Original Contribution

The American Journal of Gastroenterology (2005) 100, 1828–1834; doi:10.1111/j.1572-0241.2005.41510.x

Nosocomial Antibiotic-Associated Diarrhea Associated with Enterotoxin-Producing Strains of Methicillin-Resistant Staphylococcus Aureus

John M Boyce MD1 and Nancy L Havill MT1

1Department of Medicine, Hospital of Saint Raphael; and Yale University School of Medicine New Haven, Connecticut

Correspondence: John M Boyce, MD, Section of Infectious Diseases, Hospital of Saint Raphael, 1450 Chapel Street, New Haven, CT 06511

Received 14 September 2004; Revised  0000; Accepted 15 February 2005.





The aim of this study is to present new evidence that enterotoxin-producing strains of methicillin-resistant Staphylococcus aureus may cause nosocomial antibiotic-associated diarrhea.



We conducted a prospective study that utilized standard methods to exclude other bacterial, parasitic, and viral pathogens as causes of nosocomial diarrhea in patients with heavy growth of methicillin-resistant S. aureus in their stool. Staphylococcal enterotoxin assays were performed on S. aureus strains recovered from patients' stools and on stool specimens from affected patients. Retrospective cohort studies compared the severity of diarrhea in patients with methicillin-resistant S. aureus-associated diarrhea with that of patients whose stool did not contain the organism and with patients colonized or infected with enterotoxin-negative methicillin-resistant S. aureus strains.



During an 18-month period, 11 patients had nosocomial antibiotic-associated diarrhea associated with enterotoxin-producing strains of methicillin-resistant S. aureus. Other common bacterial, parasitic, and viral pathogens were excluded. S. aureus strains from the 11 patients produced staphylococcal enterotoxin A, A and B, or D. Eighty-nine percent of patients had the same enterotoxin(s) in stool specimens as produced by the strain recovered from their stool. Case patients had a greater number of days of diarrhea than patients without methicillin-resistant S. aureus in their stool (p < 0.001), or randomly selected patients colonized or infected with enterotoxin-negative methicillin-resistant S. aureus (p < 0.001).



Our findings provide evidence that enterotoxin-producing strains of methicillin-resistant S. aureus may cause nosocomial antibiotic-associated diarrhea. Greater recognition of this disease should result in more rapid and appropriate treatment of affected patients.

Antibiotic-associated diarrhea is a common problem, affecting 3–29% of hospitalized patients receiving antibiotics (1,2). Clostridium difficile, the most commonly identified cause of antibiotic-associated diarrhea, accounts for only 15– 25% of episodes (3), suggesting multiple other causes for this syndrome. In the 1980s and 1990s, investigators in Australia, Japan, and France reported that methicillin-resistant Staphylococcus aureus (MRSA) was a cause of antibiotic-associated diarrhea in hospitalized patients (4,5,6,7,8). Interestingly, despite the fact that MRSA currently accounts for more than 55% of S. aureus isolates recovered from intensive care unit patients in the United States (9), MRSA has rarely been reported to be a cause of antibiotic-associated diarrhea (10). Recent review articles have questioned its role as a cause of antibiotic-associated diarrhea (11,12), or have not mentioned it as a possible etiology (13,14).

We conducted a prospective study in a university-affiliated hospital to determine if MRSA is a cause of antibiotic-associated diarrhea. Our findings provide new evidence that MRSA may cause nosocomial antibiotic-associated diarrhea, and suggest that this entity frequently goes unrecognized, and that appropriate therapy is often not administered or is delayed.



Identification of S. aureus

Isolates were identified as S. aureus based on colony morphology and slide coagulase tests, and were tested for antimicrobial susceptibility by using automated microdilution broth and standardized disk diffusion methods. Methicillin resistance was confirmed by growth on commercially prepared oxacillin salt screening agar. As part of ongoing infection control surveillance activities, one or more MRSA isolates from each patient are routinely saved and frozen at -70°C.

Initial Screening Stool Specimens

During the study period (August 2001 through February 2003), stool specimens submitted to the Clinical Microbiology Laboratory for C. difficile toxin assay were tested for C. difficile toxins A and B by using an enzyme-linked immunosorbent assay (ELISA) kit (Meridian), an aliquot of stool was inoculated onto colistin–nalidixic acid (CNA) agar, a selective agar for gram-positive bacteria, and the stool specimen was refrigerated. If there was heavy growth of staphylococci on CNA agar, they were identified and tested for antimicrobial susceptibility by using methods cited above, and an aliquot of the stool specimen was frozen. Recovery of MRSA from a stool specimen was promptly reported to nursing personnel caring for the patient if the patient had no prior history of MRSA colonization or infection, so that the patient would be placed in contact precautions.

Detection of Other Common Enteric Pathogens

When stool specimens were negative for C. difficile toxins and yielded heavy growth of MRSA on CNA plates, an aliquot of stool was plated onto a blood agar and MacConkey agar plates to determine if normal aerobic gram-negative enteric bacilli were present. All stools were cultured for enteric pathogens by using blood agar, MacConkey agar, Salmonella–Shigella agar, sorbitol–MacConkey agar, and Campylobacter agar. Stool specimens were examined for ova and parasites by using standard methods.

Tentative Case Definition

Diarrhea was defined as three or more loose or liquid stools, or >250 ml of liquid stool from the rectum or from a rectal tube or colostomy bag per 24 h for at least 2 days (1,15). Diarrhea was classified as nosocomial if onset occurred three or more days following hospital admission, or was present on admission in a patient recently discharged from the hospital. Patients were tentatively classified as having MRSA diarrhea if stool from the patient met all of the following criteria: heavy growth of MRSA, scant or no normal aerobic gram-negative enteric flora, negative for C. difficile toxin A/B, absence of enteric bacterial pathogens or ova and parasites, and not attributable to medications.

Additional Assays for Enteric Pathogens

Stool specimens from 13 patients who met the tentative case definition were analyzed using additional tests to further exclude known enteric pathogens. An aliquot of stool from each patient was inoculated onto cefoxitin–cycloserine–fructose–egg yolk agar and incubated anaerobically to determine if C. difficile was present. Stool specimens were forwarded to the Virology Laboratory at Yale New Haven Hospital and cultured for adenoviruses and enteroviruses by using standard techniques (16), and tested for the presence of rotavirus by using an ELISA kit (Meridian). Stool specimens were tested for norovirus at the Yale School of Public Health by using a modification of reverse-transcriptase polymerase chain reaction (RT-PCR) methods previously described (17). Norovirus-positive control stool samples and primer sequences were provided by the Centers for Disease Control and Prevention.

Staphylococcal Toxin Assays

MRSA isolates recovered from stool specimens were tested for staphylococcal enterotoxins A-E (SEA-E) by using an ELISA kit (Tecra) and for toxic shock syndrome toxin-1(TSST-1) by using a reverse passive latex agglutination kit (Oxoid). Stool specimens from nine patients were available for enterotoxin assays, and were tested for SEA-E by using the same ELISA kit. Specimens were diluted 1:10 in saline, and filtered through a 0.2-mum filter. In several instances, stool filtrates yielded nonspecific reactions, i.e., positive for SEA-E. Additional dilutions (1:10 and 1:100) of these stool filtrates were made, and ELISA tests were repeated. Stool filtrates that still yielded indeterminant results were tested for SEA-D by using a reverse passive latex agglutination (RPLA) assay (Oxoid).

Quantitative Stool Cultures

To determine the number of S. aureus in stool specimens, 1 g of stool from each patient was added to 9 ml of saline, and two further 10-fold dilutions were made in saline. One microliter from each dilution was inoculated onto a blood agar plate, which was incubated overnight at 37°C. Staphylococcal colonies were counted, and were expressed as the number of colony-forming units (CFUs) per gram of stool.

Retrospective Cohort Study 1

To establish if patients with diarrhea due to MRSA have a greater number of loose or liquid stools than do patients who do not have the organism in their stool, retrospective cohort study 1 was conducted. Using a statistical software package (SPSS, Chicago, IL), a random sample of 50 specimens was selected from a computerized list of all stool specimens submitted for C. difficile toxin assays in 2002. Specimens that tested positive for C. difficile toxin A/B or yielded MRSA when screened on CNA agar were excluded, yielding 34 nonduplicate patients who had one or more stool specimens submitted for C. difficile toxin assay, but did not have C. difficile-associated diarrhea and did not have MRSA in their stool. Medical charts of case patients and patients without MRSA in their stool were reviewed and the following data were recorded: age, gender, date of hospital admission, date of onset of loose stools, severity and duration of diarrhea, the number of C. difficile toxin assays performed, and the presence or absence of antibiotic and antacid therapy during the 2 wk prior to onset of loose stools.

Frequency of SE Production among Randomly Selected MRSA

To determine if isolates recovered from patients with diarrhea due to MRSA produced SEA-E more frequently than MRSA isolates in general, we used SPSS statistical software to randomly select a sample of 34 MRSA isolates from a computer-generated list of all MRSA isolates recovered by the Microbiology Laboratory during the study period. The 34 randomly selected MRSA isolates were tested for SEA-E production by using the same ELISA kit used to test isolates recovered from stool specimens of case patients. Isolates that failed to produce detectable SEA-E were classified as enterotoxin negative.

Retrospective Cohort Study 2

To determine if nosocomial diarrhea also was common among patients colonized or infected with enterotoxin-negative strains of MRSA, retrospective cohort study 2 was performed. Medical records of patients whose randomly selected MRSA isolates were enterotoxin-negative were reviewed, and the same clinical variables obtained for retrospective cohort study 1 patients were recorded for patients in retrospective cohort study 2.

Statistical Analysis

Dichotomous variables were analyzed by chi2 or Fisher's Exact test. Continuous variables were analyzed by using Mann-Whitney U tests. Differences between groups were considered to be significant if p-values were less than 0.05.



In the period August 2001–February 2003, a total of 3,590 C. difficile toxin A/B tests were performed, of which 380 (10.6%) were positive. Of the 3,590 stool specimens screened on CNA agar, 321 (8.9%) from 297 patients yielded MRSA. Stool specimens from 13 patients with antibiotic-associated diarrhea met our tentative case definition for MRSA diarrhea. Subsequent assays revealed that stool specimens from all 13 patients were negative for C. difficile when cultured anaerobically. All were negative for bacterial enteric pathogens, ova and parasites, enteroviruses, adenoviruses, rotavirus, and norovirus.

Two patients whose MRSA isolates did not produce detectable amounts of SEA-E were excluded from further analyses. MRSA isolates from the remaining 11 patients tested positive for SEA (2 isolates), SEA and SEB (6 isolates), or SED (3 isolates) (Table 1). Only one MRSA isolate (from case 7) was positive for TSST-1.

Of nine stool specimens available for SE assays, four tested positive SEA and SEB, and one was positive for SEA, SEB, and SEE when tested by ELISA methods. Three of the remaining four stool specimens tested positive for SED by the RPLA method. One stool yielded indeterminant reactions by both ELISA and RPLA methods. Overall, eight (89%) of nine stool specimens tested were found to contain the same SE(s) produced by the MRSA isolate recovered from the respective patient's stool (Table 1).

Quantitative cultures of stool yielded a median of 2.2 times 108 (range: 3 times 107–2.6 times 109) CFUs of MRSA per gram of stool (Table 1).

Demographic and Clinical Characteristics of Case Patients

Five (45%) of the 11 patients were men. Patients ranged from 58–83 yr of age, and had a variety of underlying diseases and conditions. Five patients underwent a surgical procedure 6–11 days prior to onset of symptoms. All 11 patients received antimicrobial therapy in the 2-week period prior to onset of diarrhea. The duration of preceding antimicrobial therapy ranged from 2 to 21 days (median = 7 days). The antibiotics most frequently received prior to onset of diarrhea were fluoroquinolones (91% of case patients), cephalosporins (45%), and metronidazole (18%). Other antibiotics received by one patient each included clindamycin, ampicillin-sulbactam, piperacillin-tazobactam, azithromycin, amikacin, and vancomycin. Two patients who experienced onset of diarrhea prior to admission to the hospital had been recently discharged from the hospital (2 days and 16 days, respectively) prior to admission. Nine patients had onset of diarrhea a median of 9 days (range: 6–24 days) after admission. The medium highest temperature of patients during the diarrheal illness was 38.4°C (range: 38.1–40.3). Nine patients had peripheral blood white blood cells (WBCs) counts greater than 12,000/mm3. During the diarrheal illness, MRSA pneumonia developed in five (45%) patients, one of whom also became bacteremic.

Clinical Characteristics of MRSA Diarrhea

Stools were described as liquid in 10 of the 11 patients, and were green colored in a majority of patients. Fecal leukocyte examinations were performed on stool specimens from five patients; two were negative and three showed a few (25) WBCs. The maximum number of bowel movements in a 24-h period ranged from 5 to 12 (Table 1). The maximum recorded stool volume per 24-h period ranged from 475 to 8,250 ml. The total volume of stool recorded for the entire diarrheal illness ranged from 745 to 38,850 ml (Table 1). Of 10 patients without a colostomy, 7 (70%) required insertion of rectal tube, which remained in place for 3–11 days. Patients had diarrhea for a median of 7 days (range: 3–14).

Abdominal X-rays or computed tomographic scans of the abdomen were performed on 8 (73%) of the 11 patients, and revealed dilated, fluid-filled loops of small bowel in 7 (87.5%) of the 8 patients. One patient had marked dilatation of the colon, and three had evidence of localized thickening of the colonic bowel wall. Two patients underwent colonoscopy, which revealed diffuse mucosal hyperemia without pseudomembranes or ulcers. Initial diagnoses considered by patients' physicians included C. difficile-related pseudomembranous colitis or diarrhea, partial small bowel obstruction, paralytic ileus, and ischemic colitis.

One patient improved without receiving any antimicrobial therapy for diarrhea. Another patient who had recently undergone bowel surgery received several vancomycin enemas, but expired from other causes before oral vancomycin therapy could be initiated. Two patients were treated empirically with oral metronidazole despite having negative C. difficile toxin assay results on several stool specimens. Neither responded clinically to metronidazole. The latter two patients and the remaining seven patients were treated with oral vancomycin. The oral vancomycin therapy was initiated a median of 4 days (range; 1–19) after onset of diarrhea when the presumptive diagnosis of MRSA diarrhea was made. One patient's stool frequency had decreased to less than three per day before receiving oral vancomycin. For the remaining eight patients, stool frequency decreased to less than three bowel movements per day, a median of 3.5 days (range: 1–7) after oral vancomycin therapy was begun.

Retrospective Cohort Study 1

All 11 patients with diarrhea associated with enterotoxin-producing MRSA received antibiotic therapy during the 2-wk period prior to onset of symptoms, compared to 27 (79%) of 34 patients without MRSA in their stool (p > 0.05). Case patients were significantly more likely than patients without MRSA in their stool to have received a fluoroquinolone (91%vs 44%, respectively, p= 0.01). Case patients were also more likely than patients without MRSA in their stool to have received a H2 blocker or proton pump inhibitor (91%vs 62%, respectively, p > 0.05).

When compared to patients without MRSA in their stool, patients with diarrhea associated with MRSA had a significantly greater number of days of diarrhea (p < 0.001), a greater maximum number of bowel movements per 24-h period (p < 0.001), a greater maximum volume of loose or liquid stool per 24-h period (p= 0.01), and a greater total volume of stool recorded during their diarrheal illness (p= 0.006) (Table 2). Patients with MRSA diarrhea were significantly more likely than patients without MRSA in their stool to require insertion of a rectal tube (0.02), and had a rectal tube in place significantly longer (p= 0.002).

Of the 34 patients without MRSA in their stool, 19 had diarrhea (three or more liquid or loose stools per day for at least 2 days). Stool cultures for routine enteric pathogens were performed on specimens from 7 of the 19 patients, and all were negative. In 3 of the 19 patients, diarrhea may have been caused by docusate, colchicine, and polyethylene glycol.

Frequency of SE Production among Randomly Selected MRSA

Only 17 (50%) of the 34 randomly selected MRSA isolates produced SE, compared to 11 (85%) of 13 isolates recovered from patients who met our tentative case definition of MRSA-related diarrhea (p= 0.031). One randomly selected isolate produced SEA, one produced SEB, eight produced SEA and SEB, and seven produced SED.

Retrospective Cohort Study 2

Medical records were available for 16 of the 17 patients whose randomly selected MRSA isolates were enterotoxin negative. These 16 patients with enterotoxin-negative MRSA were less likely than patients with MRSA-related diarrhea to have received preceding antimicrobial therapy (62.5%vs 100%, Fisher's Exact test, p= 0.054). Only two (12.5%) of the 16 patients developed nosocomial diarrhea within 7 days before or after recovery of MRSA from various body sites. One of these two patients with diarrhea was receiving daily administration of oral lactulose, which was considered to cause the patient's loose stools. Patients with enterotoxin-negative MRSA had significantly fewer days of diarrhea (median = 0, range: 0–2) than patients with MRSA diarrhea (median = 7 days, range: 3–14) (p < 0.001), and a lower maximum number of stools per day (median = 2.0, range: 0–8) than patients with MRSA diarrhea (median = 8.1, range: 5–12) (p < 0.001). None of the patients with enterotoxin-negative MRSA required rectal tube placement for loose stools, compared to 64% of patients with MRSA diarrhea (p < 0.001).



In the 1950s and 1960s, enterotoxin-producing strains of S. aureus were considered an important cause of antibiotic-associated diarrhea among hospitalized patients (18,19,20). Following the discovery of the role of C. difficile in causing pseudomembranous colitis and antibiotic-associated diarrhea, S. aureus enterocolitis was assumed by most experts to have represented unrecognized C. difficile-related disease, and S. aureus was discounted as a cause of antibiotic-associated diarrhea. However, following the widespread emergence of MRSA (21,22), investigators in Australia described several patients with antibiotic-associated diarrhea that was attributed to MRSA (4). Japanese investigators reported that MRSA strains associated with postoperative antibiotic-associated diarrhea produced SEA and/or SEC, sometimes in combination with TSST-1 (7,23,24). In France, stools of some patients with antibiotic-associated diarrhea have yielded a predominant growth of strains of MRSA, most of which produced SEA and the bicomponent leucotoxin LukE-LukD (8).

We believe that the present study provides the most compelling evidence to date that enterotoxin-producing strains of MRSA may be a cause of antibiotic-associated diarrhea. In contrast to earlier reports (4,7,8,10,24,25,26,27), our study employed a broad range of assays that effectively excluded not only C. difficile and other bacterial enteric pathogens and parasites, but also enteric viruses including rotavirus and norovirus as causes of nosocomial antibiotic-associated diarrhea in affected patients. Unlike previous investigators, we utilized quantitative cultures to establish that a majority of affected patients have more than 1.0 times 108 CFUs of MRSA per gram of stool.

We confirmed earlier reports that MRSA recovered from the stool of affected patients produce a variety of SEs (23,24,25). A previous study documented that MRSA strains recovered from stool specimens of patients with diarrhea produced SE significantly more often than randomly selected strains, a finding confirmed in our study (8). Since in vitro studies have established that S. aureus can produce enterotoxins in an anaerobic environment (28), we suspected that enterotoxin would be present in the enteric tract of patients whose stool yielded heavy growth of enterotoxin-producing MRSA. To the best of our knowledge, our study is the first to document the presence of SE in stool specimens of patients with antibiotic-associated diarrhea. The results are analogous to C. difficile-related diarrhea, where toxins produced by pathogenic strains can be detected in the stool of affected patients.

Our finding that patients with MRSA-related diarrhea had significantly greater severity and duration of diarrhea than patients without MRSA in their stool makes it unlikely that heavy growth of enterotoxin-producing MRSA from the stool represented mere colonization of the gastrointestinal tract. The fact that only 12.5% of patients colonized or infected with enterotoxin-negative strains of MRSA developed nosocomial diarrhea provides additional evidence that enterotoxin-producing strains of MRSA were responsible for antibiotic-associated diarrhea among case patients.

There is considerable experimental and epidemiological evidence that SEs cause diarrhea. Administration of enterotoxin-producing S. aureus to chinchillas and immunocompromised mice treated with antibiotics causes enteritis (29,30). Per oral administration of purified SE often causes diarrhea in addition to vomiting in monkeys (31). SE-mediated food poisoning causes diarrhea in 68–88% of affected individuals (32,33), and ingestion of purified SE causes both diarrhea and vomiting in human volunteers (34).

Our study has several limitations. The study was retrospective in nature, and includes a small number of case patients. The frequency of antibiotic-associated diarrhea associated with enterotoxin-producing strains of MRSA may have been underestimated by our study, since medical records of some patients with heavy MRSA in their stool had insufficient documentation of the number of bowel movements that occurred per day, which made it impossible to determine if such patients met our case definition for diarrhea. Because more severely ill patients may have better documentation of clinical events, the severity of the diarrheal illness experienced by our case patients may have been greater than usual for individuals with this syndrome. The assays we used to detect SEs are not yet approved for testing stool specimens. However, the fact that SE(s) detected in stool correlated in each case with the SE produced by the patient's MRSA stool isolate suggests that our results were accurate. Another limitation of our study was that we tested MRSA isolates and stool specimens only for SEA-E. Several other SEs (G, H, J) have been described (35,36), and it is plausible one of these may have caused diarrhea in two patients (excluded from the analysis) whose stool contained S. aureus that tested negative for SEA-E. Because there are currently no established criteria for making the diagnosis of antibiotic-associated diarrhea due to MRSA, we may have excluded other affected patients who did not meet our rigorous case definition.

The paucity of reports of MRSA antibiotic-associated diarrhea from other medical centers may be due to several factors. Because it has been recommended that laboratories avoid culturing stool specimens from patients hospitalized more than 3 days (37,38), clinicians are often unaware that some hospitalized patients with antibiotic-associated diarrhea are excreting large numbers of MRSA in their stool. When stool cultures are performed on patients hospitalized for more than a few days, recovery of S. aureus probably has been interpreted by others as mere colonization of the gastrointestinal tract. Since not all strains of MRSA produce SEs, the prevalence of antibiotic-associated diarrhea due to MRSA will depend on the relative frequency of enterotoxin-producing strains in any given facility.

It is possible that prior exposure to enterotoxin-producing S. aureus, with subsequent development of antienterotoxin antibodies, may also affect the risk of developing MRSA antibiotic-associated diarrhea. Patients with food poisoning or serious staphylococcal infections caused by enterotoxin-producing S. aureus often develop serum antibodies directed against SEs (39,40). Monkeys immunized with SEB toxoid are protected against subsequent SEB challenge (41). Similarly, Kyne et al. (42) demonstrated that patients with preexisting serum antibodies directed against C. difficile toxin A are less likely to develop C. difficile-related diarrhea than those who have little or no serum antibodies to toxin A.

Our findings suggest that when patients have persistent nosocomial antibiotic-associated diarrhea that has tested negative on two or more occasions for C. difficile toxins A/B and cannot be attributed to medications, it is reasonable to culture a stool specimen (11). If there is a predominant, heavy growth of MRSA, few if any normal gram-negative bacilli, and no other known enteric pathogens, oral vancomycin therapy should be considered (26). Since overuse of oral vancomycin therapy could promote colonization of patients with vancomycin-resistant enterococci, it should be emphasized that mere recovery of MRSA from the stool of patients with diarrhea should not be considered an indication for oral vancomycin therapy.

Further studies are needed to establish practical clinical and laboratory criteria for diagnosing nosocomial antibiotic-associated diarrhea associated with MRSA. The availability of a rapid SE assay approved for testing stool specimens would contribute to appropriate diagnosis, and would facilitate epidemiological studies to determine the frequency of this disorder in various patient populations. Development of a more suitable animal model of MRSA-associated diarrhea would facilitate studies of the possible mechanisms by which SEs cause nosocomial antibiotic-associated diarrhea.


J. M. Boyce designed the study, participated in data collection and analysis, performed statistical analysis, and wrote the report. N. L. Havill participated in data collection and analysis, performed bacteriologic and parasitologic tests on stool samples, performed staphylococcal enterotoxin assays, and critically reviewed the manuscript. Enteric virus assays and cultures of stool specimens were performed on a fee-for-service basis at the Yale New Haven Hospital Virology Laboratory. M. Hojjati participated in data collection.



  1. Less than 30% of cases of nosocomial antibiotic-associated diarrhea can be attributed to C. difficile-related disease, suggesting that there are other unrecognized causes of this syndrome.
  2. Studies from Australia, Japan, and France have provided evidence that enterotoxin-producing strains of methicillin-resistant S. aureus (MRSA) may cause antibiotic-associated diarrhea, but this syndrome is seldom recognized in American hospitals where MRSA is highly prevalent.
  3. 68–88% of patients with S. aureus enterotoxin-mediated food poisoning have diarrhea as part of their illness, and ingestion of purified staphylococcal enterotoxin by human volunteers causes diarrhea as well as vomiting.

Unique contributions of our study include:

  • We excluded other potential causes of nosocomial diarrhea (including viruses) more completely than previous investigators.
  • First study to characterize the severity of MRSA-related diarrhea.
  • First study to quantify the load of MRSA in the stool of affected individuals.
  • First study to demonstrate the presence of staphylococcal enterotoxin in the stool of patients with MRSA-related antibiotic-associated diarrhea.
  • Cohort studies provided evidence that large numbers of MRSA in the stool did not represent mere colonization of affected patients, and that patients with nonenterotoxin-producing MRSA were less likely to have diarrhea.


  1. McFarland, LV. Diarrhea acquired in the hospital. Gastroenterol Clin N Amer 1993;22: 563–577.
  2. McFarland, LV, Surawicz, CM, Greenberg, RN, et al. Prevention of B-lactam-associated diarrhea by Saccharomyces boulardii compared with placebo. Am J Gastroenterol 1995;90: 439–448. | PubMed | ISI | ChemPort |
  3. Bartlett, JG. Antibiotic-associated diarrhea. Clin Infect Dis 1992;15: 573–581. | PubMed | ISI | ChemPort |
  4. McDonald, M, Ward, P, Harvey, K. Antibiotic-associated diarrhoea and methicillin-resistant Staphylococcus aureus. Med J Aust 1982;1: 462–464.
  5. Morita, H, Tani, M, Adachi, H, et al. Methicillin-resistant Staphylococcus aureus (MRSA) enteritis associated with prophylactic cephalosporin administration and hypochlorhydria, after subtotal gastrectomy. Am J Gastroenterol 1991;86: 791–792.
  6. Yoshida, D, Fukunari, H, Hojo, I, et al. Enterocolitis due to methicillin-resistant Staphylococcus aureus - report of two cases. Bull Tokyo Med Dent Univ 1992;39: 31–34.
  7. Kodama, T, Santo, T, Yokoyama, T, et al. Postoperative enteritis caused by methicillin-resistant Staphylococcus aureus. Surg Today 1997;27: 816–825.
  8. Gravet, A, Rondeau, M, Harf-Monteil, C, et al. Predominant Staphylococcus aureus isolated from antibiotic-associated diarrhea is clinically relevant and produces enterotoxin A and the bicomponent toxin LukE-lukD. J Clin Microbiol 2000;37: 4012–4019.
  9. Centers for Disease Control and Prevention National Nosocomial Infections Surveillance (NNIS) System report, data summary from January 1992 - June 2001, issued August 2001. Am J Infect Control 2001;29: 404–421.
  10. Schiller, B, Chiorazzi, N, Farber, BF. Methicillin-resistant Staphylococcal enterocolitis. Am J Med 1998;105: 164–166.
  11. Bartlett, JG. Antibiotic-associated diarrhea. N Engl J Med 2002;346: 334–339. | Article | PubMed | ISI |
  12. Hogenauer, C, Hammer, HF, Krujs, GJ, et al. Mechanisms and management of antibiotic-associated diarrhea. Clin Infect Dis 1998;27: 702–710. | Article | PubMed | ChemPort |
  13. Cunha, BA. Nosocomial diarrhea. Crit Care Clinics 1998;14: 329–339.
  14. Procop, GW. Gastrointestinal infections. Infect Dis Clin North Am 2001;15: 1073–1108.
  15. Bliss, DZ, Johnson, S, Savik, K, et al. Acquisition of Clostridium difficile and Clostridium difficile-associated diarrhea in hospitalized patients receiving tube feedings. Ann Intern Med 1998;129: 1012–1019. | PubMed |
  16. Landry, ML Hsiung, GD. Primary isolation of viruses. In: Specter S.S., Hodinka R, and YoungS., eds. Clinical virology manual, 3rd Ed. American Society for Microbiology Press: Washington, DC, 2000: 27–42.
  17. Fankhauser, RL, Monroe, SS, Noel, JS, et al. Epidemiologic and molecular trends of "Norwalk-like viruses" associated with outbreaks of gastroenteritis in the United States. J Infect Dis 2002;186: 1–7.
  18. Altemeier, WA, Hummel, RP, Hill, EO. Staphylococcal enterocolitis following antibiotic therapy. Ann Surg 1963;157: 847–858.
  19. Surgalla, MJ, Dack, GM. Enterotoxin produced by micrococci from cases of enteritis after antibiotic therapy. JAMA 1955;158: 649–650.
  20. Casman, EP, Bergdoll, MS,Robinson, J. Designation of staphylococcal enterotoxins. J Bacteriol 1963;85: 715–716.
  21. Cookson, BD, Phillips, I. Epidemic methicillin-resistant Staphylococcus aureus. J Antimicrob Chemother 1988;21(suppl C):57–65.
  22. Boyce, JM. Methicillin-resistant Staphylococcus aureus. Detection, epidemiology, and control measures. Infect Dis Clin North Am 1989;3: 901–913.
  23. Takesue, Y, Yokoyama, T, Kodama, T, et al. Toxin involvement in methicillin-resistant Staphylococcus aureus enteritis in gastroenterological surgery. Gastroenterol Jpn 1991;26: 716–720.
  24. Mayumi, T, Takezawa, J, Takahashi, H, et al. IL-15 is elevated in the patients of postoperative enterocolitis. Cytokine 1999;11: 888–893.
  25. Watanabe, H, Masaki, H, Asoh, N, et al. Enterocolitis caused by methicillin-resistant Staphylococcus aureus: molecular characterization of respiratory and digestive tract isolates. Microbiol Immunol 2001;45: 629–634.
  26. Konishi, T, Idezuki, Y, Kobayashi, H, et al. Oral vancomycin hydrochloride therapy for postoperative methicillin-cephem-resistant Staphylococcus aureus enteritis. Surg Today 1997;27: 826–832.
  27. Takeuchi, K, Tsuzuki, Y, Ando, T, et al. Clinical studies of enteritis caused by methicillin-resistant Staphylococcus aureus. Eur J Surg 2001;167: 293–296.
  28. Belay, N, Rasooly, A. Staphylococcus aureus growth and enterotoxin A production in an anaerobic environment. J Food Prot 2002;65: 199–204.
  29. Tan, T-L, Drake, CT, Jacobson, MJ, et al. The experimental development of pseudomembranous enterocolitis. Surg Gynecol Obstet 1959;108: 415–420.
  30. Kakamura, Y, Aramaki, Y,Kakiuchi, T. A mouse model for postoperative fatal enteritis due to Staphylococcus infection. J Surg Res 2001;96: 35–43.
  31. Bergdoll, MS. Monkey feeding test for staphylococcal enterotoxins. Methods Enzymol 1988;165: 324–333.
  32. Eisenberg, MS, Gaarslev, K, Brown, W, et al. Staphylococcal food poisoning aboard a commercial aircraft. Lancet 1975;2: 595–599.
  33. Holmberg, SD, Blake, PA. Staphylococcal food poisoning in the United States. JAMA 1984;251: 487–489.
  34. Raj, HD, Bergdoll, MS. Effect of enterotoxin B on human volunteers. J Bacteriol 1969;98: 833–834.
  35. Dinges, MM, Orwin, PM, Schlievert, PM. Exotoxins of Staphylococcus aureus. Clin Microbiol Rev 2000;13: 16–34. | PubMed | ISI | ChemPort |
  36. Balaban, N, Rasooly, A. Staphylococcal enterotoxins. Int J Food Microbiol 2000;61: 1–10.
  37. Siegel, DL, Edelstein, PH, Nachamkin, I. Inappropriate testing for diarrheal diseases in the hospital. JAMA 1990;263: 979–982.
  38. Valenstein, P, Pfaller, M,Yungbluth, M. The use and abuse of routine stool microbiology: A College of American Pathologists Q-probes study of 601 institutions. Arch Pathol Lab Med 1996;120: 206–211.
  39. Felsenfeld, O, Nasuniya, N. Staphylococcal antitoxin values in the sera of permanent residents and visitors in Thailand. J Trop Med Hyg 1964;67: 300–303.
  40. Jozefczyk, Z. Specific human antibodies to enterotoxins A, B, and C1 of Staphylococcus aureus: Their increased synthesis in staphylococcal infection. J Infect Dis 1974;130: 1–7. | PubMed |
  41. Bergdoll, MS. Immunization of rhesus monkeys with enterotoxoid B.J Infect Dis 1966;116: 191–196.
  42. Kyne, L, Warny, M, Qamar, A, et al. Asymptomatic carriage of Clostridium difficile and serum levels of IgG antibody against toxin A. N Engl J Med 2000;342: 390–397. | Article | PubMed | ISI | ChemPort |


We thank the laboratory staff of the Emerging Infections Program, Yale School of Public Health, for performing RT-PCR assays for norovirus on stool specimens.