| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Pediatric Imaging |
1 From the Department of Radiological Sciences (S.J.W., J.W.S.), Divisions of Plastic Surgery (M.A.S., H.K.) and Neurosurgery (J.L.), and Department of Anesthesiology (C.T.M.A.), UCLA School of Medicine, Los Angeles, Calif. From the 1996 RSNA scientific assembly. Received February 23, 2000; revision requested April 8; final revision received July 20; accepted August 15. Address correspondence to S.J.W., Department of Radiology, Children’s Memorial Hospital, 2300 Children’s Plaza, Box No. 9, Chicago, IL 60614-3394 (e-mail: SWestra@childrensmemorial.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Transcranial Doppler US was performed through the temporal squama, fontanels, and existing skull defects prior to and immediately following cranioplasty. Twenty-four studies were performed in six patients with multisuture synostosis, and 61 studies were performed in 26 patients with single-suture synostosis. Study findings were compared with those of 23 control subjects and were characterized as normal or abnormal on the basis of age-specific normal criteria for RI.
RESULTS: In multisuture synostosis, results of six of the nine preoperative transcranial Doppler US studies were abnormal. During postoperative follow-up, three recurrences requiring reoperation occurred, one of which was detected with abnormal transcranial Doppler US findings. In single-suture synostosis, results of seven of the 26 preoperative transcranial Doppler US studies were abnormal, and all occurred in young infants with sagittal and unicoronal synostosis. Immediate effects of surgery were variable. All patients with sagittal synostosis had a significant immediate postoperative increase in RI, which normalized during postoperative follow-up. There was no significant difference in RI between patients with successfully treated craniosynostosis and control subjects.
CONCLUSION: Transcranial Doppler US can be used to identify patients with craniosynostosis with decreased intracranial compliance, and it is a suitable noninvasive test to monitor the effects of surgery on compliance.
Index terms: Brain, blood flow, 10.761 • Brain, growth and development, 10.143, 10.145 • Brain, US • Skull, abnormalities, 10.143 • Skull, growth and development, 10.143 • Ultrasound (US), Doppler studies • Ultrasound (US), in infants and children
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
However, there is uncertainty in the literature (6,7) as to what constitutes elevation of ICP in infants and young children. The pathologic relevance of long-standing low-grade elevation of ICP, as may be seen in some patients with single-suture synostosis, is also unclear (8). Most of the plastic development of the skull occurs before the age of 2 years, and it is assumed that early intervention has the best results, both for reaching acceptable aesthetic results and for the prevention of neurodevelopmental sequelae (9).
Clinical and radiologic criteria for monitoring the effects of mildly elevated ICP are unreliable (9–11). Recording of abnormal visually evoked potentials before the onset of papilledema has been proved helpful for early identification of surgical candidates (12), but to our knowledge no noninvasive test is currently available to directly measure ICP.
Transcranial placement of pressure transducers in the extradural space for prolonged ICP monitoring has been used by some authors (4,5,13,14) in the surgical decision-making process. Others (15) have used pressure measurements during lumbar puncture with fluid infusion to identify patients with pathologic cerebrospinal fluid (CSF) outflow resistance. These tests have not been widely accepted due to their invasive nature and are therefore not suitable as screening tests.
Transcranial Doppler ultrasonography (US), with calculation of the resistive index (RI) of basal cerebral arteries, is an accepted noninvasive method to quantify perfusion of intracranial arteries in adults (16,17) and children (18–21). Several authors (1,22,23–25) have described the use of RI to study the relationship between cerebral perfusion pressure and ICP, to investigate intracranial compliance, and to study the effects of treatments aimed at decreasing ICP and improving brain perfusion. A further refinement of transcranial Doppler US involves the short application of light pressure over the acoustic window with the transducer, thereby momentarily increasing ICP, with simultaneous monitoring of the Doppler waveform (23). This pressure provocation test has increased the accuracy of transcranial Doppler US to identify elevated ICP. To our knowledge, it is the only noninvasive test that is clinically available to assess dynamic intracranial volume-pressure relationships (compliance).
We hypothesized that decreased intracranial compliance, as measured with transcranial Doppler US, is common in infants and children undergoing surgery for craniosynostosis. We expected the highest prevalence of compliance abnormality in patients with multisuture synostosis. In repair procedures for coronal, metopic, lambdoid, and multisuture synostosis, existing compression on the brain is released, whereas in sagittal synostosis repairs, preexisting brain compression may temporarily worsen. We studied the use of transcranial Doppler US to evaluate these perioperative alterations in brain compression and intracranial compliance.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The study group consisted of six children with multisuture synostosis (group 1) and 26 children with single-suture synostosis (group 2). Group 1 consisted of only boys, three with Apert syndrome (11 studies), one with Crouzon syndrome (two studies), and two with nonsyndromic cloverleaf deformity of the skull (11 studies); age range at initial study was 35 days to 19 months (median age, 4 months). Group 2 consisted of 14 boys and 12 girls; age range at initial study was 3–36 months (median age, 7.5 months). Group 2 was further subdivided as follows: sagittal suture synostosis (group 2a, n = 13; 32 studies) and other single-suture synostosis (group 2b; unilateral coronal suture synostosis, n = 6, 13 studies; bilateral coronal suture synostosis, n = 3, eight studies; metopic suture synostosis, n = 2, four studies; and unilateral lambdoid synostosis, n = 2, four studies).
Control Subjects
There is uncertainty in the literature about the upper limit of normal for RI values in the age ranges of our study groups, as investigators (3,21–24,26) of most published series report values obtained in healthy premature or term newborn infants. Therefore, we also examined a control group of 23 subjects with normal ICP by clinical criteria. The control group consisted of subjects with adequately shunted hydrocephalus (Arnold-Chiari malformation, n = 5; aqueductal stenosis, n = 2), other brain malformation (such as agenesis of corpus callosum, n = 3), cerebral atrophy (n = 4), benign external hydrocephalus (n = 3), the cystic stage of periventricular leukomalacia (n = 1), prior successful drainage of subdural hygromas and embolization of dural arteriovenous fistula (n = 1 each), or no demonstrable intracranial abnormality (n = 3).
The age distribution of our control group was 11–90 days in 11 subjects, 3–11 months in six, and 1.0–2.9 years in six. All postoperative studies in patients with craniosynostosis who did not exhibit signs or symptoms of elevated ICP were compared with those of the control population. The age distribution in this postoperative group was 11–80 days in 10 patients, 3.0–11.9 months in 35, and 1.0–2.9 years in 16.
Transcranial Doppler US: Pre- and Postoperative Studies
All transcranial Doppler US studies were performed by the same author (S.J.W.). Initially, a screening real-time gray-scale examination of the brain was performed to determine the presence of ventriculomegaly. The anterior cerebral artery was insonated through the greater fontanel in all patients in whom this acoustic window existed. In older children with closed fontanels, any existing postoperative craniotomy defect, in combination with transtemporal scanning as described by Seibert et al (27), was used. A 3-MHz phased-array transducer with a small rectangular footplate, operating at a Doppler frequency of 2.3 MHz, was used for all studies.
By using color Doppler flow imaging, the pericallosal branch of the anterior cerebral artery was identified in the midsagittal plane, and this vessel was subsequently insonated with spectral Doppler US at its vertical course immediately anterior to the genu of the corpus callosum. The insonation angle was kept at less than 60° during the recording of the spectral waveform, and angle correction was not used. The Doppler sample size and wall filter were set at their lowest levels, and the pulse repetition frequency (velocity scale) and baseline were set to optimally display the spectrum over the full height of the scale without aliasing. After a stable waveform was obtained, the image was frame-frozen, and the RI was determined, with cursors identifying the peak systolic velocity, PSV, and the end-diastolic velocity, EDV. The RI was expressed in percentage points by using the following equation: RI (percentage) equals [(PSV minus EDV) divided by PSV] multiplied by 100.
In 54 of the 85 craniosynostosis studies and in 14 of the 23 control studies, the RI measurement was subsequently repeated during the application of uniform continuous pressure with the transducer on the acoustic window for up to 5 seconds while the Doppler spectrum was being recorded. From these recorded spectra, the pressure resistive index (PRI) was obtained. The instantaneous effects of pressure application and release on the Doppler spectral waveform were also documented.
In addition to anterior cerebral artery sampling, Doppler spectra were obtained in 62 of 85 craniosynostosis studies from one or both middle cerebral arteries through the temporal squama. Pressure provocation was performed in these studies by means of digital application of pressure on the greater fontanel when patent or on an available (surgical) skull defect when the fontanels were closed, while the middle cerebral artery Doppler spectrum was simultaneously being recorded.
Transcranial Doppler US: Intraoperative Studies
Transcranial Doppler US was performed intraoperatively in four of the six group 1 patients (seven studies) and in 25 of the 26 group 2 patients (25 studies). A preoperative study was first performed, with the patient under general anesthesia, through the intact skin in patients in whom an open fontanel or surgical skull defect existed. In patients without accessible acoustic windows, a variety of measures were used. During the initial months of our study, the anterior cerebral artery was interrogated through a small midline burr hole through the vertex, which was made by the attending neurosurgeon as part of the initial stages of the cranioplasty procedure. Later in the study, transtemporal scanning of the middle cerebral arteries was performed. Immediately after the performance of the sutural release and calvarial remodeling, studies were repeated through available skull defects, with the transducer surface placed on the intact dura.
In 19 of 32 intraoperative studies, we performed pressure provocation by compressing the intact dura with the ultrasound transducer, and the pressure applied was subjectively rated by the operator as being similar to that used in patients who were examined with the conventional transfontanellar technique. In six studies, immediate postoperative measurements were performed after closure of the skin, with the patients still under general anesthesia, or in the postoperative recovery room.
Statistical Analysis: Reference Standard
At least three RI measurements were obtained at baseline (RI) and during pressure application (PRI). The RI and PRI values of each vessel were averaged, and the SDs were calculated. The values of both middle cerebral arteries were compared with one another and with that of the anterior cerebral artery to determine whether lateralization of RI and PRI abnormalities existed. For each study without lateralization, a weighted mean of the RI and PRI was calculated by averaging the obtained RI and PRI values for all examined basal intracranial arteries. These values formed the basis for our comparison between patient groups. The differences between mean pre- or postoperative RI or PRI values obtained in the study groups and those obtained in the control group were expressed as P values, which were calculated by using a Student t test. Within the study groups, the differences between RI or PRI values obtained before and immediately after surgical intervention and between the latter and those obtained at postoperative follow-up were expressed as P values, which were calculated by using a paired t test of repeated measurements. Statistical significance was set at a P value of less than .05.
We used the age-specific normal RI data published by Bode and Wais (28) and Bode (29) as our reference standard. To determine the cutoff points between normal and abnormal RI values, we made the assumption that an abnormally low RI had no relevance for the purpose of our study, which was to identify patients with an abnormally decreased intracranial compliance. By setting the cutoff point for RI at the population mean plus 1.645 SDs, we designated 5% of the normal population as having an abnormally elevated RI (type I error, 
= .05).
We compared the values of Bode and Wais and Bode with those obtained in our control population and tested for statistical significance of differences between both populations at the P value of less than .05 by using the Student t test. To our knowledge, since no normative data are published in the literature for PRI, we determined the upper limit of normal for the three age groups on the basis of the values of our own control population. These cutoff points were subsequently applied, in conjunction with the cutoff points for baseline RI derived from data of Bode and Wais and Bode, to determine normality or abnormality of results of transcranial Doppler US studies in individual patients with craniosynostosis. Findings from studies in which the pressure provocation test was performed in addition to baseline recordings were considered abnormal when abnormal PRI values were obtained, irrespective of the baseline RI values.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
|
|
At postoperative follow-up, no abnormal RI values were encountered in this group. The patient with unilambdoid synostosis had normal RI values preoperatively, but immediately following repair mild ventriculomegaly was noted, and an abnormal RI was recorded from the ipsilateral middle cerebral artery and the anterior cerebral artery, with a normal contralateral middle cerebral RI value. No postoperative follow-up study findings were available in this patient. With the exception of this patient, no lateralization of RI abnormalities in the middle cerebral arteries was found in the subgroup of eight patients who underwent unilateral repairs. Ventriculomegaly was noted preoperatively in two patients with bicoronal synostosis who were older than 1 year, and this resolved following cranioplasty in one patient, whereas additional ventriculoperitoneal shunt placement was required in the other.
In infants younger than 1 year, no significant difference in RI values was found between those of the populations of patients with craniosynostosis in whom surgery was successful, those of our control population, and the normative data of Bode and Wais (28) and Bode (29). In children older than 1 year, there was no significant difference in RI values between patients with craniosynostosis in whom surgery was successful and control subjects, but mean RI values of both groups were significantly higher than those in the corresponding control group reported by Bode and Wais and Bode (Table 1).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Shapiro et al (31) directly examined intracranial compliance in hydrocephalic infants by injecting small aliquots of fluid through recently placed ventricular drains and measuring simultaneously occurring changes in ICP. Hydrocephalus and herniation of the hindbrain are not infrequent complications of severe craniosynostosis (32,33). Experimental study findings in an animal model for unicoronal synostosis have confirmed that CSF conduction and resorption abnormalities may be responsible for regional underperfusion of the brain and thereby may play a role in the pathogenesis of brain damage (34). Lumbar infusion study findings have confirmed the presence of abnormal CSF dynamics in patients with craniosynostosis, especially in the more severe forms (15).
Using epidural pressure monitors, Renier et al (4) and Renier (5) have shown a high prevalence (47%) of elevated ICP in patients with multisuture synostosis, but even in the single-suture synostosis group, ICP was elevated in a substantial number (14%) of patients. Elevated ICP in sagittal synostosis and unicoronal or unilambdoid synostosis was more frequently seen after 1 year of age than before, reflecting the fact that early in life the remaining open sutures were able to accommodate for brain growth. In bicoronal synostosis, elevated ICP, when present, may manifest earlier in life (4,5). After the period of rapid brain growth, which is virtually complete at the end of the 2nd year (9), no further increase of the prevalence of elevated ICP occurs. At that time, an equilibrium between volume and pressure will have been reached, and older children who were not treated in a timely manner may actually have a normal cranial volume for age (35), despite marked deformities.
In our single-suture synostosis study group, the incidence of elevated preoperative RI was 27% (seven of 26 patients), and abnormal values were recorded in four of 13 patients with sagittal synostosis and in three of six patients with unicoronal synostosis. Five of these seven patients were younger than 6 months, and the remaining two were 15 and 20 months of age. Rifkinson-Mann et al (36) described the transcranial Doppler US test in single-suture craniosynostosis and found a high (70%) prevalence of abnormal RI. Since these authors did not specify their cutoff points for abnormal RI, and apparently did not use pressure provocation, it is difficult to compare their results with ours.
In infants and children with craniosynostosis, there is controversy as to the indication for and the appropriate timing of surgical repair (4–6,9,13,35). True primary craniosynostosis must be differentiated from craniosynostosis secondary to microcephaly, benign posterior plagiocephaly due to positional molding, and normal variants (dolichocephaly in preterm infants), which do not require surgery (37). In infants with multisuture synostosis, where continued brain growth is compromised by the fused calvarium, there is general agreement that early surgical intervention is indicated.
The observation of regional cerebral perfusion and metabolic abnormalities depicted on single photon emission CT and positron emission tomographic scans in children with single-suture synostosis (38,39) has prompted some to surgically treat these not solely for aesthetic reasons but also to ensure normal brain development. However, in this group, a direct relationship between the presence or severity of the abnormality and associated neurologic sequelae is not proved (8,40–42). This issue has been compounded by worrisome reports in both the lay press and scientific literature about the overdiagnosis of isolated lambdoid synostosis, resulting in unnecessary operations having been performed in infants with benign positional deformities (40, 43,44). Transcranial Doppler US with pressure provocation may, in a noninvasive way, be used to screen patients with mild skull deformities for intracranial compliance abnormalities, which would strengthen the indication for surgery in those patients.
With this study, we made interesting observations regarding the relationship between the type of synostosis, the nature of the surgical intervention, and the RI values. The fact that the largest number of patients with abnormal RI values was found in the multisuture synostosis group was to be expected. There was a partial normalization of the RI values following surgery, but a small number of patients with multisuture synostosis developed a recurrence and required further surgery. Transcranial Doppler US appears to be an ideal noninvasive test to screen these patients periodically (Fig 2).
In some patients with single-suture synostosis, we observed a sudden increase in RI immediately following cranioplasty. This may be explained by the abrupt changes in biomechanical relationships within the skull resulting from the surgical repair. It also appears likely that the insult of the surgery may actually unmask clinically silent or aggravate subtle intracranial compliance abnormalities. The most profound immediate postoperative changes were seen in the sagittal synostosis group, and these changes included a large increase in RI, to the point of absence or reversal of diastolic flow in a substantial number (31%) of patients (Fig 3).
Most surgeons use a variation of the so-called Pi procedure, during which a 
-shaped skull segment involving both coronal sutures and the central portions of both parietal bones is removed. The strip of bone containing the sagittal suture is shortened anteriorly and reattached to the frontal bones, leading to reduction of the anteroposterior skull diameter (7,45). The immediate effects of this procedure are well known to craniofacial surgeons: The brain (dura) feels more taut, is less compliant, and can be seen to bulge out laterally. As an indirect sign of the acutely elevated ICP during this procedure, the anesthesiologist may note a transient bradycardia due to increased vagal tone (Cushing reflex).
Transcranial Doppler US adds a level of sophistication in intraoperative monitoring to that derived from direct palpation of the dura and has in several of our patients directly influenced surgical treatment (Fig 3). The method of repair performed in patients with other types of cranial deformity typically does not exert the same degree of clinically apparent intradural compression as that seen with the Pi procedure. This clinical impression is borne out by our data.
The fact that diastolic flow reversal occurred so frequently during the sagittal repair procedure is of some concern, as the brain parenchyma is dependent on forward flow during systole and diastole for tissue oxygenation and metabolic activity. Because these operations are performed primarily for aesthetic considerations, any risk of ischemic brain damage during the intraoperative period is to be avoided. However, we found these changes to be only transient, as they had disappeared at follow-up. Any untoward effects of this short period of diastolic flow reversal on neurologic development are difficult to prove. Clinical outcomes generally have been described (46,47) as excellent in this group of patients. Prognosis for neurodevelopmental outcome in multisuture (syndromal) synostosis is more guarded. The development of mental retardation, especially in patients with Apert syndrome, appears to be unrelated to the timing and type of surgical intervention, for which a primary brain malformation may be implicated (4,6,7).
Our study has a number of limitations that result from our retrospective research design. Follow-up was not performed for all abnormal intraoperative transcranial Doppler US results, and the timing of the performed postoperative studies was variable. These factors may have introduced a selection bias into our study results.
A more fundamental limitation is that no standard was used for the determination of ICP and compliance abnormalities, since pressures were not measured directly. Transcranial Doppler US is used to assess ICP and compliance only indirectly, and many additional hemodynamic (ie, cardiac output, arterial compliance, hyperventilation, effects of general anesthesia) and intracranial parameters (presence of hydrocephalus and other signs of CSF conductance and resorption abnormality) may influence transcranial Doppler US results. We tried to circumvent the problem of the lack of a reference standard by studying a control group in conjunction with our study group and by referring to published normative RI data.
Because we were not able to compare our results with a reference standard, we could not determine the sensitivity of our method. Due to our setting of the RI cutoff criteria on the basis of published normative data, our specificity may be estimated at 95%, but again we could not determine specificity directly against an independent reference standard.
Nonetheless, we believe that results of this study have demonstrated the feasibility and clinical applicability of perioperative transcranial Doppler US in patients with craniosynostosis. A future prospective study in which a comparison of RI without and with pressure provocation is made, in correlation with intracranial volume measurements on CT scans and epidural pressure measurements without and with CSF infusion, might further clarify the complex relationship between intracranial volume, pressure, and CSF conduction abnormalities that exist in patients with craniosynostosis.
In conclusion, the indications and appropriate timing for craniosynostosis surgery remain controversial, and transcranial Doppler US can fill the clinical need for a noninvasive screening and monitoring test for ICP and compliance in patients undergoing craniosynostosis repair. Our study findings confirm findings of previous investigations that global ICP in single-suture craniosynostosis is elevated in a substantial number of patients. In most patients with sagittal synostosis and in a substantial number of patients with unicoronal and multisuture synostosis, cranial vault surgery may lead to transient aggravation of the intracranial compliance abnormality. Within a few hours to days after surgery, a new equilibrium will have been established, characterized by normalization of ICP, compliance, and perfusion.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, S.J.W.; study concepts, all authors; study design, S.J.W.; definition of intellectual content, all authors; literature research, S.J.W.; clinical studies, S.J.W., M.A.S.; data acquisition, C.T.M.A., H.K., S.J.W., M.A.S.; data analysis, all authors; statistical analysis, S.J.W., J.W.S.; manuscript preparation and editing, S.J.W.; manuscript review, H.K., M.A.S., J.L.; manuscript final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
= baseline RI values, 
= RI values with pressure provocation, arrows = surgical interventions. (a) Early cranial vault remodeling was indicated in this infant with severe intracranial hypertension and hydrocephalus. Abnormal preoperative transcranial Doppler US findings normalized postoperatively, but at the age of 3.5 months an abnormal pressure response was noted at transcranial Doppler US, and a second skull remodeling procedure was performed at age 8 months because of signs of intracranial hypertension and stagnation of psychomotor development. This operation resulted in marked aesthetic improvements and a normal neurodevelopmental outcome at 20 months, and RI and PRI values normalized immediately postoperatively and remained normal. (b) Two consecutive cranial vault operations were required at ages 6 and 9 months in an infant with hydrocephalus, which progressed between the operations but which was no longer present after the second operation. The first operation led to improvement of compliance abnormality, as detected preoperatively at transcranial Doppler US, whereas the second operation was associated with an abnormal immediate postoperative RI, which normalized at follow-up.
