Introduction

Differently from simple craniosynostoses, the management of which is well established and the results predictable, the management of complex synostoses has to be tailored on the single subjects, as the complex interaction of bone anomalies of the skull and face, distortion and compression of the intracranial cerebrospinal fluid (CSF) spaces, vascular and nervous parenchymal structures, and venous outflow of the space may require urgent treatment at any phase of the postnatal life [9, 12, 13, 15, 55, 66]. Possible ocular and respiratory severe associated functional and anatomical disorders may also dictate the time of the surgical correction [33, 34, 39, 66]. Consequently, the neurosurgeon and the cranio-maxillofacial surgeon are required to have an understanding of the physiopathogenetic mechanisms concurring to the clinical expression of these particularly difficult to deal pathological conditions [34, 37, 68]. The present paper will focus on the influence of the CSF dynamics anomalies, impaired venous circulation, caudal descent of the cerebellar tonsils, and upper airways volumetric alterations on this decision-making process.

The multifactorial etiology of abnormal intracranial pressure in complex craniosynostoses and its management according to the different pathogenetic factors

Abnormal intracranial pressure: general features

Abnormal intracranial pressure values have been documented in 47–67 % of children with complex craniosynostoses [9, 12, 33, 54] The relationship between complex craniosynostoses and increased intracranial pressure (ICP) is multifactorial and actually incompletely understood. Historically, the number of fused sutures has been related to ICP [54]; however, since early 1990s, many authors have highlighted the lack of a predictable correlation between intracranial volume and intracranial pressure, being the intracranial volume normal in most children with complex craniosynostoses or even enlarged [18]. In a recent study, De Jong et al. compared the MR intracranial volumes of 103 complex craniosynostoses children at different ages with normal values in age-matched normal children taken from the literature. No statistically significant difference was found among the two groups in the intracranial volumes, whereas the total CSF volume was significantly higher in children with craniosynostoses, suggesting that this last might be the main driving force for the compensatory skull growth [18]. Impaired venous outflow from hypoplastic cranial foramina and impaired respiration due to associated midface hypoplasia and airway compromise have been propounded as potential important contributors in syndromic cases to elevated intracranial pressure interacting in a vicious cycle also with the alterations of CSF dynamics [34, 55].

The detection of elevated intracranial pressure cannot be easy [52, 65]. Noninvasive signs of elevated intracranial pressure, such as papilledema on funduscopic examination, increased optic nerve sheath diameters measurements at ecography, and skull markings (copper beaten appearance) on plain skull films, may lack of sufficient sensitivity and specificity in children with craniosynostoses. Tuite et al. [69] demonstrated that the specificity of papilledema as an indicator of elevated ICP was 87 %; however, its sensitivity was only 30 %. The significance of papilledema was even less reliable in the youngest patients. The association of optic nerve sheath diameter measurements has a little contributed to improve the diagnostic value of fundoscopy. Optic nerve sheath measurements have not improved this detection rate. Driessen et al recently addressed optic nerve sheath measurements in a series of 128 children with syndromic craniosynostosis. Overall, compared to fundoscopy, optic nerve sheath measurement provided a sensitivity of only 11 %, though the specificity was as high as 97 %. Positive and negative predictive values were 46 and 80 % [20]. X-ray digital markings, as an isolated finding on plain radiographs, have similarly been shown not to be a sensitive indicator of elevated intracranial pressure. In a classic paper, Davidoff demonstrated a 15–20 % incidence of digital markings in normal 5–9-year-old children. In children with complex craniosynostosis, only the association among digital markings, erosion of the sella, or suture diastasis resulted in sensitivity for elevated ICP as high as 90 % in all age groups [17]. Planning a prolonged monitoring of the ICP in children with complex craniosynostoses is also burdened by many difficulties for the poor cooperation of the youngest patients and for the lack of universally accepted scales of normal and abnormal ICP values in children [65]. Most authors suggest to consider children with craniosynostoses as comparable to children above 1 year/with closed sutures [21, 29, 53, 61] and to evaluate the ICP in them according to adult parameters (normal range, below 10 mmHg; borderline, 10–15 mmHg; abnormal, above 15 mmHg). However, this scheme is only partially transferable to children with complex craniosynostoses due to their reduced cerebral compliance and the mostly chronic, periodic alteration of the intracranial pressure, which render absolute ICP values unreliable [21, 29, 53].

The alternative methods that have been proposed take into account the occurrence of Lundberg A waves (rises in ICP above 50 mmHg lasting 5 min) and Lundberg B waves (rises in ICP up to 50 mmHg of 0.5–2-min duration), even though again, such a parameter cannot be applied to borderline patients, such is the frequent case of children with complex craniosynostoses [65]. In order to try to overcome this problem Eide et al., proposed to consider the frequency of ICP elevations of 20 or 25 mmHg lasting 0.5 to 1 min, reporting a significant reduction of false negative results [21–23].

Alterations of the CSF dynamics

The presence of a ventricular dilation is a common finding in children with syndromic craniosynostoses [15, 16]; most series report the presence of ventriculomegaly in 30–70 % of patients with Crouzon’s and Pfeiffer syndrome and in 40–90 % of children with Apert syndrome [12, 15, 16, 19]. The extreme variety of these figures is related to the fact that many series join in their report patients with progressive hydrocephalus and nonprogressive ventriculomegaly, in some case including patients with simple ventricular distortion, which should be considered a normal anatomic pattern in children with syndromic craniosynostosis [12, 15, 54]. When only active forms of hydrocephalus are concerned, a rate of 12.1–15 % has been reported [15]. The exact nature of the hydrocephalus associated with complex craniosynostoses is not yet completely understood. Several pathogenetic mechanisms can play a role, apparently differently interacting in different syndromic grounds. In Crouzon’s syndrome, the progressive fusion of cranial base synchondroses produces alterations in the skull base with consequent stenosis of the jugular foramina [34, 55, 66]. The resulting venous hypertension leads to increased CSF hydrostatic pressure. The contemporary premature fusion of the cranial vault sutures contribute to this mechanism, as for example for the closure of the lambdoid sutures, which leads to a crowded posterior fossa compression of the sigmoid sinuses and secondary hindbrain herniation [11, 27, 34]. According to Cinalli et al., two groups of Crouzon syndrome babies can be recognized when CSF dynamics alterations are considered: the first, when simultaneous closure of sagittal and coronal sutures occur, that usually present with a pseudo-tumor-like state, having normal or small size ventricles but clinical signs of increased intracranial pressure; in this group of patients, ventricle size might increase after cranial vault remodelling, requiring, in a selected proportion of cases, a direct surgical treatment of the progressive hydrocephalus. The second group of patients in these authors’ experience were those with sequential closure of the sagittal and coronal sutures; in these patients, a variable degree of ventricular dilation was present at diagnosis depending from the residual skull vault compliance [12]. When a cloverleaf skull deformity is present (mostly in children with Pfeiffer syndrome), the perinatal fusion of multiple sutures both of the skull base and of the cranial vault exacerbate all the factors potentially responsible for a disturbance of the CSF dynamics. Venous outflow impairment, due to the jugular foramina stenosis, hindbrain caudal displacement, and constriction of the posterior fossa structures within a hypoplasic posterior cranial fossa all would contribute to obstruct the major CSF pathways [13, 33, 34]. It is therefore not surprising the high incidence in this group of active hydrocephalus early in the course of the disease (Fig. 1). Crowding of the posterior fossa is, in most cases, an acquired disorder as demonstrated by experimental findings [41], estimates of the posterior fossa volume [59, 68], and well-documented cases of postnatally developing upwards and downwards tonsillar herniation [37, 57, 67]. The concept of mechanical CSF outflow obstruction as primary due to posterior fossa crowding is, however, challenged by the cases of hindbrain herniation in many patients with craniosynostosis without an associated hydrocephalus [12, 66]. Consequently, a major role of venous pathways obstruction was proposed by Hoffman and Hendrick [36] already in the late 1970s. Cinalli et al. [12] subsequently demonstrated a stenosis of jugular foramina early in life in a significant proportion of patients affected by Crouzon syndrome, only few patients with active hydrocephalus in this experience presenting with normal venous outflow early in life. The finding that ventricular dilation occurred mostly in babies with still partially opened sutures complies with the hypothesis of venous hypertension as a primary event; as in a totally rigid skull, the latter should rather induce a pseudo-tumor-like state [34]. In spite of these convincing data, some doubt still exists on the primary role of venous hypertension in the development of active hydrocephalus in this subset of patients. Indeed, a not negligible proportion of infants with Crouzon syndrome, actually present with a significant increase in intracranial pressure but normal-sized ventricles [12]. On these grounds, most authors currently favor an immediate sequence, or even coexistence of the two mechanisms, by assuming that venous hypertension causes a CSF absorption deficit as well as brain swelling leading to the tonsillar herniation [27], whose further progression is favored by the progressive closure of the cranial vault and skull base sutures.

Fig. 1
figure 1

Pathogenetic factors in the development of hydrocephalus in complex craniosynostoses: 4-months-old Crouzon child. a 3D-CT reconstruction showing the almost complete closure of the two lambdoid sutures and of the sagittal suture. b 3D-CT reconstruction of the cranial base showing a restriction of the jugular foramina, almost completely occluded on the left side. c, d Angio-MR of the same patient showing a stenosis of the right transverse and sigmoid sinuses with the lack of visualization of the left jugular vein. The presence of multiple transcranial compensating venous channels is also documented. e, f Axial CT images showing the development of an active hydrocephalus, with enlargement of the temporal horns of the lateral ventricles and of the third ventricle. g Sagittal MR T2-weighted view confirming the presence of an active hydrocephalus associated with a Chiari I malformation

Full size image

Children with Apert syndrome significantly differ from other syndromic forms in terms of alteration of the CSF dynamics. Although often the coronal sutures close early in the first phases of the disease, both sagittal and lambdoid sutures as well as cranial base ones usually close late along the disease course, resulting in a less severe jugular foramina stenosis and venous hypertension; this is also the reason why, though ventriculomegaly is a common finding in children with Apert syndrome, active hydrocephalus occurs infrequently or late [12, 34] (Fig. 2).

Fig. 2
figure 2

Apert syndrome: the factors that favor the lack of development of an active form of ventricular dilation. a 3D-CT reconstraction showing the patency of both the lambdoid sutures. b 3D-CT reconstruction of the skull base showing normal size of both jugular foramina. c Oblique angio-MR reconstruction showing normal course of both jugular veins in the absence of compensating venous circles. d Axial MR image documenting the absence of an active ventricular dilation. e, f Prolonged ICP monitoring showing normal ICP values and waves both in a representative sample trace and the bar graphic of the whole recording (48 h)

Full size image

Management strategies for CSF dynamics disorders

The relatively complex and mutifactorial etiology of CSF dynamics disorders in children with complex craniosynostoses has lead to a significant and still ongoing discussion on what would be the best treatment when an active hydrocephalus develops in these children. A prolonged intracranial pressure monitoring is advised whenever clinical or radiological signs of altered CSF and venous circulation are suggested by MR/MR angiographic examinations [65]. The role of posterior fossa decompression, aimed to address the two main postulated causes of increased intracranial pressure, namely, the posterior fossa spaces constriction and the main venous outflow compression, has been questioned [11–13, 34].

The control of the hydrocephalus has been reported as not consistent in largest series [34]. Extrathecal ventricular shunts have been long considered as the only feasible way to deal with active hydrocephalus in these children. However, there are many concerns related to the use of extrathecal shunts in this subset of patients. Apart from the extrathecal shunt morbidity, common to the other hydrocephalic conditions, extrathecal shunts have two other major concerns in children with complex craniostenosis. The first is that extrathecal shunts do not address the venous hypertension, with the possibility of continuing intracranial hypertension, in spite of an apparently working shunt. The second concept is that the drained CSF spaces will be early filled by the growing brain, while at the same time the skull will be withdrawn of an important growing stimulus; consequently, the spontaneous tendency of the cranial vault and base sutures to close will be even favored [12, 15, 34]. Furthermore, in case of cranial reconstruction, the procedure should be planned at least with a 2-month time distance from the shunting procedure, in order to avoid the possible bone retrusion due to the reduction in intracranial pressure, induced by the shunt [15]. Endoscopic third ventriculostomy has been proposed as a further alternative for the management of the hydrocephalus in a selected proportion of children with complex craniosynostosis. This choice should be taken into consideration in case of restriction/distortion of the Sylvian aqueduct and/or when a caudal dislocation of the cerebellar tonsils, possibly associated to constriction of posterior fossa intracranial structure is documented [19]. In a recent paper, Di Rocco et al. reviewed the results of their experience in performing ETV in a series of 11 children with complex craniosynostosis, selected on the above-mentioned criteria. ETV was performed after cranial expansion in four cases and as first line treatment in seven cases. At a mean follow-up of 53 months, a stable control of the hydrocephalus was documented in 6 of 10 available patients (one patient died for a severe pulmonary infection one year after surgery) [19]. In spite of the relatively low number of children, this series, confirms that endoscopic third ventriculostomy might be a valid option for the management of an obstructive featuring hydrocephalus in complex craniostenosis.

Abnormal patterns of the intracranial venous circulation

Alterations of the cranial vault as well as skull base venous structures can contribute to increased intracranial pressure in syndromic craniosynostoses, independently from their eventual contribution to alterations of the CSF dynamics, or to cerebellar tonsils herniation [66]. Taylor et al. analyzed the venous drainage angiographic findings of 23 patients, all of whom had either a craniosynostosis-related syndrome (18 patients) or a nonsyndromic multisutural synostosis (five patients). Twenty-one patients had experienced raised ICP invasively documented. Nineteen of them did not have an associated hydrocephalus. A significant degree of stenosis (51–99 % reduction in normal diameter or the complete absence of flow) at the level of the complex represented by the sigmoid sinus, intraosseous portion of the jugular sinus, and jugular bulb was documented in 18 cases, unilaterally in 7 and bilaterally in 11. A lesser degree of narrowing (1–50 % reduction in normal diameter) was noted in a further three cases. In 11 of the 18 cases with more severe narrowing of the sigmoid/jugular complex, a florid venous collateral circulation in the region of the stylomastoid emissary veins on one or both sides coexisted, with transosseous venous drainages in all cases [66]. The possibility of venous outflow obstruction to be associated with increased intracranial pressure, without the development of hydrocephalus, has been confirmed by the MR venography study of Rollins et al. who documented a venous outflow obstruction in 11 of 17 children with complex craniosynostosis, of whom 2 did not have hydrocephalus in spite of clinical signs of increased intracranial pressure; the main venous outflow collaterals in this series was through the posterior condylar veins [57]. A further evidence of the relevance of sigmoid/jugular bulb stenosis in the pathogenesis of increased ICP has been provided by the study of Rich et al. These authors compared the measurements of the jugular foramina obtained from reformatted helical CT scans in 12 children with complex or syndromic craniosynostosis and raised ICP with those obtained in two control groups of children with respectively nonsyndromic (10 cases) and syndromic (9 cases) craniosynostosis, but normal ICP. Children with raised ICP had significantly narrower jugular foramina than did the age-matched control subjects. No significant difference existed between the two control groups [55]. Different hypotheses have been advanced for the etiology of venous obstructions in children with syndromic craniosynostosis. Skull base abnormalities, constricting the normal venous pathways, remain the favored one, seen the more frequent involvement of the sigmoid giugular complex [34, 55, 66]. Another mechanism that has been called into question is the overexpression of fibroblast growth factors (FGFR1–3) at the level of the vascular endothelia, that could determine endothelial proliferation and differentiation in the sigmoid sinuses and jugular complex with consequent narrowing of their lumen. Immunohistochemistry and experimental studies seem to confirm this possibility [45]. The contribution of venous circulation obstruction to ICP increase is considered to vary with time; it is more frequently seen in the first years of life to be gradually reducing and becoming rare after the 6th–7th year of age, as the result of the opening up of a collateral venous circulation, including the development of transosseous connections [34, 55].

Management strategies for the abnormal venous pathways contribution to raised ICP

The management of raised ICP when apparently mainly related to an obstacle in skull base venous circulation is a controversial issue. Theoretically, the direct decompression of the venous pathways at the level of their exit from the skull base should represent the treatment of choice. However, direct decompression may be partial, when concerning the only true intraosseous venous channels, as the decompression would not address the usually coexisting stenosis at the level of the sigmoid sinus and passage from the sigmoid sinus to the jugular bulb [55]. The alternative of a by-pass between the transverse sinus and the jugular vein, originally proposed by Sainte-Rose et al. [58], though reported as successful in the short term, has been not performed extensively and is considered as a relatively high-risk procedure. Interventional radiological stenting, reported as successful in stenosis of venous sinuses of different etiologies, cannot be taken into account in children with complex craniosynostoses, due to the fact that the stenosis, as previously stated, is related to a bone/vascular walls hyperplasia [34, 55]. On these grounds, the most accepted procedure still remains a cranial vault expansion directed at the posterior parietal and occipital region. Clinical and ICP monitoring data are in favor of the efficacy of this procedure (Fig. 3). However, drawbacks are represented by the fact that this surgery, either performed with a free bone flap technique, or through a fixed expansion and even with the use of distractors, does not eliminate a prominent stenosis of the jugular foramina, so that patients with this anomaly remain at high risk of a persisting increased intracranial pressure as well as of developing a secondary hydrocephalus [34].

Fig. 3
figure 3

a Occipital expansion free-floating procedure: intraoperative image of the patient of Fig. 1 showing the marked retrusion of the occipital bone documented at the CT at the level of the craniocervical junction. b ICP monitoring representative sample trace of the prolonged ICP recording performed before occipital expansion, showing pathological ICP values and waves. c Prolonged ICP monitoring: bar graphic of the whole recording (48 h) performed before occipital expansion showing pathological ICP values (>12 mmHg, red bars) for 90 % of the recording time. d Occipital bone at the end of the free floating occipital bone detachment procedure, showing its elevation from the dural level, favored by the increased intracranial pressure. e, f Postoperative ICP recording (after occipital expansion) documenting a normalization of ICP values, both on single trace and bar graphic of the whole recording

Full size image

Upper airways obstruction in children with complex craniosynostoses

The contribution to increased ICP

The causal correlation between upper airways obstruction and raised ICP in children with syndromic craniosynostoses has long been recognized. Hayward et al. contemporarily measured ICP and arterial blood pressure in a series of 11 children with complex craniosynostoses during sleep. ICP was elevated during quiet sleep only in five cases, whereas during active sleep, pathological ICP values were recorded in all the patients. This leaded to a significant reduction of cerebral perfusion pressure (CPP). During quiet sleep, the mean baseline CPP was in the normal limits (mean baseline CPP = 53.3 mmHg; range, 34–70 mmHg), falling during active sleep to a mean of 32.6 mmHg (range, 23–52 mmHg). All patients experienced obstructive breathing problems that appeared to be temporally related with abnormal ICP recordings and consequently CPP. Arterial blood pressure showed only modest increases, suggesting a relatively insufficient compensation of it to the decreased CPP [33]. The same authors reported, in a personal communication, improvements of ocular findings of increased ICP after management of airways obstruction. Expiratory obstruction and occlusion of the neck venous structures, due to the related abnormal activity of the accessory respiratory muscles, are the main mechanisms considered to contribute to increased ICP in this context [34].

Respiratory problems related to upper airways obstruction

Upper airway obstruction is related to distortion and displacement of mid-third and lower third of the craniofacial bones, which can compromise both nasopharyngeal and oropharyngeal spaces.

The causes of upper airways obstruction include reduction of the pharyngeal overall space, increased length and thickness of the velum, decreased length of the hard palate, marked reduction of the posterior cranial base, with a lesser degree reduction of the anterior cranial base, choanal stenosis or atresia, midnasal stenosis, tracheal cartilaginous sleeve, and maxillary hypoplasia [14, 43, 47]. These changes are already present in infancy and tend to be more pronounced with craniofacial growth [50]. The main consequences of this complex upper airways obstructive disorder are the high occurrence rate of inflammatory diseases and, more significantly, sleep-related disordered breathing. This last includes obstructive sleep apnea (OSA), central sleep apnoea (CSA) periodic breathing and hypoventilation [2]. The commonest sleep disorder seen in children with syndromic craniosynostoses is OSA. OSA is a disorder of breathing during sleep characterized by prolonged partial upper airway obstruction (obstructive hypopnea) and/or intermittent complete obstruction (obstructive apnea) that disrupts normal ventilation during sleep and normal sleep patterns [2, 48]. In children with craniosynostosis, OSA has been described since the early 1980s and is actually considered to occur with an incidence of 40 up to 85 % of the cases [1–3, 14, 32, 43, 47, 51, 62]. The evaluation methods of OSA include clinical assessment, RX teleradiography in lateral projection, CT scan, magnetic resonance imaging, fibroendoscopy and polisomnography, and parents’ questionnaires. Of the number of questionnaires that are present in the literature, only a few can be applied in children, and a very limited proportion has demonstrated their reliability in pediatric patients with syndromic craniosynostosis [30, 40, 64, 71].

Bannik et al. tested the reliability of the OSA-18 questionnaire, a questionnaire structured to test healthy children with OSAS in a pediatric population of patients with syndromic craniosynostoses. All the patients had been contemporarily submitted to polisomnography. A 70 % reliability of the test was documented in children with syndromic craniosynostoses, which was significantly reliable in discriminating this population from healthy children [7].

In another study, the same group prospectively evaluated the reliability of the Brouillette score, a score based on a parental questionnaire taken by parents at home, in a series of 78 children with complex craniosynostoses. Results were compared with those of ambulatory polisomnography. The Brouillette score had a negative predictive value of 90 % and a sensitivity of 55 % compared with polisomnography, suggesting that it can be used to select patients for polisomnography, but it has a low reliability in children with otherwise positive polisomnographic findings [8]. The relatively low rate of OSAs documented by polisomnographic examinations in this study could be ascribed to the limits of using ambulatory polisomnography, which lacks various signals, in particular nasal flow signs, which are normally gathered during clinical registrations. A complete polisomnography examination should include electroencephalogram, electro-oculogram, submental electromyogram (EMG), and bilateral anterior tibialis EMG. Respiratory measurements include chest wall and abdominal movement using chest wall and abdominal belts; nasal airflow measurements using nasal air pressure transducer and/or oronasal thermal sensor, oxygen saturation (SaO2), trans-cutaneous carbon dioxide, and end-tidal carbon dioxide [48].

Only two studies are present in the literature in which a complete polisomnographic examination has been performed in children with syndromic craniosynostoses. Nelson et al. reported an incidence of 72 % of obstructive upper airways disturbances (18/25 cases), which was documented with complete polisomnographic studies in 12 cases (6 having undergone a previous tracheostomy). The mean preoperative respiratory disturbance index was 33.4–37.57 (range, 1.8–109.2), indicative of moderate obstructive sleep apnea. In the series of Al-Saleh et al., 26 of 35 children (74 %) had abnormal PSG results. Of these, 7 of 26 had evidence of mild OSA, 7 of 26 had moderate OSA, 10 of 26 had severe OSA, and two patients had CSA in addition to moderate to severe OSA [46].

In addition to PSG, multiple imaging techniques have been used to evaluate the upper airways in patients with OSAS. Static techniques include cephalometric radiography, CT, and MR imaging. Dynamic techniques include fluoroscopy, somnofluoroscopy, cine CT, fast CT and MR imaging, and fluoroscopic MR; the main limit of cephalometric evaluations is that they are usually applied to patients in upright position and can show craniofacial structures only in two dimensions; the upper airways areas changing during breathing and decreasing in the supine position. CT, performed in the supine position, provides information about airway cross-sectional area and site of collapse, resulting in particular value when performed in different phases of respiration. In spite of the number of papers referring to CT volumetric studies of the upper airways in patients with OSAS, only scanty information is present in the literature about the relationship between CT and polysomnographic findings in children with craniosynostoses [74]. The difficulties in performing a dynamic evaluation of upper airways in pediatric patients, due to their scarce cooperation, might surely have contributed to this lack of available data. Although on limited number of patients a substantial correspondence between clinical and CT findings seems to be confirmed (Fig. 4). Endoscopy of the rhinopharyngeal and laryngeal region are actually advocated only in doubtful cases in order to document any associated malformation of this regions [48].

Fig. 4
figure 4

Upper airways restriction in syndromic craniosynostoses: 4-year-old Crouzon child. a Lateral view of the patient before fronto-orbito-maxillary distraction showing the retrusion of the fronto-orbital region and maxillary complex. b, c Sagittal midline view of the cranio-cervical CT with upper airways space reconstruction, before fronto-orbito-maxillary distraction advancement, showing a significant reduction of the volume of the upper airways. d Lateral view of the same patient after fronto-orbito-maxillary monobloc distraction advancement showing the significant improvement of both the fronto-orbital and maxillary region retrusion. e, f Sagittal midline view of the cranio-cervical CT with upper airways reconstruction showing a significant increase in the upper airways volume after the distraction advancement procedure

Full size image

Management options

Management options for OSA in children with syndromic craniosynostoses include choanal dilatation, [38, 43], continuous positive airway pressure (CPAP) [31, 35, 38], nasopharyngeal airways [1, 31], palatal surgery [43], adenotonsillectomy [3, 35], midface advancement surgeries [24, 38, 42], or tracheostomy [35, 38, 43, 62]. Adenoidectomy and tonsillectomy are usually ineffective; it is seen that the main causes of upper airways obstruction are midface hypoplasia and the reduction of respiratory spaces. Removing the soft tissue components of upper airways determines only a modest increase in the upper airways space [25]. CPAP has similar limits; its conceptual role is to maintain a constant pressure along the upper airways during inspiration and expiration. However, its impact on upper airways obstructive sleep apneas in children with complex craniosynostoses might not be sufficient in the long term, as its most significant effect is enlargement of the lateral pharyngeal walls, again not dealing with the leading cause of upper airways obstruction in this population [8, 25]. Indirect effects of CPAP include increased vagal tone, increased cardiac output, and decreased systemic vascular resistance, which might contribute to the temporary benefit, which has been described. Further problems with CPAP are the relatively low tolerance and difficulty to use it correctly by the pediatric population [25]. Among the other temporary measures, oral devices like tongue retaining and mandibular repositioning appliances and the so-called stage I surgical maneuvers are the most frequently adopted ones. Stage I surgical procedures include nasal septoplasty and turbinectomy. Tongue surgery is to be considered in children with associated macroglossia, and palato pharyngoplasty in dealing with the mechanical enlargement of the mucosal retropharyngeal space. Static maxillo-mandibular advancement also has several limits. One of them is that when performed before the end of maxillo-facial bone structure growth (7th year of life), it yields a high incidence of early recurrences. The introduction and continuous development of osteodistraction devices has partially changed this perspective. The concept of gradually reaching an enlarged volume certainly stays with a higher possibility of a more stable ossification at the wanted stage [8]. However, results of maxillo-mandibular osteodistraction on obstructive disorders of the upper airways have been variable in large series. Flores et al. examined the clinical courses and analyzed the cephalograms of 20 patients, who underwent Le Fort III distraction. They found that nasopharyngeal and velopharyngeal airspaces are enlarged, and respiratory conditions improved after Le Fort III distraction all cases. Nine of 10 patients with significant airway compromise experienced improvement in their symptoms of OSA or had their tracheotomy removed [26]. On the other side, Witherow et al. reported that only six of the 14 children (43 %) in their series had a resolution of their OSA after monobloc with external distraction, in spite of a satisfactory anatomical enlargement of the upper airways, which was obtained in all cases. The other eight patients remained dependent on tracheostomy or CPAP (mean follow-up, 24 months) [73]. Arnaud et al. showed a respiratory improvement measured by oxygen level in 14 of 16 patients after monobloc with internal distraction. In severe cases, removal of tracheostomy was possible in four of six (67 %) children (mean follow-up, 2.5 years) [4, 5]. Nelson et al. studied 18 patients with syndromic bilateral coronal synostosis and OSA; in 15 of them, a tracheostomy or CPAP was required before midface advancement. After midface advancement, five patients were decanulated and in six CPAP was discontinued (mean follow-up, 3.2 years) [46]. In the series of Bannink et al., only 6 of 11 patients with moderate to severe OSA had a satisfactory control of the obstructive respiratory symptoms, whereas the procedure was ineffective in the remaining 5 patients (mean follow-up, 3.5 years) . In this last series, pharyngeal collapse seemed to play a role in the persistence of respiratory disturbances. The authors suggested performing an endoscopic exam of the upper airways in all children candidates of maxillo-mandibular advancement, in order to prognostically identify those who could benefit of the procedure in the long term [8].

A technical limit of the traditional distraction devices is that they cannot be used in infants. In fact, the cranial bones of young children are thin and brittle, and the risk of intracranial injuries and/or skull/maxillary bone fractures, both at the time of the system implantation, as well as during the period of osteodistraction, is relatively high [4, 5]. To overcome these problems, some authors have proposed the use of an alternative system distracting the midface through transfacial pins. Preliminary encouraging clinical results date back at the beginning of this century [4, 49]. Mitsukawa et al. described the results of this technique in five infants (mean age, 18.4 months) with syndromic craniosynostoses and severe OSA in the first months of life. At a mean follow-up of 3 years (min., 2 years; max, 4.6 years), the frequency of OSA (both studied with PSG and cephalograms) was markedly reduced after surgery, respiratory conditions were greatly improved, and tracheotomy was avoided in all cases. In all of them, there was a scar of a few millimeters in bilateral cheeks where transfacial pins were inserted, but there were no major complications associated with distraction [49]. Finally, tracheostomy, once considered the only treatment able to long-term overcome the problem of severe upper airways obstruction in the very young, is actually considered only in most severe cases and in those not responding to the above-mentioned surgical procedures [25].

Chiari malformation

The association between Chiari malformation (CM) and syndromic craniosynostosis has been recognized for several decades.

Cinalli et al. examined 95 patients with syndromic craniosynostosis and found CM in 70 % of those with Crouzon syndrome, 75 % of those with oxycephaly, 50 % of those with Pfeiffer syndrome, and 100 % of those with the Kleeblattschädel deformity, whereas it was present in only 1.9 % of children with Apert syndrome. Cerebellar tonsillar descent of <2 mm below the basion–opisthion line was used by the authors as the diagnostic criterion for a Chiari malformation, perhaps leading to the high number of diagnosed cases in this experience [11]. Other reports, however, have substantially confirmed these data [7, 9, 33]. Francis et al. [27] found an associated CM in 5 of 10 patients with Crouzon syndrome. Fearon and Rhodes [24] found that 84 % of the 28 patients they treated for Pfeiffer syndrome had associated CM. Strahle et al. [63] reported an incidence of 35 % of Chiari malformations when all children with multisuture synostoses were considered; however, when only pansynostosis patients were selected, this incidence raised up to 80 %. Earlier closure of the lambdoid suture is actually considered as the main contributor for the higher rate of CM in patients with Crouzon and Pfeiffer syndromes if compared with those with Apert syndrome. It has also been observed that lambdoid suture synostosis occurs significantly earlier in the cases of Crouzon syndrome associated with CM compared with Crouzon’s patients without CM [10, 11].

Most authors consider CM associated with complex craniosynostosis as an acquired disorder with a multifactorial etiology. Premature and progressive fusion of cranial vault and/or cranial base sutures is one of the main pathogenetic factors called into the field; the observation that, in most cases, hindbrain herniation is not present at birth, but develops paralleling the modification of the skull shape secondary to premature closure of the lambdoid and cranial base sutures (usually between 3 and 6 months of age) supports this hypothesis. Closure of the skull base sutures determines jugular foramina stenosis, with consequent venous hypertension; the latter, as mentioned before, is further worsened by the crowding of the posterior cranial fossa, due to the consequent compression of the transverse and sigmoid sinuses; finally, both crowding of the posterior cranial fossa and venous hypertension are the main causes of the development of hydrocephalus in these patients, acting on its side in a loop for the worsening of the cerebellar tonsils descent (Fig. 5). Interestingly, shunt positioning for the management of the hydrocephalus might lead to a worsening of the Chiari I malformation. Similarly to what may happen in infants shunted for different causes, lowering of the intracranial CSF volume and pulsations favors the fusion of skull vault and skull base sutures, increasing the venous turgor and pressure, and contributing to the worsening of the caudal descent of the cerebellar tonsils [11]. Among pathogenetic mechanisms, genetic mutations have also been considered. Mulliken et al. tried to find a correlation between the mutation observed and the presence or not of CM. They suggested that in Crouzon syndrome, the patients affected by CM and syringomyelia present a variety of mutations that spreads over exons IIIa and IIIc of the FGFR2 gene [44]. The involvement of exons IIIa and IIIc of FGFR2 in Crouzon’s patients with CM was confirmed by a subsequent study, which described a novel, previously unreported FGFR2 mutation that has been found to be associated only with Crouzon syndrome with Chiari I and syringomyelia, not with Crouzon syndrome without CM [28].

Fig. 5
figure 5

Chiari I as an acquired and progressive disease in syndromic craniosynostoses: Crouzon syndrome. a Prenatal MR showing a normal posterior fossa and normal position of the cerebellum. b Postnatal MR of the same patient at 1 month of age; the restriction of the posterior fossa space and the initial descent of the cerebellar tonsils through the foramen magnum is documented. Active hydrocephalus is present. c, d Sagittal T1 (c) and coronal T2 (d) MR images of the same patient after two months (3 months of age) showing a progression of both the Chiari I malformation and of the ventricular dilation with compensating acrocephaly and bulging of the temporal fossae

Full size image

Management of Chiari I malformation

The management of the Chiari I malformation in children with complex craniosynostoses is dependent on the clinical and radiological condition as well as on the age of the patient at the time of diagnosis. In children with active ventricular dilation, hydrocephalus should be managed as first step. There is a growing body of evidence that ETV can be considered as a valid alternative to VP shunt implantation in these cases [19]. Children in the first years of life, with controlled hydrocephalus or without a symptomatic ventricular dilation, should undergo primarily a cranial vault expansion [36]. Surgical maneuvers centered selectively on the foramen magnum are indeed destined to failure for two reasons: (1) they do not deal even minimally with any of the three factors that contribute to the development of the Chiari malformation (namely, craniocerebral disproportion, venous hypertension, and hydrocephalus) and (2) a rapid regrowth of the removed foramen magnum bone would lead to an early failure of the temporary craniocervical junction decompression [11]. In the more frequent condition of a dominant early involvement of the lambdoid sutures, occipital cranial vault remodeling and expansion should be considered as the preferred first surgical procedure; in fact, it allows to primarily enlarge the posterior compartment space, protecting posterior fossa intracranial structures as well as to decompress the main dural venous sinuses. A midline suboccipital craniectomy should be added, widely opening the lateral margins of the foramen magnum, in order to avoid failures related to bone regrowth [60]. Opening of the dura is not advised, unless the patients present clinical signs of severe compression of the brain stem; severe bleeding from the dural edges has been described and related to the significant collateral venous circulation at the level of the foramen magnum [11, 63]. Different techniques have been proposed for posterior cranial vault expansion. Free bone flaps are considered in the first months of life mostly hoping on the possibility of the brain pulsations to progressively enlarge the intracranial space. Advantages are the relatively easy technique and the possibility to avoid detaching the bone from the main venous sinuses, with a consequent reduction of the blood losses; the main disadvantage is the lack of possibility to predict the amount of cranial space enlargement [60]. Posterior distraction advancement has been proposed as an alternative with the main aim of overcoming this problem; similarly to free bone flaps, it allows to avoid the split of the bone from the dural sinuses and to reduce both operative times and blood losses [72]. The limit is that it easily lead to bone fractures during distraction in the very young, with an up to 30 % rate of complications related to dislodgement of the implants; moreover, even in the absence of complications, at least a second surgery is required in order to remove the distractors [60]. Before the introduction of distraction devices, fronto-orbital advancement was considered as an early obliged second surgical step to obtain an adequate enlargement of the intracranial space. Actually, most authors defer this treatment, unless severe respiratory conditions coexist. Posterior cranial vault expansion has been demonstrated to adequately deal with the clinical problems related to craniocerebral disproportion in early life, Fronto-orbito-maxillary advancement can indeed be performed earlier than static advancement procedures, with a much lower risk of recurrence. One of the main advantages of this kind of protocol is the possibility of avoiding a reoperation in the fronto-orbital region, which is associated with a relatively high rate of complications.

When Chiari malformation is diagnosed in children beyond the end of cranio-maxillofacial growth, observation is adopted in asymptomatic cases, whereas in the presence of clinical symptoms, opening of the foramen magnum with/without duroplasty can be considered as a therapeutic surgical maneuver. This is because both craniocerebral disproportion and venous sinuses compression are much less severe than in younger children; moreover, the risk of bone regrowth is significantly lower [11].

Ocular problems in children with complex craniosynostoses

Optic nerve and orbit-related eye-globe diseases are both common concerns in children with syndromic craniosynostoses [39]. Papilledema is mostly related to the presence of an increased intracranial pressure. In a series of 84 children with syndromic craniosynostoses, Bannink et al. found papilledema in 51 % of the patients. It usually had a subtle clinical course, unrelated with any specific clinical symptom; exorbitism and ventricular dilation were the only significantly related factors [6]. Due to above-mentioned multifactorial origin of increased intracranial pressure, surgical treatment, either of an eventually associated hydrocephalus, or of the craniocerebral disproportion and venous hypertension is not expected to permanently solve the problem of papilledema; yearly fundoscopic examinations should therefore be advised both at the time of diagnosis as well as in the follow-up clinical evaluation and after the different steps of surgical treatment up to the end of craniomaxillary growth (7 years of life).

On a different note the synostotic disease causes significant abnormalities of the orbital space and might severely condition the function of intraorbital structures. As a distinctive feature, the bony orbit is shallow owing to retrusion of both the supraorbital and infraorbital margins; compensatory bulging of the temporal lobe into the posterolateral wall of the orbits further reduces the volume of the already shallow orbit, resulting in ocular proptosis. Hypertelorism and antimongoloid slant of the palpable fissure may also be present. High incidences of strabismus and amblyopia reflective error and astigmatism have been reported [39]. Kreiborg et al. recently presented data suggesting that, in spite of the fact that ocular proptosis and hypertelorism characterize both disorders, differences exist in the type of orbital dystopia between children affected by Crouzon and Apert syndromes. In Crouzon syndrome, ocular proptosis is primarily caused by retrusion of the lateral and inferior orbital margins with a very short orbital floor. In Apert syndrome, the eye globe actually protrudes in relation to the cranial base and to the orbit, probably resulting from marked protrusion of the lateral orbital wall. The implications account for some of the differences encountered. Asymmetry of the orbital position is associated with Apert syndrome frequently. Exotropia is found in Crouzon syndrome, whereas the V pattern is more characteristic in Apert syndrome with divergent upgaze and esotropic downgaze. Subluxation of the eye globe is found in some cases of Crouzon syndrome but is not found in Apert syndrome. Structural alterations of the extraocular muscles have been associated with some cases of Apert syndrome, suggesting that ocular motility disturbances in Apert syndrome may not be caused solely by mechanical factors. Absence of the superior rectus and other extraocular muscles has been recorded [39].

Shallow orbits and secondary proptosis are the main contributors for the relatively high risk of these children to develop corneal ulcers. Ueek et al. described a 50 % incidence of corneal ulcers in a series of six children with Apert syndrome dealt with at their Institution in a period of 6 years. Morbidity in these authors experience was high and included decreased visual acuity, opacified corneas, amblyopia, and blindness [70].

Management of ocular problems

The management of papilledema is complex and corresponds to the complex management of bone abnormalities and increased intracranial pressure in these patients.

Surgical treatment of orbit-related ocular problems is substantially coincident with the management of fronto-orbito-maxillary retrusion. As specified in other papers of this issue, this include fronto-orbital and midface advancement, correction of hypertelorism, and other, eventually needed, adjunctive procedures on soft periorbital tissues. Contemporary fronto-orbital and maxillo-facial advancement can be performed through a single step “mono-bloc” detachment of the two structures from the skull base, both dealing with the roof and floor of the orbit retrusion, or separately, with orbital floor retrusion dealt with through a Le Fort III osteotomy. External and internal distraction devices are actually the preferred mode to obtain a gradual gain of the correct space both for the orbital and maxillofacial regions after bone detachment. When hypertelorism (more frequently in Apert syndrome) coexist, a variation of Le Fort III osteotomy known as facial bipartition [13] can correct the interorbital distance as well as the anteroposterior and vertical midface hypoplasia. This involves removal of a triangular segment of bone from the midline of the osteotomized midface [56].

One of the risks of fronto-orbito-maxillary advancement is that complete correction of the nasomaxillary deficiency often requires more advancement than would be desirable at the orbital margins, and this can result in enophthalmos. Both the nose and maxilla are often deficient vertically, and full correction requires a vertical elongation of the midface, a movement that can further increase orbital volume and enophtalmos. Of paramount importance in orbital osteotomies is the management of the medial canthal ligament. This is best left attached to its bony base, and great care must be taken in subperiosteal dissection to avoid encroaching on the area of the medial canthus. Medial canthopexy has to be performed when medial canthus detachment occurs [56].