Practical Algorithm for the Management of Multisutural Craniosynostosis with Associated Chiari Malformation and/or Hydrocephalus

Introduction: The association between multisutural craniosynostosis with Chiari malformation (CM), venous hypertension, and hydrocephalus is widely described in the literature, especially in children with paediatric craniofacial syndromes. Some efforts have been done in the last years to understand the complex pathogenetic mechanisms underlying this association, and several theories have been proposed. In particular, it is now accepted that the hypothesis of the overcrowding of the posterior fossa due to precocious suture fusion is the cause of the cerebellar herniation in syndromic and non-syndromic patients, against the theory of intrinsic cerebellar anomalies, ventriculomegaly, and venous hypertension. However, whatever the pathophysiological mechanism, it is still unclear what the best management and treatment of CM and hydrocephalus are in multisutural craniosynostosis patients. The aim of this study was to report our 25 years’ experience in treating paediatric patients affected by these rare pathologies in order to propose a simple and effective therapeutic flow chart for their management. Materials and Methods: We retrospectively collected data of each patient who underwent a cranial vault remodelling (CVR) for complex multisutural craniosynostosis in our institution in the last 25 years, while monosutural craniosynostosis was excluded. We recorded data concerning type of craniosynostosis and craniofacial syndromes, presence of ventriculomegaly, and CM at presentation and clinical and radiological follow-up. Therefore, we evaluated the final outcomes (improved, stable, deteriorated) of these patients and created a practical flow chart that could help physicians choose the best surgical treatment when different pathological conditions, as Chiari malformation I (CMI) or hydrocephalus, affect complex craniosynostosis children. Results: Thirty-nine patients (39 out of 55; 70.9%), with an isolated multisutural craniosynostosis at presentation, underwent a two-step CVR as first surgery; 36 patients (92.3%) had an improved outcome, 2 patients (5.1%) had a stable outcome, and 1 patient (2.56%) had a deteriorated outcome. Other eight children (8 out of 55; 14.5%) had a radiological evidence of asymptomatic CMI at presentation. In this group, we performed CVR as first surgery. As for the final outcome, 7 patients had an improved outcome (87.5%) with good aesthetic result and stability or resolution of CMI. Finally, 7 patients (7 out of 55; 12.7%) presented a various combination of CMI and ventriculomegaly or hydrocephalus at presentation. Among them, 3 patients had an improved outcome (42.8%), and 4 patients had a deteriorated outcome (57.1%). Discussion: The prevalence of one pathological condition with associated symptoms over the others was the key factor leading our therapeutic strategy. When craniosynostosis is associated with a radiological CM, the assessment of clinical symptoms is of capital importance. When asymptomatic or pauci-symptomatic, we suggest a CVR as first step, for its efficacy in reducing tonsillar herniation and solving CM symptoms. When craniosynostosis is associated with ventricular enlargement, the presence of intracranial hypertension signs and symptoms forces physicians to first treat hydrocephalus with a ventriculo-peritoneal shunt or endoscopic third ventriculostomy. For patients with various degrees and severity of ventriculomegaly and associated CM, the outcomes were very heterogeneous, even when the same therapeutic strategy was applied to patients with similar starting conditions and symptoms. This is maybe the most unexpected and least clear part of our results. Despite the proposed algorithm comes from a clinical experience on 85% successfully treated patients with multiple craniosynostosis, more extensive and deep studies are needed to better understand CM and hydrocephalus development in such conditions.

hydrocephalus are in multisutural craniosynostosis patients. The aim of this study was to report our 25 years' experience in treating paediatric patients affected by these rare pathologies in order to propose a simple and effective therapeutic flow chart for their management. Materials and Methods: We retrospectively collected data of each patient who underwent a cranial vault remodelling (CVR) for complex multisutural craniosynostosis in our institution in the last 25 years, while monosutural craniosynostosis was excluded. We recorded data concerning type of craniosynostosis and craniofacial syndromes, presence of ventriculomegaly, and CM at presentation and clinical and radiological follow-up. Therefore, we evaluated the final outcomes (improved, stable, deteriorated) of these patients and created a practical flow chart that could help physicians choose the best surgical treatment when different pathological conditions, as Chiari malformation I (CMI) or hydrocephalus, affect complex craniosynostosis children. Results: Thirty-nine patients (39 out of 55; 70.9%), with an isolated multisutural craniosynostosis at presentation, underwent a two-step CVR as first surgery; 36 patients (92.3%) had an improved outcome, 2 patients (5.1%) had a stable

Introduction
Multisutural craniosynostosis (CS) is defined as any CS that involves more than one of the sutures of the skull. The diagnosis of CS is made by both clinical and radiological evaluation, mainly with head 3D computed tomography (CT) scans [1,2]. There are several genetic syndromes that can be associated with CS and, within them, we can distinguish between the ones involving multiple sutures, such as Apert, Pfeiffer, Crouzon, and Antley-Bixler syndromes, and the ones associated with isolated coronal synostosis, such as Saethre-Chotzen syndrome and Muenke syndrome [3,4].
Apert syndrome is an autosomal-dominant inherited disorder that occurs in 1:100,000 newborns [3]. It is most commonly caused by specific de novo missense mutations in FGFR2-gene (Ser252Trp and Pro253Arg) [5,6]. Generally, it is characterized by the premature fusion of both coronal sutures, resulting in a brachycephalic skull shape, associated with hypertelorism and midface hypoplasia.
Crouzon-Pfeiffer syndrome is an autosomal dominant inherited disorder with an incidence of 1:25,000 [3,5,6]; it is mostly caused by a mutation in FGFR2-gene and, in rare cases, by a mutation in FGFR1 as well as a mutation in FGFR3 with acanthosis nigricans [7]. Most patients have bilateral coronal suture synostosis, but different skull sutures might be involved, sometimes resulting in pansynostosis. Within CS syndromes, patients with Crouzon-Pfeiffer syndrome have the highest incidence of Chiari malformation I (CMI) (up to 72.7%) [8].
CMI is defined by the radiological finding of more than 5 mm cerebellar tonsillar herniation (TH) through the foramen magnum (FM). The so-called acquired CM is a pathological cerebellar TH caused by several conditions that determine reduction in intracranial volume, especially in the posterior fossa (PF), without any primary abnormality of the brain [9][10][11][12].
A stable enlargement of the ventricular system is defined as ventriculomegaly, while a progressively increasing size of the ventricular system, associated with signs and symptoms of intracranial hypertension, is defined as hydrocephalus. On the basis of symptom intensity, we can further distinguish a mild and severe hydrocephalus [13]. The first association between CMI and CS was made by Saldino et al. [14] in 1972, while more recent studies have demonstrated the link between TH and ventriculomegaly [15]. By then, different hypotheses have been advocated to explain these associations.
The simplest one suggests that an obstruction of cerebrospinal fluid (CSF) at the level of the 4th ventricle may be the cause of ventricular enlargement and the subsequent increased intracranial pressure. This condition may lead to TH and CMI [16]. The evidence that not all the patients with CS and CMI have ventricular enlargement and vice versa, suggests that a more complex pathomechanism is underlying [8].
The premature fusion of cranial vault sutures in syndromic CS might be the cause of reduction of intracranial volume (cephalocranial disproportion), leading to an overcrowded PF and hence to a cerebellar TH [10]. This theory is also supported by the evidence that the TH in CS patient is not present at birth but develops together with the premature closure of cranial sutures in the first 6 months of life [8].
On the other hand, a cephalocranial disproportion and an overcrowded PF could be the causes of a CSF flow obstruction at the level of the FM and finally the cause of ventriculomegaly or, in the worst case, of obstructive hydrocephalus [15]. The last theory, however, is discredited by the evidence that just a PF decompression often fails to restore normal CSF circulation and to reduce hydrocephalus [17,18].
As for the role of venous outflow obstruction and intracranial venous hypertension in the development of intracranial pressure in these children, Rich et al. [19] speculated that a big role is played by jugular foramen narrowing, that can commonly be seen in syndromic CS. Venous outflow obstruction has also been linked to hydrocephalus development in patients with syndromic CS [20][21][22].Whereas there is a large consensus in literature in treating patients with symptomatic CMI and CS, less clear are the indications for type, order, and timing of surgical intervention. For these reasons, we designed the present study and proposed practical therapeutic flow chart, that could help physicians choose the best surgical treatment when different pathological conditions, with various degrees of severity, affect complex CS children.

Materials and Methods
A retrospective study was conducted with patients who had undergone a two-step cranial vault remodelling (CVR) procedures at G. Gaslini Children's Hospital in the last 25 years. We identified 55 patients with multisutural CS, both syndromic and nonsyndromic.
Syndromic CS was defined as CS in the setting of a genetic diagnosis of Apert, Antley-Bixler, Crouzon, Muenke, Pfeiffer, Noonan, Baller-Gerold, and Saethre-Chotzen syndrome. For each patient, we collected data about age, sex, sutures involved, timing of surgical procedures, and clinical symptoms. Through magnetic resonance imaging (MRI) or CT scan analysis, we evaluated the presence and the extent (in millimetre) of pre-operative cerebellar herniation and the presence of ventricular enlargement. Ventricular enlargement was established by an Evan's index >0.3, calculated on axial CT scans, and CM was defined as evidence in CT and MRI brain scans of TH under the FM greater than 5 mm.
Eventually, we evaluated clinical, cosmetic, and radiological outcomes with a post-operative medical examination and MRI brain scan within 1 year after surgery. After that, we classified children in three categories (deteriorated, stable, and improved) based on the final outcome.

Definition of Final Outcome
We defined as Improved, patients with complex CS (with or without associated hydrocephalus or CMI) who underwent surgical treatment and had 1-year follow-up: 1. Post-operative good aesthetic outcome (assessed on patients'/ parents' judgement of the aesthetic result); 2. Radiological and clinical evidence of CM/hydrocephalus resolution or improvement, and therefore not requiring further surgery. We defined as Stable, patients with complex CS (with or without associated hydrocephalus or CMI) who underwent surgical treatment with stability of pre-operative symptoms at postoperative 1-year follow-up.
We defined as Deteriorated, patients with complex CS (with or without associated hydrocephalus or CMI) who underwent surgical treatment and had at 1 year follow-up: 1. Post-operative poor aesthetic outcome (assessed on patients'/ parents' judgement of the aesthetic result) 2. Radiological and clinical evidence of CM/hydrocephalus of new onset or worsening of the previous conditions, therefore requiring an adjunctive surgery.

Surgical Procedures
A perioperative preparation with high-dose tranexamic acid has been performed in every patient [23]. A bicoronal incision executed in prone position is useful for both the posterior and the anterior expansion. In our institution, as an alternative to sutural stripping and distractors, we perform a CVR by creating an interosseous gap between a free-floating parieto-occipital bone flap and the skull [24]. This technique, in combination with a linear dura mater cauterization, can guarantee a slow and controlled ossifications of the gap and therefore a remodelling and expansion of cranial vault [25]. This technique can be adopted as a single procedure or prior to a fronto-orbital advancement. In patients in whom we also performed a fronto-orbital advancement, the second step was planned as a separate second surgery or, more recently, in the same session with the first step.
With the so-called floating forehead technique the orbital roof (bandeau) is lifted, reshaped, and advanced by 2.5 cm bilaterally [26]. Frontal bony opercula are fixed to supraorbital bar after being rotated and reshaped through linear osteotomies, in order to expand the anterior cranial vault volume and to create a symmetric forehead shape. The swelling of the temporal fossa is corrected by lifting a small bone flap and placing it back in position. Absorbable sutures reduced average hospitalization times to 4-6 days. Postoperative therapy with morphine for pain control (primary management with via i.v. PCA for 48 h and subsequent switch to non-opioid oral therapy), combined with albumin and diuretics to reduce oedema, has been utilized.

Results
Results of our study are summarized in Table 1. In our series of 55 children with CS, 25 were females (45.4%) and 30 were males (54.5%). Between all of them, seventeen (30.9%) patients had isolated CS; in 4 children (7.2%), we found aspecific genetic mutations without a defined syndromic diagnosis, while the others (61.8%) had a genetic diagnosis of Apert, Antley-Bixler, Crouzon, Muenke, Pfeiffer, Noonan, Baller-Gerold, and Saethre-Chotzen syndrome. Among 55 patients with CS, 47 patients had an improved outcome (85.4%), 2 patients   Other eight children (8 out of 55; 14.5%), besides complex CS, had a radiological evidence of asymptomatic CMI at presentation. In this group, we performed CVR as first surgery. As for the final outcome, 7 patients had an improved outcome (7 out of 8, 87.5%) with good aesthetic result and stability (2 out of 7) or resolution (5 out of 7) of CMI. One patient (number 54) had a deteriorated outcome (1 out of 7, 12.5%), requiring additional surgery of PF decompression after CVR; subsequently, he had a resolution of CMI and a good aesthetic result at 1-year follow-up.
On the other hand, just 1 patient (1 out of 55; number 13 of Table 1; 1.8%) had severe hydrocephalus at presentation, besides complex CS. This child underwent ventriculo-peritoneal shunt (VPS) as the first surgery and, after hydrocephalus resolution, underwent a second surgery of CVR. At 1-year follow-up, the patient presented an improved outcome. Finally, 7 patients (7 out of 55; 12.7%) presented a various combination of CMI and ventriculomegaly or hydrocephalus at presentation, besides complex CS (number 12, number 15, number 16, number 17, number 29, number 40, and number 46 of Table 1). Among them, 3 patients had an improved outcome (3 out of 7; 42.8%) and 4 patients had a deteriorated outcome (4 out of 7; 57.1%).
In this heterogeneous group, 2 patients (number 29 and number 40), who presented both mild ventriculomegaly and asymptomatic CMI, were treated performing CVR, with a good aesthetic outcome. Two other patients (number 16 and number 46 of Table 1), who presented the same pathological condition (namely, both mild ventriculomegaly and asymptomatic CMI), In this study, 4 patients had wound dehiscence repaired for cosmetic reasons, of which 2 had bony decubitus and were revised surgically with drilling of the protrusion. Five patients had intraoperative dural tears repaired during the procedure itself without further complications. Two patients had central venous catheter sepsis. Nine patients had periorbital lacerations during orbital bandeau, and 1 patient suffered from hemorrhagic shock from disseminated intravascular coagulation. Only 1 patient had immediate post-op epidural haemorrhage after posterior cranial fossa decompression. Finally, one patient had laceration of venous lakes pathologic circles during duroplasty. In VPS cases, one had malfunctioning catheter after 1 year and one valve obstruction after 3 follow-up years.

Discussion
The management of patients with CS and associated hydrocephalus or CMI has always been challenging. Even in patients without evidence of ventricular dilatation or TH at the presentation, there is a possibility of a late development of these conditions, especially after a CVR. While in children with frank symptoms of hydrocephalus or CMI, the choice for first surgery is easy to make, the bet can be more challenging in children with mild symptoms or when radiological evidence is not consistent with clinical manifestations.
Current recommendations for management of patients with co-existing CS and acquired CMI (with or without associated symptoms) are to perform either, through the modified prone position, a simultaneous CVR and PF decompression or a two-staged procedure [9,17]. In the first case, it is important to mention that the modified prone position may cause significant pressure at the craniovertebral junction due to the cervical hyperextension [17,27,28].
Furthermore, the increased surgical duration leads to additional perioperative anaesthesia complications risk  Algorithm for Multisutural Craniosynostosis with Associated Chiari [29]. On the other hand, the alternative two-staged procedures entail a longer hospital stay and additional surgical and anesthesiological risks [30].
In the last few years, a different therapeutic strategy has been reported as safe and effective in the resolution of CMI in patients with syndromic CS, especially the asymptomatic ones. In the 2012, Levitt et al. [31] reported the case of a 3-month-old boy, with phenotypic characteristics consistent with Crouzon syndrome, premature closure of the left coronal, sagittal, and bilateral lambdoid sutures, who underwent circumferential posterior vault craniectomy, with complete radiological resolution of pre-operative CMI.
Whereas there is a large consensus in literature in treating patients with symptomatic CM and CS, less clear are the indications for type, order, and timing of surgical intervention. Over the years, different groups have advocated the efficacy of simple CVR in reducing TH and solving CM symptoms [32,33], without a posterior cranial fossa decompression. On the other hand, some authors recommend a posterior cranial fossa expansion surgery before CVR [9,34,35].
The evidence that posterior cranial fossa expansion in patients with concomitant CM and ventricular enlargement it is not always sufficient for hydrocephalus resolution [17,20] suggests that maybe both venous hypertension and crowded posterior cranial fossa are responsible for the ventricular enlargement seen in complex CS children. When clinical and radiological evidences are unmistakable, diagnosis and treatment of hydrocephalus are easy and must be performed before every other surgical procedure [20,21].
On the other hand, clinical experience suggests that in patients with CS and ventricular enlargement, a CVR can be a solution for both conditions, leaving the patient shunt-independent in most cases [36,37]. Another important thing that has to be considered, however, is the evidence that a slowly progressive or latent hydrocephalus, caused by the limitation of ventricular expansion forces in the rigid sinostotic skull, can be developed in children after CVR and the subsequent cranial expansion [20]. Hence, the right therapeutic choice in small patients with CS and concomitant mild ventricular enlargement is never easy to make.
Our knowledge of CS has greatly expanded over the years and, consequently, has also changed its therapeutic and surgical management. Over the past two decades, the management of this pathology in our institution has been more and more pointed towards the resolution of both CS and potential associated conditions as a first goal. Consequently, it was usually performed a double step CVR, as unique surgical treatment, without a PF decompression, VPS placement, or ETV.
This minimizes surgical and anaesthesiologic risks, hospital stay, and overall post-operative complications. Of course, this solution is not applicable indiscriminately to all patients, and every therapy has to be tailored on each child. As demonstrated in our series, indeed, this strategy is possible only if associated conditions, such as hydrocephalus or CMI, are mild and slowly progressive; a VPS or ETV or a PF decompression should be reserved for those cases with rapid and serious intracranial hypertension.
We created a simple and effective therapeutic strategy flow chart (Fig. 1) that can help physicians in daily practice and decision-making in children with CS associated with CMI or hydrocephalus. The prevalence of one pathological condition with associated symptoms over the others was the key factor leading our therapeutic strategy. Ahead a child with a complex CS, the first important step is to evaluate the presence of associated pathological conditions by performing, when feasible, an MRI brain scan or, at least, a CT brain scan. Thereafter, an extensive clinical examination is required in order to understand if one of associated conditions is symptomatic and what kind of symptoms (hydrocephalus or CM) are predominant over the others.
In our series, just 1 out of 39 children with an isolated CS at presentation who underwent CVR developed a post-operative symptomatic CMI, deserving a posterior cranial fossa decompression (patient number. 19), and reporting a final deteriorated outcome. Even so, we are aware that the late post-CVR, development of one of the associated conditions is unpredictable and always possible; therefore, we recommend a strict clinical and radiological follow-up after surgery for these patients. It is interesting to notice that, despite the percentage reported in the literature [36], none of our patients with isolated CS who underwent CVR developed postoperative hydrocephalus.
When CS is associated with a radiological CM, the assessment of clinical symptoms is of capital importance. When asymptomatic or pauci-symptomatic, we suggest a CVR as first step, in line with the theory of CVR efficacy in reducing TH and solving CM symptoms [32,33]. Seven out of 8 patients who underwent a CVR as first surgery had a good final outcome with resolution or stability of TH, as demonstrated at 1-year post-operative follow-up.
Even if in our series there is no children with a symptomatic CM and complex CS at presentation, in our opinion a posterior cranial fossa decompression, followed by CVR, is the treatment of choice. In our experience, when CS is associated with ventricular enlargement, the presence of intracranial hypertension signs and symptoms forces physicians to first treat hydrocephalus with a VPS or ETV. In our series, only one child presented with severe hydrocephalus in addition to complex CS. His outcome improved at 1-year follow-up after VPS and subsequent CVR.
In our series, there are no patients with an isolated mild ventriculomegaly at presentation, but in our opinion, a CVR as first surgery is a good choice, in order to reduce the crowding of posterior cranial fossa and the obstruction of CFR outflow at 4th ventricle [17,20,38]. Interestingly enough, however, in our series, 7 out of 8 patients with various degree and severity of ventriculomegaly had also an associated CM. In these cases, outcomes were very heterogeneous, even when the same therapeutic strategy was applied to patients with similar starting conditions and symptoms. This is maybe the most unexpected and least clear part of our results.
In fact, in the four children with both mild ventriculomegaly and asymptomatic CMI at presentation, CVR was performed as the first surgery without distinctions. In two of them (patient number 29 and patients number 40), we registered a total radiological resolution of associated conditions, with a good final outcome. For the other 2 patients, we found that one (patient number 16) required an ETV for a post-operative development of hydrocephalus symptoms, while the other (patient number 46) required a PF decompression because of post-operative development of symptomatic CMI. It is difficult to explain the reason behind the opposite course of children with the same clinical and radiological presentation. This evidence confirms that different pathological mechanisms underlie CM and hydrocephalus in complex CS patients, and therefore, the same therapeutic strategy cannot be effective on everyone. Despite the proposed algorithm coming from a clinical experience on 85% successfully treated patients with multiple CS in the last 25 years, more extensive and deep studies are needed to better understand CM and hydrocephalus development in such conditions.

Limitations
Our study has several limitations. In the first place, it is a single-centre retrospective study, and although the results seem encouraging, it carries the limitation of the relatively low number of cases and of being nonrandomized. The final outcomes have been partially analysed on the basis of patients' or parents' judgements of the aesthetic result, that cannot be easily objectified. At last, the final percentage of success (85%) demonstrates that there is still a lot to understand about patients with complex CS and that the perfect therapeutic strategy is still to be defined.

Conclusions
Different and complex pathological mechanisms underlie the development of CM and hydrocephalus in complex CS patients, and a good part of them is still unknown. This postulate leads to the conclusion that a single and rigid therapeutic strategy cannot be effective on all patients. A more flexible approach is needed, with a careful planning of every single therapeutic step tailored on patient-specific conditions. Despite the proposed algorithm coming from a clinical experience on up to 85% successfully treated patients with multiple CS in the last 25 years, more extensive and deep studies are needed to better understand CM and hydrocephalus development in such conditions.

Statement of Ethics
Ethics approval was not required since this was a retrospective study and was not required due to local/national guidelines. Patients routinely give written consent to the processing of their clinical and radiological data anonymously for scientific purposes upon hospital admission. The study was conducted according to the Declaration of Helsinki for medical research involving human subjects.
Aruta, Marco Pavanello, Marco Ceraudo, and Andrea Bianconi have participated to drafting the manuscript. Marco Pavanello, Francesca Secci, and Gianluca Piatelli revised it critically. Pietro Fiaschi, Marco Pavanello, and Gianluca Piatelli supervised the project. All authors read and approved the final version of the manuscript.

Data Availability Statement
All data generated or analysed during this study are included in this article. Further enquiries can be directed to the corresponding author.