This paper considers some of the changes in practice that have occurred in the last 5 years. There have been significant improvements in parental involvement in care. Not all changes have been based on evidence from research: practice has also been affected by changing technology and pressure by industry and other groups. Among the research-based changes were: an awareness of confidentiality, individualised developmental care, increased use of inhaled nitric oxide, therapeutic hypothermia, less postnatal steroids (although the dosage used is not evidence-based), sucrose as analgesia and permissive hypotension.
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.
Five years ago, when I started my morning ward round at 8.30 am in our neonatal nursery, it was brightly lit, a radio may have been playing softly in the background and, if the weather was fine, sunlight streamed in through the window. The sickest babies, clearly visible in their incubators, were ventilated using conventional intermittent positive pressure ventilation and their monitors only made a sound when the alarms went off. One nurse looked after two or three babies and went quietly about her work. Yesterday my ward round started half an hour later (because of the European Working Time Directive), the nursery was darkened, the blinds on the windows were closed. About half the babies were receiving mechanical ventilation, but the others were supported by nasal continuous positive airway pressure (CPAP). They were nursed in incubators, but I could not see them because the incubators were covered with rather attractive cloths, like the cages of budgerigars at night. High on the wall there was an electronic sign featuring an ear (SoundEar), which lit up when there was a sound louder than a whisper – it was continuously alight during the ward round; in two of the rooms, the nurses had switched it off. There were three nurses looking after five babies. Several parents sat next to their babies (as they had 5 years previously), but they were asked to leave when the ward round was not considering their baby (which did not happen 5 years ago). Some had spent the night in the hospital and, whereas previously we had just two bedrooms for parents on the unit, we now have accommodation for six families from the more distant parts of this or another neonatal network (another innovation). Parents at home looked at pictures of their baby on a secure website. Five years ago, I reviewed paper charts and x rays were taken out of envelopes and viewed on a light box. Yesterday, I accessed a (almost) paperless data system on a screen next to each cot, on which I could also see the baby’s x rays, MRIs and (soon) ultrasound scans.
Marshalling my ideas to write this review, I reflected on the nature of the factors that had driven these changes in clinical practice. Some were due to research. Sometimes a fresh analysis of published work changed our understanding of the underlying disease mechanisms. Sometimes there was a government alert, not always evidence-based (for example, the advice to use pH paper to test gastric aspirates or for pregnant women to avoid all alcohol). Sometimes there was accumulating clinical experience and even commercial pressure (for example, the increasing use of nasal CPAP). Sometimes the drive for change came from user groups (for example, the gentler approach to nursing based on NIDCAP, the Newborn Individualized Developmental Care and Assessment Program).1 2
Listing and discussing all the changes that have taken place over the last 5 years would have made for a paper that was long and dull. Instead, I have focussed on two areas: parent-centred care and changes in clinical practice.
The parents’ perspective
Protocols guide more and more of our practice. Ours are structured with a clearly referenced evidence base (where available) and are now available on the internet for viewing by parents as well as colleagues (http://www.lwh.me.uk/).
By accessing a secure website, “Babylink” (http://www.babylink.info/liverpool/NeonatalUnit/Welcome1.aspx), parents and relatives are able to see pictures of their baby and to read notes written by the nurses. The site went live in February 2007 and parents’ views were evaluated a month later. All parents accessed the diary daily, 85% had shared the password with friends and family, and all reported finding information on the website useful. One said, “People at my work have been able to see my baby and also I have friends that live in Australia so they have been able to see him as well”.
Our practice regarding the presence of parents on the ward during ward rounds changed because of a study done in our own unit.3 Formerly parents stayed by their babies’ cots during ward rounds and nursing handovers – we thought that this was in a spirit of fostering openness. But, when we asked parents what they felt about being present when their baby and others were being discussed, the results surprised us. About half had overheard conversations about other babies or thought discussions about their own baby had been overheard. Those who had actually overheard discussions about other babies were concerned about these experiences. It seemed that although parents expected some sharing of information between families on the unit, about half were concerned about confidentiality. Consequently families now stay in the ward for discussions about their own babies only.
The improved accommodation for parents is not evidence-based but was funded by our neonatal network (not evidence-based either). Both seem like good ideas. The practice of darkening the intensive care rooms and covering the cots of critically ill babies was nurse-led and an interpretation of individualised developmental care. Sizun and Westrup4 reviewed the evidence base 3 years ago and set out nicely the theoretical framework underlying developmental care: the evidence that implicates stress as an adverse factor on the very immature developing brain, probably through a corticosteroid effect, provides a biologically plausible mechanism for the benefits claimed by the proponents of individualised developmental care. Sizun and Westrup called for more research “urgently”. The call was echoed by a recent Cochrane review,5 which observed limited benefits with respect to necrotising enterocolitis, improved family outcome and moderate-severe chronic lung disease, but an increase in mild lung disease and in the length of stay by infants receiving developmental care. Although the authors concluded that “before a clear direction for practice can be supported, evidence demonstrating more consistent effects of developmental care interventions on important short- and long-term clinical outcomes is needed”, individualised care has been enthusiastically and uncritically adopted by neonatal nurses.
I have picked out here some aspects of clinical practice that have changed in the last 5 years because of research evidence.
Hypoxaemic respiratory failure
Our treatment of babies born at or near to term who are in hypoxaemic respiratory failure has changed dramatically in recent years. Extracorporeal membrane oxygenation (ECMO) reduced mortality to about a third for babies without congenital diaphragmatic hernia, but we now use it relatively infrequently. This is because of our increased use of inhaled nitric oxide (iNO) for hypoxaemic respiratory failure in term babies, most often due to persistent pulmonary hypertension. iNO is easily available, effective and safe (although not cheap). Its biological basis is established.6 In hypoxaemic term and near term infants, oxygenation improves in about 50% of babies who are given iNO. The effect is fast and dramatic: within 30–60 min after starting iNO, the mean oxygenation index (OI) decreases by about 15 points and mean Pao2 increases by about 50 mm Hg.6 When a combined endpoint of death or the need for ECMO was considered, the evidence was that the improvement was almost entirely due to a reduced need for ECMO although mortality was not reduced.6 Based on this evidence, we treat between 15 and 20 babies each year with iNO in an initial concentration of 20 ppm for term and near term infants who are in hypoxaemic respiratory failure.
The causes (and therefore the pathophysiology) of respiratory failure in preterm infants are different from those for the infant at term. A recent overview of randomised controlled trials (RCTs) on premature babies found no significant effect of iNO on mortality or bronchopulmonary dysplasia (BPD).7 But the nature of clinical practice is such that in desperation one sometimes turns to therapies where the evidence is less than that of a meta-analysis. In an interesting single centre study by Schreiber et al, a 2×2 factorial design was used to investigate the effect of iNO (or oxygen placebo) and high frequency oscillatory ventilation (or conventional ventilation).8 Although the OI for entry to this study was not particularly high (only about 7), there was a significant reduction in the combined outcome of death or BPD (relative risk (RR) 0.76, 95% confidence intervals (CI) 0.60 to 0.97) and a significant reduction in the incidence of serious intraventricular haemorrhage (IVH) or periventricular leukomalacia (PVL) (RR 0.51, 95% CI 0.27 to 0.97). Probably because of the improvement in these biomarkers of brain damage, this study also showed a significant reduction at 2 years of age in the frequency of neurodevelopmental disability due to a decrease in low Bayley scores; cerebral palsy was similar between the groups.9
Although another large multicentre RCT of NO (the INNOVO trial) found no difference between the groups in the incidence of major disability at 1 year of age,10 one is always influenced by research done on one’s own unit. Subhedar and Shaw found that preterm infants treated with iNO had improved oxygenation mostly in the first 2 h after starting treatment.11 This was due to a reduction in pulmonary artery pressure as assessed by echocardiography within 30 min of starting treatment. Two further pieces of evidence emerged last year. Kinsella et al reported that low-dose (5 ppm) iNO reduced to about half the overall incidence of BPD among infants with birth weights between 1000 g and 1250 g. Furthermore, its administration reduced brain damage (IVH, PVL or ventriculomegaly).12 Using 20 ppm, Ballard et al found that infants who received iNO had improved survival without BPD, were discharged sooner and received supplemental oxygen therapy for a shorter time. We therefore sometimes use iNO even for preterm babies.13
High frequency positive pressure ventilation (HFPPV) is available on our unit but rarely used. This is partly because of equivocal evidence for its effectiveness. One meta-analysis14 showed that the only advantage for HFPPV compared with conventional intermittent positive pressure ventilation was a reduced incidence of pneumothorax, and the incidence of pneumothorax is very low in our own unit anyway (2.6% of ventilated babies in 2005). A review by Wunsch and Mapstone15 reported a similar lack of effect. Another meta-analysis observed an increase in the incidence of IVH and concluded, “The small amount of data that exists suggests that harm might outweigh any benefit”.16 But there is another reason for our lack of use of HFPPV. When a technology is used infrequently, its practitioners become less familiar and comfortable with it.
Five years ago, when Ian Laing wrote a similarly titled article,17 he focussed on temperature control and mentioned hypothermia for babies at or close to term who are delivered in a severely depressed state because of a presumed hypoxic-ischaemic encephalopathy (HIE). We now use this treatment for asphyxiated term babies born within this neonatal network. But why was total body hypothermia adopted before the completion of the follow-up phase of several studies? There are several likely reasons for clinicians moving out of equipoise. Firstly, published RCTs18 19 20 21 have reported encouraging results, especially for infants with moderate brain injury. Just over 550 infants with severe HIE have been enrolled in four RCTs. A meta-analysis of these studies with death as an outcome showed a significant improvement (RR 0.77, 95% CI 0.60 to 0.99); if there was a moderate HIE, meta-analysis of the two studies with outcome at 18 months19 20 showed RR 0.51 (95% CI 0.32 to 0.81).22 Secondly, although the TOBY trial, a multicentre RCT of total body hypothermia, finished recruiting in November 2006 and will not report until follow-up has been completed, its data monitoring committee took the unusual step of recommending that a register of cooled babies should be kept, the implication being that no serious adverse effects had been reported. The National Perinatal Epidemiology Unit took up the recommendation and clinicians who took part in the trial agreed that cooling should be undertaken in line with the trial’s protocol. Thirdly, clinicians, who had gained experience in achieving whole body cooling by nursing the baby on a special mattress, discovered for themselves that the procedure was easy to do and seemed safe. Finally, there is a lack of safe and effective alternative interventions for asphyxiated term babies. The implication is that clinicians want to intervene actively if their action is safe and relatively easy to pursue even though the effectiveness of the intervention may be incompletely proven.
It is interesting to consider why this has happened in this way, in contrast to the uptake of other treatment options which have been subjected to RCTs but which have not been generally adopted. For example, a review of studies of delayed cord clamping showed less need for transfusion and less intraventricular haemorrhage.23 Furthermore, the mean regional cerebral oxygenation of babies born after delayed cord clamping was higher at 4 h (70% vs 66%) and 24 h (71% vs 68%) than of those treated conventionally.24 There are likely to be several reasons why this intervention has not been generally adopted, but an important factor may be the lack of a charismatic neonatal champion in the delivery room.
Several other changes have taken place without the support of evidence from RCTs.
Postnatal corticosteroids lead to earlier extubation and are associated with a reduction in the combined incidence of chronic lung disease or death, irrespective of when steroids are given postnatally.25 26 27 However, concerns about long term side effects (neurodevelopmental impairment and impaired growth)28 have led us, like others,29 to reduce our use of postnatal steroids. When an immature baby is ventilated at very high and increasing levels of oxygen (above Fio2 0.85) and is several weeks old and we are concerned about the development of hypoxaemic respiratory failure – the threshold depends on the individual baby and on the attending consultant – we now use a lower dose regime proposed by Doyle and colleagues.30 Our initial dose of dexamethasone is 0.15 mg/kg 12-hourly usually for 3 days, followed by 3 days at 0.10 mg/kg 12-hourly, and then 0.05 mg/kg 12-hourly. The duration of each dose level is adjusted according to clinical response.
An important lesson to be learnt from this experience is that drugs used during this critical period of development need to be subjected to the same testing that new drugs must undergo before their introduction for the treatment of adults. There is, however, a concern that an important and valuable pharmaceutical intervention may have been lost, leading to the conclusion that rigorous pharmacokinetic studies need to be undertaken at different gestations. The changing physiology of the immature human means that it is unlikely that one therapeutic protocol will be appropriate for all premature babies. Our approach needs to be more sophisticated. Drug administration regimes need to be tailored to the individual taking account of gestation, weight and postnatal age.
We recently changed our approach to basic analgesia and now use oral sucrose. It is a shame that this very natural and gentle approach has taken so long to come into general clinical usage. The first reports of its clinical effectiveness were in the mid-1990s,31 but the practice is much older. Following a paper by Markestad showing the effectiveness of sucrose as a treatment for infant colic in 1997,32 Bell wrote a commentary reminding us that the apparent pain relieving effect of sweet things was not exactly news: generations of Jewish baby boys have had their circumcisions eased by a drop of sweet wine immediately before the operation, and being put to the breast “for the sweetest of milks” immediately afterwards. Furthermore, Bell reminded us, the value of chewed dates (70% sugar as glucose and fructose) put inside a baby boy’s mouth before circumcision was known to the prophet Muhammad.33 In 2004, the practice was subjected to meta-analysis and the authors concluded that sucrose was safe and effective for reducing procedural pain from single painful events, such as heel picking and venepuncture.34 Only in December 2006 was this recognised by a new hospital guideline, which pointed out that sucrose/glucose, although effective in term infants, has not been studied adequately in preterm infants and drew attention to a comment that sucrose may adversely affect neurobehavioral development if given frequently to preterm infants.35 I would suggest that even greater caution should be exercised in relation to the long periods of exposure to opiates that very immature babies are currently subjected to as part of well meant efforts to reduce the stress and discomfort of being ventilated: the issue has been studied by Anand et al36 and reviewed by Aranda et al.37
We now accept lower levels of mean blood pressure (BP), “permissive hypotension”, before intervening. Although BP is easy to measure, it is not the best guide to tissue perfusion: blood flow also depends on peripheral resistance (flow = pressure/resistance). Preterm infants have a lower BP on the first day than subsequently, but this does not appear to be generally harmful. Left ventricular output,38 superior vena cava flow (and therefore cerebral blood flow)39 and cerebral oxygen delivery40 also increase during the days after birth. However, on day 1, cerebral fractional oxygen extraction is significantly higher than on subsequent days.41 It therefore seems that an initial low BP may not be harmful: reduced cerebral oxygen delivery is compensated by increased oxygen extraction. But what level of BP requires treatment? As BP falls, cerebral electrical activity decreases. In very immature babies this is shown by the EEG interburst interval, which lengthens with diminished cerebral perfusion and decreased cerebral oxygen delivery. At a mean arterial pressure of 23 mm Hg there is a marked increase in interburst interval (fig 1).42 We have interpreted this observation as indicating not that cerebral damage occurs below this value but that there is decreased metabolism resulting in appropriately decreased substrate requirement. In view of this compensatory response to diminished oxygen and substrate provision, we feel that a mean BP as low as 25 mm Hg is safe for the smallest babies.
Improving the outcome for very immature babies must be at the forefront of our endeavours and there has not been space to discuss changes in our understanding of the disease processes that affect these very vulnerable children, for example, we now know that many have white matter injury and poor brain growth.43 I sincerely hope that a similarly titled review in 5 years’ time will describe changes that will have resulted in a significantly improved outcome for infants at the margins of viability.
I am grateful to Bill Yoxall, Nim Subhedar and Fred Weindling, who commented so usefully on an early draft of this paper, and to Ian Laing and an unnamed reviewer for their very constructive comments.
Competing interests None.
Provenance and Peer review Commissioned; externally peer reviewed.