TOC | GI
NEJM Volume 350:38-47; January 1, 2004
Nathan M. Thielman, M.D., M.P.H., and Richard L. Guerrant, M.D.
An otherwise healthy 23-year-old man presents after the acute onset of watery diarrhea that has persisted for two days. He reports associated nausea and cramping but no emesis and is febrile, with a temperature of 38.7°C (101.7°F). How should he be evaluated and treated?
The Clinical Problem
In the United States, an estimated 211 million to 375 million episodes of acute diarrhea occur each year (1.4 episodes per person per year); such episodes are responsible for more than 900,000 hospitalizations and 6000 deaths annually.3,4
Acute diarrhea, defined as an increased frequency of defecation (three or more times per day or at least 200 g of stool per day) lasting less than 14 days, may be accompanied by nausea, vomiting, abdominal cramping, clinically significant systemic symptoms, or malnutrition. We focus here on acute infectious diarrhea in immunocompetent adults in industrialized countries.
Strategies and Evidence
The Foodborne Diseases Active Surveillance Network (FoodNet) of the Centers for Disease Control and Prevention (CDC) collects data on the incidence of diarrhea attributable to nine enteropathogens in 13 percent of the U.S. population (37.4 million people) living in nine states. Of these, the pathogens responsible for the most cases of diarrhea in 2002 were
Thorough clinical evaluation of a patient who presents with acute diarrhea is essential in order to guide a cost-effective, evidence-based approach to initial diagnostic testing and therapy. In six studies conducted between 1980 and 1997, the diagnostic yield of stool cultures ranged from 1.5 to 5.6 percent.6 The estimated cost of $952 to $1,200 per positive culture can be reduced through improved selection and testing of the specimens submitted. As is indicated in Figure 1, the initial clinical evaluation of the patient with acute diarrhea should focus on the assessment of the severity of the illness, the need for rehydration, and the identification of likely causes on the basis of the history and clinical findings.
Because most diarrheal illnesses are self-limited or viral, and nearly half last less than one day,3 microbiologic investigation is usually unnecessary for patients who present within 24 hours after the onset of diarrhea, unless such patients are dehydrated or febrile or have blood or pus in their stool.
The characteristics of the illness, the epidemiologic setting, and the public health implications help to determine whether and what types of fecal testing are appropriate. The major epidemiologic and clinical features associated with common enteropathogens are listed in Table 1.
In the United States, most diarrheal illnesses occur during the winter months; these illnesses are commonly associated with noroviruses (in families and in outbreaks, including those in nursing homes)7 or rotaviruses (in young children). Since these illnesses are typically short-lived (lasting one to three days) and self-limited, supportive fluid therapy is usually sufficient. Systemic illness, fever, or bloody stools should prompt routine stool testing for salmonella, shigella, campylobacter, and (especially if there are bloody stools) Shiga toxinproducing E. coli. When diarrhea is persistent or recurring, or if the history (for instance, of fever or tenesmus) is equivocal, fecal testing by means of microscopy for polymorphonuclear leukocytes or by means of immunoassay for the neutrophil marker lactoferrin can quickly provide additional laboratory evidence to support a presumptive diagnosis of inflammatory diarrhea.
In equivocal cases, the negative predictive value of fecal lactoferrin testing may help to determine the need for routine bacteriologic culture for campylobacter, salmonella, and shigella.8 In developed countries, the sensitivity and specificity of fecal leukocytes for inflammatory diarrhea are 0.73 and 0.84, respectively, and the sensitivity and specificity of lactoferrin for inflammatory diarrhea are 0.92 and 0.79, respectively; in resource-poor countries, the ranges are broader.9
Other testing is necessary in special circumstances (Figure 1). In patients who present with an acute onset of bloody diarrhea (especially without fever) or the hemolyticuremic syndrome, fecal testing for Shiga toxin should be performed, along with cultures for E. coli O157:H7. Bloody diarrhea in recent immigrants or travelers returning from regions where E. histolytica is endemic (such as tropical Africa, Asia, or Latin America) should also prompt fecal testing for that pathogen.10 A report by a patient of recent ingestion of shellfish should prompt special cultures for vibrio species. Persistent abdominal pain, fever, mesenteric adenitis, erythema nodosum, or other immunologic manifestations may suggest Yersinia enterocolitica or Y. pseudotuberculosis, culture for which is enhanced by cold enrichment.
The microbiologic evaluation of hospitalized patients or those with recent exposure to antibiotics in whom diarrhea develops should focus on the diagnosis of toxigenic C. difficile,11 the most common recognized cause of nosocomial diarrhea. Testing for other pathogens in patients who have been hospitalized for more than 72 hours is discouraged, except in an outbreak or in patients 65 years of age or older with preexisting medical conditions, patients with human immunodeficiency virus (HIV) infection, patients with neutropenia, and patients with suspected systemic spread of enteric infection.12
Exposure of a traveler or hiker to untreated water and illnesses that persist for more than seven days should prompt evaluations for protozoal pathogens, especially giardia and cryptosporidium. The testing of one or two specimens by means of enzyme immunoassay has a sensitivity of more than 95 percent and is considerably better than older methods based on microscopical examinations. Less common but readily treatable protozoal causes of persistent diarrhea in normal or immunocompromised persons include Cyclospora cayetanensis and Isospora belli (which can be identified by means of acid-fast staining of stool specimens).
In patients with advanced HIV disease, especially those who have fewer than 50 CD4+ lymphocytes per cubic millimeter, stool studies for cryptosporidium, microsporidia, cyclospora, and isospora and blood cultures or biopsies for Mycobacterium avium complex and cytomegalovirus should be considered.13
Fecal evidence of inflammation in persistent diarrhea should arouse concern about inflammatory bowel disease,14 particularly if the presence of a treatable inflammatory pathogen such as shigella, salmonella, campylobacter, C. difficile, or E. histolytica has been ruled out.
Approaches to Therapy
Rehydration and Nutrition
Regardless of the causative agent, initial therapy should include rehydration. Unless the patient is comatose or severely dehydrated, oral rehydration with a glucose-based electrolyte solution is preferred. The standard formulation recommended by the World Health Organization (WHO) or a newer reduced-osmolarity formula for children can be lifesaving in resource-poor regions and is valuable in the industrialized world for infants, the elderly, immunocompromised patients, and anyone with profuse watery diarrhea (information is available on the Web at http://www.who.int/child-adolescent-health/New_Publications/CHILD_HEALTH/WHO_FCH_CAH_01.22.htm). Most adults in developed countries who have acute diarrhea, including travelers, should be encouraged to drink fluids and take in salt in soups and salted crackers. Nutritional support with continued feeding improves outcomes in children. The use of a "BRAT" diet (bananas, rice, applesauce, and toast) with avoidance of milk products (since a transient lactase deficiency may occur) is commonly recommended, although supporting data are limited.15
Nonspecific Symptomatic Therapy
Nearly 400 over-the-counter products are promoted in the United States for their antidiarrheal properties, but few have been demonstrated to be effective in randomized, controlled trials. Only in the cases of loperamide, bismuth subsalicylate, and kaolin has sufficient evidence of efficacy and safety been collected to permit labeling as an antidiarrheal agent.16
Loperamide, the antimotility agent of choice for adults, inhibits intestinal peristalsis and has antisecretory properties, but unlike other opiates (such as codeine, diphenoxylate, and paregoric), it does not penetrate the nervous system and has no substantial potential for addiction.17 When it is used with antibiotics for traveler's diarrhea18 or bacillary dysentery,19 it may reduce the duration of diarrhea by as much as one day. Because antimotility agents have been implicated in prolonged fever in human volunteers with shigellosis,20 toxic megacolon in patients with C. difficile infection, and the hemolyticuremic syndrome in children infected with Shiga toxinproducing E. coli,21 these agents should be avoided in patients with bloody diarrhea or suspected inflammatory diarrhea.
Although it has not been as effective as loperamide in direct comparisons,22 bismuth subsalicylate can alleviate stool output in children23 or the symptoms of diarrhea, nausea, and abdominal pain in patients with traveler's diarrhea.24 Bismuth subsalicylate significantly decreased the median duration of experimental norovirus illness from 27 hours to 20 hours.25 There are insufficient data from clinical trials to support the use of adsorbents such as kaolinpectin, activated charcoal, and attapulgite. Given the limited scientific evidence, limited clinical benefit, and concern about the toxicity of antidiarrheal agents, pediatric guidelines discourage their use in children.26,27
Whereas appropriate antibiotics are effective in the treatment of shigellosis, traveler's diarrhea, C. difficileassociated diarrhea, and, if given early, campylobacteriosis, antibiotics may prolong the duration of shedding of salmonella28 or C. difficile29 and may increase the risk of life-threatening complications of Shiga toxinproducing E. coli infection.30 The clinical benefit should be weighed against the cost, the risk of adverse reactions, and the risk of harmful eradication of normal flora or the induction of Shiga-toxin production.31 Antibiotic therapies for specific pathogens are listed in Table 2.
Randomized, controlled trials in patients with shigellosis have demonstrated that appropriate antimicrobial therapy shortens the average duration of diarrhea by 2.4 days or more, decreases the duration of fever and tenesmus, and reduces the excretion of infectious organisms.32,33 Antimicrobial therapy is of greatest benefit when it is started early, and it probably reduces person-to-person spread. Increased resistance now limits the effectiveness of trimethoprimsulfamethoxazole, ampicillin, and nalidixic acid, which were once considered optimal therapies for shigellosis. Clinical trials have established the efficacy of newer fluoroquinolones for adults with shigellosis,33,34 and accumulating data suggest that ciprofloxacin is also safe and effective in children with shigellosis.35
In immunocompetent patients with campylobacteriosis, erythromycin reduces carriage of the organism but reduces diarrhea only when it is given within four days after the onset of symptoms.36,37 Immunocompromised patients (particularly those who have the acquired immunodeficiency syndrome [AIDS] and those who are pregnant, who have an increased risk of complications of campylobacter infection) and patients with severe, prolonged, or relapsing illness may benefit from antimicrobial therapy. Dramatic increases in the rate of fluoroquinolone resistance (contemporaneous with the widespread use of fluoroquinolones in poultry feeds) now limit the usefulness of these agents in campylobacter infections.38
In uncomplicated salmonella gastroenteritis, antibiotics are not recommended, since they typically do not alleviate the diarrhea39 and have been associated with prolonged carriage28 and even relapse.40 However, because bacteremia occurs in 2 to 4 percent of patients, persons who are at increased risk for dissemination or metastatic seeding should receive antibiotics. These patients include those 12 months of age or younger and those older than 50 years; those with lymphoproliferative disorders, cancer, hemoglobinopathies, or AIDS; transplant recipients; patients with vascular grafts, artificial joints, degenerative joint diseases, or valvular heart disease; and patients who are taking corticosteroids.
Clues to the diagnosis of Shiga toxinproducing E. coli infection include bloody diarrhea (especially if the patient is afebrile) and a history of eating rare hamburger meat or seed sprouts. Given the concern that antimicrobial agents such as trimethoprimsulfamethoxazole and quinolones induce Shiga-toxin production31 and may increase the risk of the hemolyticuremic syndrome,30 patients with suspected or documented infection with Shiga toxinproducing E. coli should receive supportive care. Risk factors for the development of the hemolyticuremic syndrome with this infection remain controversial and are incompletely understood.41
Empirical Antibiotic Treatment for Community-Acquired Illness
Empirical antibiotic treatment in adults who present with severe, community-acquired diarrhea (Table 3) reduces the average duration of illness by one to two days.28,42,43 However, as noted above, the potential benefits must be weighed against the potential harm, such as prolonged fecal shedding of certain pathogens, the increased risk of relapse, and the increased risk of complications. In patients who have evidence of inflammatory diarrhea (fever, tenesmus, dysentery, or fecal leukocytes or lactoferrin) and in whom diarrhea is not thought to be attributable to Shiga toxinproducing E. coli or fluoroquinolone-resistant campylobacter, empirical treatment with an agent such as a fluoroquinolone antibiotic (or for children, trimethoprimsulfamethoxazole), pending fecal testing, is reasonable.6,44 Given the increasing incidence of fluoroquinolone-resistant campylobacter,38,45 empirical therapy in immunocompromised or severely ill patients or travelers with inflammatory diarrhea (especially those who have traveled to southern Asia, where the rate of quinolone resistance exceeds 80 percent) might include erythromycin or azithromycin.
Although most cases of traveler's diarrhea are self-limited, several clinical trials have demonstrated that fluoroquinolones or other agents many of which are no longer used because of increased resistance reduce the duration of diarrhea from three or four days to one or two days with as little as a single dose.46 Although such data support a recommendation that travelers carry appropriate antibiotics for the self-treatment of diarrhea, most experts do not recommend antimicrobial prophylaxis for travelers (even those who are immunocompromised, whose highest risk may be for infection with pathogens that are not covered by the standard antimicrobial agents used). Use of bismuth subsalicylate (one to two 262-mg tablets taken four times daily) decreases the incidence of traveler's diarrhea by 35 to 65 percent and appears to be safe when used for up to three weeks but may result in the blackening of the tongue and the stools.47,48
The treatment of C. difficileassociated diarrhea has been discussed in detail previously.11 Many patients have improvement if therapy with broad-spectrum antibiotics can be discontinued. Oral metronidazole (which is less expensive) and oral vancomycin are effective in the treatment of C. difficileassociated diarrhea, with response rates of more than 90 percent and a mean time to improvement of three days, but relapses are common.11
The protozoal pathogens that are most frequently associated with diarrhea that persists for more than 7 to 10 days, both in the United States and worldwide, are giardia and cryptosporidium. Giardia infections typically respond to metronidazole (the use of larger doses, if they are tolerated, may result in fewer treatment failures) or tinidazole. Cryptosporidium infection, in contrast, is notoriously difficult to treat, particularly in patients with AIDS and low CD4+ lymphocyte counts, in whom the most effective strategy is to facilitate immunologic reconstitution with optimal antiretroviral therapy.49 Nitazoxanide, which was recently approved by the Food and Drug Administration (FDA) for use in children, may reduce oocyst shedding and diminish diarrhea.50 Since these hardy oocysts are difficult to filter or to kill with chlorine, infections caused by them are best prevented by heating water or food to near-boiling temperatures. Finally, cyclospora infection, which can be detected by means of acid-fast staining of fecal specimens, is readily treated with trimethoprimsulfamethoxazole.51,52
Areas of Uncertainty
The inability to determine the specific causative agent when a patient is first seen greatly complicates management. Technology that will permit rapid, pathogen-specific diagnostic testing could mitigate the inappropriate use of antibiotics, help to track sources and outbreaks, and identify resistant pathogens or fecal contamination of food or water. Whether certain antibiotics increase risk whereas others prevent the hemolyticuremic syndrome in patients with Shiga toxinproducing E. coli infections remains controversial.30,41
Drugs such as calmodulin inhibitors, chloride-channel blockers, enkephalinase inhibitors, and agents that block prostaglandin-mediated pathways are under study as antisecretory agents; at present, their role remains uncertain. A poorly absorbed antibiotic, rifaximin, is effective in the treatment of traveler's diarrhea53 and may prove useful for other bacterial diarrheal illnesses but has not been approved by the FDA. Improved therapy must also address the repair of damaged or inflamed mucosa, since the associated reduction in absorptive function and resulting malnutrition may impair child development and potentially reduce the efficacy of therapies for associated diseases.
Several groups have published guidelines for the management of acute diarrhea in adults.6,44,54 The guidelines that the Infectious Diseases Society of America published in 2001, with which we still concur (available at http://www.idsociety.org/), detail diagnostic strategies, treatment recommendations, and public health implications for adults, listing the quality and strength of evidence supporting each recommendation.6 Several pediatric guidelines emphasize fluid management and the fact that antidiarrheal compounds are not useful.26,27,55
Conclusions and Recommendations
In healthy patients with an acute onset of diarrhea, such as the patient described in the vignette, epidemiologic and clinical evaluation should be used to guide laboratory evaluation and management. For example, a recent history of eating raw oysters should prompt consideration and special culture for vibrio, and a recent history of antibiotic use should prompt testing for C. difficile toxin. Bloody diarrhea or a history of blood in the stool or of the ingestion of undercooked beef should prompt testing for Shiga toxinproducing E. coli; the presence of bloody diarrhea would also warrant testing for shigella, salmonella, and campylobacter. A history of immunocompromise, sickle cell disease, severe atherosclerosis, or intravascular prosthetic material should prompt the testing of blood and stool cultures for salmonella. Recent travel to a developing country would suggest infection with pathogens that typically cause traveler's diarrhea and would warrant empirical therapy with a single dose of a quinolone antibiotic (unless the travel was to southern Asia, which should arouse concern about infection with quinolone-resistant campylobacter). One might also consider empirical treatment for inflammatory diarrhea whose cause was suspected, on the basis of the exposure history, to be shigella or campylobacter infection (Table 2 and Table 3).
If none of the above pathogens were present, but symptoms worsened or persisted for more than five days (especially if there was fever), we would order a fecal examination for inflammation to direct further testing for invasive pathogens (including salmonella, shigella, campylobacter, C. difficile, or E. histolytica); if the examinations were negative for these pathogens, we would evaluate the patient for inflammatory bowel disease. Patients with illness lasting more than seven days should be tested for parasites (including giardia, cryptosporidium, cyclospora, and isospora). Finally, patients with recurrent or refractory giardiasis should be tested for possible immune compromise, such as immunoglobin A deficiency, and patients with refractory cryptosporidiosis, cyclosporiasis, or isosporiasis should be tested for HIV infection.
Drs. Thielman and Guerrant are named as coinventors on a patent held by the University of Virginia for the use of glutamine derivatives in oral rehydration therapy. Dr. Guerrant is named as a coinventor on a patent held by the University of Virginia for a fecal lactoferrin assay. Neither coinventor receives any patent royalties.
From the Department of Medicine, Division of Infectious Diseases and International Health, Duke University Medical Center, Durham, N.C. (N.M.T.); and the Department of Medicine, Division of Infectious Diseases and International Health and the Center for Global Health, University of Virginia School of Medicine, Charlottesville (R.L.G.).
Address reprint requests to Dr. Guerrant at the Center for Global Health, HSC #801379, University of Virginia School of Medicine, Charlottesville, VA 22908, or at firstname.lastname@example.org.