Doppler measurements of blood flow velocity were obtained from the superior mesenteric artery (SMA), coeliac axis, and anterior cerebral artery (ACA) of 19 infants with suspected necrotising enterocolitis, which was classified as confirmed (n=9) or unconfirmed (n=8). Infants with confirmed disease were compared with controls who were either enterally fed or who were receiving intravenous fluids. SMA velocity was significantly higher in the infants with confirmed necrotising enterocolitis (36.5 cm/s) than in unfed controls (20.4 cm/s) or infants with unconfirmed enterocolitis (19.6 cm/s). Three infants with confirmed disease had data from before the onset of symptoms. One had low SMA velocity on the first day of life, and one showed no increase in SMA velocity after enteral feeds were started. SMA velocity is increased when infants develop symptoms of necrotising enterocolitis, suggesting that total gut ischaemia is not present at the time that the disease is clinically apparent, although it may precede the onset of symptoms and play a part in the pathogenesis of the disorder. Children Nationwide Neonatal Centre King's College Hospital School of Medicine, London S T Kempley H R Gamsu Correspondence to: Dr S T Kempley, Frederick Still Ward, King's College Hospital, London SE5 9RS. Accepted 9 March 1992 Epidemiological studies have suggested a bewildering variety of risk factors for necrotising enterocolitis. Identified risk factors include birth asphyxia,l-4 chronic intrauterine hypoxia,5 6 apnoea,2 4 7 8 umbilical catheterisation,' 4 7 9respiratory distress and ventilation,4 7 8 patent ductus arteriosus,3 polycythaemia,8 10 hypoglycaemia,8 and hypothermia.10 It is widely believed that intestinal ischaemia is the unifying mechanism that explains the ability of these many different factors to injure the neonatal bowel. Risk factors might cause ischaemia directly or as a consequence of the 'diving reflex' when blood flow is diverted from the abdominal organs during hypoxia, thus preserving cerebral oxygenation. However, the concept of necrotising enterocolitis being predisposed to by gut ischaemia is not straightforward. Clustering of cases has been used to support the idea of an infective aetiology." Although many agents, including bacteria and viruses, have been implicated in some outbreaks,'2 other studies have not identified infective agents to be consistently associated with necrotising enterocolitis.7 It has also been proposed that mucosal immaturity alone may be the most important factor in its aetiology.13 14 The fact that necrotising enterocolitis is rare in the absence of enteral feeding demonstrates the importance of intraluminal factors. Breast milk exerts a protective effect whereas hyperosmolar feeds,2 or hyperosmolar drugs such as high doses of vitamin E,'5 may be risk factors. In order to examine the hypothesis that necrotising enterocolitis is due to intestinal ischaemia, Doppler ultrasound has recently been used to study the characteristics of blood flow in the superior mesenteric artery of infants in high risk groups. We have demonstrated a specific reduction in superior mesenteric artery (SMA) blood flow velocity in infants who are small for gestational age, who experienced chronic intrauterine hypoxia16; and Coombs et al have shown absent diastolic flow and reduced velocity in infants with patent ductus arteriosus.17 However, none of the infants in these studies went on to develop necrotising enterocolitis. We therefore set out to measure SMA blood flow velocity using Doppler ultrasound in infants who developed necrotising enterocolitis. By studying from birth a large number of infants who were at risk we also hoped to have data that preceded the onset of this condition in a number of infants. In order to document the specificity of our findings for the mesenteric circulation, we also measured velocity in the coeliac axis and anterior cerebral artery (ACA).

The interaction of an infant's spontaneous respiration and intermittent positive pressure ventilation can be an important determinant of outcome; infants who actively breathe out develop pneumothoraces.1 Both active breathing out2 and pneumothoraces3 are more common in infants ventilated with prolonged inspiratory times. Fast rate ventilation (more than 60 breaths/minute) during which the inspiratory time is shortened, reduces the incidence of active breathing out4 and pneumothoraces.5 As the infant is weaned from ventilation the rate is often reduced by increasing both the inspiratory time and expiratory time until the inspiratory time is increased to one second. This increase could stimulate active breathing out and reduce the effective-

Aim: To compare the effects of inhaled and systemic steroids on growth in very low birthweight (VLBW) infants with chronic lung disease (CLD).Methods: Sixteen babies with CLD randomly received inhaled budesonide (100 mug four times daily for 10 days via Aerochamber) or systemic steroids (dexamethasone 0.5 mg/kg/day, reducing over nine days). Linear growth (lower leg length, LLL) was measured by knemometry twice weekly.Results: The gestational age, birth weight, postnatal age, and LLL velocity (LLLvel) were similar between the two groups at the start of treatment. At the end of the treatment period, LLLvel was reduced in the dexamethasone group (mean -0.01 mm/day) but had increased in the budesonide group (mean 0.48 mm/day). Mean weight gain was non-significantly lower in the dexamethasone group (5.8 g/kg/day) compared to the budesonide group (mean 12.7 g/kg/day).Conclusion: Inhaled budesonide has less short term effects on growth than systemically administered dexamethasone.

AIMS To evaluate whether serial Doppler measurements of superior mesenteric artery (SMA) blood flow velocity after the first enteral feed could predict early tolerance to enteral feeding in preterm infants. METHODS When clinicians decided to start enteral feeds, Doppler ultrasound blood flow velocity in the SMA was determined before and after a test feed of 0.5 ml milk. The number of days taken for infants to tolerate full enteral feeding (150 ml/kg/day) was recorded. RESULTS Fourteen infants (group 1) achieved full enteral feeding within seven days. Thirty infants (group 2) took 8–30 days. There was no difference in the preprandial time averaged mean velocity (TAMV) between the groups at a median age of 3 (2–30) days. In group 1, there was a significant increase in TAMV (p<0.01) above the preprandial level at 45 and 60 minutes, but this did not occur in group 2. An increase in TAMV by more than 17% at 60 minutes has a sensitivity of 100% and a specificity of 70% for the prediction of early tolerance to enteral feeds. CONCLUSIONS There is a significant correlation between an increase in mean SMA blood flow velocity and early tolerance of enteral feeding. Doppler measurements of SMA blood flow velocity may be useful for deciding when to feed high risk preterm infants.

The aim of this paper is to determine whether antenatal detection of small-for-gestational-age (SGA) babies influences 2-year outcomes. All low-birth-weight (<2,500g) infants born in South-EastThames region, England from September 1, 1992 to August 31, 1993 were identified at birth. Antenatal "suspicion" and ultrasound assessment confirming growth restriction was categorized as "detection" of SGA. Postnatally, infants were classified as SGA if they had a birth weight for given gestation below the 10th centile. At 2 years, those below 32 weeks' gestation and a random 25% sample of infants of 32 weeks' gestation or more underwent pediatric assessments. Of 49,787 births, 3,456 (6.9%) were of low birth weight. One thousand four hundred and fifty one (42.5%) were SGA, of whom 611 (42%) were detected antenatally by ultrasound scan. At 2 years, 1,008 (75.8%) of 1,358 expected infants were assessed, 379 (37.6%) were SGA at birth, and 188 (49.6%) were confirmed antenatally. Although undetected infants had higher mean birth weights and gestational ages, they had a higher proportion of perinatal deaths (12.6 vs. 6.4%, RR 1.96: CI 1.32-2.86) than detected infants. At 2 years, detected SGA infants had smaller head circumferences (p = 0.026), a higher prevalence of febrile convulsions (8.0 vs. 3.1 %: p = 0.040) and lower scores on the locomotor (DQA) scale of Griffith's developmental test (p = 0.021) compared with undetected SGA infants. Despite detected SGA fetuses having lower weights and gestation at birth than undetected fetuses, they had significantly lower mortality without a parallel increase in severe 2-year neuro-developmental, clinical, or growth morbidity.

Human milk is often inadequate nutritionally for preterm infants. We investigated the effect of adding a commercially prepared milk fortifier to human (maternal or bank) milk and measured changes in lower leg length velocity (LLLvel) using knemometry, weight gain and biochemical indices of nutrition. Babies were allocated to one of three feed groups, in a semi‐randomized fashion, to receive human milk alone (group I), fortified human milk (group II) or a preterm formula (group III). The birthweights (median and R) and birth gestations (median and R) of the three groups were as follows: group I 1099 g (654‐1248 g) and 28 wk (26‐32 wk); group II 838 g (742‐1340 g) and 31 wk (28‐36); group III 1136g (624‐1552 g) and 32 wk (27–36 wk). All babies who received fortified milk either showed significant ( p = 0. 0004) acceleration in LLLvel during the period studied, or maintained their pre‐study period velocity. This increase in LLLvel was comparable to that achieved by a group of babies given a standard preterm infant formula ( p <0.001). By comparison, the control group's change in LLLvel was more modest ( p = 0.04). Babies who received human milk with the fortifier added had the lowest serum levels of alkaline phosphatase at the end of the study period when compared to the other two groups. Other biochemical indices were similar in the three feed groups. No adverse clinical events were encountered which could be attributed to the use of the breast milk fortifier.

Theophylline administration has been shown to attenuate erythropoietin (EP) production in adults; the effect of caffeine is not known. Our aim was to determine whether caffeine and theophylline had similar effects on EP production in the premature newborn. If caffeine was found to have a greater effect, this would influence prescribing habits. Fifty preterm infants (mean gestational age 28 weeks) who had clinically significant apnoea were randomized to receive theophylline (4 mg/kg then 2 mg/kg twice daily) or caffeine (10 mg/kg then 2.5 mg/kg once daily). The methylxanthines were continued at least until discharge from the NICU and the dosage altered to keep the levels within the therapeutic range. As an assessment of EP production, serum EP concentrations were measured. Blood for EP, haemoglobin, reticulocyte count, theophylline and caffeine levels was obtained prior to treatment and at least during weeks 3 and 7. There was no significant difference in the mean EP level in the two groups taken prior to treatment at a median age of 2 days of life. There were similar falls in haematocrit and haemoglobin in the two groups during the study period compared to pre-treatment values. At that time, however, the median reticulocyte count was higher in the caffeine compared to the theophylline treated infants (P < 0.05). This was associated with a rise compared to baseline (median 10.0–0.2 mU/ml) in the mean EP levels in the caffeine group and a decrease from a median of 10.1 to 8.3 mU/ml in the theophylline group, but the EP levels in the two groups at week 7 did not differ significantly.

The intestinal circulation normally meets the bowel's metabolic needs, but adaptive responses may sacrifice flow to the intestine to preserve the oxygenation of other organs. Severe or prolonged hypoxia associated with birth asphyxia or intrauterine growth restriction may predispose to necrotizing enterocolitis (NEC) by repeated or persistent redistribution of blood flow away from the bowel. Immature autoregulation of gut blood flow in the face of such challenges may explain the susceptibility of neonates to gut ischaemia. Other factors reducing intestinal blood flow may interact with the increased metabolic demands imposed by enteral feeding to contribute to the development of NEC. The intestinal circulation normally meets the bowel's metabolic needs, but adaptive responses may sacrifice flow to the intestine to preserve the oxygenation of other organs. Severe or prolonged hypoxia associated with birth asphyxia or intrauterine growth restriction may predispose to necrotizing enterocolitis (NEC) by repeated or persistent redistribution of blood flow away from the bowel. Immature autoregulation of gut blood flow in the face of such challenges may explain the susceptibility of neonates to gut ischaemia. Other factors reducing intestinal blood flow may interact with the increased metabolic demands imposed by enteral feeding to contribute to the development of NEC.

This study examines the effect of pregnancy on fetal outcome and maternal renal function in 17 women with Type 1 diabetes mellitus and nephropathy attending a joint diabetic-antenatal clinic between 1985 and 1993. There were 7 successful pregnancies in 6 women with moderate renal impairment, mean pre-pregnancy serum creatinine 165 mu mol l(-1) (Group 1), and 12 in 11 women with proteinuria and preserved renal function (Group 2). Median gestation of pregnancy was 31+3 weeks in Group 1 and 36+4 weeks in Group 2 (p < 0.05). All babies in Group 1 required neonatal intensive care for a median of 19 days (range 8-271) as compared to only 5 of 13 in Group 2 whose median stay was 13 (7-17) days (p < 0.05). There was one late death in Group 1. Longitudinal creatinine data in those with moderate renal impairment suggest no systematic adverse long-term effect of pregnancy on maternal renal function, although differing changes in renal function were observed during pregnancy. The generally favourable outcome achieved relied heavily upon neonatal care expertise.

OBJECTIVES--To assess the contribution of children with different birth weights to special educational needs within a single health district, and to determine whether this pattern changed over the time when the survival of very low birthweight (VLBW) infants was increasing. SETTING--An inner London health district. STUDY DESIGN--A cohort of children born to local parents between January 1974 and December 1980 was selected from birth notifications, including only those infants who survived for more than one month. Community child health records were then inspected to identify children from the cohort who had been formally assessed for special educational needs before their 8th birthday. The risk of special educational needs was compared for the years 1974-77 and 1977-80 (the first and second halves of the period studied). SUBJECTS--The infant cohort consisted of 31,846 children. Altogether 260 (0.8%) of these were later assessed formally. RESULTS--VLBW infants were 4.4 times more likely to be assessed than normal birthweight infants. Formal assessment within the district occurred in three of 68 VLBW infants from the first half of the period studied, and three of 120 from the second half. CONCLUSION--Although VLBW infants are at higher risk, an increase in their survival was not associated with any increase in their contribution to the group with special educational needs within our district. Their contribution, as a group, to the total number of children with special educational needs is very small.

A prospective study of the outcome of care of a regional cohort of very low birthweight (< 1500 g) and very preterm (< 32 weeks) infants was carried out. Its aims were to assess the ability of the CRIB (clinical risk index for babies) score, rather than gestational age or birthweight, to predict mortality before hospital discharge, neurological morbidity, and length of stay, and to access CRIB score as an indicator of neonatal intensive care performance. 676 live births fulfilled the criteria and complete data were available for 643 (95%). Compared with gestation and birthweight, CRIB was better for the prediction of mortality, was as good for the prediction of morbidity, and was not as good for the prediction of length of stay. CRIB adjusted mortality did not demonstrate better performance in units providing the highest level of care. Either the CRIB score was not sensitive to performance or the level 3 hospitals in this study were performing badly. On the basis of this analysis purchasers and providers of neonatal intensive care cannot yet rely on the CRIB score as a performance indicator.

Congenital diaphragmatic hernia is associated with significant mortality and morbidity. The aim of this study was to compare a series of tests with respect to prediction of outcome. Tidal volume and compliance of the respiratory system (CRS) were measured preoperatively and on the first and second postoperative days. The maximum and modified ventilation indexes and the maximum Paco2 were noted for the first 6 hours of life and the first 6 hours postoperatively. In addition, it was recorded whether the stomach was within the ipsilateral hemithorax preoperatively. Twenty infants were studied (median gestational age, 38 weeks; range, 31 to 40), six of whom had a poor outcome, ie, they died or remained oxygen-dependent after 28 days. A CRS of less than 0.18 mL/cm H2O/kg was the most accurate predictor of poor outcome, with 66% sensitivity and 100% specificity. The authors conclude that lung function measurement are useful in the assessment of infants with congenital diaphragmatic hernia.

High frequency oscillation (HFO) as rescue treatment for preterm infants with severe respiratory failure has been assessed and prognostic factors identified. Thirty six infants with a median gestational age of 27 weeks were studied. Immediately before transfer to HFO, the infants were receiving an inspired oxygen concentration of > or = 85% and/or a mean airway pressure of > or = 12 cm H2O and had a median alveolar-arterial oxygen gradient (A-aDO2) of 73.28 kPa (range 49.34-89.91). Seventeen infants subsequently died. Comparison of those 17 with the remaining 19 infants demonstrated that respiratory distress syndrome and persistent fetal circulation were associated with a significantly better outcome than pulmonary airleak. The A-aDO2 after two and six hours on HFO was significantly higher in those infants who survived compared with those who died. We conclude that a diagnosis of pulmonary airleak and failure to show early improvement in respiratory status indicate a poor prognosis when HFO is used as rescue treatment.