In this retrospective cohort study we aimed to evaluate the postoperative management of plasma sodium fluctuations in children who underwent resection of sellar or suprasellar tumors since centralization of care at the Princess Máxima Center in the Netherlands. A similar evaluation was performed in a national cohort before the care for pediatric oncology patients had been centralized [26]. A first finding of this study is a decrease in both the delta sodium as well as in the minimum and maximal value of sodium was seen in the years following 2018. Since centralization of care, the most severe outliers were observed in 2018, which is the first year after centralization. In addition, the change in plasma sodium concentrations per 24 h in children with early postoperative AVP-D is lower than was reported in a comparable pediatric cohort of patients treated for (supra)sellar tumors in the Netherlands before centralization of care [26]. Furthermore, the absolute lowest and highest plasma sodium concentrations deviate less from the normal range, when compared to this previous cohort of patients with postoperative AVP-D. These findings seem to suggest an improved trend in the management of plasma sodium balance over the years since centralization in this vulnerable patient group, probably attributable to the improved expertise of all members of the multidisciplinary team after centralization of care and the collaborative implementation of a clinical protocol. Following centralization of care, the variability in plasma sodium concentration steadily decreased over successive years. Importantly, no causal relationship between centralization of care and improved sodium management can be established directly, since this study only described the cohort post-centralization. Also, we no longer found an association between early postoperative AVP-D and the occurrence of short-term neurological events, in contrast to findings from the national cohort before centralization [26]. This may partly be explained by our smaller sample size, which limits the ability to detect such associations. It must also be emphasized that these are results from a heterogenous group of (supra)sellar tumors. However, improved management of plasma sodium levels in patients with early postoperative AVP-D may also have played a role, as smaller fluctuations in sodium concentration reduce the risk of short-term neurological sequelae. Prior to centralization, protocols for clinical management of postoperative AVP-D were center-specific, and a limited number of patients per center were treated annually, potentially contributing to variation in recognition, management and clinical outcomes. With the care for pediatric (supra)sellar tumors now largely centralized across the Netherlands, all team members, from PICU nurses and ward physicians to the on-call pediatric endocrinologist, have developed expertise in the management of these patients with postoperative AVP-D.

The observed incidence of early postoperative AVP-D in our current cohort was 68.5%, which is similar to the previously reported prevalence of 67.5%, prior to centralization of care [26]. This emphasizes that AVP-D remains a prevalent and clinically relevant complication of pediatric (supra)sellar tumor resection. Furthermore, all patients with a postoperative triphasic response subsequently developed permanent AVP-D, and a triphasic response was observed in more than half of the patients with permanent AVP-D.

For management of the fluid balance in children at risk for AVP-D post-surgery, in our center a protocol has been developed in collaboration with the PICU, the department of endocrinology and pediatric oncology, which already starts preoperatively, and then continues peri-operatively and in the postoperative period (Supplement 1). In this protocol, the pediatric endocrinologist is the coordinating physician throughout all phases of care; from outpatient clinic to inpatient clinic, operating theater, PICU and back to inpatient clinic and dismissal to outpatient clinic, hereby providing continuation of care. In each phase, a different attending physician coordinates the plasma sodium concentration and fluid balance which may differ from attending resident to anesthesiologist, pediatric intensivist or pediatric neuro-oncologist. As the patient frequently changes from department within this treatment phase, it is crucial that there is a continuous coordinating physician that oversees the different stages of AVP-D development and desmopressin requirements. In our expert center, next to collaborative protocol writing, clinical lessons are provided to instruct and teach the team, that includes attending nurses, specialized nurses, (young) residents and supervising staff, about the pitfalls in AVP-D management. Pitfalls that should be recognized are the presence of adipsia (impaired thirst perception), peri-operative hypotension necessitating extra (intravenous) fluids, co-medication influencing diuresis or plasma sodium concentration (e.g. steroids or vincristine) and the presence of hypertonic dehydration (in the latter high sodium concentrations and concentrated urine are present, whereas, in AVP-D, high sodium concentrations with diluted urine are present). Severe shifts in plasma sodium concentration or rapid correction of hypo- or hypernatremia may affect neural integrity and may lead to acute neurological complications, such as cerebral edema following rapid correction of hypernatremia or central pontine myelinolysis resulting from overly rapid correction of hyponatremia [29, 30]; none of these complications were observed in the current study sample. Before centralization of care, a maximum, individually observed, delta plasma sodium concentration during the first ten postoperative days of patients with postoperative AVP-D of 46 mmol/L/24 h, was observed [26]. Although the decreased maximum delta here reported in 2018 (30 mmol/L/24 h) is still too high, it is shown that further improvement of this maximum delta could be observed over the subsequent years to a max of 14 mmol/L/24 h in 2023. These changes reflect an improvement in the management of changes in fluid and sodium balance postoperatively with prevention of too rapid or too steep changes in plasma sodium.

Despite centralization of care and improved expertise of the team, management of post-operative AVP-D remains challenging as can for example be seen in the outliers of serum sodium concentrations in 2021. These outliers were caused by a single case of a very young child (age 2 months) at time of surgery. In this specific case, the AVP-D was probably caused by severe shifts in intracranial pressure, rather than pituitary (stalk) damage. After 5 weeks, the AVP-D normalized and desmopressin could be stopped. This illustrates that in children with very young age (difficult to interpret feelings of thirst) or in case of unexpected AVP-D (with intact pituitary stalk and function) increased awareness is necessary to prevent severe changes in plasma sodium concentrations.

Several limitations of our study should be acknowledged. The retrospective nature of this study limits the ability to study the correlation between early postoperative AVP-D and the occurrence of short- and long-term neurological effects in depth, and has therefore been limited to an explorative analysis for associations. Additionally, plasma sodium concentrations are outcomes of point-of-care testing, as continuous monitoring of plasma sodium is not feasible, and may not always capture all fluctuations. Significant outliers could therefore potentially be missed. However, plasma sodium levels were measured routinely and frequently in each patient, following the local protocol, and were measured more frequently in response to abnormal values and changes in treatment, to ensure that potential extremes are either captured or prevented and will thus be mostly included in our data. We explored whether there was an association in this cohort between tumor type or pituitary stalk sacrifice and severe sodium fluctuations, but could not find such an association. A limitation of this exploration was the fact that not in all surgical reports it was explicitly stated whether the pituitary stalk was sacrificed or not. As expected, all patients with an explicitly stated sacrificed pituitary stalk, developed permanent AVP-D. The lack of association between tumor type and severe sodium fluctuations might be explained by the fact that postoperative AVP-D management is similar, regardless of tumor pathology. Also, the lack of association between noted pituitary stalk sacrifice in the surgical report and the severity of sodium fluctuations might be explained by the fact that more careful sodium/fluid monitoring is conducted by the team in these patients, given that postoperative AVP-D is anticipated when the neurosurgeon reports pituitary stalk transection. Moreover, due to the rareness of pediatric (supra)sellar brain tumors and AVP-D, the cohort size remains limited and lacks power for further statistical analyses. Although findings of this cohort were compared to the previous findings of the national cohort in the Netherlands prior to centralization of care, we did not perform statistical analyses for comparison. We reasoned it would not have been appropriate to use statistical tests for comparisons of outcomes of two different cohorts, since plasma sodium measurements have not been performed following similar protocols nor in similar labs in the years before centralization compared to after centralization of care. Furthermore, in our study, we chose to exclude patients who were already diagnosed with AVP-D (and thus already on maintenance desmopressin treatment) prior to tumor resection, since these patients are expected to have more stable plasma sodium levels postoperatively, as also reported by Kruis et al. [26]. Interestingly, even when only evaluating the postoperative course of plasma sodium levels of patients who are at highest risk of a complicated course (patients with newly diagnosed AVP-D), we see less severe plasma sodium fluctuations in the first 10 postoperative days. We believe this comparison might be considered a strength of this study, since this method allowed for a meaningful evaluation of care and provided clinically relevant insights into the effects of system-level changes in delivery of care for rare diseases.

The majority of patients with a postoperative triphasic response were diagnosed with at least hypothyroidism and ACTH deficiency at three months postoperatively. The presence of a triphasic response appears to be a valuable clinical indicator for the development of long-term endocrine dysfunction, although it is now based on descriptive data. In our experience, patients exhibiting this response are considerably more likely to experience multiple hormone deficiencies during follow-up, including permanent AVP-D.

In conclusion, findings of this study indicate that centralization of care for pediatric suprasellar brain tumors requiring neurosurgery may be associated with improved early postoperative management of sodium balance. This study enhances the awareness of the risks inherent to (supra)sellar tumor surgery and underlines the importance of centralization of care for rare diseases.