122 Surg Neurol 1989;31:122-8

Cranial Doppler Ultrasonography Correlates with Criteria for Ventriculoperitoneal Shunting

William M. Chadduck, M.D., Joanna J. Seibert, M.D., James Adametz, M.D.,

Charles M. Glasier, M.D., Mark Crabtree, M.D., and Cynthia Ann Stansell, M.D.

Departments of Neurosurgery and Radiology, Arkansas Children’s Hosptial, University of Arkansas for Medical Sciences, Little Rock, Arkansas

Chadduck WM, Seibert JJ, Adametz J, Glasier CM, Crabtree M, Stansell CA. Cranial Doppler ultrasonography correlates with criteria for ventriculoperitoneal shunting. Surg Neurol 1989;31:122-8.

Introduction

Using pulsed Doppler ultrasound through the sonic win- dow of the anterior fontanelle in infants, 46 children with ventriculomegaly had sequential evaluations of Resistive Index for comparison with the need for ventriculoperi- toneal shunting. Hydrocephalus was associated with my- elomeningocele in 17 patients and posthemorrhagic, postmeningitic, or congenital hydrocephalus accounted for the rest. The Resistive Index was obtained by dividing the differente between the peak systolic and end diastolic frequenties by the systolic frequency. Criteria for shunt- ing included massive ventriculomegaly, increasing ven- triculomegaly with an abnormal increase in head circum- ference, and clinical evidente of symptomatic hydrocephalus. Those patients requiring shunts had a Resistive Index of 84 2 13% prior to shunting. The Resistive Index fel1 to 72 ? 11% postshunting with the differente being statistically significant at p < 0.001. Moreover, the postshunting values compared favorably to data on 29 normal newborns, having Resistive Indices of 71 ? 7% and to the Resistive Indices of three patients never shunted who had values of 74%, 66%, and 71%. Variations included primarily low Resistive Indices in myelomeningocele patients whose ventricles were de- compressed by the leaking spinal lesions. Resistive Index used in conjunction with clinical observations may be extremely helpful not only in predicting the need for ventriculoperitoneal shunt, but also in evaluating chil- dren suspected of shunt malfunctions.

KEY WORDS: Pulsed Doppler ultrasonography; Ventri- culomegaly; Ventriculoperitoneal shunting; Resistive Index

The flow velocity of blood within intracranial arteries can be assessed using pulsed ultrasound through the sonic window of the anterior fontanelle in infants, or by transcranial ultrasonography in older children and adults. Because estimations of blood flow using the Doppler effect are in part dependent upon the angle of incidence of the pulsed ultrasound probe with respect to the vessei, absolute velocity measurements may be difficult to obtain. Pourcelot {B] has described an index of cerebral vascular resistance minimizing the effect of probe/vessel angle. The Resistive Index (RI) described by Pourcelot is obtained by dividing the differente between peak systolic and end diastolic frequenties by the systolic frequency. Archer et al [l) have demon- strated an excellent correlation between RIS of neonatal cerebral arteries and cerebral vascular resistance. Sei- bert et al 191 have described normal ranges of RIS in neonates as wel1 as the correlation of abnormally high RIS with a variety of abnormal clinical conditions, including patent ductus arteriosus, brain edema, cardiac ischemia, seizures, asphyxia, pneumothorax, and intra- cranial pathological processes. Daneman et al [4] have also found reproducible values for RIS in neonates and have described a number of clinical conditions associ- ated with abnormally high values for RI. Finally, Hill and Volpe {bl have demonstrated a decrease in pulsatile flow in the anterior cerebral arteries in infantile hydro- cephalus. The present study correlated classica1 clinical criteria for hydrocephalus with RI, determining by transfontanelle pulsed Doppler measurements, predic- tive indicators of the need for ventriculoperitoneal shunts in infants with ventriculomegaly.

Presented at the Pediatrie Sectional Meeting of the American Association of Neurological Surgeons, Chicago, Illinois, December 8-11, 1987.

Patients and Methods

Address reprint requests to: William M. Chadduck, M.D., Depart- ment of Neurosurgery, Arkansas Children’s Hospital, 800 Marshall, Little Rock, Arkansas 72202.

Received June 23, 1988; accepted September 21, 1988.

Using an Acuson 128 sector 519 scanner, 46 neonates with ventriculomegaly were evaluated on two or more occasions before and after ventriculoperitoneal shunt- ing. Three additional patients with mild ventri-

8 1989 by Elsevier Science Publishing Co., Inc. 0090-3019/89/$3.50

Cranial Doppler Ultrasonography Surg Neurol 123 1989;31:122-8

Figure 1. Transfontanelle parasagittal real-time rrhrasonogran of an infant with severe bydroceph- ah5 shows placement of a cursor in the region of the proximaal pericallosal brancb of tbe left anterior cerebral artery. Pulsed Dopp.!er studies obtained from tbis region of tbe vessel, indirated an RI of 131%. Tbe dotted line indicates tbe direction of tbe pulsed ultrasound, and tbe region of tbe vessel assessed lies between tbe two parallel horizontal Lines.

culomegaly but not requiring shunts were also studied with ultrasound evaluations. Resistive Indices were obtained on 628 vessels, and the values of the anterior and middle cerebral arteries and internal carotid arteries bilaterally were averaged. Real-time ultrasound studies of ventricular size and pulsed Doppler measurements of RI were made through the sonic window of the anterior fontanelle. Real-time ultrasound localization of each vessel allowed placement of cursors for measurement of flow velocities at diesignated regions of each artery to be studied. The anterior cerebral arteries (Figure 1) were visualized in a parasagittal view, whereas the middle cerebral and internal carotids were best seen in the corona1 plane. Hydrocephalus was associated with my- elomeningocele in 17 patients, and posthemorrhagic, postmeningitic, or congenital hydrocephalus accounted for the rest. One child had a Galen’s vein aneurysm in addition to hydrocephalus, but none had any other factors (patent ductus arteriosus, cardiac abnormalities, pneumothorax, abnormalities of arterial blood gases) known to affect the RIS. Al1 shunted patients received Pudenz-Schulte medium pressure ventriculoperitoneal shunt systems. The criteria for shunting included massive ventriculomegaly, increasing ventriculomegaly with an abnormal rate of increase in head circumfer- ence, and/or clinical evidente of symptomatic hydro- cephalus (lethargy, bradycardia, and tense fontanelle), oftentimes with clinical improvement following ventric- ular tapping. In those infants needing ventricular or lumbar taps, intracranial pressure and RIS were mea- sured before and after the taps. Head circumference measurements were plotted on standard charts, using

data from the National Center for Health Statistics percentiles CS]. Ventriculomegaly was assessed by high resolution transfontanelle ultrasonography and/or com- puted tomography (CT) scanning of the brain.

Statistical differences between groups were analyzed by the chi-square method.

Results One hundred and ninety-six examinations were per- formed on 46 infants having ventriculomegaly associ- ated with myelomeningocele intraventricular hemor- rhage, postmeningitic hydrocephalus, or congenital hydrocephalus. The three patients never shunted had RIS of 74%, 66%, and 71%. Moreover, serial evalu- ations of ventricular size and head circumference showed the patients to have stable ventriculomegaly and changes in head circumference conforming to normal growth curves, and their RI values compared favorably to a study of 29 full-term infants, judged to be normal by the neonatologists, who had RIS of 7 1 ? 7%. Figure 2 shows data on patient JB, a myelomeningocele patient who has not required shunting after more than a year of follow-up. As shown in Figure 2, sequentia1 high resolu- tion ultrasound studies of the ventricles did not show significant enlargement. Head circumference measure- ments on the standard growth chart increased at a normal rate and remained in a normal range. The RIS also remained in a normal range.

By contrast, the average RI of patients requiring ventriculoperitoneal shunts just prior to the shunting procedure was 84 & 13% and fel1 to 72 ? 11% after

124 Surg Neurol 1989;31:122-8

Chadduck et al

Figure 2. (A) Initial high resobtion nltraso- nogram, parasagittal uiew, of the uentrict’e of JB, a myelomeningocele patient, shows a ventricle of flor- mal size. At the time of this study, the head circumference was in the 25th percentile, and the Rl averaged 74%. (B) Cranial sonogvam, parasa- gittal uiew, of tbe ventricle of tbe same patient at age 5 montbs shows mild uentriculomegaly associ- ated witb an RI of 71%. Tbe cbild’s bead circum- ference of 40.5 cm remained at tbe 29th percentile. At 1 year of age tbe cbild bad not required a shunt.

shunting. These differences are statistically significant increased from 68% to 82% and then dropped to 76%

(p < 0.001). Not only did the RI values significantly postshunting. Table 1 summarizes the results. drop following shunting, but they also fel1 to a range equal to those of normal infants (71 + 7%). Figure 3 shows sequentia1 high resolution ultrasound studies of Discussion the ventricles of LL, a myelomeningocele patient whose The RI values in this series of patients having ventri- ventricles showed progressive enlargement, whose head culomegaly clearly correlated with the ultimate need for circumference accelerated abnormally, and whose RIS ventriculoperitoneal shunting. An impressive cor-

Cranial Doppler Ultrasonography Surg Neurol 1989;31:122-8

125

B

relation was found comparing the RI of normal full- term infants, patients not requiring shunts, and patients following treatment with shunts, al1 being in a range significantly lower than those of infants with clinically significant untreated hydrocephalus. It should be noted, however, that although the average values between the groups are different to a statistically significant degree, the ranges of RIS found in each of the groups necessi- tates that clinical correlations account for the variations. The most consistent variation was found in myelomen-

Figure 3. (A) High resolution cranial ultra- sound, parasagittal view, of LL made at the time of myelomeningocele closure shows mild ventri- culomegaly. The RI was 68%; the head circumfer- ence was 34 cm (50th percentilei. (B) High resolu- tìon ultrasonic assessment of ventricular size of patient LL made prior to sbunting shows dramatic uentricuiomegaly. The RI was 82%, tbe bead cìrcumference 37 cm, (81th percentile). (C) (next page) Postoperative uhrasonogram, para- sagittal view, of tbe same patient shows a marked decrease in tbe size of tbe ventricle. Tbe byperecboic cboroid plexus and ventricular catheter are seen. Tbe RI at tbe time of tbis study bad decreased to 75%.

Table 1. Resuh of Patients Requiring Ventriculoperitoneal Shunts

Clinical status Number of patients Resistive Index

Normal Ventriculomegaly

nor requiring shunt Ventriculomegaly

prior to shunting Ventriculomegaly

after shunting

29 71 * 7% 3 70 + 4%

43 84 t 13%’

43 72 -- 11%

126 Surg Neurol 1989;31:122-8

Chadduck et al

ingocele patients who had measurement of RIS prior to closure of leaking myelomeningoceles. Despite the fact that these infants occasionally had dramatic ventri- culomegaly with abnormally large head circumferences, the RIS preshunting were normal. We atrribute this to the fact that the ventricular system was decompressed by the leaking myelomeningocele. Many of the myelo- meningocele patients had simultaneous myelomen- ingocele repair and placement of ventriculoperitoneal shunts if they had marked ventriculomegaly or large ventricles with high RIS. One child with a leaking encephalocele had a higher RI after shunting than before, but the postshunting values were normal, 72%, and the discrepancy was explained on the basis of ventricular decompression through the leaking enceph- alocele prior to its closure and simultaneous shunting. In another instance, an infant, CI, had no decrease in RIS after placement of a shunt, but was found to have an isolated ventricle due to megachoroid and required subsequent placement of a contralateral shunt. Another child with a nonfunctioning ventriculoperitoneal shunt had normal RIS but had tracking of the ventricular fluid al1 along the shunt tract and a failure of the pumping assembly to refill. The decompression of the ventricular system along the shunt tract was presumed to account for the normal RIS. In two other instances, we were unable to account for imperfect correlations.

An example of the sensitivity of the technique is given by the case of TD, a newborn with an intraventri- cular hemorrhage who had marked hydrocephalus. The RIS were elevated, averaging 86%. Ten days later, the

RIS had decreased to normal values associated with a program of intermittent spinal drainage. A recurrence of elevated RIS correlated with disproportionate en- largement of the head and progressive ventri- culomegaly, tense fontanelle, and episodic bradycardia. After shunting, RIS fel1 to 75%, but later increased to 91%. Because the patient had an elevated ventricular fluid protein concentration and the pump assembly functioned normally, a program of pumping the shunt 40 times per shift was elected. One month later, the head circumference had increased only 1/2 cm and RIS had fallen to the normal range (76%).

Vascular factors may also affect the RI. For example, DJ, a newborn with high output cardiac failure caused by a Galen’s vein aneurysm (Figure 4), had surgical ligation of the lesion. The immediate postsurgical RI was normal, 62%, having been extremely low (24% and 36%) due to extremely low resistance in the anterior cerebral arteries prior to closure of the arteriovenous fistula. Two weeks later, when ventriculomegaly was noted, the question arose as to whether or not the patient had encephalomalacia or hydrocephalus. Se- quential Doppler studies showed a gradual increase in RI to SO%, associated with tense fontanelle and pro- gressive enlargement of the ventricles. Bilateral ventri- culoperitoneal shunts were placed, and the RI returned to a normal range of 62% prior to the child’s discharge from the hospital. Again, an excellent correlation was obtained, but the findings in this patient confirm that fact that vascular factors such as arteriovenous fistulae may effect the RIS. Black et al [3] have used Doppler

Cranial Doppler Ultrasonography Surg Neurol 1989;31:122-

127

Figure 4. (A) A preoperative right lateral cere- bral angiogram of the patient uith the Galen’s uein aneurysm demonstrates the lesion. RI averaged only 30%, consistent with high Jow through the arteriovenous fistula. 03) Postoperative cerebral angiogram of the patient with a Galen’s vein aneurysm shows ventriculomegaly, evident by the sweep of the anterior cerebral arteries, and the presence of uentricular shunts. At the time of thìs study, RIS bad returned to an immediatepostopera- tive vahe of 629~0, bauing risen to 80% witb progressiue enlargement of tbe ventricles.

B

imaging of cerebral arteriovenous malformations as a surgical adjunct. Moreover, in older patients having obstructive cerebral vascular lesions, the Doppler flow velocities vary with the degree of stenosis [73. Among the patients included in this paper, al1 but this one were presumed to have normal cerebral vascular anatomy.

Another child, CE, with intraventricular hemorrhage was managed by repeated lumbar punctures and ven-

tricular taps. The RIS were used as an indicator of the need for drainage procedures, being in the range of 88% prior to tapping. The RI dro ,ped to a range of 74%-76% following drainage. Stab,e ventriculomegaly and periventricular leukomalacia were documented, and the RIS remained in a normal range. At 5 months of age, the RI was 66% without shunting. It remains impera- tive, therefore, to allow for the course of the evolving

Chadduck et al 128 Surg Neurol 1989;31:122-8

pathology to be treated by standard means before considering a shunt procedure, but the effectiveness of that treatment program can be monitored using the RI. This is especially true in patients with intraventricular hemorrhage, whereas those having hydrocephalus asso- ciated with myelomeningocele would not have a rapidly changing pathophysiologic process that would affect the need for a shunt.

It has been pointed out by Bel1 et al [2] that ventriculomegaly in the newborn with myelodysplasia should be evaluated by other means than palpating the anterior fontanelle and measuring the head circumfer- ence. Because those traditional evaluations may be misleading, they have suggested a lateral ventricular ratio obtained from ultrasonograms to predict those children developing hydrocephalus and therefore re- quiring shunts. The use of pulsed Doppler studies to determine RI adds stil1 another noninvasive means of predicting the need for shunts in myelodysplastic pa- tients.

The correlation with RI and the need for shunting has been so reliable in this study that we have also used it for the evaluation of children suspected of shunt malfunctions. Preliminary studies indicate that the RIS of older children are significantly lower than those of the newborn. We are, therefore, attempting to establish age-related normal values to compare with those of patients having functioning ventriculoperitoneal shunts, and to evaluate how RI values may be altered in those patients ultimately requiring shunt revision.

At times, the decision for shunting, especially in a patient with myelomeningocele or posthemorrhagic ventriculomegaly, requires sequential evaluations over a period of time to document progressive ventri- culomegaly, and therefore progressive insult to the bram. If indeed the RI can be predictive of the need for ventriculoperitoneal shunting, the appropriate decision

may be made earlier, presumably with benefit to the patient. Our results thus far indicate that the RI, obtained in the absente of other significant hemody- namic abnormalities, is a reliable indicator of the need for shunting and one that we can recommend in the management of infants with ventriculomegaly.

We would like to thank Sandi L. Corbitt and Alyson B. Rodgers, ultrasound technicians, for their expertise in performing the doppler studies.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

Archer LNJ, Evans DH, Paton JY, Levene MI. Controlled hyper- capnia and neonatal cerebral artery Doppler ultrasound wave forms. Pediatr Res 1986;20:218-2 1.

Bel& WO, Sunmer TE, Volberg FM. The significante of ventri- culomegaly in the newborn with myelodysplasia. Child’s Nerv Syst 1987;3:239-41.

Black KL, Rubin JM, Chandler WF, McGillicuddy JE. Intraopera- tive imaging of AVMs and aneurysms. J Neurosurg 1988;68: 635-9.

Daneman A, Soto G, Hellmann J. Doppler sonography of intra- cranial vessels in neonates. Presented at the The Society for Pediatrie Radiology, Washington, D.C., April 1986.

Hammill PVV, Drizd TA, Johnson CL, Reed RB, Roche AF, Moore WM. Physical growth: National Center for Health Statistics percentiles. Am J Clin Nutr 1979;32:607-29.

Hill A, Volpe JJ. Decrease in pulsatile flow in the anterior cerebral arteries in infantile hydrocephalus. Ped 1982;69:4-7.

Mattle H, Grolimund P, Huber P, Sturzenegger M, Zurbrugg, HR. Transcranial doppler sonographic findings in middle cerebral artery disease. Arch Neurol 1988;45:289-95.

Pourcelot L. Applications cliniques de l’examen Doppler transcu- tane. In: Peronneau P, ed. Velocinietre ultrasonone Doppler. Paris. Institut National de la Sante et de la Recherche Medicale, 1975:213.

Seibert JJ, McCowan TC, Chadduck WM, Glasier CM, Williamson SL, Taylor BJ, McConnell JR, Rodgers AB, Corbitt SL, Adamett IR. Leithiser RL Tr. Clinical and exnerimental studies of correlation between duplex”pulsed doppler and intracranial pressure in the neonate. Radiology (in pres+