.A. and Wykrzykowska, Dr. J.J. |
| Faculty: | Medical Sciences |
| Date Deposited: | 16 May 2023 13:00 |
| Last Modified: | 16 May 2023 13:00 |
| URI: | https://umcg.studenttheses.ub.rug.nl/id/eprint/3522 |
Actions (login required)
 |
View Item |