Definitions for Vasoplegia Associated With On-Pump Cardiac Surgery: A Systematic Review With Meta-Analysis

Abstract

OBJECTIVES:

Vasoplegia is a common complication of cardiac surgery that uses cardiopulmonary bypass and contributes to morbidity and mortality, yet a consensus definition does not exist. The objective of this study was to evaluate the diagnostic criteria and definitions used to characterize vasoplegia and how different criteria influence incidence estimates.

DATA SOURCES:

Ovid Embase, Ovid MEDLINE, Scopus, Web of Science Core Collection, ClinicalTrials.gov, Ovid Cochrane Central Register of Controlled Trials, and the World Health Organization’s International Clinical Trials Registry Platform clinical trials registry.

STUDY SELECTION:

Randomized clinical trials and observational studies reporting on vasoplegia in adults undergoing any type of cardiac surgery that used cardiopulmonary bypass.

DATA EXTRACTION:

Proportional meta-analysis using a random-effects model and the inverse variance method was used to calculate the pooled incidence of vasoplegia and its clinical outcomes.

DATA SYNTHESIS:

A total of 68 studies encompassing 56,580 patients were identified, from which 63 unique vasoplegia definitions were used. Blood pressure (n = 57 studies, 84%) and cardiac output (n = 50 studies, 74%) were among the most common criteria used in vasoplegia definitions; however, there was a vast variety of threshold values applied within these criteria and all other criteria comprising the definitions. The pooled incidence of vasoplegia was 21% (95% CI, 17–25%), acute kidney injury was 32% (95% CI, 21–45%), and mortality was 12% (95% CI, 9–16%). Subgroup analysis revealed that transplantation and left ventricular assist device implantation surgeries, and those with baseline left ventricular ejection fraction less than 40% had a significantly greater incidence of vasoplegia.

CONCLUSIONS:

The published literature varies greatly in the criteria used to define vasoplegia associated with on-pump cardiac surgery. Generation and adoption of a unified definition for vasoplegia must be an international priority.

KEY POINTS

Question: What are the criteria and definitions used to characterize vasoplegia associated with cardiac surgery and how do different criteria impact incidence estimates?

Findings: This systematic review identified 63 unique vasoplegia definitions used in the literature, with the most common criteria encompassing blood pressure, cardiac output, and vasopressor use. There was a high amount of heterogeneity in the calculated incidence and clinical outcomes.

Meaning: Definitions for vasoplegia vary across published literature, obscuring incidence and outcome estimates, requiring careful interpretation of research findings. A standardized definition for vasoplegia is needed.

First described in 1994 (1), vasoplegia remains one of the most common complications associated with cardiac surgery that uses cardiopulmonary bypass (2). Hypotension from systemic vasodilation and lack of adequate vascular tone in this form of circulatory failure impairs the ability of tissues to effectively receive oxygenated blood and remove metabolic waste byproducts. Exposure to hypotension in critical illness, even for brief periods of time, is associated with organ dysfunction (3). As such, patients who develop vasoplegia after cardiac surgery experience more acute kidney injury, spend longer durations in the ICU, and have greater mortality than those who do not develop vasoplegia (4, 5).

Diagnosing and managing vasoplegia can be challenging for clinicians as many features overlap with other shock syndromes, principally, the “low cardiac output syndrome” that is often encountered after cardiac surgery (6). Generally, vasoplegia is considered a syndrome of arterial hypotension despite adequate cardiac function and circulatory volume (7). However, specific hemodynamic threshold values for variables such as cardiac output and mean arterial pressure are heterogeneous. Furthermore, no clinical practice guidelines or consensus recommendations exist surrounding vasoplegia, resulting in a lack of a standardized definition.

Variability in the criteria used to characterize vasoplegia poses challenges for clinicians and complicates the performance and interpretation of research surrounding vasoplegia. Epidemiologic estimates including incidence, risk factor assessment, and outcome quantification are dependent on clear criteria for case ascertainment and are hence compromised when one does not exist (8). Additionally, applicability of clinical trial findings is impacted by criteria used for screening and eligibility determination (9). Finally, pooled estimates of epidemiological variables, effects of interventions, and outcomes evaluation in meta-analysis are often obscured when studies using different definitions are aggregated. This last issue positions patient care in a particularly vulnerable situation as pooled estimates from meta-analysis are used to grade the certainty of available evidence and guide the application of research findings to the bedside (10). Therefore, we aimed to provide a comprehensive overview of the diagnostic criteria and definitions used to characterize vasoplegia associated with cardiac surgery. We also sought to assess how different criteria used in definitions influence incidence estimates generated from the study of vasoplegia.

MATERIALS AND METHODS

This was a systematic review and meta-analysis utilizing prognosis systematic review methods (11) designed to evaluate, in adult human persons undergoing cardiac surgery with the use of cardiopulmonary bypass, how variable criteria used to characterize vasoplegia influence incidence estimates. The study protocol was registered a priori in the PROSPERO database (no. CRD42025642557) and the study was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (12) and the Meta-analyses of Observational Studies in Epidemiology (13) guidance statements.

Search Strategy

The biomedical literature was comprehensively searched for the concepts of vasoplegia and cardiac surgery, utilizing a combination of keywords and standardized index terms, in collaboration with a skilled medical librarian study investigator (D.J.G.). The searches were performed on October 14, 2024, from database inception in ClinicalTrials.gov (2000+), Ovid Cochrane Central Register of Controlled Trials (1991+), Ovid Embase (1974+), Ovid MEDLINE (1946+ including Epub ahead of print, in-process and other nonindexed citations), Scopus (1788+), Web of Science Core Collection (Science Citation Index Expanded 1975+ and Emerging Sources Citation Index 2015+), and the World Health Organization’s International Clinical Trials Registry Platform clinical trials registry (2005+). The search was restricted to human subjects but did not have any restriction to language. The full search strategies can be found in eTable 1 (https://links.lww.com/CCX/B592).

Eligibility Criteria and Study Selection

Studies reporting on vasoplegia in adults undergoing any type of cardiac surgery that used cardiopulmonary bypass were eligible for inclusion. Prospective and retrospective observational studies (longitudinal and nonlongitudinal) were eligible. We also included randomized trials only if one arm was subject to a nonactive intervention (e.g., placebo or standard of care) such that outcomes of vasoplegia could be adjudicated, or if eligibility for the trial was conducted in a way that systematically assessed the incidence of vasoplegia in enrolled individuals. Single-arm cohort studies, population-based studies, and case series reporting on at least ten patients that systematically assessed the incidence of vasoplegia were also eligible. Studies reporting on off-pump cardiac surgery were eligible for inclusion only if the comparator group consisted of on-pump cardiac surgery and incidence and/or outcomes were compared. Studies were excluded if they reported on solely off-pump cardiac surgery, did not present detail regarding the vasoplegia definition used, or reported on cardiogenic shock, low cardiac output syndrome, or other nonvasoplegia syndromes. Case reports were excluded due to the inability to systematically assess the incidence or outcomes of vasoplegia, as were conference abstracts due to insufficient detail to adjudicate the selection procedures and definitions of vasoplegia.

The results from the systematic search were loaded into Covidence Software (Melbourne, VIC, Australia) and the automatic de-duplication procedure was applied. The titles and abstracts for each search result were independently assessed for possible inclusion by two reviewers (P.M.W., E.D. Wieruszewski), with discrepancies resolved by discussion with a third investigator (E.D. Wittwer) serving as an arbitrator. The full-text files of eligible studies were retrieved, and the inclusion and exclusion criteria were independently applied by two investigators (P.M.W., E.D. Wieruszewski), with discrepancies resolved by a third investigator (E.D. Wittwer). The reasons for exclusion were documented during the full-text review stage. Backward and forward citation searching of reference lists of included studies was performed, as well as review of reference lists of published relevant systematic and narrative reviews for any potential missing studies.

Data Extraction

A modified version of the CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies (14) for prognostic factor studies (15) was used to accommodate this review and was used to guide the data extraction procedures. Data regarding the publication details, years of study, study design and methodology, subject characteristics including their demographics, types of surgical procedures performed, and outcomes characteristics, the intervention and comparator groups, if applicable, and the definitions of vasoplegia were independently extracted manually from the full-text files of each eligible study by two investigators (P.M.W., E.D. Wieruszewski) and collated. The primary outcome was the incidence of vasoplegia. Secondary outcomes included acute kidney injury, arrhythmias, and mortality. Discrepancies were resolved through discussion or consultation with another study investigator if resolution could not be achieved. Authors of original articles were not contacted in the event of missing data.

Bias Assessment

The Quality in Prognostic Factor Studies (QUIPS) (16) tool was adapted for assessment of incidence and outcomes by removing the prognostic factor measurement and adjustment for other prognostic factor domains. This modified QUIPS tool was applied independently by two reviewers (P.M.W., E.D. Wieruszewski) to assess the risk of bias for all included studies in the following domains: study participation, study attrition, outcome measurement, and statistical analysis and reporting. The risk of bias was judged as low, moderate, or high in each domain, and an overall risk of bias judgment was made for each included study. Discrepancies were resolved by a third investigator (E.D. Wittwer).

Data Synthesis and Statistical Analysis

The details regarding each included study were summarized qualitatively in tabular and narrative form. If a study reported on multiple variations of vasoplegia (e.g., “mild,” “moderate,” etc.), each individual definition was reported and used in qualitative summaries. A sunburst diagram was constructed to visualize the various criteria used in definitions among included studies, where the concentric rings represented a hierarchy from: 1) use of vasopressors, 2) types of hemodynamic criteria, and 3) eligibility window and duration of exposure.

Meta-analysis was performed to calculate the pooled incidences using a random-effects model with the inverse variance method. For studies reporting on multiple variations of vasoplegia (e.g., “mild,” “moderate,” etc.), the overall incidence, if not provided, was computed, and the definition used in heterogeneity assessment was the most inclusive one provided. If a study reported on a definition that used mutually exclusive criteria and reported incidence for different definitions of vasoplegia independently, a separate entry was made in the meta-analysis. The same methods were applied to calculate the pooled estimates for all the binary outcome measures including acute kidney injury, arrhythmias, and mortality. The between-study statistical heterogeneity was calculated using the I² statistic, and the cutoff ranges recommended by the Cochrane Collaboration were applied to assess the degree of statistical heterogeneity (17). Evidence for heterogeneity of subgroup effect beyond chance was determined from the p value of a chi-square test (p_interaction), with a p_interaction threshold below 0.1 used to determine statistical significance (17, 18). Because only proportional meta-analysis was performed and no comparisons were made, funnel plots were not constructed to assess publication bias. All analyses were performed using R, Version 4.5.0 (R Foundation for Statistical Computing, Vienna, Austria, 2025) and meta version 8.1-0.

Planned Subgroup and Sensitivity Analyses

We performed a priori defined subgroup and sensitivity analyses in an attempt to elucidate potential explanations for heterogeneity. Subgroup analyses included transplant surgery (vs. nontransplant surgery), left ventricular assist device (LVAD) surgery (vs. non-LVAD surgery), and preoperative left ventricular ejection fraction (< 40% vs. > 40%). The study mean or median preoperative left ventricular ejection fraction, or the value that was present in at least 50% of the subjects was used for group assignment. We hypothesized that patients with reduced ejection fraction heart failure and those undergoing transplant or LVAD surgery would have greater incidence of vasoplegia and experience worse outcomes. Sensitivity analyses included the various hemodynamic threshold criteria in each individual vasoplegia definition, only observational studies, and only studies judged to be of low risk of bias.

RESULTS

A total of 2608 entries were identified by the systematic search. Following automatic de-duplication and exclusion of irrelevant entries in the title and abstract screening, 105 full-text files were retrieved. Of these, 37 were excluded, most commonly due to studying the wrong population (n = 15) and not presenting a definition of vasoplegia (n = 13; eFig. 1, https://links.lww.com/CCX/B592). Of the 68 included studies (4, 5, 9, 19–83), 54 were observational studies (4, 5, 19–26, 28–38, 40, 41, 44, 46, 47, 49, 51–56, 59, 61–66, 68–75, 77–82) and 14 were randomized trials (9, 27, 39, 42, 43, 45, 48, 50, 57, 58, 60, 67, 76, 83) (eTable 2, https://links.lww.com/CCX/B592). Studies were performed in 24 different countries across Europe (n = 28), North America (n = 19), South America (n = 9), Asia (n = 7), West Asia (n = 3), and Oceania (n = 2). The included patients were treated from 1992 to 2023 (study years not reported in five studies) (30, 34, 46, 70, 80), and spanned publication years from 1996 to 2024. A total of 56,580 patients were encompassed across the 68 included studies. Of the included studies, 15 were of patients undergoing cardiothoracic transplantation (22, 23, 26, 29, 31, 32, 36, 38, 40, 44, 57, 61, 63, 66, 79), eight undergoing LVAD implantation (35, 47, 50, 68, 74, 75, 77, 82), and the remainder a mix of various cardiac surgical procedures. The full details of each included study are presented in eTable 3 (https://links.lww.com/CCX/B592).

Bias Assessment

The overall risk of bias assessment identified most studies (n = 50, 74%) to be of low risk of bias, while 15 studies (22%) were of moderate risk of bias, and 3 (4%) were of high risk of bias (eFig. 2, https://links.lww.com/CCX/B592). The most common reasons influencing a judgment of moderate risk of bias were related to the selection of participants and included overly restrictive exclusion criteria or very high rates of exclusion. Similarly, the studies judged to be high risk of bias did not describe the selection procedures at all and excluded participants without detailing the exclusionary criteria.

Diagnostic Criteria for Vasoplegia

Across the 68 included studies, there were 63 unique vasoplegia definitions reported (Fig. 1; eTable 3, https://links.lww.com/CCX/B592). Seven studies reported multiple definitions (e.g., mild, moderate, etc.) (4, 21, 30, 32, 37, 66, 75). The definition in most studies required the administration of at least one vasopressor paired with hemodynamic criteria (n = 53) (4, 5, 19–28, 30–32, 35–42, 45, 47, 48, 50–55, 57, 59, 61–64, 66, 67, 70–79, 81–83), while 12 studies did not require vasopressor administration (9, 29, 33, 34, 43, 44, 46, 58, 60, 65, 68, 80), and four studies did not have any hemodynamic criteria in their definitions (50, 56, 68, 69) (Table 1). There was a vast variety of hemodynamic and physiologic measurements (eTable 4, https://links.lww.com/CCX/B592), vasopressors or other interventions (eTable 5, https://links.lww.com/CCX/B592), and duration of vasopressor exposure/hemodynamic derangement and window of eligibility for the diagnosis (eTable 6, https://links.lww.com/CCX/B592) across the definitions used.

Figure 1.

Sunburst diagram of the criteria composing vasoplegia definitions in published literature. The concentric rings present a hierarchy from the interior of: 1) vasopressor use, 2) specific hemodynamic criteria used, and 3) eligibility window for the diagnosis and duration of vasopressor exposure/hemodynamic derangement required. BP = blood pressure, CO = cardiac output, SVR = systemic vascular resistance.

TABLE 1.

High-Level Summary of Criteria Used in Vasoplegia Definitions

Variable	No. of Different Values/Thresholds	Total No. of Studies (n = 68), n (%)
Hemodynamic and physiologic measurements
Blood pressure abnormality	12	57 (84)
Cardiac output	15	50 (74)
Systemic vascular resistance	14	32 (47)
Central cardiac pressures	14	14 (21)
Left ventricular ejection fraction	3	7 (10)
Intervention requirements
Vasopressor requirement	39	57 (84)
One or more vasopressors	31	45 (66)
Two or more vasopressors	7	11 (16)
Three or more vasopressors	1	1 (1.5)
Fluid administration	5	13 (19)
Duration and window for eligibility
Duration of vasopressor exposure	9	24 (35)
Duration of hemodynamic derangement	5	5 (7)
Window for eligibility	13	36 (53)
Exclusions to make the diagnosis	9	14 (21)

Full list of specific criteria and threshold values can be found in the Supplemental Digital Content (eTables 4–6, https://links.lww.com/CCX/B592).

Blood pressure (n = 57 studies, 84%) and cardiac output (n = 50 studies, 74%) were the most commonly present hemodynamic criteria across all definitions (Table 1). However, within these domains and all other criteria domains, there was a considerable assortment of threshold values used. Also, some studies were nondescriptive and simply included “hypotension” (n = 4), “low systematic vascular resistance” (n = 1), “low cardiac output/index” (n = 4), “preserved left ventricular function” (n = 1), and “normal cardiac function/contractility” (n = 5), without providing specific threshold values (eTable 4, https://links.lww.com/CCX/B592).

While a vasopressor requirement was one of the most common criteria used in definitions (n = 57 studies, 84%), the number and types of agents, combinations, and dose thresholds were widely variable. Some definitions required “a vasopressor” without further specifying requirements (eTable 5, https://links.lww.com/CCX/B592).

Only 36 studies (53%) specified a window of eligibility for vasoplegia diagnosis, with most restricting to the first 24–48 hours postoperatively (eTable 6, https://links.lww.com/CCX/B592). However, some studies restricted this window to the operating room only or just the first 6 hours postoperatively, while others were up to 3–4 days after the operation. Finally, exclusions, such as ruling out sepsis, cardiogenic shock, or other syndromes that may explain hypotension were specified only in a minority of studies.

Vasoplegia Incidence

Of the included studies, 64 (94%) reported the incidence of vasoplegia and were eligible for the meta-analysis (4, 5, 19–38, 40–66, 68–82). Two studies used two mutually exclusive definitions and evaluated the incidence independently and thus each definition was entered into the meta-analysis separately (21, 37). The overall pooled incidence of vasoplegia across the 66 definitions encompassing 82,197 patients was 21% (95% CI, 17–25%; eFig. 3, https://links.lww.com/CCX/B592), noting that there was substantial statistical heterogeneity (I² = 99%). The overall pooled incidence was similar in sensitivity analyses of only low risk of bias studies, observational studies, and different vasopressor and hemodynamic components of vasoplegia definitions (Fig. 2). In subgroup analysis, studies reporting on transplantations or LVAD implantation had higher vasoplegia incidence (34%; 95% CI, 27–41%) than studies reporting on other surgeries (16%; 95% CI, 13–20%; p_interaction < 0.001; eFig. 8, https://links.lww.com/CCX/B592). Similarly, studies of patients with baseline left ventricular ejection fraction less than 40% had higher vasoplegia incidence (26%; 95% CI, 20–33%) than those with baseline left ventricular ejection fraction greater than 40% (16%; 95% CI, 12–21%; p_interaction = 0.01; eFig. 9, https://links.lww.com/CCX/B592). All meta-analyses, including subgroup and sensitivity analyses for vasoplegia incidence are presented in eFigs. 3–31 (https://links.lww.com/CCX/B592) and summarized in Figure 2 and eTables 7–9 (https://links.lww.com/CCX/B592).

Figure 2.

Summary of pooled estimates for vasoplegia incidence from meta-analyses. CO = cardiac output, LVAD = left ventricular assist device, LVEF = left ventricular ejection fraction, MAP = mean arterial pressure, SVR = systemic vascular resistance.

Vasoplegia Outcomes

Of the included studies, 32 (47%) reported on at least one of the vasoplegia outcomes of interest (4, 9, 22–24, 29–31, 35, 36, 39–41, 44, 46–48, 52, 53, 55, 56, 59, 63, 64, 66, 67, 73, 75, 78, 79, 82, 83). The pooled incidence of acute kidney injury (n = 14 studies) was 32% (95% CI, 21–45%; eFig. 32, https://links.lww.com/CCX/B592). Acute kidney injury was higher in studies reporting on transplantations or LVAD implantations than other surgeries (eFig. 35, https://links.lww.com/CCX/B592), as well as those studying patients with baseline left ventricular ejection fraction less than 40% (eFig. 36, https://links.lww.com/CCX/B592); however, the p_interaction for subgroup differences was only significant for the latter. The pooled incidence of arrhythmias (n = 7 studies) was 34% (95% CI, 20–53%; eFig. 37, https://links.lww.com/CCX/B592). The pooled incidence of mortality (n = 31 studies) was 12% (95% CI, 9–16%; eFig. 40, https://links.lww.com/CCX/B592). All meta-analyses, including subgroup and sensitivity analyses for vasoplegia outcomes are presented in eFigs. 32–44 (https://links.lww.com/CCX/B592) and summarized in Figure 3.

Figure 3.

Summary of pooled estimates for outcomes of vasoplegia from meta-analyses. LVAD = left ventricular assist device, LVEF = left ventricular ejection fraction.

DISCUSSION

Summary of Key Findings

In this systematic review, we found a wide variety of criteria used in definitions of vasoplegia associated with on-pump cardiac surgery. Indeed, among the 68 included studies, 63 unique vasoplegia definitions were used. The pooled incidence of vasoplegia was calculated to be 21%; however, there was substantial statistical heterogeneity. Vasoplegia incidence was found to be higher in studies reporting on LVAD implantation or transplantation surgery, and those with baseline left ventricular ejection fraction less than 40%. Due to the large variation in reported threshold criteria used among definitions, it was not possible to conclude any relationship on incidence estimates.

Relationship to Previous Studies

A prior review reported on definitions of vasoplegia used in cardiac surgery studies (84); however, some notable differences in methodology and findings with our study warrant discussion. While their meta-regression on the proportion of patients with heart failure receiving surgery did not explain the variation in vasoplegia incidence, we chose to assess preoperative ejection fraction in our subgroup analyses. Therefore, we found that those with ejection fraction less than 40% undergoing on-pump surgery had significantly higher incidence of vasoplegia, which is consistent with prior studies demonstrating that heart failure, specifically accompanied by reduced ejection fraction, is associated with increased risk for postoperative vasoplegia (21, 30).

Our study used a comprehensive systematic search strategy and included databases and registers beyond only MEDLINE, which identified 16 additional articles for inclusion that the Papazisi et al (84) study did not. Notably, while our search identified all the entries included in the Papazisi et al (84) review, we excluded 47 of the studies they included. These studies were excluded from our study because they were either case reports, small case series, or studies that did not systematically assess the incidence of vasoplegia, sampled on biomarkers (85) or drug exposure (86) while not being designed to assess vasoplegia incidence, combined on- and off-pump surgeries without separate reporting (87), or reported on nonvasoplegia syndromes such as septic shock (88), nonseptic systemic inflammatory response syndrome (89), the “hypotension, acidosis, vasodilation syndrome” (90), or patients with low cardiac output (91). While there is considerable variability in the specific criteria used for vasoplegia definitions, we paid special attention to the granular details of study design and selection procedures to ensure that our study enriched the population to those with vasoplegia, an important exercise to minimize selection bias at the level of entering studies into the meta-analysis (92).

Another study assessed vasoplegia definitions in patients undergoing heart transplantation, and found hemodynamic criteria to be variable among a limited number of studies (93). While no meta-analysis was performed, the incidence of vasoplegia and mortality across studies was highly variable in the four included studies. Another review of vasoplegia in heart transplantation calculated a raw percentage of vasoplegia to be 29% using the total number of cases divided by the number of surgeries reported in the included studies (94). While that study did not use formal meta-analytic methods, the estimate was similar to our pooled estimate, however, that work did not capture the magnitude of statistical heterogeneity present. Patients undergoing heart transplant often present with unique risk factors increasing propensity for development of vasoplegia and worse outcomes including use of preoperative inotropes and mechanical circulatory support, repeat cardiac surgery, longer cardiopulmonary bypass times, and greater amount of blood product transfusion than typical cardiac surgery (94). Additionally, patients who had an LVAD as a bridge to transplantation appear to be at particularly high risk of vasoplegia when they undergo heart transplant surgery (22, 23, 66).

Implications of the Findings

Our study highlights the inconsistency and variability in the application of criteria to define vasoplegia associated with on-pump cardiac surgery. Additionally, because outcomes were sparsely reported, we were unable to assess for the relationship of varying hemodynamic threshold criteria on subsequent outcomes. This highlights important remaining uncertainties related to vasoplegia, and which criteria should be included in a definition. Furthermore, the optimal cardiac output and mean arterial pressure targets during the treatment of vasoplegia remain unknown. Studies often refer to vasoplegia incidence as a range, although our study suggests that, in addition to other variables, certain cardiac surgical populations including patients undergoing LVAD implantations and transplantation surgeries, as well as patients with baseline left ventricular ejection fraction less than 40% experience vasoplegia more often than other groups. These estimates are critical for targeted identification of at-risk patients given the contribution of vasoplegia on postoperative outcomes. Although our study was not designed to compare outcomes of vasoplegia to those without vasoplegia, our calculated incidence of acute kidney injury of 32% is considerably higher than national registry estimates of 1–12% among all cardiac surgeries (95). Similarly, our pooled estimate of 12% mortality is higher than most cardiac surgical procedures when evaluated on a population level (95). While assessment of vasoplegia management strategies was beyond the scope of this systematic review, our findings suggest that attempts to quantify treatment effects of specific interventions using meta-analytic methods may be obscured by variable criteria and definitions used across studies.

Strengths and Limitations

Our study findings are the result of a comprehensive a priori specified protocol and analysis. We followed rigorous methods for search design, article selection, bias adjudication, and quantitative analysis as recommended by the Cochrane Collaboration. Our use of a sunburst diagram allows for global visualization of the hierarchy of the components used to diagnose vasoplegia and the frequency at which each specific definition is used.

Several limitations warrant discussion. We chose to include randomized trials, which generally limit the ability to calculate incidence. However, including such studies allowed us to use subjects who were randomized to nonactive interventions (e.g., placebo) in the pooled outcomes meta-analysis. Additionally, this allowed us to include several randomized trials that performed strict selection procedures to appropriately assess vasoplegia incidence. Accordingly, to confirm the robustness of our primary findings, sensitivity analyses removing randomized trials resulted in similar pooled estimates for incidence and outcomes. The high degree of statistical heterogeneity in our study poses challenges for meta-analysis and may be argued as a reason for omitting the calculation of pooled estimates altogether. However, this finding was largely expected given the review question, and one of our objectives was to assess the influence of specific criteria on the pooled incidence estimates. Last, because of sparse outcome reporting, we did not believe it was appropriate to attempt to assess the relationship between varying hemodynamic threshold criteria and clinical outcomes.

CONCLUSIONS

Criteria used to define vasoplegia associated with on-pump cardiac surgery vary greatly among the published literature. The estimation of vasoplegia incidence and its associated outcomes is obscured by arbitrary criteria and threshold values used in definitions. Generation and adoption of the optimal hemodynamic targets and a unified definition for vasoplegia should be a priority for cardiac surgical, anesthesia, and critical care societies.

Supplementary Material

cc9-7-e1361-s001.pdf ^{(2MB, pdf)}