The multi-disciplinary management of complex congenital and acquired tracheo-oesophageal fistulae.

AIM OF THE STUDY: Complex tracheo-oesophageal fistulae (TOF) are rare congenital or acquired conditions in children. We discuss here a multidisciplinary (MDT) approach adopted over the past 5 years. METHODS: We retrospectively collected data on all patients with recurrent or acquired TOF managed at a single institution. All cases were investigated with neck and thorax CT scan. Other investigations included flexible bronchoscopy and bronchogram (B&B), microlaryngobronchoscopy (MLB) and oesophagoscopy. All cases were subsequently discussed in an MDT meeting on an emergent basis if necessary. MAIN RESULTS: 14 patients were referred during this study period of which half had a congenital aetiology and the other half were acquired. The latter included button battery ingestions (5/7) and iatrogenic injuries during oesophageal atresia (OA) repair. Surgical repair was performed on cardiac bypass in 3/7 cases of recurrent congenital fistulae and all cases of acquired fistulae. Post-operatively, 9/14 (64%) patients suffered complications including anastomotic leak (1), bilateral vocal cord paresis (1), further recurrence (1), and mortality (1). Ten patients continue to receive surgical input encompassing tracheal/oesophageal stents and dilatations. CONCLUSIONS: MDT approach to complex cases is becoming increasingly common across all specialties and is important in making decisions in these difficult cases. The benefits include shared experience of rare cases and full access to multidisciplinary expertise.

Double-blind randomized clinical trial of percutaneous endoscopic gastrostomy versus radiologically inserted gastrostomy in children.

BACKGROUND: The aim of this RCT was to determine whether radiologically inserted gastrostomy (RIG) in children is associated with more complications than percutaneous endoscopic gastrostomy (PEG). METHODS: Children at a single tertiary children's hospital requiring a primary gastrostomy were randomized to PEG or RIG. Patients were followed by assessors blinded to the insertion method. Complications were recorded, assigned a severity score, and analysed by zero-inflated Poisson regression analysis on an intention-to-treat basis, adjusting for length of follow-up. RESULTS: Over a 3-year period, 214 children were randomized (PEG, 107; RIG, 107), of whom 100 received PEG and 96 RIG. There was no significant difference in the number of complications between PEG and RIG groups (P = 0·875), or in the complication score: patients undergoing RIG had a 1·04 (95 per cent c.i. 0·89 to 1·21) times higher complication score than those who underwent PEG (P = 0·597). Only age had an independent significant effect on complication score, with older patients having a 0·97 (0·95 to 1·00) times lower complication score per year. CONCLUSION: PEG and RIG are both safe methods of gastrostomy insertion with a low rate of major complications. Registration number: NCT01920438 ( http://www.clinicaltrials.gov).

Umbilical hernia following gastroschisis closure: a common event?

PURPOSE: To assess incidence and natural history of umbilical hernia following sutured and sutureless gastroschisis closure. METHODS: With audit approval, we undertook a retrospective clinical record review of all gastroschisis closures in our institution (2007-2013). Patient demographics, gastroschisis closure method and umbilical hernia occurrence were recorded. Data, presented as median (range), underwent appropriate statistical analysis. RESULTS: Fifty-three patients were identified, gestation 36 weeks (31-38), birth weight 2.39 kg (1-3.52) and 23 (43 %) were male. Fourteen patients (26 %) underwent sutureless closure: 12 primary, 2 staged; and 39 (74 %) sutured closure: 19 primary, 20 staged. Sutured closure was interrupted sutures in 24 patients, 11 pursestring and 4 not specified. Fifty patients were followed-up over 53 months (10-101) and 22 (44 %) developed umbilical hernias. There was a significantly greater hernia incidence following sutureless closure (p = 0.0002). In sutured closure, pursestring technique had the highest hernia rate (64 %). Seven patients underwent operative hernia closure; three secondary to another procedure. Seven patients had their hernias resolve. One patient was lost to follow-up and seven remain under observation with no reported complications. CONCLUSIONS: There is a significant umbilical hernia incidence following sutureless and pursestring sutured gastroschisis closure. This has not led to complications and the majority have not undergone repair.

Esophageal atresia and malrotation: what association?

INTRODUCTION: Esophageal atresia/tracheo-esophageal fistula (EA/TEF) has an incidence of approximately 1:3,500. The incidence of malrotation is thought to be 1:200-500. We attempted to define the incidence of a combination and discuss the implications. METHODS: This was a retrospective review of all patients admitted to a single institution with a diagnosis of EA or EA/TEF or TEF between April 1981 and January 2013. Patients were included if the position of the duodeno-jejunal flexure (DJF) was determined by upper GI contrast study (UGIS), surgery or post-mortem. RESULTS: Case notes were reviewed for 235 patients. In the EA type A group, 3/28 (11 %; 95 % CI 3.7-27.2 %) had malrotation, significantly higher than the reported incidence of malrotation in the general population (p = 0.0008). All three patients in this group were symptomatic with one patient found to have a volvulus at emergency surgery. In the type C group, 6/196 (3 %, 95 % CI 1.4-6.5 %) had malrotation, significantly higher than the incidence reported for the general population (p = 0.0033) but not significantly different to that of the type A group (p = 0.0878). There were no patients with malrotation identified in any other EA/TEF type. In total, 9/235 (3.8 %; 95 % CI 2.0-7.2 %) patients with EA had malrotation, significantly higher than the 5/1,050 (0.48 %) reported for the general population (p = 0.0002). CONCLUSION: There is a high incidence of malrotation in patients with pure EA. In the type A group an attempt to identify the DJF position at gastrostomy siting and/or performance of UGIS in the neonatal period should be undertaken. There should also be a low threshold for UGIS in all EA/TEF patients.

The role of parenteral nutrition following surgery for duodenal atresia or stenosis.

PURPOSE: In our institution, some children routinely receive parenteral nutrition (PN) following surgery for duodenal atresia/stenosis, while others do not. Our aim was to compare growth and infection rate between these two treatment strategies. METHODS: This was a retrospective study of all children undergoing surgery for duodenal atresia/stenosis over 7 years. RESULTS: Of the 54 children, 19 commenced PN soon after surgery (the 'Initial PN' group). Of the remaining 35 children, 13 (37 %) subsequently required PN (the 'Delayed PN' group). The remaining 22 never received PN (the 'Never PN' group). The proportion of patients experiencing clinically suspected sepsis was higher in those receiving PN ('Initial' plus 'Delayed'; 41 %) compared with those who never received PN (14 %; p = 0.04). The 'Initial PN' and 'Never PN' groups did not show a significant change in weight Z score over time. However, the 'Delayed PN' group showed a significant decrease in weight Z scores from the time of operation to the time of achieving full enteral feeds, and failed to catch up by the time of last follow-up. CONCLUSION: Children with duodenal atresia/stenosis can be managed without PN. However, a third of these children subsequently require PN, lose weight centiles, and have a high rate of sepsis.

Randomized clinical trial of glutamine-supplemented versus standard parenteral nutrition in infants with surgical gastrointestinal disease.

BACKGROUND: Addition of glutamine to parenteral nutrition in surgical infants remains controversial. The aim of this trial was to determine whether glutamine supplementation of parenteral nutrition in infants requiring surgery would reduce the time to full enteral feeding and/or decrease the incidence of sepsis and septicaemia. METHODS: A prospective double-blind multicentre randomized clinical trial was performed in surgical infants less than 3 months old who required parenteral nutrition. Patients were allocated to treatment or control groups by means of minimization. Infants received either 0·6 g per kg per day alanyl-glutamine (treatment group) or isonitrogenous isocaloric parenteral nutrition (control group) until full enteral feeding was achieved. Primary outcomes were time to full enteral feeding and incidence of sepsis. Cox regression analysis was used to compare time to full enteral feeding, and to calculate risk of sepsis/septicaemia. RESULTS: A total of 174 patients were randomized, of whom 164 completed the trial and were analysed (82 in each group). There was no difference in time to full enteral feeding or time to first enteral feeding between groups, and supplementation with glutamine had no effect on the overall incidence of sepsis or septicaemia. However, during total parenteral nutrition (before the first enteral feed), glutamine administration was associated with a significantly decreased risk of developing sepsis (hazard ratio 0·33, 95 per cent confidence interval 0·15 to 0·72; P = 0·005). CONCLUSION: Glutamine supplementation during parenteral nutrition did not reduce the incidence of sepsis in surgical infants with gastrointestinal disease. REGISTRATION NUMBER: ISRCTN83168963 (http://www.controlled-trials.com).

'See one, practise on a simulator, do one' - the mantra of the modern surgeon.

Minimally invasive techniques are now well established in paediatric surgery. Training has traditionally been based on an apprenticeship model, as for open surgery. More recently the constraints of litigation, finance and restriction of doctors' working hours have led to a need to rethink this training. Simulation to learn and improve skills is by no means a new concept, but has been suggested as a way to address the above issues because it provides an ideal platform for acquiring the necessary skills for modern laparoscopic surgery. This paper explores some of the current issues of learning minimally invasive surgical skills in a simulated environment, and suggests that such simulation should not be seen in isolation but as a part of a wider and encompassing curriculum of learning for the 21st-century surgeon.

Surgical repair of incarcerated inguinal hernia in children: laparoscopic or open?

PURPOSE: The management of Incarcerated Inguinal Hernia (IIH) in children is challenging and may be associated with complications. We aimed to compare the outcomes of laparoscopic vs. open repair of IIH. METHODS: With institutional ethical approval (09SG13), we reviewed the notes of 63 consecutive children who were admitted to a single hospital with the diagnosis of IIH between 2000 and 2008. Data are reported as median (range). Groups were compared by chi-squared or t-tests as appropriate. RESULTS: · Open repair (n=35): There were 21 children with right and 14 with left IIH. 2 patients also had contralateral reducible inguinal hernia. Small bowel resection was required in 2 children. · Laparoscopic repair (n=28): All children had unilateral IIH (19 right sided, 9 left sided). 15 children (54%) with no clinical evidence of contralateral hernia, had contralateral patent processus vaginalis at laparoscopy, which was also repaired. The groups were similar with regard to gender, age at surgery, history of prematurity, interval between admission and surgery, and proportion of patients with successful preoperative manual reduction. However, the duration of operation was longer in the laparoscopy group (p=0.01). Time to full feeds and length of hospital stay were similar in both groups. Postoperative follow-up was 3.5 months (1-36), which was similar in both groups. 5 patients in the group undergoing open repair had serious complications: 1 vas transaction, 1 acquired undescended testis, 2 testicular atrophy and 1 recurrence. The laparoscopic group had a single recurrence. CONCLUSION: Open repair of incarcerated inguinal hernia is associated with serious complications. The laparoscopic technique appears safe, avoids the difficult dissection of an oedematous sac in the groin, allows inspection of the reduced hernia content and permits the repair of a contralateral patent processus vaginalis if present.

Gastric emptying is accelerated following laparoscopic Nissen fundoplication.

AIM OF THE STUDY: The effects on gastric motility following Nissen fundoplication in children are poorly documented. Some paediatric surgeons advocate additional procedures at the same time as fundoplication, such as a pyloroplasty, to enhance gastric emptying. The aim of this study was to determine whether laparoscopic Nissen fundoplication without pyloroplasty affects gastric emptying. METHODS: Gastric emptying was measured before laparoscopic Nissen fundoplication in 8 children after ingestion of a standardised volume of milk for age mixed with 150 mg of (13)C-octanoic acid. None of the patients had a gastrostomy insertion at the time of fundoplication and 2 patients had neurological impairment. Breath samples were collected by breathing into a mask at baseline and every 15 minutes up to 3 hours, and were analysed for (13)CO (2)/ (12)CO (2) ratio by mass spectrometry. Gastric emptying time (t (1/2)) was derived from the curve of (13)CO (2)/ (12)CO (2) ratio against time. The test was repeated in 6 children following Nissen fundoplication at the time of full feeds. Data are reported as mean +/- SD and were analysed by the Mann-Whitney test. RESULTS AND CONCLUSIONS: There were 4 males and 4 females; mean age at surgery was 3.3 +/- 3.0 years. Mean gastric emptying time was 59 +/- 17 min prior to laparoscopic Nissen fundoplication and 45 +/- 4 min following surgery (p = 0.03). Gastric emptying was accelerated in all except one patient. Gastric emptying for liquids is accelerated following Nissen fundoplication in children. Procedures aimed at improving gastric emptying time such as pyloroplasty or pyloromyotomy might not be justified at the time of laparoscopic Nissen fundoplication.

Laparoscopic and conventional appendicectomy in children: outcomes in English hospitals between 1996 and 2006.

BACKGROUND: Laparoscopic appendicectomy is increasingly used in children. This national retrospective study compared outcomes of paediatric open and laparoscopic appendicectomy. METHODS: Length of stay, readmission rates and mortality in children undergoing open and laparoscopic appendicectomy in English NHS Trusts between 1 April 1996 and 31 March 2006 were compared. Procedures coded as emergency excision of appendix (OPCS-4 H01) on the Hospital Episode Statistics (HES) database in patients less than 15 years of age were included. Multivariate analysis was used to identify independent predictors of length of hospital stay and mortality. RESULTS: Eighty-nine thousand, four-hundred and ninety-seven (89,497) appendicectomies were studied; of which, 2,689 (3%) were performed laparoscopically. The percentage of laparoscopic cases rose from 0.6 to 8.4% between 1996 and 2006 (Pearson's r = 0.954, P < 0.001). Length of stay (median 3, interquartile range 2 days, P = 0.068) and 28-day readmission rates were similar (6.3 vs. 7.2%, respectively; P = 0.072) between groups. No independent hospital stay advantage for laparoscopy was observed (P = 0.121). No difference in 30-day mortality (P = 0.986) or 365-day mortality (P = 0.598) was demonstrated. CONCLUSION: Hospital stay, readmission rates and mortality are similar following laparoscopic and open appendicectomy in children.

Current status of prenatal diagnosis, operative management and outcome of esophageal atresia/tracheo-esophageal fistula.

Ultrasonographic features suggestive of esophageal atresia with or without tracheo-esophageal fistula (EA/TEF) are only in a small minority of fetuses with EA/TEF (<10%) identifiable on prenatal scans.The prenatal diagnosis of EA/TEF relies in principle, on two nonspecific signs: polyhydramnios and absent or small stomach bubble. Polyhydramnios is associated with a wide range of fetal abnormalities, but most commonly it pursues a benign course. Similarly the sonographic absence of a stomach bubble may point to a variety of fetal anomalies.The combination of polyhydramnios and absent stomach bubble in two small series offers a modest positive predictive value of 44 and 56% respectively. Prenatal scanning for EA/TEF identifies a larger proportion of fetuses with Edwards syndrome; there is also a higher proportion of isolated EA in comparison to postnatal studies.Current ultrasound technology does not allow for a definite diagnosis of EA/TEF and therefore, counseling of parents should be guarded.Postnatal diagnosis of EA is confirmed by the failure to pass a firm nasogastric tube into the stomach; on chest X-ray, the tube is seen curling in the upper esophageal pouch. Corrective surgery for EA/TEF is well established and survival rates of over 90% can be expected.

Umbilical polyp in infants and children.

AIM OF THE STUDY: The umbilical polyp is a rare congenital lesion resulting from the persistence of omphalomesenteric duct (OMD) enteric mucosa at the umbilicus. Exploration of the abdomen to exclude the presence of associated OMD remnants is controversial. The aim of this study was to evaluate the need for peritoneal cavity exploration in children with umbilical polyp. METHODS: All umbilical lesions (n = 53) excised between 1995 and 2005 in a single institution were reviewed to identify patients with umbilical polyp (n = 13). This is characterised histologically by the presence of gastrointestinal mucosa. A follow-up study of patients with umbilical polyp was performed. Data are reported as median (range). RESULTS AND CONCLUSIONS: All 13 patients underwent excision of an umbilical polyp at a median age of 15.1 months (3.1 - 80.5). All presented with a discharging polyp (associated with bleeding in 9) which did not respond to topical silver nitrate. Median diameter of the lesions was 0.5 cm (0.2 - 1). Histology revealed the presence of small bowel mucosa in 11 (associated with pancreatic tissue in 1 and gastric mucosa in 1) and large bowel mucosa in 2. All patients underwent inspection and probing of the base of the polyp after its excision. In 6 patients an associated OMD anomaly was suspected and exploration of the peritoneal cavity was performed (mini-laparotomy in 5 and laparoscopy in 1). No OMD anomaly was found. The 7 children who did not undergo exploration of the abdominal cavity remain asymptomatic after 5.8 years (0.9 - 13.7) follow-up. An umbilical polyp can be present in the absence of other OMD anomalies. Exploration of the peritoneal cavity in children with an umbilical polyp does not seem to be necessary.

Surgical skills training in the laparoscopic era: the use of a helping hand.

The aim of this study was to evaluate whether a training programme involving an assessment exercise performed on a laparoscopic trainer model leads to an improvement in the acquisition of laparoscopic skills in surgical trainees. Subjects were recruited from a cross-section of surgical trainees at the Great Ormond Street Hospital, Department of Surgery and the Institute of Child Health. All subjects completed both a baseline laparoscopic surgical skills questionnaire and three exercises on a new laparoscopic trainer model. Thirteen subjects completed both the baseline questionnaire and all three assessment exercises. These subjects exhibited a wide range of previous experience in laparoscopic surgery. Sixty-nine percent of subjects showed a significant improvement in the assessment exercise score with training (ANOVA; P = 0.01). Sixty-two percent of subjects showed a greater improvement between exercises 2 and 3 than between exercises 1 and 2. The difference in score between exercises 1 and 2 was not statistically significant (P = 0.597), whereas the difference in score between both exercises 2 and 3 and exercises 1 and 3 was statistically significant (P = 0.018 and P = 0.005, respectively). The double glove training model is thus a simple, inexpensive, and easily reproducible tool that elicits a significant improvement in laparoscopic surgical skills in surgical trainees with a broad range of previous laparoscopic experience. It can therefore be used as part of a training programme to facilitate the acquisition of laparoscopic skills in a paediatric surgery setting.

Pilot study of new training model for laparoscopic surgery.

The aim of this study was to evaluate whether an assessment exercise performed on a laparoscopic trainer model reliably reflects previous laparoscopic experience and can therefore be used to accurately assess laparoscopic skills in surgical trainees. Subjects were recruited from a cross-section of surgical trainees and students at the Great Ormond Street Hospital for Children and the Institute of Child Health. Subjects were required to complete a baseline laparoscopic surgical skills questionnaire and an exercise on a new laparoscopic trainer model. Nine subjects completed both the baseline questionnaire and the exercise. These subjects exhibited a wide range of previous experience in laparoscopic surgery. Subjects with higher self-assessment scores had the lowest exercise scores (i.e. better scores; P=0.003). Furthermore, the exercise score was strongly negatively correlated with the baseline number of training modalities received (P=0.007) and the laparoscopic experience score (P=0.027). The assessment exercise on a novel laparoscopic trainer was capable of differentiating between subjects with little laparoscopic experience and those with more extensive previous laparoscopic training. The correlation between the exercise score and measured baseline variables suggests that the scoring system used in this model is sensitive and specific to measuring skills relevant to laparoscopic surgery.

Absorption of carbon dioxide during laparoscopy in children measured using a novel mass spectrometric technique.

BACKGROUND: Carbon dioxide (CO(2)) is absorbed during pneumoperitoneum and may cause adverse haemodynamic effects. The aim of this study was to measure the elimination of exogenous CO(2) during laparoscopy in children. METHODS: Ten children [27.6 (56.5) months; mean (SD)] undergoing laparoscopic and nine [24.5 (17.3) months] undergoing open surgery were studied. Breath samples were collected at the line for end-tidal CO(2) and analysed for (13)CO(2)/(12)CO(2) ratio expressed as deltaPDB (difference from standard), by isotope-ratio mass spectrometry. The proportion of absorbed CO(2) was calculated comparing exhaled (13)CO(2)/(12)CO(2) before and during CO(2) pneumoperitoneum. RESULTS: (13)CO(2)/(12)CO(2) in medical CO(2) was -32.7 (2.1) deltaPDB. (13)CO(2)/(12)CO(2) in breath of patients undergoing open procedures was -24.3 (2.4) deltaPDB at the start of operation and did not change during the operation (P > 0.2). (13)CO(2)/(12)CO(2) in breath of patients undergoing laparoscopy was -21.5 (5.4) deltaPDB at the start of insufflation, and decreased during pneumoperitoneum by 2.5 (1.6) deltaPDB, indicating absorption of exogenous CO(2). The percentage of expired CO(2) absorbed rose to 15.5 (7.7)% after 30 min of pneumoperitoneum and decreased rapidly after desufflation. CONCLUSION: After 10 min of laparoscopy 10-20% of expired CO(2) derives from the exogenous CO(2). CO(2) absorption can be measured using a simple mass spectrometric technique.

A multicenter, randomized, double-blind, placebo-controlled trial of the prokinetic agent erythromycin in the postoperative recovery of infants with gastroschisis.

BACKGROUND/PURPOSE: The recovery of gut function after repair of gastroschisis is frequently prolonged, and these infants are prone to complications associated with parenteral nutrition. This trial was designed to investigate the effect of the prokinetic agent, erythromycin, on the attainment of full enteral feeding in infants after primary repair of uncomplicated gastroschisis. METHODS: A multicenter, randomized, double-blind, placebo-controlled trial was used to investigate the effect of enteral erythromycin (3 mg/kg/dose 4 times daily) compared with placebo on the attainment of full enteral feeding tolerance after primary repair of uncomplicated gastroschisis. Eleven neonatal surgical units in the United Kingdom participated in the study. The primary end-point was the time taken to achieve continuous enteral feeding at 150 mL/kg/24 hours sustained for 48 hours. RESULTS: Of 70 eligible infants, 62 were recruited and randomly divided. There were 30 patients in group I (placebo) and 32 in group II (erythromycin). The groups were comparable in terms of mean gestational age, mean birth weight, extent of evisceration, and degree of intestinal peel. There was no statistically significant difference between the 2 groups in the time taken to achieve full enteral feeding (27.2 v 28.7 days; P =.75). Similarly, no significant differences were found in the incidence of catheter-related sepsis, duration of parenteral nutrition, or time to discharge between the 2 groups. CONCLUSIONS: Enterally administered erythromycin at a dose of 3 mg/kg 4 times daily conferred no advantage in the time taken to achieve full enteral feeding after primary repair of uncomplicated gastroschisis.

Range of motion after arthroplasty for the stiff osteoarthritic knee.

In 28 of 1656 total knee arthroplasties (TKAs) (1.7%) performed for osteoarthritis at this institution, the preoperative arc of motion was 60 degrees or less (average 47.5 degrees; range, 20-60 degrees). The outcome of 22 of the 28 TKAs (21 patients) is reported at a mean follow-up of 52.9 months (range, 24-144 months). Eighty-two percent of the cases were managed with standard soft-tissue releases and posterior cruciate-retaining implants and 18% with a cruciate-substituting design. The joint score rose from 28.8 to 82.2 and the Knee Society Score from 24.6 to 77 points. The mean postoperative arc of motion improved by 46 degrees-93.5 degrees. At latest follow-up, 68.2% of knees achieved maximal flexion of 90 degrees or more. Manipulation was performed in 22.7%. Complications were minimal. A functionally useful range of motion is possible after total knee arthroplasty in the majority of stiff osteoarthritic knees, often without the need for posterior cruciate substitution.

Body mass index as a predictor of outcome in total knee replacement.

The clinical and radiographic outcomes of 326 total knee replacements (TKR) in 285 osteoarthritic patients with body mass index (BMI) greater than 30 kg/m2 were compared with the results of a matched group of 425 TKR in 371 patients with BMI less than 30 kg/m2. At an average follow-up of 75.9 (48-144) months the Knee Society score (KSS) in the obese patients had increased by 41.9 points, and the joint score by 43.7. In the non-obese group the KSS rose by 40.2 points and the joint score by 42.6 points. Although patients with BMI greater than 40 kg/m2 achieved a lower final KSS the 'absolute improvement' appeared to be independent of BMI. Of the obese patient group 4.9% underwent a revision of their TKR, compared with 3.1% of the non-obese group. Although linear osteolysis (radiolucency) rates were comparable, focal osteolysis rates were 5 times those of control subjects when the BMI exceeded 40 kg/m2. Ten-year survivorship figures were similar for both obese and non-obese patients.

Review article: erythromycin as a prokinetic agent in infants and children.

Erythromycin has been used as an antibiotic for more than four decades, but only in the last 10 years have other therapeutic benefits of this agent been exploited. Animal and human studies have demonstrated a prokinetic effect on the gastrointestinal tract at sub-antimicrobial doses (typically a quarter or less of the antibiotic dose). A limited number of studies have been performed in children to investigate this action. A review of this literature is particularly pertinent given the frequency of clinical problems related to gastrointestinal dysmotility in children and the limited availability of prokinetic agents in paediatric practice, compounded by the recent withdrawal of cisapride. The prokinetic effects of erythromycin have been investigated in infants with dysmotility associated with prematurity, in low birth-weight infants recovering from abdominal surgery, and in older children with a variety of other gastrointestinal disorders. Only one randomized placebo-controlled trial has been conducted. All except one of these studies have shown a beneficial effect of erythromycin in either promoting tolerance of enteral feeds or enhancing a measured index of gastrointestinal motility. Erythromycin appears to be equally effective when given orally (as ethylsuccinate or estolate) or intravenously (as lactobionate). Significantly, no serious adverse effects have been reported in studies in which erythromycin has been used for its prokinetic effects, although fatal reactions have followed the intravenous administration of erythromycin to neonates in antibiotic doses.