Sample Chapter

April 2003

Section 4

The word diarrhea is derived from the Greek words for "flowing through." For most persons, diarrhea means the frequent passage of loose stools.¹ This definition includes two major components: loose stool consistency (pourable stools) and increased stool frequency (more than two bowel movements daily). Physicians often include a third component: increased stool weight (> 200 g/24 hr), but patients are poor estimators of stool output. In addition, some patients report diarrhea when they have fecal incontinence, even if stools are solid; every patient complaining of diarrhea should be asked about fecal incontinence.

Diarrhea is a universal human experience. Most persons have had acute infectious diarrhea at some time during their lives. The incidence of acute diarrhea is roughly 5% to 7% annually.² Infectious diarrhea is associated with contaminated food and water; it is typically spread via fecal-oral transmission. Chronic diarrhea (lasting > 4 weeks) is also common, with a prevalence of approximately 5% in the United States.³ It is less likely to be caused by infection and is more likely to be a symptom of other disorders.

Diarrhea results from excess water in the stool.⁴ To understand the pathophysiology of diarrhea, it is necessary to understand how water is transported across the mucosa of the gastrointestinal tract. Water moves in response to osmotic gradients established by the absorption of salts (mainly sodium chloride, but also potassium and bicarbonate salts) and nutrients (e.g., monosaccharides, amino acids, and fatty acids). Salts and nutrients move both passively, in response to electrochemical gradients across the mucosa, and actively, in response to molecular pumps located in the enterocyte membranes.⁵

Each day, a typical person ingests about 2 L of fluid and produces 7 to 8 L of secretions (i.e., saliva, gastric juice, bile, pancreatic juice, and succus entericus). Thus, a total volume of 9 to 10 L enters the upper intestine daily. Most of the water is absorbed in the jejunum, along with nutrients. Absorption of residual nutrients and salts in the ileum results in a reduction of the volume of luminal contents entering the colon to only 1 to 1.5 L, a 90% reduction in the volume of fluid entering the intestine each day. The colonic mucosa can absorb salt against large electrochemical gradients and can reclaim 90% of the fluid passing the ileocecal valve each day, making the overall efficiency of small bowel and colonic water absorption about 99%.

Diarrhea develops if the overall efficiency of absorption declines by as little as 1%. This can occur under the following circumstances: the rate of intestinal nutrient and salt absorption decreases; net electrolyte secretion develops (an unusual circumstance except with severe diarrhea such as the diarrhea associated with cholera, in which stool output can exceed 10 L/day); transit through the intestine speeds up, thereby limiting the time available for absorption; or poorly absorbable substances are ingested and increase intraluminal osmotic activity, causing the retention of water within the intestine (osmotic diarrhea).⁶

Common problems that primarily cause a reduction in the rate of intestinal nutrient and salt absorption include mucosal diseases, such as celiac disease; inflammatory diseases that disrupt the integrity of the intestinal mucosa (e.g., Crohn disease); and infections with pathogens that produce toxins that affect enterocyte function.

Isolated acceleration of transit is a poorly recognized mechanism of diarrhea, although diarrhea was historically attributed to it. Some patients with so-called functional diarrhea have rapid intestinal transit, which is likely to be important in the pathogenesis of their condition. Many patients with chronic idiopathic diarrhea have normal rates of fluid and electrolyte absorption when measured under perfusion conditions during which motility effects are neutralized, suggesting that motility must be playing a role in the pathogenesis of their diarrhea under ordinary circumstances.⁷ Accelerated transit is also a major factor in diarrhea associated with some endocrine disorders (e.g., hyperthyroidism, carcinoid syndrome, or other peptide-secreting tumors).

Poorly absorbed substances that can induce osmotic diarrhea include lactose in lactase-deficient individuals. Osmotic diarrhea can also occur with ingestion of sufficient quantities of other poorly absorbed carbohydrates (e.g., fructose and the sugar alcohols mannitol and sorbitol) and ions such as magnesium, phosphate, and sulfate.

Mechanisms that reduce the overall efficiency of absorption may coexist in various disease states. For instance, in celiac disease, loss of intestinal villi results in reduced salt and water absorption, as well as reduced nutrient absorption. Thus, increased stool water in this condition results from both a reduced rate of electrolyte absorption and the increased intraluminal osmotic activity of poorly absorbed substances. Transit may accelerate in many diarrheal states because of stimulation of peristalsis by increased intraluminal volumes.

Fecal Osmotic Gap

Figure 1. Fecal Electrolyte Concentrations

As the rate of intestinal salt absorption decreases, the concentration of salts in stool rises to the point where it approaches plasma osmolality (290 mOsm/kg), the osmolality that intestinal contents maintain beyond the proximal jejunum. If the rate of salt absorption is unimpaired but nutrients are malabsorbed or poorly absorbable substances are ingested, fecal salt concentrations decrease because most of the available osmotic space is occupied by the poorly absorbed substance. This is the basis for the calculation of the fecal osmotic gap [see Figure 1].⁸ In this calculation, the contribution of electrolytes to stool osmolality is estimated by doubling the concentration of sodium and potassium (the predominant cations in stool water) to account for unmeasured anions (mostly fatty acid anions, bicarbonate, or chloride). This value is then subtracted from 290 mOsm/kg (the putative osmolality of gut contents) to determine the contribution of nonelectrolytes to fecal osmolality. When electrolytes constitute most of luminal osmolality, the calculated fecal osmotic gap will be low (< 50 mOsm/kg). When poorly absorbable substances are present, the fecal osmotic gap will be large (> 100 mOsm/kg). Watery diarrhea with a low fecal osmotic gap is classified as secretory diarrhea; diarrhea with a large fecal osmotic gap is classified as osmotic diarrhea. These categories are most helpful in the evaluation of patients with chronic diarrhea.

For clinical purposes, diarrhea can be classified as either acute (duration < 4 weeks) or chronic (duration > 4 weeks). Chronic diarrhea is further categorized as watery, inflammatory, or fatty, based on the characteristics of the stools.³ This classification allows the physician to direct evaluation and management more effectively, because diarrheal diseases can be distinguished by the duration of illness and type of stools produced.

Most acute diarrheas (those lasting < 4 weeks) are caused by infections and are self-limiting. Most are caused by viruses (e.g., adenovirus, Norwalk agent, rotavirus), but some are caused by bacteria (e.g., Campylobacter, Salmonella, Shigella, Escherichia coli) and others by protozoa (e.g., Giardia lamblia, Entamoeba histolytica) [see Table 1].²

The epidemiology of acute infectious diarrhea depends on the circumstances of infection and where one contracts the infection. For example, a history of recent travel, particularly to developing countries, makes a diagnosis of traveler's diarrhea likely. Previous antibiotic use or residence in an institution where antibiotic use is common (e.g., hospitals, nursing homes) are risk factors for Clostridium difficile infection. Children in day care facilities and their contacts, persons engaging in promiscuous sexual activity, and users of illicit intravenous drugs are all at enhanced risk for contracting infectious diarrhea. Consumption of potentially contaminated food and drink is another risk factor for infectious diarrhea. With the globalization of commerce and mass processing of food, esoteric infections from overseas and large outbreaks of food-borne diarrhea have become more common.

Pathogenic infections cause diarrhea by one of four mechanisms: (1) enterotoxins that subvert the regulatory mechanisms of enterocytes, (2) cytotoxins that destroy enterocytes, (3) adherence to the mucosa by organisms (so-called enteroadherent organisms) that alter enterocyte function as a result of physical proximity to the mucosa, and (4) invasion of the mucosa by organisms that provoke an inflammatory response by the immune system.⁹ In general, patients with cytotoxin-mediated diarrheas and those with invasive organisms experience more toxicity and have more abdominal pain than patients with enterotoxin-mediated diarrhea or enteroadherent infections.

Another mechanism for acute diarrhea is ingestion of a preformed toxin.¹⁰ Several species of bacteria, such as Staphylococcus aureus, C. perfringens, and Bacillus cereus, can produce toxins that produce so-called food poisoning (i.e., vomiting and diarrhea) within 4 hours of ingestion. In such cases, the bacteria do not need to establish an intraluminal infection; ingestion of the toxin alone can produce the disease. Symptoms subside after the toxin is cleared, usually by the next day; evidence of toxicity (e.g., fever) is minimal.

Other potential causes of acute diarrhea include food allergies and medication reactions. Food allergies are rarely recognized as causes of diarrhea in adults unless the diarrhea is associated with urticaria or other allergic symptoms. Medications often produce diarrhea as a side effect; this association is typically appreciated by the patient because of the temporal relation between drug ingestion and diarrhea.

Finally, acute diarrhea may represent the initial stages of chronic diarrhea. This is significant because acute and chronic diarrhea have very different sets of causes. However, patients with chronic diarrhea often do not seek help during the initial weeks of their illness unless the diarrhea is severe or complicated by dehydration, symptomatic electrolyte disorders, or fever.

Laboratory Testing

Figure 2. Evaluation of Acute Diarrhea

Extensive laboratory testing is not necessary for most patients with acute diarrhea; it should be reserved for those with toxicity, dehydration, or persistence of diarrhea for longer than would be expected, given its probable cause [see Figure 2]. In such patients, a complete blood count should be obtained to assess for hemoconcentration, anemia, or leukocytosis. Patients with viral diarrheas typically have normal white blood cell (WBC) counts and differentials, although lymphocytosis may be seen. Invasive bacterial infections typically produce leukocytosis with many immature WBCs, but salmonellosis can induce leukopenia. Serum electrolyte, blood urea nitrogen (BUN), and serum creatitine measurements can define the metabolic impact of diarrhea, which includes hypokalemia, hyponatremia, hyperchloremic acidosis, hypomagnesemia, and prerenal azotemia.

Stool testing is of value in patients with blood in the stool, dehydrating diarrhea, prolonged diarrhea, or dysentery and in patients who present during an outbreak of diarrhea. Stool cultures are sensitive and specific, but they are expensive. Some experts recommend obtaining cultures only in patients who have leukocytes (or the leukocyte marker lactoferrin) in the stool, because the yield of pathogenic bacteria will be higher in that group.¹¹ Others dispute this.¹² Laboratories routinely test for Salmonella, Shigella, Campylobacter, and E. coli serotype O157:H7. Special cultures for tuberculosis, Yersinia, Aeromonas, or Plesiomonas may need to be requested in appropriate circumstances. Examination of stool for ova and parasites has variable utility, depending on the pretest probability of these infections. For example, such testing might be very useful in a day care worker with diarrhea, but it would be of little help in a patient with hospital-acquired diarrhea. Enzyme-linked immunosorbent assay (ELISA) stool testing for giardiasis and serologic testing for amebiasis are more accurate tests for those specific infections in most settings. Patients treated with antibiotics in the 3 months preceding onset of diarrhea and those who develop diarrhea in institutional settings should have a stool sample analyzed for C. difficile toxin.

In patients who are toxic, in those who have blood in their stools, or in those who have persistent acute diarrhea, sigmoidoscopy or colonoscopy should be considered. In most patients without rectal bleeding, sigmoidoscopy is probably adequate as an initial evaluation, because most patients with colitis will have involvement of the left side of their colons. In patients with bleeding or those with AIDS, colonoscopy is preferable because some opportunistic infections and lymphomas are seen only in the right colon.¹³ Mucosal biopsies should be performed in either case, particularly if the colon is grossly inflamed, because the pathologist can readily distinguish self-limited colitis from chronic ulcerative colitis early in the course of the disease.¹⁴ Abdominal x-rays or computed tomography should be obtained in toxic patients to confirm a diagnosis of colitis, to determine its extent, and to look for evidence of ileus or megacolon.

Because most cases of acute diarrhea are self-limited, most patients do not need specific therapy. Instead, judicious replacement of fluid and electrolyte losses is sufficient. This can be accomplished by intravenous fluids or oral rehydration solutions. Oral rehydration solutions are based on the concept that nutrient absorption accelerates the absorption of sodium and fluid by the jejunum.¹⁵ Initially, rehydration formulas used glucose as the absorbable nutrient, but recent work has shown that cereal-based oral rehydration solutions can be more efficient. Oral rehydration solution does not reduce fecal losses (it may actually increase stool output) but, instead, increases net fluid and electrolyte absorption. These solutions cannot be used if vomiting precludes ingestion; intravenous rehydration must be used in those situations. Sports drinks (e.g., Gatorade) are designed to offset fluid and electrolyte losses from sweating and do not contain sufficient sodium to replace fecal losses. Solutions that more closely approximate World Health Organization rehydration solution are now commercially available (e.g., Rehydralyte, Resol, Ricalyte).

Diet Most patients seek advice about altering their diets when suffering from diarrhea. Other than the provision of adequate water and salt, no specific instructions are needed. Some physicians routinely restrict dairy products in patients with diarrhea on the theory that these patients may have temporary lactase deficiency. This need not be done unless there is clinical evidence of lactose intolerance (e.g., exacerbation of diarrhea or flatus with ingestion of dairy products).

Antibiotics Empirical antibiotic therapy for acute diarrhea may be appropriate under certain circumstances (e.g., in travelers with diarrhea, during local outbreaks of bacterial or protozoal diarrhea, or in patients who are frail or toxic). However, experts discourage the routine use of empirical antibiotic therapy because of its lack of demonstrable efficacy in many infections and because of concerns about precipitating complications, such as hemolytic-uremic syndrome in patients with E. coli serotype O157:H7.¹⁶ A recent meta-analysis suggests that this latter point is not supported by the literature.¹⁷ When indicated, fluoroquinolones or trimethoprim-sulfamethoxazole are commonly used for empirical therapy.

Nonspecific antidiarrheal agents, such as opiates, can reduce stool frequency and stool weight, and they may reduce associated symptoms, such as abdominal cramps.¹⁸ Concerns about slowing the clearance of pathogens from the intestine by reducing peristalsis largely have not been borne out. Intraluminal agents, such as bismuth subsalicylate (Pepto-Bismol) and adsorbants (e.g., kaolin) are also sometimes used [see Table 2].

Campylobacter A frequent cause of acute ileocolitis in the United States,¹⁹Campylobacter infection is usually acquired by eating undercooked chicken; it has an incubation period of up to 1 week. Ulceration of the colonic mucosa and bloody diarrhea may occur with this infection. Antibiotics, such as erythromycin or perhaps a fluoroquinolone (although fluoroquinolone resistance is increasingly reported with Campylobacter species), shorten the course of the illness if given within the first few days of symptoms.

Salmonella enteritidis and S. cholerasuis These nontyphoidal Salmonella species are spread via contaminated food or water and cause acute gastroenteritis, ileocolitis, or colitis characterized by watery diarrhea.²⁰ Antibiotic therapy with a fluoroquinolone, trimethoprim-sulfamethoxazole, or ampicillin should be reserved for severely ill patients or patients with compromised immunity (e.g., infants, the elderly, pregnant women, and AIDS patients).²¹

Salmonella typhi This organism causes typhoid fever, a form of enteric fever.²² The propensity of S. typhi to produce bacteremia distinguishes it from other enteric pathogens. When the infection is limited to the intestine of an otherwise healthy person, no specific therapy is indicated, because antibiotics may paradoxically prolong excretion of the organism and increase relapses. When the infection becomes systemic and the patient is very ill, therapy is necessary, especially if the organism produces a metastatic endovascular infection. Fluoroquinolones are most often used. The diagnosis of a carrier state is is made when stool cultures are found to be positive for a period longer than 1 year.

Shigella Shigella species are invasive organisms, but they also produce an enterotoxin that reduces water and electrolyte absorption.²³ Shigellosis commonly causes a watery diarrhea initially (this watery diarrhea is most likely related to the Shigella enterotoxin). This is followed by a bloody diarrhea, which results from colitis produced by invasion of the colonic mucosa. Treatment with a fluoroquinolone is recommended for most patients with shigellosis.

E. coli serotype O157:H7 This organism has become a common cause of food-borne infection in the United States.²⁴ It produces toxins similar to those produced by Shigella.²⁵ Infection with this organism causes a hemorrhagic segmental colitis. The disease often occurs in large outbreaks from contamination of widely distributed foods, such as hamburger meat. Patients can become quite ill; hemolytic-uremic syndrome is a well-recognized complication. Antibiotics do not seem to improve the course of the illness and may be associated with the development of hemolytic-uremic syndrome in children, although this is controversial.^16,17

Clostridium difficile C. difficile has become the most common cause of nosocomial diarrhea in many institutions.²⁶ In nonhospitalized adults, carriage rates for this organism are low, but it is spread easily from person to person by spores. Suppression of the normal bacterial flora of the colon by antibiotic therapy can result in the overgrowth of C. difficile, if it is present. In institutional settings, the organism can be distributed efficiently to a large pool of susceptible patients by health care workers who do not wash their hands. The disease produced can range from a simple, self-limited diarrhea to a fulminant colitis.

Treatment for 2 weeks with metronidazole, 250 mg four times daily, or vancomycin, 125 to 500 mg four times daily, is effective against C. difficile. Relapses occur in up to 25% of patients, however, probably because of residual spores. Ingestion of probiotic bacteria or the nonpathogenic yeast Saccharomyces boulardii may reduce relapse rates. In most instances of relapse, longer periods of antibiotic therapy are indicated.

Other nosocomial diarrheas Noninfectious causes of nosocomial diarrhea include medications—particularly elixirs that contain sorbitol or mannitol as noncaloric sweeteners and cancer chemotherapeutic drugs—enteral feeding, and paradoxical diarrhea in patients with fecal impaction. Infections with organisms other than C. difficile also occur in institutions, particularly extended-stay facilities. An important cohort of hospital patients that may develop infectious diarrhea are those who are immunocompromised by diseases such as AIDS, or by drugs used to prevent transplant rejection or to treat inflammatory diseases. These patients are often infected with opportunistic pathogens, including viruses (e.g., cytomegalovirus, herpesvirus), bacteria (e.g., Mycobacterium avium complex), and parasites (e.g., Cryptosporidium species, Strongyloides species).²⁷ In addition, bone marrow transplant recipients may develop acute diarrhea from graft versus host disease.

Parasites Acute diarrhea in noninstitutionalized patients can be from parasitic infection.²⁸ The likelihood of parasitic disease as a cause of acute diarrhea is profoundly influenced by geography and epidemiologic features. For example, giardiasis is a common infection in some areas but not others, probably because of variability in the effectiveness of water treatment. Ingestion of as few as a dozen cysts of G. lamblia may result in an infection. This accounts for the frequency of person-to-person transmission of this disease. ELISA for Giardia antigen is superior to microscopic inspection of stool (so-called ova and parasites testing) for the detection of giardiasis. Therapy with metronidazole is effective in most patients, but reinfection can occur.

Amebiasis is also common in some areas. Persons with amebiasis may be asymptomatic or extremely ill from spread of infection to other organs, such as the liver. Diagnosis is typically made by microscopic examination of fresh stools, but ELISA shows promise in distinguishing the pathogenic species E. histolytica from nonpathogenic amebae. The colonoscopic appearance of amebiasis often is distinctive, and the organism can be identified in colonic biopsy specimens.

Other parasites that may cause acute diarrhea include Cryptosporidium, Isospora, Cyclospora, and Strongyloides species, as well as Trichuris trichiura (whipworm). Special tests that may be necessary to identify these parasites include concentration of stool samples, acid-fast stains of stool, and mucosal biopsy. If these organisms are suspected, consultation with the laboratory staff allows use of the proper diagnostic tests.

In contrast to acute diarrhea, in which infection is the overwhelmingly likely cause of illness, chronic diarrhea has an extensive and daunting list of possible causes [see Table 3].³ The simplest approach to diagnosis is to classify chronic diarrhea by the characteristics of the stools. Three categories of chronic diarrhea are recognized: watery, inflammatory, and fatty. Watery diarrheas can be subdivided further into osmotic and secretory diarrheas on the basis of stool analysis.

Osmotic diarrheas result from the ingestion of an osmotically active, poorly absorbable substance that necessitates the retention of water within the gut lumen to maintain isosmolar conditions.²⁹ In practical terms, osmotic diarrheas are caused by ingestion of osmotic laxatives (magnesium, phosphate, and sulfate salts; sugar analogs, such as lactulose; sugar alcohols, such as mannitol or sorbitol; and polyethylene glycol) and carbohydrate malabsorption. The ingestion of osmotic laxatives may be purposeful (i.e., laxative abuse) or accidental, as when excess magnesium is ingested as part of a mineral supplement or multivitamin tablet. Carbohydrate malabsorption is most often the result of acquired lactase deficiency (a normal development after adolescence) or mucosal disease, such as celiac sprue, that interferes with nutrient absorption.

Secretory diarrhea has a much larger differential diagnosis than osmotic diarrhea [see Table 3].

Congenital chloridorrhea Rarely, congenital absence of a transporter mechanism results in diarrhea. This is the case in congenital chloridorrhea, in which the chloride-bicarbonate exchanger in the ileum is not active.³⁰ Under such conditions, chloride becomes poorly absorbable in the distal bowel and obligates intraluminal water retention.

Chronic infections Some bacterial infections can last long enough to produce chronic secretory diarrhea.³¹ These include Aeromonas and Pleisiomonas species, enteropathogenic E. coli, C. difficile, M. tuberculosis, and Yersinia enterocolitica. A special situation is small bowel bacterial overgrowth syndrome, in which structural problems, such as jejunal diverticulosis, or motility problems, such as those seen in scleroderma, result in proliferation of bacteria in the jejunum.³² Although this bacterial overgrowth disrupts digestive processes and may produce fatty diarrhea, it also may reduce water and salt absorption, producing secretory diarrhea. Infection with parasites, such as G. lamblia, E. histolytica, and Cryptosporidium, also can produce chronic diarrhea.³³

Inflammatory bowel disease Typically, inflammatory bowel diseases (e.g., ulcerative colitis, Crohn disease) produce inflammatory diarrhea, with blood and pus in the stool (see below). Watery diarrhea can occur, however, especially when the distal colon is not involved. One form of inflammatory bowel disease that typically produces a watery diarrhea is microscopic colitis syndrome (lymphocytic colitis and collagenous colitis), in which the mucosa is inflamed but not ulcerated.³⁴ Although diverticulitis usually presents as an acute illness, some patients who have smoldering diverticulitis with relatively low-grade inflammation (and, in some cases, low-grade obstruction) will present with chronic secretory diarrhea, which is probably mediated by inflammation-linked cytokines. Vasculitis and systemic inflammatory diseases may also be associated with secretory diarrhea.

Drugs Drug therapy is a key cause of secretory diarrhea. Many drugs cause diarrhea as a side effect. These include antibiotics; cardiovascular agents, such as beta-adrenergic antagonists, digitalis, and quinidine; cancer chemotherapy; nonsteroidal anti-inflammatory drugs (NSAIDs); and colchicine. Thus, in taking the history of a patient with chronic diarrhea, it is critical to formulate a detailed drug list, including over-the-counter and alternative medications. A special category of drug-induced secretory diarrhea is surreptitious ingestion of stimulant laxatives.

Other causes Disordered motility or regulation can produce secretory diarrhea.³⁵ Secretory diarrhea associated with disordered motility can occur in patients who have undergone vagotomy or sympathectomy, patients with autonomic neuropathy from diabetes or amyloidosis, and probably patients with irritable bowel syndrome.³⁶ In the United States, irritable bowel syndrome is the most commonly diagnosed cause of chronic diarrhea. This diagnosis is often incorrect, however, and may delay appropriate treatment.

Malabsorption of bile acid in the ileum occurs in many diarrheal diseases as a result of ileal disease or resection and may be secondary to other processes, such as vagotomy, cholecystectomy, or rapid transit past the ileum. In a relatively small number of patients, idiopathic bile acid malabsorption is a cause of diarrhea.³⁷

Endocrine causes of secretory diarrhea include hyperthyroidism, Addison disease, and a group of rare tumors of the endocrine cells of the gut, including gastrinomas, carcinoid tumors, vasoactive intestinal polypeptide-secreting tumors, somatostatinomas, and medullary carcinoma of the thyroid.³⁸ These tumors produce peptides and other mediators that affect intestinal mucosal and muscle function and thereby lead to diarrhea. In most cases, rapid intestinal transit seems to be the major mechanism producing diarrhea, although this remains controversial.³⁹

Other tumors that produce secretory diarrhea include colon cancer (through an uncertain mechanism), villous adenoma of the rectum, lymphoma, and mastocytosis. Mastocytosis (and probably some lymphomas) produce diarrhea through the release of histamine or other mediators that affect gut function. Infiltration of the mucosa by mast cells or lymphoid cells also may play a role in some cases.

Secretory diarrhea can also be idiopathic.⁴⁰ Idiopathic secretory diarrhea occurs in both sporadic and epidemic forms and may be caused by an as-yet unidentified infection.

An accurate medical history is even more important for patients with chronic diarrhea than for those with acute diarrhea. In addition to all the issues that should be discussed with patients who have acute diarrhea [see Acute Diarrhea, above], the history of patients with chronic diarrhea should include long-term trends in body weight, current appetite and food intake, a review of previous medical problems and surgeries, potential secondary gains from illness, previous evaluations and treatments of diarrhea, and a detailed review of systems to look for clues to systemic illnesses [see Table 4].

A principal diagnostic distinction in chronic diarrhea is between diarrhea associated with irritable bowel syndrome and diarrhea associated with other functional or organic problems. Irritable bowel syndrome is characterized by abdominal pain associated with defecation and an altered bowel habit.⁴¹ Variable stool consistency and intermittent constipation are common. Painless diarrhea was once considered a type of irritable bowel syndrome but is no longer; other causes of diarrhea should be sought in such cases.

As in acute diarrhea, routine laboratory testing is indicated to help determine the severity of chronic diarrhea [see Acute Diarrhea, above]. Unlike acute diarrhea, in which stool analysis is typically not used, stool analysis plays a key role in the assessment of chronic diarrhea by allowing categorization of the type of diarrhea, thereby limiting the number of conditions to be considered.³ The stool analysis can be obtained through either a random sample or a timed collection. The value of a timed collection is that it allows the physician to quantitate stool output accurately. However, stool analysis obtained through a random sample still can provide many diagnostic clues.

Stool characteristics to measure include sodium and potassium concentrations, osmolality, and pH. Measurement of stool electrolyte concentrations allows calculation of the fecal osmotic gap [see Fecal Osmotic Gap, above]. This can be used to determine whether watery diarrhea is osmotic or secretory. Measurement of actual stool osmolality is of value only in detecting samples that have been contaminated (unintentionally or deliberately) with water or dilute urine and therefore have an osmolality less than 290 mOsm/kg. Stool osmolality rises rapidly in vitro because of bacterial fermentation, so the actual measurement should not be used to calculate the fecal osmotic gap. The pH of stool water can indicate whether or not carbohydrate malabsorption is present. Carbohydrates (or sugar alcohols) that are not absorbed in the small bowel and so reach the bacterial flora of the colon are fermented into short-chain fatty acids that reduce fecal pH, usually to less than 6. Thus, acid stools suggest carbohydrate malabsorption.⁸

Other helpful tests include fecal occult blood testing and examination of stool for leukocytes (or a surrogate chemical test, such as fecal lactoferrin concentration), which can be used to identify an inflammatory diarrhea. Fatty diarrhea can be identified by measurement of stool fat, although careful interpretation of the results is sometimes necessary [see Steatorrhea, below].

When appropriate, a laxative screen should be obtained. Measurement of laxatives by chemical or chromatographic methods can detect surreptitious laxative ingestion.

Completion of the stool analysis allows the clinician to characterize chronic diarrhea as watery (whether secretory or osmotic), inflammatory, or fatty. The subsequent evaluation depends on this categorization.

Evaluation of Watery Secretory Diarrhea

Figure 3. Evaluation of Chronic Secretory Diarrhea

Secretory diarrhea has a broad differential diagnosis, so a thorough evaluation is needed [see Figure 3].

Stool testing Infection should be excluded by stool culture for bacteria, stool assay for C. difficile toxin, and other stool tests for parasites, including ELISA for giardiasis. Biopsies of the small bowel or colon may be necessary to find the pathogens, especially in patients with AIDS or other immunodeficiencies. Quantitative culture of small bowel aspirate is the best test for detecting small bowel bacterial overgrowth.

Imaging and endoscopic tests Structural diseases (e.g., short bowel syndrome or fistula, mucosal diseases, inflammatory bowel disease, and tumors) should be sought by radiographic and endoscopic testing. Small bowel radiography remains important in these patients. CT scans can detect small bowel and colonic disease, as well as problems extrinsic to the gut that may cause diarrhea, such as endocrine tumors. Inspection of the colonic mucosa by colonoscopy or sigmoidoscopy is essential in patients with secretory diarrhea, both to evaluate for gross changes and to obtain biopsy specimens to look for evidence of microscopic colitis syndrome. Biopsies should be performed even if the gross appearance of the colon is normal, because of the prevalence of microscopic colitis syndrome in patients with chronic watery diarrhea (see below). A long endoscope that can reach the jejunum to obtain biopsy specimens and aspirates is a valuable adjunct when other studies are unrevealing. The role of capsule endoscopy in the evaluation of patients with chronic diarrhea is uncertain, and it does not allow for biopsy of abnormalities that are seen.

Serum peptide measurement Because diarrheagenic endocrine tumors are very rare, the measurement of serum peptides (e.g., gastrin, vasoactive intestinal polypeptide, calcitonin, and glucagon) or urinary secretagogue metabolites (e.g., 5-hydroxyindoleacetic acid or metanephrine) should be restricted to patients with symptoms consistent with tumor syndromes or those in whom a diagnosis remains elusive after initial testing.⁴² More common endocrine problems, such as diabetes, hyperthyroidism, or Addison disease, should be excluded with appropriate blood tests.

Bile acid absorption measurement Ileal resection or ileal disease can result in the escape of sufficient bile acid into the colon to increase luminal bile acid concentrations above 3 to 5 mmol. At those concentrations, bile acids reduce colonic mucosal water and electrolyte absorption; alternatively, they may stimulate secretion, resulting in increased stool water. In most circumstances, bile acid malabsorption can be inferred from a history of ileal resection or disease. More controversial is the concept that bile acid malabsorption occurring in the absence of ileal resection or obvious ileal disease is responsible for idiopathic secretory diarrhea.⁴³ Although bile acid malabsorption can be documented in many of these patients, administration of bile acid-binding resins does not always mitigate diarrhea in those patients, casting doubts on bile acid malabsorption as a cause of their diarrhea.³⁷ Therefore, in patients with secretory diarrhea that appears to be idiopathic, it is more practical to give a therapeutic trial of bile acid-binding resins than to measure bile acid malabsorption directly.

Evaluation of Watery Osmotic Diarrhea

Figure 4. Evaluation of Chronic Osmotic Diarrhea

The differential diagnosis of osmotic diarrhea is more limited than that of secretory diarrhea, so the evaluation is simpler [see Figure 4]. If stool water has low electrolyte concentrations (and therefore a high fecal osmotic gap), some other substance is taking up the osmotic space and is holding water in the lumen. In practice, this substance is usually ingested magnesium or malabsorbed carbohydrates.

Fecal magnesium Magnesium can be measured accurately in stool water. Excretion of more than 15 mmol (30 mEq) daily or a concentration of greater than 45 mmol/L (90 mEq/L) in a random stool sample strongly suggests magnesium-induced diarrhea.⁴⁴ This may be intentional (surreptitious laxative ingestion) or accidental (use of magnesium-containing antacids or mineral supplements).

Carbohydrate absorption Carbohydrate malabsorption can occur from ingestion of poorly absorbable carbohydrates, such as lactose in a person with lactase deficiency, or from reduced carbohydrate absorption as a result of small bowel mucosal disease. In addition to lactose, common causes of osmotic diarrhea include excessive ingestion of fructose (often used as a sweetener in commercial food products), ingestion of poorly absorbed sugar alcohols (such as mannitol and sorbitol, which are used as low-calorie sweeteners), and use of inhibitors of carbohydrate absorption, such as acarbose. Because malabsorbed carbohydrate is rapidly fermented by colonic bacteria, gas and bloating are frequent symptoms in these patients. Diagnosis is made on the basis of a finding of low stool pH (typically less than 6) and a thorough dietary history.

Patients with WBCs or blood in the stool are classified as having inflammatory diarrhea. Causes may include inflammatory bowel disease, infections, ischemia, radiation enteritis, and neoplasia [see Table 3]. Sometimes these conditions produce a watery, secretory diarrhea without blood or pus in the stool, so they must be considered in the evaluation of patients with nonbloody diarrhea.

Imaging and endoscopic tests

Figure 5. Evaluation of Chronic Inflammatory Diarrhea

Evaluation should start with radiographic and endoscopic tests to look for structural problems [see Figure 5]. Sigmoidoscopy or colonoscopy should be considered first, because colitis is a common cause of inflammatory diarrhea. Biopsies should be performed to properly categorize colitis. CT has proved useful in many of these patients because of its ability to visualize inflammatory changes in the small bowel and colon and to identify complications of inflammation, such as abscess.

Infections that may produce chronic diarrhea, such as C. difficile, cytomegalovirus, amebiasis, and tuberculosis, need to be excluded by culture, biopsy, or serologic testing. It is important to realize that infection may complicate the courses of established disorders, such as ulcerative colitis or Crohn disease. Patients with AIDS need an especially careful search for opportunistic infections (see below).

Steatorrhea

Figure 6. Evaluation of Chronic Fatty Diarrhea

Excessive fat in the stools implies a problem with fat solubilization, digestion, or absorption in the small intestine [see Figure 6]. Steatorrhea is usually defined as stool fat output of more than 7 g/24 hr or more than 9% of daily intake. These criteria may not be valid in patients with diarrhea, however, because voluminous stools per se can increase fat excretion. In one study, artificially induced diarrhea produced mild steatorrhea of up to 14 g/24 hr in 35% of normal persons.⁴⁵ Thus, in patients with diarrhea, fecal fat excretion of 7 to 14 g/24 hr has a low specificity for defective fat absorption. The threshold for diagnosing steatorrhea also should be corrected for fat intake, because some patients with diarrhea have anorexia and some patients with steatorrhea have hyperphagia. When possible, fat intake should be estimated from diet diaries maintained during the collection period. Finally, measurement of fat excretion can be compromised by ingestion of poorly absorbed fat substitutes, such as olestra.

Qualitative estimation of fat excretion by Sudan stain of a fecal smear can be used when a timed collection or quantitative analysis is not possible. Semiquantitative methods employing assessment of the number and size of fat globules correlate well with quantitative analysis of fat excretion.

The fecal fat concentration may provide a clue to the etiology of steatorrhea. The major causes of steatorrhea are mucosal diseases (e.g., celiac disease), pancreatic exocrine insufficiency (e.g., chronic pancreatitis), and lack of bile acids (e.g., advanced biliary cirrhosis). Mucosal diseases are often associated with reduced fluid and electrolyte absorption; as a result, fat is diluted by unabsorbed water. Furthermore, in mucosal disease, fat still can be digested to fatty acids, which can inhibit water absorption in the colon. In contrast, diseases that alter fat solubilization or digestion typically do not alter mucosal water and electrolyte absorption; as a result, unabsorbed fat is disbursed in a smaller stool volume. Fecal fat concentrations of greater than 9.5 g/100 g strongly suggest pancreatic or biliary steatorrhea.⁴⁶

Imaging and endoscopic tests If the cause of steatorrhea is not obvious from the patient's history and the results of fecal fat assessment, the next step is evaluation of the absorptive surface of the small intestine by endoscopic, histologic, and radiographic tests. During endoscopy, small bowel biopsy specimens should be obtained for histologic analysis, and small bowel contents should be aspirated for quantitative culture to assess for small bowel bacterial overgrowth. Indirect tests, such as measurement of antigliadin, endomysial (antireticulin) antibodies, or tissue transglutaminase antibodies for the diagnosis of celiac disease or breath tests for bacterial overgrowth, have not displaced endoscopic testing as the gold standard for diagnosis of these conditions. these tests may, however, be useful in some cases. Small bowel radiography and CT are valuable adjuncts for structural assessment in these patients.

If the absorptive surface is normal, attention should shift to luminal problems with fat solubilization or digestion. Testing for pancreatic exocrine insufficiency is rarely done, because of unwillingness to use duodenal intubation tests. Indirect tests, such as the bentiromide test and measurement of stool chymotrypsin activity, have limited sensitivity and specificity. The best test for pancreatic exocrine insufficiency may be a therapeutic trial of pancreatic enzyme supplementation. If this is done, a large dose of enzymes should be administered and fat excretion should be measured to assess the response to therapy. Likewise, testing for the adequacy of the solubilization of fat by bile salts is rarely done. If necessary, duodenal bile salt concentration can be measured.

Nonspecific therapy is used for patients with chronic diarrhea in three situations: (1) as a temporizing or initial therapy before diagnostic testing, (2) after diagnostic testing has failed to result in a diagnosis, and (3) when a diagnosis has been made but no specific treatment is available or specific treatment fails.¹⁸

Antibiotic therapy Antibiotics are less useful in chronic diarrhea than in acute diarrhea because bacterial infection is less likely to be the cause of chronic diarrhea. Nevertheless, many clinicians try an empirical course of metronidazole or a fluoroquinolone before starting an extensive evaluation.

Symptomatic therapy with antidiarrheal drugs is often required in patients with chronic diarrhea. Loperamide or diphenoxylate with atropine can be tried initially [see Table 2]. In patients with chronic diarrhea, routine dosing (e.g., two tablets before each meal or at bedtime) is more effective than as-needed dosing after passing loose stools. More potent opiates, such as codeine, opium, or morphine, are underutilized in patients who do not respond to loperamide or diphenoxylate with atropine. Although these are controlled substances because of the possibility of abuse, abuse is unlikely in closely monitored patients with chronic diarrhea. Dosing should be started at a low level (e.g., codeine, 30 mg q.i.d.; deodorized tincture of opium, 3 drops q.i.d.; or morphine, 2 mg q.i.d.) and titrated up gradually to an effective dose. Stool-modifying agents, such as psyllium, can alter stool consistency but do not reduce stool weight. They may be of special help in patients with coexisting fecal incontinence.

Irritable bowel syndrome and functional diarrhea Patients with chronic diarrhea in whom no other etiology is established are commonly diagnosed with irritable bowel syndrome or functional diarrhea. Irritable bowel syndrome is characterized chiefly by abdominal pain that is associated with altered bowel function, including constipation, diarrhea, or alternating diarrhea and constipation.⁴¹ A diagnosis of functional diarrhea is made when patients do not have prominent abdominal pain and have no evidence of other specific causes for diarrhea. Obviously, these diagnoses are reliable only if a thorough evaluation has been done to exclude other causes of diarrhea. For example, most cases of diarrhea from malabsorption of bile acid or carbohydrates are characterized as functional diarrhea or irritable bowel syndrome because specific testing for those disorders is not done. Thus, careful consideration of alternative diagnoses should precede a diagnosis of irritable bowel syndrome or functional diarrhea in patients with chronic diarrhea.

Nevertheless, there are certain clues to the diagnosis of irritable bowel syndrome or functional diarrhea that should be sought by the physician. Features that suggest a diagnosis of irritable bowel syndrome include a long history of diarrhea, dating back to adolescence or young adulthood; passage of mucus; and exacerbation of symptoms with stress. Historical points that argue against irritable bowel syndrome include recent onset of diarrhea, especially in older patients; nocturnal diarrhea; weight loss; blood in stools; voluminous stools (> 400 g/24 hr); and blood tests indicating anemia, leukocytosis, or low serum albumin concentration or a high erythrocyte sedimentation rate.

The treatment of irritable bowel syndrome is discussed in detail elsewhere [see 4:XIV Gastrointestinal Motility Disorders].

Microscopic colitis syndrome This disorder, which subsumes the diagnoses of lymphocytic colitis and collagenous colitis, is a frequent cause of chronic diarrhea.³⁴ Microscopic colitis syndrome is characterized by chronic watery diarrhea and microscopic evidence of mucosal inflammation in the presence of normal gross colonoscopic findings. Histologic findings in both lymphocytic colitis and collagenous colitis include intraepithelial lymphocytic infiltration and chronic inflammation in the lamina propria without crypt destruction. Collagenous colitis and lymphocytic colitis are distinguished by the presence or absence of a thickened subepithelial collagen layer.

The cause of microscopic colitis syndrome is uncertain. It is associated with many autoimmune disorders and immunologically mediated diseases, such as celiac disease, which suggests that immune dysregulation is important. Bacterial antigens within the colonic lumen may play a role as well. NSAIDs have been implicated in some reports.

Women are more likely than men to have collagenous colitis; lymphocytic colitis occurs equally in men and women. Diarrhea is of moderate severity (typically, 500 to 1,000 g/24 hr) and is characteristically secretory in nature because it results from the failure of the colonic mucosa to absorb water and salt. Diagnosis is made by obtaining biopsy material from normal-appearing mucosa during sigmoidoscopy or colonoscopy.

Treatment options include bismuth subsalicylate, budesonide, 5-aminosalicylate drugs, prednisone, and azathioprine.³⁴ Bile acid-binding drugs also have been reported to be successful in reducing diarrhea.⁴⁷ Microscopic colitis can have a remitting and relapsing course, and symptomatic therapy with opiate antidiarrheal drugs may be all that is needed. There is no evidence that microscopic colitis is a risk factor for colon carcinoma; no surveillance program is currently recommended.

Laxative abuse Although rarely suspected, laxative abuse occurs regularly in four groups of patients: (1) those with anorexia or bulimia, (2) those who obtain secondary gain from illness (e.g., disability payments or attention from relatives), (3) those with Munchausen syndrome, and (4) those who are dependent on others for health care and who are poisoned by their caregivers (caregivers who do this are usually motivated by the desire to demonstrate their devotion to the patients).⁴⁸ Physicians need to consider surreptitious laxative abuse in patients who confound diagnosis and who are in one of those categories.

Detection of laxative abuse requires a high index of suspicion. Clues include the presence of hypokalemia in a patient who is able to eat (suggesting stimulant laxative abuse or concurrent ingestion of diuretics), melanosis coli (brownish pigmentation in the colonic mucosa resulting from ingestion of anthraquinone laxatives) in a patient being evaluated for chronic diarrhea, or a large fecal osmotic gap (seen with magnesium ingestion). Most laxatives can be detected in stool water by chemical techniques. Adulteration of stool by added water or hypotonic urine is evidenced by low stool osmolality (< 250 mOsm/kg). The addition of hypertonic urine is evidenced by impossibly high stool osmolality (> 600 mOsm/kg) and the presence of a negative fecal osmotic gap resulting from high urine sodium or potassium concentrations. Negative fecal osmotic gaps may also be calculated in patients ingesting laxatives containing phosphate or sulfate.

Before confronting patients with the diagnosis of laxative abuse, testing should be confirmed on another stool specimen, and appropriate psychiatric consultation should be available, because some of these patients become suicidal when confronted, and all of them need counseling. In cases of laxative poisoning by a caregiver, legal proceedings need to be instituted to separate the patient from the caregiver. Outcome studies in laxative abuse patients are few. One study suggested that nearly half of the patients sought further medical attention elsewhere for chronic diarrhea.⁴⁹

Postsurgical diarrhea Diarrhea can occur after several different kinds of operations. Peptic ulcer surgery is less common than it used to be, but new kinds of gastric operations, such as gastric bypass for obesity, produce similar complications. Dumping syndrome is the term used to describe a condition characterized by postprandial flushing, hypotension, diarrhea, and hypoglycemia.⁵⁰ This syndrome results from unregulated gastric emptying, osmotic shifts of fluid into the gut, and the rapid release of peptide hormones from the small intestine. Dumping syndrome can occur after vagotomy (intentional or accidental), pyloroplasty, gastrojejunostomy, and gastric resection. It can be treated with dietary modifications, antidiarrheal drugs [see Table 2], and injection of the somatostatin analog octreotide. Gastric surgery may also predispose patients to bacterial overgrowth in the small intestine, abnormally rapid intestinal transit, bile acid malabsorption, and pancreatic exocrine insufficiency from inadequate stimulation of the pancreas.

Bowel resection can result in loss of surface area sufficient to impair absorption of nutrients or water and salt. Lesser degrees of resection can result in diarrhea if an area of specialized function is removed.⁵¹ For example, resection of the terminal ileum and right colon reduces bile acid absorption and the ability to absorb sodium against a large electrochemical gradient; these defects cannot be overcome by other areas of the intestine. With time, intestinal adaptation can overcome impaired electrolyte absorption, but intestinal adaptation cannot reverse loss of these specialized functions.

Ileostomy diarrhea is said to occur when stoma output exceeds 1,000 ml/24 hr. It may be caused by loss of absorptive surface area if a substantial length of bowel has been resected; it may also result from stomal stenosis, partial bowel obstruction, bacterial overgrowth, recurrent bowel disease, medications, or intraperitoneal infection.⁵² A special situation occurs in patients with ulcerative colitis who have had an ileoanal anastomosis with creation of an ileal reservoir pouch. These patients may develop inflammation of the pouch (so-called pouchitis) from bacterial overgrowth or recurrent inflammatory bowel disease.⁵³ Pouchitis can be treated with antibiotics such as metronidazole, anti-inflammatory drugs such as mesalamine, or ingestion of probiotic bacteria. Ordinary ileostomy diarrhea can be treated successfully with antidiarrheal opiate drugs.

Postcholecystectomy diarrhea occurs in as many as 20% of patients. It may be delayed in onset, and it is rarely severe. Diarrhea may occur as a result of ileal bile acid malabsorption at night, when the migrating motor complex may sweep bile acid past the absorptive sites in the terminal ileum, but some cases may have other causes.⁵⁴ Postcholecystectomy diarrhea is best treated with bile acid-binding agents given at bedtime. Opiate antidiarrheal drugs may be needed in refractory cases.

Diabetic diarrhea Up to 30% of patients with long-standing diabetes mellitus may experience chronic diarrhea.⁵⁵ This diarrhea has been attributed to autonomic neuropathy and dysregulation of motility, but definite evidence of neuropathy is not always evident. If steatorrhea is present, three conditions that occur with increased prevalence in diabetic patients should be considered: (1) small bowel bacterial overgrowth, (2) pancreatic exocrine insufficiency, and (3) celiac disease. Other causes that need to be considered include medications, such as acarbose, and ingestion of so-called dietetic foods containing sugar alcohols (e.g., sorbitol or mannitol).

When watery diarrhea is present, treatment with clonidine, an alpha₂-adrenergic agonist drug, may have special value. When clonidine cannot be tolerated because of its hypotensive effect or when it does not work, opiate antidiarrheal drugs may be effective. Fecal incontinence from diabetic sensorimotor neuropathy may complicate diarrhea; this form of diarrhea is important to diagnose, because therapies to mitigate incontinence, such as biofeedback training, may have a dramatic effect on quality of life.⁵⁶

Diarrhea in patients with AIDS Diarrhea in AIDS patients is likely to result from opportunistic infections or lymphoma. A careful search for the cause of diarrhea can lead to targeted therapy that may cure the diarrhea.⁵⁷ Colonoscopy is preferable to sigmoidoscopy because it allows visualization and biopsy of the right colon and ileum, which are often the sites of infection. It is possible that HIV-1 may directly produce diarrhea (so-called AIDS enteropathy), but in most cases, a specific infection can be identified.

Idiopathic secretory diarrhea The diagnosis of idiopathic secretory diarrhea can be made after an exhaustive evaluation fails to reveal a cause of chronic secretory diarrhea. This condition often begins suddenly in previously normal persons and is distinguished from acute secretory diarrhea by its persistence beyond 4 weeks. It occurs in two forms, epidemic and sporadic.

Epidemic idiopathic secretory diarrhea occurs in outbreaks that are seemingly related to contaminated food or water.⁵⁸ The initial report of this disorder described an epidemic of chronic diarrhea in Brainerd, Minnesota, and the condition has consequently become known as Brainerd diarrhea. Several outbreaks have been described in detail since the initial epidemic, and although the epidemiology suggests an infectious cause, no organism has been isolated.

Sporadic idiopathic secretory diarrhea affects patients in an identical fashion as the epidemic form, but it does not seem to be acquired easily by family members or others.⁴⁰ Many patients describe a history of travel to local lakes or recreational sites, but they are the only members of their parties that become ill.

Both forms of idiopathic secretory diarrhea begin abruptly and reach maximum intensity shortly thereafter. Fever is unusual. Weight loss of up to 20 lb characteristically occurs in the first few months of the illness but does not become progressive thereafter. Empirical trials of antibiotics and bile acid-binding drugs are ineffective, but nonspecific opiate antidiarrheal drugs provide some relief. Idiopathic secretory diarrhea is self-limited and usually disappears within 2 years of onset. The resolution of diarrhea is gradual, occurring over 2 to 3 months.

Diarrhea of obscure origin This condition is said to be present when chronic diarrhea has evaded diagnosis in spite of an evaluation for structural problems. Patients are often referred to centers interested in diarrheal diseases, where a specific cause for their diarrhea is often identified. Common diagnoses in these patients include fecal incontinence, drug-induced diarrhea, surreptitious laxative ingestion, microscopic colitis syndrome, bile acid-induced diarrhea, pancreatic exocrine insufficiency, carbohydrate malabsorption, sporadic chronic idiopathic secretory diarrhea, and, rarely, endocrine tumors. Most of these conditions can be recognized through a careful history, an appropriate index of suspicion, proper testing, or a well-conducted therapeutic trial. Failure to make a diagnosis is usually the result of not thinking through the differential diagnosis of chronic diarrhea and not appreciating the evidence at hand.