Microdiverticulitis: A Rare But Distinct Cause of Small Bowel Obstruction in Early Infancy

By Arthur Cooper, Jane Chatten, John T. Boyle, C. Everett Koop, James A. O'Neill, Jr, Harry C. Bishop, Louise Schnaufer, Moritz M. Ziegler, and Arthur J. Ross, III

New York, New York and Philadelphia, Pennsylvania

�9 During a 20-year period, eight term infants admitted to a large children's hospital for t reatment of dehydration associated with diarrhea (6) and vomiting (2), and who ranged in age from 10 to 60 days at onset of symptoms and 18 to 75 days at presentation, developed distal small bowel obstruction following apparently successful rehydration. With the exception of a single infant who was dead on arrival, and another whose obstruction went unrecognized, all infants came to operation shortly after obstruction was confirmed. In each case, a severely inflamed area of distal ileum (7} or proximal colon (1) was found at autopsy (2} or operation (6) to be the cause; perforation was present in four of the cases. Resection of the diseased segments of intestine, and primary anastamosis, were performed in all six operated cases; reoperation was required in four of the six for leaks (3) and adhesions (1). All but two survived. Pathologically, the resected intestinal segments showed a unique pattern of injury: numerous punctate ulcers were apparent, which undermined the muscularis mucosae, without evidence of necrosis. Regenerating epithelium extended through these defects, resulting in the presence of glandular invaginations, which were surrounded by a brisk inflammatory response: hence the term "'microdiver- ticulitis.'" W e believe this lesion represents a beginning or furtive at tempt at repair of severely inflamed, but viable intestine, and that it is a rare but true cause of small bowel obstruction in early infancy, separate and distinct from necrotizing enterocolitis. �9 1987 by Grune & Stratton, Inc.

INDEX WORDS: Microdiverticulitis; enterocolitis; regional enteritis; Crohn's disease.

S MALL BOWEL OBSTRUCTION in infants who have survived the first weeks of life is most often

due to hernia, volvulus, total colonic aganglionosis, and stricture due to necrotizing enterocolitis. All but the latter result from preexisting congenital lesions. How- ever, there is yet another cause of small bowel obstruc- tion in infants that appears to be acquired, and is the

From the Division of Pediatric Surgery, Babies" Hospital and Columbia University College of Physicians and Surgeons, New York, and the Departments of Pathology, Pediatrics, and Surgery, Children's Hospital of Philadelphia and University of Pennsylvania School of Medicine, Philadelphia.

Presented at the 35th Annual Meeting of the Surgical Section of the American Academy of Pediatrics, Washington, DC, November 1-2, 1986.

Address reprint requests to Arthur Cooper, MD, Department of Surgery, Babies" Hospital, Columbia-Presbyterian Medical Center, 3959 Broadway, New York, NY 10032.

�9 1987 by Grune & Stratton, Inc. 0022-3468/87/2206-0021503.00/0

subject of the present report. Previously called cicatriz- ing enterocolitis, chronic idiopathic ulcerative ileitis, and regional enteritis of early infancy by prior authors in describing isolated similar cases, we have called this condition "microdiverticulitis," in recognition of its most salient histologic feature.

MATERIALS AND METHODS

Since 1967, several thousand infants have been admitted to the Children's Hospital of Philadelphia for dehydration associated with acute gastroenteritis. From this group of patients, we have identified eight (six term, two preterm) infants admitted for treatment of severe dehydration associated with diarrhea (6) or vomiting (2) who developed distal small bowel obstruction following apparently suc- cessful treatment by means of bowel rest, parenteral rehydration, and, usually, intravenous antibiotics. In all cases, the infants had been unresponsive to multiple formula changes, and were referred only after repeated attempts at oral rehydration and refeeding had failed.

Retrospective review of hospital charts and pathologic material constituted the basis for the study.

RESULTS

The patients ranged in age from 10 to 60 days (mean, 25 _+ 16 days) at onset of symptoms, and 18 to 75 days (mean 38 + 17 days) at time of presentation. That they were both profoundly dehydrated and mal- nourished at the time of admission is indicated by the fact that weight on admission (mean 2,776 _+ 739 g) was, on the average, some 10% to 15% less than weight at birth (mean 3,173 +_ 979 g). Details of presentation and findings are shown in Table 1.

With the exception of the single infant who was dead on arrival, and another infant in whom high-grade small bowel obstruction either failed to become mani- fest or went unrecognized (probably the latter), all infants came to operation shortly after obstruction developed and was confirmed by x-ray. This generally occurred within a week or so of admission (mean age 46 _+ 23 days), by which time the patients had long since been adequately rehydrated (mean weight 3,034 _+ 636 g). In each case, a severely inflamed, but anatomically incompletely obstructed, area of distal ileum (7) or proximal ascending colon (1) were found at autopsy (2) or operation (6) to be the cause; perforation was present in three of the cases, and possibly a fourth case as well, in which it was impossi- ble to distinguish a walled-off perforation from severe inflammation. Resection of the diseased segments of

552 Journal of Pediatric Surgery, Vol 22, No 6 (June), 1987: pp 552-555

MICRODIVERTICULITIS 553

Table 1. Details of Presentation, Findings, M a n a g e m e n t , and Outcome in Microdiverticulit is

Case No. 1 2 3 4 5 6 7 8

Born 11/71 6 /72 11/75 2 /76 8/76 4 /84

Sex F M M M M F

Race 8 W B W B W

Home Rural Suburb Urban Suburb Urban Urban

Diet Isomil Similac Similec Similac Similac Similac Weight

Birth (g) 3,940 3,750 1,860 "'36 w k " 4,000 2,780

Admission (g) 3,400 3,400 1,580 2,770 3,240 2,370

Age

Onset (d) 20 21 21 15 10 14

Admission (d) 30 33 24 18 35 44

Resection (d) 35 35 (None) 18 (DOA) 54

Discharge (d) 118 102 32 (died) 46 (died) - - 115 Cultures

Salmonella/Shi . . . . . .

gella

Enteropathic E . . . . Normal flora

coil

Virus - -

Other - C. difficile toxin

Preoperative Temperature

WBC

% Immature X-rays

Barium -enema Resection

Specimen

Length (cm) Perforation

Course

Reoperation

Reinfection

Other

Status

12/85 6 /86

M M

W 8

Suburb Urban

Similae Nursoy

3 , 9 8 0 1 , 9 0 0

3,720 2,350

35 60

42 75

50 86 74 110 (died)

- +

- Camp + Pseud

- Aero + E coil

-- Yeast + Enterob + Klebs

+ B frag

37.5 37.2 Afebrile Afebrile 36.9 38.5 36.8 12.5 12.5 8.1 12.2 5.7 8.3 3.7

46 "many" 0 ? 3 21 25

SBO SBO "'negative" PSBO PS80 PSBO Free air

Normal Normal Normal

Ileum Ileum (None) Ileum (Postmortem) Ileum + +

Colon Colon

2 0 § 32 45 + O 20 + 1 11

+ (2) -- -- + -- --

Leak Adhesions (None) Leak (None)

(POD 3) (POD 17) (POD 6)

Salmonella Salmonella Serratia

(d 106) (d 72) and Candida

E coil sepsis sepsis

Gut normal Dehydration Fetal

postmortem hydantoin

syndrome Lost to Lost to Died Died Dead on arrival Cystic

follow-up follow-up fibrosis

Ileum Ileum + +

Colon Colon 1 0 + 0 3 + 12

+- + (2)

(None)

Alive

Dehissence

(POD 6)

Leak (POD 1 O)

Pseud,

E cell,

Enterob,

Klebs, and 8 frag

sepsis

Died

*Previous hospitalization elsewhere for diarrhea and dehydration.

Abbreviations: POD, postoperative day; SBO, small bowel obstruction; PSBO, partial small bowel obstruction.

bowel, and primary anastamosis, were performed in all six operated cases; reoperation was required in four of the six for leaks (3) or adhesions (1). All but two survived. Among these, the postoperative courses, while most often stormy, were free of relapse. Micro- biological studies were not generally helpful in estab-

lishing a specific etiology for the preceding gastroen- teritis; preoperative leukocyte counts, while variable, included large numbers of immature forms. Details of management and outcome are also shown in Table 1.

Pathologically, the resected intestinal specimens showed a unique pattern of injury: numerous small

554 COOPER ET AL

punctate ulcers were apparent that undermined the muscularis mucosae, without evidence of necrosis. Regenerating epithelium extended through these defects, resulting in the presence of glandular invagi- nations, which were surrounded by a brisk inflamma- tory response. Examples of the gross and microscopic findings are shown in Fig 1.

Grossly, these ulcers are sharply demarcated, linear or oval, and transversely oriented with respect to the bowel wall, but have no special relationship to the mesenteric attachment. They may be inconspicuous by virtue of their small size, boggy mucosa rolling over their edges, and their tendency to occur in the depths of the mucosal folds. The bowel itself is moderately dilated and indurated. It is fragile at operation, devel- oping small perforations with handling.

Microscopically, microdiverticulitis refers to the presence of epithelium in the submucosa, or below an imaginary line across a gap in the muscularis mucosae. As the result o f sectioning of these undermining ulcers at different levels, any of the following may be inter- preted as microdiverticulitis: (1) obvious ulcers with

Fig 1. Gross (A) and microscopic (B) appearance of microdi- verticulitis (see tex t } .

wide inflamed bases, and mucosal edges curling inward over the base; (2) an inflamed slit in the submucosa completely roofed by intact mucosa unsup- ported by muscularis mucosae; and (3) a round or oval inflamed space containing remnants of epithelium in the submucosa under an intact muscularis mucosae. These patterns must be distinguished from tangential sections of intact bowel. The associated inflammation is mainly mononuclear, and sometimes faintly granulo- matous or palisading in ulcer bases, but granulomas are not seen deeper or in the regional nodes: inflamma- tion extends deeply only under the ulcers. Closely spaced ulcers may create pseudopolyps of the mucosal island type, but granulation tissue polyps do not occur. Finally, mucosa between the ulcers is not inflamed.

As yet, no etiologic agent has been identified, although one patient (case 6) has subsequently been found to have cystic fibrosis.

DISCUSSION

One of us (C.E.K.) described in 1947 a case of cicatrizing enterocolitis in a newborn infant. 1 While this case is often cited as the first report of this phenomenon in the medical literature, there are sev- eral features that lead us to believe that the current group of patients differs in many important respects from the index case. First, this infant passed scant meconium and became symptomatic shortly following birth, suggesting a congenital etiology for his symp- toms. Second, mucosal ulceration was present only in the sigmoid colon; the distal ileum and proximal colon were relatively spared, although mucosal inflamma- tion was unquestionably present. Third, there were several areas of frank necrosis; this does not correspond to the histologic pattern observed in the more recent patients. Finally, there was marked hypertrophy of the muscular coats of the colon; while no mention was made of ganglion cells, it must be noted that this report predated those of Swenson on Hirschprung's disease by a year or more. Taken together, these findings seem more suggestive of the enterocolitis that may accom- pany Hirschprung's disease than microdiverticulitis, although an undermining ulcer was present in the sigmoid colon of the index case. Unfortunately, the pathologic material from this case is no longer avail- able for review.

Since the publication of this early report, others have described individual cases in which the clinical and pathologic findings seem much more like those in the present series. 26 In 1971, Miller and Larsen, in reviewing this literature and adding six cases of their own, clearly understood the disease process to be a distinct entity, v However, in none of these instances was the disease process recognized as truly unique.

MICRODIVERTICULITIS 555

Rather, it was believed either to be a variant of Crohn's disease, or the intractable diarrhea of infancy described by Avery and coworkers in 1968. 8 However, the clinical presentation and pathologic characteristics have been so constant and distinct in our experience as to suggest that this entity must be considered a disease in its own right.

It has occurred in term or healthy preterm infants who have been bottle fed at home. The onset of the disease is marked by severe diarrhea with or without vomiting; fever and leukocytosis are usually present. The diarrhea is unresponsive to dietary changes, bowel rest, or antibi- otics; it progresses within a week or so to high-grade distal small bowel obstruction, compelling surgery. The treatment has been resection, end-to-end anastamosis, nutritional support, and intravenous antibiotics.

Unfortunately, resection has not uniformly resulted in immediate cure, especially in instances when all gross disease has not been removed. Indeed, postopera- tive courses in such patients have been prolonged by persistent diarrhea, anastamotic leaks, adhesions, and both gastrointestinal and intraabdominal infections: intraoperative frozen section of the margins of resec- tion might have been useful in indicating the need for further resection in these cases, but none were obtained. Fortunately, both gastrointestinal and sys- temic symptoms resolved during hospitalization. Deaths have been due to nosocomial infection or inefficacy of medical support rather than persistence or recurrence of bowel lesions.

In general, microdiverticulitis is a misnomer for the appearance of undermining ulcers cut at odd angles, but two observations suggest possible anatomic sub- strates for these lesions. First, the muscularis mucosae is usually absent over lymphoid aggregates in normal or diseased bowel, and epithelium often streams inward in such areas. Second, small gaps in the muscu- laris mucosae may be found occasionally in normal bowel with no associated inflammation, scar, or lym- phoid aggregates. Gaps such as these could serve as entry points for earlier inflammation of the submucosa and the subsequent appearance of undermining in patients with enteritis of any cause.

A number of pathogenetic theories may be advanced that could account for the restricted age of presenta- tion and the rarity of this condition among diarrheal diseases of infancy and childhood. Of these, defects in host defense seem most likely. Gut associated lym- phoid tissue is sparse in infants; a follicular distribu- tion of the ulcers could therefore be overlooked. An unusual number of random gaps may occur in the muscularis mucosae of these patients, resulting in the presence of minute congenital diverticula. A relative immunodeficiency might also be present at the time of acquisition of gut flora that could contribute to the development of the lesion. Finally, it is possible that the disease may be associated with cystic fibrosis, as suggested by our case 6 and an additional case that has been reported previously. An unrecognized extrinsic pathogenic organism or substance seems an unlikely cause of the lesions. However, the disease may repre- sent simply a common reaction to intestinal infection that is rarely severe enough to come to pathologic examination. Least likely etiologies are early Crohn's disease or a variant of necrotizing enterocolitis.

We believe this lesion represents beginning, or fur- tive, repair of acutely focally inflamed by viable intes- tine. It is a rare reaction pattern of the small bowel in infancy, distinct from Crohn's disease or neerotizing enterocolitis. The punctate character of the inflamma- tion likely reflects an inherent weakness of the muscu- laris mucosae at sites of still-undeveloped lymph folli- cles, and/or infection localizing in such areas. The diagnosis should be considered whenever gastrointesti- nal motility fails to recover promptly following rehy- dration in an infant with severe dehydration due to diarrhea or vomiting. I f laparotomy is performed, intra-operative frozen sections of the margins of resec- tion should probably be obtained to ensure complete removal of the diseased segment of intestine. However, if this should necessitate too extensive a resection, the ends of the bowel loops should be exteriorized so as to avoid an anastamotic leak and the potentially fatal complication of intraabdominal sepsis.

REFERENCES

l. Koop CE, Perlingiero JG, Weiss W: Cicatrizing enterocolitis in a newborn infant. Am J Med Sci 214:27-32, 1947

2. Uebelhart R: Enteritis regionalis in Saulingsalter. Schweiz Med Wschr 84:1336-1369, 1954

3. Mozer J: Enterite necrosante, oedematcuse et stenosante chez le nouveau-he, thesis. Geneve, ! 955

4. Walter LE, Chaffin L: Regional ileitis in infancy. West J Surg 65:354-357, 1957

5. Borer F: Crohn's disease in infants. Gastroenterologia 95:48- 56, 1961

6. Harrison HE, Spear GS, Dorst JP: Chronic idiopathic ulcer- ative colitis in infancy. J Pediatr 78:538-546, 1971

7. Miller RC, Larsen E: Regional enteritis in early infancy. Am J Dis Child 122:301-311, 1971

8. Avery GB, Villavicencio O, Lilly JR, et al: Intractable diar- rhea in early infancy. Pediatr 41:712-721, 1968