Cerebrospinal fluid drainage in ... - Wiley Online Library

Acta Pædiatrica ISSN 0803–5253

REGULAR ARTICLE

Cerebrospinal fluid drainage in posthaemorrhagic ventricular dilatation leads to improvement in amplitude-integrated electroencephalographic activity Monika Olischar ([email protected])1,2 , Katrin Klebermass1 , Barbara Hengl1 , Rod W Hunt2 , Thomas Waldhoer3 , Arnold Pollak1 , Manfred Weninger1 1.Division of General Pediatrics and Neonatology, Department of Pediatrics and Adolescent Medicine, Medical University of Vienna, Austria 2.Department of Neonatology, The Royal Children’s Hospital, Melbourne, Australia 3.Department of Epidemiology, Center for Public Health, Medical University of Vienna, Austria

Keywords Amplitude-integrated electroencephalography (aEEG), Cerebral function monitor (CFM), Posthaemorrhagic ventricular dilatation (PHVD) Correspondence Monika Olischar, M.D., Division of General Pediatrics and Neonatology, Department of Pediatrics and Adolescent Medicine, University of Vienna, Waehringer Guertel 18–20, 1090 Vienna, Austria. Tel: +61-3-424936220 | Fax: +61-3-93455315 | Email: [email protected] Received 26 December 2008; accepted 29 January 2009. DOI:10.1111/j.1651-2227.2009.01252.x

Abstract Aim: Progressive posthaemorrhagic ventricular dilatation (PHVD) may induce abnormal amplitude-integrated electroencephalographic (aEEG) activity prior to clinical deterioration or significant cerebral ultrasound changes. These abnormalities might be ameliorated with cerebrospinal fluid (CSF) drainage. The aims of this study were to investigate the occurrence of aEEG-abnormalities with progressive PHVD in relation to clinical and cerebral ultrasound changes and to evaluate whether CSF drainage results in aEEG improvement. Methods: aEEG and cerebral ultrasound scans were performed in 12 infants with PHVD, before and after CSF drainage, until normalization of aEEG occurred. Results: aEEG was abnormal with progressive PHVD in all patients. Concurrently, 60% of the patients were clinically stable without deterioration in ultrasonographic cerebral abnormalities. Post drainage, continuous pattern was restored in all but one patient, whereas the frequency of discontinuous pattern decreased in nine patients and burst-suppression pattern decreased in all but one patient. Low-voltage pattern was only observed in one patient who suffered severe grade IV IVH and died one week after EVD placement. Sleep-wake cycling matured in 75%. Conclusion: These findings demonstrate the impact of CSF drainage on compromised aEEG-activity associated with PHVD. aEEG changes indicative of impaired cerebral function were apparent before clinical deterioration or major ultrasound changes. These changes were reversible with CSF drainage. aEEG should therefore be used in addition to clinical observation and ultrasound when monitoring PHVD.

INTRODUCTION Intraventricular haemorrhage (IVH) is still a major complication of preterm birth (1). The ensuing posthaemorrhagic ventricular dilatation (PHVD) is known to be associated with subsequent white matter damage and permanent neurodevelopmental disability (2–4). In a rat model of neonatal PHVD (5), white matter loss and reduced motor performance occurred in animals with dilated lateral ventricles. Further animal studies have clearly demonstrated that early treatment of silicon oil and kaolin-induced hydrocephalus prevents irreversible brain injury (6,7). Although randomized controlled trials in human infants have failed to show a clear benefit of early intervention by early repeated cerebrospinal fluid (CSF) taps in PHVD (8,9), a retrospective study in 95 infants showed that infants who have received Abbreviations aEEG, amplitude-integrated electroencephalography; CFM, cerebral function monitor; GA, gestational age; PHVD, posthaemorrhagic ventricular dilatation; CSF, cerebrospinal fluid; EVD, external ventricular drain; IVH, intraventricular haemorrhage; RI, resistive index.

1002

late intervention defined as drainage once the p 97 + 4 mm line was crossed, required ventriculo-peritoneal (VP) shunt insertion more often than those infants who were treated earlier (10). An ongoing randomized prospective intervention study is further assessing the role of earlier intervention. Little is known about the optimal treatment modality and the optimal timing in treatment of the injured developing nervous system. Currently, intervention is performed when there is deterioration in clinical findings, such as an increase in head circumference, a bulging fontanelle and an increasing suture width. Moreover, sonography plays a key role in monitoring PHVD. Changes in ventricular size, compromised flow in the anterior cerebral artery, an increase of the resistive index (RI) and graded fontanelle compression during Doppler ultrasound scanning have all been used as markers of impending ischemic injury with evolving ventriculomegaly (11–14). However, these parameters do not measure brain function during development of PHVD and the current modes of assessment of PHVD may detect changes at a late stage, after significant deterioration in brain function. Thus, a tool that detects abnormality of cerebral activity before

C 2009 The Author(s)/Journal Compilation C 2009 Foundation Acta Pædiatrica/Acta Pædiatrica 2009 98, pp. 1002–1009

Olischar et al.

significant deterioration has occurred could be an important advance to improve outcomes of PHVD. Amplitudeintegrated electroencephalography (aEEG) allows continuous neurophysiological surveillance of cerebral function. In two patients previously published (15), we observed an increased discontinuity in background pattern activity and a loss of sleep-wake cycles in aEEG when progressive PHVD occurred. These two infants were otherwise clinically stable and cerebral ultrasound did not show any significant compromise of cerebral perfusion. In this study we tested the hypothesis that aEEG-activity abnormalities occur during progressive PHVD and are detectable before clinical or cranial ultrasound changes. To further confirm the utility of aEEG in PHVD, we determined whether aEEG-activity recovered after CSF drainage.

PATIENTS AND METHODS In the Neonatal Unit at the University Children’s Hospital, patients developing progressive PHVD were assessed for the need for CSF drainage by serial cranial ultrasound scans and close clinical surveillance, including measurement of head circumference. All infants admitted to the Neonatal Intensive Care Unit at the University Children’s Hospital Vienna from October 2003 until December 2006, who developed PHVD, were screened prospectively for the purpose of the present observational study. Only those patients who were treated with an external ventricular drain (EVD), were included in this study. PHVD was defined by cranial ultrasound as the progressive increase in ventricular size following peri/intraventricular haemorrhage. Severe PHVD was defined as a ventricular index crossing the p 97 + 4 mm line according to Levene (16). Gestational age (GA) was determined either from the date of the mother’s last menstrual period or from antenatal ultrasound scans. For infants with IVH, the study design included daily cranial ultrasound scans monitoring the development and/or progression of ventricular dilatation. IVH was graded according to the classification of Papile (17). Scans were performed directly before and after the placement of the CSF drainage system until there was a distinct reduction in ventricular size. aEEG was placed on all of these infants. Patients showing progressive PHVD were examined every two days for changes in aEEG background activity, the occurrence of sleep-wake cycles and seizure activity. A minimum of 4 h traces were obtained perioperatively (for placement of an EVD) until the background activity normalized. An EVD system was placed on the Neonatal Intensive Care Unit when the clinical condition of the patient showed a marked deterioration (increase in head circumference of greater than 2 cm/week, a bulging fontanelle or increasing suture width) or when the ventricular index crossed the p97 + 4 mm line according to Levene (16) indicating severe PHVD. The amount of CSF drained daily was calculated according to the CSF production rate of 0.5 to 1.0 mL/kg/h, but adapted individually to daily changes in the ventricular width measured by cranial ultrasound. All infants were

aEEG in posthemorrhagic hydrocephalus

mechanically ventilated at the time of the placement of the EVD. Postoperatively, rapid weaning of ventilation was attempted. However, when mechanical ventilation was necessary for a prolonged period, midazolam and opiates were given intravenously. Due to blockage or dislocation of the draining device, some patients had to undergo more than one EVD procedure during their clinical course. In order to exclude any possible effect of such a complication on aEEG-activity, data were only collected from the first placement. For the calculation of the relationship between background patterns and sleep-wake cycles before and after CSF drainage, only the last measurement before the intervention and the measurement after the EVD placement showing greatest recovery were taken into account. Cranial ultrasound scans were performed using an Acuson 128XP (Acuson Corp., Mountain View, CA, USA) with a 7.5 MHz transducer. The scans were conducted and assessed by the attending medical staff and reviewed by the investigators (K Klebermass, B Hengl, M Olischar). The ventricular index was measured in the coronal plane from the lateral wall of the body of the lateral ventricle to the falx. In addition, maximum ventricular width was measured in the coronal plane at the level of the foramen of Monro, and a quantification of the cerebral blood flow was performed, measuring the PourcelotRI (18) in the anterior cerebral artery. RI values above 0.85 were classified as abnormally elevated. The aEEG was recorded as a single channel EEG from biparietal surface disk electrodes using a Cerebral Function Monitor (CFM 5330, Lectromed Devices Ltd., Letchworth, UK) for the first measurements and then, with its acquisition, the Olympic CFM 6000 (Olympic Medical, Seattle, Washington, USA), which provided a simultaneous rawEEG signal. For each patient only one type of machine was used. The technique of the CFM has been described in detail elsewhere (19). In brief, the obtained signal is filtered, rectified, smoothed and amplitude-integrated before it is displayed at slow speed (6 cm/h) at the bedside. The quality of the recording is monitored by continuous impedance tracing. The aEEG tracings were assessed by the investigators (K Klebermass, B Hengl, M Olischar) for the relative duration of five background patterns according to Hellstrom¨ Westas et al. (continuous, discontinuous, burst-suppression, low voltage and inactive/flat) (19), the appearance of sleepwake cycles and the presence of seizure activity. Two of the three investigators were not blinded during assessment of aEEG tracings, as they were part of the clinical treating team. However a third rater (B.H.) assessed the tracings blinded to the clinical information and there was uniform agreement regarding the patterns and their duration. The relative duration of each of the five aEEG patterns as described by Hellstrom-Westas et al. (20) in percent was calculated as the ¨ ratio between the duration of the pattern and the duration of the entire recording, as we previously described (21). Sleepwake cycles were recognized as cyclical sinusoidal variations of both amplitude and continuity of aEEG-activity with minimum epoch duration of 20 min (22). Furthermore, we distinguished immature and developed sleep-wake cycles,


1003


Olischar et al.

Table 1 Patients characteristics Patient

GA at birth

Birth weight

IVH grade

Ventilation

Sedation

ACM

Shunt

Death

1 2 3 4 5 6 7 8 9 10 11 12

25 + 5 25 + 5 27 + 3 26 + 3 27 + 5 29 + 3 25 + 3 25 + 1 25 + 5 26 + 4 29 + 5 24 + 3

940 645 1150 700 980 1500 716 730 950 1080 1485 643

4 4 4 2 3 2 3 3 3 4 3 4

yes no yes no yes no yes no no yes no no

Yes No Yes No yes yes yes no no yes no no

yes no yes no yes no no no no yes no yes

no yes no yes yes no yes no yes yes yes no

yes no yes no no no no no no no no yes

median 26 + 4

median 960

n=5

n=6

n=5

n=7

n=3

GA in weeks + days; Birth weight in grams; Sedation = midazolam and opiates; ACM = Anticonvulsant medication consisting of phenobarbital as bolus intravenous application.

according to Hellstrom-Westas et al. (20), and classified ¨ seizure activity into single seizures, repetitive seizures and status epilepticus (20). Previously published reference values for aEEG in neurologically normal and clinically stable preterm infants below 30-weeks GA were used for comparison (21). The voltage criteria of our previous work were applied to dichotomize traces as normal or abnormal, as these provide relative durations of patterns thought to be normal according to GA. aEEG was classified as normal when (1) background patterns appeared to be appropriate for GA (i.e. within 5th to 95th percentile) according to our own reference values (21), 2). Sleep-wake cycles were present and (3) there were no seizures recorded. Conversely, aEEG was considered abnormal when (1) background pattern was inappropriate for GA, (2) there was an absence of sleep-wake cycling or (3) seizures were recorded. The use of anticonvulsants, intravenous sedation and analgesics were noted. Handling or routine nursing care periods were marked on the tracing. Those infants who presented with additional periventricular leukomalacia, cerebral malformations, central nervous system infection or metabolic disorders and hydrocephalus of any other etiology, were excluded. The study was approved by the local ethics committee. Informed parental consent was obtained in all cases.

RESULTS Patients During the study period (October 2003 to December 2006, 39 months), 40/723 (5.5%) patients admitted to the Neonatal Intensive Care Unit acquired IVH (10/40 IVH grade I, 19/40 IVH grade II, 5/40 IVH grade III and 6/40 IVH grade IV). Nineteen of the 30 patients (63%) with IVH ≥ grade II developed progressive PHVD. The inclusion criteria were met for 12 newborns, as 12/19 patients required CSF drainage (63%) and could be monitored prior to and after the placement of a CSF drainage system (IVH grade II

1004

n = 2, IVH grade III n = 5, IVH grade IV n = 5). Some patients had to undergo more than one EVD procedure during their clinical course (6 one EVD, 5 two EVDs, 1 three EVDs). In these cases data were collected from the first placement only. The EVD insertion was performed at a median of 18 days of life (range 10 to 51). The median duration of aEEG recordings was 10 h (range 4 to 20 h). Patients characteristics are summarized in Table 1. aEEG-activity is abnormal in patients with PHVD All patients showed abnormal aEEGs according to our definition with the occurrence of PHVD. One patient fulfilled all three criteria for abnormality (abnormal background pattern for GA, lack of sleep-wake cycling (SWC) and presence of seizure activity), five patients showed a combination of two of the criteria (2/5 presented with abnormal background pattern for GA and seizure activity, 3/5 presented with abnormal background pattern for GA and a lack of SWC) and six patients showed one of the criteria (3/6 showed abnormal background activity for their GA, 3/6 had a lack of SWC). Relationship of aEEG-activity to clinical seizures At the time of the detection of abnormal aEEG-activity, 7/12 patients were clinically stable. Specifically, the infants showed no variations in their pulse or respiratory rates and none had increasing apnoea. Five of 12 patients showed clinical seizures, which could only be confirmed by aEEG in three cases (1 patient had a single seizure, 1 patient had repetitive seizures, 1 patient with a large grade IV IVH had status epilepticus). No aEEG seizures were detected that lacked a clinical correlate. Relationship of aEEG-activity to cranial ultrasound changes The number of cranial ultrasound scans performed prior to EVD placement was four per patient (range 3 to 11). All patients showed ventricular enlargement, seven of these were


Olischar et al.


Table 2 Data before CSF-drainage including aEEG measurements, RIs and ventricular indices according to Levene (17) in the course of developing PHVD

Ultrasound

aEEG Result

Patient

Gestational age at birth

Day of scan

RI

LI

max VW

1

25

6 8 13

0.85 0.78 0.80

3 3 3

17 19 20

2

25

2 4 6 8 12 14 17 19 24 28 35

0.63 0.69 0.76 0.73 0.75 0.71 0.83 0.77 0.86 0.87 0.88

2 2 2 2 2 2 2 2 2 2 2

10.6 11.1 11.4 11.3 11.5 12.3 10.8 13.5 12.6 13.6 14.1

x x x x x x x x

10 12 18 20 26 34 36 40 42 48 50 56

0.80 0.78 0.78 0.76 0.79 0.75 0.80 0.84 0.76 0.78 0.79 0.82

3 3 3 3 3 3 3 2 3 3 3 3

23.6 19.6 17.2 18.3 18.5 19.3 22.3 16.2 23.8 23.3 19.4 19.6

x x

3

27

Normal

abnormal

x x x x x x x x x x

26

8 10 12

0.84 0.78 0.74

3 2 3

15.7 11.7 15.9

x x x

5

27

8 10 14 18 20 22

0.75 0.77 0.72 0.80 0.79 0.77

2 2 2 2 2 2

14.3 14.6 14.9 14.9 13.8 14.5

x x x x x x

6

29

4 6 8 12 15

0.70 0.61 0.76 0.81 0.79

1 2 2 2 2

8 14.9 13.1 13.6 12.4

x x x x x

7

25

2 5 7 9 13 17

0.70 0.67 0.75 0.84 0.75 0.84

2 2 2 3 3 3

11.4 11.5 12.2 14.3 16.5 15.7

x

2 4 5 7

0.64 0.74 0.81 0.69

2 2 2 2

11.8 12.3 12.1 12.2

25

x x x x x x x x x

aEEG Result

Patient

Day of scan

RI

LI

max VW

Normal

9

25

2 5 8 10 13

0.61 0.84 0.70 0.73 0.69

2 2 3 3 3

13.8 14 14.7 15.1 14.8

x

2 6 8

0.54 0.80 0.80

2 3 3

14.3 20.3 18.9

x

10

x x x

Ultrasound

Gestational age at birth

x x x

4

8

Table 2 Continued

26

abnormal

x x x x x x

11

29

3 5 7 14

0.72 0.85 0.73 0.71

3 3 3 3

16 16.6 18.8 22.8

x x x x

12

24

2 4 6 8

0.91 0.87 0.90 0.88

3 3 3 3

14.7 14.5 14.9 14.5

x x x x

RI = Resistive index; LI = Levene index; Levene index – 1: 97th percentile, 3: >97th percentile + 4mm line; max VW = maximum ventricular width measured in coronal plane at level of foramen of Monro.

classified as severe PHVD according to Levene index grade 3. Levene indices and RI measurements for each individual recording are shown in Table 2. At the time of detection of abnormal aEEG-activity, 7/12 did not show any ultrasonographic deterioration. Specifically there was no progression in ventricular dilatation i.e. there was no significant change in either the ventricular width or Levene’s indices. Of the 12 patients enrolled, six (50%) had an abnormal aEEG on every recording after entry into the study. In five of the remaining six, aEEG was classified as normal initially, but became abnormal as the RI increased (see Table 2). Ten patients showed RI’s within the normal range (