Ernst, Angela Johanna ORCID: 0000-0003-2375-1889
(2025).
Development and validation of two-step algorithms to predict the outcome after kidney transplantation in the Eurotransplant region.
PhD thesis, Universität zu Köln.
This is the latest version of this item.
All available versions of this item
-
Development and validation of two-step algorithms to predict the outcome after kidney transplantation in the Eurotransplant region. (deposited UNSPECIFIED)
- Development and validation of two-step algorithms to predict the outcome after kidney transplantation in the Eurotransplant region. (deposited 20 May 2025 07:04) [Currently Displayed]
![]() |
PDF
Disserttion_Ernst_AJ_revision_e-signed.pdf Download (7MB) |
Abstract
The shortage of donor kidneys in Eurotransplant (ET) countries is undeniable, which is why the decision to discard an organ should not be made without solid reasoning. An evidence base, particularly with regard to the role of nephropathogical assessments, is still lacking. With additional pathological evaluation the cold ischaemia time is prolonged, and it is not available in all transplantation clinics and centers. The aim of this study was therefore to develop flexible clinico-pathological scores for the predict delayed graft function (DGF) and one-year death-censored graft loss (1y-tl) after kidney transplantation, to validate and compare them. As the modelling procedure depends on the set of pre-selected potential predictors, a second aim was to apply five statistical and machine-learning variable selection methods and compare results in terms of predictive discriminative power of the derived models. In terms of generating a good prediction model, machine learning methods were not superior to classical statistical models in selecting adequate predictors. Two two-step algorithms were developed which, with limitations, reflect the current practice in six ET countries (Austria, Belgium Austria, Belgium, Croatia, Germany, Hungary and the Netherlands). In order to address the prolonged cold ischaemia time in nephropathological evaluation and a lack of availability of these, scores with optional nephropathology were developed. With regard to the ability to discriminate cases of DGF the 2-Step scores was not inferior to the models of Irish, Balaz and Chapal. Neither was the 2-Step scores for 1y-tl inferior to that of Snoeijs, Port, De Vusser and Miller. The flexible 2-Step scores developed should provide good results for clinical practice in Eurotransplant and should not be inferior to more complex scores. However, they can and should be adapted for the increasing use of perfusion pumps and donation after cardiac death, as well as the risk group of donors with marginal quality kidneys.
Item Type: | Thesis (PhD thesis) | ||||||||||
Translated title: |
|
||||||||||
Translated abstract: |
|
||||||||||
Creators: |
|
||||||||||
URN: | urn:nbn:de:hbz:38-783018 | ||||||||||
Date: | 13 May 2025 | ||||||||||
Language: | English | ||||||||||
Faculty: | Faculty of Medicine | ||||||||||
Divisions: | Faculty of Medicine > Medizinische Statistik und Bioinformatik | ||||||||||
Subjects: | General statistics Medical sciences Medicine |
||||||||||
Uncontrolled Keywords: |
|
||||||||||
Date of oral exam: | 19 March 2025 | ||||||||||
Referee: |
|
||||||||||
Refereed: | Yes | ||||||||||
URI: | http://kups.ub.uni-koeln.de/id/eprint/78301 |
Downloads
Downloads per month over past year
Export
Actions (login required)
![]() |
View Item |