Statistics from Altmetric.com
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.
A perspective on the paper by Schulte et al (see page 31)
Despite the many advances of newborn intensive care over the past 20 years, periventricular haemorrhage (PVH) remains a significant cause of morbidity and mortality for the preterm infant. About 15% of infants with birth weight less than 1500 g develop PVH,1 and its presence significantly increases the risk of neurodevelopmental impairment.
New insights have been gained into the pathophysiology of PVH. The germinal matrix, a fragile network of blood vessels lining the ventricular system, is prone to bleeding in the preterm infant. The beagle puppy model of PVH has provided insight into our current understanding of the pathogenetic role of ischaemia and reperfusion.2 In the human preterm infant, depression of cerebral blood flow, associated with initial reduction in myocardial performance and presence of a patent ductus arteriosus, provides an environment in which ischaemia and reperfusion are likely, and PVH occurs more commonly under these circumstances.3
The absence of one unifying aetiological pathway to PVH has left those who practice neonatal medicine without a specific therapeutic strategy that has the capacity to decrease the incidence of PVH. Many therapeutic agents have been investigated over the past 30 years, in the hope of developing such a strategy.
One such agent is etamsylate (diethylammonium 1,4-dihydroxy-3-benzenesulphonate). This non-steroidal drug was shown to be effective in reducing blood loss from menorrhagia4 and after trans-urethral resection of the prostate.5 The benefits of etamsylate in reduction of bleeding in these settings led to the postulate that it may be of benefit in reducing PVH.
The precise mechanism of action of etamsylate is unknown. It has been shown to reduce bleeding time and blood loss from wounds.6 This appears to relate to increased platelet aggregation mediated by a thromboxane A2 or prostaglandin F2α dependent mechanism.7,8 It has also been associated with decreased concentrations of 6-oxoprostaglandin F1α, a stable metabolite of prostacyclin. Prostacyclin is a potent vasodilator, and may be implicated in reperfusion; it is also a disaggregator of platelets.9 Whereas prostaglandins themselves may have a role in regulating cerebral blood flow, etamsylate appears to have no effect on cerebral blood flow.10 When tested in the beagle puppy,11 etamsylate reduced PVH and its administration was associated with reduced levels of 6-oxoprostaglandin F1α. Etamsylate was also thought to stabilise capillaries, reinforcing capillary membranes by polymerising hyaluronic acid.12 No significant adverse effects have been related to the use of etamsylate in humans.
Etamsylate has been the subject of a number of clinical trials, and the first published trial13 did show some reduction in PVH in the etamsylate treated group. However, the distribution of haemorrhage was different between the etamsylate and placebo groups, and the overall mortality related to extensive PVH did not differ between the two groups. Two large multicentre randomised controlled trials followed,14,15 and the developmental outcome data from one of these trials is reported in the current edition.16 In the reporting of the early morbidity and mortality, these two large trials gave conflicting results. Benson et al14 reported that etamsylate reduced both incidence and extension of PVH, although overall mortality was unaffected. The EC etamsylate trial group15 showed no such difference.
The key outcomes of any treatment in the perinatal period are mortality and long term morbidity. In fact, the etamsylate trials were among the first to recognise this important premise, which is why the reporting by Schulte et al16 of their long term neurodevelopmental outcomes is so important. One of the most impressive aspects of this report is a retention rate of subjects over a period of several years approaching 100%, and the absence of confounding by interobserver error. Their finding of improved cognitive outcome in etamsylate treated infants is interesting, with etamsylate appearing to have a more protective effect on girls than boys, albeit in small numbers. However, when death and significant impairment (general cognitive index < 70) are combined, there is no significant difference between infants treated with etamsylate and those who received placebo.
A systematic review of randomised controlled trials of etamsylate (including data from Schulte et al, giving a total of over 500 infants) suggests a reduction in the risk of any PVH in preterm infants treated with etamsylate (risk rate (RR) 0.78, 95% confidence interval (CI) 0.63 to 0.97), but no significant difference in the risk of grade 3 or 4 PVH (RR 0.55, 95% CI 0.27 to 1.12) and more importantly, no overall reduction in mortality or longer term neurodevelopmental or neurosensory impairment.17 A similar message came out of the systematic review of prophylactic indomethacin, with promise of short term benefit but no significant improvement in mortality or longer term outcomes.18 At present, the randomised controlled trials of etamsylate have not shown a significant improvement in either mortality or neurodevelopmental outcome. Consequently, clinical use of etamsylate cannot be recommended in the preterm population.
Long term neurodevelopmental follow up has rightly become the outcome measure of greatest interest in perinatal trials. But how long is long enough? In most cases, “long term” refers to an assessment at 2 years of age. To their credit, Schulte et al16 report follow up at around 4 years of age. Are there subtle differences in neurodevelopmental performance that may not be measurable until much later in childhood, when more detailed psychometric testing is possible? If such differences were to exist, they may be clinically important, impinging on the child’s ability to perform later in life. Follow up to school age is difficult, and rarely achieves the retention of a cohort managed by Schulte et al.16 Problems of cohort retention aside, the single most important barrier to the conduct of trials, with outcome assessment at a meaningful age of childhood, is the difficulty in acquiring funding that will allow staff to be employed until a trial’s completion. Government and other funding bodies must be made aware of the importance of funding trials to meaningful completion, and the responsibility for imparting this message lies with us, the clinical researchers. Even longer term follow up (of the order of five or eight years) of preterm infants enrolled into trials in the perinatal period is an almost overwhelming prospect, but true differences in significant, albeit subtle, neurodevelopmental outcome may not be detectable until this stage.
Advances in neonatal neurology have also led us to a point where our understanding of perinatal brain injury has extended beyond the injury detectable by ultrasound alone. The more diffuse encephalopathy of prematurity is common and is characterised by injury to both white and grey matter, with reduction in volumes of these tissue types.19 This finding further challenges us to unravel the pathophysiology of preterm brain development and injury. Diffuse encephalopathy of prematurity may well be the reason why PVH does not always correlate well with neurodevelopmental outcome. The therapeutic strategies that optimise brain development in the preterm infant will undoubtedly by multifaceted. The potential benefit of strategies used in the perinatal period must continue to be evaluated by well constructed randomised controlled trials, the primary outcome measure of which is timely and meaningful long term neurodevelopmental outcome.
With thanks to Dr Susan Jacobs and Dr Peter Davis for their thoughtful comments on this manuscript.
A perspective on the paper by Schulte et al (see page 31)