 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
PEDIATRICS Vol. 108 No. 3 September 2001, pp. 597-607
,
,
From the * Department of Pediatric Neurology, Southampton
General Hospital, Southampton; Objective. Posthemorrhagic
ventricular dilation (PHVD) is a complication of intraventricular
hemorrhage in preterm infants and is associated with a high risk of
long-term disability. Furosemide and acetazolamide are used widely in
the treatment of PHVD in the hope of avoiding the need for placement of
a ventriculoperitoneal shunt, but these drugs have not been evaluated
in a controlled trial. This article reports a multicenter, randomized,
controlled trial designed to test the hypothesis that these drugs would
reduce the rate of shunt placement (or death) and increase survival to
1 year of age without disability.
Methods. Between 1992 and 1996, 177 infants who were less
than 3 months past term and had ventricular width >4 mm above the 97th
centile following intraventricular hemorrhage were assigned randomly to either standard therapy or standard therapy plus drug therapy with
acetazolamide (100 mg/kg/d) plus furosemide (1 mg/kg/d). Infants who
were enrolled in the trial had a median gestational age of 28.6 weeks
and were enrolled at a mean postnatal age of 3.6 weeks. Forty-four
percent were reported to have a cerebral parenchymal lesion on
ultrasound scan at randomization. The primary outcome measure of death
or shunt placement (known in all but 1 infant) occurred in 56 of 88 infants who were allocated to drug plus standard therapy compared with
46 of 88 who were allocated to standard therapy. The risk ratio was
1.23 (95% confidence interval: 0.95-1.59). Neurodevelopmental
outcome information at a corrected age of 1 year (known in all but 3 of
149 surviving infants) included disability or neuromotor impairment in
54 of 67 infants (81%) who were allocated to drug plus standard
therapy and 52 of 69 infants (66%) who were allocated to standard
therapy. Seventy-two of 85 infants (85%) who were allocated to drug
therapy either died or were disabled or impaired at 1 year compared
with 62 of 89 infants (70%) who were treated with standard therapy
(risk ratio: 1.22; 95% confidence interval: 1.03-1.4376). The excess risk of these adverse outcomes was greater among infants who did not
have a cerebral parenchymal lesion seen on ultrasound examination at
trial entry.
Conclusions. These results suggest that the use of
acetazolamide and furosemide in preterm infants with PHVD is
ineffective in decreasing the rate of shunt placement and is associated
with increased neurologic morbidity. This treatment therefore cannot be
recommended.
Institute of Primary Care and General
Practice, Sheffield University, Sheffield; § Medical Statistics Unit,
London School of Hygiene and Tropical Medicine, London; Department
of Paediatrics, John Radcliffe Hospital, Oxford; and ¶ National
Perinatal Epidemiology Unit, Oxford, United Kingdom.
ABSTRACT
Top
Abstract
Methods
Results
Discussion
References
Slowly progressive dilation of the cerebral ventricles
occurs in approximately 35% of all infants who develop
intraventricular hemorrhage (IVH).1 Of these infants with
posthemorrhagic ventricular dilation (PHVD), approximately 15% require
shunt insertion for control of raised intracranial
pressure.2 The prevalence of PHVD per 1000 infants born at
<32 weeks' gestational age has been documented in western Sweden,
increasing from 7.0 per 1000 in 1973 to 1978 to 25.4 in 1983 to 1986 and then declining again to 13.7 in 1991 to 1994.3 This
presumably is attributable to increasing survival of infants of
progressively lower birth weight and gestation followed by improvements
in neonatal care (eg, widespread use of antenatal steroids and
surfactant). Similar changes have been seen in North
America.4,5 Severe IVH with its complications and
periventricular leukomalacia remain the major determinants of brain
injury in preterm infants,1 and the neurodevelopmental
outcome of infants with PHVD is extremely poor.3,6,7
Prevention and treatment of PHVD therefore is of considerable
importance.
Surgical treatment of PHVD in the form of ventriculoperitoneal shunt
insertion is effective in treating symptoms or signs of raised
intracranial pressure but has a high incidence of shunt blockage or
infection in small, ill infants8 and carries lifelong
risks associated with late infection or shunt failure. Endoscopic third
ventriculostomy has been used9 but is unlikely to be an
effective form of treatment in the majority of cases because of the
communicating nature of the hydrocephalus. In the United Kingdom,
moderate to severe PHVD usually is managed at first by intermittent
removal of cerebrospinal fluid (CSF), but randomized trials have
suggested that serial removal of CSF (by ventricular taps with or
without a reservoir or by lumbar puncture) does not reduce the
progression to eventual shunt placement.10-12 In the
large Ventriculomegaly Trial Group study (n = 157), in which infants were randomized at a mean age of 19 days to early tapping
or conservative management in which taps were deferred until head
growth was excessive, the rate of shunt placement was 62% in both
groups. At follow-up examination at 30 months, 48% of survivors scored
<70 on the Griffiths developmental scales, 90% had neuromotor
impairment, 76% had marked disability, and 56% had multiple
disabilities.7
Nonsurgical methods of treating PHVD avoid the risks of surgery.
Isosorbide and glycerol are limited in their effect,13-15 but acetazolamide and furosemide are less dehydrating and for many
years have been considered an effective
treatment.1,1316-19 One uncontrolled study suggested
that these drugs reduce the rate of shunt placement by
50%.19 They have considerable potential for side effects,
however, including metabolic acidosis, carbon dioxide retention, poor
feeding, electrolyte disturbance, osteopenia, hypercalciuria and
nephrocalcinosis, diarrhea, and vomiting.
Given the doubts about the effectiveness and adverse effects of
acetazolamide and furosemide, we undertook an international multicenter, randomized, controlled trial of this treatment. We previously reported the preliminary findings20 including brief outcome information in 105 infants. This article presents the
results for the entire trial population followed up to 1 year (corrected for prematurity).
Our preliminary report20 contains a detailed
account of the methods used. A shorter account is given below.
Trial Entry
The criteria for trial entry were age <3 months beyond the
expected data of delivery, evidence of germinal layer or IVH on cranial
ultrasound scan before trial entry, evidence of progressive dilation of
the lateral cerebral ventricles, and ventricular index >4 mm above the
97th centile for age.21 The ventricular index is defined
as the distance in millimeters between the midline and the lateral
border of the smaller lateral ventricle in the coronal plane at the
level of the foramen of Munro. When the parents agreed, after
discussion assisted by trial information sheets, written informed
consent to participation in the trial was obtained from them. Clinical
information about the infant was reported to the Clinical Trial Service
Unit by telephone. The infant then was allocated randomly either to
standard therapy or to drug therapy plus standard therapy by means of a
computerized minimization algorithm to ensure balance between therapy
groups with respect to referral center and the presence of cerebral
parenchymal lesions on cranial ultrasound scan. The clinician in the
referring hospital then talked with the parents about the allocation
and gave them further written information.
Drug Therapy and Standard Therapy Schedules
Acetazolamide therapy was increased daily to a full dose from
day 4 onward of 100 mg/kg/d. Furosemide 1 mg/kg/d divided into 2 doses
was given from day 1. Supplements of 8.4% sodium bicarbonate solution
(4 mL/kg/d) and potassium chloride (1 mmol/kg/d) also were given from
day 1 and adjusted according to plasma concentrations of sodium,
potassium, and bicarbonate. Drug therapy for 6 months was recommended.
Standard therapy of ventricular dilation was provided in the referring
hospitals. Written clinical guidelines on this were supplied to
participating centers, but, as in any pragmatic multicenter trials,
final clinical decisions were at the discretion of referring
clinicians.
Criteria for Shunt Insertion
Shunt insertion was recommended for infants in both therapy
groups in whom 2 of the following criteria were met: head size at least
1.5 cm above the 97th centile, head growth at least 1.5 cm per week for
2 weeks, and the presence of specified symptoms or signs of raised
intracranial pressure.
Outcome Measures
The primary endpoint was death or shunt placement before a
corrected age of 1 year. Death or disability at 1 year was the principal secondary measure of outcome. These outcomes were documented at first postnatal discharge from hospital and at a clinic visit at a
corrected age of 1 year. An appropriate pediatrician or community pediatrician in the child's district applied the Vineland Social Maturity Scales and undertook a neurodevelopmental examination. When possible, this was an individual who was blind to the neonatal management. Centile position for height, weight, and head circumference was calculated relative to 1990 British growth data.22
Definitions of Neurodevelopmental Status
Neurodevelopmental outcome was described in terms of the number
of infants whose Vineland age-equivalent scores were below the mean and
also the number of infants who had the presence or absence of
neuromotor, sensory, and/or developmental impairments and disabilities.
Neuromotor impairment was considered to be present when there were
abnormal reflexes or tone changes or left/right asymmetry or absent
pincer grip. Neuromotor disability was based on age-equivalent standard
scores for the motor skills area of the Vineland Social Maturity
Scales.23 A score of 50 to 69 was designated a moderate
neuromotor disability, and a score of <50 was designated as a severe
neuromotor disability. Motor delay alone was based on an age-equivalent
standard score of <70 for the motor skills area, with no signs of
motor impairment. Global delay was defined as moderate when the
composite age-equivalent standard score on the Vineland Social Maturity
Scales was 50 to 69 and severe when <50. Infants whose developmental
disabilities prevented their achieving any items on a Vineland subscale
were assigned an age-equivalent score of 20 on that subscale. All
children with a hearing aid were classified as having sensorineural
deafness. Vision loss was considered moderate when the child had
activity that suggested low vision or difficulty with fixation or
tracking and severe when there was only response to light or the child was blind.
Statistical Analysis
An expected frequency of shunt placement of 60% was estimated
from the previous trial of early removal of CSF in PHVD.6 A reduction of the percentage of infants who required shunting from
60% to 40% was judged to be clinically important. A total sample size
of 191 was calculated to be sufficient to give an 80% power of
detecting such a change in rate of death or shunt placement at
P < .05 (2- sided significance testing). The analyses are based on the groups as randomly allocated (ie, intention to treat).
Differences between the randomized groups are presented as relative
risks or as differences between means or medians, with 95% confidence
intervals (CI). Statistical significance was tested by
Approval was obtained from the local research ethics committee for all
participating centers. Trial recruitment began in 1992, and the
opinions of the independent Data Monitoring Committee were sought at
intervals. In August 1996, they recommended that outcome data be shown
to the Trial Steering Group. The Steering Group met on September 2, 1996, and decided that the trial could not prove any advantage to the
drug plus standard therapy arm; recruitment was halted 4 days later.
The trial enrolled 177 infants from 55 centers worldwide. The
primary outcome measure of shunt placement and/or death is known for
all but 1 infant (99.4%). Neurodevelopmental status at 1 year is known
for 146 of 149 (98%) survivors (Fig 1).
METHODS
Top
Abstract
Methods
Results
Discussion
References
2, t tests, and nonparametric tests
as appropriate. The effect on outcome of birth weight, gestational age,
postnatal age, ventricular index, and head circumference at trial entry
was examined by logistic regression. Prespecified stratified analyses
were based on the presence or absence of parenchymal lesions at trial
entry.
RESULTS
Top
Abstract
Methods
Results
Discussion
References
View larger version (25K):
[in a new window]
Fig. 1.
Flow diagram for enrolled infants.
Infants in the 2 therapy arms (88 allocated to drug plus standard therapy and 89 to standard therapy) were closely comparable at trial entry (Table 1). Just under half of the infants in each therapy arm had an associated lesion of the cerebral parenchyma on ultrasound scan at trial entry.
|
Of those assigned drug plus standard therapy, 87 (99%) received acetazolamide and 82 (93%) received furosemide (Table 2). Acetazolamide was given for a median of 35 days (Table 2). Among the 89 infants who were allocated to standard therapy, only 4 (5%) received acetazolamide. Furosemide alone was given to 18 (20%) in that therapy group, predominantly for systemic fluid balance or pulmonary disease rather than PHVD. Adverse effects of treatment with acetazolamide and furosemide were reported in 38 of the infants who were assigned drug plus standard therapy (Table 2). In 23 infants, these effects required permanent cessation of drug therapy. Nearly 60% of infants in each therapy arm had a renal ultrasound examination after trial entry and nephrocalcinosis was detected in a significant excess of the infants in the drug plus standard therapy arm (24% vs 4%; difference: 19%; 95% CI: 9%-30%).
|
Table 3 shows that 99 (56%) had therapeutic CSF removal after trial entry. The number of taps and number of infants who received taps by lumbar puncture or by ventricular puncture or by ventricular reservoir puncture are listed in Table 3 and were similar in the 2 therapy arms, although there was a suggestion of taps being started earlier in the standard therapy arm. The number of infants who received a surgical intervention for hydrocephalus other than a shunt (11% vs 8%) and the number of infants with central nervous system infection (15% vs 12%) were similar in the 2 therapy arms.
|
Dependence on ventilation (38% vs 37%), oxygen supplementation (55% vs 62%) after trial entry, and mean number of days between trial entry and discharge from hospital (69.0 vs 67.6 days after excluding deaths before hospital discharge) were very similar in the 2 therapy arms (Table 4). The rate of head growth between trial entry and first shunt placement, known for 81 of 83 infants who had a shunt placed, also was similar in the 2 therapy arms (Fig 2), as were the ventricular index and the head circumference at shunt placement (Table 4). The excess number of infants who required shunt insertions (43 vs 40) and of infants who required shunt revision (23 vs 14) in the drug plus standard therapy arm were not statistically significant (Table 4).
|
|
There were 18 (20.5%) deaths in the drug plus standard therapy arm, including 15 before discharge from hospital, and 11 (12.4%) in the standard therapy arm, including 10 before discharge from hospital and 1 that occurred at more than 1 year of age after term (Table 4). There was no excess in either therapy arm of infectious, respiratory, or gastrointestinal causes of death, and no deaths were attributed directly to drug therapy (data available on request). Four deaths in each therapy arm were judged by the referring clinician to be related directly to hydrocephalus or its treatment.
Death or shunt placement, the primary outcome measure, occurred in 56 (63.3%) infants in the drug plus standard therapy arm compared with 46 (52.2%) infants in the standard therapy arm after excluding from the
analysis the 1 infant in the standard therapy arm for whom this measure
of outcome was not available (Table 4). This difference did not reach
statistical significance (difference: 11.1%; 95% CI: 
3.2%-25.2%;
risk ratio: 1.23, 95% CI: 0.95-1.59; P = .15; Fig
3). In a multiple logistic regression,
ventricular index at trial entry was the only factor that was
significantly predictive of this outcome measure (coefficient: 0.107;
standard error: 0.045; P < .017).
|
Among survivors, neurologic, visual, and hearing abnormalities; length; weight; and head circumference were fairly evenly distributed in the 2 therapy groups (Table 5). Disability without motor abnormality (delay or impairment) was seen in only 2 infants, 1 with moderate visual impairment and 1 with seizures and on anticonvulsant drug treatment. Vineland Social Maturity Scales overall scores were available for 132 of the 146 survivors assessed. The complete scales were not administered to 13 children, and 1 child who had profound disabilities failed to achieve even a minimal score on the scales. A higher percentage of Vineland overall age-equivalent scores were below the mean score of 100 in the drug plus standard therapy arm (62 of 67 vs 62 of 79; relative risk: 1.179; 95% CI: 1.03-1.35; P < .05; Figs 3 and 4). This was true also for the motor scale subscore (66 of 67 vs 70 of 79; relative risk: 1.11; 95% CI; 1.04-1.19; Figs 3 and 4). The relative risk for the presence of motor impairment, delay, or disability (see Methods for definitions) in infants who survived to be assessed at a corrected age of 1 year (54 of 67 vs 52 of 79) was 1.33 (95% CI: 1.07-1.64; P = .009; Fig 3).
|
|
The principal secondary outcome measure of either death in the first year of life or the presence of neurodevelopmental disorder at 1 year was more common in infants who were allocated to drug plus standard therapy (72 of 85 vs 62 of 89; Table 6). After excluding the 3 infants whose 1-year neurodevelopmental status was not assessed, the risk ratio for death, impairment, or disability of any type was 1.22 (95% CI: 1.03-1.43; P = .02). Analysis of outcome status at 1 year also was done in the subgroups with and without cerebral parenchymal lesions at trial entry (Table 7). The excess risk of death, impairment, or disability was seen in infants who had not had parenchymal lesions at trial entry (n = 98; risk ratio: 1.41; 95% CI: 1.05-1.89) but was not so clear in those who had had such lesions (n = 79; risk ratio: 1.08; 95% CI: 0.94-1.24; Fig 3). However, a test for interaction between treatment allocation and the presence or absence of lesions was not significant (P = .76).
|
|
| |
DISCUSSION |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
The group of infants who were allocated randomly to treatment with
furosemide and acetazolamide in this trial had a higher number of shunt
insertions and a higher number of deaths. In our preliminary report on
this trial, the primary outcome measure of shunt insertion, death, or
both was significantly worse in infants who were allocated to drug plus
standard therapy than in those who were allocated to standard therapy
(risk ratio: 1.42; 95% CI: 1.06-1.90).20 In the full
data set reported here, the difference between therapy arms in the
estimate of risk ratio for these outcomes had fallen to 1.23 and fell
just short of a statistically significant deleterious effect of
allocation to drug plus standard therapy (Fig 3). This change may reflect the phenomenon of "regression to the truth"
a diminution of the observed treatment difference when stopping a trial early compared with when all of the data are in.24 Ventricular size at trial entry was predictive of shunt placement or death in the
first year, suggesting that a greater severity of PHVD was more likely
to be followed by these outcomes.
None of the measures of outcome suggest any advantage from using drug treatment. Indeed, the risk ratio for death, impairment, or disability at 1 year, the principal secondary outcome measure, remained, as in our preliminary report,20 statistically significantly elevated in the infants who were allocated to drug plus standard therapy at 1.22 with a higher risk ratio of 1.41 in the subgroup without parenchymal lesions at trial entry (Fig 3). This also was reflected in a significantly higher percentage of Vineland Social Maturity Motor Subscale and overall age-equivalent scores below the mean for age (Fig 4). These adverse effects of drug treatment on neurodevelopmental outcome may be underestimated, because all 3 infants who were excluded from the analysis because this information was not available were treated with drug plus standard therapy and were at high risk of impairment or disability because of cerebral parenchymal lesions at trial entry. Drug therapy therefore was associated with an adverse neurodevelopmental effect and was, at best, completely ineffective in its primary therapeutic goal of avoiding shunt placement.
Clinicians might have been more inclined to use other treatments or alter their criteria for shunt insertion because they expected greater efficacy of drugs given in addition to standard therapy, especially as concealment from parents and physicians of treatment allocation was impossible because of the need to monitor electrolytes and acid/base balance during drug therapy. Criteria for removal of CSF before shunt placement (Table 3) and for shunt placement (Table 4, Fig 2), however, seem to be similar in the 2 therapy arms. Earlier use of CSF removal, seen as a slight trend in the standard therapy arm of the present trial (Table 3), was found to be of no benefit previously when subjected to randomized, controlled trial.6
The findings of this trial should be generalizable because the enrolled infants are likely to be representative of all infants with PHVD of this degree: a questionnaire that was circulated 2 years after the trial opened showed that the most common reason for nonenrollment of an infant was failure to fulfill the entry criteria (37 of 44 hospitals). Withholding of parental consent (8 hospitals) and overlooking the existence of the trial (6 hospitals) were much less commonly cited as reasons. Correspondence between treatment allocation and treatment received was good; 99% of those who were allocated to drug plus standard therapy received acetazolamide compared with 5% of those who were allocated to standard therapy (Table 2). The use of furosemide for indications other than PHVD in 20% of those in the standard therapy arm is inevitable in a pragmatic trial. Follow-up information was relatively complete up to the corrected age of 1 year with the primary outcome known for all but 1 (99.4%) and the neurodevelopmental outcome known for all but 3 (98.3%) infants.
The lack of therapeutic benefit of these drugs as used leads one to question whether an alteration in the regimen might have been associated with a better outcome. First, with regard to efficacy of the chosen combination of drugs, experimental evidence suggests that acetazolamide and furosemide are synergistic in reducing CSF production, and the largest study to suggest benefit, albeit uncontrolled, used the 2 drugs in combination,19 whereas the published evidence that acetazolamide monotherapy may suffice is weak.25 The use of a higher dose or lower threshold for drug therapy would increase the already substantial incidence of adverse effects, but a lower threshold also would be difficult to justify given these adverse effects and that only 45% of infants in the standard therapy arm required shunt insertion. A lower dose would be unlikely to be more efficacious.
Second, regarding duration of therapy, we recommended continuation of drug therapy for 6 months but accepted cautious weaning in cases with stable or decreasing ventricular size for 4 weeks. Only 5 (6%) required a second course. Furthermore, a temporary benefit from abbreviated drug therapy would have been apparent as an increase, compared with infants in the standard therapy arm, in the interval between trial entry and shunt insertion, but this was not seen (mean: 58.5 vs 63.0 days).
Third, regarding adverse effects, nephrocalcinosis, which may follow acetazolamide monotherapy, can be caused by furosemide and was seen in this trial but led to termination of drug therapy in only 5 (6%) infants. The use of sodium bicarbonate and potassium chloride rather than Polycitra (the preferred preparation in North America in this clinical context) to maintain electrolyte and acid/base balance was determined by the nonavailability of Polycitra in most participating countries. Polycitra would be expected to be equally as liable as sodium bicarbonate to cause carbon dioxide retention. Although Polycitra may be less prone to causing sodium retention, this rarely is a problem in this population of infants who frequently require supplements to correct sodium "leak" from the kidneys, and it seems improbable that this difference would have altered the outcome as lack of efficacy and not lack of tolerability was the principal problem.
The mechanism whereby neurodevelopmental outcome at 1 year was worse in the group that received drugs is unclear. Telephone randomization should have minimized selection bias at trial entry, and adjustment for potential confounders did not alter the effect of treatment allocation on outcome. The excess of deaths did not conform to a pattern that pointed to a specific biological mechanism, and, in particular, drug treatment was not associated with an increase in the number of infants who required ventilation or in the number of respiratory deaths.
The rationale for treatment with these drugs is a reduction of the rate of CSF formation, but these drugs have other effects. In an uncontrolled study of 12 infants, a single 50 mg/kg dose (half the daily dose used in the present trial) increased cerebral blood flow velocity and, when given intravenously, increased intracranial pressure and transcutaneous PCO2 without changing blood pressure or heart rate.26 It has the potential to cause cerebral vasodilation, impairment of autoregulation, and, thus, cerebral injury. Some data from experimental studies suggest that acetazolamide may be toxic to the developing oligodendrocyte. Alterations in cerebral perfusion pressure therefore may underlie the adverse effects seen in this trial. This suggestion is supported by the greater worsening of neurodevelopmental outcome in infants did not have cerebral parenchymal lesions at trial entry compared with those who did (Table 7); possibly those with lesions had already experienced the adverse consequences of abnormal cerebral perfusion, whereas drug treatment increased the likelihood of this occurring in those without lesions at trial entry.
The present trial suggests that infants with PHVD are at high risk of neurodevelopmental disorder. Has this changed? Infants in the previous ventriculomegaly group trial,6 which compared early CSF taps with taps deferred until head growth became excessive, had similar characteristics at trial entry (mean gestational age: 28.4 weeks; mean postnatal age: 17.5 days; mean ventricular index: 17.6 mm; mean head circumference: 28.4 cm; cerebral parenchymal lesion present in 47%) to those in the present trial. Mortality in the first year was similar between the earlier and present trials (19% vs 16%), but the rate of shunt insertion in survivors, which was used as an outcome measure in the earlier trial, was lower in the standard therapy arm of the present trial than in the 2 arms of the earlier trial (44% vs 62%). Neurodevelopmental outcome at 1 year, assessed by similar methods, was assessed as normal in a higher percentage of survivors in the present trial (27% vs 11.5%); this effect is more marked when only the standard therapy arm of the present trial is considered (34% vs 11.5%).
PHVD may be attributable either to white matter atrophy secondary to oligodendrocyte injury or to increase in CSF pressure resulting from obstruction to its flow. Drug or surgical treatment that aims to alter CSF hydrodynamics can be expected to be effective only against the latter. Although distinction between these 2 causes of PHVD was not possible at trial entry, the subgroup that required shunt placement had at least 2 of excessive head size, excessive head growth, and other symptoms of raised intracranial pressure. It is reasonable, therefore, to assume that this subgroup had obstructed CSF flow.
The use of explicit criteria for shunt placement in the present trial (see Methods) may have prevented surgical procedures that otherwise would have been undertaken unnecessarily because of the difficulty of determining both the relative contributions of white matter injury and CSF obstruction to the PHVD and also the precise point at which surgical intervention is required. It is not clear that this accounts for the lower rate of shunt insertion than that reported in the Ventriculomegaly Trial Group study. The differences between these trials, which opened 8 years apart, also need to be viewed in light of a 4-fold increase in the incidence of PHVD up to 1994,27 at which point the present trial opened, coinciding with an increase in the proportion of low birth weight infants as a percentage of all new cases of cerebral palsy from 32% to >50%.28 The incidence of PHVD subsequently has declined,3-5,29 and rates of cerebral palsy in recent birth cohorts of premature infants also may prove to be lower. Both the decline in incidence of PHVD and the lower rates of impairment and disability in the present trial may be attributable to improvements in neonatal care, such as the more widespread use of antenatal steroids (whose neurodevelopmental effects remain the subject of active debate) and of exogenous surfactant therapy.
Drug treatment of PHVD was accepted in clinical practice when the present trial was undertaken yet proved to be completely ineffective and to be associated with a significantly worse neurodevelopmental outcome. Randomized, controlled trials are a powerful method of revealing such associations, which otherwise are particularly likely not to come to the attention of clinicians in conditions with high morbidity and low incidence.
Other current treatment options include intraventricular fibrinolytic therapy with tissue plasminogen activator, which exposes infants to the risk of intracranial hemorrhage,12 may be limited in its effect by tissue plasminogen activator inhibitor in the CSF,30 and, in animal models, needs to be started within 24 hours of hemorrhage to be effective.31,32 Third ventriculostomy is less effective in young infants generally8,9 and is ineffective when hydrocephalus is present despite communication between the ventricles and the extra cerebral CSF spaces, as in 80% of infants with PHVD.6
At present, conservative management using explicit criteria for shunt insertion seems to be the route of least risk. Certainly, removal of CSF may be the only available remedy for symptoms of raised intracranial pressure when shunt criteria are met but surgery is too hazardous to embark on in a small, sick infant. Many infants with PHVD will have therapeutic removal of CSF performed before that stage is reached, but there still is insufficient evidence available to determine whether CSF removal at the stage of excessive head growth improves outcome. Future trials could explore different thresholds and strategies for removal of CSF. Substantial future reductions in the adverse consequences of PHVD, however, are likely to come from continuing changes in neonatal care that contribute to its prevention rather than to its treatment.
| |
APPENDIX: INTERNATIONAL PHVD DRUG TRIAL GROUP |
|---|
Writing Committee: C. Kennedy, S. Ayers, M. Campbell, D. Elbourne, P. Hope, and A. Johnson; Steering Group: C. Kennedy (chairman), M. Campbell, A. Darroch, ?. Darroch, D. Elbourne, A. Grant (until March 1994), P. Hope, A. Johnson, C. Newsome, and M. Wearmouth; Data Monitoring Committee: E. Alberman (chair), D. Field, and P. Yudkin; data coordination: S. Ayers, E. Aylett, D. Elbourne, J. Fooks, K. Frederick, and A. Johnson; statistical analysis: M. Campbell, and D. Elbourne.
Referral hospitals a) non-UK: University of Copenhagen, Copenhagen, Denmark (G. Greisen); Center Hospitalier, Arras, France (B. Theret); Aglaia Kyriakou Children's, Athens, Greece (H. Dellagrammaticas); National Maternity, Dublin, Ireland (F. Gorman); Bellinson Medical Center, Petach Tiqva (S. Davidson), Bnai-Zion Medical Center, Haifa (D. Bader), and Sapir Medical Center, Kfar-Saba, Israel (R. Regev); Center Hospitalier, Luxembourg (M. Schroell); Christchurch, New Zealand (B. Darlow); Aker University, Oslo, Norway (A. Whitelaw); de Santa Maria, Lisbon (J. Costa); de Santo Antonio, Porto (O. Cunha), Maternidade Dr. Alfredo da Costa, Lisbon (D. Fino), and Maternidade Julio Diniz, Porto, Portugal (M. Areias); Kangdang Kerbau, Singapore (N.K. Ho); La Paz, Madrid, Spain (J. Quero Jiminez); University of Lund, Lund (N. Svenningsen), and University of Halmstead, Halmstead, Sweden (D. Andersson); Kinderspital, Zurich (E. Boltshauser), Ostschwiez Kinderspital, St. Gallen (J. Micallef), and Universitatsspital, Zurich, Switzerland (H. Ulrich-Bucher).
Referral hospitals b) UK: Basingstoke District General (R. Walters), Bristol Maternity, City (P. Fleming); Nottingham (T. Stephenson); Dorset County, Dorchester (R. Clifford); Gloucestershire Royal Infirmary, Gloucester (M. Webb); Guy's, London (J. Rissik); Hope Salford (M. Robinson); John Radcliffe, Oxford (P. Hope); Jubilee Maternity, Belfast (M. Reid); Kent and Canterbury, Canterbury (N. Martin); Leeds General (M. Levene); Liverpool Maternity, Liverpool (R. Cooke); New Cross, Wolverhampton (J. Anderson); North Tees General (I. Verber); Northwick Park, Harrow (R. Thomas); Odstock, Salisbury (D. Stratton); Princess Anne, Southampton (M. Hall); Princess Margaret, Swindon (P. Rowlandson); Princess Mary Maternity, Newcastle on Tyne (E. Hey); Queen's Medical Center, Nottingham (J. Grant); Queen's Park, Blackburn (J. Benson); Rosie Maternity, Cambridge (J. Rennie, A. Kelsall), Royal Berkshire, Reading (A. Boon); Royal Cornwall, Truro (G. Taylor); Royal Free, London (V. Van Seneren); Royal Hampshire County, Winchester (D. Schapira); Royal Maternity, Belfast (H. Halliday); Royal United, Bath (J. Osborne, P. Rudd); Royal Victoria Infirmary, Newcastle on Tyne (M. Ward Platt); Rush Green, Romford (S. McKenzie); St. George's, London (P. Hamilton); South Cleveland, Middlesborough (S. Sinha); Southmead, Bristol (B. Speidel); Victoria, Blackpool (R. Stevens); Waveney Maternity, Belfast (J. Jenkins).
Other participating centers a) non-UK: de Pediatria "J P Garrahan," Buenes Aires, Argentina; St. Polten, Austria; University Ziekenhuis Antwerpen, Edegem, Belgium; Clinico Universidad, Santiago; Universidad de Catolica, Santiago, Chile; St. Finbarr's, Cork, Ireland; Intercommunal, Poissy, Maternite Regionale, Nancy, Hopital Bretonneau, Tours, Hopital Nord, Amiens, and Hopital Nord, Saint-Priest en Jarez, France; Hvidovre, Denmark; University of Turku, Turku, Finland; Groot Ziekengasthuis, ME-Hertgenbosch, Holland; National University, Reykjavik, Iceland; Barzilai Atniel, Ashkelon, Israel; Dunedin, Green Lane, Auckland, and Waikaito, Hamilton, New Zealand; Alesund Central, Haukeland, Bergen, and Rogaland, Stavangar, Norway; S Francisco Xavier, Lisbon, Garcia de Orte, Almada, and de S Joao, Porto, Portugal; General de Galicia, Santiago, 12 de Octubre, Madrid, Spain; Central, Karlstad, Lannsjukhuset Ryhov, Jonkoping, Malmo City, Sweden.
Other participating centers b) UK: All Saints, Chatham; Ashington General; Edgware General; and Fazackerly, Liverpool; Grimsby Maternity; Heatherwood, Ascot; King's College School of Medicine, London; Mayday, Croydon; North Tyneside General, North Shields; Ninewells, Dundee; Pembury, Tunbridge Wells; Poole General; Royal Gwent, Newport; Royal Hospital for Sick Children, Glasgow; Royal Maternity, Glasgow; Royal Preston; St. Helier, Carshalton; St. James' University, Leeds; St. Mary's, Newport, IOW; St. Mary's, Portsmouth; St. Peter's, Chertsey; Southern, Glasgow; Ulster, Belfast; University of South Manchester; Warrington District General; Whittington, London; Wythenshaw, Manchester.
| |
ACKNOWLEDGMENTS |
|---|
We gratefully acknowledge all of the infants who were enrolled in the trial and their parents, the Clinical Trial Service Unit, and the financial support provided throughout by Action Research. The Department of Health of the United Kingdom funds the National Perinatal Epidemiology Unit.
| |
FOOTNOTES |
|---|
Received for publication Oct 24, 2000; accepted Jan 24, 2001.
Reprint requests to (C.R.K.) Department of Paediatric Neurology, Mailpoint 21, Southampton General Hospital, Southampton SO16 6YD, England. E-mail: crk1{at}soton.ac.uk
| |
ABBREVIATIONS |
|---|
IVH, intraventricular hemorrhage; PHVD, posthemorrhagic ventricular dilation; CSF, cerebrospinal fluid; CI, confidence interval.
| |
REFERENCES |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
survey form age equivalent and standard score with
the Bayley Mental Development Index.
Percept Mot Skills
1994;
79:203-220 [Medline]This article has been cited by other articles:
 |
|
 |
|
  A. Whitelaw, D. Evans, M. Carter, M. Thoresen, J. Wroblewska, M. Mandera, J. Swietlinski, J. Simpson, C. Hajivassiliou, L. P. Hunt, et al. Randomized Clinical Trial of Prevention of Hydrocephalus After Intraventricular Hemorrhage in Preterm Infants: Brain-Washing Versus Tapping Fluid Pediatrics, May 1, 2007; 119(5): e1071 - e1078. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A J Brouwer, F Groenendaal, A van den Hoogen, M Verboon-Maciolek, P Hanlo, K J Rademaker, and L S de Vries Incidence of infections of ventricular reservoirs in the treatment of post-haemorrhagic ventricular dilatation: a retrospective study (1992-2003) Arch. Dis. Child. Fetal Neonatal Ed., January 1, 2007; 92(1): F41 - F43. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Whitelaw, I. Pople, S. Cherian, D. Evans, and M. Thoresen Phase 1 Trial of Prevention of Hydrocephalus After Intraventricular Hemorrhage in Newborn Infants by Drainage, Irrigation, and Fibrinolytic Therapy Pediatrics, April 1, 2003; 111(4): 759 - 765. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 OTHER ARTICLES NOTED (Nov 01 to 18 Oct 02) Evid. Based Nurs., January 1, 2003; 6(1): e1 - 1. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2001 American Academy of Pediatrics. All rights reserved. |
||
|
||
