gms | German Medical Science

65th Annual Meeting of the German Association for Medical Informatics, Biometry and Epidemiology (GMDS), Meeting of the Central European Network (CEN: German Region, Austro-Swiss Region and Polish Region) of the International Biometric Society (IBS)

06.09. - 09.09.2020, Berlin (online conference)

Article

Send article

A multivariable model for improved prediction of kidney failure requiring kidney replacement therapy based on routine laboratory parameters

Meeting Abstract

Search Medline for

  • Helena Zacharias - University Medicine Greifswald, Greifswald, Germany
  • Michael Altenbuchinger - University of Hohenheim, Hohenheim, Germany
  • Ulla Schultheiß - University of Freiburg, Freiburg, Germany
  • Johannes Raffler - Helmholtz Zentrum München – German Research Center for Environmental Health, Neuherberg, Germany
  • Fruzsina Kotsis - University of Freiburg, Freiburg, Germany
  • Sahar Ghasemi - University Medicine Greifswald, Greifswald, Germany
  • Ibrahim Ali - Salford Royal Hospital and University of Manchester, Salford, United Kingdom
  • Marie Metzger - Centre for Research in Epidemiology and Population Health (CESP), INSERM UMRS 1018, Université Paris-Saclay, Université Versailles Saint Quentin, Villejuif, France
  • Inga Steinbrenner - University of Freiburg, Freiburg, Germany
  • Peggy Sekula - University of Freiburg, Freiburg, Germany
  • Ziad Massy - Centre for Research in Epidemiology and Population Health (CESP), INSERM UMRS 1018, Université Paris-Saclay, Université Versailles Saint Quentin, Villejuif, France; Ambroise Paré University Hospital, APHP, Boulogne-Billancourt/Paris, France
  • Christian Combe - Centre Hospitalier Universitaire de Bordeaux, Bordeaux, FranceUniv Bordeaux Segalen, Bordeaux, France
  • Philip Kalra - Salford Royal Hospital and University of Manchester, Salford, United Kingdom
  • Florian Kronenberg - Medical University of Innsbruck, Innsbruck, Austria
  • Bénédicte Stengel - 9Centre for Research in Epidemiology and Population Health (CESP), INSERM UMRS 1018, Université Paris-Saclay, Université Versailles Saint Quentin, Villejuif, France
  • Kai-Uwe Eckardt - Charité-Universitätsmedizin Berlin, Berlin, GermanyFriedrich-Alexander Universität Erlangen Nürnberg, Erlangen, Germany
  • Anna Köttgen - University of Freiburg, Freiburg, Germany
  • Matthias Schmid - University of Bonn, Bonn, Germany
  • Wolfram Gronwald - University of Regensburg, Regensburg, Germany
  • Peter Oefner - University of Regensburg, Regensburg, Germany

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie. 65th Annual Meeting of the German Association for Medical Informatics, Biometry and Epidemiology (GMDS), Meeting of the Central European Network (CEN: German Region, Austro-Swiss Region and Polish Region) of the International Biometric Society (IBS). Berlin, 06.-09.09.2020. Düsseldorf: German Medical Science GMS Publishing House; 2021. DocAbstr. 423

doi: 10.3205/20gmds327, urn:nbn:de:0183-20gmds3275

Published: February 26, 2021

© 2021 Zacharias et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License. See license information at http://creativecommons.org/licenses/by/4.0/.

Outline

Text

Identification of chronic kidney disease (CKD) patients, who are at risk of progressing to kidney failure requiring kidney replacement therapy (KRT), frequently also designated as end-stage kidney disease (ESKD), is important for clinical decision-making and clinical trial design and enrollment.

We report a new 6-variable risk model based on routine laboratory parameters that predicts progression to ESKD requiring KRT initiation in patients with CKD stages G3 together with A1-3 or G1-2 together with A3. To develop this model, we analyzed data from 4,915 patients from the longterm prospective observational German Chronic Kidney Disease (GCKD) cohort study. During an observation period of 3.71 ± 0.88 years, 200 of the 4,915 patients (4.07%) progressed to initiation of KRT, defined as initiation of long-term dialysis or kidney transplantation. A LASSO Cox proportional hazards model was trained and tested in a resampling approach and achieved a median concordance (C) index of 0.899 (95% CI, 0.871-0.920). Comprising the variables serum creatinine, albumin, cystatin C, and urea, as well as hemoglobin and the urine albumin-to-creatinine ratio, the new risk model outperformed the 5-year, 4-variable Tangri risk equation, currently the golden standard to predict ESKD events in CKD patients, which achieved a median C index of 0.866 (95% CI, 0.830-0.897) in the resampling approach. Likewise, the novel 6-variable risk equation always yielded positive net reclassification improvements (NRI) in comparison to the Tangri risk equation evaluated one, two, three, and four years after the baseline visit.

This 6-variable risk score further outperformed the Tangri risk score in three independent, international validation cohorts comprising, in total, 3,064 CKD patients, with improvements in C statistics ranging from 0.010 to 0.018.

In conclusion, the proposed risk model based on easily accessible routine laboratory parameters led to a marked improvement in the distinction of CKD patients likely to progress to ESKD requiring KRT.

The authors declare that they have no competing interests.

The authors declare that a positive ethics committee vote has been obtained.