- Email: [email protected]
Complications after surgical intervention for necrotizing enterocolitis: A multicenter review
Complications After Surgical Intervention for Necrotizing Enterocolitis: A Multicenter Review By Jeffrey R. Horwitz, Kevin P. Lally, Henry W. Cheu, W. David Vazquez, Houston, Texas; San Antonio, Texas; Indianapolis, Indiana; Cincinnati, 0 Necrotizing enterocolitis (NEC) is a serious condition affecting predominantly the premature infant. The purpose of this study is to report a multicenter experience of complications in 252 infants requiring surgical therapy for NEC. Data from eight institutions for the years 1980 through 1990 were collected and analyzed for infants undergoing surgical therapy for NEC. Records were reviewed for gestational age, birth weight, age at operation, indications for operation, degree of intestinal involvement, operation(s) performed, complications, and 30-day mortality rates. A total of 264 infants underwent surgical intervention for NEC during the study period. Complete information was available for 252 patients. The mean gestational age was 31 f 5 weeks and the mean birth weight was 1,552 f 823 g. The mean age at operation was 18 f 35 days. Pneumoperitoneum was the most common indication for operation (42%). The 30-day survival rate was 72%. Eighty-one percent of patients underwent primary laparotomy, whereas peritoneal drainage was performed in 48 (19%) patients. Postoperative complications were identified in 119 (47%) patients. The most common postoperative complications were sepsis (9%), intestinal strictures (9%), and short gut (8%). Wound infections occurred in 6%, and the incidence of intraabdominal abscess formation was only 2.3%. Gestational age ~27 weeks (P c .005) and birth weight
WORDS:
Necrotizing
enterocolitis,
operative
compli-
From the Departments of Surgery, The University of Texas Medical School and Hermann Children S Hospital, Houston, TX; Wiqord Hall USAF Medical Center, San Antomo, TX; J. W Riley Hospital for Children, Indianapolis, IN; Children’s Hospital Medical Center, Cincinnati, OH; and Children’s Hospital of Pittsburgh, MageeWomen’s Hospital, Mercy Hospital of Pittsburgh, and the Western Pennsylvania Hospital, Pittsburgh, PA. Presented at the 1994 Annual Meeting of the Sectzon on Surgery of the American Academy of Pediatrics, Dallas, Texas, October 21-23, 1994. Address reprint requests to Kevm P. La&, MD, Division of Pediatric Surgery, Umversity of Texas Medical School, 6431 Fannin, Suite 6.264, Houston, TX 77030. Copyright o 1995 by W B. Saunders Company 0022-3468/9513007-0018$03.0010 994
Jay L. Grosfeld, and Moritz M. Ziegler Ohio; and Pittsburgh, Pennsylvania
N
ECROTIZING enterocolitis (NEC) is the most common, life-threatening gastrointestinal emergency seen in patients admitted to most newborn intensive care units.* Preterm, low birth weight infants appear to be at the greatest risk. Medical management consists of bowel rest, fluid rescusitation, intravenous nutrition, correction of acid-base imbalances, and the use of broad-spectrum antibiotics. In 25% to 50% of patients, medical management is unsuccessful and surgical therapy is required.2 Survival after operation has been reported to range between 44% and 87%.3-11 At operation, many patients have intestinal necrosis and perforation, predisposing them to a high incidence of complications and intraabdominal sepsis. Although numerous publications have concentrated on the factors associated with the development and mortality of NEC, there is limited information concerning factors that contribute to complications in patients undergoing operative intervention for NEC. The purpose of this study was to evaluate factors associated with development of postoperative complications in 252 infants requiring operative intervention for NEC. MATERIALS
AND METHODS
All available records for infants with NEC who underwent operation from January 1, 1980 through December 31, 1990 at Hermann Children’s Hospital, Houston, Texas, Wilford Hall USAF Medical Center, San Antonio, Texas; the J.W. Riley Hospital for Children, Indianapolis, Indiana; Cincinnati Children’s Hospital, Cincinnati, Ohio; Children’s Hospital of Pittsburgh, Pittsburgh, Pennsylvania; Magee-Women’s Hospital, Pittsburgh, Pennsylvania; Mercy Hospital of Pittsburgh, Pittsburgh, Pennsylvania; and the Western Pennsylvania Hospital, Pittsburgh, Pennsylvania were reviewed. Patients from the Pittsburgh hospitals encompassed the years 1980 through 1985, and the Wilford Hall data were from 1985 through 1990. The data obtained included gestational age, birth weight, age at the time of operation, indications for operation, degree of intestinal involvement, and the operative procedure(s) performed. Operative complications were specifically searched for and survival was recorded. Operative death was defined as death within 30 days of the initial surgical intervention. The data were analyzed with x2 analysis. A Pvalue of less than .05 was considered significant. RESULTS
A total of 264 infants underwent surgical intervention for NEC during the study period. Complete information was available for the 252 infants that form the basis of this study. The excluded patients lacked information regarding birth weight, gestaJournalofPed/atric
Surgery,
Vol30,
No 7 (July),
1995:
pp 994-999
NEC:
COMPLICATIONS
PRIMARY
Fig 1. laparotomy
LAPAROTOMY
Distribution and initial
AFTER
(~204)
of patients peritoneal
SURGERY
995
PERITONEAL
(n = 252) drainage.
who
INITIAL DRAINAGE
underwent
(n-48)
primary
tional age, or age at operation. Mean gestational age was 31 + 5 weeks (range, 23 to 42 weeks), with 70 infants (28%) 27 weeks’ gestation or less. The mean birth weight was 1,552 + 823 g (range, 470 to 4,000 g), with 77 infants (31%) weighing less than 1,000 g. The mean age at the time of operation was 18 + 35 days (range, 1 to 383 days). One hundred seventeen patients (46%) underwent operation before and including day 10 of life, and 135 (54%) were operated on after day 10. Pneumoperitoneum was the most common indication for operation (42%). In 92 infants (37%), surgical intervention was performed for progressive clinical deterioration (based on a combination of clinical, laboratory, and radiological parameters). In the remaining patients (21%) there were other indications for operation. Information regarding distribution of disease was available for 218 patients. A combination of small bowel and colonic involvement was described in 81 (37%) patients. In these patients the terminal ileum was the most common site of small bowel disease. In 79 (36%) patients, only the small bowel was affected. This group tended to have more diffuse small bowel involvement. In the remaining 58 (27%) infants, the colon was the sole site of disease. Primary Exploration Two hundred four infants (81%) underwent a primary laparotomy. One hundred eighty-nine (93%) had birth weights greater than 1,500 g. The major indications for operation were clinical deterioration (40%) and pneumoperitoneum (36%). One hundred seventy-six infants (70%) underwent resection and creation of stomas. Nineteen (7.5%) patients were explored but the degree of intestinal involvement precluded any resection. Seven patients (3.4%) underwent proximal intestinal diversion only and 2 infants (1%) had resection followed by primary anastomosis (Fig 1). Peritoneal Drainage Forty-eight infants (19%) had peritoneal drainage. Portions of data from these infants have been previ-
ously published.12 These patients were from the same institution(s) and were selected for peritoneal drainage in a nonrandomized fashion based on the individual surgeon’s preference. These infants were considered too unstable to undergo laparotomy. Fortyone (85%) had a birth weight of less than 1,500 g. Pneumoperitoneum was the indication for operation in 32 (67%) of these infants. Twenty-nine of these infants (60%) had drainage alone. Nineteen infants (40%) underwent subsequent laparotomy an average of 2.3 days after drainage (range, 1 to 9 days). Thirteen (68%) had a resection and creation of stomas, 5 (26%) had an exploration only, and 1 (6%) had a resection and primary anastomosis (Fig 1). SurvivaE Overall operative survival rate was 72%. Age at operation (before or after day 10) and the indication for operation (pneumoperitoneum versus clinical deterioration versus other) were not associated with significant differences in survival rates. Gestational age below 27 weeks (P < .OOS) and birth weight under 1,000 g (P < .005) were associated with significant decreases in operative survival. One hundred fifty-four of the 204 (75%) primary laparotomy patients survived compared with 27 of the 48 (56%) drainage patients. Sixteen of the 29 (55%) patients treated with drainage alone survived. Survival for infants weighing 2 1,000 g was 82% (122 of 149) after primary laparotomy and 62% (16 of 26) after peritoneal drainage. However, among the very low birth weight (VLBW) infants ( < 1,000 g), the survival rate after peritoneal drainage (54%) was similar to that after primary laparotomy (58.2%). Patients with isolated colonic disease had a significantly better survival rate compared with cases of small bowel disease alone and combined small-large bowel involvement (88% versus 75% and 64%, P < .05). Complications Postoperative complications occurred in 119 (47%) patients. The most common complications were sepsis, intestinal strictures, and short gut syndrome (Table 1). Wound infections occurred in 15 (6%) Table
1. Complications
Sepsis
23 (9%) 23 (9%)
Stricture Short gut Wound Infection Stoma Bowel obstruction Renal failure Dissemmated intravascular lntraabdommal
(47%)
abscess
22 15 12 9 8 coagulation
(8 7%) (5.9%) (4.7%) (3.5%) (3.1%)
8 (3.1%) 6 (2 3%)
996
HORWITZ
infants. Interestingly, only 6 (2.3%) infants developed an intraabdominal abscess after operation. Infants who developed sepsis postoperatively had a 30-day mortality rate of approximately 50%. Complication rates were not affected by the indication or the age at the time of operation. In contrast to mortality, gestational age and birth weight were not associated with the incidence of postoperative complications (Fig 2). Complications were identified in 63% (30 of 48) of the patients who underwent initial peritoneal drainage compared with 44% (89 of 204) of the patients who had primary laparotomy. Seventeen of the 29 (59%) patients who had drainage alone developed a postoperative complication. Among peritoneal drainage patients, sepsis was the most common postoperative complication (19%), followed by strictures (15%) and fistulae (8%). Short gut (9.8%), strictures (7.3%), wound infections (7.3%), and sepsis (6.8%) were the most frequent postoperative complications in the primary laparotomy group. Infants 2 1,000 g had similar complication rates after peritoneal drainage or primary laparotomy. There was, however, a significant increase in complications among the VLBW infants treated with peritoneal drainage compared with primary laparotomy (82% versus 38%, P < .OS). Sepsis was the most common complication among the VLBW infants (14.2%), whereas short gut (9%) and strictures (9%) were the most common complications after operation in the 2 1,000 g infants (Fig 3). Infants with isolated small bowel involvement had significantly higher complication rates compared with isolated colonic and combination small and large bowel cases (61% versus 37% and 41%, P < .OS). Short gut syndrome (18%) and sepsis (15%) were the most common complications in the isolated smal1 bowel group. In the isolated colonic cases, wound infections (12%) and strictures (10%) were the most frequent complications, whereas sepsis occurred in only two (3.4%) patients. Similarly, sepsis accounted for only 3.7% of complications in the combination small and large bowel group. 60 50 40 %
30 20 10 0
* lII;L, 4000
I
1000-1500
BIRTH
Fig 2.
Mortality
WEIGHT
(0) and complication
* =p
1501-2000
“S >I000
grams
22000
(grams)
(N) rates
by birth
weight.
ET AL
16 14 12 %
10 * 6 4 2 0
Fig 3. >l,OOOg.
I
SEPSIS Specific
STX
complications
by birth
SHORT GUT weight.
n,
~1,000
g; 0,
DISCUSSION
Necrotizing enterocolitis occurs in approximately 2% of all patients admitted to neonatal intensive care units.13 Medical therapy is effective in more than 50% of these infants.3Jj,8J1 Infants who progress to perforation are a select group of critically ill patients evidenced by their significant mortality rates. Several publications with smaller numbers of operative NEC cases have documented significant postoperative complication rates.5~14J5Our operative complication rate was nearly 50% of the entire population. Because of the retrospective nature of this review, the total number of complications may have been higher because some could be missed during the chart review process. This would make our 47% figure a minimum complication rate that occurred during the study period. Postoperative complications were not related to the indication for operation. Complications developed in 53% of patients with pneumoperitoneum and in 45% when clinical deterioration was the primary indication for operation. Although it is generally believed that operation before perforation should improve outcome, both Ricketts3 and Kosloskel’j found that survival was similar in their NEC patients with and without intestinal perforation. Complications occurred with relative consistency across all birth weight groups. This emphasizes the importance of aggressive surveillance of all patients in the early postoperative period. Although the incidence of postoperative complications were relatively constant across birth weight groups, the types of complications differed. VLBW infants were more likely to have septic complications compared with infants 2 1,000 g (14.2% versus 6.8%) while having similar incidences of both intestinal strictures and short gut syndrome. The premature, low birth weight infant has an increased susceptibility to infection. Studies have shown significant reductions of neutrophil migration and phagocytosis17 in addition to deficiencies in immunoglobulin
NEC:
COMPLICATIONS
AFTER
SURGERY
production18 in these patients. This immunological immaturity may be involved in the increased incidence of postoperative septic complications. Complications occurred more frequently after peritoneal drainage compared with primary laparotomy, which probably reflects the severity of disease among the subgroup of drainage patients. The drainage patients were considered too unstable to undergo primary laparotomy. Sepsis was the most frequent complication among drainage patients (19%), occurring in 6.8% of patients that had a primary laparotomy. After laparotomy and resection, infants are usually able to begin enteral feedings within 10 to 14 days. In this study, drainage patients were maintained on parenteral nutrition for an average of almost 33 days after operation. The need for prolonged central venous access combined with extended periods of TPN may have been important contributing factors to the increased incidence of septic complications. Only two patients in this review underwent primaIy anastomosis. Both survived, and neither developed a complication. Harberg et all9 performed a primary anastomosis on 27 consecutive operative NEC patients. Their overall survival rate was 89%, and no postoperative strictures or anastomotic leaks were encountered. Favorable results with resection and primary anastomosis have also been reported by other groups.20,21 In contrast, Cooper et a122found no improvement in survival for primary anastomosis compared with traditional resection. The limited number of patients in our series treated with resection and primary anastomosis makes any evaluation of associated postoperative morbidity in this subset of patients invalid. Infants with isolated small bowel involvement had a statistically significant increase in postoperative complications compared with both isolated colonic and combination small and large bowel cases. Small bowel involvement in the combination cases normally consisted of only the distal ileum, whereas in the small bowel only group the amount of proximal intestinal involvement was greater. Although we lacked complete information regarding postoperative feeding, with less small bowel disease, the isolated colonic and combination patients should have tolerated earlier enteral feedings and would not have required several months of parenteral nutrition. This may have decreased their risk for infectious complications. Interestingly, sepsis accounted for 15% of the complications in the small bowel only group, compared with only 5% in the other two groups. The incidence of postoperative strictures in our surviving patients was nearly 13%. The reported incidence of intestinal strictures after all cases of NEC ranges from 10% to 25% in combined data from
997
a number of series.23-26 Reviews of operative NEC cases, in smaller series, have documented postoperative stricture formation in 14% to 32% of patients.5,9,10,1j.27 Strictures occurred nearly twice as often in infants treated with peritoneal drainage, compared with infants who underwent laparotomy and bowel resection. Janik et alz6 also reported an increased rate of stricture formation in their patients treated with peritoneal drainage compared with patients who underwent laparotomy (33% versus 6%). Had the drainage patients undergone primary laparotomy, the damaged areas prone to stricture formation may have been resected. Short gut syndrome occurred in 20 (9.8%) of the patients that had primary laparotomy compared with 2 (4.1%) patients who underwent peritoneal drainage. It is possible that peritoneal drainage, by providing a window to complete the rescuscitation and stabilize these seriously ill patients, allowed intestine of questionable viability to improve before laparotomy. Despite a perforation rate of approximately 41%, the incidence of wound infection (6%) and intraabdominal abscess (2.3%) formation was extremely low. In Rickett’G study, only 7 of 86 (8%) surviving patients had a wound complication. Similar findings of infrequent wound complications have been reported by Kosloske and Martin28 and Kurscheid and Holschneider.27 A comparable series of perforation and necrosis in older children and adults reported the incidence of intraabdominal abscess formation to be nearly 19%.29 The reasons for these differences are unclear. Although we had inadequate culture data, others have reported positive cultures in 88% to 92% of all neonatal gastrointestinal perforations30-33 of which the majority were NEC. Although the culturepositive rate is high, no single organism was predominant and anaerobes were isolated in only 15% to 50%.30,31,33This contrasts results from adult studies that found a high rate of retrieval of anaerobic organisms from intraabdominal sites of infection.34 The lower number of anaerobes may contribute to the low incidence of abscess formation. Abscess formation in the abdominal cavity requires bacterial localization through fibrin deposition and decreased clearance.35 The exact nature of the peritoneal response in the premature infant is largely unknown. Certainly, in neonates, the omentum is less well developed, and the abdominal cavity is small and has a different anatomy compared with adults. It is unknown to what degree these factors are important, but it is clear that the premature infant’s response is probably different as evidenced by the success that has been reported in the VLBW infants treated with primary drainage alone for perforated NEC.12J6s37
HORWITZ
998
ET AL
REFERENCES 20. Griffiths DM, Forbes DA, Pemberton PJ, et al: Primary anastomosis for necrotizing enterocolitis: A 1Zyear experience. J Pediatr Surg 24:515-518,1989 21. Sparnon AL, Kiely EM: Resection and primary anastomosis for necrotizing enterocolitis. Pediatr Surg Int 2:101-104, 1987 22. Cooper A, Ross AJ, O’Neill JA, et al: Resection with primary anastomosis for necrotizing enterocolitis: A contrasting view. J Pediatr Surg 23:64-68,1988 23. Kosloske AM, Burstein J, Bartow SA: Intestinal obstruction due to colonic strictures following neonatal necrotizing enterocolitis. Ann Surg 192:202-207,198O 24. Schwartz MZ, Richardson CJ, Hayden CK, et al: Intestinal stenosis following successful medical management of necrotizing enterocolitis. J Pediatr Surg 15:890-899,198O 25. Radhakrishnan J, Blechman G, Shrader C, et al: Colonic strictures following successful medical management of necrotizing enterocolitis: A prospective study evaluating early gastrointestinal contrast studies. J Pediatr Surg 26:1043-1046,199l 26. Janik JS, Ein SH, Mancer K: Intestinal strictures after necrotizing enterocolitis. J Pediatr Surg 16:438-443,198l 27. Kurscheid T, Holschneider AM: Necrotizing enterocolitis (NEC)-Mortality and long term results. Eur J Pediatr Surg 3:139-143,1993 28. Kosloske AM, Martin LW Surgical complications of neonatal necrotizing enterocolitis. Arch Surg 107:223-228,1973 29. Greenall MJ, Evans M, Pollock AV: Should you drain a perforated appendix. Br J Surg 65:880-882,1978 30. Bell MJ: Perforation of the gastrointestinal tract and peritonitis in the neonate. Surg Gynecol Obstet 160:20-26, 1985 31. Bell MJ, Ternberg JL, Bower RJ: The microbial flora and antimicrobial therapy of neonatal peritonitis. J Pediatr Surg 15:569-573, 1980 32. Emanuel B, Zlotnik P, Raffensperger JG: Perforation of the gastrointestinal tract in infancy and childhood. Surg Gynecol Obstet 146:926-928,1978 33. Kosloske AM, Ulrich JA: A bacteriologic basis for the clinical presentation of necrotizing enterocolitis. J Pediatr Surg 15:558-564,198O 34. Swenson RM, Lorber B, Michaelson TC, et al: The bacteriology of intra-abdominal infections. Arch Surg 109:398-399,1974 35. McRitchie DI, Girotti MJ, Glynn MFX, et al: Effect of systemic fibrinogen depletion on intra-abdominal abscess formation. J Lab Clin Med 118:48-55,199l 36. Ein SH, Shandling B, Wesson D, et al: A 13-year experience with peritoneal drainage nnder local anesthesia for necrotizing enterocolitls perforation. J Pediatr Surg 25:1034-1037,199O 37. Morgan LJ, Shochat SJ, Hartman GE: Peritoneal drainage as primary management of perforated NEC in the very low birth weight infant. J Pediatr Surg 29:310-315,1994
1. Ghory MJ, Sheldon CA: Newborn surgical emergencies of the gastrointestinal tract. Surg Clin North Am 65:1083-1098, 1985 2. Kosloske AM Necrotizing enterocolitis in the neonate. Surg Gynecol Obstet 148:259-269,1979 3. Ricketts RR: Surgical therapy for necrotizing enterocolitis. Ann Surg 200:653-657,1984 4. Grosfeld JL, Cheu H, Schlatter M, et al: Changing trends in necrotizing enterocolitis: Experience with 302 cases in two decades. Ann Surg 214:300-307,199l 5. Spigland N, Yazbeck S, Desjardins JG: Surgical outcome for necrotizing enterocolitis. Pediatr Surg Int 5:355-358,199O 6. Cikrit D, Mastandrea J, West KW, et al: Necrotizing enterocolitis: Factors affecting mortality in 101 surgical cases. Surgery 96:648-655,1984 7. Jackman S, Brereton RJ, Wright VM: Results of surgical treatment of neonatal necrotizing enterocolitis. Br J Surg 77:146148,199O 8. Pokorny WJ, Garcia-Prats JA, Barry YN: Necrotizing enterocolitis: Incidence, operative care and outcome. J Pediatr Surg 21:1149-1154,1986 9. Robertson JFR, Azmy AF, Young DG: Surgery for necrotizing enterocolitis. Brit J Surg 74:387-389,1987 10. Kosloske AM: Surgery for necrotizing enterocolitis. World J Surg 9:277-284,1985 11. O’Neill JA, Holcomb GW Jr: Surgical experience with neonatal necrotizing enterocolitis (NNE). Ann Surg 189:612-619, 1979 12. Cheu HW, Sukarochana K, Lloyd DA: Peritoneal drainage for necrotizing enterocolitis. J Pediatr Surg 23:557-561, 1988 13. Amoury RA: Necrotizing enterocolitis: A continuing problem in the neonate. World J Surg 17:363-373,1993 14. Cikrit D, West KW, Schreiner R, et al: Long-term follow-up after surgical management of necrotizing enterocolitis: Sixty-three cases. J Pediatr Surg 21:533-535,1986 1.5. Ricketts RR, Jerles ML: Neonatal necrotizing enterocolitis: Experience with 100 consecutive surgical patients. World J Surg 14:600-605,199O 16. Kosloske AM: Indications for operation in necrotizing enterocohtisrevisited. J Pediatr Surg 29:663-666,1994 17. Al-Hadtihy H, Addison IE, Goldstone AH, et al: Defective neutrophil f&ction in low-birth-weight, premature infants. J Clin Path01 34:366-370,1981 / 18. Ball&w M, Cates KL, Rowe JC, et al: Development of the, immune @tern in the very low birth weight (less than 1500 g)\ premature\ infants: Concentrations of plasma immunoglobulins and patterns of infection. Pediatr Res 20:899-904,1986 19. Harberg’FI,, McGill CW, Saleem MM, et al: Resection with primary anastomosis for necrotizing enterocolitis. J Pediatr Surg l&743-746,1983
Discussion Moderator: Can you differentiate any differences between surgical deaths and deaths related to the more global issues of prematurity and the very low birth weight babies? J.R. Horwitz (response): As far as deaths are concerned, the purpose of the study was identifying risk factors in patients for complications. As the slide
showed, mortality seems directly related to age and morbidity seems to be constant across all of the age groups. It does not appear that age is a risk factor for developing morbidity after operation for NEC. R.J. Touloukian (New Haven, CT): This study and survey would give you a wonderful opportunity to compare the modalities of treatment in the various
NEC:
COMPLICATIONS
AFTER
SURGERY
institutions. Did you detect any significant management styles that differed from one institution to another? J,R. Hun&. (response): Unfortunately, we did not have information as far as the different treatment modalities of different institutions. A. Kosloske (Columbus, OH): Multicenter studies are admirable, but sometimes they raise as many questions as they answer. Your study covered a period of 11 years in eight different institutions, which, if you divide it out, is an average of about three cases per year per institution. That’s not very much to draw general conclusions on. My question is, what about the patterns of occurrence? Pittsburgh and Indiana had reported their NEC experience previously, and those institutions had many more than three cases per institution per year. Can you comment on that?
999
JR. Hun&.. (response): As I said, the Indiana patients were from 1980 through 1990, and I believe there were approximately 127 patients. We also had 60 patients from Cincinnati Children’s Hospital, and then the patients from Pittsburgh and from Hermann and Wilford Hall. Dr Grosfeld’s paper was their second group of patients from 1980 through 1990. The Cincinnati cases had never been published before. The Pittsburgh cases were the same patients that we looked at, just in a different fashion. I understand the problem with multicenter information over a large period of time. Hopefully we were able to get an idea of the morbidity in these patients and what some of the risk factors involved are. I think the most important thing from this is that age does not have much of an effect on morbidity, although it has a significant effect on mortality.
WORDS:
Necrotizing
enterocolitis,
operative
compli-
From the Departments of Surgery, The University of Texas Medical School and Hermann Children S Hospital, Houston, TX; Wiqord Hall USAF Medical Center, San Antomo, TX; J. W Riley Hospital for Children, Indianapolis, IN; Children’s Hospital Medical Center, Cincinnati, OH; and Children’s Hospital of Pittsburgh, MageeWomen’s Hospital, Mercy Hospital of Pittsburgh, and the Western Pennsylvania Hospital, Pittsburgh, PA. Presented at the 1994 Annual Meeting of the Sectzon on Surgery of the American Academy of Pediatrics, Dallas, Texas, October 21-23, 1994. Address reprint requests to Kevm P. La&, MD, Division of Pediatric Surgery, Umversity of Texas Medical School, 6431 Fannin, Suite 6.264, Houston, TX 77030. Copyright o 1995 by W B. Saunders Company 0022-3468/9513007-0018$03.0010 994
Jay L. Grosfeld, and Moritz M. Ziegler Ohio; and Pittsburgh, Pennsylvania
N
ECROTIZING enterocolitis (NEC) is the most common, life-threatening gastrointestinal emergency seen in patients admitted to most newborn intensive care units.* Preterm, low birth weight infants appear to be at the greatest risk. Medical management consists of bowel rest, fluid rescusitation, intravenous nutrition, correction of acid-base imbalances, and the use of broad-spectrum antibiotics. In 25% to 50% of patients, medical management is unsuccessful and surgical therapy is required.2 Survival after operation has been reported to range between 44% and 87%.3-11 At operation, many patients have intestinal necrosis and perforation, predisposing them to a high incidence of complications and intraabdominal sepsis. Although numerous publications have concentrated on the factors associated with the development and mortality of NEC, there is limited information concerning factors that contribute to complications in patients undergoing operative intervention for NEC. The purpose of this study was to evaluate factors associated with development of postoperative complications in 252 infants requiring operative intervention for NEC. MATERIALS
AND METHODS
All available records for infants with NEC who underwent operation from January 1, 1980 through December 31, 1990 at Hermann Children’s Hospital, Houston, Texas, Wilford Hall USAF Medical Center, San Antonio, Texas; the J.W. Riley Hospital for Children, Indianapolis, Indiana; Cincinnati Children’s Hospital, Cincinnati, Ohio; Children’s Hospital of Pittsburgh, Pittsburgh, Pennsylvania; Magee-Women’s Hospital, Pittsburgh, Pennsylvania; Mercy Hospital of Pittsburgh, Pittsburgh, Pennsylvania; and the Western Pennsylvania Hospital, Pittsburgh, Pennsylvania were reviewed. Patients from the Pittsburgh hospitals encompassed the years 1980 through 1985, and the Wilford Hall data were from 1985 through 1990. The data obtained included gestational age, birth weight, age at the time of operation, indications for operation, degree of intestinal involvement, and the operative procedure(s) performed. Operative complications were specifically searched for and survival was recorded. Operative death was defined as death within 30 days of the initial surgical intervention. The data were analyzed with x2 analysis. A Pvalue of less than .05 was considered significant. RESULTS
A total of 264 infants underwent surgical intervention for NEC during the study period. Complete information was available for the 252 infants that form the basis of this study. The excluded patients lacked information regarding birth weight, gestaJournalofPed/atric
Surgery,
Vol30,
No 7 (July),
1995:
pp 994-999
NEC:
COMPLICATIONS
PRIMARY
Fig 1. laparotomy
LAPAROTOMY
Distribution and initial
AFTER
(~204)
of patients peritoneal
SURGERY
995
PERITONEAL
(n = 252) drainage.
who
INITIAL DRAINAGE
underwent
(n-48)
primary
tional age, or age at operation. Mean gestational age was 31 + 5 weeks (range, 23 to 42 weeks), with 70 infants (28%) 27 weeks’ gestation or less. The mean birth weight was 1,552 + 823 g (range, 470 to 4,000 g), with 77 infants (31%) weighing less than 1,000 g. The mean age at the time of operation was 18 + 35 days (range, 1 to 383 days). One hundred seventeen patients (46%) underwent operation before and including day 10 of life, and 135 (54%) were operated on after day 10. Pneumoperitoneum was the most common indication for operation (42%). In 92 infants (37%), surgical intervention was performed for progressive clinical deterioration (based on a combination of clinical, laboratory, and radiological parameters). In the remaining patients (21%) there were other indications for operation. Information regarding distribution of disease was available for 218 patients. A combination of small bowel and colonic involvement was described in 81 (37%) patients. In these patients the terminal ileum was the most common site of small bowel disease. In 79 (36%) patients, only the small bowel was affected. This group tended to have more diffuse small bowel involvement. In the remaining 58 (27%) infants, the colon was the sole site of disease. Primary Exploration Two hundred four infants (81%) underwent a primary laparotomy. One hundred eighty-nine (93%) had birth weights greater than 1,500 g. The major indications for operation were clinical deterioration (40%) and pneumoperitoneum (36%). One hundred seventy-six infants (70%) underwent resection and creation of stomas. Nineteen (7.5%) patients were explored but the degree of intestinal involvement precluded any resection. Seven patients (3.4%) underwent proximal intestinal diversion only and 2 infants (1%) had resection followed by primary anastomosis (Fig 1). Peritoneal Drainage Forty-eight infants (19%) had peritoneal drainage. Portions of data from these infants have been previ-
ously published.12 These patients were from the same institution(s) and were selected for peritoneal drainage in a nonrandomized fashion based on the individual surgeon’s preference. These infants were considered too unstable to undergo laparotomy. Fortyone (85%) had a birth weight of less than 1,500 g. Pneumoperitoneum was the indication for operation in 32 (67%) of these infants. Twenty-nine of these infants (60%) had drainage alone. Nineteen infants (40%) underwent subsequent laparotomy an average of 2.3 days after drainage (range, 1 to 9 days). Thirteen (68%) had a resection and creation of stomas, 5 (26%) had an exploration only, and 1 (6%) had a resection and primary anastomosis (Fig 1). SurvivaE Overall operative survival rate was 72%. Age at operation (before or after day 10) and the indication for operation (pneumoperitoneum versus clinical deterioration versus other) were not associated with significant differences in survival rates. Gestational age below 27 weeks (P < .OOS) and birth weight under 1,000 g (P < .005) were associated with significant decreases in operative survival. One hundred fifty-four of the 204 (75%) primary laparotomy patients survived compared with 27 of the 48 (56%) drainage patients. Sixteen of the 29 (55%) patients treated with drainage alone survived. Survival for infants weighing 2 1,000 g was 82% (122 of 149) after primary laparotomy and 62% (16 of 26) after peritoneal drainage. However, among the very low birth weight (VLBW) infants ( < 1,000 g), the survival rate after peritoneal drainage (54%) was similar to that after primary laparotomy (58.2%). Patients with isolated colonic disease had a significantly better survival rate compared with cases of small bowel disease alone and combined small-large bowel involvement (88% versus 75% and 64%, P < .05). Complications Postoperative complications occurred in 119 (47%) patients. The most common complications were sepsis, intestinal strictures, and short gut syndrome (Table 1). Wound infections occurred in 15 (6%) Table
1. Complications
Sepsis
23 (9%) 23 (9%)
Stricture Short gut Wound Infection Stoma Bowel obstruction Renal failure Dissemmated intravascular lntraabdommal
(47%)
abscess
22 15 12 9 8 coagulation
(8 7%) (5.9%) (4.7%) (3.5%) (3.1%)
8 (3.1%) 6 (2 3%)
996
HORWITZ
infants. Interestingly, only 6 (2.3%) infants developed an intraabdominal abscess after operation. Infants who developed sepsis postoperatively had a 30-day mortality rate of approximately 50%. Complication rates were not affected by the indication or the age at the time of operation. In contrast to mortality, gestational age and birth weight were not associated with the incidence of postoperative complications (Fig 2). Complications were identified in 63% (30 of 48) of the patients who underwent initial peritoneal drainage compared with 44% (89 of 204) of the patients who had primary laparotomy. Seventeen of the 29 (59%) patients who had drainage alone developed a postoperative complication. Among peritoneal drainage patients, sepsis was the most common postoperative complication (19%), followed by strictures (15%) and fistulae (8%). Short gut (9.8%), strictures (7.3%), wound infections (7.3%), and sepsis (6.8%) were the most frequent postoperative complications in the primary laparotomy group. Infants 2 1,000 g had similar complication rates after peritoneal drainage or primary laparotomy. There was, however, a significant increase in complications among the VLBW infants treated with peritoneal drainage compared with primary laparotomy (82% versus 38%, P < .OS). Sepsis was the most common complication among the VLBW infants (14.2%), whereas short gut (9%) and strictures (9%) were the most common complications after operation in the 2 1,000 g infants (Fig 3). Infants with isolated small bowel involvement had significantly higher complication rates compared with isolated colonic and combination small and large bowel cases (61% versus 37% and 41%, P < .OS). Short gut syndrome (18%) and sepsis (15%) were the most common complications in the isolated smal1 bowel group. In the isolated colonic cases, wound infections (12%) and strictures (10%) were the most frequent complications, whereas sepsis occurred in only two (3.4%) patients. Similarly, sepsis accounted for only 3.7% of complications in the combination small and large bowel group. 60 50 40 %
30 20 10 0
* lII;L, 4000
I
1000-1500
BIRTH
Fig 2.
Mortality
WEIGHT
(0) and complication
* =p
1501-2000
“S >I000
grams
22000
(grams)
(N) rates
by birth
weight.
ET AL
16 14 12 %
10 * 6 4 2 0
Fig 3. >l,OOOg.
I
SEPSIS Specific
STX
complications
by birth
SHORT GUT weight.
n,
~1,000
g; 0,
DISCUSSION
Necrotizing enterocolitis occurs in approximately 2% of all patients admitted to neonatal intensive care units.13 Medical therapy is effective in more than 50% of these infants.3Jj,8J1 Infants who progress to perforation are a select group of critically ill patients evidenced by their significant mortality rates. Several publications with smaller numbers of operative NEC cases have documented significant postoperative complication rates.5~14J5Our operative complication rate was nearly 50% of the entire population. Because of the retrospective nature of this review, the total number of complications may have been higher because some could be missed during the chart review process. This would make our 47% figure a minimum complication rate that occurred during the study period. Postoperative complications were not related to the indication for operation. Complications developed in 53% of patients with pneumoperitoneum and in 45% when clinical deterioration was the primary indication for operation. Although it is generally believed that operation before perforation should improve outcome, both Ricketts3 and Kosloskel’j found that survival was similar in their NEC patients with and without intestinal perforation. Complications occurred with relative consistency across all birth weight groups. This emphasizes the importance of aggressive surveillance of all patients in the early postoperative period. Although the incidence of postoperative complications were relatively constant across birth weight groups, the types of complications differed. VLBW infants were more likely to have septic complications compared with infants 2 1,000 g (14.2% versus 6.8%) while having similar incidences of both intestinal strictures and short gut syndrome. The premature, low birth weight infant has an increased susceptibility to infection. Studies have shown significant reductions of neutrophil migration and phagocytosis17 in addition to deficiencies in immunoglobulin
NEC:
COMPLICATIONS
AFTER
SURGERY
production18 in these patients. This immunological immaturity may be involved in the increased incidence of postoperative septic complications. Complications occurred more frequently after peritoneal drainage compared with primary laparotomy, which probably reflects the severity of disease among the subgroup of drainage patients. The drainage patients were considered too unstable to undergo primary laparotomy. Sepsis was the most frequent complication among drainage patients (19%), occurring in 6.8% of patients that had a primary laparotomy. After laparotomy and resection, infants are usually able to begin enteral feedings within 10 to 14 days. In this study, drainage patients were maintained on parenteral nutrition for an average of almost 33 days after operation. The need for prolonged central venous access combined with extended periods of TPN may have been important contributing factors to the increased incidence of septic complications. Only two patients in this review underwent primaIy anastomosis. Both survived, and neither developed a complication. Harberg et all9 performed a primary anastomosis on 27 consecutive operative NEC patients. Their overall survival rate was 89%, and no postoperative strictures or anastomotic leaks were encountered. Favorable results with resection and primary anastomosis have also been reported by other groups.20,21 In contrast, Cooper et a122found no improvement in survival for primary anastomosis compared with traditional resection. The limited number of patients in our series treated with resection and primary anastomosis makes any evaluation of associated postoperative morbidity in this subset of patients invalid. Infants with isolated small bowel involvement had a statistically significant increase in postoperative complications compared with both isolated colonic and combination small and large bowel cases. Small bowel involvement in the combination cases normally consisted of only the distal ileum, whereas in the small bowel only group the amount of proximal intestinal involvement was greater. Although we lacked complete information regarding postoperative feeding, with less small bowel disease, the isolated colonic and combination patients should have tolerated earlier enteral feedings and would not have required several months of parenteral nutrition. This may have decreased their risk for infectious complications. Interestingly, sepsis accounted for 15% of the complications in the small bowel only group, compared with only 5% in the other two groups. The incidence of postoperative strictures in our surviving patients was nearly 13%. The reported incidence of intestinal strictures after all cases of NEC ranges from 10% to 25% in combined data from
997
a number of series.23-26 Reviews of operative NEC cases, in smaller series, have documented postoperative stricture formation in 14% to 32% of patients.5,9,10,1j.27 Strictures occurred nearly twice as often in infants treated with peritoneal drainage, compared with infants who underwent laparotomy and bowel resection. Janik et alz6 also reported an increased rate of stricture formation in their patients treated with peritoneal drainage compared with patients who underwent laparotomy (33% versus 6%). Had the drainage patients undergone primary laparotomy, the damaged areas prone to stricture formation may have been resected. Short gut syndrome occurred in 20 (9.8%) of the patients that had primary laparotomy compared with 2 (4.1%) patients who underwent peritoneal drainage. It is possible that peritoneal drainage, by providing a window to complete the rescuscitation and stabilize these seriously ill patients, allowed intestine of questionable viability to improve before laparotomy. Despite a perforation rate of approximately 41%, the incidence of wound infection (6%) and intraabdominal abscess (2.3%) formation was extremely low. In Rickett’G study, only 7 of 86 (8%) surviving patients had a wound complication. Similar findings of infrequent wound complications have been reported by Kosloske and Martin28 and Kurscheid and Holschneider.27 A comparable series of perforation and necrosis in older children and adults reported the incidence of intraabdominal abscess formation to be nearly 19%.29 The reasons for these differences are unclear. Although we had inadequate culture data, others have reported positive cultures in 88% to 92% of all neonatal gastrointestinal perforations30-33 of which the majority were NEC. Although the culturepositive rate is high, no single organism was predominant and anaerobes were isolated in only 15% to 50%.30,31,33This contrasts results from adult studies that found a high rate of retrieval of anaerobic organisms from intraabdominal sites of infection.34 The lower number of anaerobes may contribute to the low incidence of abscess formation. Abscess formation in the abdominal cavity requires bacterial localization through fibrin deposition and decreased clearance.35 The exact nature of the peritoneal response in the premature infant is largely unknown. Certainly, in neonates, the omentum is less well developed, and the abdominal cavity is small and has a different anatomy compared with adults. It is unknown to what degree these factors are important, but it is clear that the premature infant’s response is probably different as evidenced by the success that has been reported in the VLBW infants treated with primary drainage alone for perforated NEC.12J6s37
HORWITZ
998
ET AL
REFERENCES 20. Griffiths DM, Forbes DA, Pemberton PJ, et al: Primary anastomosis for necrotizing enterocolitis: A 1Zyear experience. J Pediatr Surg 24:515-518,1989 21. Sparnon AL, Kiely EM: Resection and primary anastomosis for necrotizing enterocolitis. Pediatr Surg Int 2:101-104, 1987 22. Cooper A, Ross AJ, O’Neill JA, et al: Resection with primary anastomosis for necrotizing enterocolitis: A contrasting view. J Pediatr Surg 23:64-68,1988 23. Kosloske AM, Burstein J, Bartow SA: Intestinal obstruction due to colonic strictures following neonatal necrotizing enterocolitis. Ann Surg 192:202-207,198O 24. Schwartz MZ, Richardson CJ, Hayden CK, et al: Intestinal stenosis following successful medical management of necrotizing enterocolitis. J Pediatr Surg 15:890-899,198O 25. Radhakrishnan J, Blechman G, Shrader C, et al: Colonic strictures following successful medical management of necrotizing enterocolitis: A prospective study evaluating early gastrointestinal contrast studies. J Pediatr Surg 26:1043-1046,199l 26. Janik JS, Ein SH, Mancer K: Intestinal strictures after necrotizing enterocolitis. J Pediatr Surg 16:438-443,198l 27. Kurscheid T, Holschneider AM: Necrotizing enterocolitis (NEC)-Mortality and long term results. Eur J Pediatr Surg 3:139-143,1993 28. Kosloske AM, Martin LW Surgical complications of neonatal necrotizing enterocolitis. Arch Surg 107:223-228,1973 29. Greenall MJ, Evans M, Pollock AV: Should you drain a perforated appendix. Br J Surg 65:880-882,1978 30. Bell MJ: Perforation of the gastrointestinal tract and peritonitis in the neonate. Surg Gynecol Obstet 160:20-26, 1985 31. Bell MJ, Ternberg JL, Bower RJ: The microbial flora and antimicrobial therapy of neonatal peritonitis. J Pediatr Surg 15:569-573, 1980 32. Emanuel B, Zlotnik P, Raffensperger JG: Perforation of the gastrointestinal tract in infancy and childhood. Surg Gynecol Obstet 146:926-928,1978 33. Kosloske AM, Ulrich JA: A bacteriologic basis for the clinical presentation of necrotizing enterocolitis. J Pediatr Surg 15:558-564,198O 34. Swenson RM, Lorber B, Michaelson TC, et al: The bacteriology of intra-abdominal infections. Arch Surg 109:398-399,1974 35. McRitchie DI, Girotti MJ, Glynn MFX, et al: Effect of systemic fibrinogen depletion on intra-abdominal abscess formation. J Lab Clin Med 118:48-55,199l 36. Ein SH, Shandling B, Wesson D, et al: A 13-year experience with peritoneal drainage nnder local anesthesia for necrotizing enterocolitls perforation. J Pediatr Surg 25:1034-1037,199O 37. Morgan LJ, Shochat SJ, Hartman GE: Peritoneal drainage as primary management of perforated NEC in the very low birth weight infant. J Pediatr Surg 29:310-315,1994
1. Ghory MJ, Sheldon CA: Newborn surgical emergencies of the gastrointestinal tract. Surg Clin North Am 65:1083-1098, 1985 2. Kosloske AM Necrotizing enterocolitis in the neonate. Surg Gynecol Obstet 148:259-269,1979 3. Ricketts RR: Surgical therapy for necrotizing enterocolitis. Ann Surg 200:653-657,1984 4. Grosfeld JL, Cheu H, Schlatter M, et al: Changing trends in necrotizing enterocolitis: Experience with 302 cases in two decades. Ann Surg 214:300-307,199l 5. Spigland N, Yazbeck S, Desjardins JG: Surgical outcome for necrotizing enterocolitis. Pediatr Surg Int 5:355-358,199O 6. Cikrit D, Mastandrea J, West KW, et al: Necrotizing enterocolitis: Factors affecting mortality in 101 surgical cases. Surgery 96:648-655,1984 7. Jackman S, Brereton RJ, Wright VM: Results of surgical treatment of neonatal necrotizing enterocolitis. Br J Surg 77:146148,199O 8. Pokorny WJ, Garcia-Prats JA, Barry YN: Necrotizing enterocolitis: Incidence, operative care and outcome. J Pediatr Surg 21:1149-1154,1986 9. Robertson JFR, Azmy AF, Young DG: Surgery for necrotizing enterocolitis. Brit J Surg 74:387-389,1987 10. Kosloske AM: Surgery for necrotizing enterocolitis. World J Surg 9:277-284,1985 11. O’Neill JA, Holcomb GW Jr: Surgical experience with neonatal necrotizing enterocolitis (NNE). Ann Surg 189:612-619, 1979 12. Cheu HW, Sukarochana K, Lloyd DA: Peritoneal drainage for necrotizing enterocolitis. J Pediatr Surg 23:557-561, 1988 13. Amoury RA: Necrotizing enterocolitis: A continuing problem in the neonate. World J Surg 17:363-373,1993 14. Cikrit D, West KW, Schreiner R, et al: Long-term follow-up after surgical management of necrotizing enterocolitis: Sixty-three cases. J Pediatr Surg 21:533-535,1986 1.5. Ricketts RR, Jerles ML: Neonatal necrotizing enterocolitis: Experience with 100 consecutive surgical patients. World J Surg 14:600-605,199O 16. Kosloske AM: Indications for operation in necrotizing enterocohtisrevisited. J Pediatr Surg 29:663-666,1994 17. Al-Hadtihy H, Addison IE, Goldstone AH, et al: Defective neutrophil f&ction in low-birth-weight, premature infants. J Clin Path01 34:366-370,1981 / 18. Ball&w M, Cates KL, Rowe JC, et al: Development of the, immune @tern in the very low birth weight (less than 1500 g)\ premature\ infants: Concentrations of plasma immunoglobulins and patterns of infection. Pediatr Res 20:899-904,1986 19. Harberg’FI,, McGill CW, Saleem MM, et al: Resection with primary anastomosis for necrotizing enterocolitis. J Pediatr Surg l&743-746,1983
Discussion Moderator: Can you differentiate any differences between surgical deaths and deaths related to the more global issues of prematurity and the very low birth weight babies? J.R. Horwitz (response): As far as deaths are concerned, the purpose of the study was identifying risk factors in patients for complications. As the slide
showed, mortality seems directly related to age and morbidity seems to be constant across all of the age groups. It does not appear that age is a risk factor for developing morbidity after operation for NEC. R.J. Touloukian (New Haven, CT): This study and survey would give you a wonderful opportunity to compare the modalities of treatment in the various
NEC:
COMPLICATIONS
AFTER
SURGERY
institutions. Did you detect any significant management styles that differed from one institution to another? J,R. Hun&. (response): Unfortunately, we did not have information as far as the different treatment modalities of different institutions. A. Kosloske (Columbus, OH): Multicenter studies are admirable, but sometimes they raise as many questions as they answer. Your study covered a period of 11 years in eight different institutions, which, if you divide it out, is an average of about three cases per year per institution. That’s not very much to draw general conclusions on. My question is, what about the patterns of occurrence? Pittsburgh and Indiana had reported their NEC experience previously, and those institutions had many more than three cases per institution per year. Can you comment on that?
999
JR. Hun&.. (response): As I said, the Indiana patients were from 1980 through 1990, and I believe there were approximately 127 patients. We also had 60 patients from Cincinnati Children’s Hospital, and then the patients from Pittsburgh and from Hermann and Wilford Hall. Dr Grosfeld’s paper was their second group of patients from 1980 through 1990. The Cincinnati cases had never been published before. The Pittsburgh cases were the same patients that we looked at, just in a different fashion. I understand the problem with multicenter information over a large period of time. Hopefully we were able to get an idea of the morbidity in these patients and what some of the risk factors involved are. I think the most important thing from this is that age does not have much of an effect on morbidity, although it has a significant effect on mortality.