Vasoplegia in Cardiac Surgery: A Systematic Review and Meta-analysis of Current Definitions and Their Influence on Clinical Outcomes
Olga Papazisi, Marnix M. van der Schoot, Remco R. Berendsen, M. Sesmu Arbous, Saskia le Cessie, Olaf M. Dekkers, Robert J.M. Klautz, Nándor Marczin, Meindert Palmen, Eric E.C. de Waal
Abstract
<h3>Objectives</h3> To identify differences in the reported vasoplegia incidence, intensive care unit (ICU) length of stay (LOS), and 30-day mortality rates as influenced by different vasoplegia definitions used in cardiac surgery studies. <h3>Design</h3> A systematic review was performed covering the period 1977 to 2023 using PubMed/MEDLINE, Embase, Web of Science, Cochrane Library, and Emcare and a meta-analysis (PROSPERO: CRD42021258328) was performed. <h3>Setting and Participants</h3> One hundred studies defining vasoplegia in cardiac surgery patients were systematically reviewed. Sixty studies with 20 or more patients, irrespective of design, reporting vasoplegia incidence, ICU LOS, or 30-day mortality were included for meta-analysis. <h3>Interventions</h3> Cardiac surgery on cardiopulmonary bypass. <h3>Measurements and Main Results</h3> Studies were categorized depending on the used mean arterial pressure (MAP) thresholds. Random intercept logistic regression models were used for meta-analysis of incidence and mortality. Random effect meta-analysis was used for ICU LOS. One hundred studies were reviewed systematically. MAP and cardiac index thresholds varied considerably (<50-80 mmHg and 2.0-3.5 L·min<sup>−1</sup>m<sup>−2</sup>, respectively). Vasopressor dosages also differed between definitions. The reported incidence (60 studies; mean incidence, 19.9%; 95% confidence interval [CI], 16.1-24.4) varied largely between studies (2.5%-66.3%; I<sup>2</sup> = 97%; p < 0.0001). Meta-regression models, including the MAP-threshold, did not explain this heterogeneity. Similarly, the effect of vasoplegia on ICU LOS, and 30-day mortality was very heterogeneous among studies (I<sup>2</sup> = 99% and I<sup>2</sup> = 73%, respectively). <h3>Conclusions</h3> The large variability in vasoplegia definitions is associated with significant heterogeneity regarding incidence and clinical outcomes, which cannot be explained by factors included in our models. Such variations in definitions leads to inconsistent patient diagnosis and renders published vasoplegia research incomparable.