Br. J. Anaesth. (1987), 59, 672-677

ANAESTHESIA FOR CONGENITAL HYPERTROPHICPYLORIC STENOSISA Review of 350 Patients

N. J. M A C D O N A L D , G. J. FITZPATRICK, K. P. MOORE, W. S. WRENAND M. KEENAN

Congenital hypertrophic pyloric stenosis, with anincidence of 3 per 1000 live births (Editorial,1984), is the second commonest reason for surgicalintervention in the first 6 months of life.Ramstedt's pyloromyotomy is the treatment ofchoice.

Bennet (1968) and Touloukain and Higgins(1983) have described the complex electrolyte andmetabolic disturbances which occur in thiscondition, and which stem from the loss ofhydrogen, chloride, sodium, potassium and waterin the vomitus. Initially, the kidney compensatesby excreting sodium and potassium bicarbonate,but when hypokalaemia and hyponatraemia super-vene an acid urine is produced, aggravating thehypochloraemic hypokalaemic alkalosis. A com-pensatory respiratory acidosis is usually present.Lactic acidosis secondary to hypovolaemic shockoccurs in severe cases. These starving infants areprone to hypoglycaemia.

Although the management of this condition hasimproved considerably, some published seriesshow a small but significant mortality whichranges from 0.32% to 2% (Bell, 1968; Daly andConn, 1969; Kiely et al., 1981; Gray, Gear andStevens, 1984). We report the experience of theanaesthetic management of congenital hyper-trophic pyloric stenosis in 350 consecutivepatients over an 8-year period in one hospital.

NEIL J. MACDONALD,* M.B., D.C.H., F.F.A.R.CS.I.; GERARDJ. FITZPATRICJC, M.B., F.F.A.R.CS.I.; KEVIN P. MOORE,M.B., F.F.A.R.CS.I., F.F.A.R.C.S.; WILLIAM S. WREN, M.B.,F.F.A.R.CS.I., F.F.A.R.CS.J MARGUERITE KEENAN, M.B.,F.F.A.R.CS.I.; Department of Anaesthetics, Our Lady'sHospital for Sick Children, Dublin 12, Ireland. Accepted forPublication: September 9, 1986.

'Present address: Department of Anaesthesia, RegionalHospital, Newcastle Road, Galway, Ireland.

Correspondence to G. F.

SUMMARYThe anaesthetic management of 350 consecutivepatients with congenital hypertrophic pyloricstenosis over an 8-year period is reviewed. Theanaesthetic technique is discussed and thecomplications encountered reviewed. The mor-bidity rate was 15.9%. The anaesthetic relatedmorbidity rate was 3.7%. One patient in the seriesdied (0.27%) 8 weeks after pyloromyotomy, as aresult of an underlying myopathy.

PATIENTS AND METHODS

The names of all patients who underwentpyloromyotomy between January 1, 1977 andDecember 31,1984 were taken from the operatingtheatre record books. All of the case notes werelocated and included in the study.

Fourteen patients (4%) had been admitted tohospital for other reasons when the symptoms ofpyloric stenosis developed. The diseases presentin this group of patients are listed in table I. Forthe purposes of the study they were designatedgroup A. The statistics for the variables "timefrom admission to surgery" and "days spent inhospital after operation" were calculated both forall patients (350) and with group A excluded(336).

ManagementAfter admission to hospital the infants were

resuscitated with appropriate i.v. fluids andelectrolyte solutions. A nasogastric tube (size 8French gauge) was passed and the stomach keptempty by frequent aspiration. Estimations ofblood glucose concentration were made 2-4 hourlyusing reagent strips (Ames Dextrostix).

Dow

nloaded from https://academ

ic.oup.com/bja/article/59/6/672/245772 by guest on 19 July 2022

CONGENITAL HYPERTROPHIC PYLORIC STENOSIS: REVIEW 673

TABLE I. Conditions present in patients who were in hospital whenihe symptoms of pyloric stenosis first occurred

ConditionNo.

patients

Congenital cardiac anomalyTracheo-oesophageal fistulaCystic fibrosisCongenital myopathySeborrhoeic dermatitisViral meningitis

Total

821111

14

Initial i.v. therapy was always with an elec-trolyte/dextrose solution, usually 2.5 % dextrosein saline 0.45 g at 150 ml kg"1 day"1. Additionali.v. fluid with physiological saline (sodiumchloride 0.9 g%) was given to replace nasogastriclosses. Potassium chloride 2 mmol dl"1 was addedto the i.v. fluids once urine flow had beenestablished. Occasionally, an infusion of plasma(10-15 ml kg"1) was required for patients pre-senting in hypovolaemic shock.

Frequent assessments, both clinical and bio-chemical, were mandatory and resuscitative effortsadjusted accordingly. The infant was deemed fitfor surgery when fluid and electrolyte balance wasrestored to normal. Premedication was withatropine 0.02 mg kg"1 i.m., but was omitted ifbody temperature exceeded 37.5 °C.

The patient was transferred to the operatingtheatre in a neutral thermal environment. Heatloss was minimized during surgery by keeping theenvironmental air temperature at 23-25 °C, andby using a thermostatically-controlled heatedwater mattress.

Before the induction of anaesthesia the naso-gastric tube was aspirated (and left in place).Anaesthesia was induced either by inhalationor with a rapid sequence i.v. technique using a"sleep dose" of thiopentone. Suxamethonium1-2 mg kg"1 was administered, the trachea in-tubated and anaesthesia maintained with 66%nitrous oxide in oxygen supplemented with avolatile anaesthetic agent. Muscular paralysis wasusually achieved with a competitive neuro-muscular blocking agent or with intermittentinjections of suxamethonium. Ventilation wascontrolled mechanically using a suitable ven-tilator, or manually via the Jackson-Rees modifi-cation of Ayres' T-piece (both techniques areused in this teaching hospital). Arterial pressure,the electrocardiogram and rectal temperature were

monitored routinely. A precordial stethescopewas used frequently.

At the end of surgery neuromuscular blockadewas reversed and the trachea extubated when theinfant was fully awake and moving all four limbsvigorously. The patient remained under closeobservation in the recovery unit for at least 1 h,and during this period blood glucose concen-tration was estimated using a "Dextrostix". Thepatient returned to an infant ward staffed bynurses experienced in the postoperative care ofinfants.

I.v. fluids were continued with 2.5% Dextrosein saline 0.45 mg% at 4 ml kg"1 h"1 until oralfeeding was re-established.

The current practice is to leave the nasogastrictube in situ for 12 h. If, at 24 h, there is novomiting, a trial feed of dextrose 100 ml,electrolyte and water is given. If this is toleratedthen feeds are recommenced, starting withone-quarter strength and gradually building up tofull strength (Leahy and Fitzgerald, 1982).

RESULTS

Of the 350 patients in the series, 294 (84%) weremale and 56 (16 %) were female. The mean values,standard deviations and ranges for age (weeks)and weight (kg) are given in table II. Threehundred and fifteen (90 %) patients were full term(38-42 weeks gestation) at birth and 11 (3.1%)were born at less than 35 weeks gestation.

Pyloric stenosis developed in eight patients whowere admitted to hospital because of congenitalheart disease (table I). (This hospital is the

TABLE II. Mean, SD and range for age, weight, time fromadmission to surgery, duration of anaesthesia, time on i.v. fluids

after operation and days spent in hospital after operation

Variable

Age (weeks)Weight (kg)Time from admission to

surgery (h)All patientsGroup A excluded

Duration of anaesthesia(51 % of patients) (min)

Time on i.v. fluids post-op.(h)

Time in hospital post-op.(days)

All patientsGroup A excluded

Mean

6.23.76

51.34644.3

48.2

6.45.9

SD

3.40.76

55.936.213.6

25.5

6.44.8

Range

1-301.8-8.4

2-2002-336

15-100

6-240

2-662-̂ 13

Dow

nloaded from https://academ

ic.oup.com/bja/article/59/6/672/245772 by guest on 19 July 2022

674 BRITISH JOURNAL OF ANAESTHESIA

National Centre for paediatric heart disease, with500-600 referrals per year.)

The mean interval between admission andsurgery was 51.3 h (table II). One hundred andforty-five (41.4%) infants were not dehydrated onadmission. However, of these, 86 (59.3%) hadbeen actively treated in other hospitals before theirtransfer (table III). One hundred and ninety-three(55.1%) patients had chloride concentrationsbelow normal (table III). Seventy-one (20.3%)had serum potassium values below normal (tableIII). Atropine was given i.m. in 287 (82 %) infantsas a premedicant, and omitted in 63 (18%), 26 ofwhom had a rectal temperature of 37-38 °C. Theother 37 were brought to theatre unpremedicatedand received atropine i.v. at induction.

Anaesthesia was induced i.v. in 210 (60%)patients and by inhalation in 140 (40%). Therelative proportions of i.v. and inhalation in-ductions in each year are displayed in figure 1. In1977, anaesthesia was induced i.v. in six (14.6%)patients; in 1984, 35 (85.4%) received an i.v.induction. One hundred and ninety-nine (57%)patients were ventilated automatically and 151(43%) were manually ventilated using theJackson-Rees modification of Ayre's T-piece.

The duration of anaesthesia was recorded in179 (51 %) patients and was not available in 171(49%). The mean value was 44.3 min (SD 13.4,range 15-100). The mean time for which i.v. fluidswere required after operation was 48.2 h (SD25.5, range 6-240) (table II).

TABLE III. Degree of dehydration, and tenon chloride andpotassium concentrations, on admission. (Of those not dehydrated,59.5% had been actively treated in another hospital before

transfer here)

Degree of dehydrationNoneMildModerateSevere

Serum chloride (mmol litre"1)> 9 590-9580-90< 8 0

Serum potassium (mmol litre"1)>3.53.0-3.5<3.0

No. patients

145 (41.4%)154(44%)40(11.4%)11(3.2%)

157(44.9%)88(25.1%)68(19.5%)32(10.5%)

276(78.7%)42(12%)32 (9.3%)

Morbidity and mortalityOne patient (0.27%) known to have a rare

progressive muscle disorder died 8 weeks afterpyloromyotomy. This patient was the only infantin the series to have his operation performed underlocal infiltration anaesthesia (using 3 ml of 0.25 %lignocaine).

The morbidity rate was 15.9%. This wasarbitrarily divided into surgical and anaestheticmorbidity. The surgical morbidity rate was12.2%; this included wound infections, wounddehiscence, persistent vomiting, delayed feedingand urinary tract infections. The anaesthetic

100'

75-

50-

25-

85.4/!

14§Z179%

78.8/!

8757,

bj2

t 1.9".

MX

1980 1981 1982 1983 1984

FIG. 1. Percentage of inhalation (black columns) and i.v. inductions in each year from 1977 to 1984.

Dow

nloaded from https://academ

ic.oup.com/bja/article/59/6/672/245772 by guest on 19 July 2022

CONGENITAL HYPERTROPHIC PYLORIC STENOSIS: REVIEW 675

TABLE IV. Causes of morbidity related to anaesthesia

Cause

Upper respiratorytract infection

Lower respiratorytract infection

No.patients

3

5

Comment

Consisted of mild pyrexia, nasalsnuffles, and/or pharyngitis. Therewas no clinical or radiologicalevidence of lower respiratory tractinfection.

Two patients had a preoperative upperresDiratorv tract infection. One

Delayed recovery fromanaesthesia

Postoperative cyanoticattack

Postoperative stridor

patient had had respiratory distresssyndrome requiring controlled ventila-tion as a neonate. One patient hadcongenital heart disease and cardiacfailure before operation.

See text

See text

Persisted for 6 h after operation.Treated with humidified oxygen.

related morbidity rate was 3.7%. This includedall patients in whom morbidity was definitely orpossibly related to anaesthesia. These are listed intable IV. All of these patients made a full recovery.

DISCUSSION

No patient died as a result of congenitalhypertrophic pyloric stenosis or pyloromyotomyin this hospital during an 8-year period in which350 pyloromyotomies were performed.

This compares favourably with several reportsin which a small but definite mortality wasrecorded (Bell, 1968; Kiely et al., 1981; Gray, Gearand Stevens, 1984). One patient, suffering frommyotubular myopathy and requiring mechanicalventilation from birth, died 8 weeks afterpyloromyotomy. The operation was performedunder local infiltration analgesia and both surgeryand anaesthesia were uneventful.

The morbidity rate was 15.9%. All patientsmade a full recovery. Gray, Gear and Stevens(1984) report a complication rate of 20%;Dubousset (1973) had two cases of gram-negativesepticaemia from which both patients subse-quently died, in a series of 109 patients.

Four infants developed postoperative com-plications which warrant further discussion.

Patient 1. A 49-day-old full term male infantweighing 3.6 kg presented with pyloric stenosis.

He was mildly dehydrated on admission andbiochemical analysis revealed a hypochloraemicalkalosis. Twenty-four hours later, followingcorrection of hydration with i.v. fluids, he wasdeemed fit for surgery. Atropine 0.06 mg i.m. wasgiven as a premedicant. Before induction ofanaesthesia the nasogastric tube was aspirated andanaesthesia induced with halothane and nitrousoxide in oxygen. Following the administration ofsuxamethonium 6.25 mg, the trachea was in-tubated. The lungs were ventilated manuallyusing the Jackson-Rees modification of Ayre'sT-piece, and anaesthesia was maintained withhalothane and nitrous oxide in oxygen. Surgerywas uneventful and lasted 20 min. The tracheawas extubated once the infant was fully awake andmoving all limbs. Shortly after extubation of histrachea he was returned to the ward where, 30 minafter anaesthesia, he had an apnoeic attack andbecame cyanosed. He responded to manualventilation with 100% oxygen via a face mask.

Patient 2. A 17-day-old 4.06-kg male infantpresented for Ramstedt's pyloromyotomy. Afteraspiration of the nasogastric tube, anaesthesia wasinduced with thiopentone 12.5 mg and trachealintubation was facilitated with suxamethonium6.25 mg. Cricoid pressure was maintained fromthe start of the induction until correct placementof the tracheal tube was confirmed by auscultation.The lungs were ventilated mechanically using a

Dow

nloaded from https://academ

ic.oup.com/bja/article/59/6/672/245772 by guest on 19 July 2022

676 BRITISH JOURNAL OF ANAESTHESIA

Servo 900C ventilator. Anaesthesia was main-tained with 0.5% halothane and nitrous oxide inoxygen. Pancuronium 0.3 mg was administered.Surgery and anaesthesia were uneventful. Theduration of anaesthesia was 45 min. Upon ces-sation of anaesthesia, and after adequate ant-agonism of neuromuscular blockade, withneostigmine 0.24 mg and atropine 0.12 mg—asshown by an adequate response to peripheralnerve stimulation—the infant developed a jerky,inadequate ventilatory pattern. Blood-gas analysiswas normal. He was transferred to the intensivecare unit and required ventilatory assistance.Following 7 h of i.v. fluid therapy and intermittentpositive pressure ventilation, he was weaned fromIPPV and the trachea was extubated. The patientmade an uneventful recovery. Blood sugar con-centration was normal throughout the intra- andpost-operative periods.

Patient 3. A 42-day-old full-term female whoweighed 4.7 kg presented with a 3-day history ofvomiting and was mildly dehydrated on ad-mission. She was resuscitated for a 24-h period,before undergoing surgery. Atropine 0.08 mg wasgiven i.m. as a premedicant. A rapid sequence i.v.induction was used in this patient with thio-pentone 20 mg and suxamethonium 7.5 mg.After intubation of the trachea, the lungs wereventilated mechanically, anaesthesia was main-tained with nitrous oxide in oxygen, and neuro-muscular blockade was achieved withpancuronium 0.4 mg. This child was givenmorphine 0.4 mg during the operation.

After antagonism of neuromuscular blockadewith neostigmine 0.28 mg and atropine 0.14 mg,the infant remained drowsy and unresponsive.Spontaneous ventilation was inadequate. Nalor-phine 0.5 mg and later naloxone 40 ng failed toproduce significant improvement in the level ofconsciousness, or ventilation.

Blood sugar concentration, serum electrolyteconcentrations and the chest x-ray were normal.Blood-gas analysis showed a respiratory acidosis(pH7.23: Pco, 8 kPa; Pao, 37 kPa; HCO,24.6 mmol litre"1). She was transferred to theintensive care unit for controlled ventilation.Eighteen hours after operation spontaneousbreathing was adequate and her trachea was extu-bated. She remained drowsy for a further 24 h.There were no sequelae.

Patient 4. A 3.7-kg male infant presented forRamstedt's pyloromyotomy. Again, a rapid se-quence induction was used with thiopentone 20 mg

and suxamethonium 5 mg. Anaesthesia was main-tained with nitrous oxide in oxygen, and pan-curonium 0.4 mg was the neuromuscular blockingagent used. The lungs were mechanically ven-tilated. In this patient fentanyl 12.5 ug was givenduring the operation. On completion of surgery,residual neuromuscular blockade was antagonizedwith neostigmine 0.2 mg and atropine 0.1 mg.However, the infant failed to resume spontaneousbreathing and required ventilatory assistance for60 min. Subsequent recovery was uneventful.

Three of the patients in this series were inextremis on admission as a result of severedehydration. They had marked hypochloraemicalkalosis with serum chloride concentrationsranging from 44 to 67 mmol litre"1, and a serumpotassium concentration of less than 2.5 mmollitre"1. Admission to the intensive care unit andmeticulous correction of metabolic derangement,which took 3-7 days, resulted in uneventfulsurgery and anaesthesia. Two of these threepatients were discharged home within 12 days ofsurgery. The third developed a septicaemia on the10th day after operation, but was eventuallydischarged home well after 5 weeks. It is essentialthat any underlying metabolic derangement anddehydration are completely corrected beforesurgery is contemplated. As pointed out by Stevenand Allen (1973), Ramstedt's pyloromyotomy isan elective operation. Emergency surgery for thiscondition is never necessary, nor is it desirable.

Anaesthetic technique

In 1973, Steven and Allen recommendedinhalation induction as the method of choice forpyloromyotomy. The widespread use of Sellick'smanoeuvre (Sellick, 1961) associated with a rapidsequence i.v. induction in adults, has spread topaediatric anaesthesia and has been recommendedas the technique of choice by Dierdorf andKrishna (1981), and Battersby and his colleagues(1984). As can be seen from figure 1, anaesthesiais induced in the vast majority of patients in thisunit by this method. In our series both techniquesproved safe and no case of aspiration occurred.The safety of induction depends upon ensuring,as far as possible, that the stomach is emptied, byprior aspiration of the nasogastric tube.

The nasogastric tube was left in place. Itfacilitated the surgeon's testing the integrity ofthe bowel wall after pyloromyotomy. A smallvolume of air was injected down the nasogastrictube and the surgeon manipulated it into the

Dow

nloaded from https://academ

ic.oup.com/bja/article/59/6/672/245772 by guest on 19 July 2022

CONGENITAL HYPERTROPHIC PYLORIC STENOSIS: REVIEW 677

duodenum and occluded the bowel lumen a shortdistance proximal and distal to the incision.Mucosal perforation would have been indicatedby leakage of air.

Anaesthesia was maintained with 60 % nitrousoxide in oxygen, and supplemented in 171 patients(48 %) with a low concentration of a volatile agent.Halothane was used in the majority of patientsalthough, more recently, we have used enfluraneor isoflurane. Neuromuscular blockade was main-tained in the majority of patients with pan-curonium, although in 1984 we used atracuriumsatisfactorily in nine patients.

The use of opioids in infants is a controversialissue. In our series 11 patients received an opioidduring the surgical procedure. Of the five patientswho received morphine, three weighed less than5 kg, and delayed recovery occurred in one ofthese. The six patients who received fentanyl allweighed less than 5 kg and one case of delayedrecovery occurred in this group. Opioid drugsshould be used with caution.

It is important to emphasize the need for closeobservation in the postoperative period. Poten-tially fatal complications can occur during thistime. Hypoglycaemia is always a potential hazard(Pagliara etal., 1973) and was the cause of fatalityin Kiely's series (Kiely et al., 1981).

The postoperative feeding regimen, advocatedby Leahy and Fitzgerald in 1982, has resulted ina smoother postoperative course with less vomit-ing. Previously, feeding was attempted 4 h afteroperation, but this was frequently associated withvomiting.

The morbidity was evenly distributed over thefirst 7 years. However, there was no morbidityrecorded in 1984. There was no significant changein surgical or anaesthetic management in 1984 andwe attributed this morbidity rate to random-ization. Both surgical and anaesthetic morbidityhave occurred since completion of the study in1985.

In summary, we report on 350 consecutivepyloromyotomies in which there were no deathseither directly or indirectly attributable to theprocedure. There was, however, a small butsignificant morbidity.

The importance of thorough preoperativepreparation, careful anaesthesia and close super-vision in the postoperative period is stressed.

REFERENCES

Battersby, E. F., Face, E., Glover, W. J., Hatch, D. J., James,I., Mackersie, A., and Sumner, E. (1984). Anaesthesia forpyloromyotomy. Anaesthesia, 39, 381.

Bell, M. J. (1968). Infantile pyloric stenosis—experience with305 cases at Louisville Children's Hospital. Surgery,64, 983.

Bennet, E. J. (1968). Pyloric stenosis. Clin. Anaesth., 3, 275.Daly, A. M., and Conn, A. N. (1969). Anaesthesia for

pyloromyotomy—A review (The Hospital for Sick Children,Toronto). Can. Anaesth. Soc. J., 16, 316.

Dierdorf, S. F., and Krishna, G. (1981). Anesthetic manage-ment of neonatal surgical emergencies. Anesth. Analg., 60,204.

Dubousset, A. M. (1973). Anaesthesia for pyloric stenosis.Anesth. Analg. Fr., 30, 1067.

Editorial (1984). Lancet, 1, 888.Gray, D. W., Gear, M. W., and Stevens, D. W. (1984). The

results of Ramstedt's operation—room for complacency?Ann. R. Coll. Surg. Engl., 66, 280.

Kiely, E. A, Deasy, J., Hart, P., Kirwan, W. O., and Brady,M. P. (1981). Congenital hypertrophic pyloric stenosis—Areview. Ir. Med.J., 74, 161.

Leahy, A., and Fitzgerald, R. J. (1982). The influence ofdelayed feeding on postoperative vomiting in hypertrophicpyloric stenosis. Br. J. Surg., 69, 658.

Pagliara, A. S., Karl, I. E., Haymond, M., and Kipnis, D. M.(1973). Hypoglycaemia in infancy and childhood—Part I.J. Paediatr., 82, 365.

Sellick, B. A. (1961). Cricoid pressure to control regurgitationof stomach contents during induction of anaesthesia. Lancet,2,404.

Steven, I. M., and Allen, T. (1973). Congenital hypertrophicpyloric stenosis: the anaesthetist's view. Anaesth. Intens.Care, 1, 544.

Touloukain, R. J., and Higgins, E. (1983). The spectrumof serum electrolytes in hypertrophic pyloric stenosis.J. Paediatr. Surg., 18, 394.

Dow

nloaded from https://academ

ic.oup.com/bja/article/59/6/672/245772 by guest on 19 July 2022