Introduction

In 1980, Barash et al. reported extubation in the operating room after congenital heart surgery in children had positive consequences on lung function and survival, but there were no Fontan patients among their collective [1]. General advantages of early extubation such as an avoidance of laryngotracheal trauma, accidental extubation, mucous plugging, oversedation, and increased incidence of pulmonary infections [2, 3] seem to outweigh the potential necessity to suppress systemic inflammation and stress response after cardiopulmonary bypass (CPB) by giving high-dose opioids [4, 5].

The exclusion of the right heart from the venous system and the resulting passive blood flow through the pulmonary vascular bed make the Fontan circulation a unique, still not fully understood entity. It has already been shown that venous blood return from the body to the pulmonary arteries is enhanced by spontaneous ventilation [6,7,8]—an important effect after Fontan operation (FO) as right ventricular suctioning effect is missing [9]. Several studies report early extubation shortens length of hospital stay, favorably influences hemodynamics, and reduces complications after FO [10,11,12].

However, these reports do not provide evidence that extubation on the operating table is superior to later extubation in the intensive care unit (ICU).

Materials and Methods

Patient Characteristics

Between 2013 and 2016, 114 consecutive children with a single ventricle malformation, 71 (62%) boys and 43 (38%) girls, mean age of 3.8 ± 2.3 years, and mean weight of 15.2 ± 6.8 kg underwent FO in our institution. The most common malformations were hypoplastic left heart syndrome (HLHS) (n = 56, 49.1%), HLHS variants (n = 15, 13.2%), hypoplastic right heart syndrome (HRHS) (n = 12, 10.5%), HRHS variants (n = 13, 11.4%), and double-inlet left ventricle (DILV) (n = 9, 7.9%). HLHS-variant group included unbalanced atrioventricular canal with hypoplastic left ventricle, double-outlet right ventricle with hypoplastic left ventricle, and transposition of the great arteries with hypoplastic left ventricle. HRHS-variant group included unbalanced atrioventricular canal with hypoplastic right ventricle or tricuspid atresia with transposition of the great arteries.

The first 53 children (ICUE group) were extubated in the ICU within a median time of 195 min after surgery (range 30–515 min). After a change in strategy in January 2015 to extubation on the operating table, the following 61 children were extubated in the operating room immediately after chest closure (ORE group). One child from the ORE collective developed laryngospasm immediately after extubation and was re-intubated (its data were placed in the ICUE-group analysis). Therefore, the final ICUE group consists of 54 children, whereas the ORE group consists of 60 children. The ICUE group was further subdivided into children who were ventilated ≤ 119 min (ICUE early, n = 9), 120–239 min (ICUE intermediate, n = 27), and ≥ 240 min (ICUE late, n = 18). Surgery and anesthesia were performed by the same two pediatric cardiac surgeons and two anesthesiologists according to the same and constant protocol.

Anesthetic Management

ORE Group

Anesthetic management in all children consisted of intravenous induction of anesthesia with sodium thiopental (5–7 mg/kg), remifentanil (0.5 µg/kg/min) and cisatracurium besilate (0.2 mg/kg).

After endotracheal intubation, sevoflurane was added to the oxygen-enriched air mixture (0.5–1.5%) and a continuous infusion of dexmedetomidine was started (1.0 µg/kg/h up to a total dose of 1.5 µg/kg and then 0.1 µg/kg/h). Analgesia was maintained with remifentanil (0.2–0.5 µg/kg/min). During CPB, sevoflurane was administered through the extracorporeal circuit. On termination of CPB, a single dose of piritramide (0.1 mg/kg) and metamizole (15 mg/kg) was administered. When deemed necessary, dopamine (3–7 µg/kg/min) was initiated. ORE criteria included no significant bleeding and stable satisfactory hemodynamics during closure of the skin. Remifentanil and sevoflurane were then discontinued. Having regained protective reflexes and a regular breathing pattern, the endotracheal tube was removed on the operating table and oxygen was administered via nasal cannula.

ICUE Group

The anesthetic protocol as described in the ORE group was followed except that sufentanil (anesthesia induction: 0.5–1.0 µg/kg, and then continuous infusion: 3–5 µg/kg/h) was used instead of remifentanil and dexmedetomidine. At the end of surgery, sevoflurane was terminated and midazolam (0.1 mg/kg) administered. Sufentanil was discontinued after admission to the ICU. Morphine (5–50 µg/kg/h) and midazolam (0.1–0.3 mg/kg/h) infusions were used until the children were extubated.

Surgical Procedure

All 114 children underwent a total cavopulmonary connection operation using an extracardiac conduit without fenestration. The surgery was performed on a “beating heart” during normothermic CPB after bicaval and aortic cannulation. Eighteen (15.8%) children received cardioplegic cardiac arrest or a short period of induced ventricular fibrillation in order to perform additional intracardiac procedures. During FO, the inferior caval vein (IVC) was divided from the right atrium and anastomosed to a polytetrafluoroethylene tube (PTFE), 18 mm diameter in 112 (98.2%) cases and 20 mm in one child. Opposite the Glenn anastomosis, the right pulmonary artery (RPA) was opened longitudinally and the superior end of the PTFE conduit was anastomosed “side-to-end” to the RPA. In one patient with an absent hepatic segment of the inferior vena cava, the anastomosis was performed between the stem of the hepatic veins and the RPA, using a 16-mm PTFE conduit. After weaning from the CPB, modified ultrafiltration was performed in all patients.

ICU Management

The ICU children’s treatment consisted mainly of dopamine (3–7 µg/kg/min) and milrinone (0.25–0.75 µg/kg/min) infusion as well as fluid therapy with crystalloid solutions. Analgesia was provided with piritramide (0.05–0.1 mg/kg, 4–6 times per day), paracetamol (15 mg/kg, four times per day), as well as metamizole (10–15 mg/kg, four times per day) and dexmedetomidine (0.2–1 µg/kg/h). Furosemide (1–6 mg/kg in several doses), spironolactone (1–2 mg/kg, one time per day), and hydrochlorothiazide (1 mg/kg, two times per day) were administered depending on the volume status of the child.

Cefuroxime (100 mg/kg/day) was administrated for 3 days including the day of the operation (perioperative prophylaxis).

Every child received supplementary oxygen (if still intubated: 30–50%; if extubated: 3–6 l/min). In the ICUE group, when the child regained adequate reflexes and a spontaneous breathing pattern after reduction of respiratory support and sedation, the child was extubated.

The postoperative infection was diagnosed if a secondary increase of infection parameters with fever was observed. The amounts of pleural and pericardial effusions were analyzed. A right pleura drainage tube was routinely placed during surgery. Left pleura drainage was placed in the postoperative course if needed. Drainage tubes were removed when the daily effusions fell below 5 ml/kg.

Data Collection

All data were collected and analyzed retrospectively. Pre-, peri-, and postoperative data were taken from the medical records in the hospital data bank system. Hemodynamic parameters (heart rate, arterial blood pressure, central venous pressure, urine output) were collected for each of the following 24 postoperative hours after the arrival on the ICU. From the mean values for each postoperative hour, one mean value for each parameter was calculated and compared between the groups. The mean values of the blood gas analysis (BGA) parameters taken between third and fourth hour after FO were analyzed. The period of time in which the children received inotropic support with dopamine (> 1 µg/kg/min) and milrinone (> 0.05 µg/kg/min) was recorded. To analyze the effect of the extubation in the operating room on the postoperative course in the high-risk patients, from the whole study population was selected a group of children with the higher transpulmonary gradient (≥ 5 mmHg) or higher end diastolic pressure in the single ventricle (≥ 9 mmHg).

Statistical Analysis

Mean values and standard deviations were used for continuous variables with normal distribution and median and range values in case the data were not normally distributed. Categorical variables were presented by using number of patients and frequencies. The groups were compared using the Mann–Whitney U Test for continuous variables and χ2-test as well as the exact Fisher test for categorical variables. In case of a comparison between three groups with continuous variables, the Kruskal–Wallis Test was used. p Values < 0.05 were considered to be statistically significant. All statistical analyses were performed with IBM SPSS Statistics v24.0.

Results

Preoperative Characteristics

Except for more RPA balloon dilatations in the ORE group, there were no significant differences between the groups regarding the preoperative demographic, cardiac catheterization, and echocardiographic data (Table 1).

Table 1 Preoperative characteristics
Full size table

Intraoperative Characteristics

There were no differences in the duration of the operation (ICUE group: 168.7 ± 40.6 vs. ORE group: 159.4 ± 46.6 min; p = 0.167), CPB (ICUE: 48.5 ± 19.0 vs. ORE: 49.0 ± 29.8 min; p = 0.330), or aortic cross-clamping (ICUE: 22.8 ± 21.28 vs. ORE: 21.8 ± 13.9 min; p = 0.673) between the groups.

Within the ICUE group, the children who stayed intubated for ≥ 240 min (n = 18) had a significantly longer CPB time than patients who were extubated within the first two hours (n = 9) after the operation (59.0 ± 25.2 vs. 38.0 ± 11.7 min, respectively; p = 0.025). Table 2 contains other surgical details.

Table 2 Surgical details
Full size table

Postoperative Characteristics

The hospital survival rate of all patients was 100%. Children who were extubated on the operating table had lower mean heart rate (106.5 ± 4.0 vs. 120.3 ± 6.0 bpm; p < 0.001), lower mean systolic (90.6 ± 2.1 vs. 93.5 ± 6.4 mmHg; p = 0.024), lower mean diastolic (49.6 ± 1.4 vs. 53.6 ± 4.3 mmHg; p < 0.001), and lower mean central venous pressure (10.4 ± 0.7 vs. 11.4 ± 1.4 mmHg; p = 0.001) within the first 24 h after FO than ICUE children (Fig. 1).

Fig. 1
figure 1

Hemodynamics within first 24 h after Fontan operation [intensive care unit extubation group (ICUE)—solid line, operating room extubation group (ORE)—dotted line; *p < 0.05]

Full size image

However, systolic blood pressures of ORE children exceeded those of ICUE children within the first 7 h after operation (88.6 ± 2.5 vs. 85.6 ± 2.6 mmHg; p = 0.036). The diuresis of ORE children was lower compared to ICUE children (3.7 ± 0.8 vs. 4.8 ± 1.0 ml/kg/h; p < 0.001). Among ICUE children, a shorter period of mechanical ventilation was accompanied with slower heart rate [114.2 ± 7.5 (ICUE early) vs. 119.1 ± 6.1 (ICUE intermediate) vs. 124.9 ± 7.0 bpm (ICUE late); p < 0.001], higher systolic pressure (97.3 ± 4.8 vs. 93.0 ± 6.5 vs. 91.2 ± 7.6 mmHg; p = 0.017), higher diastolic pressure (55.8 ± 4.2 vs. 53.4 ± 4.3 vs. 52.7 ± 5.0 mmHg; p = 0.047), lower central venous pressure (10.6 ± 1.5 vs. 10.8 ± 1.7 vs. 12.8 ± 1.6 mmHg; p < 0.001), and lower urine output (4.2 ± 1.8 vs. 5.0 ± 1.6 vs. 5.0 ± 1.0 ml/kg; p = 0.014).

The ORE children required less inotropic support than ICUE children: they received dopamine over 9.8 ± 11.3 h vs. 12.8 ± 11.7 h; (p = 0.033) with significantly higher doses (Fig. 2). The requirement for intravenous crystalloid fluid administration within the first 24 h after FO was significantly lower among ORE children than ICUE children (54.1 ± 31.2 vs. 73.8 ± 46.1 ml/kg; p = 0.019), whereas the need for metamizol was higher (40.0 ± 13.8 vs. 21.5 ± 16.3 mg/kg; p < 0.001). The use of piritramide and paracetamol was almost the same. The ORE children did not require antibiotic therapy for as long as the ICUE children did (4.7 ± 3.4 vs. 5.8 ± 4.6 days; p = 0.037).

Fig. 2
figure 2

Dopamine administration after Fontan operation

Full size image

Although the total period of chest drainages left in the thorax did not differ significantly, the amount of pleural effusions was higher in the ICUE group than in the ORE group (Fig. 3). There were a similar number of children requiring a secondary chest drainage [23 (38%) vs. 22 (41%); p = 0.902].

Fig. 3
figure 3

Fluid loss via chest drainages

Full size image

The comparison of the blood gas analysis, performed between the third and fourth hour after the operation revealed that values of the ICUE children were shifted towards metabolic acidosis, whereas ORE children’s blood gases were usually normal (Table 3). The results of the postoperative parameters analysis in the high-risk patient group are presented in Table 4.

Table 3 Postoperative data
Full size table
Table 4 Postoperative parameters in the high-risk patient group
Full size table

The length of the ICU stay (ICUE group: 2.3 ± 2.5 day, ORE group: 2.6 ± 1.6 day; p = 0.179) and hospital stay (ICUE group: 15.2 ± 7.3 day, ORE group: 17.4 ± 6.3 day; p = 0.640) was not influenced by the earlier extubation. All children were discharged home. At discharge, the ORE children required significantly lower doses of furosemide than ICUE children (0.5 ± 0.6 vs. 0.7 ± 0.6 mg/kg/day; p = 0.023). ICUE children who were ventilated for ≥ 240 min showed a significantly higher maximum postoperative level of ALT than ORE children (363.1 ± 551.6 vs. 174.7 ± 472.1 U/l; p = 0.029). The comparison between ICUE children ventilated for ≤ 119 min and ≥ 240 min also revealed higher maximum liver enzyme elevations among the longer ventilated subgroup (ALT: 31.4 ± 29.1 vs. 363.1 ± 551.6 U/l; p = 0.009, AST: 48.0 ± 13.5 vs. 422.1 ± 603.2 U/l; p = 0.019, GGTP: 17.3 ± 9.9 vs. 42.9 ± 28.3 U/l; p = 0.005) as well as longer administration of dopamine (8.1 ± 6.8 vs. 16.2 ± 9.2 h; p = 0.020).

Discussion

Despite the slight difference of merely 3 h between ORE and ICUE, extubation on the operation table had significant effects on the postoperative course after FO emphasizing how readily the Fontan circulation is affected by spontaneous ventilation.

Central venous pressure and SV output are the main driving forces that push blood to the Fontan patient’s pulmonary arteries [13]. Natural inspiration which creates negative intrathoracic pressure was shown to increase blood flow in the pulmonary arteries [14] and consequently represents a supportive factor for systemic blood return [9].

Hypotension and tachycardia, observed in the ICUE collective, are symptoms of a low cardiac output syndrome (LCOS) [15]. Postoperative positive pressure ventilation impedes venous blood mobilization from its reservoirs and raises central venous pressure [16] and transpulmonary pressure gradient (TPG), thereby reducing SV preload and cardiac output [17]. Another factor that plays a role in LCOS pathogenesis after FO is a sudden unloading of the SV in patients who had an additional lung perfusion source, resulting in a decrease of myofilamental strain and calcium release (“disuse hypofunction”) [18]. Additionally, an activation of inflammatory cascades caused by CPB contributes to the LCOS [19].

The ICUE group’s blood pressure began to surpass the ORE group few hours after ICU arrival and ICUE children showed higher total blood pressures over 24 h. The reason for this certainly lies in the more aggressive treatment of their initial hypotension in the form of higher dopamine and crystalloid volume administration, and as a consequence a higher urine production was observed. Kawaguchi et al. recently reported on constantly higher blood pressures and urine production when early extubation was performed [10]. However, their ICUE collective was sedated and ventilated for another 116 h which is hardly comparable to a mean prolonged ventilation of 195 min in our study group. Nevertheless, their ORE group showed lower heart rates and a reduced need for fluids as well.

The central venous pressure of ICUE children was constantly elevated during the first 24 h. Also liver enzymes showed higher elevations in the children with longer time of mechanical ventilation. These findings suggest that suctioning effects induced by negative intrathoracic pressure are not only able to enhance cardiac output but also to effectively prevent venous stasis and end organ damage. Our findings correlate with the observations of Lofland and Mutsuga who both noted a decline in central venous pressure after extubation of Fontan patients [20, 21]. Based on these observations, it appears logical that ORE children also showed significant reductions in pleura effusions. Prolonged fluid loss via chest drainages is eventually caused by elevated pulmonary artery pressure [22,23,24] and represents one of the most important risk factors for long hospital stays [25] and late Fontan failure [26].

We were able to support these findings by doing a sub-analysis of the ICUE group comparing three different durations of mechanical ventilation and the same differences as in the ICUE–ORE comparison were observable: with shorter time of mechanical ventilation, the lower heart rate as well as central venous pressure and the higher blood pressure were observed.

The ICUE children had a metabolic acidosis 3–4 h after FO more often, which is also a sign of the LCOS. Hypoxemia and acidosis are both conditions which trigger a pulmonary vasoconstriction [27], causing a mPAP and TPG elevation which further exacerbates the LCOS. Bearing in mind that CPB-caused cytokine storm additionally impedes pulmonary blood flow [2, 4, 28, 29] raises further doubts about the practice of prolonged mechanical ventilation.

In our study, extubation on the operating table additionally reduced the requirement of antibiotics, which is probably related to the reduced number of airway infections. Ventilator-associated pneumonia represents the most common and serious infection after cardiac surgery and is associated with adverse outcomes [30].

Moreover, a link between the length of intraoperative extracorporeal circulation and postoperative need for ventilation was found: The longer CPB was running, the later extubation could be performed within the ICUE group. This causality appears plausible as the inflammatory response to CPB and hypothermia also leads to formation of interstitial edema in the lung that impedes gas exchange [3, 4].

This study has important limitations: it is a retrospective, observational study and involves a relatively small number of patients from a single center. There were important differences between the groups regarding the operative anesthetic management of these patients.

Conclusions

Extubation on the operating table after Fontan operation appears to be safe and may offer early physiological advantages to patients with a venous pressure-driven pulmonary circulation, reducing the need for postoperative therapeutic interventions and avoiding complications associated with endotracheal intubation. This regimen seems to promote faster hemodynamic adaptation to the Fontan circulation.