Article Text

Download PDFPDF

European perspective on the diagnosis and treatment of posthaemorrhagic ventricular dilatation
  1. AJ Brouwer1,
  2. MJ Brouwer1,
  3. F Groenendaal1,
  4. MJNL Benders1,
  5. A Whitelaw2,
  6. LS de Vries1
  1. 1Department of Neonatology, Wilhelmina Children's Hospital, University Medical Center Utrecht, Utrecht, The Netherlands
  2. 2School of Clinical Science, University of Bristol, Southmead Hospital, Bristol, UK
  1. Correspondence to LS de Vries, Department of Neonatology, Wilhelmina Children's Hospital, University Medical Center, Room KE 04.123.1, 3508 AB Utrecht, The Netherlands; l.s.devries{at}


Background Posthaemorrhagic ventricular dilatation (PHVD) is a serious complication of prematurity with subsequent disabilities. The diagnostic and therapeutic approaches to PHVD vary among neonatal centres.

Aim To gain more insight into the different diagnostic criteria and treatment policies on PHVD among neonatal intensive care units across Europe.

Methods A PHVD questionnaire was designed and sent to neonatologists in 37 European centres.

Results A response was obtained from 32/37 (86%) centres located in 17 European countries. An overall estimated incidence of 7% was reported for severe intraventricular haemorrhages (grades III or IV according to Papile) among premature neonates born below 30 weeks’ gestation. Approximately half of these infants developed PHVD, of whom three-quarters required intervention. Ultrasound measurements of ventricular size were most commonly used to diagnose PHVD (94%). No consensus existed on which ventricular parameters needed to be enlarged and when to start treatment of PHVD. Early intervention (ie, initiated after the ventricular index (VI) exceeded the 97th percentile (p97) according to Levene) was provided in 8/32 centres (25%), whereas 23/32 centres (72%) first started therapy once the VI had crossed the p97+4 mm line and/or when neonates presented with a progressive increase in head circumference or with clinical symptoms of raised intracranial pressure. Wide variation was seen with respect to the applied therapy modalities for cerebrospinal fluid drainage.

Conclusion This survey shows that diagnostic and therapeutic approaches to neonates with PHVD vary considerably. Uniform diagnostic criteria would facilitate studies to assess optimal timing and mode of intervention.

View Full Text

Statistics from

A substantial number of survivors of preterm birth exhibit subsequent disabilities. A variety of intracranial lesions are known to account for a spectrum of cognitive and behavioural problems in 25–50% of the very low birthweight infants who survive, while major motor deficits occur in 5–10%.1 Among the different lesions that may cause brain injury, severe intraventricular haemorrhage (IVH) grades III and IV according to Papile et al2 and subsequent posthaemorrhagic ventricular dilatation (PHVD) pose a significant threat to the developing neonatal brain.3,,5 However, preterm infants who develop progressive PHVD in the absence of associated parenchymal lesions may have a normal neurodevelopmental outcome. In a recent retrospective cohort study, we found that 90% of the preterm born infants that were treated for PHVD following an IVH grade III had a developmental quotient (DQ) greater than 85 at 2 years corrected age.6 As was to be expected, outcome was not so promising for those infants with PHVD following a grade IV haemorrhage. They more often had cognitive and motor impairments, and 49% went on to develop cerebral palsy; nevertheless, 67% of these infants with an IVH grade IV had a DQ greater than 85 at the age of 2 years. Results of other studies also suggest that the extent of associated parenchymal lesions is the main predictor especially for motor outcome in neonates with progressive PHVD.7 8

What is already known on this topic

  • Severe IVH and subsequent PHVD pose a significant threat to the developing neonatal brain.

  • At present there is no consensus on the diagnostic criteria and optimal timing of intervention for PHVD.

What this study adds

  • This survey offers an insight into the different perspectives regarding the diagnosis and treatment of PHVD among 32 European neonatal centres.

  • Diagnostic and therapeutic approaches towards PHVD vary widely. The majority of centres initiate treatment once the VI has crossed the p97+4 mm line.

At present there is no consensus about the optimal timing of intervention for PHVD. In the same retrospective study we observed that early intervention was associated with a reduced need for ventriculoperitoneal shunt placement and a trend towards a better cognitive outcome.6 In the prospective randomised Drainage, Irrigation and Fibrinolytic Therapy (DRIFT) study, severe cognitive disability (mental development index <55) was significantly less common at the age of 2 years in those infants with PHVD who had received DRIFT intervention.9 In addition to differences in timing and mode of intervention, diagnostic approaches to PHVD seem to vary among neonatal centres.

The aim of this study was to gain more insight into the diagnostic criteria and different treatment policies for PHVD among neonatal intensive care units across Europe by conducting a survey on this topic.


The survey was conducted between May and August 2010. An online questionnaire was designed, and the invitation to participate was sent by email to neonatologists with a known interest in neonatal neurology in 37 European neonatal centres. Questions did not request the personal opinion of the recipients but explored the formal protocols for the diagnosis and treatment of PHVD.

The questionnaire (see appendix, available online only) included sections related to the diagnostic criteria for PHVD and the timing and mode of intervention. In addition, participants were asked how many very preterm infants (gestational age <30 weeks) were admitted annually, how many of them developed a severe IVH (grade III or IV), how many subsequently developed PHVD and how many received intervention.

As this study related to unit policies and not to clinical information of individual patients, requirements for informed consent and ethics committee approval did not apply.


A positive response was obtained from 32/37 (86%) European centres, located in 17 different countries.

Incidence of PHVD

Incidence rates of severe IVH (grades III and IV) and PHVD were reported by 27/32 respondents. Most respondents provided estimates of the past 1–3 years. Severe IVH was diagnosed in less than 10% of the neonates below 30 weeks’ gestational age in most centres (range 3–20%). Approximately half of these infants subsequently developed PHVD, of whom almost three-quarters required any type of cerebrospinal fluid (CSF) drainage (figure 1). The majority of centres treated less than five preterm infants with progressive PHVD each year.

Figure 1

Estimated incidence rates of severe intraventricular haemorrhage (IVH) (grades III or IV) and posthaemorrhagic ventricular dilatation (PHVD) in very preterm infants below 30 weeks’ gestational age (GA) in 27/32 of the responding centres; a distinction is made between infants with progressive PHVD who received any type of intervention and infants with PHVD who did not receive intervention.

Diagnosis of PHVD

The diagnosis of PHVD was based on measurements of ventricular size using cranial ultrasound in 94% of the centres (figure 2). In infants with an IVH, cranial ultrasound was conducted twice a week in most centres (range daily to once a week). Once PHVD had been diagnosed, cranial ultrasound was performed more frequently, mainly daily or every other day; a few centres, however, continued performing cranial ultrasound once a week (figure 2).

Figure 2

Measured ventricular parameters on cranial ultrasound to diagnose posthaemorrhagic ventricular dilatation. (A) Measurements of the anterior horn width (AHW), the maximal diagonal width of the anterior horn, the ventricular index (VI), the distance between the falx and the lateral wall of the anterior horn and the frontal horn ratio, the ratio between the VI and corresponding hemispheric width, in the coronal plane at the level where the AHW appears maximal. (B) Measurement of the thalamo-occipital distance, the distance between the outermost point of the thalamus at its junction with the choroid plexus and the outermost part of the occipital horn posteriorly, in the parasagittal plane.

In more than half of the centres (n=18), the diagnosis of PHVD was exclusively based on the ventricular index (VI) and corresponding reference curve according to Levene.10 In 10 centres, the VI was measured in combination with additional ventricular dimensions. The anterior horn width (AHW) was assessed by eight respondents, using either the reference curve of Davies et al11 or, Couchard et al.12 A few centres measured the frontal horn ratio (FHR) 13 or took occipital horn size (thalamo-occipital distance TOD)11 into account. The diagnosis of PHVD was based on visual assessment of ventricular shape and the presence of ballooning on either cranial ultrasound or CT in two centres (table 1).

There was wide variation among the respondents regarding the criteria to diagnose PHVD. The thresholds for ventricular size that were most often applied were either a VI above the 97th percentile (p97) according to Levene or a VI exceeding the p97+4 mm line (table 2).10

In addition to measurements of the lateral ventricles, seven respondents took into account the size of the third ventricle; six respondents measured both third and fourth ventricular size.

Table 1

Diagnostic approach to PHVD; n (%)

Table 2

Criteria to diagnose PHVD; n (%)

Timing of intervention

The decision whether or not to start intervention, and when, was mainly based on ultrasound measurements of ventricular size in 20 centres. A rapid increase in head circumference (1.0–2.0 cm per week, depending on the centres’ criteria) and clinical symptoms of raised intracranial pressure (ICP) (eg, a bulging fontanel, apnoeas and/or seizures) were the main indications for CSF drainage in three and six centres, respectively. For another three respondents, the decisive factor was the presence of cerebral blood flow velocity abnormalities, recorded using Doppler ultrasound measurements.

Early intervention, defined as intervention once the VI has crossed the p97 line according to Levene, was used by a quarter of the respondents.10 In the other centres, therapy was first started after the VI had exceeded the p97+4 mm line (late intervention), when a more marked increase in AHW was noted, and/or when neonates presented with a progressive increase in head circumference or with clinical symptoms of raised ICP. Isolated dilatation of the occipital horns (TOD >24 mm) was regarded as an indication for treatment in two centres (table 3).

Table 3

Criteria to start intervention in neonates with progressive PHVD; n (%)

Treatment modalities

When CSF drainage was considered necessary in neonates with progressive PHVD, the majority of respondents (26/32, 81%) started with lumbar punctures (LPs). Six centres (19%) started with placement of either a subcutaneous ventricular reservoir (n=3) or an external drain (n=3). Osmotic diuretics (isosorbide or acetazolamide) were administered before or in addition to CSF drainage in two centres (table 4). In general, the infants’ gestational age did not influence the treatment offered.

Table 4

Applied treatment modalities in neonates with progressive PHVD; n (%)

Most centres performed a few LPs (median 3, range 1–20) before switching to any surgical intervention, depending on the effectiveness of the LP in decreasing ventricular size or ICP. Five centres did not limit the number of LPs, but continued to do LPs until they were no longer effective due to a lack of communication or until the infant's weight and CSF protein content were appropriate to place a ventriculoperitoneal shunt.

All centres consulted a neurosurgeon once ventricular taps or the placement of a ventricular reservoir, subgaleal shunt, external drain or ventriculoperitoneal drain were required. In four hospitals, infants had to be transferred to another centre for neurosurgical intervention. In the other hospitals, either a paediatric neurosurgeon (16/28, 57%) or a general neurosurgeon (12/28, 43%) was on service for neurosurgical interventions. Prophylactic antibiotics (ie, amoxicillin plus clavulanate, ampicillin, cefazolin, cefuroxim, flucloxacillin in combination with gentamycin, teicoplanin or vancomycin) were administered in 66% of the centres (21/32) before neurosurgical intervention (three missing).

Neonates with isolated occipital horn dilatation were treated in two of 32 centres (6%) with either an osmotic diuretic agent (isosorbide or acetazolamide) or a combination of an osmotic diuretic agent and additional CSF drainage in the case of progressive dilatation or signs of raised ICP.

Ventricular reservoir

Twenty-one centres used a ventricular reservoir (66%). Punctures from the reservoir were mainly performed by the attending neonatologist (n=18), resident (n=11) and/or neurosurgeon (n=6). In a few hospitals, physician assistants (n=2), advanced neonatal nurse practitioners (n=2) and nurses (n=1) also performed punctures.

Reservoir punctures were in 90% of the centres (19/21) carried out under strict aseptic conditions, including hand scrubbing and the use of a cap, mask and sterile gown and gloves, whereas the protocol in two hospitals only prescribed hand washing and the use of gloves. While 38% of the respondents (8/21) allowed the CSF to drop freely from the needle following insertion into the reservoir, 62% (13/21) attached the needle to a syringe and used gentle suction. The most common approach was to withdraw 1 ml/min, aiming for 10 ml/kg (n=14). Others either stopped the procedure when 10 ml/kg CSF had dropped freely from the needle (n=2) or used a time limit of 15 min for CSF withdrawal, irrespective of the amount of fluid that had been removed (n=4) (one missing).

Ventriculoperitoneal shunt

With respect to the placement of a ventriculoperitoneal shunt, different indications were reported (table 5). A combination of bodyweight and concentration of protein in the CSF was most frequently considered when deciding whether insertion of a ventriculoperitoneal drain could be performed (17/32, 53%). Either bodyweight or the CSF protein concentration was the main criterion in eight of 32 (25%) and five of 32 (16%) centres, respectively (two missing). The minimal bodyweight required to insert a ventriculoperitoneal drain varied between the hospitals from 1000 to 3000 g. Thresholds for CSF protein concentration ranged between 1.0 g/l and 1.5 g/l. In one centre, the concentration of erythrocytes in the CSF was evaluated in addition to the bodyweight, and this had to be less than 100/mm3.

Table 5

Indications for placement of a ventriculoperitoneal drain; n (%)


This questionnaire offers insight into the variety of different perspectives regarding the diagnosis and treatment of PHVD among 32 neonatal centres with a special interest in neonatal neurology in 17 European countries. From this survey, it can be concluded that there is considerable variation in the diagnostic and therapeutic approaches to neonates with PHVD.

According to these data, the pooled estimated incidence of severe IVH (grades III or IV) was 7% among neonates born below 30 weeks’ gestation. In the literature, the incidence of severe IVH in preterm infants with a gestational age less than 30 weeks or a birth weight less than 1500 g varies between 7% and 18%.14,,18 Almost half of the very preterm infants that were diagnosed with severe IVH developed subsequent PHVD, which is in line with previous reported data.1 Of these neonates, approximately three-quarter required intervention for their PHVD.

Measuring ventricular size on cranial ultrasound was the most common method to diagnose PHVD. However, there was no consensus on which ventricular parameters needed to be enlarged and to what extent. The same applied to the timing and mode of intervention. Some centres advocated early intervention, initiated when the VI had crossed the p97 line according to Levene but had not yet reached the p97+4 mm line, whereas others started intervention only once the VI had exceeded the p97+4 mm threshold or in neonates who showed symptoms of an increased ICP.10 When CSF drainage was deemed necessary, LPs were most often considered as the intervention of first choice, whereas placement of a ventriculoperitoneal shunt, in general, was the last treatment option after failure of other therapeutic interventions. Consensus was missing regarding the use of other treatment modalities for CSF drainage.

Differences in the organisation of neonatal care among hospitals may to some degree account for the reported variety. Whether or not an infant has to be transferred to another hospital for neurosurgical treatment, as well as the availability of specialised neonatal intensive care unit care, will be taken into account when deciding for a specific type of intervention or the placement of a particular ventricular device. Ethical considerations and personal beliefs may also play a role. Neurosurgeons may be less willing to place a shunt or reservoir in extremely low birthweight infants or in infants who lack clinical symptoms of a raised ICP (eg, a bulging fontanel) or a rapid increase in head circumference, even though a raised ICP is unlikely to be present in the context of these signs in the preterm infant.19 Concerns for an increased risk of infection following the placement of a ventricular device may also affect decisions that have to be made. Due to a high number of missing answers (41%), it is not possible to draw conclusions from this survey regarding the infection rate associated with reservoirs and ventriculoperitoneal drains, which varied significantly from 0% to 50%.

For the past few decades, the amount of research focusing on gaining more evidence for distinct treatment approaches has been increasing. Several studies have investigated the optimal timing of intervention,20 21 and currently, the prospective randomised ELVIS trial (Early vs Late Ventricular Intervention Study, trial number ISRCTN43171322) is conducted to assess the potential beneficial role of early intervention (ie, initiated once the VI has crossed the p97 line according to Levene) over late intervention (ie, initiated after the VI has exceeded the p97+4 mm line).10 A few randomised controlled trials have been conducted to gain evidence for the role of diuretics.22 23 Others have focused on the specific benefits of several modalities of CSF drainage, but conclusions regarding the optimal treatment approach varied.9 24,,26 A common concern among neonatal centres is the risk of infections associated with neurosurgical interventions and devices. Results of recent studies that considered infections following neurosurgical treatment are, however, reassuring and suggest that concerns for ensuing infections should not be a limiting factor.27,,29

Several limitations should be considered with regard to this survey. The centres participating in this survey only represent some of the neonatal centres across Europe. Therefore, no conclusions can be drawn regarding the diagnostic and therapeutic approaches for PHVD in other European hospitals or outside Europe. In addition, the provided incidence rates for severe IVH and PHVD should be interpreted as estimates. Management in individual centres may have been limited by the available treatment modalities and specialists. From this questionnaire it was, however, not possible to determine to what extent this did influence the choice of a particular approach to infants with PHVD.

In general, the results of this study underline the need to collaborate on an international consensus on the diagnosis and treatment of neonates with PHVD. More uniformity in diagnosing PHVD would be a significant step forward. The management of infants with PHVD is controversial because the benefits of any specific treatment regimen have not yet been established. More research is needed to define the optimal timing of intervention and the most preferable treatment strategy. The use of cranial ultrasound on a regular basis is recommended to monitor the rate of progression of PHVD and to evaluate the effectiveness of therapy. As most hospitals treat only a few neonates with progressive PHVD each year, the centralisation of care in specialised centres with neonatal intensive care and neurosurgery on site will be important to gain more experience, develop expertise and maintain consistency in the management of infants with PHVD, thus optimising care. In addition, collective registration of data regarding the infection rates associated with the different ventricular devices would be of great value, because little is known about these specific infection risks.

At present, several centres in Europe are involved in the ELVIS study. Awaiting the results of this randomised trial, we hope the ELVIS protocol may be an impetus to a European protocol for the diagnosis and treatment of PHVD to achieve optimal care and neurodevelopmental outcome of these neonates.


The authors would like to thank all the respondents who contributed to this survey by answering the questionnaire: Dr Elke Griesmaier, Innsbruck University Hospital, Austria; Dr Monika Olischar, Medical University of Vienna, Austria; Dr Luc Cornette, AZ-St-Jan Bruges-Ostend AV, Belgium; Dr Gunnar Naulaers, University Hospitals, Leuven, Belgium; Dr Gorm Greisen, Rigshospitalet, Copenhagen, Denmark; Dr Sampsa Vanhatalo, Helsinki University Hospital, Finland; Dr Veronique Pierrat, Hôpital Jeanne de Flandre, Lille, France; Dr Olivier Claris, Hôpital Femme Mère Enfant, Lyon, France; Dr Pierre Gressens, Dr Angela Kaindl, Hôpital Robert Debré, Paris, France; Dr Axel Heep, Bonn University Hospital, Germany; Dr Ursula Felderhof, Dr Mattias Keller, University Hospital, Essen, Germany; Dr Helen Bouza, Aghia Sophia Children's Hospital, Athens, Greece; Dr Vassiliki Soubasi, Hippokratio General Hospital, Thessaloniki, Greece; Dr Geraldine Boylan, Dr Brendan Murphy, Cork University Maternity Hospital, Ireland; Dr Monica Fumagalli, Dr Luca Ramenghi, Ospedale Maggiore Policlinico, Mangiagalli and Regina Elena, IRCCS, University of Milan, Italy; Dr Guiseppe Buenocore, Azienda Ospedaliera Universitaria Senese, Santa Maria alle Scotte General Hospital, Siena, Italy; Dr Linda de Vries, Wilhelmina Children's Hospital/University Medical Centre, Utrecht, The Netherlands; Dr Dag Bratlid, Dr Jon Skranes, St Olav's University Hospital, Trondheim, Norway; Hospital de Santa Maria, Lisbon, Portugal; Dr Ana Vilan Lopez, Hospital de São João, Porto, Portugal; Dr Damjan Osredkar, University Children's Hospital, Ljubljana, Slovenia; Dr Alfredo Garcia-Alix, Dr Adelina Pellicer, La Paz University Hospital, Madrid, Spain; Dr Ulrika Ådén, Dr Mats Blennow, Karolinska University Hospital, Stockholm, Sweden; Dr Gregory Lodygensky, University Hospitals, Geneva, Switzerland; Dr Anita Truttmann, University Hospital, Lausanne, Switzerland; Dr Heike Rabe, Brighton and Sussex University Hospital, Brighton, UK; Dr Marianne Thoresen, St Michael's Hospital, Bristol, UK; Dr Andrew Whitelaw, Southmead Hospital, Bristol, UK; Dr Michael Weindling, Liverpool Women's Hospital, UK; Dr Frances Cowan, Dr David Edwards, Hammersmith Hospital, London, UK; Dr Neil Marlow, Dr Nikki Robertson, University College London, UK; Dr Mary Anthony, John Radcliffe Hospital, Oxford, UK.


View Abstract


  • AJB and MJB contributed equally to this manuscript.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.