The role of ICP overnight monitoring (ONM) in children with suspected craniostenosis
Secondary craniostenosis is a relevant problem pediatric neurosurgeons are confronted with and poses challenges regarding reliable diagnosis of raised ICP, especially in case of absent or questionable papilledema. How to identify children with elevated ICP is still controversial and diagnostics vary. We report on our experience with computerized ICP ONM in relation to imaging derived parameters.
Thirty-four children with primary or secondary craniostenosis and clinical suspicion of raised ICP were investigated. We compared clinical signs, history, and radiographic assessment with the results of computerized ICP ONM. Differences were significant at a p < 0.05.
Baseline ICP was significantly higher in patients with combined suture synostosis, who also had a higher rate of questionable papilledema. Children with narrowed external CSF spaces in MRI had significantly higher ICP levels during REM sleep. Mean RAP was significantly elevated in patients with multi-suture synostosis, indicating poor intracranial compensatory reserve. Syndromal craniostenosis was associated with elevated ICP, RAP was significantly lower, and skull X-rays showed more impressions (copper beaten skull). RAP increased with more severe impressions only to decline in most severe abnormalities, indicating exhaustion of cerebrovascular reserve at an upper ICP breakpoint of 23.9 mmHg. Headaches correlated to lower ICP and were not associated with more severe X-ray abnormalities.
Narrowed external CSF spaces in MRI seem to be associated to elevated ICP. Skull X-rays can help to identify patients at risk for chronically elevated ICP. Severe X-ray changes correlate with exhausted cerebrovascular reserve as indicated by RAP decline. Only ICP monitoring clearly identifies raised ICP and low brain compliance. Thus, in cases with ambiguous imaging, ONM constitutes an effective tool to acquire objective data for identification of surgical candidates.
KeywordsCraniosynostosis ICP monitoring Intracranial hypertension Secondary craniostenosis
In 1982, Renier et al. performed ICP (intracranial pressure) monitoring in 92 cases with craniosynostosis. They report on “normal” ICP levels in one-third of patients, borderline results in another third, and clearly pathological measurement in the last third . Identifying children at risk for raised intracranial pressure is a relevant problem pediatric neurosurgeons are confronted with and poses challenges regarding reliable diagnosis, especially in case of absent or questionable papilledema. How to identify children with elevated ICP is still controversial and diagnostics vary from department to department.
Studies with intraoperative monitoring of ICP at primary surgery have been conducted, showing a mean ICP pre-decompression of 14.7 mmHg. The postoperative course, though, is largely unknown . The incidence of secondary coronal synostosis approximates 10% in scaphocephaly after craniectomy not involving the coronal sutures, with 1% requiring surgical decompression in this series .
In untreated nonsyndromic unicoronal synostosis, raised ICP shows no correlation with clinical and radiographic findings . Syndromic craniosynostosis is associated with more frequent ICP elevation. In Apert syndrome, up to 83% as cases had pathological intracranial hypertension. These cases usually show higher rates of additional ICP raising pathologies like venous hypertension, airway obstruction, and hydrocephalus .
Signs of raised ICP in patients with suspected secondary craniostenosis can include headaches, fatigue, and papilledema. In nonsyndromic single-suture synostosis, signs of postoperative intracranial hypertension are reported to occur in about 6.2% of patients. The rate in patients with sagittal suture synostosis is highest, followed by metopic and unicoronal synostosis. Children in need of re-operation were younger at first surgery. In syndromic synostosis, the incidence of secondary ICP elevation was between 10 and 37% .
Non-invasive methods of indirect assessment of ICP include CT and MRI with their known severe limitations. ONSD constitutes a valid sonographic tool to assess ICP; however, cut-off values for the scenario of secondary craniostenosis are not yet defined. Ophthalmological evaluation for papilledema is one of the most important examinations with the known limitations regarding compliance, age dependency, and especially the known low sensitivity, implying that absence of papilledema does not rule out ICP elevation .
Elevated mean ICP and increased B-wave height as well as frequency are considered to be meaningful indicators of clinically relevant raised ICP . The definition of pathological ICP-findings is not standardized, however. Some authors suggest a mean ICP greater than 15 mmHg or more than three plateau waves in 24 h as evidence of intracranial hypertension . Others define the threshold at 20 mmHg  or use a combination of baseline as well as overnight ICP and plateau waves . Evaluation of invasive ICP monitoring showed in cases with inconclusive ophthalmologic findings that this technique can help to identify children at risk, which was more frequent in syndromic craniosynostosis . In a study investigating syndromic craniosynostosis, invasive monitoring showed a prevalence of 33.9% raised ICP . Complex craniosynostoses show intracranial hypertension in 47–64% of cases.
Secondary craniostenosis is an issue which seems to be underrepresented in literature regarding clinical management and follow-up of patients after surgery for single craniosynostosis. In trigonocephaly, an incidence of 1.9% for raised ICP is reported in patients after surgical correction. Stagnation of head circumference correlates to intracranial hypertension .
In sagittal synostosis, prevalence of secondary craniostenosis with elevated ICP is reported in about 5% and in 4% of nonsyndromic children . Interestingly, a rate of postoperative secondary coronal suture synostosis has been reported postoperatively in 89% of cases after nonsyndromic sagittal suture repair with only a 3% incidence of intracranial hypertension . Other studies show an incidence of 6.9% of intracranial hypertension16 Thus, the rate of secondary craniostenosis seems not high, but annual follow-up after surgical repair of synostosis is recommended .
Some surgical techniques have been associated with higher incidence of postoperatively raised ICP, as shown especially for modified strip craniectomy . In these cases, the role of skull X-ray, especially after isolated sagittal synostosis repair, has been evaluated and beaten copper appearance seems to be suggestive for development of raised ICP.19 Syndromal multi-suture synostosis, like in Saethre-Chotzen is more prone to postoperative recurrence of raised ICP [1, 26].
The goal of this study was to evaluate—in the setting of presumed secondary craniostenosis after primary surgery of single-suture and syndromic synostosis or in primary craniostenosis (untreated craniosynostosis in infancy)—the correlation between routine diagnostics, comprising clinical evaluation, skull X-ray and MRI, and computerized ICP overnight monitoring (ONM), which offers apart from the ICP values further calculated parameters for diagnostic assessment like ICP amplitude and the correlation index between changes of ICP and changes of ICP amplitude, called RAP as a parameter describing intracranial reserve capacity .
Materials and methods
Thirty-four children with clinical suspicion of mostly secondary but also primary craniostenosis (older children with craniosynostosis which were primarily left untreated) and suspected raised ICP were included. We compared clinical signs, symptoms, radiographic assessment of skull X-rays, and MRI results with the results of computerized ICP ONM. ICP monitoring was performed using ICM+® (Cambridge Enterprise, https://icmplus.neurosurg.cam.ac.uk/).
During ONM, ICP and associated parameters were calculated for total monitoring time, and separately for the so-called baseline periods (non-REM sleep) and during REM sleep with increased vasogenic ICP dynamics as published previously.24 ICP amplitude was calculated and expressed as AMP (first harmonic of ICP amplitude after Fourier transformation), and RAP is the correlation coefficient between 30 samples of 10 s ICP averages and corresponding AMP values.
Twenty-five cases were monitored at the University Hospital of Wuerzburg and 9 cases at University Hospital of Tuebingen. Standard postoperative follow-up of patients after craniosynostosis surgery comprises regular outpatient visits, photo documentation, morphometric head measurements, and skull X-rays. Ophthalmologic assessment was performed every six months. Clear papilledema constituted an indication for surgery without further ICP assessment.
The used grading system for evaluation of skull X-rays is based on previously unpublished own data . Skull X-rays were evaluated based on interpretation of Impressiones digitatae by Davidoff 1936  and Macaulay 1951 . Four areas were examined separately: frontal, parietal, temporal, and occipital. Grade 0 is chosen for a skull without impressiones. Grade I constitutes radiographs with mild impressiones. In grade II, we include skull X-rays with impressiones significantly over the average of grade I cases in the first two life years. Grade III is chosen when extreme impressions are seen in all quadrants. Also, localized and generalized impressiones can be differentiated, the latter comprising abnormalities in at least two areas .
In Wuerzburg, skull X-rays play a major role in assessing the likelihood of secondary craniostenosis and were routinely performed as a screening tool. In cases with suspicious findings, MRI imaging was often performed, followed by ICP ONM. If skull X-rays showed minor changes without any clinical signs or symptoms, there was only clinical and ophthalmologic follow-up. CT scans were not performed on a regular basis.
In Tuebingen, skull X-rays were only performed when clinical signs and symptoms pointed towards secondary craniostenosis; otherwise, follow-up was comparable with Wuerzburg.
MRI scans were graded in a blinded retrospective fashion with regard to the existence of normal or narrowed external CSF spaces alongside the cortical surface.
The mean age was 6.7 years (2.3–15, interquartile range 3.75). The underlying synostosis was in 14 cases (41.2%) sagittal, in 9 cases (26.5%) coronal, in 8 cases (23.5%) combined sagittal and coronal, in one combined frontal and sagittal, one pansynostosis, and one parietotemporal suture synostosis. Twenty-three cases (67.6%) were non-syndromal, 6 (17.6%) were associated with Saethre-Chotzen-syndrome, two with Crouzon syndrome, one with microdeletion syndrome, one with neurofibromatosis type 1, and one with monosomy 9p.
Headaches were reported in 52.9% of patients. Three patients (8.8%) had a questionable non-unambiguous papilledema.
In 4 cases (12%), a skull X-ray was not available. X-rays were classified as normal in 3 cases (9%), in 10 cases (29%) as grade 1, in 9 cases (27%) as grade 2, and in 8 cases (24%) as grade 3. There was no significant difference in the frequency of pathological skull X-ray findings between non-syndromal and syndromal children. Combined sagittal and coronal synostosis showed significantly lower scores than isolated coronal craniosynostosis (mean 1.29 ± vs 2.43±, p = 0.047). There was no significant difference between other combined or isolated craniosynostosis types.
MRI was classified as normal or abnormal (with narrowed external CSF spaces) in 12 (35%) and 13 (39%) cases, respectively. Nine children did not receive MRI. There was no significant difference in the frequency of pathological MRI findings between non-syndromal and syndromal children or between combined and isolated craniosynostosis.
In children with moderate skull X-rays, narrowed external CSF spaces were more common than in mild or none (82 vs 36% p = 0.030). Children with severe skull X-ray abnormalities (group 3) on the other hand show significantly less narrowed external CSF spaces (82 vs 50%, p = 0.040).
Pathological skull X-ray findings correlated with a higher frequency of questionable papilledema, although the severity of the findings did not (none vs mild p = 0.002, none vs moderate p = 0.003, none vs severe p = 0.003).
Regarding total monitoring time, mean ICPtotal in all cases was 18.0 ± 4.1 mmHg (range 13.1–32.0), mean RAPtotal was 0.58 ± 0.11 (0.38–0.83), and mean AMPtotal (first harmonic of ICP amplitude after Fourier transformation) was 1.59 ± 0.60 (0.90–6.0) mmHg.
Baseline ICP during non-REM sleep (ICPbase) was 15.7 mmHg (11–26, standard deviation 3.21), mean RAPbase 0.50 (0–1, 0.13), and mean AMPbase 1.15 ± 0.27 (1.0–2.0) mmHg.
During REM sleep, mean ICPREM was 20.8 mmHg (14–38, 5.12), with a mean maximum ICP of 38.28 (19–70, 9.72), mean RAPREM was 0.69 (0–1, 0.13), and mean AMPREM was 2.08 (1.0–6.0, 1.0) mmHg.
When comparing isolated coronal synostosis with combined cases, the differences in ICP were not significant anymore. However, mean RAPbase was significantly higher (0.43 ± 0.09 vs 0.54 ± 0.11, p = 0.040).
Comparing syndromal (n = 11) and non-syndromal children (n = 23), median ICPtotal was significantly higher in syndromal cases with values of 19.5 ± 5.46 vs 16.4 ± 2.94 mmHg (p = 0.033). The same was seen with mean ICPREM (23.39 ± 7.00 vs 19.56 ± 3.50 mmHg, p = 0.039). All mean RAP parameters (total, baseline, REM) were significantly lower in syndromic children (e.g., mean RAPtotal: 0.49 ± 0.07 vs 0.62 ± 0.10, p < 0.0001).
All mean RAP parameters were significantly higher in mild and moderate craniosynostosis as compared with severe (e.g., RAPtotal: mild 0.61 vs 0.53 p = 0.036, moderate 0.65 vs 0.54 p = 0.026). During baseline and REM sleep, the differences of ICP were not significant. In mild and moderate compared with severe abnormalities, the difference of mean RAPbase was significant (0.42 ± 0.104 vs 0.54 ± 0.117, p = 0.021) (see Fig. 3). During baseline and REM sleep, the differences of ICP were not significant.
Choosing an ICP cut-off of 20 mmHg over the total measuring time to dichotomize our data resulted in a significant difference in median as well as mean RAP during REM sleep (mean: 0.71 ± 0.10 vs 0.58 ± 0.17; p = 0.013, median: 0.76 ± 0.10 vs 0.65 ± 0.16; p = 0.031). This correlation was also true the other way around. An RAP during REM sleep under 0.6 was associated with significantly higher ICP during total monitoring time as well as REM sleep (total: 21.5 ± 6.56 vs 16.53 ± 2.76 p = 0.013; REM: 25.3 ± 8.70 vs 19.8 ± 3.54, p = 0.015).
Non-invasive imaging techniques like skull X-ray and MRI can help identify children with craniosynostosis at risk for chronically elevated ICP by describing indirect signs possibly related to raised intracranial pressure. However, images cannot replace pressure recordings, especially in cases of absent or questionable papilledema, since an as much unambiguous finding as possible should back any decision to a major surgical intervention. This study shows that ONM of ICP constitutes a safe and unequivocal tool to prove or rule out chronically or periodically (at REM sleep) elevated ICP. Similar findings showed a case series of intraoperative ICP monitoring in delayed sagittal synostosis cases and showed a rate of 81.8% cases with significantly raised ICP .
As described earlier, the cut-off values for pathological ICP monitoring are considered to lie between 15 and 20 mmHg and 3–5 plateau waves over a 24-h monitoring period [3, 12, 17, 22]. In our series, ICP values were generally high with a mean ICP over the total monitoring time of 18 mmHg.
Tamburrini et al. using intracranial pressure monitoring with calculation of derived parameters showed that the role of copper beaten appearance on skull X-ray has a low sensitivity and specificity regarding identification of children with intracranial hypertension. Furthermore, as confirmed by this study, they saw the highest diagnostic value in prolonged—not just intraoperative—ICP monitoring. The diagnostic limitation [23, 24] of skull X-ray findings to rule out pathological intracranial pressure can be re-evaluated by this study which provides a good correlation of imaging data not to ICP but to the derived parameter RAP.
RAP as an index of compensatory reserve was increased in patients with syndromic craniosynostosis, pathological X-ray findings (group 1 and group 2), and more than one premature closed suture, indicating an increasing impairment of reserve capacity. Most importantly, the most severe X-ray findings in group 3 correlated to a decrease of RAP.
One problem with skull X-rays is the shortcoming of imaging classification systems, which are always subjective as long as they do not contain measurable variables. Furthermore, more obvious impressiones digitatae are physiologically seen between five and eight years of age . In our study, the senior surgeon that scored the images was blinded to the case information and to the ICP results, so we considered the correlations found as valid, which is a relevant finding. However, in the individual patient’s case, when the surgeon is affected by clinical information, a skull X-ray information might not be sufficient to decide alone about surgery.
The fact that RAP had a good correlation to increasing skull X-ray abnormalities (none–mild–moderate) indicates that the low compensatory reserve situation, which per se results in an unphysiological intracranial “tightness” and higher ICP dynamics, is probably a major driving force for the chronic changes of the bone. This supports the use of skull X-ray as an easy screening approach to the problem.
The finding that RAP decreases or was lower in those cases with the most severe skull X-ray changes can be explained by either a compensatory effect of the thinning of the skull by gaining some additional intracranial space or creating unexpected elasticity of the skull (by creating holes in the bone), which both might increase the intracranial compliance again. In favor of this hypothesis is the fact that the external CSF spaces were more often normal in the group of patients with severe X-ray changes than in those with mild or moderate changes.
Alternatively, it has been described in TBI patients that the positive correlation of higher ICPs to higher AMP values (which results in higher RAP values) has an upper breakpoint, where the compensatory reserve is exhausted, autoregulation fails, which results in loss of correlation of ICP amplitude (AMP) to mean ICP, with a decreasing RAP despite higher ICP values .
The ICP values associated with an upper break point of RAP in TBI cases are described in the range of 19–25 mmHg . When ICP reaches a critical level the AMP-ICP relationship can thus become negative [9, 19]. Indeed, choosing an ICP cut-off of 20 mmHg over the total measuring time to dichotomize our data resulted in a significant difference in median as well as mean RAP during REM sleep. This correlation was also true the other way around. An RAP during REM sleep under 0.6 was associated with significantly higher ICP during total monitoring time as well as REM sleep.
The fact that the breakpoint-associated ICP values in group 3 were lower than in group 2—and were corresponding more to those described in TBI patients—might indicate that there is a more compromised pathophysiology as in TBI, e.g., on the venous outflow side in the specific setting of secondary or primary craniostenosis in older children. This means, that in group 2, cases reserve capacity seems to be, although already compromised, still greater and later exhausted than in group 3, so that decompensation happens at higher ICPs. In cases with none or mild radiographic features, no reliable breakpoint could be identified.
In support of the hypothesis that a lower RAP is a sign of more severe compromise of reserve capacity is the fact that syndromic children with multiple suture synostosis had more severe changes in skull X-ray, higher ICPs, and lower RAP values.
Up to this day, it is difficult to recommend a uniformly valid follow-up algorithm for children with craniosynostosis. Especially in non-syndromal craniosynostosis, no proper guidelines exist concerning the diagnosis of possible chronically elevated ICP. Long-term follow-up of nonsyndromic craniosynostosis shows that children lie within the normal range of intelligence but a discrepancy between verbal and performance IQ has been reported .
Hyward elaborated on the interaction between ICP, CPP, and obstructive sleep apnea in complex craniosynostosis in syndromal children . Even in mild craniosynostosis, high rates of developmental delay and learning disability have been shown . It remains however unclear, if those neurocognitive deficits are either related to chronically elevated ICP or to underlying genetics or even result from early ICP increases in the first months of life before initial treatment.
Thus, we conclude that ICP monitoring is able to create objective data on the existence of generally or periodically raised ICP in cases with moderate to severe skull X-ray changes. The addition of ICP amplitude derived computed information on intracranial reserve capacity (RAP Index) enhances the diagnostic gain by the identification of an exhausted status. Higher RAP especially during REM sleep can be considered a sign for a diminished compensatory capacity, but with still some reserve. Existence of a rather lower RAP in moderate to severe skull X-ray changes indicates that the reserve capacity is already exhausted, and cerebrovascular integrity affected by rising ICP, and that surgery should be performed rather earlier than later.
Compliance with ethical standards
Conflict of interest
On behalf of all authors, the corresponding author states that there is no conflict of interest.
- 12.Eley KA, Johnson D, Wilkie AO, Jayamohan J, Richards P, Wall SA (2012) Raised intracranial pressure is frequent in untreated nonsyndromic unicoronal synostosis and does not correlate with severity of phenotypic features. Plast Reconstr Surg 130:690e–697e. https://doi.org/10.1097/PRS.0b013e318267d5ae CrossRefPubMedGoogle Scholar
- 21.Ruane EJ, Garland CB, Camison L, Fenton RA, Nischal KK, Pollack IF, Tamber MS, Grunwaldt LJ, Losee JE, Goldstein JA (2017) A treatment algorithm for patients presenting with sagittal craniosynostosis after the age of 1 year. Plast Reconstr Surg 140:582–590. https://doi.org/10.1097/PRS.0000000000003602 CrossRefPubMedGoogle Scholar
- 22.Spruijt B, Joosten KF, Driessen C, Rizopoulos D, Naus NC, van der Schroeff MP, Wolvius EB, van Veelen ML, Tasker RC, Mathijssen IM (2015) Algorithm for the management of intracranial hypertension in children with syndromic craniosynostosis. Plast Reconstr Surg 136:331–340. https://doi.org/10.1097/PRS.0000000000001434 CrossRefPubMedGoogle Scholar
- 25.Wohner J, Collmann, H. (1990) Das Schädelinnenrelief bei Craniostenosen im Röntgenbild. Wuerzburg, WuerzburgGoogle Scholar
- 26.Woods RH, Ul-Haq E, Wilkie AO, Jayamohan J, Richards PG, Johnson D, Lester T, Wall SA (2009) Reoperation for intracranial hypertension in TWIST1-confirmed Saethre-Chotzen syndrome: a 15-year review. Plast Reconstr Surg 123:1801–1810. https://doi.org/10.1097/PRS.0b013e3181a3f391 CrossRefPubMedPubMedCentralGoogle Scholar