Premature Fusion of the Sagittal Suture as an Incidental Radiographic Finding in Young Children

Abstract

Background:

Craniosynostosis (CS) typically develops prenatally and creates characteristic changes in craniofacial form. Nevertheless, postnatal forms of craniosynostosis have been described. The purpose of this study was to determine the prevalence of incidentally-identified, but temporally premature, cranial suture fusion in normocephalic children.

Methods:

Computerized tomography (CT) scans obtained from children age 1 to 5 years old evaluated in our emergency department between 2005– 2016 were reviewed for evidence of CS. Patients with prior VP shunt, brain or cranial abnormality, or known syndromes were excluded. The presence of CS and cranial index (CI) was assessed by a panel of three craniofacial surgeons and one pediatric neurosurgeon. Demographic information, fusion type, reason for the CT scan, and past medical history were recorded as covariates. Cranial shape and intracranial volume were calculated using previously validated automated system.

Results:

331 patients met the inclusion criteria. The mean age was 2.4 ± 1.3 years. Eleven (3.3%) patients were found to have a complete (n=9) or partial (n=2) fusion of the sagittal suture. All patients had a normal CI (0.80, range 0.72–0.87) and a grossly normal head shape. Only 2 fusions (18.2%) were documented by the radiologist. Cranial shape analysis performed in 5 of the 11 patients showed subtle phenotypic changes along the scaphocephaly spectrum in 4 patients, with a normal shape in the remaining case.

Conclusion:

Sagittal fusion is present in 3.3% of otherwise phenotypically-normal children aged 1–5 years. The clinical significance of this result is unclear, but routine screening of affected patients is paramount.

INTRODUCTION

The overwhelming majority of patients develop craniosynostosis before birth and are born with characteristic changes in the craniofacial skeleton that herald the diagnosis and underlying pathology.^1–3 Nevertheless, there are well documented exceptions to this rule. Postnatal forms of craniosynostosis can occur and the resultant changes in cranial shape may be subtle or even imperceptible to even an experienced observer.^4–6 One example is progressive postnatal pansynostosis, a condition in which multiple cranial sutures fuse after birth rendering a morphologically normal, but volumetrically undersized, cranial shape.^6,7 Postnatal cranial suture fusions have been described in association with various metabolic conditions,^8,9 ventriculoperitoneal shunt placement for hydrocephalus,¹⁰ and certain genetic conditions.¹¹ These forms are often termed “acquired” or “secondary” craniosynostosis to distinguish them from fusions that occur in-utero, although this terminological distinction is of dubious significance.

In 2010, the Oxford Craniofacial Unit reported described a group of patients with isolated sagittal synostosis who had no scaphocephaly phenotype.⁴ The authors termed this condition normocephalic sagittal craniosynostosis and they hypothesized, given the lack of characteristic secondary cranial shape changes, that the affected suture had closed after birth.⁴ Ruane et al, reported a larger series of patients with the same findings, and recommended reserving surgery only for those with objective evidence of elevated intracranial pressure.¹² Our group recently reported on an additional 19 patients with this condition and found that most had subtle changes in cranial form that were along the spectrum of scaphocephaly.¹³ Given this observation, we suggested the term occult scaphocephaly to more accurately describe the phenotype of this patient cohort. One disturbing aspect in each of these studies was that most of the patients were identified by CT scan taken for reasons other than cranial shape changes or suspicion of craniosynostosis. This raised an important question as to whether postnatal fusion of isolated cranial sutures, especially the sagittal suture, is more ubiquitous that previously realized. The purpose of this investigation was to determine the prevalence of fused cranial sutures, other than the metopic suture, identified incidentally on CT scans of phenotypically normal (i.e., no cranial shape changes) children.

PATIENTS AND METHODS

Following IRB approval, we queried a radiology database of CT scans taken on patients age 1–5 years evaluated in our emergency room between 2005 and 2016. The inquiry specifically included only patients whose scans were read as “normal” on the radiologist’s report, thus excluding all patients with: cerebral and/or cranial shape or structural abnormality; hydrocephalus or the presence of a shunt; patients with morphologic evidence of prior cranial surgery. All scans were additionally reviewed individually and collectively by three craniofacial surgeons (G.F.R, A.K.O and J.R.B) and one pediatric neurosurgeon (R.F.K) to verify that there were no identifiable intra- or extracranial abnormalities, and to identify the presence or absence of fused cranial sutures. Cranial index (CI) was measured and calculated on the CT axial images. Additional data collected included: patient demographics, age at the time of CT scan, past medical history, suture fusion findings, final radiologist opinion/read, reason for the imaging, and the patient medical record. All patients with a known syndrome or medical diagnosis were excluded. Cranial shape and volume were calculated using our previously described automated system.^14–18 In addition, we measured the intracranial volume and compared it to an age-matched normative database.¹⁹

RESULTS

A total of 331 CT scans were evaluated, with a mean age at imaging of 2.93 SD ±1.31 years. Most of the patients were male (54.1%) and the most common indication to receive a CT scan was trauma, reported in 179 patients (51.1%). Other indications included headaches (9.4%), seizures (9.1%), and altered mental status (6.04%), among others. Of the total cohort of patients, 17 (5.14%) had significant past medical history including chronic headaches (n=7), developmental delay without an underlined syndrome (n=8), and non-febrile seizures (n=2).

Eleven patients (3.3%) were found to have fusion of the sagittal suture; nine of the CT scans showed complete fusion [Figure 1], and two had a partial (30–40%) closure [Figure 2]. None of the scans demonstrated endocortical scalloping, enlarge ventricles, or visible evidence of cerebral crowding [Figure 3]. CI was normal in all patients with a mean of 0.80 (range, 0.72–0.87). Surprisingly, only 18.2% (n=2) of the formal radiology reports in the affected patients documented the presence of the cranial suture fusion. The majority of patients were male (63.6%) and 45.46% (n=7) had their CT scan between two and three years of age [Table 1]. Indication for CT scan in patients who had craniosynostosis included trauma (n=7), followed by non-febrile seizures (n=2), headache (n=1) and dog bite (n=1). Past medical history of this subset of patients was remarkable only for congenital cardiac anomalies and Rett syndrome in two different patients.

Figure 1:

3D reconstruction of 2-year old patient with complete fusion of the sagittal suture; (left) Vertex view, (right) frontal view.

Figure 2:

3D reconstruction and CT scan of 4-year old patient with partial (30%) fusion of the sagittal suture; (top, left) Vertex view, (top, right) posterior view. (bottom) Axial CT cut at vertex showing fusion of posterior but not anterior sagittal suture.

Figure 3:

Axial CT images of 5-year old patient with complete closure of the sagittal suture showing no evidence of endocortical scalloping (left) or obvious brain compression (right).

Table 1:

Craniosynostosis prevalence per age group

Age group (months)	Sagittal suture fused (%)
12–23 (n=104)	0 (0%)
24–35 (n=84)	5 (6%)
36–47 (n=65)	2 (3%)
48–59 (n=49)	3 (6%)
59–61 (n=29)	1 (3.5%)

Intracranial volume analysis of the affected patients was compared to normative data. Mean ICV was 1204.2 ± 112.2 cubic centimeters (cc³). Ten of the 11 patients with sagittal synostosis fell into the 95% prediction interval around the predicted ICV value for the normative subjects. Only one subject was below the prediction interval and this was in the patient who had Rett Syndrome, a condition that can be associated with low ICV. [Figure 4] Due to the poor quality of the CT scan taken at the ED, only 5 (45.5%) of the CT scans from fused patients could undergo cranial shape analysis. Four of the patients showed subtle phenotypic changes along the spectrum of scaphocephaly, while the remaining case was deemed completely normal in form. [Figure 5]

Figure 4:

Intracranial volume of all patients with incidentally diagnosed sagittal fusion (red) as compared to the normative cohort (blue).

Figure 5:

Representative cranial shape analysis (CSA) of a patient with a fused sagittal suture and grossly normal head shape shows subtle phenotypic changes along the spectrum of scaphocephaly.

DISCUSSION

The results of this study show that a small, but significant, percentage (3.3%) of children under 6 years of age will develop fusion of the sagittal suture at an age that is historically considered premature. This finding raises questions regarding the pathophysiology of the sagittal craniosynostosis and, more importantly, about the clinical significance of this observation. Besides the metopic sutures, which normally closes in infancy,^20,21 the sagittal and other cranial sutures should remain patent until after skeletal maturity.^22,23 It is conceivable that the observed fusions merely represent normal variation in the temporal closure of the sagittal suture. If that were the explanation here, however, one would expect the prevalence to rise steadily with increasing age- no such effect was observed. In fact, the prevalence stayed relatively constant for each age group observed. We would posit, therefore, that sagittal suture fusion in this group of young children (<6 years of age) is premature in its onset and technically, should be considered pathological and, a priori, termed craniosynostosis. Nevertheless, none of the affected patients manifest any visible or anthropometric (i.e. cephalic index <0.70) changes characteristically observed with sagittal craniosynostosis. We did observe some mild phenotypic changes along the scaphocephaly spectrum, but this was only noted using very sophisticated proprietary imaging software we have previously described.^14–19 Thus, this group of patients falls into the subset of “normocephalic” sagittal craniosynostosis as described by Morritt et al. and Ruane and colleagues.^4,12

While an incidental suture fusion in this age group is abnormal, it is unclear if it has any clinical importance. None of the 11 CT scans that demonstrated a fused suture showed radiographic finding suggestive of elevated ICP, and volume analysis on this subset of patients was normal with the exception of the one patient with Rett’s syndrome. Three of the affected patients did presented to the ER for symptoms that could be consistent with elevated ICP (i.e., seizure, headache), but subsequent records failed to show that these issues continued or were persistent. Morritt and coworkers reported raised ICP in 44% of patients with isolated SC with normal CI.⁴ This number falls between that reported by Inagaki et al (69%)²⁴ and Ruane and coworkers (25%).¹² The latter group has largely advocated observation in the absence of clear ICP elevation (i.e., papilledema and/or latent visual evoked potentials), but these tests may not have sufficient accuracy in young patients on which to rely. Tuite et al found that only 22% of patients 8 years of age or younger with ICP elevations by measurement had papilledema, giving the test a 22% sensitivity in this age group.²⁵ Although some reports demonstrate a strong correlation of visual evoked potentials and ICP elevation,^26–29 other authors have found high inherent inter-individual variability in normal persons that makes these tests unreliable.^29,30 The overall poor sensitivity of non-invasive monitoring tests, especially in young children, coupled with wide variability in reported ICP for patients with this diagnosis and the negative consequences of unrecognized ICP elevation mandate that all affected patients should at minimum undergo very careful and comprehensive clinical monitoring. Regardless of the management approach taken in these patients, however, the bigger issue is how to identify them. Our data would suggest that these patients are ubiquitous throughout the pediatric population and that the relatively small number of reported cases are those that were identified through mere happenstance.

An especially provocative finding from our investigation was that only the sagittal suture was affected in this cohort of children. The sagittal suture is the most common type of single-suture craniosynostosis,^32–34 with a reported incidence in the general population of between 0.2–0.25%.³³ This statistic applies principally to children in whom sagittal fusion was suspected because of an abnormal, or scaphocephalic, cranial shape. It is recognition of the characteristic head shape that typically draws attention to the child and prompts confirmatory imaging. Nevertheless, the phenotypic severity of patients with craniosynostosis can vary widely depending on the age of fusion, number of fused sutures and even genetic factors.^4,35 Patients who fuse cranial sutures later in life often lack many of the sentinel cranial shape changes and may be more likely to go undetected and undiagnosed. This study provides the first prevalence estimate of children who have radiographic fusion of their sagittal suture but without discernable abnormalities of cranial form. Based on our findings, the prevalence of this entity is significantly more common than its more phenotypically-obvious counterpart. In our opinion, the absence of any clear cranial shape changes in the affected group indicated that the fusion occurred after birth. Unfortunately, the cross-sectional study design makes it impossible to discern that age that fusion occurred.

The results reported herein do not explain why premature fusion occurred only in the sagittal suture. Many variables have been associated with an increased risk of developing craniosynostosis including genetic and epigenetic factors,^35,36 metabolic conditions,^8,9 and environmental influences such as maternal smoking,^36,37 fetal exposure to certain medications³⁹ and anti-metabolites.⁴⁰ Our study design does not allow us to rule out all of these possibilities, but we limited our review to children who had none of these risk factors in their medical history. We recently reported a nearly 50% incidence of acquired craniosynostosis in patients who underwent a ventriculoperitonal shunt (VPS) for hydrocephalus in infancy, presumably due to deleterious mechanotransduction effects brought about by supraphysiologic cranial expansion (hydrocephalus) followed by rapid decompression and collapse after VPS placement on the patency of the cranial sutures.^41–45 Interestingly, the sagittal suture was by far the most affected of the cranial sutures, and was fused in 90% of the affected cases.

Prior investigations provide some undergirding to the idea that the sagittal suture has a heightened predisposition to close prematurely. The only other midline suture, the metopic, is genetically programmed to normally close in the first year of life.^20,46 It is conceivable that the sagittal suture shares some of the same (albeit not as robust) molecular influences as its collinear neighbor and may have a heightened propensity for closure naturally or in response to certain environmental influences or mechanical forces. It is noteworthy that the structure of the two midline sutures, sagittal and metopic, is different from the more transverse sutures; coronal(s) and lambdoid. It is noteworthy that the structure of the two midline sutures, sagittal and metopic, is different from the more transverse coronal(s) and lambdoid sutures; the osteogenic edges of the former sutures normally abut one another, while those of the latter sutures tend to overlap.^21,46 If our observations merely represent the tail of the normal temporal range of sagittal suture fusion, we would expect that the incidence of sagittal suture fusion would have increased progressively with increasing age at CT- this was not observed. As noted previously, the prevalence was relatively constant between 1–5 years of age. This suggests that there is some other predisposing variable in this small group of children that accounts for our findings. The genome wide association studies (GWAS) have identified intron regions of BBS9 and non-coding regions regulating BMP2 expression associated with sagittal non-syndromic craniosynostosis.⁴⁸ Single nucleotide variants both in these loci identified by GWAS and in genes known to be associated with syndromic CS have also been identified in cases of non-syndromic sagittal craniosynostosis.⁴⁹ It is possible that some or even all of the affected patients in our study have these variations, but the retrospective study design and cost limitations prohibited testing for them.

The implications of our study are profound and far reaching. It is inconceivable that the observations reported here are limited or unique to this cohort of patients, and we posit that they are representative of the pediatric population at large. If these findings are confirmed, a small but significant percentage of normal children under the age of 5 years have a premature and undocumented sagittal suture fusion of unknown clinical significance. While none of the CT scans in our affected patients had findings suggestive of elevated ICP (e.g., endocortical erosions), detailed cranial shape analysis identified subtle cranial shape changes in some of the fused patients that appear to fall along the spectrum of scaphocephaly. We reported similar findings in a larger group of patients with sagittal synostosis and otherwise anthropometrically-normal head shapes, a condition we called occult scaphocephaly (to account for these minor shape changes)¹³ and others have termed “normocephalic sagittal synostosis”.⁴ As noted above, our current investigation leads us to conclude that this entity is vastly more common than what is reported in the literature, and that only a small fraction of affected children are ever identified and diagnosed. Unfortunately, the normal head shape in these patients makes clinical detection nearly impossible, and widespread radiographic imaging is impractical due to the cost and the recent concerns about the oncogenic effects of ionizing radiation. Fortunately, the CT-derived shape analysis protocols that we used to identify the subtle shape changes in this cohort are now formatted for use with three-dimensional (3D) photogrammetry^18,19 and this modality may offer a practical radiation-free manner of detecting affected children. Currently, the equipment for 3D photogrammetry equipment is quite large and expensive, but there are efforts underway to apply the technology using a smaller device, such as a cell phone. Such innovation would allow routine screening of pediatric head shapes to detect patients affected with this condition.

The findings reported here create an ethical dilemma for our institution that will resonate throughout the pediatric community. All of these patients were seen for reasons other than concern over craniosynostosis and the presence of a fusion went largely unidentified and were not acted on at that setting. Moreover, the CT images were reviewed for this study two years after they were obtained. Consequently, our institution and others must now wrestle with the decision of whether to review all CT images taken in the ER for such incidental sagittal suture fusions and, if unreported fusions exist (and we suspect they do), should patients be notified of the results and clinically evaluated? Recalling patients this long after the diagnosis could have been made has potential medicolegal consequences. Based on these results, we are instituting some changes to our procedures including: 1) more meticulous documentation of cranial suture patency on routine radiographs; 2) the need for clinical monitoring of patients with incidental findings of craniosynostosis before the cessation of brain development. These are questions that our institution has addressed in the affirmative. We would suggest other pediatric facilities consider reviewing their CT images to identify affected patients.

Our study has a number of limitations that require acknowledgement. First, the CT imaging was done in the course of patient care in our emergency department and, accordingly, we cannot predict the exact age at which these fusions occurred. Second, we limited our analysis to CT images obtained between 1–5 years of age. Although the sagittal suture should be patent until maturity, we limited our study to this age group as there is little dispute that fusion between the ages of 1–5 is de facto abnormal. Furthermore, suture width decreases with age and the inclusion of older patients could obscure the radiographic distinction between open and close sutures. Lastly, the clinical impact of a fusion arising at a later age (>5 years) would, in our opinion, be minor given the limited amount of brain growth remaining. It is, therefore, possible that the prevalence of radiographic sagittal fusion may actually be greater at later ages that we did not include in our review. Another potential limitation of our study is the use of the axial CT images to calculate CI, creating the possibility of error due to the tilt of the axial cuts. Careful review of the CT image cuts did not appear to show any evidence of this. Finally, the retrospective nature of the study precluded genetic testing and a more expansive clinical follow-up of the affected patients. This is largely because the study design was retrospective in nature and the data procurement process was completed nearly 2 years prior to this writing.

CONCLUSION

Postnatal premature fusion of the sagittal sutures occurs in approximately 3.3% of otherwise healthy children with normal appearing head shape. The implications of this finding are uncertain, but we caution other pediatric centers to establish protocols for identifying and evaluating these patients when the fusion is present on routine imaging.