Diarrhea is best defined as passage of loose stools often with more frequent bowel movements. For clinical purposes, the Bristol Stool Form Scale works well to distinguish stool form and to identify loose stools. Laboratory testing of stool consistency has lagged behind. Acute diarrhea is likely to be due to infection and to be self-limited. As diarrhea becomes chronic, it is less likely to be due to infection; duration of 1 month seems to work well as a cut-off for chronic diarrhea, but detailed scientific knowledge is missing about the utility of this definition. In addition to duration of diarrhea, classifications by presenting scenario, by pathophysiology, and by stool characteristics (e.g. watery, fatty, or inflammatory) may help the canny clinician refine the differential diagnosis of chronic diarrhea. In this regard, a careful history remains the essential part of the evaluation of a patient with diarrhea. Imaging the intestine with endoscopy and radiographic techniques is useful, and biopsy of the small intestine and colon for histological assessment provides key diagnostic information. Endomicroscopy and molecular pathology are only now being explored for the diagnosis of chronic diarrhea. Interest in the microbiome of the gut is increasing; aside from a handful of well-described infections because of pathogens, little is known about alterations in the microbiome in chronic diarrhea. Serological tests have well-defined roles in the diagnosis of celiac disease but have less clearly defined application in autoimmune enteropathies and inflammatory bowel disease. Measurement of peptide hormones is of value in the diagnosis and management of endocrine tumors causing diarrhea, but these are so rare that these tests are of little value in screening because there will be many more false-positives than true-positive results. Chemical analysis of stools is of use in classifying chronic diarrhea and may limit the differential diagnosis that must be considered, but interpretation of the results is still evolving. Breath tests for assessment of carbohydrate malabsorption, small bowel bacterial overgrowth, and intestinal transit are fraught with technical limitations that decrease sensitivity and specificity. Likewise, tests of bile acid malabsorption have had limited utility beyond empirical trials of bile acid sequestrants.
Diarrhea is a common symptom of many different disturbances of gut function. It has been estimated that the average individual living in an industrialized country has an episode of acute diarrhea every 18 months. Chronic diarrhea is less frequent; 3–5% of the population may have diarrhea lasting more than 1 month in any given year.[1, 2] Incidence rates may be higher in places with less advanced sanitation systems, but this point has not been well studied.
The physician's goals are to mitigate the symptom and its potential complications, such as dehydration, to sort out the many conditions that may cause self-limited diarrhea from those that require therapy and to provide effective treatment for the underlying problem. These goals are particularly challenging when diarrhea becomes chronic and therefore less likely to resolve spontaneously. Diagnosis is difficult in part because of the variety and number of diagnostic studies that can be applied to these patients.
In an effort to help practitioners with the diagnosis of chronic diarrhea, the American College of Gastroenterology proposed creation of a working party to the World Gastroenterology Organisation for the 2013 World Congress of Gastroenterology in Shanghai to examine the definition, classification, and diagnosis of chronic diarrhea. One major goal of this working group was to examine the evidence behind definitions of chronic diarrhea and to develop consensus definitions for this symptom complex. Another goal was to establish classification schemes for chronic diarrhea to assist with more expeditious evaluation of patients. Finally, we made an attempt to assess diagnostic tests used in patients with chronic diarrhea to allow gastroenterologists to make an informed selection of these tests, thereby leading to more cost-effective management for these patients. Throughout this report, gaps in knowledge and opportunities for clinical research will be highlighted.
Definition of chronic diarrhea
The dictionary definition of “diarrhea” is “an intestinal disorder characterized by abnormal frequency and fluidity of fecal evacuations”. Despite this fairly simple definition, diarrhea means different things to different individuals.
Most patients consider loose stools—more than frequency—as the key characteristic of diarrhea. Patients may report diarrhea as being present when stools are looser than normal (i.e. liquid or unformed), even when stool frequency is not increased. The Bristol Stool Form Scale (Fig. 1) works well in the clinic to assess consistency and correlates roughly with intestinal transit times.[5, 6] Stool consistency is difficult to quantitate with laboratory tests though and reports of the degree of looseness are subjective. Increased stool frequency is easier to quantitate, and this characteristic is often used by physicians as the sole definition of diarrhea. Stool frequency depends in large part on diet and varies considerably from person to person. On average, normal individuals on a typical American diet produce 25 bowel movements per month. The range of normal is said to be from two or three bowel movements per day to three bowel movements per week depending upon diet. Among patients presenting to a tertiary hospital for evaluation of chronic diarrhea who were found to have normal stool weights (< 200 g/24 h), some patients had high frequency evacuations with minimal change in consistency, while others had unformed stools with normal bowel movement frequency. Patients with higher stool weights tended to have both loose stools and increased frequency. This suggests that while stool frequency and consistency may be independently determined when stool weight is low, the increased water content characteristic of more voluminous diarrheas tends to change both measures.
Stool weight is often used in “scientific” definitions of diarrhea. On average, normal individuals in Western countries produce approximately 100 g of stool per day, with the upper limit of the 95% confidence interval set at roughly 200 g/day. Higher stool outputs can be normal if the diet is loaded with fiber as in other parts of the world: outputs as high as 300 g may be normal when a high-fiber diet is consumed. Conversely, about 20% of patients referred to one tertiary referral center for chronic diarrhea had a normal stool weight. This may reflect frequent small bowel movements, variation in daily frequency, or the presence of fecal incontinence. Urgency and fecal incontinence may drive patient perception of the severity of “diarrhea” and often go unvoiced if not specifically elicited by the physician. Because the leading component of stool weight is stool water, increased stool weight is largely driven by increased stool water output. This, in turn, depends on the underlying pathophysiology-causing diarrhea, and so stool weight is a surrogate for the severity of the abnormality-causing diarrhea (but may not reflect the impact of the disease on the patient).
This review focuses on chronic diarrhea. The duration of diarrhea may be used as a clue to etiology. Acute diarrhea typically is due to bacterial or viral infections that run a course of a week or less. Most of these cases are either self-limited or are easily treated with antimicrobials. Persistent diarrhea (duration 2–4 weeks) has been distinguished mainly by pediatric gastroenterologists and raises the question of a protozoal infection. Even these infections frequently abate within 1 month of onset. A duration of 4 weeks or more has been set arbitrarily as the threshold duration for a diagnosis of chronic diarrhea by several sources.[1, 10] Although some infections, such as those because of Clostridium difficile, Aeromonas, or Yersinia, may last longer than 4 weeks, most patients with diarrhea lasting longer than 4 weeks will be found to have a non-infectious cause of their problem.
Most of the studies used to define a 4-week cut-off for a diagnosis of chronic diarrhea were done in industrialized countries. Information from less developed parts of the world is needed to evaluate how practical such a cut-off is in those regions.
The differential diagnosis for chronic diarrhea is enormous, with a large number of diagnostic tests available that can be used to evaluate these patients. Classifying the patient with chronic diarrhea into a subcategory helps to direct the diagnostic work-up, thereby limiting the number (and cost) of tests needed and increasing the efficiency of the evaluation.[11, 12] Pathophysiologically, chronic diarrhea can be divided into six broad categories (Table 1). Patients can be placed into one of these categories using clues from the history (including details of onset, pattern, duration, epidemiology, past medical history, family history, medications), physical examination (e.g. rash, evidence of malnutrition), and the diagnostic tests that will be discussed later. However, in practice, such categorization is not always used when evaluating every patient. For example, in a patient presenting with diarrhea, weight loss, and dermatitis herpetiformis, the clinician will focus on a diagnosis of celiac disease without first categorizing the patient. Furthermore, some conditions that cause chronic diarrhea do not fit cleanly into one of these categories. An example of this is small intestinal bacterial overgrowth, which can cause secretory or osmotic diarrhea and which can be caused by slow intestinal transit or result in accelerated transit. Moreover, some causes of diarrhea exploit more than one mechanism simultaneously. For example, Crohn's disease (CD) may produce diarrhea by both inflammatory and secretory mechanisms.
On the other hand, in patients in whom a specific diagnosis is not suggested by their clinical features, a systematic approach to classification of the type of diarrhea as discussed later can be very helpful. One step in classifying the type of chronic diarrhea is to distinguish osmotic from non-osmotic causes. Osmotic diarrhea is caused by the ingestion of a non-absorbable substance that results in the presence of osmotically active compounds in the gastrointestinal (GI) tract, which obligate fluid retention within the lumen to maintain osmotic equilibrium with plasma, resulting in diarrhea. Osmotic diarrhea is suggested in a patient whose diarrhea is clearly postprandial and resolves with fasting, and can be confirmed by measurement of stool electrolytes and calculation of the osmotic gap as discussed later. Carbohydrate malabsorption is further suggested by a low stool pH, indicating bacterial fermentation of the unabsorbed substrate.
Another step in classification is to determine the presence or absence of steatorrhea, which is suggested by significant weight loss and frankly greasy or oily stools that are malodorous and float on the toilet water. Some patients may report oil droplets or fat floating on the surface of the toilet water. Steatorrhea can be confirmed by stool analysis, ideally with a timed fecal collection while taking a known quantity of fat in the diet. Inflammatory diarrhea is suggested by the presence of obvious blood or pus in the stools, significant abdominal pain, or fever. It can be confirmed by the presence of fecal leukocytes or leukocyte proteins (such as lactoferrin or calprotectin), or the demonstration of inflammation by direct visualization of the colonic mucosa by endoscopy or histology.
Motility disorders cause diarrhea through either accelerated GI transit (e.g. post-vagotomy diarrhea) or by slowing transit, thereby predisposing to small intestinal bacterial overgrowth (e.g. scleroderma). Motility-related diarrhea can be either secretory or osmotic.
Secretory diarrhea is suggested in a patient with large-volume watery diarrhea that persists despite fasting; it can be confirmed by stool analysis showing increased stool volume with no increase in the osmotic gap. Hormone secreting tumors are often considered in a patient with secretory diarrhea, but these are relatively uncommon causes. Indiscriminant use of tests for these tumors results in a large number of false-positive results and results in unnecessary patient anxiety and cost because of further investigations.
Irritable bowel syndrome (IBS) and functional diarrhea are the most common causes of chronic diarrhea in Western populations, but the underlying pathophysiology remains poorly understood. IBS is defined by the Rome Committee as a chronic condition characterized by abdominal pain and altered bowel habits; the pain characteristically is in association with a change in stool form or frequency, and is relieved by defecation. Studies have validated this definition as identifying a cohort of patients who rarely develop some other bowel condition. In contrast, functional diarrhea—defined as recurrent or continuous passage of loose or watery stools without abdominal pain or discomfort—is less well distinguished from many other diarrheal diseases and easily can be confused with these other conditions.
Finally, factitious diarrhea is due to the intentional ingestion of a substance that causes diarrhea or adulteration of a stool sample with water or urine to exaggerate its volume. Factitious diarrhea can fit into any of the proposed categories depending on which substance is ingested or used to adulterate the stool.
History and physical examination
While most cases of chronic diarrhea will require detailed evaluation, the efficient use of resources and optimum management will be facilitated by obtaining a careful history and performing a thoughtful physical examination. Key historical features to elicit are demographics, mode of onset, patterns of symptoms, presence or absence of pain, and clinical features, including associated comorbidity. Because the commonest diagnosis in at least some areas is IBS, features which are atypical for IBS are especially useful in indicating whom to investigate more extensively. In addition, chronic diarrhea may be due to some systemic disease, and so the physician must be alert to the presence of symptoms and disease outside of the GI tract.
Age of onset and gender are important. Some conditions, notably IBS, show a striking peak in the third and fourth decade, average age (± standard deviation [SD]) at onset was 38 ± 14 years in one large study. In contrast, microscopic colitis is much commoner in the seventh and eighth decade (see Fig. 2), with an average (± SD) age at onset of 57 ± 13 years in a recent series. AIDS-related diarrhea is commoner in younger patients, but most other etiologies can present at any age.
Female predominance is noted in patients with IBS, with a sex ratio of 3 : 1 in those consulting with primary care physicians. The sex ratio is less dramatic in those with diarrhea-predominant IBS. Microscopic colitis also has a female preponderance, while most other conditions causing chronic diarrhea do not show as striking an influence of gender when this has been examined.
Comorbidities and drug history
Prior to patient consultation, it is useful to have an overview of the patient's previous illnesses, comorbidities, and drug history. A good referral letter listing prior consultations is helpful. The typical IBS patient may have had multiple previous consultations in the last 5 years for both GI and non-GI symptoms. This may reflect abnormalities of pain processing sometimes associated with psychological distress that adds to the severity of a multiplicity of symptoms.[20, 21] Patients with celiac disease often have had previous iron deficiency anemia or other autoimmune disorders. It is worth noting that the prodrome before diagnosis may be years in CD that typically relapses and remits. Systemic diseases, such as diabetes mellitus, hyperthyroidism, and Addison's disease, may produce chronic diarrhea as a complication; careful review of the patient's history is mandatory.
Drug history and its relation to onset of diarrhea is important because drugs can cause diarrhea through several mechanisms, including direct pharmacological effects (e.g. β-blockers, metformin, magnesium-containing antacids) or indirect effects (e.g. proton pump inhibitors causing microscopic colitis or antibiotics causing C. difficile colitis) (Table 2).
Many infectious illnesses that can be chronic—like giardiasis—may begin with an acute illness, but some acute infections can trigger presentation of different conditions including IBS, celiac disease, inflammatory bowel disease (IBD), and rarely lactose intolerance. Whether the acute infection actually initiates the condition or makes the underlying chronic condition more obvious is uncertain for most, although at least for IBS, prior sufferers have been excluded from most studies. Onset in association with acute infection can be confirmed by positive stool culture demonstrating a pathogen, but this may not be available, especially if diarrhea occurs while traveling. Other conditions like lymphoma or celiac disease usually develop insidiously.
Pattern of diarrhea
Precise details of stool form using the Bristol Stool Form scale (with visual aids if need be) are useful in understanding the patient's symptoms (Fig. 1). It is important to recognize steatorrhea that significantly alters diagnostic possibilities. Fat laden stools are pale and bulky, often float and are sticky, typically needing several flushes of the toilet, a useful feature that patients may recognize, as most do not examine their stools minutely. Erratic and unpredictable bowel movement are typical of IBS patients, 81% of whom reported > 3 stool forms per week compared with just 41% of those with organic causes of diarrhea. However, history is often unreliable, and in difficult cases, it may be useful for the patient to complete a 1-week stool diary. A study of IBS and diarrhea that recorded symptoms real-time using a web-based system, which is less subject to bias than a diary, showed that loose and watery stools were reported on just 29% of days with normal consistency stool on 42% of days. Variable stools also can be seen in patients with diet-driven diarrhea (e.g. lactose intolerance), but clinical experience suggests that loose stools occur nearly every day for many organic disorders, like microscopic colitis, although there are no directly comparable studies in these conditions.[2, 38]
Another useful feature in the diagnosis of IBS is the periodicity of symptoms with bouts lasting a few days and remitting for a further few days.[19, 39] This short-lived fluctuation would be unusual in most organic diseases, with CD, for example, typically waxing and waning over weeks or months rather than days.
Nocturnal diarrhea has been considered to be an “alarm” feature, suggesting the likelihood of an organic process and the need for more extensive investigations. However, more recent studies suggest that nocturnal symptoms occur in similar proportions of patients with IBS and with a typical organic problem, microscopic colitis (40% and 39%, respectively).[38, 41] Nocturnal diarrhea is recognized as common in diarrhea associated with diabetic autonomic neuropathy and also a feature of postinfectious bile salt malabsorption.[34, 43]
Of course, the key symptom of IBS is abdominal pain, the onset of which is associated with a change in stool frequency or form and which is relieved by defecation. Of course, Pain is not specific for IBS, however. Painless diarrhea is no longer recognized as a form of IBS and should prompt a careful consideration of other diagnoses.
Associated features suggesting need for further investigations
This is an indication for further examination, usually by colonoscopy, although minor anal canal bleeding because of trauma of frequent defecation is extremely common in all diarrheal diseases.
This is common in many diarrheal diseases, often induced by restrictive diets as patients learn to avoid certain foods that aggravate the diarrhea. Weight loss is reported in over half of patients with celiac disease and is more often due to food avoidance than malabsorption. This is likely to be also true in IBD when eating induces pain. While IBS patients typically do not lose weight, this was reported in 15% of IBS patients in one series and so is not a reliable discriminator.
Age > 50 years
The incidence of colon cancer rises over the age of 50, and so most patients presenting with new onset of diarrhea over the age of 50 years will have a colonoscopy for screening purposes, even though colon cancer is a rare cause of chronic diarrhea and the incidence of colon cancer in such patients is low (1–2%). A family history of colon cancer, especially presenting at an early age, would also be a strong indication to perform a colonoscopy, if only to relieve anxiety. Colonoscopy with mucosal biopsy also may be indicated in older patients to look for microscopic colitis. Microscopic colitis is strongly age-related and accounts for 10% of all cases of chronic diarrhea in patients over 70 years of age in Sweden.
While many diarrheal diseases cause mild crampy abdominal pain, which may be relieved by defecation, pain in IBS is often reported as very severe and incapacitating. Crohn's ileocolitis is often associated with pain localized to the right iliac fossa, while most cases of malabsorption are associated with dull, poorly localized discomfort. Inflammatory diseases affecting the rectum may be associated with tenesmus (painful defecation).
Associated psychological factors
While anxiety, panic disorders, and depression are well known to be associated with IBS,[21, 45, 46] being present in 67%, it is less well appreciated that anxiety may prompt presentation to a physician for many illnesses. For example, in a recent study comparing patients with celiac disease to those with IBS (both presenting with diarrhea), anxiety was equally increased in both groups, the Hospital Anxiety and Depression Score being 10.2 ± 1.0 and 9.3 ± 0.9, respectively, both significantly greater than healthy controls at 6.5 ± 0.7. Thus, anxiety may precipitate consultation without necessarily causing the underlying condition. Similarly, it has been reported that diabetics have increased reporting of several GI symptoms, including diarrhea in 18% (normal 12%) and early satiety in 12% (normal 5%). However, anxiety was present in 28% of diabetics (normal 20%), and when the dataset was controlled for anxiety, only early satiety remained significantly more common, suggesting that some symptoms were anxiety-driven.
Patients with chronic diarrhea may have symptoms related to other organ systems that may relate to their underlying illnesses. For example, dysuria might be a symptom related to an enterovesical fistula in CD, and symptoms of pellagra might be due to carcinoid syndrome. In other patients multiple, recurring clinically significant complaints about pain and GI, neurological, and sexual symptoms are related to somatization disorder, a psychiatric syndrome.
Somatization may be a particularly difficult problem in IBS patients. IBS patients often report multiple comorbidities, including psychological disorders such as panic attacks; urinary symptoms such as dysuria, nocturia, and frequency and urgency of micturition; gynecological symptoms such as dyspareunia and chronic pelvic pain; and musculoskeletal problems, including chronic fatigue syndrome and fibromyalgia. PHQ12 Somatic Symptoms scale (PHQ12SS) is a useful measure of non-GI symptoms. A PHQ12SS score > 6 identifies patients with IBS with a sensitivity of 66.4% and specificity of 94.7%, and a positive likelihood ratio of 13.2. A low score makes IBS unlikely and should prompt a search for other diagnoses.
Dietary causes of diarrhea
It is important to take a dietary history to assess intake of foods that might cause diarrhea. A high intake of wheat fiber and certain fruits can cause diarrhea, and many patients with chronic diarrhea respond to dietary restriction of these items. Recent work in Australia has emphasized the importance of poorly absorbed carbohydrates and popularized a diet that restricts FODMAPs (Fermentable Oligosaccharide, Disaccharide and Monosaccharides, and Polyols). Foods containing these substances have been shown in double-blind, controlled trials to trigger abdominal symptoms, such as flatulence, bloating, abdominal discomfort, and diarrhea. FODMAPs include fructose, lactose, fructo- and galacto-oligosaccharides (fructans and galactans), and polyols (sorbitol, mannitol, xylitol, and maltitol) that have limited or no absorption in the small intestine in many individuals (polyols often are used as artificial sweeteners in foods and beverages). When taken in excess, FODMAPs enter the colon and are rapidly fermented, producing short-chain fatty acids, carbon dioxide, hydrogen, and methane, which stimulate colonic motility and can cause diarrhea.[49, 50] Carbohydrate malabsorption can be the major determinant of stool weight in malabsorption syndromes and is the main source of colonic gas production.
Fructose absorption in the small intestine is limited due to its mechanism of absorption by facilitated diffusion. In recent years, there has been a marked increase in consumption of fructose and fructans, particularly in the United States, where high-fructose corn syrup is widely used as a sweetener in commercial food products. Fresh fruits also are now readily available throughout the year, and many (particularly grapes and stoned fruits, such as plums, mangos, and cherries) can cause diarrhea if ingested in excess.
Lactose is a disaccharide that is the major carbohydrate in milk. It must be hydrolyzed to glucose and galactose to be absorbed by the mucosa. Mammals typically no longer ingest lactose after weaning and so production of the lactase enzyme normally is downregulated by adulthood in most mammals and most human populations, rendering them lactose-intolerant. About 10 000 years ago, a mutation developed in the lactase promoter, T/C (-13910), which prevents this normal post-weaning switch off. This mutation first developed in northwest Europe and has spread southward and eastward, so that geographic and racial origin is now a strong predictor of adult lactose tolerance. The prevalence of lactose intolerance is lowest in northwest Europeans and their descendants (10%), 40% in southern Mediterranean and Middle Eastern patients, and up to 90% in patients from the Orient. A separate mutation appears to account for lactase persistence in southern Africa. Diarrhea is likely to be due to lactose intolerance only if the patient ingests > 12 g/day (240 mL of milk or its equivalent in other dairy foods). Milk ingested in large volumes that may empty rapidly into the small intestine is more likely to cause symptoms than when it is incorporated into solid foods.
Excessive alcohol intake, particularly as beer, impairs intestinal water absorption and can cause diarrhea. Prolonged alcohol abuse can also cause chronic pancreatitis and lead to pancreatic exocrine insufficiency leading to diarrhea secondary to malabsorption. Caffeine causes jejunal secretion and may be responsible for diarrhea in some patients.
Physical examination is of limited value in most patients with chronic diarrhea beyond assessment of hydration and nutritional status. General examination should take note of evidence of weight loss and signs of malnutrition, such as anemia, vitamin deficiency, clubbing, or lymphadenopathy. Abdominal examination is usually unremarkable apart from vague tenderness. Abdominal masses are rare, but fullness in the right iliac fossa may be felt in ileocolonic CD. Perineal inspection and rectal examination is useful to exclude any induration or local tenderness that might suggest CD or an anal fissure. Assessment of voluntary squeeze is useful in assessing patients with urgency to see if there is any sphincter defect because incontinence may lead to a complaint of diarrhea. Table 3 lists rare physical findings that can be of diagnostic value.
Table 3. Physical findings of interest in chronic diarrhea
Muscle wasting, edema
Urticaria pigmentosa, dermatographism
Mast cell disease (mastocytosis)
Pinch purpura, macroglossia
Migratory necrotizing erythema
Malignant atrophic papulosis
Thyroid nodule, lymphadenopathy
Medullary carcinoma of the thyroid
Tremor, lid lag
Right-sided heart murmur, wheezing
Endocrine tumor, amyloidosis
Inflammatory bowel disease, yersinosis
HIV, lymphoma, cancer
Chronic mesenteric ischemia
Anal sphincter weakness
Many types of radiological imaging studies are useful in the diagnosis of chronic diarrhea in selected patients. The choice of imaging study is based on history, clinical presentation, and type of diarrhea. In addition, because of increasing concern for exposure to ionizing radiation, it is incumbent to select these studies with care and to avoid repetitive studies when possible. When pancreatic insufficiency is suspected, a plain abdominal radiograph that shows pancreatic calcifications is diagnostic of established chronic pancreatitis (although computerized tomography [CT] scanning is more sensitive). In high-volume secretory diarrheas, radiography may demonstrate intestinal air fluid levels or a paucity of bowel gas, suggesting a fluid-filled bowel. Standard abdominal CT scan is useful in detecting extraintestinal causes of chronic diarrhea, such as neuroendocrine tumors and chronic pancreatitis, but is a poor test for small bowel mucosal disease.
Small bowel radiography
Small bowel barium studies have historically been used in the diagnosis of chronic diarrhea and abdominal pain. When bacterial overgrowth is suspected as a cause of malabsorptive diarrhea, small bowel barium study is useful in identifying multiple jejunal diverticula, altered small bowel anatomy, blind loops, gastrocolic fistula, slow small bowel transit, and small bowel strictures. It also allows localization of small bowel lesions to guide further management. Findings that suggest small bowel mucosal disease include jejunal dilation, fold thickening, and intussusceptions. These findings are non-specific and are present in celiac disease, tropical sprue, scleroderma, and hypoalbuminemia, as well as other mucosal diseases. Findings that are more specific for celiac disease include decreased jejunal-fold pattern and increased ileal-fold pattern (ilealization of jejunum or so-called reversal of fold pattern). With the introduction of abdominal CT scans, use of fluoroscopic barium studies has declined. Classic radiographic findings are listed in Table 4.
Table 4. Radiographic findings in chronic diarrhea and malabsorption syndromes
CT and magnetic resonance (MR) enterography protocols using intravenous and high-volume negative (or neutral) oral contrast have revolutionized radiological examination of the small bowel wall and also allow examination of extraintestinal structures.[59-61] They are useful in the diagnosis of chronic diarrhea because of small bowel CD, particularly when out of reach of standard endoscopy. Findings in active CD include mucosal enhancement, mural thickening, proliferation of mesenteric fat, and dilated vasa recta. Findings characteristic of active celiac disease include reversal of fold pattern, intussusception, hyposplenism, and mesenteric adenopathy. Findings in eosinophilic gastroenteritis depend on the site of small bowel involvement: fold-thickening, ulcers, and polyps are found with mucosal involvement; bowel wall thickening, decreased distensibility, and strictures are found with muscle involvement; and ascites, adenopathy, and omental thickening are found with serosal involvement. CT enterography also is useful in the detection of small bowel tumors, such as carcinoids, that can be detected when they are as small as 5 mm in size. CT enterography is being used in developing countries, but technical limitations in such settings pose difficulties. MR enterography provides high-quality imaging without ionizing radiation but is more costly, time-consuming, and more difficult to interpret because of motion artifacts from bowel movement. These limitations are likely to change with newer high-speed MR scanners. MR enterography is likely to become more widely used as concerns with iatrogenic radiation exposure grow. MR cholangiography and pancreatography largely has supplanted endoscopic techniques as the diagnostic technique of choice for pancreatobiliary disease.
Nuclear medicine imaging
Neuroendocrine tumors are a rare cause of secretory diarrhea. These tumors can be small and difficult to diagnose. Radioligand scintigraphy (e.g. OctreoScan) is useful in diagnosing neuroendocrine tumors that express somatostatin receptors, such as gastrinomas and carcinoid tumors. The newer tomographic hybrid scanner, single-photon emission computed tomography-CT provides better localization and may increase accuracy. Use of higher doses of radioligands may improve tumor detection rates. Positron emission tomography combined with CT or MR is useful in detecting and localizing small bowel lymphomas and for monitoring response to therapy.
Endoscopy and colonoscopy
Endoscopy is a commonly used diagnostic test in the evaluation of chronic diarrhea. Upper endoscopy is indicated when there is a history of chronic diarrhea with weight loss, positive celiac serologies, and/or vitamin and mineral deficiencies to suggest small bowel mucosal disease (chronic infection, celiac disease, tropical sprue, eosinophilic gastroenteritis, CD, radiation, amyloidosis, common variable immunodeficiency syndrome, lymphangiectasia, graft vs host disease). Findings of nodularity, fissuring, or scalloping in the duodenum are suggestive of villous atrophy from any cause. Endoscopic visualization of villous atrophy is improved with water emersion, zoom magnification, optical band imaging, and confocal microscopy.[69, 70] Systematic evaluation of the utility of endomicroscopy in patients with chronic diarrhea is needed. The duodenum may appear normal with milder degrees of inflammation (villous blunting or intraepithelial lymphocytosis with normal architecture) and therefore should be biopsied when small bowel mucosal disease is suspected. Duodenal biopsy may provide a specific diagnosis such as in giardiasis and other protozoal infections, Whipple's disease and other infections, combined variable immunodeficiency, and graft versus host disease. In other diseases, duodenal biopsy findings are non-specific, but suggestive, such as in celiac disease, tropical sprue, and CD. In such cases, the diagnosis is made with supporting serologies or in response to a specific treatment. In celiac disease, diagnostic yield improves with targeted biopsies of abnormal appearing mucosa, when four duodenal biopsies are obtained and when both the bulb and distal duodenum are biopsied.[71, 72] Upper endoscopy also permits collection of duodenal aspirate for quantitative culture, the current “gold standard” for a diagnosis of small intestinal bacterial overgrowth.
Colonoscopy with ileoscopy is indicated in patients with watery, inflammatory, or elusive diarrhea to assess for IBD, microscopic colitis, infections such as C. difficile, ischemia, villous adenoma, or mastocytosis. Colonoscopy with ileoscopy has highest specificity when compared with CT imaging and capsule endoscopy in the diagnosis of CD. Ileoscopy is also useful in the diagnosis of infections that cause chronic diarrhea such as tuberculosis or yersinosis. The colon appears grossly normal with microscopic colitis; a sufficient number of biopsies should be obtained to increase the chances of making a diagnosis. In contrast, C. difficile typically produces grossly apparent pseudomembranous colitis that is difficult to mistake for anything else. Approximately 15% of cases of non-bloody chronic diarrhea have histological findings in the colon; biopsies from the left colon are sufficient to detect the causative condition in the vast majority of these cases.
Enteroscopy and capsule endoscopy
Most small bowel mucosal diseases that cause chronic diarrhea can be diagnosed by standard upper endoscopy and duodenal biopsy. Rarely, celiac disease has a patchy distribution that is missed on duodenal biopsy must be diagnosed by push enteroscopy with jejunal biopsy. Newer technologies of wireless capsule endoscopy and device-assisted (deep) enteroscopy allow complete examination of the small bowel.
Capsule endoscopy is more sensitive than standard endoscopy in the detection of villous atrophy with good interobserver agreement,[75, 76] but it misses milder inflammatory lesions. As tissue sampling is not possible with capsule endoscopy, it cannot provide a specific diagnosis for other small bowel mucosal diseases that cause chronic diarrhea. The diagnostic yield is particularly low in patients with chronic diarrhea and no laboratory or imaging studies to suggest organic disease. Capsule endoscopy is useful in refractory celiac disease for the detection of ulcerative jejunitis and enteropathy associated T-cell lymphoma (EATL). Capsule endoscopy has similar sensitivity as CT enteroscopy, ileocolonoscopy, or small bowel radiography in small bowel CD, but lower specificity limits its utility in that setting. One of the risks of capsule endoscopy is retention within the intestine. This occurred in 1.3% of patients in one retrospective study from Sweden; risk factors included IBD and tumors. The risk of capsule retention in known CD may be as high as 13%.
Device-assisted enteroscopy (balloon or spiral overtubes) allows complete examination of the small bowel and has the advantage of permitting tissue sampling but is invasive and time-consuming. It is useful to assess for ulcerative jejunitis and EATL in refractory celiac disease and for tissue diagnosis of other causes of chronic diarrhea when suspicious lesions are identified at capsule endoscopy or radiological small bowel imaging studies.
Endoscopic retrograde cholangiopancreatography and endoscopic ultrasound
Detection of chronic pancreatitis and neuroendocrine tumors of the pancreas that might cause chronic diarrhea can be facilitated by endoscopic retrograde cholangiopancreatography (ERCP) or endoscopic ultrasound (EUS).[81, 82] EUS permits biopsy of lesions and ERCP can be used to treat certain forms of chronic pancreatitis.
Histopathology of the small bowel and colon has an important role in the evaluation of patients with chronic diarrhea. Biopsy of both abnormal and normal appearing mucosa may be indicated.
Colorectal and terminal ileal biopsy
Biopsy is essential in the diagnosis of colitis to confirm inflammation, to identify its causes, and to identify other abnormalities (Table 5). A common differential is between infectious colitis and IBD. Some features such as crypt abscesses are non-specific. The histological features of infectious or acute self-limited colitis are preservation of normal crypt architecture and predominantly acute inflammation in the lamina propria; inflammation may be patchy. In contrast, features of active ulcerative colitis are contiguous mucosal inflammation with abnormal crypt architecture with branched crypts, acute and chronic inflammatory cells in the lamina propria, increased inflammatory cells at the crypt bases, and basilar plasmacytosis with basal lymphoid aggregates.[84, 85] In fact, basilar plasmacytosis has been identified as a predictor of clinical relapse. In ulcerative colitis, when inflammatory changes resolve (“mucosal healing”), crypt distortion often remains.
Table 5. Interpretations of pathological findings on colorectal biopsy
aIncludes lymphocytic colitis and collagenous colitis
The histological features of colitis in CD are focal inflammation and architectural distortion, transmural inflammation, and epithelioid granulomas that can be found in 10–20% of cases. While diffuse inflammation can be seen, it is less common. The focal nature of Crohn's inflammation may be confused with infectious colitis early in the course of disease. In approximately 10% of patients with IBD, endoscopic and histological features are insufficient to distinguish ulcerative colitis and CD with confidence, such patients are described as having indeterminate colitis. Segmental colitis associated with diverticulitis can show focal inflammation, granulomas, and may mimic CD or ulcerative colitis.
Sometimes specific pathogens can be seen on biopsy, such as ameba trophozoites, schistosomes with associated granulomas, and viral inclusions because of cytomegalovirus (CMV) infection. Granulomas can be seen in tuberculosis, yersinosis, and schistosomiasis.
Pseudomembranes on biopsy can be seen with C. difficile infection but also with ischemia and Shiga toxin Escherichia coli infections. Other more common features of ischemia are loss of mucosa in the upper lamina propria, hemorrhage, and a superficial polymorphonuclear infiltrate. There is minimal inflammation; necrosis can be seen later in severe cases. In radiation colitis, there is ischemia and fibrosis, with scarring, neovascularization with blood vessels in the upper mucosa, and hyalinized blood vessels.
Biopsy of normal colonic mucosa is important in the evaluation of patients with chronic watery diarrhea. It may detect occult IBD, pseudomelanosis coli (a sign of laxative use), and is necessary to diagnose microscopic colitis (both collagenous colitis and lymphocytic colitis), eosinophilic colitis, and enteritis, and may yield a diagnosis in 10–30% of cases.[87, 88] In a series from Brazil, histology was abnormal in 32.1% and identified occult IBD, microscopic colitis, eosinophilic enterocolitis, spirochetosis, and schistosomiasis in addition to findings of borderline significance. The histological features of collagenous colitis are normal architecture, increased intraepithelial lymphocytes (IELs) (more than 20 IELs per 100 absorptive cells) in the surface mucosa and crypts, increased lamina propria lymphocytes, and a thickened subepithelial collagen band (> 10 microns). Lymphocytic colitis has the same features but no thickened collagen band. Other changes that can be seen with microscopic colitis are surface epithelial flattening, cryptitis, and Paneth cell metaplasia. Most cases will be detected with biopsy of the left colon, but it is important to obtain biopsies above the rectum. Usually four random biopsies of left colon will be adequate. Biopsies of the right colon are rarely necessary to make the diagnosis. One caution in interpreting random colon biopsies from the cecum is to distinguish mild inflammatory changes that can be seen in normal individuals from true colitis. A reasonable practice is to obtain four to five random biopsies from throughout the colon and place them in a single bottle.
Terminal ileal biopsies
Ileitis can be detected in patients with CD, even without visible inflammatory changes and in some patients with ulcerative colitis in whom it is termed “backwash” ileitis. The yield of biopsy of the terminal ileum in patients with diarrhea is highest in those with right lower quadrant pain; it may aid diagnosis in 10–20% of cases. In patients without right lower quadrant pain, the yield will be much lower.
Small bowel biopsy
Small bowel biopsy plays an important role in the evaluation of patients with fat malabsorption as well as the evaluation of chronic watery diarrhea. It can detect mucosal disease such as celiac disease, CD, eosinophilic enteritis, tropical sprue, and infections with organisms such as Mycobacterium tuberculosis, Strongyloides stercoralis, Giardia intestinalis (formerly G. lamblia), Cystoisospora belli (formerly Isospora belli), Cyclospora cayetanensis, and Cryptosporidium spp. Small bowel biopsy in patients with HIV associated diarrhea who are immunosuppressed can detect any of the previously listed organisms and also microsporidiosis and Mycobacterium avium intracellulare (MAI). Uncommon small bowel diseases that can be recognized by biopsy include Whipple's disease, collagenous sprue, autoimmune enteropathy, and lymphoma. Classical findings of these conditions are listed in Table 6. In addition, finding normal small bowel mucosa in a patient with malabsorption may point to pancreatic insufficiency as a cause.
Table 6. Interpretation of pathological findings on small bowel biopsy
EM, electron microscopy; PAS, periodic-acid Schiff stain.
Whipple's disease (bacilli on EM), Mycobacterium avium-intracellulare (acid-fast bacilli)
Small bowel biopsies in some diseases are diagnostic—particularly for very rare diseases such as abetalipoproteinemia, collagenous sprue, amyloid, and Whipples's disease. Small bowel biopsy can be suggestive but non-diagnostic in celiac disease and CD. Flattened mucosa (i.e. villous atrophy or blunting) can be seen in all of these conditions, as well as with viral gastroenteritis and nonsteroidal anti-inflammatory drugs use; thus, clinical correlation is always important. Histological characteristics that suggest celiac disease are variable blunting or flattening of the villi that can be patchy and IELs (more than 30 IELs/100 enterocytes). However, these findings are non-specific and can be seen with infections and other inflammatory diseases, such as CD.
The current gold standard in biopsy diagnosis of celiac disease is to take four to six biopsies in the duodenum; one should be from the bulb and the others in the second and third portions of the duodenum. While serology can suggest celiac disease, biopsy remains mandatory to confirm the diagnosis. It is usually not necessary to rebiopsy to document healing unless there is a poor response to a gluten-free diet or there is doubt about the initial diagnosis.
Eosinophilic gastroenteritis involves the stomach and small bowel most commonly. There may be peripheral eosinophilia. Histological characteristics are an eosinophilic infiltrate, > 10–50/high power field, that can be variable in location but is usually in the mucosa. The small bowel is involved in 75% of cases. However, increased eosinophils can be seen in other situations, such as parasitic infections and CD.
Pathological analysis of tissue samples is undergoing a renaissance as molecular techniques allow more information to be extracted from specimens than just morphology. These techniques include expanded special stains and genetic analysis of both mucosa and associated microorganisms. Application of these methods to samples from patients with chronic diarrhea is just beginning.
In developed countries and in the normal, immunocompetent host, infections are unusual causes of chronic diarrhea. However, in developing countries chronic bacterial, mycobacterial, and parasitic infections are common. Additionally, there are special situations where intestinal infections are more frequently responsible for chronic diarrhea. Some of these situations include diarrhea in immigrants from endemic areas, immunocompromised subjects, patients with HIV/AIDS infection, men who have sex with men, and in individuals with chronic traveler's diarrhea.
Immunocompetent subjects in developed countries
In immunocompetent individuals in developed countries, giardiasis, amebiasis, yersinosis, and C. difficile infections may be chronic. Strongyloides is occasionally seen but is quite unusual. In immunocompetent patients with chronic diarrhea, these five pathogens should be sought if the cause of diarrhea is not readily apparent. Giardia is most reliably detected with a stool enzyme-linked immunosorbent assay (ELISA) assay. Ameba and Strongyloides are sought with serological tests and stool examination for ova and parasites; no more than three stool samples should be sent for microscopic examination. C. difficile is most reliably detected with a stool DNA amplification assay; only a single stool sample need be done for that diagnosis. Less reliable tests include enzyme-linked immunoassays for C. difficile toxins A and B that suffer from poor sensitivity.
In patients receiving immunosuppressant medications or those with HIV/AIDS infection, the likelihood of chronic infections is much greater. Many common enteropathogens that cause acute, self-limited diarrhea in immunologically normal individuals can cause chronic diarrhea in these patients. These pathogens include Salmonella, Shigella, Campylobacter, E. coli, Yersinia, and others. These infections can last many weeks in the immunosuppressed host. Traditionally, these infections have been sought with standard stool cultures; however, new molecular techniques may prove to be better in time, making standard stool cultures obsolete. Immunosuppression may cause atypical presentations of infectious diarrhea. For example, patients with subclinical Strongyloides infection may develop hyperinfection syndrome or disseminated strongyloidiasis when exposed to high-dose corticosteroids for treatment of asthma or chronic obstructive pulmonary disease.
With HIV/AIDS potential infectious etiologies are related to the degree of immunosuppression. With lesser degrees of immunosuppression (CD4 count > 200 cells/mm3), the usual pathogens mentioned in the preceding paragraph predominate. If the CD4 count is < 200 cells/mm3, the potential spectrum is much wider. In addition to the enteropathogens mentioned, mycobacterial and protozoan infections become more likely. These include MAI, cryptosporidia, cyclospora, cystoisospora, and microsporidia. Viral infections, such as CMV and Herpes simplex virus, and fungal infections, such as candidiasis and histoplasmosis, should be considered if other pathogens are not found.
In immunocompromised subjects, a staged work-up is recommended. Initially, stools should be sent for routine bacterial culture (to detect Salmonella, Shigella, and Campylobacter), a test for C. difficile, ELISA testing for giardia and cryptosporidium, and for ova and parasite examination to look for Stronglyoides and other parasites. If these tests are unrevealing, upper GI endoscopy and either flexible sigmoidoscopy or colonoscopy with biopsies should be done. The biopsies should be examined microscopically and cultured for viral pathogens.
Men who have sex with men
In men who have sex with men (especially those who practice rectal intercourse), organisms that cause proctitis need to be considered, especially when rectal or anal pain, tenesmus, or rectal bleeding is present. These organisms include Chlamydia trachomatis serovars L1, L2 and L3 (cause of lymphogranuloma venereum), amebiasis, Herpes simplex virus, Neisseria gonorrhoeae, and Treponema pallidum. In addition to stool cultures, sigmoidoscopy with biopsies should be done routinely in these individuals.
Chronic “traveler's diarrhea”
Traveler's diarrhea is usually acute and self-limited. In approximately 3% of patients, the diarrhea is persistent and chronic. In this situation, chronic intestinal infection must be excluded, and stool samples should be sent for bacterial culture and microscopy, ELISA tests for giardiasis and cryptosporidiosis, and a test for C. difficile, especially if the patient previously received antibiotics. If these investigations are negative, small intestinal biopsy should be considered to look for unusual infections or tropical sprue. Tropical sprue should be included in the differential diagnosis, especially if the traveler spent a long time in an endemic region. If all of these tests are negative, some suggest empirical treatment with antibiotics aimed at bacterial enteropathogens and, if that does not work, a course of therapy for protozoal pathogens. Others would proceed directly with small bowel biopsy to look for tropical sprue. Some patients develop IBS after a bout of traveler's diarrhea.
There have been many outbreaks of diarrhea in which some individuals developed chronic diarrhea. These outbreaks have characteristics of a point-source epidemic, usually associated with potential food or water contamination. One of the first outbreaks occurred in Brainerd, Minnesota, giving the condition its common name, “Brainerd diarrhea”. Despite the obvious concern for an infectious cause, state-of-the-art microbiological evaluation of an ongoing outbreak failed to detect an agent for the disease. An identical syndrome also can occur sporadically with no indication of direct person-to-person spread. While this condition shares some similarity to postinfectious IBS, patients have no pain, complain of continuous watery diarrhea that is moderately voluminous, and eventually get better (after months or years). It is unclear whether this represents ongoing infection with a novel agent or some sort of long-lasting reaction to an acute infection or a self-limited form of functional diarrhea.
Serological testing is used to support specific diagnoses in many subspecialties, including rheumatology and hepatology, but has had limited use in the diagnosis of chronic diarrhea. There are three areas in which serological testing has been considered: (i) celiac disease; (i) IBD; and (iii) autoimmune enteropathy. The usefulness of serological tests in these conditions is variable and ranges from indispensable (celiac disease) to just suggestive (IBD).
Serological tests are basic to the diagnosis of celiac disease. Serological testing has provided new insights into the epidemiology of celiac disease, expanding our understanding of its worldwide prevalence. It has also given us a more comprehensive picture of the spectrum of the disease, with the recognition of latent and silent celiac disease. Given the recent recognition of an entity now termed “non-celiac gluten sensitivity,” serological testing has assumed a greater importance in diagnosis of “true” celiac disease.
Immunoglobin A anti-tissue transglutaminase (anti-TTG) and anti-endomysial antibody (anti-EMSA) are very sensitive and specific for a diagnosis of celiac disease. Immunoglobulin A (IgA) anti-TTG is the preferred single test for detection of celiac disease in individuals over the age of 2 years and should be done, while the patient is consuming gluten. Total serum IgA should be measured at the same time to exclude IgA deficiency that might cause a falsely negative test (present in 2–3% of patients with celiac disease). Antigliadin antibodies generally are positive in celiac disease, but are not specific, and therefore not helpful for diagnosis. Anti-gliadin antibodies occur frequently (10–15%) in patients with IBS and other intestinal diseases, such as CD. Their presence speaks to near universal exposure to gluten in the Western diet and may reflect changes in intestinal permeability. Some patients with positive anti-TTG or anti-EMSA tests have no symptoms and no definite biopsy evidence of celiac disease; these individuals are said to have “potential” or “latent” celiac disease. Patients with potential celiac disease have no history of a previous diagnosis, whereas those with latent celiac disease have had a previous diagnosis of celiac disease. Many of these patients have a family history of celiac disease and also may have symptoms suggestive of celiac disease, including chronic diarrhea, abdominal pain, and anemia. The significance of these findings remains a subject of debate.[108, 109]
In an individual with steatorrhea, dermatitis herpetiformis, or other signs/symptoms strongly suggestive of celiac disease, serological testing is indicated. Small intestinal biopsy should be considered in such patients, even if serological tests are negative. The role of serological testing is less clear in individuals with a clinical picture of diarrhea-predominant IBS or intermittent diarrhea. Several studies have examined the question of how frequently celiac disease presents as IBS and have reached different conclusions. A meta-analysis concluded that celiac disease (positive serologies and biopsies) was fourfold more common among IBS patients than normal, approximately 4%. A subsequent large study from the United States found that positive serology tests for celiac disease were common in both controls and non-constipated IBS patients, and found no difference in the prevalence of biopsy-proven celiac disease (0.44% and 0.41%, respectively). Routine testing for celiac disease in diarrhea-predominant IBS may be cost-effective in populations with a greater than 1% background prevalence of celiac disease.
HLA-DQ heterodimers DQ2 or DQ8 are present in almost all patients with celiac disease. Therefore, the negative predictive value of tests for these tissue types is > 99%. While HLA-DQ testing should not be used for diagnosis of celiac disease because of a high prevalence of these tissue types in most populations (30–40%), absence of HLA-DQ2/DQ8 can be used to exclude the disorder when there are equivocal histological findings or conflicting serological and pathological results, or when a patient is on a gluten-free diet and refuses to resume gluten consumption before testing.[106, 114]
Serological tests commonly used in managing IBD measure antibodies targeting a yeast used in food production (anti-Saccharomyces cerevisiae antibodies, directed against Saccharomyces cerevisiae, a yeast used in winemaking, baking, and brewing), an intracellular neutrophil component (perinuclear anti-neutrophil cytoplasmic antibodies, directed against various proteins inside neutrophils), and bacterial components (anti-OmpC, outer membrane porin C of E. coli, and anti-CBir-1, flagellin). Reactivity to these antigens may represent cross-reactivity to a molecular mimic, exposure to cell components normally inaccessible to the immune system, or altered permeability for these antigens.
Studies specifically examining the role of IBD serology in the differential diagnosis of diarrhea are limited. The predictive value of a test depends on the pretest probability of the diagnosis in the patient being studied. In the setting of bloody diarrhea, fever, and abnormal imaging, the likelihood of IBD is fairly high and so the predictive value of a positive IBD serology is substantial, but it is not clear that serological tests add much diagnostic certainty to that provided by standard tests, such as colonoscopy or enterography. Moreover, in a patient with watery diarrhea and no other IBD alarm signs, the likelihood of IBD will be low and the false-positive rate high. Using IBD serological testing in this clinical setting rarely is useful and frequently causes confusion and additional unnecessary testing. It has been suggested that combining different serological tests with genetic and inflammatory markers can improve the performance of blood tests for IBD, but this remains to be established prospectively in a cohort with a prevalence of IBD similar to the general population.
Autoimmune enteropathy is a rare condition marked by intractable diarrhea, malabsorption, and histological changes on small intestinal biopsy that resemble but are not pathognomonic for celiac disease. It is often confused with celiac disease but does not respond to gluten withdrawal or other dietary manipulations, and the histology is subtly different than that seen in celiac disease. Originally, it was considered a pediatric disease, but it does occur in adults. Specialized research laboratories offer serological testing, such as anti-enterocyte antibodies. These may be helpful in confirming the diagnosis, but because this is such a rare and variable disease, performance characteristics are not well defined and a positive test is not necessary for establishing a diagnosis. Steroids and other immunosuppressives frequently are required for treatment.
In a small number of patients secretory diarrhea is caused by circulating agents that lead to water and electrolyte secretion at the mucosal level and often also decrease intestinal transit time. Many of these conditions are classic syndromes where a neuroendocrine tumor produces and releases a peptide hormone or a neurotransmitter. The Verner–Morrison syndrome (vasoactive intestinal polypeptide [VIP]oma or watery diarrhea—hypokalemia—hypochlorhydria syndrome) is caused by excessive production of VIP;[120, 121] serotonin, substance P, and tachykinins are the culprits in the malignant carcinoid syndrome. In gastrinoma (Zollinger–Ellison syndrome [ZES]) the large amounts of gastric acid secretion overwhelm the absorptive capacity of the intestine. Neuroendocrine tumors are rare, the incidence is about 2.5 per 100 000 population per year, and two-thirds are non-functioning and do not release a tumor product into the circulation. The classic syndromes with endocrine activity per se are thus much rarer: it is estimated that there is one VIPoma per 10 million people per year and one ZES per 2 million people per year.
Radioimmunoassays (RIAs) are available to determine elevated levels of many of these hormones (e.g. gastrin, VIP, calcitonin), but because of the rarity of these syndromes, the rate of false-positive laboratory results remains a problem. With an extended panel of assays including motilin, neurotensin, pancreatic polypeptide, somatostatin, substance P, VIP, gastrin-releasing peptide, and calcitonin, the high rate of false-positive results precludes any clinical utility for the evaluation of patients with chronic diarrhea. Cholecystokinin (CCK) could now also be added to this list. In other words, if peptide hormone levels were used as a diagnostic tool for secretory diarrhea, the yield of true-positives would be very low and false-positives very high. This changes when secretory diarrhea accompanies a tumor mass found in the pancreas, duodenal wall, ileum, or parathyroid. In this setting, RIAs for VIP, gastrin, somatostatin, and calcitonin may be useful.
In the presence of small bowel tumors or symptoms, and signs of flushing, hepatomegaly or valvular heart disease, the malignant carcinoid syndrome should be suspected. Determination of the serotonin metabolite, 5-hydroxyindoleacetic acid, in 24-h urine is still the most reliable test. When the carcinoid syndrome is present (e.g. diarrhea, flush) metastases of the primary tumor usually have already increased total tumor mass and provided access for tumor secretion to the systemic circulation.
Chemical analysis of stool
Conceptually, chemical analysis of stool in patients with chronic diarrhea can allow insight into pathophysiology and could lead to expedited diagnosis. The differential diagnosis of chronic diarrhea includes dozens of entities, and therefore, diarrhea can be daunting to evaluate. Different etiologies cause different types of diarrhea. When an obvious diagnosis is not evident, categorizing patients with chronic diarrhea as having watery diarrhea, fatty diarrhea, or inflammatory diarrhea based on clinical presentation and simple stool analysis can result in a less imposing differential diagnosis (Table 1). The stool analysis involves inspecting the stool, measuring stool weight, stool electrolytes, fecal pH, and fat content, and checking for the presence of blood and white blood cells. Additional studies that can be done selectively include measurement of carbohydrate excretion, fecal chymotrypsin or elastase, and screening of stool and urine for laxatives.
Measurement of stool weight gives guidance about the severity of diarrhea and the need for fluid or electrolyte repletion. Stool weight in functional diarrhea or IBS typically is < 500 g/24 h; higher outputs suggest a more substantial disruption of normal absorptive physiology. Assay of fecal electrolytes allows the physician to distinguish osmotic and secretory diarrhea based on calculation of the fecal osmotic gap. Secretory diarrhea produces stools that are rich in electrolytes because intraluminal fluid retention is due to incomplete absorption of electrolytes. Electrolyte absorption is unaffected in osmotic diarrhea, and so stools have low concentrations of absorbable electrolytes. The contribution of fecal electrolytes (sodium, potassium, and associated anions) to stool osmolality can be judged by adding measured fecal sodium and potassium concentrations, doubling that number to account for unmeasured anions, and subtracting that number from 290 mOsm/kg (the osmolality of intraluminal contents that have equilibrated with body fluids) (measured stool osmolality is affected by fermentation that occurs in the collection container and should not be used for this determination). If the fecal osmotic gap is small, most of the osmolality of stool water is accounted for by electrolytes; this suggests that water and electrolyte absorption is impaired. If the fecal osmotic gap is large, some non-electrolyte contributes substantially to fecal osmolality. This usually is a poorly absorbed substance that holds fluid within the gut lumen, such as magnesium or carbohydrate. Experiments using laxatives that produce either secretory or osmotic diarrhea suggest that a fecal osmotic gap < 50 mOsm/kg is characteristic of a secretory diarrhea and a fecal osmotic gap of > 75 mOsm/kg is characteristic of osmotic diarrhea. Thus, calculation of the fecal osmotic gap allows one to classify most cases of watery diarrhea into secretory diarrhea or osmotic diarrhea that impacts on diagnosis and further evaluation.
Fecal pH normally is close to 7.[8, 127] Fermentation of carbohydrate in the colon by the colonic flora produces short-chain fatty acids that may lower fecal pH. This is particularly true in normal volunteers given a large dose of lactulose by mouth. Thus, low fecal pH may be a clue to the possibility of carbohydrate malabsorption. However, unlike normal subjects given lactulose, patients with carbohydrate malabsorption may not always have a low fecal pH probably because the amount of carbohydrate malabsorbed may be less than in normal volunteers given lactulose or because buffering by mucosal bicarbonate secretion may be more effective.
The presence of blood or pus in the stool raises the possibility of an inflammatory diarrhea because of colitis or ileitis. Not every inflammatory diarrhea causes blood or pus in the stool; however, there has to be mucosal disruption as part of the inflammatory process.
Stool fat content should be measured to see if steatorrhea is present. Steatorrhea indicates a problem with fat absorption in the small intestine because of either mucosal disease or luminal factors, such as insufficient bile acid concentration, small bowel bacterial overgrowth, or pancreatic exocrine insufficiency. While it is helpful to measure stool fat output on a timed stool collection so that steatorrhea can be quantitated, stool fat content (g/100 g stool) tends to be elevated in many patients with steatorrhea. This allows detection of steatorrhea on a spot stool specimen by means of chemical tests or microscopic inspection of stool stained with a lipophilic stain (e.g. Sudan III) if a timed collection cannot be obtained.
There are few data regarding the utility of this sort of stool analysis in the evaluation of chronic diarrhea. A recent study looked at the value of measuring stool electrolytes, pH, and fat content in a group of patients referred to one center with chronic diarrhea that was difficult to diagnose or treat. This study suggested that results of stool analysis could be used to identify six patterns of stool composition with important potential impacts on diagnosis and the selection of tests for further evaluation (Table 7). This might permit an algorithmic approach to the further evaluation of these patients.
Table 7. Patterns of stool composition in chronic diarrhea
IBS, irritable bowel syndrome.
Stool weight < 200 g/24 h
● No objective evidence of diarrhea
Change in stool frequency, intermittent diarrhea, fecal incontinence, treatment with antidiarrheal drugs during collection
● Hyperdefecation (increased frequency without excess volume)
Possible IBS, proctitis, abnormal rectal reservoir function
● Abnormal consistency (unformed to runny stools)
● Elevated fecal osmotic gap
Presumed mild carbohydrate malabsorption or excess Mg intake from supplements
Malabsorption or maldigestion
Secretory diarrhea without steatorrhea (stool weight > 200 g/24 h)
Microscopic colitis or other cause of secretory diarrhea
Carbohydrate malabsorption without steatorrhea
● High fecal osmotic gap
Ingestion of poorly absorbed carbohydrates, malabsorption
● pH not always < 5.5
Steatorrhea with or without carbohydrate malabsorption
Small bowel mucosal disease, small intestinal bacterial overgrowth, bile acid deficiency
Because of ingestion of poorly absorbed ions (e.g. magnesium, phosphate, sulfate) or osmotically active polymers (e.g. polyethylene glycol)
Unclassified (stool weight > 200 g/24 h
Blood or pus suggests inflammatory causes of diarrhea
A major limitation to this study is the selection of patients included in the analysis. Many of these patients had previous evaluations, and it is likely that the distribution of etiologies is different than a population-based sample would be. It does point out the potential for stool analysis to help with the evaluation of these patients, however, and suggests avenues for further research.
Additional stool studies that can be done on a selective basis include surrogate measures for fecal leukocytes (lactoferrin or calprotectin) and pancreatic exocrine function (chymotrypsin or elastase). Microscopy to look for fecal leukocytes is operator-dependent and semiquantitative. Measurement of the leukocyte enzymes, lactoferrin, or calprotectin is reproducible and quantitative, and can be used to follow the course of IBD. Technical issues need to be addressed. Measurement of pancreatic enzyme activity in stool is less predictive of pancreatic exocrine insufficiency than more traditional pancreatic function tests that involve duodenal intubation. At best, reduced fecal enzyme activity is only suggestive of a diagnosis of exocrine pancreatic insufficiency when steatorrhea is present.
Factitious diarrhea because of laxative ingestion remains a difficult clinical problem. It occurs in several scenarios that must be recognized to facilitate diagnosis (Table 8). In theory, stool analysis can discover magnesium, phosphate, sulfate, polyethylene glycol, senna, and bisacodyl abuse. Urine testing can be used to find absorbed stimulant laxatives, such as senna or bisacodyl. The performance of commercial stool assays for laxatives is not ideal. Dilution of a stool sample with water can be detected by measuring stool osmolality; because there is no mechanism to dilute luminal contents below isotonicity, measured stool osmolality below 290 mOsm/kg is due to addition of water or dilute urine to stool. Addition of urine to stool can be discovered by finding substantial amounts of creatinine in stool water.
Table 8. Groups of patients with laxative abuse (2)
Usually adolescent to young adult women; concerned about weight or manifesting an eating disorder; may binge eat, vomit, or exercise excessively to neutralize excessive food intake
Patients may have disability claim pending; illness may induce concern or caring behavior in others
Patients who relish being a diagnostic challenge; may undergo extensive testing repeatedly
Polle's syndrome (Munchausen syndrome by proxy)
Dependent child or adult poisoned with laxatives by parent or caregiver to show effectiveness as caregiver; may have history of sibling who died with chronic diarrhea
Hydrogen breath tests (HBTs): The recognition that hydrogen (H2) is produced in mammals only as a result of bacterial metabolism of carbohydrate has led to the development of novel technologies to detect malabsorption of carbohydrates and/or small bowel bacterial overgrowth. Bacterial metabolism of carbohydrate may occur under two circumstances: (i) if a carbohydrate normally absorbed by the small intestine is not absorbed and passes into the colon, where bacteria ferment the carbohydrate to short-chain fatty acids, CO2, CH4, and H2; and (ii) if there is a significant increase in luminal bacteria in the small intestine that degrades nutrients before they can be absorbed, again producing short-chain fatty acids, CO2, and H2. Under either scenario, the CH4 and H2 diffuse across the mucosa into the blood stream and ultimately are excreted by the lungs. Hydrogen and CH4 in the exhaled breath can be easily quantified by a mass spectrometer.[54, 133]
The first clinical application of breath hydrogen tests was to diagnose malabsorption of lactose. Lactase deficiency leads to the failure to hydrolyze lactose into its component monosaccharides, glucose, and galactose, with passage of lactose into the colon and generation of H2 or CH4 by colonic bacteria. Other sugars and sugar alcohols, such as fructose, lactulose, or sorbitol, can be used as substrates and yield information about malabsorption or small bowel transit time. Technical issues include the type of substrate given, the dose of substrate, time course of sampling, and criteria for a positive test.
Breath hydrogen testing also can be used to assess small intestine bacterial overgrowth (SIBO). This condition is associated with anatomic, functional, or motility abnormalities of the intestine, such as strictures, achlorhydria, motility disorders, or scleroderma. Symptoms related to SIBO include diarrhea, bloating, weight loss, malabsorption, and anemia. The diagnostic “gold standard” is quantitative culture of luminal fluid from small intestine. However, this test is uncommonly performed in clinical practice. Bacteria in the small intestine can ferment carbohydrate and different substrates for breath tests have been tested. The 14C-D-xylose breath test (with measurement of radioactive CO2 in the breath) has been the most rigorously tested and performs fairly well, but it is only available in selected academic centers. HBTs using glucose or lactulose as substrates, are the most commonly used tests for SIBO. The reported sensitivity and specificity for these tests varied from 27% to 93% and from 30% to 86%, respectively, for glucose HBT, and from 17% to 89% and 44% to 100%, respectively, for lactulose HBT (compared with quantitative small bowel cultures from proximal small bowel).[134-140] The wide range in sensitivity and specificity suggest potential problems with the reliability of these tests and pitfalls in basing clinical decisions on them.
Many factors may affect the accuracy of HBTs. For example, in addition to carbohydrate malabsorption, increased oral bacterial flora or failure to adhere to a low-fiber diet can result in false-positive tests. A false-negative result may also occur either because of recent or concurrent antibiotic treatment or because of the lack of H2-producing bacteria in an individual's microbiome. Variables in the test protocols, such as dosage of carbohydrate administered, method of collection, the amplitude of increase in H2 considered as positive test can all affect the results.[54, 141]
Rapid intestinal transit is the most important confounding variable in applying either glucose or lactulose breath hydrogen tests to the diagnosis of SIBO. Because lactulose is a non-absorbable disaccharide, it normally passes into the colon where it is quickly fermented. Normally, the rise in exhaled H2 concentration signals the arrival of lactulose in the cecum, and the time from ingestion of lactulose represents the oral-cecal transit time (OCTT). If the lactulose encounters bacteria in the small intestine an “early” or “double” peak pattern may be recognized and has been interpreted as indicating the presence of SIBO. The reliability of that interpretation has been questioned. A combined study linking a glucose or lactulose breath hydrogen test with a scintigraphic intestinal transit scan that provides an independent measure of OCTT may improve the diagnostic accuracy of HBTs for SIBO. Such an approach improved the specificity of a lactulose HBT from 70% to 100%, although the sensitivity was still limited.[139, 143] Therefore, the combination of a breath test and scintigraphy may make it feasible to improve the reliability of the HBT, with either lactulose or glucose. At this point in time, lactulose HBT by itself should not be used to diagnose SIBO.
Bile acid testing
In normal subjects, more than 95% of the bile acids secreted by the liver are reabsorbed in the ileum before reaching the cecum. When the enterohepatic circulation is disrupted, diarrhea may occur due to reduction of absorption or stimulation of secretion by excess bile acid in the colon.
Classically, three types of bile acid malabsorption (BAM) have been recognized as listed in Table 9. Type 1 malabsorption typically occurs in diseases of the ileum, most commonly CD, or after surgical resection of the ileum. In general, > 50 cm of the ileum needs to be lost before clinically significant BAM will occur. In this situation, malabsorbed dihydroxy bile acids, such as chenodeoxycholic acid and deoxycholic acid, inhibit colonic sodium absorption and stimulate chloride secretion causing diarrhea. Malabsorbed bile salt may also increase colonic permeability and motility, thereby providing another mechanism for diarrhea.
Type 2 bile acid malabsorption is also known as idiopathic bile acid malabsorption (IBAM); its prevalence is controversial. IBAM was thought to be a rare cause of chronic diarrhea but has been reported much more frequently since the introduction of SeHCAT as a diagnostic tool (see later). The prevalence of BAM in patients with chronic diarrhea ranges from 33% to 60% in several reports.[43, 145-149] In trying to sort out cause and effect, it is important to recognize that diarrhea induced in normal subjects can cause mild bile acid malabsorption, so some degree of BAM might result from just the presence of diarrhea.
Type 3 BAM is a grab-bag of diagnoses that have been associated with bile salt-induced diarrhea. Of these, the most common is post-cholecystectomy diarrhea occurring in 10–20% of individuals after removal of the gallbladder. Sometimes, this may be described as a loosening of stool consistency or an improvement in constipation rather than diarrhea. The pathophysiology is unclear. Because bile acids are no longer stored in the gallbladder, they may reside in the gut for a greater time and may be more subject to bacterial dehydroxylation that would increase production of diarrheogenic bile acids. Alternatively, altered motility may play a role. Colonic transit time has been shown to be decreased after cholecystectomy. The migrating motor complex may sweep intestinal content including bile acids rapidly through the ileum into the colon, leading to BAM type III. However, it is unclear how frequently BAM actually occurs after cholecystectomy.[152-154]
In most settings specific diagnostic tests for BAM are limited, and therefore, a confident diagnosis may be difficult to obtain. The 14C-glycocholate breath test has been abandoned because it is laborious and neither very sensitive nor specific. Direct measurement of fecal bile acid output involves complicated research methods not applicable to clinical use and does not predict response to bile acid binders in patients with chronic diarrhea.
In Europe measurement of whole body retention of a radioactive bile acid, SeHCAT (selenium-75-homocholic acid taurine), is the most widely available test for BAM. The radio-labeled bile acid is absorbed and then excreted at the same rate as native cholic acid. Following oral administration of the tracer, whole body counting using a standard gamma camera is done after 3 h to establish a baseline and again at 7 days to measure retention in the body. The results usually are expressed as percentage retained at 7 days, although half-life also can be calculated. Experts suggest that patients with chronic diarrhea who have SeHCAT retention of < 5% at 7 days (indicating severe bile acid malabsorption) often respond to bile acid binding drugs, whereas patients with retention of > 10% at 7 days (indicating more normal ileal bile acid absorption) are less likely to respond to those drugs.
There has been some controversy over the accuracy of SeHCAT in the diagnosis of BAM. There is a wide variation of normal and abnormal ranges in different centers and variable false-positive and false-negative rates.[156-160] Some have suggested that the SeHCAT test is of “no value” in the routine evaluation of chronic diarrhea. Although that may be an extremely critical view, it is important to recognize the limitations of the test. Given that BAM may be a manifestation of diarrhea instead of the cause, it is important to recognize that SeHCAT cannot delineate primary and secondary BAM related to diarrhea. Nevertheless, it has gained wide spread acceptance in Europe (the radioisotope is not available in the United States; consequently, American physicians have no experience with this test).
The development of a simple blood test for bile acid malabsorption would be a major advance. Normally serum bile acid levels are very low because of efficient clearance of bile acid from portal blood by the liver and so it would be impossible to detect lower than normal levels that might result from bile acid malabsorption. When bile acid is malabsorbed by the ileum, however, the liver compensates by increasing synthesis, which is reflected by increased serum levels of C4 (7α-hydroxy-4-cholesten-3-one), a precursor in the bile acid synthetic pathway and a reliable reflection of CYP7A1 activity, the rate limiting step in bile acid synthesis. While increased serum C4 is consistent with ileal bile acid malabsorption, this assay needs further validation before it can be accepted for widespread use as a measure of BAM.
For physicians outside Europe response to empirical treatment with bile acid-binding drugs, such as cholestyramine, colestipol, or colesevelam, may be a simpler “diagnostic” test. Failure to respond to a therapeutic trial of bile acid binders makes BAM an unlikely cause of diarrhea. However, if patients respond only to a large amount of bile acid-binder (e.g. more than 12 g of cholestyramine), it is unclear whether this should be considered a positive or a negative trial. One must recognize the possibility of a falsely concluding that BAM is present because of the well-recognized constipating effect of bile acid-binders.
Pancreatic function tests
The main reason to test pancreatic exocrine function in a patient with diarrhea is to determine whether pancreatic exocrine insufficiency is the cause of diarrhea. Historically, this was done by direct testing: intubating the stomach to remove gastric acid and intubating the duodenum to recover duodenal contents after stimulating pancreatic secretion with secretin or a combination of secretin and CCK (secretin test or secretin-CCK test). The test was cumbersome both for the patient and technician, required fluoroscopy to position the tubes properly, was poorly standardized, and was plagued by intermittent shortages of the stimulants. Its performance characteristics were excellent, however, with 95% sensitivity when the patient had advanced chronic pancreatitis. Intubation testing survives in its original form at a few research centers and has been modernized by collecting fluid directly from the pancreatic duct at the time of endoscopic retrograde cholangiography or EUS.[162, 163]
Another form of direct testing does not involve intubation but instead looks at fecal excretion of pancreatic enzymes. Measurement of fecal elastase 1, chymotrypsin, and lipase has been used for this purpose.[162, 164] Fecal elastase 1 has the best performance characteristics but is far from perfect, working best when the pretest probability of pancreatic exocrine insufficiency is high. Test performance characteristics are such that these tests should not be used for screening purposes.
One group of indirect tests of pancreatic function relies on looking for the biochemical effects of maldigestion of natural or synthetic substrates. For example, fat or nitrogen excretion in stools collected for 48–72 h will be elevated in patients with pancreatic exocrine insufficiency. These tests will be abnormal in patients with advanced pancreatic exocrine insufficiency but may not reflect lesser degrees of pancreatic exocrine dysfunction. They also may be abnormal in any cause of malabsorption and therefore are not specific for pancreatic exocrine insufficiency. Acceptability is limited because these tests involve quantitative stool collection over 2 or 3 days. Analyses of spot samples of stool for qualitative fat excretion (using Sudan stain) or for acid steatocrit have been proposed as alternatives, but sensitivity and specificity are reduced.[162, 164, 166] These excretory tests can be used to follow the course of treatment but should not be used to make a diagnosis.
The other group of indirect tests involves ingestion of synthetic substrates that can be cleaved by pancreatic enzymes and measures recovery of the hydrolysis products in the urine or breath. The best studied of these tests is the bentiromide (N-benzoyl-L-tyrosyl-paraaminobenzoic acid [NBT-PABA]) test. Chymotrypsin hydrolyzes the substrate, and PABA is recovered in the urine. In advanced pancreatic insufficiency, the sensitivity is 80–90% but is much lower (37–46%) with less severe impairment, making it no better than the flip of a coin for diagnosis. The substrate is no longer available in the United States. Similar results have been seen with the dual-label Schilling test and fluorescein-dilaurate (pancreolauryl) test. In contrast, good correlation of test results with fecal fat excretion has been noted with 13C-labeled triglyceride breath tests.[167-169] Unfortunately, these tests are abnormal with any cause of fat malabsorption because production of 13CO2 depends not only on hydrolysis but also absorption and metabolism of the labeled fatty acid.
At present, there is no single, simple, reliable, and accurate test for pancreatic exocrine insufficiency. Advanced disease is not difficult to diagnose with imaging studies, and these largely have supplanted other forms of testing in most centers. MR imaging of the pancreas before and after secretin is being evaluated as a test that would combine both structural and functional elements and eventually may prove to be useful. Currently, diagnosis of pancreatic exocrine insufficiency depends on clinical intuition based on history, imaging of the pancreas, supportive evidence from indirect testing for the consequences of maldigestion (e.g. steatorrhea), and the results of a therapeutic trial of pancreatic enzyme replacement.
The definition of chronic diarrhea is still a matter of some debate, with definitions of loose stool form, increased stool frequency or stool weight, and duration of symptoms at best arbitrary. It may be impossible (and perhaps unwise) to set more stringent criteria for clinical purposes, as ultimately, evaluation and management depends upon the patients' definitions of diarrhea as much as ours.
Similarly, classification schemes need to be judged by their utility in facilitating the care of individual patients who may not always fit into distinct categories. This is subject to investigation, however, and perhaps, this review will stimulate efforts to do so.
It would be ideal if the diagnostic evaluation and management of chronic diarrhea could be reduced to a simple algorithm. At present, this is impossible because of the number of possible diagnoses, limited appreciation of the pretest probabilities of these diagnoses (which depend in a critical way on geography and specifics of individual patients), variability of test availability and performance in different places, and the costs involved with diagnostic evaluations. In the future, research—and advances in our ability to aggregate information about these patients—may make this task easier. For now taking a good history, thinking over the more likely diagnoses and targeting testing is the best way forward.