Background

Timely initiation of vasoactive and inotropic medications is crucial for hemodynamic management in both surgical settings and intensive care unit (ICU) [1,2,3]. While effective in improving hemodynamic parameters, these medications also have significant side effects, including increased myocardial oxygen consumption, myocardial ischemia, arrhythmia, ischemic perfusion injury, and multi-organ failure [4, 5]. Emerging evidence suggests that excessive doses of these drugs could be detrimental during surgical procedures [6, 7]. The severity of illness quantification is vital for critical care physicians, as it can guide patient prognostication and family counseling.

The vasoactive-inotropic score (VIS), first proposed in 2010 [1], has since been extensively used to quantify cardiovascular support after cardiac surgery in pediatric patients. Davidson et al. [8], in a study on 70 infants (≤ 90 days of age), found that a higher VIS at 48h post-cardiothoracic surgery was strongly correlated with prolonged mechanical ventilation and extended stays in both the ICU and the hospital. This easy-to-calculate bedside tool has been validated as an independent predictor of adverse outcomes, such as duration of mechanical ventilation, length of stay (LOS) in the ICU and hospital, and mortality [1, 8, 9]. For cardiac surgical patients, the VIS represented a significant advancement in assessing hemodynamic needs at specific time points post-surgery, facilitating meaningful comparisons across patients and institutions. Not limited to pediatric settings, the VIS also predicts adverse postoperative outcomes in adults. For instance, a study by Koponen et al. [10] conducted a noteworthy study that aimed to retrospectively evaluate the association between the highest VIS in the first 24 h post-ICU admission and a composite poor outcome in 3213 adult cardiac surgical patients, elucidating a linear increase in the odds of adverse primary postoperative outcomes with escalating ICU-VIS scores.

To date, most studies regarding the application of the VIS have been conducted in the respective authors’ countries and regions [11,12,13,14]. A recent systematic review published by Belletti et al. [15], which searched from 2010 to 2019, examined the evolution, clinical utility, and pitfalls of the VIS. Despite differing VIS calculation timings and cutoff values across studies, all concurred on its value as an outcome predictor. However, no comprehensive qualitative or quantitative reviews have yet been conducted specifically on the association between VIS and surgical patient outcomes. This study aims to address this gap, with primary outcomes focusing on the quantitative relationship between the VIS and surgical patient outcomes. Secondary outcomes will consider the optimal the VIS cutoff value for predicting these outcomes.

Methods

Protocol and registration

The systematic review and meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [16] and were registered at PROSPERO (registration number CRD42022359100) [17].

Search strategy

Three reviewers (YTS, WW, YTY) independently searched databases including PubMed, Web of Science, EMBASE, Scopus and Cochrane Library, Google Scholar, and CNKI from November 2010, when the VIS was first published, to December 2022. To maximize the chances of identifying relevant studies, different combinations of search words were used as follows: “(vasoactive and inotropic score) AND (mortality OR morbidities OR complications).” No restrictions were used.

Inclusion and exclusion criteria

We included the following studies: (1) population, pediatrics and adults undergoing any type of surgery; (2) study type, any; (3) outcomes, studies evaluating the association between the VIS and short- or long-term patient outcomes; (4) publication type, any article type; and (5) language, published in English.

Excluded studies were as follows: (1) review articles, case reports, guidelines, conference abstracts, letters, or studies with insufficient data; (2) duplications; (3) studies with incomplete or incorrect data and those not analyzing the association between VIS and outcomes; and (4) grey literature.

Study selection

After deleting duplicates records by using EndNote X7 reference management software, two of the authors (WW and YTS) independently examined the titles and abstracts of all potentially relevant studies and retrieved the full-text records for eligibility. Two reviewers (YTS and WW) independently reviewed the titles and abstracts of all identified studies for eligibility, followed by a full text review. Disagreement on inclusion was resolved by consensus and after discussion with the senior reviewer (YTY).

Data extraction

The data was independently extracted by YTS and WW into an Excel table, including study information (first author, publication year, country, years of collection, sample size, type of surgery), patient demographics (age, gender), the VIS data (time points of collection, cut-off value, validity of the VIS), and cross-sectional association between the VIS and outcomes (adjusted or unadjusted odds ratio (OR) with corresponding 95% confidence intervals (CI)). The OR value adjusted to the maximum extent for potentially confounding variables was selected, for only one model could be selected from studies reporting more than one adjusted mode. Predictive validity was defined as the ability of the VIS to predict patient outcomes, including the receiver operator characteristic (ROC) and area under the curve (AUC), sensitivity, and specificity. Disagreements were resolved by discussion with author (YTY).

Study quality assessment

Quality assessment was conducted independently by two reviewers (YTS and WW) using the Joanna Briggs Institute (JBI) Critical Appraisal Checklists for analytical observational studies [18]. Any disagreement in opinion regarding quality was resolved by discussion consensus with a third investigator (YTY). JBI Critical Appraisal Checklist contains 11 questions for cohort studies, 8 for cross-sectional studies and 10 for case control studies. All questions to determine the potential risk of bias can be answered with yes, no, unclear, or not applicable. If the answer is yes, the question is assigned a score of 1. If the answer is no, unclear, or not applicable, it is assigned a score of 0. A score of 4 to 6 indicates moderate quality, whereas as score of 7 or more indicates high quality.

Statistical analysis

The primary outcomes focused on the quantitative review of the association between the VIS and patient outcomes. Meta-analysis was conducted if two or more studies provided the same effect concerning the VIS. Adjusted ORs from multivariate aggressive analysis, along with their respective 95% of CIs, were extracted from each of the studies. The reported ORs were converted into log(OR), and the 95% CIs were transformed into standard errors (SE) using a random-effects model to pool the data [19, 20]. A random-effects meta-analysis was used because of the expected heterogeneity. Heterogeneity was assessed using the Q statistic and an I2 index score, with P < 0.10 and I2 > 50% considered statistically significant. Publication bias was evaluated through a visual inspection of funnel plots. Sensitivity analysis was considered to examine the influence of each study on the stability of the meta-analysis results. Subgroup analysis was attempted if possible to address the potential sources of heterogeneity. All analysis were performed using Review Manager 5.3 (Cochrane Collaboration, Denmark).

The secondary outcomes focused on the qualitative review of the predictive value of the VIS. In this review and meta-analysis, apart from one non-cardiac surgery [21], all other included studies were cardiac surgery including the large number of different surgical procedures, and almost all of the studies included were retrospective and the outcome variable varied from each study. We could not pool sensitivity and specificity estimates across different VIS cut-off values, as these varied between studies. Therefore, we summarized the results of each article individually (Table 2).

Results

Search results

Figure 1 illustrates the different phases of the search and selection processes. A total of 21,600 records were identified, of which 58 studies met the inclusion criteria and 34 studies were included in the meta-analysis. Apart from one prospective randomized study [22] and one retrospective secondary analysis of an RCT [23], all other included studies were observational and had low to moderate risk of bias according to the assessment results of JBI checklist scores (Table S1–S3).

Fig. 1
figure 1

PRISMA flow diagram of study selection

Full size image

Study characteristics

The VIS was applied in studies across 14 countries, with the majority conducted in the USA (n = 22), followed by China (n = 9), India (n = 5), Turkey (n = 5), Germany (n = 4), Korea (n = 3), Russia (n = 2), Spain (n = 2), Canada (n = 1), Finland (n = 1), France (n = 1), Japan (n = 1), Canada (n = 1) and Australia (n = 1). Data from a total of 29.920 patients with the VIS were included, of which 11.409 (38.13%) were female patients, with a mean age range of 6 (1–90) days to 68 (19–90) years. The number of participants ranged from 32 to 8543 in each included study. The VIS was recorded in the early postoperative period, including 24 h, 48 h, and 72 h. Among the fifty-eight studies, 42 involved pediatric populations, and 16 involved adult patients (Table 1). As shown in Table 2, the optimal cutoff value of the VIS as a predictor of outcomes varied widely, and none of the studies reported an identical VIS cutoff value. The ROC area ranged from 76% to 94%, with sensitivity from 53% to 90% and specificity from 74% to 88%.

Table 1 Characteristics of included studies
Full size table
Table2 Predictive validity of the VIS
Full size table

Primary outcomes

Studies demonstrating a correlation between the VIS and outcomes such as arrhythmias, nonocclusive mesenteric ischemia (NOMI), extracorporeal membrane oxygenation (ECMO) instituted, chylothorax, or postoperative infection are not pooled estimated odd ratio for only one study provided the effect measures of outcomes concerning the VIS, and we summarized the results of each article individually in Fig. 2. A meta-analysis was conducted if two or more studies provided the same effect measures of outcomes concerning the VIS. The association of VIS with mechanical ventilation duration [18, 25, 27, 34, 35, 37, 41, 44, 50, 51, 66], mortality [3, 26, 36, 41, 43, 58,59,60, 62, 71], acute kidney injury (AKI) [22, 33, 43, 45, 52, 55, 69, 70], a low cardiac output syndrome (LCOS) [11, 51], LOS in the ICU [1, 8, 28, 41, 50, 66], poor outcomes [1, 9, 11, 41, 50, 59, 61, 66, 71], and LOS in the hospital [8, 50] were meta-analyzed (Fig. 3). We only provided a first step in validating the association between the VIS and patient outcomes and did not consider timepoints. Despite of the timepoints of the VIS being recorded at 24 h, 48 h, and 72 h after surgery, we referred to them collectively as the early postoperative period. It was concluded that the early postoperative VIS is associated with mechanical ventilation duration (OR, 1.16; 95% CI, 1.08, 1.26; p < 0.00001), mortality (OR, 1.16; 95% CI, 1.08, 1.25; p < 0.00001), AKI (OR, 1.21; 95% CI, 1.10, 1.34; p < 0.00001), LOS in the ICU (OR, 4.07; 95% CI, 1.42, 11.66; p = 0.009), and poor outcomes (OR, 1.06; 95% CI, 1.01, 1.12; p = 0.02), respectively, in Fig. 3A, B, C, G, and D.

Fig. 2
figure 2

Correlation of the included studies between VIS and outcomes

Full size image
Fig. 3
figure 3

Meta-analysis on association between the VIS and patient outcomes

Full size image

Sensitivity analysis

Considering that significant heterogeneity among studies were detected for mechanical ventilation duration (I2 = 91%), mortality (I2 = 82%), AKI ( I2 = 89%), LOS in the ICU (I2 = 94%), and poor outcomes (I2 = 73%), sensitivity analysis using leave one out was conducted. Finally, when omitting the studies of Tadros et al. [25], Lex et al. [51], Luo et al. [34], Parmar et al. [27], Sun et al. [50], and Zhang et al. [35], heterogeneity greatly decreased for mechanical ventilation duration (I2 = 41%), and significant differences still existed (OR 5.20; 95% CI 3.78, 7.16; p < 0.00001) (Fig. 4A). When omitting the studies of Chen et al. [36], Dilli et al. [3], Gaies et al. [41], and Kulyabin et al. [43], heterogeneity greatly decreased for mortality (I2 = 44%), and significant differences still existed (OR 1.08; 95% CI 1.05, 1.12; p < 0.000001) (Fig. 4B). When omitting the studies of Beken et al. [33], Kulyabin et al. [43, 22], and Singh et al. [70], heterogeneity greatly decreased for AKI (I2 = 44%), and significant differences still existed (OR 1.26; 95% CI 1.13, 1.41; p < 0.0001) (Fig. 4C). When omitting the studies of Gaies et al. [41] and Sun et al. [50], heterogeneity disappeared for poor outcomes (I2 = 0%), and significant differences still existed (OR 1.02; 95% CI 1.01, 1.04; p = 0.0004) (Fig. 4D). When omitting the studies of Poterucha et al. [66] and Sun et al. [50], heterogeneity disappeared for LOS in the ICU (I2 = 0%), and significant differences still existed (OR 3.50; 95% CI 2.25, 5.44; p < 0.00001) (Fig. 4E). Considering only two studies were included, LCOS and LOS in the hospital could not undergo sensitivity analysis. The sensitivity analyses, using a leave-one-out approach, suggest that the findings are robust. This is important because the initial I2 statistics showed high heterogeneity, meaning that the studies were quite different in some way. Removing certain studies reduced this heterogeneity while preserving the statistical significance.

Fig. 4
figure 4

Meta-analysis on association between the VIS and patient outcomes after sensitivity analysis

Full size image

Subgroup analysis

Due to the considerable difference in surgical procedures, timepoints of the VIS, and only one [66] involved adult patients, the association between the VIS and mechanical ventilation duration was only assessed in different continents subgroups. The pooled estimate was different for those in North American (OR 6.56; 95% CI 4.47, 9.62; p < 0.000001) compared with those in Asia (OR 1.14; 95% CI 1.02, 1.29; p = 0.03) as shown in Fig. 5A. considering difference in study design, timepoints of the VIS, and definition of outcome variables observed, the association of the VIS with mortality, AKI, and poor outcomes was assessed in different age subgroups. For mortality, the pooled estimate was different for pediatrics (OR 3.09; 95% CI 1.73, 5.50; p = 0.0001) compared with adults (OR 1.08; 95% CI 1.04, 1.11; p < 0.00001) as shown in Fig. 5B. For AKI, the pooled estimate was different for pediatrics (OR 1.43; 95% CI 1.20, 1.70; p < 0.0001) compared with adults (OR 1.12; 95% CI 1.02, 1.24; p = 0.02) as shown in Fig. 5C. For poor outcomes, the pooled estimate was different for pediatrics (OR 1.12; 95% CI 1.05, 1.19; p = 0.0004) compared with adults (OR 1.02; 95% CI 1.01, 1.04; p = 0.0006) as shown in Fig. 5D. After age stratification, the heterogeneity decreased. The association of the VIS with mortality was more significant in pediatrics, and age did not affect the association of the VIS with AKI and poor outcomes. There are differences in the strength of the VIS association when looked at from the perspective of geographical location and age groups. This could be indicative of other factors (healthcare system quality, population health characteristics, etc.) affecting the outcomes, and it suggests that the VIS might be more or less useful in different settings or for different age groups.

Fig. 5
figure 5

Meta-analysis on association between the VIS and patient outcomes after subgroup analysis

Full size image

Discussion

In this study, the authors concluded that the elevated VIS in the early postoperative period could predict outcomes, including AKI, mechanical ventilation duration, mortality, poor outcomes, and LOS in the ICU. It is important to note that more than three-quarters of the adverse outcomes developed 24-h postoperative. Therefore, awareness of the VIS from the time of return to the ICU can aid the medical team in risk stratification, targeted interventions, and parental counseling. The VIS, with readily available parameters, is easy to calculate at bedside. As a straightforward hemodynamic observation parameter does not require laboratory examination, it becomes an advantage for medical institutions with limited facilities or in places with grassroots levels to early identify and stratify patients in high-risk cardiovascular surgery. However, careful interpretation is required due to the considerable difference in study design, timepoints of the VIS, and definition of outcome variables observed. In surgical center, patients were typically started on vasoactive‐inotropic agents by the attending anesthesiologist and the cardiac surgeon without uniform protocol. Progress may be affected by the clinical experience of the anesthesiologist and PICU physician. We did not consider the concurrent validity assessed the performance of the VIS in comparison to other scales (such as vasoactive-ventilation-renal (VVR) score [24], total inotrope exposure score(TIES) [28]) which may prove to be a better predictors of patient outcomes in the surgical theater. Despite the low to moderate risk of bias according to JBI checklist scores, there was only one randomized controlled trial included. This could limit the strength of the conclusions. Despite these challenges and limitations, VIS can be used as predictor of patient outcomes in the surgical theater as it is easy and simple to measure without the need for prolonged follow‐up. Our preliminary effort was to merely describe the strength of the association between the VIS and patients outcomes. Understanding how these statistically significant findings translate into clinical practice would be the next step. Prospective randomized trials are necessary to directly correlate the VIS to predicting morbidity and mortality after surgical operations.

The VIS was developed to quantify vasoactive and inotropic support after cardiac surgery in pediatric patients. Gaies et al. [1] analyzed the association between the VIS and clinical outcomes in infants undergoing cardiac surgery and demonstrated that maximum (max) VIS values within the first 48 h postoperatively were associated with poor outcomes in terms of cardiac arrest, circulatory support, renal replacement therapy, neurologic injury, and death. Max VIS values at 24 and 48 h were significantly associated with increased hospitalization and prolonged weaning periods among adolescents (10–18 years) with congenital heart disease [9]. Davidson et al. [8] concluded that higher VIS values at 48 h was strongly associated with increased intubation time and long-term ICU and hospital stay in infants (90 days) after cardiovascular surgery. In comparison with the pediatric population, many more studies were published in adults as well. Kwon et al. [59] found that the increased VIS during the immediate postoperative 48 h following of-pump coronary artery bypass grafting (OPCAB) was significantly associated with long-term morbidity and mortality up to 1 year. A study [60] applied the VIS score in the adult population and demonstrated that the adverse prognostic impact of a high preoperative VIS was especially marked among candidates bridged to transplantation under mechanical circulatory support (MCS). Alam et al. [44] found a higher inotropic requirement (depicted by higher VIS) to be significantly associated with delayed extubation. Hou et al. [69] retrospectively reviewed 1935 adult patients who underwent cardiovascular surgery and showed that the max VIS was associated with postoperative AKI and the need for RRT in AKI patients.

In some studies, no significant association was found between the VIS and outcomes after adjusting for confounders. Knight et al. [63] showed no significant association between the VIS and AKI at 7 days after controlling for covariates in bilateral orthotopic lung transplant recipients. Ödek et al. [39] concluded that lower VIS (p<0.05) was associated with early extubation (EE), but when combined into a multivariate model, there was no significantly associated with EE. Talwar [48] analyzed the VIS and found it was not related to mortality in patients who underwent the bidirectional Glenn procedure. Siehr et al. [53] revealed that the VIS was not identified as a significant predictor of hospital length of stay in infants who underwent stage 1 surgical palliation consisting of a modified Norwood procedure with right ventricle to pulmonary artery conduit. Asfari et al. [54] identified that a higher VIS at 36 h was not independently associated with mortality and hospital length of stay. Although no significantly associated was shown in those studies, the results may have been affected by domestic insurance regulations and other issues during the treatment course.

The optimal cutoff value of the VIS as a predictor of adverse outcomes varied from 10 to 30, and none of the studies reported the same VIS cutoff value. This discrepancy in cutoff value is likely due to the differences in the definition of poor outcomes, baseline characteristics, and surgical procedures between studies. Liu et al. [64] found that the VIS ≥ 6 was a risk factor for noninvasive ventilation (NIV) failure in patients with post-extubation ARF after cardiac surgery. In a study by Zhang et al. [56], it was reported that the max VIS over the first 24 h was a good predictor of poor clinical outcome. VIS ≥ 8.5 (OR 1.294, 95% CI 1.210−1.384) was identified as risk factor for shunt failure. Baysal et al. [61] attributed prognostic implications to a postoperative VIS > 5.5 as an independent predictor of morbidity and mortality after coronary artery bypass grafting in their prospective evaluation. At the same time, with the understanding that VIS only allows for the hemodynamic support quantification at a single time-point, the conceptualization of a VIS index by Crow et al. [30], a VIS index ≥ 3 has been outlined to be associated with an increased risk of poor composite outcomes after cardiac surgery.

The VIS seems to have some prognostic value in predicting the potential need for ECMO in the early postoperative period. Xie et al. [72] focused on the VIS at the initiation of ECMO and at the weaning of ECMO to assess recovery of heart function and found that early initiation of ECMO could potentially lead to improved outcomes in these critically ill patients, before inotropic and vasopressor therapy escalation. Kuraim et al. [32], after infant cardiac surgery among 20/565 (3.5%) patients who were placed on veno-arterial ECMO in this early post-operative period after ICU admission, concluded that the highest VIS postoperative day 1 was associated with early ECMO. Friedland-Little et al. [29] identified a peak VIS of 27 within 48 h of surgery as most prognostic of the need for ECMO. The VIS is a good predictor, but further study is needed to determine whether replacing the addition of a third or fourth vasoactive agent with mechanical circulatory assist devices would be more beneficial to the patient’s prognosis. Our findings would serve as hypothesis-generating information to design and conduct prospective trials for validation.

Several possible underlying mechanisms linking the VIS to patients outcomes have been explored. The VIS is used to quantify the amount of cardiovascular support, and a higher VIS score may indicate more severe hemodynamic disorder [1, 41]. Hemodynamic perturbations, especially hypoperfusion, are considered to be important determinant in the development of poor outcomes after cardiovascular surgery [73]. High doses of vasoactive and inotropic medications, particularly catecholamines, have been shown to have a harmful effects on organ function and cause immune-mediated injury [74]. Cardiovascular surgery is highly traumatic, and the postoperative period is prone to excessive bleeding, LCOS, and hypovolemia. These conditions directly manifest in low blood pressure values, leading to reduced organ blood perfusion, which preferentially supplies vital organs such as the heart and brain, causing ischemic perfusion injury and multiple organ failure [58]. Vasoactive medications can increase myocardial oxygen consumption, cardiac arrhythmia, hypertension/hypotension, and peripheral and cardiac ischemia, which may be fatal [75]. Furthermore, catecholamine use has been associated with immunosuppression, bacterial growth, increased bacterial virulence, biofilm formation, insulin resistance, and hyperglycemia [76]. Catecholamine-induced metabolic changes include increased oxygen consumption, glycolysis, glycogenolysis, lipolysis, gluconeogenesis, ketogenesis, peripheral insulin resistance, and lactate release, which can lead to acidosis that may decrease the effect of inotropic or vasopressor drugs and often be a reason for increasing doses [77]. All of these factors can contribute to multiple organ dysfunction, making recovery from cardiovascular surgery more difficult.

In the era of electronic data management systems, it is suggested to program the hourly automatic calculation of the VIS into an electronic medical record system. This would allow for the incorporation start and stop dates as well as dose changes in studying the relationship between multiple clinical parameters and prognosis status over a defined time period in postoperative cardiovascular patients. Meanwhile, with the increasing reliance on digital systems and electronic health records, the potential risks associated with data breaches, unauthorized access, and cyberattacks cannot be overlooked [78, 79].

To properly interpret our study results, it is necessary to understand several limitations: (1) The VIS is a sum of the total vasopressor dose at a single timepoint, without providing information on the duration of exposure and incorporating dose magnitude into the equation. Incorporating duration and magnitude of vasopressor requirements during the postcardiac surgical period could improve the VIS sensitivity for predicting outcomes. (2) In this review and meta-analysis, almost all of the studies included were retrospective, which has certain limitations. In the case control studies, the lack of strategies to address incomplete follow up utilized accounted for main part. In cohort studies, the lack of the exposure period of interest long enough to be meaningful accounted for the main part. In cross-sectional studies, the lack of strategies to deal with confounding factors stated accounted for main part. (3) We only provided a first step in validating the strength of the association between the VIS and patients outcomes and did not consider timepoints. Despite of the timepoint of the VIS being recorded at 24 h, 48 h, and 72 h after surgery, we referred to them collectively as the early postoperative period. This could be a limitation, as the timing of measurements can sometimes affect the outcome. (4) We did not consider the concurrent validity assessed the performance of the VIS in comparison to other scales (such as SAPS III, OHCA score, RACA score, EuroSCORE, MR-proADM, APACHE II, SNAPPE-II, Aristotle basic complexity score (ABC), IS, APACHE III score, SOFA score) which also could be used for predicting outcomes in cardiovascular patients. (5) In surgical center, patients were typically started on vasoactive-inotropic agents by the attending anesthesiologist and the cardiac surgeon without uniform protocol. Progress may be affected by the clinical experience of the anesthesiologist and PICU physician. (6) Apart from two multi-center studies [45, 60], all other included studies were single center, and we only included studies published in English, thus affecting the methodological quality of this review and potential publication bias. The applicability of this meta-analysis to the broader patient population may be limited given that most studies involved were conducted in specific countries. At the same time, incomplete retrieval of identified research affected the reporting bias.

Conclusions

A higher VIS in the early postoperative period could predict outcomes, the VIS, which is easily calculated from routine work, could assist in predicting outcomes in high-risk cardiovascular surgery and may aid in risk stratification and targeted interventions. However, further prospective studies are required to assess its prognostic value, to validate its association with major adverse events, and to specify optimal doses and combinations of vasoactive inotropic medications.