Thank you for giving us the opportunity to reply to these interesting letters.
Our study demonstrated that in the ‘study population of 194 women requiring vacuum extraction the KIWI Omnicup was associated with a significantly higher failure rate than standard (metal or silastic) cups’. Dr Vacca questioned the external validity of our results, in particular, the apparently high failure rates. Our results were similar to the results of a study involving another large UK hospital, which demonstrated similar failure rates (31% for the Omnicup and 19% for the standard cups).1 Our study employed a pragmatic design and the deliveries were not performed just by our most junior doctors but by the whole team, four of whom had more than 4 years experience in obstetrics, which would qualify them as specialists in many other European countries. Certainly the ‘lack of skill’ has not been translated to an increased incidence of significant maternal trauma (significant maternal injury was higher than 10% in previous meta-analyses2 and 7–8% in our study). Furthermore, any ‘collective inexperience’ was equally distributed between the two groups.
We agree that training maybe a significant factor affecting current failure rates, but it would be simplistic to assume that it is the only factor. In the Cochrane review of soft versus rigid vacuum extractor cups,2 six out of nine included studies demonstrate failure rates >15% and four out of nine studies show failure rates >20%. This Cochrane review demonstrated that the soft cups are associated with a higher failure rate than the rigid cups; yet, the soft cups are used more often as they are associated with a lower rate of significant scalp trauma. In the Staffordshire/Wigan assisted delivery trial3 (forceps versus vacuum), the failure rate for the vacuum group (Kobayashi silicone cup) was 27%, and in the Portsmouth trial,4 where only metal cups were used, the failure rate was 13.4%.
Professor Mola's suggestion of a ‘failure rate of less than 10% with whatever instrument’ is perhaps a target that all units should aim for, rather than a description of current UK practice. Professor Mola also suggests that transverse and posterior positions of the fetal occiput should not be associated with higher incidence of failure. While this again would be ideal, it has been previously demonstrated that occipitoposterior (OP) position increases the likelihood of delivery by caesarean section after prolonged second stage.5 The majority of OP deliveries in our study were performed in theatre precisely because of the increased likelihood of failure.
It is plausible that the Kiwi Omnicup has a longer ‘learning curve’. We have used cumulative sums charts to analyse the trend of failures with the Omnicup during the study period,6 and we demonstrated that the Omnicup failure rate did not decrease over the study period. This may reflect the true performance of the Omnicup or that the learning curve may be longer for the Omnicup; each obstetrician performed on average less than ten Omnicup deliveries during the study period.
Dr Kotaska's comments provide a very interesting North American perspective to the debate. Contrary to his hypothesis, the Omnicup performed better for nonrotational deliveries. The failure rate for occipitanterior (OA) deliveries in the Omnicup group was 23.7% and for OP/occipit-transmerse (OT) deliveries it was 56.7%. The admission rate to neonatal intensive care unit (NICU) was not apparently increased after anterior metal cup deliveries. There were three admissions to NICU after anterior metal cup deliveries: one baby was admitted for a sepsis screen and two babies for observation following low Apgar scores.
A ‘tailored’ approach, similar to the one described by Johanson et al.,7 is used in most UK hospitals. When the Bird cup is not available, using the Kiwi Omnicup for malpositions is a reasonable alternative but women need to be informed of the expected higher failure rates (approximately 50% in the two UK studies).