Candidal Infection of Gastric Ulcers Histology, Incidence, and Clinical Significance ANNA-LUISE A. KATZENSTEIN, M.D., AND JOHN MAKSEM, M.D. Katzenstein, Anna-Luise A., and Maksem, John: Candidal infection of gastric ulcers: Histology, incidence, and clinical significance. Am J Clin Pathol 71: 137-141, 1979. Fungal organisms morphologically resembling Candida were found in one third of 72 consecutive surgically resected gastric ulcers. In over half of these cases the organisms were present in the mycelial form. Large numbers of fungi forming clusters were found in 13 cases (18%) and were associated with a surprisingly high postoperative mortality, (38.5%). These organisms are most likely one manifestation of debility in otherwise poor-risk patients rather than a cause of the high morbidity and mortality in these patients. Although organisms of the Candida group are probably not directly etiologic in the development of gastric ulcers, it is possible that their presence aggravates and perpetuates gastric ulceration. (Key words: Candidiasis; Ulcer; Stomach.) Department of Surgical Pathology, Washington University School of Medicine, St. Louis, Missouri All slides were stained with hematoxylin and eosin, Gomori methenamine silver, and Gram stains. The amounts of both bacteria and fungi were graded as 0 (absent), 1 -I- (rare), 2 + (moderate), and 3 + (marked). The fungal form, whether yeast or mycelial, or both, was also recorded. Although cultures were not taken, the organisms in all cases were morphologically consistent with organisms of the Candida group. The medical records of all 72 cases were reviewed with respect to age, sex, race, duration and nature of symptoms, antibiotic or steroid therapy, fever, associated diseases, and pertinent laboratory studies. SUPERFICIAL CANDIDAL OVERGROWTH in the gastrointestinal tract is a relatively common finding at autopsy, especially in debilitated patients and in patients with underlying malignancy. 2 We unexpectedly, however, found candidal infection in specimens of surgically resected gastric ulcers from several patients without known predisposing factors to infection. This observation prompted us to review all cases of gastric ulcers resected during a recent 3'/i-year period in order to assess the incidence and significance of gastric candidiasis in a surgical rather than an autopsy population. Results Pathology Materials and Methods All gastrectomy specimens coded as gastric ulcers in the Barnes Hospital Surgical Pathology files between January 1973 and June 1976 were reviewed. Of 117 cases identified, 45 were eliminated from the study for various reasons, including duodenal, marginal, or esophageal location (11), ulceration of a carcinoma or a lymphoma (6), healed to nonexistent ulcer (19), lack of medical records (5), and miscellaneous reasons (4). The 72 remaining cases formed the basis of this study. Received January 12, 1978; received revised manuscript and accepted for publication March 31, 1978. Supported in part by PHS Grant #5 T01 CA 05201-08. Address reprint requests to Dr. Katzenstein: Department of Pathology, Case Western Hospital, Cleveland, Ohio 44118. All of the ulcers were active; a layer of neutrophils and fibrin covered the base, which was composed of granulation tissue or fibrosis. The cases were divided into the following three groups, depending on the presence and number of fungal organisms; Group I-no fungi identified. There were 48 ulcers in which no fungi were found. Clusters of mixed bacteria were present on the surface of the ulcer in six cases (12.5%). Group II-scattered isolated fungi. In 11 ulcers rare yeasts or pseudohyphae were scattered on the surface of the exudate. The organisms could be seen in the hematoxylin-and-eosin stain in only one case; the remaining required special stains. There was no invasion by the fungi into deeper layers of the exudate. Mixed pseudohyphal and yeast forms were found in six cases, while only yeasts were found in five. Clusters of mixed bacteria were present within the superficial layer of exudate in two cases (18%). Group Ill-numerous fungi in clusters. Thirteen ulcers containing numerous yeasts or pseudohyphae, 0002-9173/79/0200/0137 $00.75 © American Society of Clinical Pathologists 137 138 KATZENSTEIN AND MAKSEM A.J.C.P. • February 1979 ,»^ tew -it* S5.-.»*•.« A-4; TO?*" * FIG. 1. Low magnification of a typical gastric ulcer from a patient in Group III showing a thick layer of acute inflammation andfibrincovering a fibrotic base. Numerous mycelial forms can be seen in the exudate {inset), even without special stains. Hematoxylin and eosin x 150. Inset x600. usually arranged in clusters. The organisms were seen with the hematoxylin-and-eosin stain in ten cases (Fig. 1), while special stains were required for their identification in three cases. A mixture of yeasts and pseudohyphae was present in nine cases, while only yeasts were found in the remaining four. The organisms were present within the superficial exudate in most cases, but they invaded the deep layers of exudate and even the superficial ulcer bed in four cases (Fig. 2). Clusters of bacteria were present on the ulcer surface in four cases (31%). Clinical Findings The major clinical findings for the three groups are summarized in Table 1. As shown in this table, patients with only scattered isolated fungi in their ulcers (Group II) had clinical features similar to patients with ulcers negative for fungi (Group I). Group III, however, differed from the other groups in several respects. There were twice as many female subjects as male subjects in Group III, in contrast to an approximately equal sex ratio in the other two groups. Almost one third of patients in Group III received preoperative antibiotics, compared to only 8% of Group I and none of Group II. Preoperative and postoperative fevers and postoperative sepsis were more common in Group III than in the other groups. Alcoholism and cancer were more frequent in Group III. The mortality of Group III was greater than that for Groups I and II. Since there was no significant difference in clinical features whether fungi were absent (Group I) or whether rare isolated fungi were found (Group II), these two groups were combined for statistical comparison with Group III. Using Fisher's exact test for two-by-two tables,3 the high incidence of postoperative sepsis and the high mortality of Group III were found to be statistically significant (P < .01), compared with combined Groups I and II. The differences in preoperative antibiotic therapy and preoperative fever were highly suggestive but were not statistically significant (P > .05). Differences in the other features were also not significant. The relationship of mortality to duration of symptoms is shown in Table 2. The overall percent mortality was GASTRIC ULCER CANDIDIASIS Vol. 71 . No. 2 generally higher for those patients presenting acutely than for those with more chronic symptoms. Nevertheless, those patients in Group III with an acute or subacute presentation still had a higher mortality than those patients in the other two groups presenting with similar acute or subacute symptomatology. The numbers were too small, however, for statistical significance. The cause of death in most cases was a combination of multiple factors including sepsis, pneumonia, pulmonary emboli, renal or hepatic failure. Disseminated candidiasis was not found in any case. Table 3 summarizes in detail the clinical feature of individual patients in Group III. Although four patients had received preoperative antibiotics, one had received corticosteroids, four had an underlying malignancy, and three were alcoholics, it is noteworthy that in five patients in Group III none of the known predisposing factors for candidal infection of the gastrointestinal tract could be identified. The postoperative courses of five patients in Group III were complicated by bacterial sepsis. Candida albi- 139 cans, however, was cultured from the blood of only one patient, and this patient had a central venous pressure catheter. Yeasts (C. albicans or Candida tropicalis) were isolated from the urine of four patients, from the abdominal wounds of two, and from abdominal fluid of one. In contrast, only one patient in Group II had a positive sputum culture for Candida species, and one patient in Group I had a positive urine culture. None of the cultures positive for Candida species were pure cultures; all had concomitant mixed bacterial growth. In Patient 5 specimens from an endoscopic biopsy two months before gastric resection contained numerous yeasts and pseudohyphae. Discussion Although the incidence of fungal infection of the gastrointestinal tract has been estimated to be as low as 1 in 1,000 autopsied patients from general hospitals, the incidence is much higher from hospitals specializing in the treatment of immunosuppressed patients.2 In a series reported from Memorial Hospital, 13% of patients V A FIG. 2. A silver-stained ulcer from a patient in Group III. (Left) Low magnification showing large numbers of mycelia and yeasts with superficial invasion into the ulcer base. Gomori methenamine silver, x 150. (Right) Higher magnification of same case showing typical morphology of the organisms at the base of the ulcer. Gomori methenamine silver. x350. A.J.C.P. • February 1979 KATZENSTE1N AND MAKSEM 140 Table I. Comparison of Clinical Features of Patients in Groups I, II, and III Males/females (ratio) Average age (range) Preoperative antibiotics (%) Steroids (%) Preoperative fever (%) Postoperative fever (%) Postoperative sepsis Group I (N = 48) Group II (N = 11) Group III (N = 13) Duration of Symptoms Number of Patients Number of Deaths (%) 22/26 (.85) 57 (23-82) 6/5 (1.2) 53 (35-84) 4/9 (.44) 63 (34-85) 0 1 (9%) 0 3 (27%) 4 (31%) 1 (7.7%) 3 (23%) 8 (62%) 0 5 (38.5%) 2 (18%) 1 (9%) 7 (64%) 8 (62%) 5 (38.5%) 0 Acute Group I Group II Group III Subacute Group I Group II Group III Chronic Group I Group II Group III 24 14 2 8 15 9 1 5 29 22 7 0 6(25) 3(21) 0 3 (37.5) 2(13) 0 0 2(40) 1 (3.4) 1 (4.5) 0 0 2 (18%) 1 (9%) lt(9%) 0 4 (31%) 3 (23%) 0 5 (38.5%) 4 5 2 19 (8.3%) (10%) (4.2%) (40%) 4 (8.3%) (%) Duration of symptoms Acute (<1 wk) Subacute (1-12 wks) Chronic (>12 wks) Underlying diseases Cancer Alcoholism Miscellaneous Deaths (%) Table 2. Mortality Rate Percent Related to Duration of Symptoms in Groups I, II, and III 14 (29%) 9 (19%) 22 (46%) 0 2 (4.2%) 6*(12.5%) 4 (8.3%) * Includes two chronic renal failure, one diabetes, one dermatomyositis, one systemic lupus erythematosis. one rheumatoid arthritis. t Chronic renal failure. with hematopoietic malignancies and 1.5% of patients with solid malignancies had gastrointestinal candidiasis at autopsy.2 The most common sites of involvement were the stomach and the esophagus. In disseminated candidiasis, the gastrointestinal tract is considered the most common portal of entry for the organisms, especially in leukemic patients.9 Ourfindingof organisms morphologically resembling Candida in 33% of 72 consecutive resected gastric ulcers is surprisingly high. Although candidal organisms may be cultured from the alimentary tract of up to 50% of healthy individuals,6 the presence of large clusters of organisms, as in our cases, is distinctly unusual. Moreover, in over half of our cases (62.5%) mycelia, which are considered to represent the pathogenic phase of the organism7 were present. Invasion into the ulcer bed was seen in four patients in Group III, although in most cases the organisms appeared confined to the more superficial exudate. Most patients with gastrointestinal candidiasis re- Table 3. Summary of Clinical Features of Patients in Group III Patient Duration of Symptoms Preoperative Antibiotic Therapy Preoperative Steroid Therapy Fever Preoperative Postoperative Postoperative Sepsis Underlying Disease Age Race Sex 1* 68 White F 2 Days Ampicillin None None Present None None 2* 70 White F Hours Keflin Chloramphenical None None None Present None 3* 66 Black F 1 Month Keflin None Present Present Present Miliary tuberculosis, rectal cancer 4* 34 Black F 2 Days None None None Present None Thyrotoxicosis 5* 68 White M Several days Keflex, Ancef, Kanamycin None None Present Present Cancer of the nasopharynx, alcoholism 6 85 White F 3 Days None None Present Present None None 7 67 White F 1 Day None Prednisone None None None Cancer of the breast, Hodgkin's disease 8 47 White F 2 Months None None None Present Present None 9 10 11 12 13 60 66 59 62 67 Black White White White White F M M F F 6 Weeks Hours 2 Months Weeks Hours None None None None None None None None None None None Present None None None None Present None Present None None Present None Present None None None Alcoholism Endometrial cancer Alcoholism * Died postoperatively. Vol. 71 • No. 2 141 GASTRIC ULCER CANDIDIASIS ported in autopsy series have had one or more factors predisposing to infection, including previous antibiotic or corticosteroid therapy, cytotoxic chemotherapy, radiation, or underlying malignancy.210 Candidal infection of a surgically resected gastric ulcer has been recently reported in a patient with diabetes mellitus, alcoholism, and Laennec's cirrhosis.8 In contrast, no predisposing factors to infection with Candida could be identified in five of our 13 patients in Group III with large numbers of organisms (Table 3). Three cases of massive yeast overgrowth in stomach remnants following a Billroth I resection have been reported in otherwise healthy patients.1 Although these patients had no known predisposing factors to infection, mechanical factors related to their previous surgery were probably responsible for their candidal infections. Patients in Group II showed clinical features essentially identical to those in Group I; no effect related to the presence of isolated fungal elements could be found in this group. In contrast, patients in Group III differed in several respects from patients in the other two groups. The majority of patients in Group III presented acutely, having had ulcer symptoms for less than one week. Preoperative and postoperative fevers and postoperative sepsis were more common in Group III than in the other groups. The most surprising finding, however, was a significantly higher mortality in Group III (38.5%). This finding may be partially related to the acute presentation of many of these patients. However, even when compared to similar patients in the other groups who presented as acutely (Table 2), the patients in Group III still had a higher mortality rate although this difference was not significant. It is unlikely that the candidal infection itself directly affected mortality, because disseminated candidiasis was not found in any case at autopsy. A more plausible explanation is that the fungi represent a sign of debility in a patient who is at high risk for multiple complications. One question raised by this study is what role, if any, Candida has in the development of gastric ulcers. Candidal infection of gastric ulcers is common in swine, and may be one factor in the pathogenesis of ulcers in these animals.4,5 In our patients in Group III, the acute presentation and paucity of pre-existing ulcer symptoms argue against long-standing candidal infection as a direct cause of ulceration. However, it is entirely possible that these organisms may perpetuate and aggravate preexisting ulcers. One way of studying this problem would be to examine sequential endoscopic biopsies of gastric ulcers for fungi. The practicality of this approach is affirmed by the positive endoscopic biopsy in one of our patients (Patient 5) two months before gastric resection. It is possible that a trial of oral nonabsorbable antifungal agents may be useful both in eradicating the organisms and, at the same time, in diminishing symptoms of the ulcer. References 1. Borg 1, Heijkenskjold F, Nilehn B. et al: Massive growth of yeasts in resected stomach. Gut 7:244-249, 1966 2. Eras P. Goldstein MJ. Sherlock P: Candida infection of the gastrointestinal tract. Medicine 54:367-379, 1972 3. Huntsberger D, Leaverton P: Statistical inference in the biomedical sciences. Allyn and Bacon, Boston, 1970, pp 85-102 4. Kadel WL, Kelley DC, Coles EM: Survey of yeast-like fungi and tissue changes in esophagogastric region of stomachs of swine. Am J Vet Res 30:401-408, 1969 5. Kowakzyk T: Etiologic factors of gastric ulcers in swine. Am J Vet Res 30:393-400. 1969 6. Kozinn PJ, Taschdjian CL: Enteric candidiasis. Diagnosis and clinical considerations. Pediatrics 30:71-85, 1962 7. Kozinn PJ, Taschdjian CL: Candida albicans: Saprophyte or pathogen? JAMA 198:190-192, 1966 8. Mohtashemi H, Davidson FZ: Candidiasis and gastric ulcer. Dig Dis 18:915-919, 1973 9. Myerowitz RL, Pazin GJ, Allen CM: Disseminated candidiasis. Changes in incidence, underlying diseases, and pathology. Am J Clin Pathol 68:29-38. 1977 10. Parker JC, McCloskey JJ, Knauer KA: Pathobiologic features of human candidiasis. A common deep mycosis of the brain, heart and kidney in the altered host. Am J Clin Pathol 65: 991-1000, 1976
© Copyright 2025 Paperzz