Editor—As highlighted in a recent edition ofFetal and Neonatal, there is increasing concern about the previously unreported high levels of neonatal group B Streptococcal (GBS) infections in the UK.1 ,2 It is important that we have statistics for regional variations in GBS infection so that we can to produce evidence based guidelines. It is also important that we are clear about the data on which we base our recommendations.

In the commentary which followed our recent paper,2 Nicoll and Heath1 refer to the incidence at which a risk factor based versus a screening approach to the prevention of Group B Streptococcus would be cost effective, quoting from a commentary by Isaacs.3 The original article by Mohle-Boetaniet al from which these data were derived,4 actually gives figures of > 0.65 and > 1.45/1000 live births at which a risk factor and screening based approach, respectively would be cost effective. This contrasts with those quoted by Isaacs, and Nicoll and Heath, of 0.6 and > 1.2/1000 live births.

More importantly, it should be noted that these figures are obtained from a study which used significantly different criteria for both the definitions of a risk factor and on the decision to treat. In the paper by Mohle-Boetani et al the risk factor approach for treatment involved treatment of both “teenagers or blacks who developed labour complications”.4 The latter included either a temperature of > 37.5ºC or prolonged rupture of membranes (PROM) for > 12 hours or preterm labour < 37 weeks of gestation. This is obviously a quite different population from those defined in the CDC guidelines5 where all mothers who go into preterm labour (< 37 weeks gestation) or who have PROM (> 18 hours) or have a temperature (>38ºC) would be offered treatment under a risk factor based strategy. The screening group in the paper by Mohle-Boetani et al were screened at 26–28 weeks gestation not 34–35 weeks as in the CDC guidelines, the latter interval being considered to be when colonisation status is most predictive of colonisation at delivery. Also, treatment was only given if the mothers also developed intrapartum risk factors, (temperature >37.5º C or PROM >12 hours or preterm labour).

Mohle-Boetani et al conclude “The strategy we developed is not generally applicable because different populations might have different risk factors for delivery of infants with GBS disease”. In the study population, 40% of births occurred in women who were teenagers or black.

It is important that before these figures become established in the current literature, we review the original data and the premises on which they were based. It is important to pay attention to crucial differences in the composition of different populations and the risk factors employed in different studies. As new guidelines are being developed we should not make recommendations based on incorrect information.