2. RELEVANCE TO UNDER-RESOURCED SETTINGS
2.1. MAGNITUDE OF THE PROBLEM
The regions with high operative vaginal delivery rates and where forceps is the standard instrument used should be most interested in this review.
Operative vaginal delivery rates in Latin American countries are low compared with those in most developed countries. Data from hospital deliveries in 18 countries show that rates do not exceed 6% and are below 2% for half of them (Perinatal Information System, Latin American Centre for Perinatology, PAHO/WHO, 1985 - 1995. Unpublished data). It is a region with low operative vaginal delivery rates and high Cesarean section rates (1). Nonetheless, the fact that forceps is the standard instrument used in this region (2), makes this review particularly relevant.
2.2. FEASIBILITY OF THE INTERVENTION
Training of health personnel would be among the main expected difficulties in the introduction of vacuum extraction as the instrument of choice for instrument-aided delivery in Latin American countries, or for that matter in any other region where it is not the standard instrument. It would be necessary to initiate teaching programmes in vacuum extraction, but it may be difficult to find professionals prepared to undergo the training. Also, the teaching programmes would have to be based partially on simulated situations. It is possible that one prognostic factor of vacuum delivery performance to achieve vaginal delivery and good neonatal outcomes is the training of the operators. So while learning the minimum skills needed to perform the procedure, these professionals should not use vacuum extraction in real patients. Training programmes of this nature imply costs which present real difficulties for under-resourced regions with economical constraints. These costs would need to be balanced against the potentially beneficial effects of using vacuum as the instrument of choice.
2.3. APPLICABILITY OF THE RESULTS OF THE COCHRANE REVIEW
Most trials have been conducted in hospitals where the instrumental delivery rate was more than 8 %, while in Latin American countries this figure is much lower. Lack of experience in the use of vacuum could also be a factor contributing to failure rates and neonatal outcomes.
Nevertheless, the effect on maternal morbidity of the use of vacuum extraction has been consistent in all trials and compatible with a minimum of 44% reduction in significant maternal injuries. It is improbable, and without any theoretical basis, that this effect would be significantly different in developing country settings with lower operative vaginal delivery rates and less experience in the use of vacuum.
2.4. IMPLEMENTATION OF THE INTERVENTION
In view of the fact that there is a reduction in the rate of maternal morbidity with vacuum extraction compared to the use of forceps, in settings with good experience of the use of vacuum extraction, this method should be promoted as the first choice when an instrument aided delivery has to be performed. In settings with little or no experience with the use of vacuum, training programmes in vacuum extraction at residency and senior level should be developed. The adoption of vacuum extraction as the first choice for instrument-aided delivery should be promoted only after a minimum standard of training has been reached.
2.5. RESEARCH
A collaborative randomized controlled trial that could definitively answer the remaining questions would be desirable. Experience with the use of vacuum extraction should be a selection criterion for centres to participate in this trial. Settings with little or no experience with the use of vacuum should participate only after proper training. Participation of settings with low vaginal operative delivery rates should be encouraged.
The trial should compare two policies of instrumental delivery: one in which vacuum extraction is the instrument of first choice and the other in which forceps is the first choice. The type of vacuum extraction method in this trial should be the one which is the best in terms of reducing failure rates and neonatal morbidity. The main outcomes to be studied should be neonatal outcomes and success in achieving vaginal delivery. Long-term follow-up of infants should also be considered as an important outcome to study. Such a follow-up study of the women and children participating in the Johanson et al trial (Keele 93 in the review) has been published (3). There were no significant differences between instruments in terms of bowel or urinary dysfunction although urinary incontinence was high overall (47%) at five years' follow-up. There were also no differences with regard to development and visual problems in children.
Sources of support: Latin American Centre for Perinatology (CLAP). Pan American Health Organization - World Health Organization.
Acknowledgements: Agustín Conde Agudelo for useful suggestions.