"Mrs. S" is an 85-year-old nursing home resident with a history of high blood pressure and a previous myocardial infarction who presented to the emergency department with a two-day history of fevers, crampy lower abdominal pain and bloody diarrhea. The patient had been treated for a dental abscess four weeks prior and a subsequent episode of Clostridium difficile (C. difficile) colitis. Mrs. S noted complete resolution of her diarrhea while taking her 10-day course of metronidazole (Flagyl). She had been well for a week before developing a low grade fever and watery diarrhea that became bloody within 24 hours. On initial examination, Mrs. S had a fever of 38.0° C and a moderately tender left lower quadrant. Her initial laboratory studies revealed a mild leukocytosis, and her fecal leukocytes were positive. Her chemistry panel was notable for a mild increase in her blood urea nitrogen (BUN) and creatinine, and a normal urine analysis.

What is the most likely diagnosis? What tests should be ordered? Should any empiric therapy be given?

This case involves an elderly, institutionalized woman with a two-day history of bloody diarrhea, associated fevers, mild dehydration and a recent C. difficile colitis. The differential diagnosis of bloody diarrhea, as in this case, is limited to causes of colitis. This is a commonly infectious, inflammatory bowel disease (rare in this age group) and vascular in origin. Given this patient's advanced age and history of coronary atherosclerosis, ischemic colitis is a possibility. This discussion will begin with a review of ischemic colitis and then highlight two important infectious etiologies: Escherichia coli O157:H7 (E. coli O157:H7) and recurrent C. difficile disease. Other common causes of bloody diarrhea include Campylobacter and Shigella, but E. coli O157:H7 has features that mimic cases of ischemic colitis and is, thus, important to recognize in the elderly.

Ischemic colitis. More than 90% of patients with ischemic colitis are older than 60 and have atherosclerotic disease (Greenwald et al., 2001). Ischemic colitis is generally a small vessel disease; thus, the watershed areas at the splenic flexure and the rectosigmoid junction are at greatest risk for injury where the collateral blood flow is weakest (Greenwald and Brandt, 1998).

While the cause of ischemic colitis has not been clearly delineated (Greenwald and Brandt, 1998; Greenwald et al., 2001), elderly patients typically present following episodes of hypotension, low-flow states to the colon or during shock. Surgery, including abdominal aortic aneurysm repairs, can result in an episode of colonic ischemia. Ischemic colitis has also been less commonly associated with bacterial infections (E. coli O157:H7) (Kravitz et al., 2002; Su et al., 1998), parasitic infections (Entamoeba histolytica) and viral infections (cytomegalovirus). In addition, many of the medications that elderly patients may take, such as nonsteroidal anti-inflammatory drugs, digoxin (Lanoxin) and pseudoephedrine, have also been associated with ischemic colitis.

Ischemic colitis can cause a variable degree of damage to the colon. Patients with ischemic colitis typically present with sudden-onset, mild abdominal pain and tenderness and/or guarding over the involved segment of bowel. They also have a variable amount of bloody diarrhea that begins within 24 hours of the onset of abdominal pain.

Diagnosis of ischemic colitis is best made by colonoscopy or flexible sigmoidoscopy, performed with care to avoid perforation of the involved colon. The typical findings include sharply demarcated areas between normal and involved colonic mucosa, which are segmental in distribution. Involved areas may reveal edema, submucosal hemorrhage and ulcerations. The rectum is typically spared because this area has a dual blood supply and compromise is uncommon. Barium enemas are less sensitive than colonoscopies and do not allow for biopsies; the mucosal and submucosal injuries are seen as thumbprints on barium exams and can occasionally be seen on abdominal plain films. Barium enemas must also be done cautiously with minimal distension of the colon; further colonic distension can decrease the blood flow to the involved colon and promote further injury. Angiography is rarely helpful in establishing the diagnosis of colonic ischemia because colonic injury occurs without occlusion of the major arteries that supply the colon.

The typical course of ischemic colitis is spontaneous resolution of symptoms within 24 to 48 hours and colonic healing within one to two weeks. When the colon is more severely injured, healing may take up to six months. When patients develop gangrene and perforation, their clinical course will rapidly progress and necessitate urgent surgical intervention. Severe damage can result in chronic ischemic strictures.

Initial therapy is directed at stabilizing and supporting the patient. Patients should be placed on bowel rest with administration of broad spectrum antibiotics and should be closely followed for more serious complications.

Escherichia coli. Escherichia coli O157:H7 was first recognized as a human pathogen in 1982 (Mead and Griffin, 1998). It is common in epidemics, but sporadic cases have also been reported. Healthy cattle are the main reservoir for this bacteria; human infection is most commonly acquired by the ingestion of contaminated undercooked ground beef. Outbreaks have also been associated with contaminated water, milk, juices, vegetables and fruits believed to be contaminated by the feces of infected cattle. Person-to-person transmission has been a documented concern in nursing homes and day care centers. Following an outbreak in a nursing home attributed to contaminated sandwich meat, a second wave of cases occurred due to person-to-person transmission (Carter et al., 1987).

The pathogenesis of E. coli O157:H7 is only partially understood. The bacteria adhere to the epithelial cells in the colon and produce toxins that resemble those of the Shigella dysenteriae. These toxins, known as Shiga-like toxins or verocytotoxins, are believed to have local and systemic effects. The toxins are believed to contribute to the colonic edema, submucosal hemorrhage and ulceration that can lead to bloody diarrhea. When the toxins enter the bloodstream, they can bind to the endothelial cells in the kidneys and may trigger platelet aggregation, subsequent hemolysis and platelet consumption in involved organs. Renal failure is thought to be the consequence of microvascular clotting. Although the kidneys are preferentially targeted, other organs, such as those of the central nervous system, may be involved as well.

People with this infection will also have a variable presentation. Many will be asymptomatic carriers or will have nly mild flu-like symptoms. The classic resentation of a severe E. coli O157:H7 infection begins with severe abdominal cramping and watery diarrhea followed by rapid progression to bloody diarrhea; fever may or may not be present. Fewer than 10% of these patients will go on to develop hemolytic uremic syndrome (HUS) or, rarely, thrombotic thrombocytopenic purpura (TTP) (Mead and Griffin, 1998). Children under 5 years of age are most likely to develop HUS, but patients at each end of the age spectrum are at increased risk. Today, this infection may account for 90% of all cases of HUS (Mead and Griffin, 1998).

The Centers for Disease Control and Prevention currently recommends that all patients with bloody diarrhea be screened for E. coli O157:H7; stool samples obtained less than six days from the onset of symptoms are more likely to be positive. Both the radiographic and endoscopic findings are similar to the findings in ischemic colitis. The barium enema may show colon thumbprinting indicative of the mucosal and submucosal injuries. The colonoscopy may reveal colonic edema, with areas of hyperemia, ulcerations and, rarely, pseudomembranes.

Escherichia coli O157:H7 infections are typically self-limited with resolution of symptoms within a week, but complications in the elderly are not uncommon. Elderly patients should be monitored with peripheral blood smears, serial hematocrits and urinalysis during the course of their illness.

Treatment with antibiotic therapy is controversial because in vitro studies have demonstrated an increase in toxin production following antibiotic exposure. Some clinical studies have demonstrated an increased incidence of HUS following the use of antibiotics, while other studies have not. Patients with E. coli O157:H7 should be offered supportive therapy; both elderly patients and children less than 5 years old should be monitored for progression to HUS/TTP.

Clostridium difficile. Antibiotic use is complicated by the development of diarrhea in up to 25% of patients, depending on the antibiotic used. Broad-spectrum antibiotics are more likely to cause diarrhea. Clostridium difficile is responsible for 10% to 20% of the cases of antibiotic-associated diarrhea. Eighty percent of C. difficile infections occur in patients while they are taking antibiotics, but patients remain susceptible to it for months following their antibiotic exposure. Clindamycin (Cleocin), penicillins and cephalosporins are associated with the highest rates of C. difficile colitis, but all antibiotics have been implicated. Elderly patients, immunocompromised hosts and hospitalized patients have been found to be at greater risk to develop C. difficile infections.

Clostridium difficile is widely present in the environment, especially in the hospital setting where colonization rates can be as high as 20% to 30% (Bartlett, 2002). Antibiotics change the colonic microflora allowing the C. difficile to proliferate. Clostridium difficile produces two toxins: toxin A (an enterotoxin) and toxin B (a cytotoxin). These toxins cause mucosal injury resulting in diarrhea.

Patients with C. difficile infection typically present with watery diarrhea and abdominal cramps. Symptoms seen in severe cases are bloody diarrhea, fever and abdominal pain. Patients may be dehydrated with prerenal azotemia. In rare situations, patients can develop fulminant colitis, which could be complicated by perforation and megacolon, and even result in death.

Diagnosis of C. difficile infection is established with the stool assay for toxin. While the toxin B tissue culture assay is the gold standard, most laboratories are moving to rapid enzyme immunoassays (EIA) for toxin A or both toxins A and B. While not as specific as the cytotoxic assay, they have a positive predictive value of 60% to 96% and a negative predictive value of 98% to 99% (Merz et al., 1994). A negative toxin A test does not exclude disease because some strains of C. difficile can be diagnosed only by detection of toxin B. Stool cultures for C. difficile are not sufficiently specific because they are positive in asymptomatic carriers. Colonoscopy is usually not needed to establish the diagnosis of C. difficile, but can reveal the classic yellow-white pseudomembranes. The absence of these lesions does not rule out this infection in mild cases (Kyne et al., 2001).

The standard therapy for C. difficile diarrhea is the discontinuation of the inciting antibiotic. Antimotility agents should be avoided. If patients do not respond to the discontinuation of the inciting antibiotic or if they present with severe symptoms of colitis, oral metronidazole or vancomycin (Vancocin) should be given for 10 to 14 days. Metronidazole is favored as first-line therapy because of its significantly lower cost and the avoidance of vancomycin use, which has been associated with the development of resistant enterococci strains. In more severe cases, intravenous metronidazole, high-dose oral vancomycin and vancomycin enemas are indicated.

Recurrences of C. difficile colitis can occur one week to two months following successful therapy (McFarland et al., 1999). As many as 20% of patients who had C. difficile diarrhea will relapse, and one relapse is highly predictive of future relapses (Kyne et al., 2001). Studies have shown that as many as 65% of patients with a relapse are likely to have another relapse in the future (McFarland et al., 2002). Patients are vulnerable to relapses because the fecal flora remain perturbed and the antibiotics do not target the spores that can germinate. Treatment of the first recurrence of the C. difficile infection is a repeated course of metronidazole or vancomycin, whichever was the initially successful therapy. Other approaches to treat the recurrence of C. difficile include: pulsed and/or tapering doses of antibiotic, the restoration of the colonic flora with probiotics, and the use of immunoglobulin therapy (Kyne et al., 2001; McFarland et al., 2002).

A review of ischemic colitis and E. coli O157:H7 and recurrent C. difficile infections shows there is significant overlap in the presentation for each of these three clinical entities. Bloody diarrhea is more typical with ischemic colitis and E. coli O157:H7, but certainly can occur in more severe cases of C. difficile colitis. Escherichia coli O157:H7 can be transmitted person-to-person; outbreaks of this disease have also occurred in nursing homes. Clostridium difficile infections are not uncommon and have a high relapse rate necessitating its consideration in patients who present with significant diarrheal syndromes, recent antibiotic exposure or hospitalization.

Addressing the questions raised by the case presentation of Mrs. S, it is important to note that bloody diarrhea itself presents significant risk of morbidity and mortality to older patients. These patients have the capacity to become quite ill and may progress quickly to severe colitis that requires surgical intervention. On initial presentation, stool studies should be sent off as early as possible to establish a diagnosis. Most labs will routinely culture for common enteric pathogens (Campylobacter, Shigella and Salmonella). Escherichia coli O157:H7 cultures should be done in all cases of bloody diarrhea. Stool studies for C. difficile toxin should be considered in cases of recent antibiotic use or a prior history of C. difficile infections. Empiric antibiotic coverage, in the setting of bloody diarrhea, would be reasonable in an elderly patient. Finally, colonoscopy should be considered for diagnosis if the stool studies are negative and the patient has not improved.

Finding the presumed cause of the bloody diarrhea is important when supportive therapy alone may be inadequate. Consideration of broad spectrum antibiotic use must be made when ischemic colitis is suspected. Directed antibiotic therapy would be indicated for the treatment of recurrent C. difficile disease. The role of antibiotic therapy for E. coli O157:H7 is controversial and generally not indicated. The attempt to establish the diagnosis for Mrs. S will evolve as stool cultures and stool toxin results become available. The endoscopic and radiographic studies may or may not be diagnostic but, along with the stool studies results and the clinical course, they will aid in establishing the correct diagnosis.

Dr. Watanabe is assistant professor in the department of medicine's division of internal medicine at the University of Washington School of Medicine.

Dr. Surawicz is professor in the department of medicine's division of gastroenterology at the University of Washington School of Medicine.

Bartlett JG (2002), Antibiotic-associated diarrhea. N Engl J Med 346(5):334-339.

Carter AO, Borczyk AA, Carlson JA et al. (1987), A severe outbreak of Escherichia coli O157:H7-associated hemorrhagic colitis in a nursing home. N Engl J Med 317(24):1496-1500.

Greenwald DA, Brandt LJ (1998), Colonic ischemia. J Clin Gastroenterol 27(2):122-128.

Greenwald DA, Brandt LJ, Reinus JF (2001), Ischemic bowel disease in the elderly. Gastroenterol Clin North Am 30(2):445-473.

Kravitz GR, Smith K, Wagstrom L (2002), Colonic necrosis and perforation secondary to Escherichia coli O157:H7 gastroenteritis in an adult patient without hemolytic uremic syndrome. Clin Infect Dis 3(9):e103-e105.

Kyne L, Farrell RJ, Kelly CP (2001), Clostridium difficile. Gastroenterol Clin North Am 30(3):753-777, ix-x.

McFarland LV, Elmer GW, Surawicz CM (2002), Breaking the cycle: treatment strategies for 163 cases of recurrent Clostridium difficile disease. Am J Gastroenterol 97(7):1769-1775.

McFarland LV, Surawicz CM, Rubin M et al. (1999), Recurrent Clostridium difficile disease: epidemiology and clinical characteristics. Infect Control and Hosp Epidemiol 20(1):43-50.

Mead PS, Griffin PM (1998), Eshcerichia coli O157:H7. Lancet 352(9135):1207-1212.

Merz CS, Kramer C, Forman M et al. (1994), Comparison of four commercially available rapid enzyme immunoassays with cytotoxin assay for detection of Clostridium difficile toxin(s) from stool specimens. J Clin Microbiol 32(5):1142-1147.

Su C, Brandt LJ, Sigal SH et al. (1998), The immunohistological diagnosis of E. coli O157:H7 colitis: possible association with colonic ischemia. Am J Gastroenterol 93(7):1055-1059 [see comment].

Infectious Diarrhea in the Elderly: Case Study