INCIDENCE AND ETIOLOGY
Infantile hypertrophic pyloric stenosis is a common cause of nonbilious emesis in the newborn. It is important, but sometimes difficult, to distinguish pyloric stenosis from other common causes of nonbilious emesis (eg, formula intolerance and gastroesophageal reflux) in the newborn period. The incidence of pyloric stenosis is approximately 1 to 4 cases per 1,000 live births, with a 2:1 to 5:1 male to female ratio.1 Whites have the highest incidence (2 to 4 cases per 1,000 live births) compared with Hispanics (1.8 per 1,000 live births) and African Americans (0.7 per 1,000 live births). The diagnosis is made at an older age for premature infants than for term infants.2 Current data suggest that the incidence is increasing in the United States and abroad.3
The proposed etiology of pyloric stenosis has evolved over time. According to a theory in the 1960s, it was related to diet.4 Other hypotheses focused on the pyloric muscle and its innervation. More recent theories propose a molecular cause. Studies on pylorìc innervation have suggested a mechanism similar to that of Hirschsprung's disease (ie, an absence or immature function of pyloric ganglion cells). Immaturity of the ganglion cells that are present has been proposed as an etiology of this disease.5 The number of ganglion cells in the hypertrophic pylorus has been observed to be normal or decreased by different investigators.6 Gastrin levels have also been elevated in infants with pyloric stenosis,7 but it is unclear whether this is a primary cause or a secondary effect of obstruction of the gastric outlet. Substance P may also have a role because it is a neurotransmitter. It is known to be elevated in patients with pyloric stenosis and can produce muscle hypertrophy due to its positive effect on enteric muscle contraction.8
Molecular biology has recently been used to determine additional factors that may induce pyloric stenosis. One clue to a potential molecular cause is the poor innervation of the pyloric musculature observed in patients with pyloric stenosis. Innervation is absent in young infants with pyloric stenosis. It increases but remains below normal levels 3 months after repair, and returns to normal levels by 7 months postoperatively.9 Nitric oxide is a known smooth muscle relaxant that is formed from arginine by the enzyme nitric oxide synthase. There is a marked decrease in nitric oxide synthase mRNA in the pyloric muscle of infants with pyloric stenosis compared with control subjects.10 Increased levels and expression of transforming growth factor-alpha mRNA have also been found in infants with pyloric stenosis,11 as have increased levels of epidermal growth factor mRNA.12
The typical presentation of pyloric stenosis is that of a male infant with a normal feeding history and new onset of nonbilious, "projectile" emesis. The age at presentation varies from 2 to 8 weeks, with a peak incidence at 3 to 5 weeks. The emesis may initially be attributed to self-limited gastroesophageal reflux or formula intolerance, but becomes progressively worse over a few days until nearly every feeding is forcefully vomited.
The infant is hungry immediately after emesis and generally does not appear ill early in the course of the disease. If the diagnosis is not recognized early, progressive dehydration and lethargy occur. A few infants will present with jaundice due to an indirect hyperbilirubinemia that is probably secondary to a lack of UDP-glucuronyltransferase enzyme activity.13 Infants with a history of prematurity usually present 2 weeks sooner than term infants. Their emesis is less projectile and evolves more slowly, delaying diagnosis even further.14
Gastroesophageal reflux and formula intolerance may be difficult to distinguish from pyloric stenosis. The onset of emesis in these medical conditions is generally more insidious and the emesis is less often truly projectile. Other nonsurgical causes, such as sepsis, hydrocephalus, or metabolic disorders, must also be considered in the differential diagnosis for infants with nonbilious emesis. Surgical causes such as antral web, gastric duplication, or pyloric atresia may have similar presentations but are much less likely than pyloric stenosis. The importance of determining that the emesis is not bilious cannot be overstated. Any suggestion that the emesis is yellow or green warrants immediate investigation with an upper gastrointestinal tract series to rule out the possibility of malrotation and an associated volvulus - a true surgical emergency.
When the diagnosis of pyloric stenosis is delayed, the infant can present with nonbilious projectile vomiting, visible peristaltic waves in the left upper quadrant, a palpable "olive" in the right upper quadrant, and an associated hypochloremic or hypokalemic metabolic alkalosis. When the diagnosis is suspected earlier, establishing it is more dependent on imaging studies because the clinical hallmarks of the disease, including the findings on physical examination, are subtler.15 Nevertheless, experienced examiners can diagnose pyloric stenosis in more than half of affected infants on the basis of physical examination and clinical presentation alone.16 In the current era of managed care, pediatricians commonly obtain an imaging study before referring to a surgeon who can make the diagnosis using history and examination alone.
Although it has been suggested that the diagnosis of pyloric stenosis by physical examination is a lost art that is performed only by more senior surgeons, recent literature suggests that it is not lost but just "misplaced" in the sequence of diagnostic maneuvers.16 Examination by a surgeon before obtaining an imaging study can be the most cost-effective method for diagnosing pyloric stenosis.17 The specificity of the physical examination by a surgeon may be as high as 90%, but its sensitivity has been less than 50%.17 In institutions where the surgeon examines the infant before imaging studies are obtained, the infant is prepared for surgery and a pyloromyotomy is performed without further diagnostic studies when a hypertrophic pylorus is felt. This strategy reduces the number of required radiologic studies by almost 50%. If the results of the physical examination are negative, the infant then undergoes either an upper gastrointestinal tract contrast study or ultrasound.17
Several maneuvers can increase the sensitivity of physical examination for pyloric stenosis. The physician must allow 5 to 10 minutes for examining the infant. Making the infant warm and relaxed and giving a pacifier that has been dipped in sugar water facilitate the examination. If adequate relaxation is not possible with these maneuvers, gastric decompression by insertion of a nasogastric tube may be helpful. Placing a nasogastric tube to low suction and allowing the infant to suck simple sugar water or balanced electrolyte solution may relax the abdominal wall musculature sufficiently to allow palpation of the pylorus. With his or her examining hand, the physician should identify the liver edge in the upper epigastrium and left upper quadrant and then feel the pylorus by rolling it under his or her fingers while pressing toward the vertebral column and drawing his or her hand caudally. If the physician is comfortable making the diagnosis by examination alone, no additional diagnostic tests are needed.
In the past, an upper gastrointestinal tract series was the gold standard when further imaging was necessary for diagnosing pyloric stenosis. This test is sensitive for pyloric stenosis and can also help diagnose other causes of nonbilious emesis such as gastroesophageal reflux. One potential disadvantage of an upper gastrointestinal tract series is that it fills the obstructed stomach with barium before induction of general anesthesia. This increases the risk of aspiration. In addition, the infant is exposed to the risks of ionizing radiation.
Ultrasound neither exposes the infant to ionizing radiation nor fills the stomach with barium. Ultrasound criteria that suggest pyloric stenosis include a muscle thickness of greater than 4 mm and a channel length of greater than 16 mm.18 Potential disadvantages of ultrasound are that it is operator dependent and not useful for making other definitive diagnoses if pyloric stenosis is not confirmed. An upper gastrointestinal tract series is more cost-effective than ultrasound if all infants presenting with projectile vomiting are being evaluated.19
One way to determine the optimal imaging test for pyloric stenosis is to measure the volume of the gastric aspirate obtained at the time of nasogastric tube placement. If an aspirate of less than 10 mL is obtained, an upper gastrointestinal tract series should be done, because 86% of infants have been observed to have gastroesophageal reflux. If more than 10 mL is obtained, ultrasound should be ordered, because 92% of infants will have pyloric stenosis.20 At our institution, if an olive is not palpable by the surgeon, a flat plate of the abdomen is obtained. If a large stomach bubble is noted, ultrasound is performed instead of an upper gastrointestinal tract series.
Acid-base status and serum electrolyte measurements can also be helpful in distinguishing pyloric stenosis from gastroesophageal reflux. As pyloric stenosis progresses, the loss of hydrogen chloride from the stomach sets the stage for a hypochloremic, hypokalemic, and often hyponatremic metabolic alkalosis. Initially, the loss of hydrogen ions results in renal wasting of potassium. As the alkalosis worsens and hypokalemia develops, the kidneys waste more hydrogen ions to spare potassium. This paradoxic aciduria indicates long-standing pyloric stenosis and may take days to slowly correct before surgery. An understanding of this electrolyte imbalance helps distinguish pyloric stenosis from reflux and other causes of emesis. Serum pH, excess base, and chloride levels have all been independent variables predictive of pyloric stenosis, with chloride level being the most predictive of acid-base status.21
Before surgery is performed, the dehydrated infant with pyloric stenosis should be adequately hydrated and electrolyte disorders corrected. Findings on physical examination such as a sunken fontanelle, poor skin turgor, decreased output of urine with increased specific gravity, and delayed capillary refill can be used to gauge the severity of dehydration. The severity of hypochloremia and metabolic alkalosis can help determine how much resuscitation is required. Currently, severely dehydrated infants are rare. Most have electrolytes that are near normal levels and can safely undergo surgery without extensive resuscitation. In our institution, the infant may receive anesthesia as soon as normal output of urine has been established, the serum HCO3" is 30 mEq/L or less, and the chloride is 90 to 95 mEq/L or greater.
This metabolic alkalosis is chloride sensitive and will not correct easily unless excess chloride is provided. Even if the infant does not have hyponatremia, excess sodium chloride should be provided to allow for correction of the metabolic alkalosis. The focus of electrolyte replacement should be on replacing the chloride, not the sodium. The use of maintenance fluids that contain one-quarter normal saline will correct neither the alkalosis nor the dehydration.
Our current regimen for resuscitation is to administer a bolus of 20 cc/kg of normal saline at the time the peripheral intravenous line is established and initial diagnostic tests are sent to the laboratory. After receiving the bolus, the infant is given 5% dextrose and half normal saline at a rate equal to 1.5 times maintenance. If the infant is hypokalemic and the output of urine has been established, 20 mEq/L of potassium chloride is added. For most infants, mild alkalosis will correct overnight with this regimen. Those with normal electrolytes are treated after the output of urine has been established. Severe alkalosis (CO2 greater than 35 mEq/L) and hyponatremia in infants should be corrected slowly, so that the rate of change is no greater than 10 mEq/L during a 24-hour period. This is done in an attempt to decrease the risk for pontine myelinolysis. The infant will occasionally require admission to the intensive care unit for monitoring while the electrolyte imbalance is corrected. It is our practice to place a nasogastric tube only for diagnostic purposes or when the infant continues to have emesis despite receiving nothing per mouth.
Proper anesthetic management of the infant with pyloric stenosis is essential. Anesthesia for a 1-month-old infant with obstruction of the gastric outlet can be challenging and is ideally undertaken only by anesthesiologists experienced with infants. In the operating room, the stomach is emptied with a gastric tube before induction. A rapid sequence or modified rapid sequence induction is performed to further lower the chance of aspiration. A combination of appropriate preoperative, intraoperative, and postoperative management can effectively reduce the morbidity of this patient group.
Most general surgeons have had exposure to performing the pyloromyotomy, and the results of pyloromyotomies performed in community hospitals have compared favorably with results from specialty centers.22 However, recent data from our institution suggest that the complication rate, length of hospitalization, and overall cost are decreased when infants with pyloric stenosis are treated by pediatric surgeons at specialty centers, even in uncomplicated cases. Despite these advantages, fewer than half of infants with pyloric stenosis in our area are cared for by a pediatric surgeon in a specialty center.23
The Ramstedt pyloromyotomy is the procedure that has been traditionally chosen for repair. Through a right upper quadrant incision, the pylorus is grasped, the serosa incised longitudinally, and the thickened pyloric muscle spread until mucosa is bulging between the separated halves of the pylorus. The muscle should be separated carefully on its most distal end because the mucosa of the duodenum can fold and is a common site for perforation. If the myotomy is not extended proximally enough onto the antrum of the stomach, most procedures will fail to effectively relieve the obstruction. Before closure, the integrity of the mucosa is evaluated by compressing the stomach.
Many pediatric surgeons now perform the pyloromyotomy using a laparoscopic technique. Although there is an initial learning curve, the operation can be performed safely and with a complication rate similar to that for the open procedure.24 Potential advantages of the laparoscopic approach are that it requires smaller incisions and is more cosmetic and that the time to return to full feeding may be reduced. Potential disadvantages include an increased operative time, respiratory acidosis after inducing CO2 pneumoperitoneum for visualization, and risk of hypothermia secondary to CO2 insufflation at room temperature. Most surgeons see the laparoscopic technique as an equivalent procedure if performed by those experienced with it.
Another method for approaching the pylorus is through an umbilical fold incision made similar to an umbilical hernia incision.25 This method combines the improved cosmesis achieved with laparoscopy with the comfort that surgeons have gained through performing the open technique. The superior umbilical fold is incised, a skin flap elevated superiorly, and the abdomen entered through the midline or a transverse incision. Although this procedure can be performed in most patients, it may be difficult when an especially large pylorus is encountered. After the incision has healed closed, it is remarkably cosmetic and often undetectable even at the first postoperative visit. Although it has been suggested that this approach is associated with an increased wound infection rate, prophylactic preoperative antibiotics may reduce this to a rate similar to mat observed using the traditional approach.
Although nonoperative approaches have been attempted, these remain controversial. It is known that pyloric stenosis will spontaneously resolve within weeks to months. It is possible that an infant could be treated with parenteral nutrition during this time if there was an absolute contraindication to surgery. Atropine has recently been reported to help speed the resolution of pyloric stenosis. Oral atropine was given once a day. Infants were hospitalized for 2 to 3 days, then discharged when they were able to take enough fluids. They took longer to return to full feedings, but overall costs were decreased when compared with surgery.26 Because of the success of surgical treatment, this method should be viewed with caution and generally accepted only if shown to be effective in carefully constructed trials.
Most infants can be fed within 6 hours of surgery. The use of a formal postoperative feeding schedule may decrease the length of hospitalization.27 Postoperative emesis is usually self-limited and not related to the feeding schedule. Feeding may be started as early as 6 hours postoperatively and advanced every 2 hours regardless of whether the infant has emesis.28 Although infants who are fed earlier may have more vomiting, a more rapid feeding schedule will lead to the earlier attainment of full feedings and discharge. Some have advocated an ad lib feeding schedule after finding that there is a similar amount of postoperative emesis and that discharge occurs earlier.29 Our current regimen for postoperative feeding is described in the table. We initiate feedings 6 hours postoperatively and then advance them every 2 to 3 hours.
Operative morbidity and mortality rates should be low in patients with pyloric stenosis. With appropriate management of fluids and electrolytes in the preoperative period, safe anesthesia, and intraoperative management by a surgeon with experience performing pyloromyotomy, the infant with mild to moderate electrolyte abnormalities should do well and be discharged within 1 to 2 days of presentation. Long-term effects of the disease are minimal. No differences in gastric emptying were detectable in patients observed to 25 years of age compared with age-matched control subjects.30 Some residual pyloric defect may exist. Manometric evaluations of adult volunteers with and without a history of pyloric stenosis demonstrated no difference in emptying but increased pyloric tone and force of gastric contraction in those with a history of pyloric stenosis.31 This difference is usually not appreciated clinically. Parents can thus be assured that there will be no long-term sequelae of the disease.
Regimen for Postoperative Feeding*
Infants with postoperative emesis that persists after a few days but who are gaining weight can continue to be fed. This phase is usually temporary and vomiting will often improve with the use of prokinetic agents. If projectile vomiting and failure to thrive are observed, the possibility of an inadequate pyloromyotomy should be considered. If present, reexploration and completion of the myotomy should be performed. It is rare that an infant will have a complete response initially followed by recurrent projectile vomiting due to recurrent pyloric stenosis. The surgeon should evaluate any infant with persistent emesis and failure to thrive for residual or recurrent disease.
1. Mitchell LE, Risch N. The genetics of infantile hypertrophic pyloric stenosis: a reanalysis. American Journal of Diseases in Children. 1993;147:1203-1211.
2. Schechter R, Torfs CP, Bateson TF. The epidemiology of infantile hypertrophic pyloric stenosis. Paediatr Perinat Epidemiol. 1997;11:407-427.
3. Tarn PK, Chan J. Increasing incidence of hypertrophic pyloric stenosis. Arch Dis Child. 1991;66:530-531.
4. Lynn H. The mechanism of pyloric stenosis and its relationship to preoperative preparation. Arch Surg. 1960;81:453-459.
5. Friesen S, Boley J, Miller D. The myenteric plexus of the pylorus: its early normal development and its changes in hypertrophic pyloric stenosis. Surgery. 1956;39:21-29.
6. Jona J. Electron microscopic observation in infantile hypertrophic pyloric stenosis (IHPS). / Pediatr Surg. 1978; 13:17-20.
7. Spitz L, Zail S. Serum gastrin levels in congenital hypertrophic pyloric stenosis. J Pediatr Surg. 1976;11:33-35.
8. Tam PK. Observation and perspectives of the pathology and possible aetiology of infantile hypertrophic pyloric stenosis: a histological, biochemical, histochemical, and immunocytochemical study. Ann Acad Med Singapore. 1985;14:523-529.
9. Kobayashi H, Wester T, Puri P. Age-related changes in innervation in hypertrophic pyloric stenosis. J Pediatr Surg. 1997;32:1704-1707.
10. Kusafuka T, Puri P. Altered messenger RNA expression of the neuronal nitric oxide synthase gene in infantile hypertrophic pyloric stenosis. Pediatr Surg Int. 1997;12:576-579.
11. Shima H, Puri P. Increased expression of transforming growth factor-alpha in infantile hypertrophic pyloric stenosis. Pediatr Surg Int. 1999;15:198-200.
12. Shima H, Ohshiro K, Puri P. Increased local synthesis of epidermal growth factors in infantile hypertrophic pyloric stenosis. Pediatr Res. 2000;47:201-207.
13. Roth B, State A, Heinisch HM, Gladtke E. Elimination of indocyanine green by the liver of infants with hypertrophic pyloric stenosis and the icteropyloric syndrome. J Pediatr. 1981;99:240-243.
14. Tack ED, Perlman JM, Bower RJ, McAlister WH. Pyloric stenosis in the sick premature infant: clinical and radiologic findings. American Journal of Diseases in Children. 1988;142:68-70.
15. Hulka F, Campbell TJ, Campbell JR, Harrison MW. Evolution in the recognition of infantile hypertrophic pyloric stenosis. Pediatrics. 1997;100:e9. Available at www.pediatrics.org/cgi/content/full/100/2/e9.
16. Chen EA, Luks FI, Gilchrist BF, Wesselhoeft CW Jr, DeLuca FG. Pyloric stenosis in the age of ultrasonography: fading skills, better patients? J Pediatr Surg. 19%; 31:829-830.
17. White MC, Langer JC, Don S, DeBaun MR. Sensitivity and cost minimization analysis of radiology versus olive palpation for the diagnosis of hypertrophic pyloric stenosis. J Pediatr Surg. 1998;33:913-917.
18. Keller H, Waldermann D, Greiner P. Comparisons or preoperative sonography with intraoperative findings in congenital hypertrophic pyloric stenosis. J Pediatr Surg. 1987;22:950-952.
19. Hulka F, Campbell JR, Harrison MW, Campbell TJ. Costeffectiveness in diagnosing infantile hypertrophic pyloric stenosis. J Pediatr Surg. 1997;32:1604-1608.
20. Mandelt GA, Wolfson PJ, Adkins ES, et al. Cost-effective imaging approach to the nonbilious vomiting infant. Pediatrics. 1999;103:1198-1202.
21. Shanbhogue LK, Sikdar T, Jackson M, Lloyd DA. Serum electrolytes and capillary blood gases in the management of hypertrophic pyloric stenosis. Br J Surg. 1992;79:251253.
22. Jahangiri M, Osborne MJ, Jayatunga AP, Bradley JW, Mitchenere P. Infantile hypertrophic pyloric stenosis: where should it be treated? Ann R CoI! Surg Engl. 1993;75: 34-36.
23. Prarukoff T, Campbell B, Travis J, Hirschl RB. Differences in outcome with subspecialty care: pyloromyotomy in North Carolina. J Pediatr Surg. In press.
24. Bufo AJ, Merry C, Shah R, Cyr N, Schropp KP, Lobe TE. Laparoscopic pyloromyotomy: a safer technique. Pediatr Surg Int. 1998;13:240-242.
25. Fitzgerald PG, Lau GY, Langer JC, Cameron GS. Umbilical fold incision for pyloromyotomy. J Pediatr Surg. 1990;25:1117-1118.
26. Yamataka A, Tsukada K, Yokoyama-Laws Y, et al. Pyloromyotomy versus atropine sulfate for infantile hypertrophic pyloric stenosis. J Pediatr Surg. 2000;352:338341.
27. Leinwand MJ, Shaul DB, Anderson KD. A standardized feeding regimen for hypertrophic pyloric stenosis decreases length of hospitalization and hospital costs. J Pediatr Surg. 2000;35:1063-1065.
28. Georgeson KE, Corbin TJ, Griffen JW, Breaux CW Jr. An analysis of feeding regimens after pyloromyotomy for hypertrophic pyloric stenosis. J Pediatr Surg. 1993;28:14781480.
29. Carpenter RO, Schaffer RL, Maeso CE, et al. Postoperative ad Hb feeding for hypertrophic pyloric stenosis. J Pediatr Surg. 1999;34:959-961.
30. Ludtke FE, Bertus M, Voth E, Michalski S, Lepsien G. Gastric emptying 16 to 26 years after treatment of infantile hypertrophic pyloric stenosis. J Pediatr Surg. 1994; 29:523-526.
31. Sun WM, Doran SM, Jones KL, Davidson G, Dent J, Horowitz M. Long-term effects of pyloromyotomy on pyloric motility and gastric emptying in humans. Am J Gastroenterol. 2000;95:92-100.
Regimen for Postoperative Feeding*