Introduction

Chronic lung disease (CLD) is a common adverse outcome of premature birth, particularly in those born at very early gestations [14]. CLD is associated with significant morbidity. Affected infants often have a protracted neonatal unit stay and require frequent readmissions, particularly for respiratory problems [11]. Pulmonary function abnormalities can be demonstrated even in adolescents who had CLD [9]. Corticosteroid administration can reduce CLD, if given in the first 2 weeks after birth [2]. A sensitivity analysis demonstrated that the effect was not enhanced if a dose higher than 0.5 mg/kg per day was given or the course continued for longer than 7 to 14 days. Randomised trials in which shorter courses, 3 days or less, were given have yielded conflicting results [3, 7, 16]. There is, however, increasing concern about the long-term adverse effects of dexamethasone administration [17, 18]. Those data [17, 18] may have influenced prescribing habits. The aim, therefore, of this study was to determine if corticosteroids were still prescribed and, if so, whether there was a consensus regarding the most appropriate regimen to use.

Methods

A postal survey was undertaken. One consultant paediatrician (identified from the Directory of Emergency and Special Care Units (2001) produced by CMA Medical Data, Cambridge Research Laboratories, UK) at each of the 223 units in the United Kingdom was sent a questionnaire. The questionnaire specified that this was a survey of corticosteroid use for the prevention and treatment of CLD. The paediatricians were asked which mode of delivery, timing of commencement and indications were used when administering corticosteroids. For systemically administered corticosteroids, the paediatricians were also asked to indicate the dosage and duration of administration.

Analysis

For the purposes of the analysis, units were classified as either special care or intensive care units. Units were classified as special care if they either never ventilated infants or had one intensive care cot for stabilisation purposes only. Units were classified as intensive care if they had at least two intensive care cots. Differences were assessed for statistical significance using the Chi squared test. The data are presented as a percentage of the responses to a particular question.

Results

A total of 160 completed questionnaires were received giving a response rate of 72%. This included responses from 58% special care units and 76.7% from intensive care units. Of the respondents, 33% indicated that they never prescribed any form of corticosteroid; 85.7% of respondents from special care units and 21.6% from intensive care units (p<0.0001) never prescribed corticosteroids. Only six respondents (3.75%) prescribed inhaled corticosteroids only. Of all respondents, 40% prescribed inhaled and systemic corticosteroids; inhaled corticosteroids were given to infants with established CLD and/or infants who had troublesome wheeze. Inhaled corticosteroids were not prescribed in the first 2 weeks after birth.

No respondent prescribed systemic corticosteroids during the 1st week after birth; 21.1% commenced corticosteroids during the 2nd week and 33.3% only outside the neonatal period. In all cases, dexamethasone was the systemic corticosteroid prescribed. A variety of dosages were used: 47.7% prescribed 0.5 mg/kg per day, 19.3% a larger daily dose and 7.9% prescribed 0.1 mg/kg per day or less. The duration of the course of systemic corticosteroids also varied: 15.1% prescribing 3 days, 28.6% 7 days, 22.3% 10 days, 17.8% 14 days and 3.7% 6 weeks. Others prescribed courses varying from 5 to 21 days. The majority (77.1%) prescribed corticosteroids only for ventilator-dependent infants; other indications included chronic oxygen or CPAP dependency.

Discussion

We have demonstrated that systemic corticosteroids are still prescribed on the majority of UK neonatal units, but there is little consensus regarding the optimum dosage regimen and timing of administration. The overall response rate to the questionnaire was 72%, but higher from respondents from units providing intensive care, who are more likely to care for infants at risk of or with established CLD. Thus, we feel the results of this questionnaire are representative of current corticosteroid prescribing habits in UK neonatal units.

Corticosteroids were never prescribed by 33% of respondents; the majority of these were from special care units. Indeed, 85.7% of those from special care units did not give corticosteroids. Approximately 22% of respondents from intensive care units, however, also did not prescribe corticosteroids, which must reflect concerns about long-term adverse effects [17, 18] and other adverse effects, such as increased risk of gastro-intestinal perforation [7]. The risk of adverse neurodevelopmental outcome and cerebral palsy may be increased threefold by systemic dexamethasone administration [10]. Such possible long-term effects need to be weighed up against the likelihood that if corticosteroids are given in the first 2 weeks after birth, CLD and mortality might be reduced [2].

Inhaled corticosteroids alone were prescribed by very few respondents. Inhaled compared to systemic administration of corticosteroids have a slower onset of action and smaller magnitude of effect [6, 12]. In a large randomised trial assessing the efficacy of inhaled corticosteroids, care was taken to maximise corticosteroid delivery, but the only positive effects noted were that fewer infants who had received inhaled beclomethasone subsequently received systemic glucocorticoid therapy, fewer infants were on mechanical ventilation at 28 days and fewer infants required bronchodilator therapy [4]. The limited efficacy of early administration of inhaled corticosteroids demonstrated so far is the likely explanation for the low numbers prescribing that therapy. A larger number of respondents, however, did give inhaled corticosteroids to infants with established CLD with or without troublesome wheeze. Inhaled corticosteroids have been demonstrated to improve lung function and reduce bronchodilator usage in wheezy, prematurely born infants seen at follow-up [19].

A 6-week course was rarely prescribed, although the results of a randomised trial [5] demonstrated that this length of corticosteroid administration compared to a shorter course or no corticosteroids was associated with less oxygen dependency and improved neurodevelopmental outcome, but few infants were seen at follow-up. A subsequent sensitivity analysis [2] demonstrated no advantage with respect to either CLD development or survival from prescribing corticosteroids for more than 7 to 14 days. A 3-day course of corticosteroids was also unusual. Randomised trials have not shown any major advantages of such a short course, unless it is repeated [3] or started in the first 48 h after birth [7], but that timing has been associated with an increase in intestinal perforation [7].

Corticosteroids were not prescribed during the 1st week after birth; this may reflect concerns about side-effects, but also the difficulty in predicting which infants are at highest risk of CLD. A variety of possible predictors of CLD development have been examined, including the chest radiograph appearance [8] and the results of lung function tests [15], but none to date are sufficiently accurate to be used to indicate the need to prescribe a treatment with possible long-term adverse effects. Approximately 33% of respondents prescribed corticosteroids outside the neonatal period. Meta-analysis [13] of the results of nine randomised trials in which corticosteroids were commenced after 3 weeks of age, demonstrated a significant effect on CLD at 36 weeks post-conceptional age, but no impact on mortality. The other beneficial effects were a reduction both in the failure to extubate by 28 days and the need for late rescue treatment with systemic dexamethasone, also fewer dexamethasone-treated infants were discharged home on supplementary oxygen. A greater proportion of respondents to the survey prescribed corticosteroids in the 2nd week after birth, a timing known to reduce CLD and mortality [2]; such a policy, however, was adopted by less than 50% of the respondents.

Corticosteroid administration was usually reserved for ventilator-dependent infants. Approximately 25% of respondents also prescribed corticosteroids for chronically CPAP- or oxygen-dependent infants. We did not ask the details of the magnitude or length of dependence of such infants; thus it is possible that these infants, although not ventilator-dependent, still had severe CLD.

Recent guidelines [1] suggest that until clear benefits of corticosteroids have been identified and the long-term neurodevelopmental outcomes clarified, dexamethasone use should be limited to "exceptional clinical situations" or only given within the context of randomised controlled trials. We have demonstrated that systemic corticosteroids are still prescribed on the majority of UK neonatal intensive care units. Our survey has also highlighted that there is wide variation in prescribing habits regarding the dosage regimen and timing of administration. These data highlight the need for appropriately designed studies to identify if there is a corticosteroid dosage regimen with a positive risk/benefit ratio.