Pediatric Hypertrophic Pyloric Stenosis Surgery

Pyloric stenosis involves hypertrophy of the circular muscle of the pylorus, resulting in narrowing and obstruction of the pyloric channel by compression of longitudinal folds of mucosa. Gastric distention results (see the image below). Gastric outlet obstruction results in emesis, which is characteristically nonbilious and projectile. Protracted emesis, as well as failure of the stomach to empty into the duodenum, results in progressive dehydration, electrolyte abnormalities, acid-base disorders, weight loss, and, potentially, shock.

Intraoperatively, the surgeon must pay strict attention to the serosal demarcation between the duodenum and the pylorus. The prepyloric vein, or Mayo vein, is located at this junction. The risk of duodenal perforation is prevented by stopping the distal extent of the myotomy 1-2 mm short of this point.

Pyloric stenosis involves hypertrophy of the circular muscle of the pylorus, resulting in narrowing and obstruction of the pyloric channel by compression of longitudinal folds of mucosa. Grossly, the pylorus is enlarged, resembling a tumor approximating the size and shape of an olive (ie, 2 cm long and 1 cm in diameter). Microscopically, the circular muscle hypertrophies, with increased connective tissue in the septa between the muscle bundles. An increase of chondroitin sulfate within the extracellular matrix may account for the cartilaginous quality of the pyloric tumor.

Gastric fluid loss is associated with the loss of H+ and Cl–. This fluid loss is unlike that in conditions caused by vomiting with an open pylorus, which involves losses of gastric, pancreatic, biliary, and intestinal fluid. Hypochloremic hypokalemic metabolic alkalosis is the characteristic biochemical disturbance observed in pyloric stenosis. Urinary Na+ and HCO3– losses, which compensate for Cl– losses, perpetuate this alkalosis.

With protracted vomiting, an extracellular volume deficit ensues, and urinary excretion of K+ and H+ increases in an attempt to preserve Na+ and volume. The initially alkalotic urine then becomes acidotic (paradoxic aciduria). This sign of protracted dehydration should alert the clinician to the severity of the volume and total body K+ deficit. The severity of electrolyte abnormalities depends on the duration of vomiting before resuscitation.

Greater awareness of the presenting signs of pyloric stenosis by pediatricians and primary care physicians, along with examination by means of ultrasonography (US), has resulted in earlier diagnosis and less severe electrolyte and acid-base abnormalities.

No conclusive evidence for the etiology of pyloric stenosis exists; however, both hereditary and environmental influences are believed to be contributing factors. Multiple factors, including both neural and hormonal, have been implicated but not substantiated in the development of pyloric stenosis. An association with B and O blood groups and maternal stress during the third trimester has also been suggested. Although pyloric stenosis is now believed to be acquired, cases of pyloric stenosis diagnosed antenatally and in neonates have been reported.

Since 1976, several reports and cohort retrospective studies have appeared in the literature suggesting an association between pyloric stenosis and exposure to macrolide antibiotics (erythromycin). In 2002, Cooper et al suggested that early exposure to erythromycin (at 3-13 days of life) is associated with a nearly eightfold increased risk of pyloric stenosis (adjusted incident-rate ratio, 7.88).[3] No increased risk of pyloric stenosis was observed in infants exposed to erythromycin after 13 days of life.

In 1993, Huang et al, by homologous recombination, generated mutant mice (knockout mice) lacking the neuronal nitric oxide (NO) synthase (NOS) gene.[4] NO mediates nonadrenergic noncholinergic smooth-muscle relaxation throughout the gut. The stomachs of homozygous mutant mice were larger than normal in this group, and the circular muscle layer of the stomach and pylorus was hypertrophied. Wild-type mouse stomachs contained NOS in the myenteric plexus and nerve fibers of the circular muscle layer, whereas mutant homozygous mice lacked NOS in both locations.

Applying these observations to the human condition, Huang et al hypothesized that the stomach and pylorus may be particularly dependent on NO and prone to dysfunction in its absence.[4] Although human pyloric stenosis does not appear to be due to a complete absence of neuronal NOS gene product, the absence of NOS in this area may result in pyloric smooth-muscle hypertrophy.

In a 2006 study, Huang et al collected biopsy samples of the pylorus in 13 patients with infantile hypertrophic pyloric stenosis, found decreased expression of neuronal NOS, and demonstrated that plasma nitrite levels can be valuable for diagnosing pyloric stenosis.[5]

With regard to other factors that contribute to smooth-muscle control and hypertrophy, one study of 81 pyloric stenosis pedigrees used single nucleotide polymorphism (SNP)-based linkage analysis to identify two pertinent functional genes on loci 11q14-22 and Xq23.[6] These areas are thought to play a part in the canonical transient receptor potential (TRPC) family of ion channels and may contribute to the development of pyloric stenosis in infants.

Associated anomalies, though rare, have been reported with pyloric stenosis. Approximately 4-7% of infants with pyloric stenosis have associated anomalies, with hiatal and inguinal hernias being the most common. Other anomalies include congenital heart disease, esophageal atresia, tracheoesophageal fistulas, renal abnormalities, rubella, and chromosomal abnormalities such as Turner syndrome and trisomy 18.

In 1993, Jackson et al found that 3.8% of infants (12/308) with de Lange syndrome had pyloric stenosis.[7] Infants with a developmental delay called FG syndrome[8] and those with Smith-Lemli-Opitz (SLO) syndrome, a type of cholesterol deficiency, were reported to be at increased risk for pyloric stenosis. Additionally, Liede et al proposed a convincing argument of a common genetic association between endometriosis, breast cancer, and pyloric stenosis in several families.[9]

Reports of pyloric stenosis in the United States have cited frequency figures ranging from as low as 1 case per 3000-4000 live births to as high as 8.2-12 cases per 1000 live births. The reported incidence has been decreasing.[10]

This condition is most commonly observed in whites of northern European descent, is less frequently observed in blacks, and is rarely found in patients of Asian or East Indian ancestry.[11, 12] Location also contributes to frequency, with areas where the population is more than two-thirds rural showing an increased risk of 1.79.

Pyloric stenosis is more common in males than in females (male-to-female ratio, 4:1). The highest incidence is in first-born males. A genetic predisposition is suggested in families with occurrences of pyloric stenosis reported in at least three generations. Involvement in twins has been reported, with an 85.7% concordance rate in monozygotic twins and an 8.4% concordance rate in dizygotic twins.

In 1969, Carter and Evans suggested a sex-modified polygenic inheritance of pyloric stenosis.[13] Data from more than 1200 families demonstrated a 20% risk in sons and a 7% risk in daughters of females having had pyloric stenosis, whereas data showed only a 5% risk in sons and a 2.5% risk in daughters of males with pyloric stenosis.

Another report showed a 29% increased risk associated with younger maternal age (< 20 y), whereas a maternal age exceeding 30 years was associated with a significantly decreased risk.[14]

Pyloromyotomy that is adequately performed is curative of pyloric stenosis. There have been reports of pyloric stenosis recurring despite performance of an adequate pyloromyotomy, but recurrence is considered to be a rare exception after incomplete pyloromyotomy has been ruled out.

Yoshizawa et al demonstrated in studies using US imaging that after pyloromyotomy, the pylorus changes significantly within 3 days postoperatively and returns to normal within 5 months.[15] Specific changes include the following:

• The dorsal pyloric aspect temporarily thickens from 5.1 ± 0.8 mm to 6.0 ± 0.3 mm within 3 days postoperatively and thins out to 2.8 ± 0.2 mm within 5 months
• The pyloric length decreases from 20.1 ± 2.9 mm preoperatively to 16.9 ± 2.7 mm within 3 days postoperatively and to less than 15 mm within 4 months
• The change in pyloric diameter is comparable to the change in pyloric length
• The transverse muscle thickness of the incision site changes from 4.3 ± 0.4 mm to 4.6 ± 0.4 mm within 3 days postoperatively and to 2.1 ± 0.9 mm within 7 days (normal, < 3 mm)

Several studies focused on patient outcomes with respect to advanced training and experience of surgeons. In 2002, Pranikoff et al reported that pediatric surgeons performing pyloromyotomy had a mucosal perforation rate of 0.5%, compared with a 2.9% rate for general surgeons, and that this difference in rate of mucosal perforation correlated with significant decreases in total hospital charges ($4806 ± $79 vs $6592 ± $492) and hospital stay (2.7 ± 0.1 vs 3.1 ± 0.1 d).[16]

In a 2005 study of 11,003 patients with pyloric stenosis, Safford et al stratified patient outcomes on the basis of surgeon volume and hospital volume of pyloric stenosis cases.[17] For surgeons, low volume was considered less than one procedure per year; intermediate volume, one to five procedures per year; and high volume, more than five procedures per year. For hospitals, low volume was considered fewer than five procedures per year; intermediate volume, five to 15 procedures per year; and high volume, more than 15 procedures per year.

In this study, patients operated on by low- and intermediate-volume surgeons were more likely to have complications than those operated on by high-volume surgeons.[17] Patients operated on at low-volume hospitals were 1.6 times more likely to have complications than those at intermediate- or high-volume hospitals. Procedures performed at high-volume hospitals were less expensive than those done at intermediate-volume hospitals, by a margin of $910. High-volume surgeons were more expensive than low-volume surgeons, by a margin of $511. Low-volume surgeons at low-volume hospitals had mucosal perforation rates 4-6.7 times higher than high-volume surgeons at high-volume hospitals.

It is important to note that between 1994 and 2000, the frequency of laparoscopic pyloromyotomy was likely increasing; however, the rates of open pyloromyotomy and laparoscopic pyloromyotomy were not included in procedure coding. The data have shown that for pyloromyotomy procedures, complication rates are lower and cost savings greater with high-volume surgeons operating at high-volume hospitals.

Laparoscopic pyloromyotomy has a significant learning curve. In 2005, Hendrickson et al reported an initial operating time of 70 minutes at a teaching hospital, with operating times decreasing to 15 minutes after 25 procedures.[18] A conversion rate of 8% from the laparoscopic to the open procedure was reported. Similar learning curves were reported at other centers. It has been suggested that appropriate simulation models may help shorten the learning curve for laparoscopic pyloromyotomy.[19, 20]

In 2004, Yagmurlu et al compared open pyloromyotomy (n = 225) with laparoscopic pyloromyotomy (n = 232) and found the overall complication rates to be 4.4% for the open procedure and 5.6% for the laparoscopic procedure.[21] The open approach resulted in a higher rate of mucosal perforation (3.6% vs 0.4%), and laparoscopy had a higher rate of postoperative complications, such as incomplete pyloromyotomy (0% for open vs 2.2% for laparoscopic).

In a 2002 retrospective study, Campbell et al reported on 117 patients showing a trend toward significantly higher complication rates with laparoscopic pyloromyotomy than with the open procedure (18% vs 12%).[22] Additionally, significantly higher hospital costs were associated with the laparoscopic approach.

Also in 2002, the International Pediatric Endosurgery Group noted in their guidelines for infantile hypertrophic pyloric stenosis that laparoscopic pyloromyotomy provided cost savings, decreased operating room time, reduced tissue trauma, and improved cosmetic outcome.[23] (As of March 2023, these guidelines were being revised.)

Oomen et al concluded that laparoscopy might be acknowledged as the standard of care if the major postoperative complication rate is low.[24] To achieve this, laparoscopic pyloromyotomy should be performed by pediatric surgeons with specific expertise in this procedure.

Studies from 2018 concluded that laparoscopic pyloromyotomy for hypertrophic pyloric stenosis yielded outcomes equivalent or superior to those of open pyloromyotomy, with lower complication rates and shorter hospital stays.[25, 26]  A study from 2021 found no significant differences between the two approaches with respect to duration of surgery, postoperative complications, duration of hospitalization, and weight at the time of surgery.[27]  There may be a small increase in the risk of mucosal perforation after laparoscopic pyloromyotomy as compared with open pyloromyotomy.[28]

Pyloric stenosis most often occurs in neonates and infants aged 1-10 weeks (mean, 5 wk; range, 5 d to 5 mo). Although uncommon in premature infants younger than corrected age for a full-term infant, pyloric stenosis has been detected on antenatal sonograms and could be considered in the differential diagnoses for nonbilious vomiting in the newborn. Pyloric stenosis is observed in premature infants older than corrected age for a full-term baby.

Regardless of age, projectile vomiting typically occurs and is always nonbilious but may have brown discoloration or a coffee-ground appearance due to associated gastritis, particularly if emesis has persisted for several days. The vomiting occurs within 30-60 minutes after feeding. The infant remains hungry and usually attempts to feed immediately after vomiting. Weight loss and evidence of dehydration (eg, decreased tearing and urinary output, with poor skin turgor) are present if vomiting is allowed to continue for more than a few days.

Physical examination of the infant is conducted in a warm environment with the baby quiet or sleeping. A general sense of hydration is assessed first (see Table 1 below), with particular attention paid to the baby's level of consciousness (arousability if sleeping), eyes, fontanelles, skin turgor, mucous membranes, and tearing. Infants with depressed fontanelles and decreased skin turgor have at least a 5% deficit of total body water. The lungs should be examined carefully, and signs of aspiration pneumonia should be looked for in any infant who presents with a history of vomiting.

Table 1. Clinical Findings in Dehydrated Infants With Pyloric Stenosis (Open Table in a new window)

The infant is best examined from the right, with mild pressure applied by the first three fingers of the right hand in a cephalad direction (see the image below).

Ideally, the infant should be examined with the stomach decompressed via a nasogastric or orogastric tube, which prevents the pyloric channel from being obscured by an overlying dilated stomach. Careful examination reveals an oblong, smooth, hard mass that is 1-2 cm in length. This mass is the hypertrophied pylorus, commonly referred to as an olive or pyloric tumor, and is located in the epigastrium just above the umbilicus, either in the midline or just to the right. Although a superficially located pyloric mass may be palpated with relatively gentle pressure, identification of masses lying deeper or masses in crying infants requires firmer deep palpation.

Upon identifying a suspected olive (pyloric tumor), the examiner must attempt to outline or palpate discrete borders of the mass to avoid mistaking the liver edge, a contracted rectus muscle, or the upper pole of the right kidney for the mass. With persistence and experience, the pyloric tumor should be palpated in 85-100% of cases. Difficulty in locating the mass is encountered if the mass is obscured by the liver, a distended stomach, or tense rectus muscles in crying infants.

Feeding the patient a small volume of warm sugar water may be useful in the examination, for two reasons. First, a feeding infant cannot cry and thus does not tense the abdominal muscles, thereby making the examination of the pylorus easier. Second, observation of the abdomen of the infant with pyloric stenosis after feeding often reveals visible gastric contractions occurring in a wavelike manner from left to right across the abdomen. These waves generally terminate in emesis and are often associated with, but are not pathognomonic for, pyloric stenosis.

Further examination of the abdomen is facilitated by nasogastric decompression and by lifting the lower extremities to help relax the abdominal musculature.

An electrolyte panel is essential for estimating the state of dehydration and acidosis/alkalosis in patients with pyloric stenosis. Hypochloremic hypokalemic metabolic alkalosis is the characteristic biochemical disturbance observed in pyloric stenosis.

Liver function studies are not routinely obtained or necessary for diagnosis. Jaundice occurs in approximately 2% of infants with pyloric stenosis. Although the cause is uncertain, this finding (similar to findings in Gilbert syndrome) is thought to reflect a decrease in hepatic glucuronosyltransferase activity associated with starvation, as occurs in high gastrointestinal (GI) obstruction. The jaundice resolves spontaneously and rapidly after pyloromyotomy.

Urinalysis with normalization of urinary pH (correction of paradoxic aciduria) also helps determine adequacy of resuscitation.

Although a careful history and physical examination lead to diagnosis in the vast majority of patients with pyloric stenosis, those in whom a palpable mass is not identified require further diagnostic studies. Plain abdominal radiography may show a dilated stomach bubble, which suggests the diagnosis but should not be considered a diagnostic finding.

Upper GI (UGI) contrast studies have largely been supplanted by ultrasonography (US; see below) as the study of choice for confirming pyloric stenosis. Although UGI studies have been reported as yielding an accuracy of 96%, obvious disadvantages of such studies include radiation exposure and the risk of aspiration of contrast material.

A UGI study may be helpful in ruling out gastroesophageal reflux, duodenal atresia, and malrotation in cases in which uncertainty exists as to the nature of the emesis (ie, bilious versus nonbilious) and in which a pyloric mass is indiscernible. Failure of gastric emptying demonstrated on UGI studies is not diagnostic of pyloric stenosis, because pyloric spasm and central nervous system (CNS) lesions may be associated with delayed gastric emptying. (See the image below.)

In 1977, Teele and Smith first described the use of US in the diagnosis of pyloric stenosis.[29]  Objective criteria in measuring the pylorus have increased the diagnostic accuracy of US.

On US, a diagnosis of pyloric stenosis can be made through identification of an elongated sausage-shaped mass with the following characteristics (see the image below)[30] :

• Pyloric diameter >14 mm
• Muscular thickness >4 mm
• Length >16 mm

A sensitivity of 91-100% and a specificity of 100% have been reported with these criteria. US is best performed with the stomach evacuated; food or milk curds may interfere with the study. If a hypertrophied pylorus is not demonstrated on US, a UGI examination should follow to assess for other causes of vomiting.

In 2008, Leaphart et al reported that US in infants younger than 21 days may need revising with respect to muscle thickness.[31]  Of the 314 newborns with hypertrophic pyloric stenosis studied, 60 (19%) were younger than 3 weeks, and 51 (85%) of these were diagnosed on the basis of US findings. Mean muscle thickness in this subset was 3.7 mm (vs 4.6 mm in those older than 3 wk), which is concerning because the normal cutoff is less than 4 mm. Given a possible 1-4% rate of negative exploration findings reported in the literature,[32]  the significance of these findings awaits further prospective research; however, this information could be useful in the assessment of borderline diagnoses.

Several studies supported the premise that surgeons[33, 34]  or even residents with appropriate resident-to-resident training may be sufficiently skilled to diagnose pyloric stenosis accurately with US. The advantages of this approach include a facilitated initial assessment and expedited management.

Microscopically, circular muscle hypertrophies are apparent, with increased connective tissue in the septa between the muscle bundles. An increase in chondroitin sulfate within the extracellular matrix may account for the cartilaginous quality of the pyloric tumor. Note that a histologic specimen is not obtained, nor is it necessary, for the diagnosis of pyloric stenosis.

Although medical treatment (see Medical Therapy) has been used to manage pyloric stenosis, pyloromyotomy has been firmly established as the treatment of choice for this condition. Adequate preoperative resuscitation is essential. Fluid resuscitation is guided by adequate urine output (1 mL/kg/hr) and by normalization of acid-base disturbances and electrolyte and bicarbonate levels.

Outcome studies comparing open and laparoscopic approaches to pyloromyotomy are becoming more numerous overall as the trend toward more minimally invasive procedures is becoming more important to the public.[26] In one survey, cosmetic outcome was valued highly enough that as many as 88% of parents were willing to pay additional expenses for their children to have smaller scars.[35]

With a higher public demand for minimally invasive surgery, it is important to keep abreast of trends in surgical resident training. Cosper et al reported that 93% of surgeons agreed that residents need to perform at least four open pyloromyotomies in order to become competent in the procedure; however, 44% reported that their residents performed fewer than four, a finding explained partly by the increased use of a laparoscopic approach and partly by decreased opportunities for residents in the operating room (OR) to acquire the necessary skills.[36]

In one study, the rate of complications associated with laparoscopic pyloromyotomy when a general surgery resident participates was 5.4-fold higher than that associated with performance by a pediatric resident, despite close attending supervision.[37] However, because more surgical residents are specializing, as well as because there may be fewer general surgeons who are comfortable performing pyloromyotomies, patients with pyloric stenosis may eventually experience a healthcare access problem. The need for parents to travel longer distances to a larger center that offers (or performs a high volume) of this procedure may result in further attempts at other modalities for treating pyloric stenosis.

For example, in an alternative approach used to address the pylorus externally, Zhang et al reported nine successful applications of endoscopic pyloromyotomy in resolving infantile hypertrophic pyloric stenosis.[38] After transabdominal ultrasonography (US) was used to assess the wall thickness, a 5.9-mm gastroscope was used to create a 2- to 3-mm incision (if the wall was 4-6 mm thick) or a 3- to 4-mm incision (if the wall was >6 mm thick). There were no complications of hemorrhage or perforation; however, one patient developed vomiting after 1 month, which resolved after a repeat procedure. With further research, this may prove to be a safer, more effective, and simpler alternative to the current standard of care.

For more information, see Pediatric Pyloric Stenosis, Pediatric Hypertrophic Pyloric Stenosis, and Imaging in Hypertrophic Pyloric Stenosis.

The primacy of pyloromyotomy in the management of pyloric stenosis notwithstanding, medical management of this condition remains important. Early assessment and treatment of fluid, electrolyte, and acid-base disturbances are paramount. Urgent resuscitation, rather than emergency surgical intervention, is the rule. Once the diagnosis is made, fluid resuscitation is begun. Clinical and biochemical assessments are made and repeated to guide appropriate fluid repletion.

Nonsurgical management was described originally in Europe, using a low-curd feeding of dextrose or breast milk; however, this treatment reportedly took months to complete and was associated with significant morbidity. Reports from Sweden showed that intravenous (IV) nutrition alone also usually failed. Additionally, a biochemical approach using atropine or scopolamine was followed to compensate for the lack of nitric oxide (NO) synthase (NOS) in the pylorus, which is thought to cause the hypertrophy. These anticholinergics were thought to decrease pyloric contractions; however, early success was lacking, and for some time, this approach was thought to be only of historical interest.

In a Japanese study from 1996, Nagita et al reported successfully treating 21 of 23 infants (91%) with pyloric stenosis by using IV atropine, administered at a dosage of 0.04-0.11 mg/kg/day until vomiting ceased, followed by oral atropine for 2 weeks.[39]

A subsequent Japanese study, by Kawahara et al in 2005, reported a success rate of 87% in 52 patients treated with IV and oral atropine.[40] Manometric studies were used to find the level of atropine that decreased tonic and phasic contractions in the pylorus. The dosing regimen of atropine was 0.01 mg/kg IV six times a day before feedings (median time, 1 wk), followed by 0.02 mg/kg orally after vomiting stopped and infants could tolerate formula in the amount of 150 mL/kg (median time, 44 d).

Hospital stays in this study ranged from 6 to 36 days (median, 13 d); however, despite the longer hospital stay, the costs for the medical group were similar to those for the surgical group, without the inherent risks of general anesthesia and surgery.[40] Regarding general outcomes, the two groups showed no difference in weight at age 1 year. Medical treatment for pyloric stenosis could prove useful to patients without sufficient access to surgical care or when surgery would be too risky. The authors concluded that this medical treatment of pyloric stenosis is an effective alternative to pyloromyotomy if the length of hospitalization and the necessity of continuing oral atropine are accepted.

Long-term studies have yet to be conducted to address recurrence rates, hospital costs in other countries, and caregiver compliance for atropine-treated patients with pyloric stenosis. In the United States, the Ramstedt pyloromyotomy remains the optimal treatment for pyloric stenosis.

Preparation for surgery

As noted (see Medical Therapy), pyloric stenosis is not a surgical emergency. The preoperative medical management of patients with pyloric stenosis is paramount for safe general anesthesia. Once the diagnosis is made, fluid resuscitation is begun to treat dehydration and electrolyte and acid-base disturbances. Clinical and biochemical assessments are made and repeated to guide appropriate fluid repletion.

IV therapy consists of 5% dextrose in one-half isotonic sodium chloride solution (0.45% NaCl/D5W) at 1.5 times the maintenance rate. Although children with severe dehydration should receive deficit fluid therapy with isotonic sodium chloride solution (20 mL/kg) initially, ongoing resuscitation should be performed with 0.45% NaCl/D5W to prevent rapid changes in volume and electrolyte levels, which can result in seizures. When urine output has been demonstrated, potassium chloride (10-20 mEq/L) can be added to the fluids.

As a general guideline, infants are deemed adequately resuscitated for the OR once the following criteria are met:

• Good urine output has been demonstrated
• Serum Cl ^– >100 mEq/dL
• Serum HCO ₃ ^– < 30 mEq/dL

In some patients with severe volume abnormalities (>20%) and electrolyte abnormalities (Cl–< 80 mEq/dL; HCO3– >35 mEq/dL; Na+< 120 mEq/L), resuscitation may take 48-72 hours. With serum bicarbonate levels exceeding 30 mEq/dL, the potential exists for myocardial dysfunction and respiratory depression. Patients should therefore be monitored for signs of apnea; if such signs are present, they may need to be intubated and mechanically ventilated until surgery is performed.

An alternative preoperative stabilization approach has been proposed for severely alkalemic patients to decrease the preoperative hospital stay. In a report of 16 infants with a pH exceeding 7.60, four patients received standard resuscitation for 4 days and then cimetidine at 10 mg/kg, and 12 received IV cimetidine on admission until the pH dropped below 7.50.[41] In all 16, pH reached the goal the same day cimetidine therapy was initiated, and all underwent pyloromyotomy that day. No complications associated with cimetidine were reported. This type of prompt preoperative resuscitation may help shorten hospital stays and reduce overall costs in infants with severe metabolic alkalosis.[42]

Operative details

Once the diagnosis of pyloric stenosis has been confirmed, adequate ongoing preoperative fluid resuscitation must be maintained by establishing adequate urine output (1 mL/kg/hr) and correcting acid-base disorders and electrolyte abnormalities. Regarding anesthetic induction for infants with pyloric stenosis, tracheal intubation with muscle paralysis seems to be superior to awake intubation, in that the former reduces the risk of desaturation and bradycardia due to multiple attempts at intubation.

Pyloromyotomy may be performed either as an open procedure, via a right-upper-quadrant (RUQ) horizontal incision or an umbilical incision (Tan-Bianchi operation), or as a laparoscopic procedure[2] (see the images below).

Open pyloromyotomy

In 1986, a Tan-Bianchi approach was described in which a pyloromyotomy was performed through a supraumbilical incision that afforded superior cosmesis.[43] In 2004, Blumer et al compared the umbilical approach with the RUQ approach in 237 patients and found that the umbilical approach took 3.1 minutes longer (28.5 vs 31.6 min); however, this difference was clinically irrelevant, in that there were no significant differences with respect to length of hospital stay, mucosal perforations, or wound infections.[44] The umbilical approach also was considered to provide a superior cosmetic outcome.

In 2004, Alberti et al reported modifying the Tan-Bianchi approach with a right semicircular umbilical incision, thus keeping all the incisions in the same axis, allowing for delivery of a larger pylorus, and decreasing the amount of retractor strain on the wound.[45] This approach resulted in a lower rate of hematoma formation and lower wound infection rates (0%) than supraumbilical incisions (16%), despite the use of prophylactic antibiotics with the semicircular umbilical approach.

In this modified Tan-Bianchi operation, after the pyloric channel is delivered from the abdomen, a seromuscular incision is made along the anterior border of the hypertrophied pylorus from 1-2 mm proximal to the duodenum to the distal antrum just proximal to the pylorus.[45] Great care must be taken not to incise or perforate the underlying mucosa. An alternative superficial V-shaped extension can be made at the duodenal end of the myotomy to reduce the risk of duodenal mucosal injury.

Other variations of the open abdominal approach include a technique reported by Yokomori et al in 2006, who used a semicircumumbilical incision to create a 1.5-cm sliding window with an 8-cm subcutaneous space.[46] This space was then slid 3-4 cm toward the RUQ. The abdomen was entered and the pyloromyotomy performed intracorporeally within the window. This type of incision can afford an uneventful postoperative course, along with a good cosmetic outcome.

One study compared an umbilical approach using a transverse muscle cutting incision and a vertical linea alba incision and found no difference in terms of postoperative morbidity.[47] In addition, a double-Y or Alalayet pyloromyotomy may be superior to the Ramstedt technique in decreasing postoperative vomiting and increasing weight gain in the first week following surgery.

Another consideration with pyloromyotomy may be the presence of foveolar cell hyperplasia (FCH),[48] a type of redundant mucosa, which may be evident on US in 12% of infantile hypertrophic pyloric stenosis cases. In patients diagnosed with FCH, an extended pyloromyotomy may decrease postoperative vomiting and reduce the need for gastric foveolar fold excision due to persistent postoperative vomiting.

Laparoscopic pyloromyotomy

A laparoscopic pyloromyotomy follows the same principles as an open procedure. First described by Alain et al in 1991,[49] the laparoscopic approach has been demonstrated as a safe alternative to exteriorizing the pylorus and improving cosmetic results.

The authors' approach entails creating a 5-mm camera port at the umbilicus. A 3-mm atraumatic locking grasping instrument is inserted in the RUQ through a transabdominal stab incision over the duodenum. The grasping instrument is used to stabilize the pyloric channel.

A 3-mm incision is made in the midepigastrium/left upper quadrant (LUQ), over the pyloric olive, through which a sheathed arthrotome is passed (without a trocar). Alternatively, a Bovie electrocautery with an extended tip can be used to create the pyloromyotomy. The pylorus is incised in the same fashion as with the open procedure. The hypertrophied muscle is then bluntly split with a laparoscopic Tanner pyloric spreader. (See the images below.)

Robotic-assisted pyloromyotomy

A robotic pyloromyotomy follows the same steps as the laparoscopic procedure. Whereas robotic pyloromyotomy has the advantages of not exteriorizing the pylorus and of improving cosmetic results, it also helps the surgeon with respect to dexterity, ergonomics, and precision. 

Partin in 1995 and Okada in 1998 described the use of robotic assistance for surgery in children in the form of a camera holder.[50]

Hollands in 2002 helped standardize robotic pediatric surgery by overcoming the technical challenges of using robotic surgical telemanipulators to facilitate operation in confined spaces.[51] She used the Zeus Robotic Surgical System (Computer Motion, Goleta, CA) on piglets to develop robotic approaches to enteroenterostomy, hepaticojejunostomy, and esophagoesophagostomy. Hollands in 2004 also presented the first data on robotic pyloromyotomy in human infants.[52]  

The robotic approach has been demonstrated to be a safe and noninferior alternative to laparoscopic surgery in pediatric patients. Since the initial studies, many other robotic platforms have been used in pediatric surgery to perform multiple complex procedures, including pyloromyotomy.[50]

Crystalloid resuscitation is continued postoperatively until the patient returns to full feeding. Data suggest that infants with pyloric stenosis have an increased incidence of postoperative apnea and bradycardia. These infants should be placed on an apnea and cardiac monitor for 24 hours following the operation.

A decrease in the number of hospital days after operation depends to a certain degree on how rapidly feeding is started and advanced. In a 2002 study, Michalsky et al reported that having a clinical pathway decreased surgeon variability by advancing the diet to oral feedings within 5 hours after operation.[53] Additionally, the length of hospital stay was significantly reduced (41.8 ± 9.7 vs 57.8 ± 11.7 hr), as were hospital costs ($4555 ± $464 vs $5400 ± $1017).

In a contrasting 2002 study, Puapong et al showed that feeding ad libitum after the patient is awake resulted in a significantly faster return to full-strength feeding than a controlled feeding regimen (29.1 vs 5.1 hr), along with a shorter hospital stay (38.8 ± 16.6 vs 25.1 ± 10.9 d) and a significant decrease in cost per patient ($3560 vs $2290).[54] Both studies found that early initiation of feeding led to better recovery and to cost savings.

The author has found the following seven-step feeding regimen to be safe and adequate:

1. Give the patient nothing by mouth (NPO) for 6 hours in the recovery room
2. Then, on demand, give 15 mL of Pedialyte every 2-3 hours for two feedings
3. If 15 mL is tolerated, advance to 30 mL of Pedialyte every 2-3 hours for two feedings
4. If 30 mL is tolerated, advance to 30 mL of full-strength formula every 2-3 hours for two feedings
5. If full-strength formula is tolerated, advance to spontaneous feedings
6. If vomiting occurs, repeat the step at which vomiting occurs and advance when tolerated
7. If the patient is being breastfed and the mother has bottled milk, follow the above schedule

If the patient is being breastfed and no bottled milk is available, follow the schedule below:

• Follow steps 1, 2, and 3 above
• Have the patient breastfeed on each side for 5 minutes every 2 hours for two feedings
• Have the patient breastfeed on each side for 10 minutes every 2 hours for two feedings
• Have the patient breastfeed spontaneously

Slow feeding and gentle burping help prevent wet burps postoperatively. Intermittent vomiting persisting through the first postoperative week is sometimes observed in patients with a protracted course of emesis and severe dehydration preoperatively. Vomiting lasting longer than 7 days postoperatively should alert the physician to the possibility of an incomplete pyloromyotomy. An upper gastrointestinal (UGI) study may be obtained but is useful only for demonstrating gastroesophageal reflux; the radiographic appearances of pyloric stenosis may persist for several months following complete pyloromyotomy.

Although pyloromyotomy is safe and curative and can be performed virtually without operative mortality (< 0.5%) and morbidity (< 10%), it is not without potential complications. Potential intraoperative and postoperative complications include the following:

• Bleeding
• Perforation 
• Wound infection

Duodenal or gastric perforation, the most serious complication, rarely occurs; however, if it goes unrecognized before wound closure, devastating or lethal consequences are possible. The infant with an enteric leak develops pain, distention, fever, and peritonitis. Ongoing fluid requirements, generalized sepsis, vascular collapse, and death follow if the enteric leak is not recognized and treated. Suspected perforation postoperatively requires immediate reexploration. Recognition of this complication at the time of surgery is important.

Mucosal perforation most commonly results from extending the myotomy beyond the pyloric-duodenal junction. If perforation occurs, the mucosal defect should be repaired and the myotomy completed. An omental patch may be sutured to the perforation site, and a paraduodenal drain may be considered. If any question exists about the success of the closure, a UGI study can be obtained before feedings are initiated. The patient should continue to receive antibiotics until feedings are resumed.

Bleeding is a rare complication of pyloromyotomy. Other complications that are more common but less serious include the following:

• Superficial wound infections (usually Staphylococcus aureus)
• Postoperative vomiting

Patients with wound erythema, drainage, or both undergo wound opening and debridement and antibiotic therapy. Incomplete myotomy results in ongoing gastric outlet obstruction and requires reoperation. However, ongoing emesis after pyloromyotomy does not mean an incomplete myotomy was performed. Patients with prolonged preoperative obstruction develop gastric distention and dysmotility, which may cause postoperative emesis for up to 1 week after an adequate pyloromyotomy. Oral atropine has been suggested as a viable treatment for persistent emesis after pyloromyotomy.[55]

The follow-up care regimen involves a routine postoperative visit at 1 week to check wounds and to ensure that the patient is once again gaining weight.