At our institution, the practice of early postoperative b-B with propranolol after TOF repair was started in 2011 and quickly became a common practice, with more than 50% of patients treated between 2014 and 2019. This proportion of postoperative b-B is much higher than previously reported in a North American national retrospective database review [26], in which 10% of admitted patients received postoperative b-B. However, this study’s population was not comparable with ours. Their use was based on individual intensivist’s preferences according to the patient’s clinical situation and evolution during the first postoperative hours. Pre-existing or new heart block was a contraindication to b-B administration. From a theoretical framework, the diastolic function is correlated with relaxation time; with a lower heart rate, more time is spent in diastole, and ventricular filling capacity is enhanced. This concept was also well described in adults presenting heart failure with preserved left ventricular ejection fraction [27]. Indeed, in our cohort, early postoperative b-B significantly decreased the mean heart rate from 18 h after CPB. This negative chronotropic effect lasted until at least 48 h after CPB. This means that the clinical goal of increasing relaxation time was achieved, with an assumptive improvement of their diastolic function. Many clinical aspects (peripheral perfusion, perfusion pressure, postoperative echocardiogram) may play a role in selecting patients started on early postoperative b-B; we could not point those out. However, our retrospective analysis showed that patients with a higher preoperative RV/LV ratio were more likely to receive early postoperative b-B. A thicker RV predisposes to diastolic dysfunction, and we hypothesized that this may have prompted clinicians to use early postoperative b-B.

We described the various surgical strategies for ToF repair and their determinants in a previous publication [1]. The low proportion of patients undergoing a PVSR may be explained by the poor quality of the pulmonary valve after years of unrepaired ToF. The humanitarian nature of the surgical management may also prompt the surgeon to opt for a definitive procedure with less risk for reoperation. In our cohort, patients undergoing a repair with TP were more likely to receive early postoperative b-B. We showed in a previous descriptive study on the same cohort that patients undergoing a repair with TP had a higher RVOT gradient, smaller pulmonary valve annulus, and greater RV hypertrophy [1]. These significative differences in the preoperative echocardiographic assessment may prompt the clinician to use early postoperative beta-blockers. Due to diastolic dysfunction with high filling pressures, the expected degree of PR consecutive to a TP may not be relevant in the early postoperative course, so this surgical strategy, as such, does not prohibit the use of b-B during the initial postoperative period. Consequently, patients undergoing a repair with TP also presented a significantly lower prevalence of postoperative LCOS. Their longer CPB duration than other surgical strategies may partly explain the more frequent use of b-B in this patient group. Although we did not identify significant differences in preoperative markers of RVH, PS, and filling pressures, a more severe obstruction of the RVOT not allowing the surgeon to spare the pulmonary valve could also have been contributive.

This study highlights that early postoperative b-B after complete repair of ToF is associated with a lower prevalence of LCOS at the expense of more vasoactive support. However, the occurrence of LCOS was not associated with a lower vasoactive-inotropic score. We interpret the significant association between early postoperative b-B and lower LCOS prevalence as a sign of improved diastolic function in the treated cohort. Indeed, rather than waiting for signs of a good cardiac output before starting b-B, whose indication would be questionable at that time point, we hypothesize that (in this retrospective observational study) clinicians were tempted to start b-B early with the aim to enhance the diastolic function during the critical early postoperative period. The presented data supports the notion that early postoperative b-B is beneficial in minimizing or preventing LCOS in this setting. A greater need for vasoactive drugs following the administration of b-B is anticipated because of the inhibitory effect on sympathetic activation. The higher VIS scores than usually reported in similar contexts are explained by the previously described strategy followed at our unit using milrinone and noradrenaline instead of adrenaline in the postoperative management [28]. Despite a greater need for vasoactive drugs, patients on early postoperative b-B experience the same postoperative course as those who were not treated. There were no significant differences in the length of PICU, hospital stay, or ventilation durations. The older age of our cohort partially explains the relatively short ventilation durations. While comparing the PICU length of stay across several studies, logistical factors related to our internal institutional organization and the step-down unit’s capacity should be considered as relevant modulators of this outcome measure.

The retrospective nature of our study confers various limitations. First, early postoperative assessment was based on inotropic score, while clinical and echocardiographic data were unavailable. For instance, presence or absence of hepatomegaly and hepatojugular reflux, jugular vein assessment, echocardiographic measures of diastolic function, quantification of PR, and residual dynamic PS would have provided additional insights. Mean CVP measurements were high in every postoperative 6-h interval for all patients and may represent an indirect sign of diastolic dysfunction in the absence of objective assessment. Similarly, markers of end-organ impairment would have allowed a more comprehensive analysis. Second, over 13 years, many changes may have occurred (change of cardiac surgeons, evolution of preoperative assessment, surgical techniques, and postoperative management) [29]. For instance, sedation strategies evolved with the uptake of alpha-agonist agents (clonidine, dexmedetomidine) with a negative chronotropic effect confounding the effect of b-B. We were not able to adjust for these confounding factors. Moreover, it should be acknowledged that despite a careful and restrictive scoring system, the retrospective definition of LCOS, albeit frequently used in pediatric studies [22,23,24], and probably the best research tool currently available to measure its occurrence, still suffers from some imprecision. Finally, longer-term outcomes (like functional capacity, for instance [30]) were not assessed.

The results of this study must be interpreted with an understanding of the peculiarities of the included population, which preclude any generalizability to other pediatric populations. In particular, we would not recommend early postoperative b-B for neonates and infants undergoing a ToF repair, as they do not present the same degree of RVH and fibrosis as our cohort, and, most importantly, they are more exposed to negative inotropism.

To the best of our knowledge, this is the first descriptive study about the postoperative use of b-B in a population of children with late surgical repair of ToF. Furthermore, this cohort is at the same time also one of the biggest ever described with this condition [31]. This population’s representatives are numerous in developing countries. This study brings unprecedented data that informs some knowledge gaps surrounding the medical management of this unique population and may guide both clinical care and further research. We explored potential factors leading the caring physicians to prescribe postoperative b-B. However, further efforts should aim at identifying objective criteria to guide postoperative therapy and anticipate the patients who may benefit from this treatment strategy.