Current research focuses on preventing mortality and reducing the risk of morbidity in children who undergo congenital heart disease (CHD) surgery with cardiopulmonary bypass (CPB). In order to accomplish these objectives, it is essential to monitor the diverse variables influencing these potential outcomes, such as the levels of inotrope score (IS) and/or of vasoactive-inotropic score (VIS), and their prognostic value in the post-operative period.1–3 Nevertheless, the association between VIS and the post-operative course of the patient remains controversial.

Wermousky et al.1 have quantified the degree of inotrope support received by neonates after surgical repair of transposition of the great arteries, and have reported that the maximum IS within the first 48h was an indicator of poor prognosis, of higher incidence of prolonged mechanical circulatory support, renal replacement therapy, neurological damage, of cardiac arrest and death. In addition to this, IS was also associated with longer duration of mechanical ventilation (MV), paediatric intensive care unit (PICU) and hospital stay.2 On the other hand, other authors have been unable to establish an association between VIS and the early post-operative outcomes in neonates who underwent heart surgery with CPB.4 Therefore, elevated VIS levels could not be associated with prolonged MV duration and PICU stay. Likewise, these authors have found that low cardiac output syndrome (LCOS) was not associated with MV duration or length of PICU stay.

Neonates undergoing CPB are more likely to present increased risk of morbimortality compared with other age groups. Thereby they might be considered a special group. Gaies et at.5 have studied 391 children (only one-third were neonates) who underwent CHD surgical repair, in which VIS>15 within the first 24h post-operatively was considered empirically high. A high VIS level was strongly associated with mortality at 30 days following surgery, with incidence of cardiac arrest, need for mechanical circulatory support, dialysis, neurological damage, MV duration and length of PICU stay. In another work on VIS as a short-term outcome biomarker, a maximal VIS on the second post-operative day predicted adverse outcomes in adolescents following cardiac surgery.6

There are other studies regarding the paediatric population that centre on cardiac biomarkers, such as the β-type natriuretic peptide (BNP),7 mid-regional pro-atrial natriuretic peptide (MR-proANP),8 pro-adrenomedulina (MR-proADM),9 cardiac troponin-I (cTn-I)10 and copeptin11 as early predictors of LCOS to monitor the post-operative period after CHD surgery with CPB. To date, there are no published studies that relate IS, VIS and cardiac biomarkers levels with the prediction of LCOS in children during the post-operative period after CHD surgery requiring CPB.

In order to address this knowledge gap, we have performed a study of CHD children to evaluate whether IS and VIS scores improve the predictive power for LCOS.IS and VIS have been estimated in the immediate post-operative period and during LCOS development and have been associated with the cardiac markers above mentioned predictors of LCOS in the paediatric population. These could be useful in aiding in therapeutic decision-making in clinical practice after CPB, and in improving the prognosis of the patient.

The study enrolled 117 consecutive paediatric patients (aged 10 days to 15 years) who were admitted to a PICU after corrective surgery for congenital heart disease with CPB. Two patients were excluded from the study because they perished within the first 2h following PICU admission. Six patients developed cardiogenic shock and required extracorporeal membrane oxygenation within the first 48h post-CPB. Table 1 displays the type of congenital heart disease and the patient characteristics of the entire cohort of our study.

Thirty-three patients (29%) exhibited LCOS following corrective cardiac surgery with CPB, while the other patients were included in the non-LCOS group. IS and VIS levels were significantly more elevated in LCOS patients in all the CPB time points (p<0.001) (Fig. 1). cTn-I and MR-proADM levels, after attaining a peak value at 2h post-operatively, decreased in the group without LCOS, whereas they increased in the LCOS group (these data have been published in Rev Esp Cardiol (Engl Ed). 2017; 70: 267–274).

Figure 1.

(A, B) Differences in inotrope score (IS) and vasoactive–inotropic score (VIS) among patients according to presence or absence of low cardiac output syndrome (LCOS) in the immediate postoperative period of cardiopulmonary bypass (CPB) at a Paediatric Intensive Care Unit (PICU). —— LCOS; - - - - Absence LCOS,. Represent the mean, * Significant differences (p=0.01) between LCOS and Absence of LCOS.

(0.06MB).

The results by the univariate and multivariate logistic regression analyses to predict LCOS after CHD surgery with CPB are shown in Table 2. There was no significant association between the gender of the patients, PIM-2 and the Aristotle score with the development of LCOS at 48h post-CPB. Although the univariate logistic regression analysis indicated that there was an association between IS levels and LCOS, in the multivariate logistic regression we did not observe such association. For this reason, we constructed models to predict LCOS using the VIS score, the significant variables of the univariate model, and cTn-I and ADM levels. The cut-off points for these two cardiac biomarkers were the same as established as predictors of LCOS in a previous study11 (cTn-l>14ng/ml and ADM>1.5nMol/L).

The model 1 exclusively comprised, aside from VIS, variables such as age and CPB time points. Age and CPB were excluded from the model 2 when cardiac biomarker were included in it. Comparisons between AUCs in both predictive models did not show statistical differences by Hanley and Mc Nail methods (p>0.05). Table 3 shows the sensitivity, specificity, predictive values and the likelihood ratio of the curves generated by the two models. The VIS score correlated positively with the duration of MV (rho: 0.385; p: 0.001) and with the stay at a PICU (rho: 0.465; p: 0.001).

LCOS after CPB is frequently observed after corrective surgery for CHD in children. Normally, ventricular dysfunction following cardiac surgery with CPB peaks within 8–12h post-CPB, and it gradually recovers within 24–48h post-operatively.17 However, there are patients in whom LCOS is prolonged, as demonstrated in our study, in which LCOS occurrence at 48h post-CPB was in 29% of the patients.

To diagnose LCOS is hard to achieve, especially in children, in whom the validation of estimation methods for LCOS is complex, and invasive procedures present great many limitations, particularly in younger children. In consequence, the current study includes cardiac function measurement by echocardiography, clinical variables that have been traditionally utilised as shock markers and quantitative data, such as the data obtained by arterial thermodilution.

The development of LCOS suggests a higher rate of complications, hence predicting it could positively influence in the post-operative course of the patient. Assessing the need for post-CPB vasoactive inotropic support using the IS and VIS scales to predict early LCOS can contribute to decision making in clinical practice. Accordingly, it has been observed that the prediction of LCOS sustained after 48h post-CPB was strongly associated with VIS at 2h of PICU admission, adjusted forage and duration of the CPB. These results support those indicated by Wernosky,1 who has determined the value of these scales in a group of neonate sunder going CPB.

Following PICU admission, according to results, the use of inotropic support varies through to the period after CPB, which increases in LCOS patients. This relatively facilitates LCOS prediction as the time span fore estimating LCOS expands but its practical utility diminishes. Consequently, it is pivotal to find a nearly indicator of LCOS by assessing the efficacy of IS and VIS at different time points. In this study, at 2h after PICU admission, VIS displayed a high LCOS predictive power post-CPB, with a cut-off point of 15.5.IS, at 2h post-CPB, also predicted LCOS in the univariate analysis but opposing VIS, it did not obtain significance in the multivariate analysis. Therefore, VIS appeared to be more efficient than IS, probably because the use of IS has some important limitations as this mainly excludes vasoactive medication, such as nitroprusside, nitro-glycerine, phosphodiesterase inhibitors and vasopressin. For this reason, Davidson et al.,3 who have undertaken a prospective, observational study of 70 children submitted to cardiac surgery, proposes a modified version of IS that includes, like VIS, vasoactive drugs.

The ability of the different variables assessed in our study to predict LCOS was analysed by multivariate logistic regression, obtaining important significance in age, CPB duration and VIS at 2h post-CPB, with a cut-off point >15.5 for VIS (Table 3). This cut-off point showed high specificity and high negative predictive value to predict LCOS at 48h following CPB.

cTn-I levels, which increase rapidly after cardiac injury, could function as a predictive cardiac biomarker during the post-operative period in patients with CHD submitted to cardiac surgery.18,19 Preceding studies have demonstrated that cTn-I levels at 4h post-CPB enable to predict the post-operative course of paediatric patients undergoing cardiac surgery.20 Bojan et al.19 have described that an early increase in cTn-I levels may be useful in predicting the post-operative course of neonates and infants subjected to cardiac surgery. Froese et al.,20 using a cut-off point 13ng/ml for cTn-I at 4h post-CPB, have proven that the early measurement of cTn-I may allow the early diagnosis of LCOS in CHD children undergoing surgery with CPB.

Adrenomedullin (ADM) was discovered in 1993 in human pheochromocytoma tissue and subsequently found to be a circulating hormone. Several studies have demonstrated that MR-proADM alone or combined with other parameters may be a predictor of LCOS in children undergoing CPB.21,22 Moreover, our research group has shown that a cut-off point >14ng/mL for cTn-I and a cut-off point 1.5nMol/L for ADM at 2 and 24h post-CPB, respectively, has high sensibility and specificity for determining LCOS.11

Based on previous findings, our study includes cTn-I and ADM levels as predictive variables with the object of improving the goodness-of-fit. However, the predictive power for LCOS did not rise when these cardiac biomarkers were utilised. We consider these results to be of great relevance as they enable an early estimation of the post operatory course of the patients with CHD. In addition, VIS scores were related to indirect indicators of worse evolution, such as MV duration and the length of PICU stay. The results of this work indicate that the monitoring of cardiac biomarkers in the immediate postoperative period following CHD does not provide more any further clinical information, as the ability to predict LCOS with VIS does not increase with the measurement of cardiac biomarker levels.

The present study has some limitations to consider. The wide range of CHDs required different types of surgical corrections. A solution to this drawback was to study a large sample size in a single hospital, resulting in minimal variability for patient management. LCOS was considered based on restrictive inclusion criteria. All of the data were adjusted for the age and the body-weight of the patients to prevent these variables from influencing our results.

In conclusion, our findings have shown that the value of VIS at 2h post-CPB was itself an early independent predictor of LCOS. This predictive value did not increase when associated with LCOS cardiac biomarkers. VIS was more useful than IS in making early therapeutic decisions in clinical practice after corrective surgery of CHDs by CPB. Therefore, these results could be useful to improve the prognosis of CHDs patients after corrective surgery.

BRAHMS GmbH Biotechnology Centre, Hennugsdorf, Berlin, Germany, in part supported this study. The partial sponsor of the study had no role in the study design or in analysis, and did not participate in writing the report.

Juan L. Pérez-Navero, Ignacio Ibarra de la Rosa and Maria José de la Torre-Aguilar conceived, planned the project and drafted the manuscript experiments. Carlos Merino-Cejas, Susana Jaraba-Caballero and Manuel Frias-Perez carried out the experiments and contributed to sample preparation. Mercedes Gil-Campos and Elena Gómez-Guzmán contributed to the interpretation of the results. All authors provided critical feedback and helped shape the research, analysis and manuscript.

The authors report no conflicts of interest. The authors are responsible for the contents and writing of the paper.