Jonathan Elmer1, Patrick J Coppler2, Bobby L Jones2, Daniel S Nagin2, Clifton W Callaway2. 1. From the Department of Emergency Medicine (J.E., P.J.C., C.C.); Department of Critical Care Medicine (J.E.); Department of Neurology (J.E.); Department of Psychiatry (B.L.J.), University of Pittsburgh; and the School of Public Policy & Management (D.S.N.), Heinz College, Carnegie Mellon University, Pittsburgh, PA. elmerjp@upmc.edu. 2. From the Department of Emergency Medicine (J.E., P.J.C., C.C.); Department of Critical Care Medicine (J.E.); Department of Neurology (J.E.); Department of Psychiatry (B.L.J.), University of Pittsburgh; and the School of Public Policy & Management (D.S.N.), Heinz College, Carnegie Mellon University, Pittsburgh, PA.
Abstract
BACKGROUND AND OBJECTIVES: Postarrest prognostication research does not typically account for the sequential nature of real-life data acquisition and interpretation and reports nonintuitive estimates of uncertainty. Bayesian approaches offer advantages well suited to prognostication. We used Bayesian regression to explore the usefulness of sequential prognostic indicators in the context of prior knowledge and compared this with a guideline-concordant algorithm. METHODS: We included patients hospitalized at a single center after cardiac arrest. We extracted prospective data and assumed these data accrued over time as in routine practice. We considered predictors demographic and arrest characteristics, initial and daily neurologic examination, laboratory results, therapeutic interventions, brain imaging, and EEG. We fit Bayesian hierarchical generalized linear multivariate models predicting discharge Cerebral Performance Category (CPC) 4 or 5 (poor outcomes) vs 1-3 including sequential clinical and prognostic data. We explored outcome posterior probability distributions (PPDs) for individual patients and overall. As a comparator, we applied the 2021 European Resuscitation Council and European Society of Intensive Care Medicine (ERC/ESICM) guidelines. RESULTS: We included 2,692 patients of whom 864 (35%) were discharged with a CPC 1-3. Patients' outcome PPDs became narrow and shifted toward 0 or 1 as sequentially acquired information was added to models. These changes were largest after arrest characteristics and initial neurologic examination were included. Using information typically available at or before intensive care unit admission, sensitivity predicting poor outcome was 51% with a 0.6% false-positive rate. In our most comprehensive model, sensitivity for poor outcome prediction was 76% with 0.6% false-positive rate (FPR). The ERC/ESICM algorithm applied to 547 of 2,692 patients and yielded 36% sensitivity with 0% FPR. DISCUSSION: Bayesian models offer advantages well suited to prognostication research. On balance, our findings support the view that in expert hands, accurate neurologic prognostication is possible in many cases before 72 hours postarrest. Although we caution against early withdrawal of life-sustaining therapies, rapid outcome prediction can inform clinical decision making and future clinical trials.
BACKGROUND AND OBJECTIVES: Postarrest prognostication research does not typically account for the sequential nature of real-life data acquisition and interpretation and reports nonintuitive estimates of uncertainty. Bayesian approaches offer advantages well suited to prognostication. We used Bayesian regression to explore the usefulness of sequential prognostic indicators in the context of prior knowledge and compared this with a guideline-concordant algorithm. METHODS: We included patients hospitalized at a single center after cardiac arrest. We extracted prospective data and assumed these data accrued over time as in routine practice. We considered predictors demographic and arrest characteristics, initial and daily neurologic examination, laboratory results, therapeutic interventions, brain imaging, and EEG. We fit Bayesian hierarchical generalized linear multivariate models predicting discharge Cerebral Performance Category (CPC) 4 or 5 (poor outcomes) vs 1-3 including sequential clinical and prognostic data. We explored outcome posterior probability distributions (PPDs) for individual patients and overall. As a comparator, we applied the 2021 European Resuscitation Council and European Society of Intensive Care Medicine (ERC/ESICM) guidelines. RESULTS: We included 2,692 patients of whom 864 (35%) were discharged with a CPC 1-3. Patients' outcome PPDs became narrow and shifted toward 0 or 1 as sequentially acquired information was added to models. These changes were largest after arrest characteristics and initial neurologic examination were included. Using information typically available at or before intensive care unit admission, sensitivity predicting poor outcome was 51% with a 0.6% false-positive rate. In our most comprehensive model, sensitivity for poor outcome prediction was 76% with 0.6% false-positive rate (FPR). The ERC/ESICM algorithm applied to 547 of 2,692 patients and yielded 36% sensitivity with 0% FPR. DISCUSSION: Bayesian models offer advantages well suited to prognostication research. On balance, our findings support the view that in expert hands, accurate neurologic prognostication is possible in many cases before 72 hours postarrest. Although we caution against early withdrawal of life-sustaining therapies, rapid outcome prediction can inform clinical decision making and future clinical trials.
Authors: S Backman; T Cronberg; H Friberg; S Ullén; J Horn; J Kjaergaard; C Hassager; M Wanscher; N Nielsen; E Westhall Journal: Resuscitation Date: 2018-07-24 Impact factor: 5.262
Authors: Kirstie Haywood; Laura Whitehead; Vinay M Nadkarni; Felix Achana; Stefanie Beesems; Bernd W Böttiger; Anne Brooks; Maaret Castrén; Marcus Eh Ong; Mary Fran Hazinski; Rudolph W Koster; Gisela Lilja; John Long; Koenraad G Monsieurs; Peter T Morley; Laurie Morrison; Graham Nichol; Valentino Oriolo; Gustavo Saposnik; Michael Smyth; Ken Spearpoint; Barry Williams; Gavin D Perkins Journal: Circulation Date: 2018-04-26 Impact factor: 29.690
Authors: Steffen Desch; Anne Freund; Ibrahim Akin; Michael Behnes; Michael R Preusch; Thomas A Zelniker; Carsten Skurk; Ulf Landmesser; Tobias Graf; Ingo Eitel; Georg Fuernau; Hendrik Haake; Peter Nordbeck; Fabian Hammer; Stephan B Felix; Christian Hassager; Thomas Engstrøm; Stephan Fichtlscherer; Jakob Ledwoch; Karsten Lenk; Michael Joner; Stephan Steiner; Christoph Liebetrau; Ingo Voigt; Uwe Zeymer; Michael Brand; Roland Schmitz; Jan Horstkotte; Claudius Jacobshagen; Janine Pöss; Mohamed Abdel-Wahab; Philipp Lurz; Alexander Jobs; Suzanne de Waha-Thiele; Denise Olbrich; Frank Sandig; Inke R König; Sabine Brett; Maren Vens; Kathrin Klinge; Holger Thiele Journal: N Engl J Med Date: 2021-08-29 Impact factor: 91.245
Authors: Claudio Sandroni; Sonia D'Arrigo; Sofia Cacciola; Cornelia W E Hoedemaekers; Marlijn J A Kamps; Mauro Oddo; Fabio S Taccone; Arianna Di Rocco; Frederick J A Meijer; Erik Westhall; Massimo Antonelli; Jasmeet Soar; Jerry P Nolan; Tobias Cronberg Journal: Intensive Care Med Date: 2020-09-11 Impact factor: 17.440
Authors: Jonathan Elmer; John J Gianakas; Jon C Rittenberger; Maria E Baldwin; John Faro; Cheryl Plummer; Lori A Shutter; Christina L Wassel; Clifton W Callaway; Anthony Fabio Journal: Neurocrit Care Date: 2016-12 Impact factor: 3.210
Authors: Jerry P Nolan; Claudio Sandroni; Bernd W Böttiger; Alain Cariou; Tobias Cronberg; Hans Friberg; Cornelia Genbrugge; Kirstie Haywood; Gisela Lilja; Véronique R M Moulaert; Nikolaos Nikolaou; Theresa Mariero Olasveengen; Markus B Skrifvars; Fabio Taccone; Jasmeet Soar Journal: Intensive Care Med Date: 2021-03-25 Impact factor: 17.440
Authors: Florian Ebner; Marion Moseby-Knappe; Niklas Mattsson-Carlgren; Gisela Lilja; Irina Dragancea; Johan Undén; Hans Friberg; David Erlinge; Jesper Kjaergaard; Christian Hassager; Matt P Wise; Michael Kuiper; Pascal Stammet; Michael Wanscher; Janneke Horn; Susann Ullén; Tobias Cronberg; Niklas Nielsen Journal: Resuscitation Date: 2020-05-21 Impact factor: 5.262