Editorial on “Open Versus Laparoscopic Pyloromyotomy for Pyloric Stenosis: A Meta-Analysis of Randomized Controlled Trials” by Jia et al.

The introduction of minimally invasive approaches into everyday surgical practice has brought with it a number of challenges. Firstly there are the obvious ones: the need for the surgeon to learn, develop and become proficient in new surgical skills. These include both the generic skills of minimally invasive surgery and those skills specific to the particular surgical procedure being performed. In some cases this may involve performing a well established procedure in a completely different way (inguinal hernia repair is the example that springs to mind). Only once a surgeon has demonstrated proficiency under supervision is he or she then permitted the luxury of independent operating. Beyond this, there is an individual learning curve for any particular procedure, that is, the length of time or number of cases required for an individual to reach his or her plateau of competency. The general principle appears to conform to the proverb ‘practice makes perfect’ implying that greater experience results in better outcomes.

Following this individual acquisition of skills, there is an institutional learning curve. The recognition of its very existence already indicates that minimally invasive surgery involves teamwork. Every member of the team must learn his or her role within the existing framework to ensure the success of any given surgical procedure. Whilst it is less likely that institutional inexperience will affect overall outcome for an individual patient, increased institutional experience may affect more subtle endpoints such as surgical efficiency, leading to reduced operating and anaesthetic times, for example.

In addition to these obvious physical challenges there are a number of philosophical ones. For instance, some may question whether a minimally invasive procedure is required at all. This raises several interesting dilemmas: just because we can perform an operation with minimally invasive techniques, does it mean we should? Does there need to be anything wrong with an old operation to seek a new way of doing it? (This is the ‘if it isn’t broken then don’t fix it’ attitude). Then there are the controversies related to the acceptance of a new surgical procedure: does a new operation need to be better than the old one to gain generalised acceptance or is an operation that is just as good as the old one good enough? How should the differing benefits of minimally invasive and open surgery be compared?

With the generation of well designed, large, prospective randomised controlled trials (RCTs) a number of groups from around the world have attempted to address some of these issues. Furthermore, academic surgeons and statisticians have combined forces to pool data for some procedures to improve its strength and confront sceptics. Despite convincing data in high quality academic publications it appears that the acceptance of results (as evidenced by the incorporation of these techniques into current surgical practice) is limited [1]. These are doubtless issues that will continue to vex surgeons for years to come and for which at present there remain no clear answers.

In this issue of the Journal Jia and colleagues present the latest in a series of attempts to compare the outcomes of open and laparoscopic pyloromyotomy for the treatment of infantile hypertrophic pyloric stenosis. Using the established statistical technique of meta-analysis they have pooled data from 3 recent, large, prospective RCTs comparing open and laparoscopic approaches. Such pooled comparative studies are the most important class of medical evidence (class I) we have available to us and may therefore be regarded as providing the definitive answer to some of the questions raised previously.

On the basis of their meta-analysis, Jia and colleagues conclude that both minimally invasive and open pyloromyotomy are safe and effective approaches in the child with pyloric stenosis but that a minimally invasive approach may (on the basis of non-significant trends) result in shorter recovery times. We fully agree with their findings based on both our own studies [2] [3] and our interpretation of the data. As a result of previous RCTs carried out in this field [3] [4] [5], laparoscopic pyloromyotomy has become our procedure of choice for infants with pyloric stenosis. However, there remains one specific area of contention which remains unresolved when comparing laparoscopic and open pyloromyotomy, namely the incidence of incomplete pyloromyotomy. Among the 3 large RCTs reported (all included in Jia's meta-analysis) incomplete pyloromyotomy only occurred in 2 trials [3] [4]. Overall, the incidence of incomplete pyloromyotomy in these 3 studies was zero (0/245) following open surgery and 2.5% (6/237) following a laparoscopic approach. The astute will observe that these figures are not the same as those that feature in the meta-analysis of Jia and colleagues, or in the previously published, and nearly identical, meta-analysis of Sola and Neville [6]. The reason for this is a technical one, and has to do with the statistical technique used and its inability to utilise studies which report a zero incidence of a particular complication in each arm. We have previously commented on this problem and shown that whether there is a statistical difference in the incidence of incomplete pyloromyotomy is critically dependent on the method used to compare these incidences [7]. If incidences are compared using Risk Difference as opposed to Odds Ratio, then the data from trials with zero incidence of incomplete pyloromyotomy can be utilised, whereas if Odds Ratios or Relative Risks are compared, then these trials are not included. As the RCT with a zero incidence of incomplete pyloromyotomy is the largest of the three published RCTs (it included 200 patients [5]), we believe that it is important to include data from this trial in any meta-analytical comparison. The data we have previously presented using Risk Difference question the statistical significance of the difference in incidence of incomplete pyloromyotomy [7]. As the question of incomplete pyloromyotomy is so important, it is disappointing that Jia et al. have used the “default” method for meta-analysing incidences rather than more carefully considering the nature of the available data. Hence, their conclusions are similar to those of Sola and Neville [6].

For the present, the debate will continue to run as to whether the incidence of incomplete pyloromyotomy is higher following laparoscopic pyloromyotomy compared to an open procedure. This appears to be the sticking point for the critics of a laparoscopic procedure. Only once formally addressed can the issue be laid to rest. How then to set about answering those critics, when 3 RCTs and 2 meta-analyses have failed to do so? The problem relates to the low incidence of this complication and therefore the large number of subjects required, in comparative groups, to demonstrate whether a difference (or lack of one) truly exists. Based on a difference in the incidence of incomplete pyloromyotomy of 2.5%, a sample size of over 600 patients would be required to demonstrate such a difference. The challenge for paediatric surgeons intent on demonstrating the equality of these operative approaches is now to design the large-scale studies required to prove it.

References

• 1 Ostlie DJ, St Peter SD. The current state of evidence-based pediatric surgery.  J Pediatr Surg. 2010;  45 1940-1946
• 2 Hall NJ, Van Der Zee J, Tan HL. et al . Meta-analysis of laparoscopic versus open pyloromyotomy.  Ann Surg. 2004;  240 774-778
• 3 Hall NJ, Pacilli M, Eaton S. et al . Recovery after open versus laparoscopic pyloromyotomy for pyloric stenosis: a double-blind multicentre randomised controlled trial.  Lancet. 2009;  373 390-398
• 4 Leclair MD, Plattner V, Mirallie E. et al . Laparoscopic pyloromyotomy for hypertrophic pyloric stenosis: a prospective, randomized controlled trial.  J Pediatr Surg. 2007;  42 692-698
• 5 St Peter SD, Holcomb III GW, Calkins CM. et al . Open versus laparoscopic pyloromyotomy for pyloric stenosis: a prospective, randomized trial.  Ann Surg. 2006;  244 363-370
• 6 Sola JE, Neville HL. Laparoscopic vs open pyloromyotomy: a systematic review and meta-analysis.  J Pediatr Surg. 2009;  44 1631-1637
• 7 Eaton S, Hall NJ, Pierro A. Zero-total event trials and incomplete pyloromyotomy.  J Pediatr Surg. 2009;  44 2434-2435

Correspondence