Results

The incidence of grade III–IV PVH/IVH was 13.4% in the whole study group, 15.1% in the no-tocolysis group, 10% in the magnesium sulphate tocolysis group, 5.3% in the ritodrine group, and 13.0% in the indomethacin group. In the univariate analysis (table 1), tocolytic treatment, lack of maternal antenatal steroid treatment, multiple births, PROM, vaginal delivery, low gestational age, low birth weight, Apgar scores of 0–3 at one and five minutes after delivery, RDS, PDA, mechanical ventilation, and pneumothorax were also significantly related to grade III–IV PVH/IVH.

In the multivariate logistic regression analysis (table 2), the following factors were significantly and independently related to grade III–IV PVH/IVH: lack of antenatal steroid treatment, gestational age, one minute Apgar score 0–3, RDS, PDA, mechanical ventilation, and pneumothorax. Infants exposed to ritodrine tocolysis were at significantly lower risk of grade III–IV PVH/IVH after adjustment for the other variables (odds ratio (OR) = 0.3; 95% confidence interval (CI) = 0.2 to 0.6). This effect was not noted in infants exposed to magnesium sulphate (OR = 0.8; 95% CI = 0.5 to 1.2) or indomethacin (OR = 1.0; 95% CI = 0.6 to 1.6) tocolysis. The Hosmer and Lemeshow goodness of fit test showed a p value of 0.84.

Table 3 shows the percentage of grade III–IV PVH/IVH by type of tocolytic treatment, stratified by gestational age, birth weight, and antenatal steroid treatment. The percentage of severe IVH in infants whose mothers received ritodrine tocolysis was lower than that among infants whose mothers were treated with other agents or among those not receiving any tocolytic treatment. This trend was observed in all gestational age, birth weight, and antenatal steroid treatment groups.

Discussion

The results of this study suggest that maternal tocolytic treatment with ritodrine was associated with lower incidence of grade III–IV PVH/IVH, whereas magnesium sulphate and indomethacin tocolysis did not appear to have any influence on the incidence of grade III–IV IVH in VLBW infants.

The current literature on the effect of tocolytic treatment on the incidence of neonatal PVH/IVH comprises mostly retrospective and small sample studies, and the results are controversial. The retrospective study of Groome et al 4 showed that β-mimetic tocolytic treatment may be associated with more than a twofold increase in the incidence of neonatal PVH/IVH. However, Özcanet al 8 found that ritodrine did not have a significant effect on the incidence of neonatal PVH/IVH, and Ment et al 13 even reported a decreased incidence of neonatal PVH/IVH with maternal β-mimetic tocolytic treatment.

Primary fetal side effects of β-mimetic tocolysis are reduction in arterial blood pressure and tachycardia.2 ,3 β-Mimetic tocolytic treatment is, however, well tolerated by most fetuses, and side effects are strongly dose dependent.4

Several possible mechanisms for the effect of ritodrine in reducing the incidence of grade III–IV PVH/IVH may be considered. β-Mimetic tocolytic drugs have been used for treatment of acute intrapartum fetal distress by reducing uterine contractions.17 As fetal distress is a risk factor for PVH/IVH,1 this effect may reduce the incidence of haemorrhage. β-Mimetic drugs administered to mothers may promote neonatal lung maturation, resorption of alveolar fluid, and decreased lung water content.18-21 Laroset al 22 found that maximal inspiratory pressure was lower in infants with RDS who were exposed to β-mimetic tocolysis. Ritodrine may thus theoretically decrease the risk of PVH/IVH by improving respiratory status and reduce mechanical ventilation requirements. However, we were unable to grade the severity of RDS from the data collected.

Sudden blood pressure elevation in sick VLBW infants with impaired autoregulation of cerebral perfusion may result in germinal matrix capillary rupture and subsequent PVH/IVH.1 It may be speculated that the systemic vasodilator effect of β-mimetic tocolysis2 ,3 could reduce grade III–IV PVH/IVH by preventing blood pressure elevation in VLBW infants during perinatal asphyxia. Another possible effect is that sympathomimetic agents may have a role in controlling the tone or permeability of vascular walls within the central nervous system.23 However, at present these effects are speculative, and the relation between β-mimetic tocolysis and reduction of PVH/IVH needs to be further explored.

Magnesium sulphate has several neuroprotective effects including vasoconstriction inhibition, reduction of free radical formation, and prevention of calcium influx into cells during hypoxia, thus preventing nerve cell death.24 Kuban et al 12 suggested a protective effect of maternal magnesium administration on neonatal PVH/IVH. However, Stigsonet al 7 suggested that magnesium tocolysis may increase the incidence of neonatal grade I IVH, and the study of Iannucci et al 5indirectly implied that magnesium tocolysis may be associated with increased incidence of PVH/IVH. The study of Salafiaet al 25 found an increased incidence of early PVH/IVH (< 72 hours) among infants whose mothers were exposed to less than 48 hours of antenatal steroids and received magnesium. Our study confirms the results of Levitonet al 11 who found no effect of magnesium sulphate administered before delivery on neonatal PVH/IVH. Recently Rantonen et al 26 found grade III–IV PVH/IVH in 15% of infants exposed to maternal ritodrine treatment, in 9% of infants whose mothers received tocolytic magnesium, and in none of those exposed to magnesium for pre-eclampsia. These differences were not significant.

Norton et al 6 found that indomethacin tocolysis increased the risk of neonatal PVH/IVH. The study of Iannucci et al 5suggested that indomethacin tocolysis may contribute to the occurrence of PVH/IVH. Post partum administration of indomethacin, however, lowered the incidence and severity of neonatal PVH/IVH.9 ,10 The inhibition of prostaglandin synthesis by indomethacin may protect the brain by decreasing baseline cerebral blood flow, microvasculature permeability, which occurs in response to hypoxic-ischaemic insults, and free radical formation.27This theoretically may reduce the incidence of PVH/IVH. Our study, however, did not show a significant association between indomethacin tocolysis and the incidence of grade III–IV PVH/IVH.

Ment et al 28 found that antenatal steroid treatment reduces grade III–IV PVH/IVH. Recently the EPICure study29 found that major cerebral scan abnormalities were less common in infants born before 25 completed weeks of gestation and whose mothers had received steroids for more than 24 hours before delivery. Although cerebral scan abnormalities were more common if the mother had received tocolytic agents, no details on the tocolytic treatment were given in this study, and this analysis did not include intracranial abnormalities in infants who died before discharge.

The protective role of antenatal steroid treatment on the incidence of grade III–IV PVH/IVH is confirmed by our study. It should be emphasised, however, that only 42% of the infants in our study were exposed to a full course of antenatal steroids and 17.5% to a partial course. Total exposure to antenatal steroid treatment increased over the study period from 45% in 1995 to over 62% in 1998. The relatively high proportion of infants not exposed to steroid treatment may also be associated with regional variations or late presentation by some mothers.

Low gestational age and asphyxia, as reflected by the Apgar scores, had an adverse effect on the incidence of grade III–IV IVH. However, in our logistic regression analysis the type of delivery did not affect the incidence of severe haemorrhage. The elimination of low birth weight from the final logistic model was due to the strong association between birth weight and gestational age.

Our study has several drawbacks and the results should be interpreted with some caution. The study was based on data obtained from the Israel National VLBW database and the tocolytic treatment was not randomised. Given the large number of centres participating in this study, the care of the mother and infant was not standardised, and the dose and duration of the tocolytic treatment and various management protocols are not accounted for. The database does, however, include almost every VLBW infant born in Israel during the study period, and completion of the data form is based on the use of standard definitions. Although we defined grade III–IV PVH/IVH as described by Papileet al,14 we are aware that echo density in the white matter adjacent to the ventricular wall is often not a haemorrhage.30 Therefore grade III–IV PVH/IVH should perhaps more properly be viewed as haemocephalus, with or without accompanying periventricular white matter damage. We also do not have data on the age at the time of diagnosis of PVH/IVH. We thus could not distinguish between early onset PVH/IVH, which may be caused by obstetrical factors, such as route of delivery and the use of tocolytic agents, and late onset PVH/IVH, which may be the result of non-obstetrical factors, such as mechanical ventilation and pneumothorax.

Our population based study and large sample size do, however, suggest a possible protective effect of ritodrine tocolysis on the incidence of severe PVH/IVH among VLBW infants. Further controlled studies are justified to clarify the very important clinical issue of the relation between commonly used tocolytic agents and neonatal PVH/IVH.