Statistics from Altmetric.com
Editor,—I read with interest the annotation by Lander and Desai discussing cisapride.1 They conclude that the case against cisapride as a safe and useful agent in premature neonates has not been proved.
As the three main indications for cisapride in the newborn (gastro-oesophageal reflux, postoperative ileus, and intestinal dysmotility of prematurity) are self resolving with time, the only useful study design is a randomised controlled trial (RCT).
There are two small published RCTs on the use of cisapride in the newborn. Enriquez et al randomised infants at the initiation of feeds and reported no difference between cisapride and placebo in the time taken to full feeds.2 Landeret al randomised neonates > 32 weeks’ gestation with postoperative ileus to cisapride or placebo for seven days.3 They noted a significant increase in the mean net enteral balance with cisapride. While this is encouraging, the trial design did not permit comparison of time to full feeds or hospital discharge, and it was not analysed as intention to treat. There are no published trials on the use of cisapride in premature neonates with gastro-oesophageal reflux.
It is certainly too early to state that cisapride is unsafe in premature infants; however, there is insufficient evidence of its efficacy in the newborn. It is surely no longer acceptable that a drug with no objective evidence of efficacy and potentially serious side effects should be in widespread use.
Large placebo controlled trials with clinically important outcome measures are needed to establish whether cisapride is an effective treatment for neonates with gastro-oesophageal reflux, postoperative ileus, or intestinal dysmotility of prematurity. Prospective monitoring of the QTc interval with drug discontinuation in the event of prolongation should allow measurement of cardiac risk without jeopardising the health of any infant.
Dr Lander replies:
Dr Cairns is right—robust, placebo controlled trials of cisapride are needed in both the premature and mature newborn, especially for those with problematic reflux or postoperative ileus. Unfortunately, the contraindication by the Committee on Safety of Medicines (CSM)1-1 will deter clinicians, ethics committees, and parents from these crucial trials. The British Association of Paediatric Surgeons abandoned plans for a multicentre study of cisapride against placebo after the repair of gastroschisis for this very reason. We are now looking at the efficacy of erythromycin in these children.
Measures of clinical outcome are the most useful end points in trials, but of equal importance in interpreting the clinical usefulness of results are entry criteria. How do the subjects compare with those one needs to treat? In self resolving conditions there is no need to treat routinely unless the morbidity of the time to resolution is likely to be clinically important. Cairns cites Enriquezet al who reported no significant difference in the time to establishing enteral feeds between placebo and cisapride when used routinely in 34 preterm infants.1-2 However, the following were noted: seven of 18 infants given cisapride had no significant gastric residue at any time compared with only one of 16 given placebo (p = 0.04); and only two of 18 given cisapride developed regurgitation compared with eight of 16 given placebo (p = 0.02). These observations are important. Furthermore, two babies with regurgitation were removed from their study and treated openly with cisapride with success. Although they cannot be included in the analysis, they show that the study population was limited to less worrying infants.
Unless the CSM revisits the question of prematurity, robust trials with cisapride, which is still widely used,1-3 will not get ethical approval and a good drug may be lost.
Please see “Randomised controlled trial of cisapride in preterm infants” by R J McClure, J H Kristensen, A Grauaug in the May issue of Fetal and Neonataledition (Arch Dis Child1999;80:F174–8).