Disorders of Gastrointestinal Motility Associated with Diabetes Mellitus MARK FELDMAN, M.D.; and LAWRENCE R. SCHILLER, M.D.: Dallas, Texas Gastrointestinal symptoms such as vomiting, constipation, diarrhea, and fecal incontinence occur frequently in patients with diabetes mellitus. In a survey of 136 diabetic outpatients, 7 6 % had one or more gastrointestinal symptoms, the commonest symptom being constipation (found in 6 0 % ) . In many cases these symptoms are thought to be due to abnormal gastrointestinal motility that, in turn, may be a manifestation of diabetic autonomic neuropathy involving the gastrointestinal tract. The pathophysiology of these gastrointestinal symptoms, clarified in recent studies, and the clinical features and treatment of these problems in diabetic patients are reviewed.

C O M P L I C A T I O N S involving the gastrointestinal tract cause considerable morbidity in patients with diabetes mellitus. Studies based on retrospective chart reviews have suggested a relatively low prevalence of gastrointestinal problems in patients with diabetes mellitus ( 1 , 2 ) . However, when patients attending diabetes clinics at our institutions were questioned specifically about gastrointestinal symptoms, 7 6 % had gastrointestinal symptoms (Table 1). In most patients these symptoms were chronic or frequently recurrent. Even asymptomatic patients with diabetes may have disturbed esophageal or gastrointestinal function. For example, diabetic patients may not have nausea, vomiting, or early satiety, yet may develop unexplained episodes of hypoglycemia as a result of abnormally delayed gastric emptying (3-5). In one prospective study ( 6 ) , 2 2 % of a group of asymptomatic patients with diabetes had radiologic evidence of gastric retention ( "gastroparesis diabeticorum" ). We review the clinical features, diagnosis, pathophysiology, and treatment of four of these gastrointestinal complications: gastroparesis diabeticorum, constipation, diarrhea, and fecal incontinence. Recent reports have led to a better appreciation of the clinical picture of these complications in patients with diabetes mellitus and have highlighted the relationship of these complications to diabetic autonomic neuropathy. Gastroparesis Diabeticorum CLINICAL F E A T U R E S A N D DIAGNOSIS

The commonest symptoms of gastroparesis diabeticorum are nausea and vomiting. Other symptoms in patients with gastroparesis include early satiety, bloating, and abdominal pain. Vomitus may contain food that had been ingested several hours earlier. Although some pa• F r o m the D e p a r t m e n t of Internal Medicine, Dallas V.A. Medical Center and University of Texas Health Science Center at Dallas; Dallas, Texas

378

tients vomit regularly, other have attacks of nausea and vomiting that last days or weeks and resolve gradually. During these attacks, glycemia becomes difficult to control and weight loss occurs. Between attacks, few if any symptoms are present. The prototype diabetic patient with symptomatic gastroparesis is one with long-standing, insulin-dependent diabetes that has been poorly controlled for many years. Peripheral neuropathy, manifestations of autonomic neuropathy (bladder dysfunction, sweat disorder, orthostatic hypotension, impotence), nephropathy, and retinopathy are frequently found ( 1 ) . However, there are patients in whom gastroparesis is the only diabetic complication, in whom diabetes is of fairly recent onset, or in whom diabetes is relatively easy to control. If gastroparesis is suspected, an upper gastrointestinal series should be done. Findings may include residual gastric secretions despite a prolonged fast, mild gastric dilatation, poor or no peristalsis, prolonged retention of barium, patulous pylorus, and atonic duodenal bulb ( 2 ) . These findings resemble those seen after truncal vagotomy ( 7 ) . To exclude pyloric obstruction as a cause of poor emptying of barium from the stomach, it may be necessary to compress the abdomen manually to express barium through the patent pylorus or, alternatively, to do an endoscopic examination. Gastric scanning after ingestion of a radiolabeled meal appears to be a more sensitive diagnostic test than an upper gastrointestinal series (810). It is possible to quantitate emptying of the liquid component or the solid component of the meal (or both) with this technique. PATHOPHYSIOLOGY

The healthy stomach empties liquids, digestible solids, and indigestible or poorly digestible solids. Liquids empty most rapidly, whereas solid food is first dispersed into very small particles before emptying. Indigestible solids are, for the most part, evacuated from the stomach long after ingestion of the meal, during the interdigestive (fasting) periods (11). The emptying of liquids from the stomach is proportional to the pressure gradient between the proximal stomach (fundus) and the duodenum. This pressure gradient is generated by tonic contractions of the fundic musculature. Interventions that increase proximal stomach contractions in dogs (such as the infusion of motilin) speed liquid emptying, whereas interventions that decrease proximal stomach contractions (such as the infu-

Annals of Internal Medicine. 1983;98:378-384.

Downloaded From: http://annals.org/ by a Penn State University Hershey User on 05/17/2016

©1983 American College of Physicians

sion of cholecystokinin) usually slow the emptying of liquids from the stomach (12). Diabetics patients with symptomatic gastroparesis tend to empty liquids slower than normal (13), but this difference is not striking (Figure 1). This fact suggests that patients with diabetes may have a relative inability to raise intragastric pressure by fundic contraction but that this dysfunction is mild. However, this dysfunction has not been studied directly. The emptying of both digestible and indigestible solids depends on motor activity of the distal stomach (antrum) (12). Radiologic studies show that antral peristalsis is disordered in diabetic gastroparesis ( 2 ) . To understand studies investigating this aspect of gastric motor function, it is necessary to review the current concept of gastric myoelectric activity. Smooth muscle cells generate a transmembrane potential between the inside and outside of the cell. This transmembrane potential is not constant in the distal stomach but fluctuates on a regular basis as a result of changes in ion conductance (Figure 2 ) . In the stomach, an area along the greater curvature high in the corpus has the fastest frequency of spontaneous depolarization ( 3 / m i n ) and entrains the more distal stomach, which ordinarily tends to spontaneously depolarize more slowly. This electrical activity is called the slow wave or pacesetter potential (Figure 2, top) and migrates from the pacemaker in the corpus distally through the stomach to the pylorus. When a threshold potential is reached, further depolarization (action potential or spike activity) appears on the plateau of the slow wave (see fourth slow wave, Figure 2). When this spike activity occurs, calcium ions enter the smooth muscle cell and the muscle contracts (Figure 2, bottom). Thus, contractions occur at a rate anywhere from zero (never a spike on a slow wave) to 3/min (always a spike) and travel distally through the stomach paced and spaced by the slow wave. Whether spike activity occurs on the slow wave as it passes a particular site in the stomach is determined by stimulatory and inhibitory nerves and hormones. If electrical activity is recorded for several hours by an electrode in the stomach or small intestine during fasting (14), a cyclic pattern of spike activity emerges (Table 2 ) . Phase 1 is characterized by slow wave activity but no spike activity. During phase 2, spike activity is recorded randomly on some but not all of the slow waves. Phase 3 is characterized by intense spike activity associated with every slow wave and lasts for 5 to 10 minutes. Phase 3 is usually followed by a brief phase 4, during which only Table 1. Gastrointestinal Symptoms in 136 Diabetic Outpatients Symptom

Outpatients

n(%) Constipation Abdominal pain Nausea and vomiting Dysphagia Diarrhea Fecal incontinence N o gastrointestinal symptoms

82 46 39 37 30 27 32

(60) (34) (29) (27) (22) (20) (24)

Figure 1 . Mean volume of fluid (meal plus secretions) emptied from the stomach after a 7 0 0 ml_ intragastric liquid meal ( 3 % casein hydrolysate) in 16 subjects without diabetes and 9 diabetic patients with chronic nausea and vomiting (symptomatic gastroparesis). The volume differences are due to differences in gastric emptying because secretion in response to the meal was similar in the two groups. From Feldman M, Seeman I, Schiller LR, unpublished data.

occasional spikes are recorded. The onset of phase 1 (no spikes) begins the next 2-hour cycle, and these cycles continue as long as the subject fasts (15). Ten years ago, Szurszewski (14) saw that a band of phase 3 myoelectric activity migrated aborally through the stomach and down the small intestine during fasting. By the time the phase 3 complex stops at the ileum, another "activity front" begins simultaneously in the stomach and duodenum. Similar interdigestive migrating myoelectric complexes and motor complexes occur in humans (16). Studies using barium or plastic spheres have shown that during the phase 3 activity front, luminal contents are swept forcefully by peristalsis out of the stomach down the small intestine toward the colon (11, 17). Proposed functions of interdigestive migrating motor complexes are to sweep stomach and upper small bowel free of indigestible debris, to keep the stomach and upper gut relatively sterile, and to "exercise" gastric and small intestinal smooth muscle during fasting. Dysfunction of this motor complex may lead to a failure to clear the stomach of debris and therefore the development of bezoars (18), bacterial overgrowth of the stomach and small intestine (15), and gastric and small intestinal atony. Malagelada and associates (19) studied fasting diabetic patients with symptomatic gastroparesis, asymptomatic diabetic patients, and healthy controls. Pressure transducers were swallowed and positioned in the gastric fundus, gastric antrum, and duodenum. Whereas healthy subjects and asymptomatic patients with diabetes averaged two migrating motor complexes (phase 3) per 180 minutes, most diabetics with gastroparesis had no such complexes. A similar abnormality was seen in patients with peptic ulcers who had gastroparesis after truncal vagotomy and antrectomy. Although metoclopramide (a dopamine antagonist with cholinergic effects) and bethanechol (a cholinergic agonist) significantly increased motor activity in the fundus and antrum of diabetic patients, Feldman and Schiller

Downloaded From: http://annals.org/ by a Penn State University Hershey User on 05/17/2016

• Gastrointestinal Motility

379

It must be emphasized that diabetic ketoacidosis itself also may cause nausea, vomiting, and abdominal pain (22). In this setting, acute gastric dilatation and gastroparesis may develop and resolve rapidly as ketoacidosis clears. In addition, some patients with diabetic ketoacidosis develop acute hemorrhagic gastritis that can cause vomiting, abdominal pain, or upper gastrointestinal bleeding ( 2 3 ) . TREATMENT

Figure 2. Transmembrane potential in the stomach {top) and gastric motor activity {bottom). Slow wave frequency is 3 / m i n . Only when an action potential (spike) occurs on the plateau of a slow wave does motor activity (muscle contraction) occur.

these drugs did not induce phase 3 activity. Fox and Behar (20) also reported that phase 3 complexes were absent in stomachs of diabetic patients with gastroparesis but were found in asymptomatic patients with diabetes. Whether analogous antral muscle dysfunction occurs postprandially in diabetic patients with gastroparesis is uncertain. It seems likely that antral motor activity after eating would be abnormal in these patients because emptying of digestible solids is delayed ( 8 ) , and antral motor activity is thought to be responsible for the normal emptying of solids (11, 12). However, postprandial antral motor activity has not been studied directly. The pathogenesis of diabetic gastroparesis remains uncertain. Because pharmacologic agents can increase motor activity, faulty regulation rather than unresponsive muscle probably underlies these defects. Current evidence supports the notion that vagal autonomic neuropathy may be responsible for gastric motor disturbance in diabetic patients with gastroparesis. In addition to the radiologic and gastric manometric similarities between patients with diabetes and patients who have had vagotomies, there is also histologic (21) and physiologic evidence of vagal dysfunction in diabetic patients with gastroparesis (13). For example, patients with diabetes with chronic vomiting (and also asymptomatic patients with diabetes) have markedly reduced acid secretory responses to sham feeding but near normal responses to pentagastrin (Figure 3). Because sham feeding normally elicits acid secretion via vagal stimulation, the data in Figure 3 suggest that these patients have a type of selective vagal defect (autovagotomy). Diabetic patients also have elevated serum gastrin concentrations (13) that are characteristic of patients who have had vagotomies. However, because the prevalence of autovagotomy in diabetic patients appears to be high, and clinical evidence of gastroparesis is less common, there may be additional factors which, in combination with autovagotomy, lead to episodes of intractable nausea and vomiting. These may include other aberrant regulatory factors (intrinsic nerves, hormones), uncontrolled hyperglycemia, increased sensitivity of the chemoreceptor trigger zone or vomiting center in the medulla, emotional disturbances, or other factors. 380

Treatment of symptomatic diabetic gastroparesis is often difficult and frustrating. If the patient has relatively mild attacks with long symptom-free intervals, pharmacologic therapy can be reserved for use at the onset of the attacks. Phenothiazines such as prochlorperazine (either by mouth, injection, or suppository) can be used to suppress nausea and vomiting. The patient with continuous symptoms or frequent attacks needs more aggressive therapy and may have to be hospitalized. The initial phase of hospital care should be devoted to adequate fluid and electrolyte repletion, control of hyperglycemia with insulin, and nasogastric suction to "decompress" and empty the stomach (24). If the patient remains symptomatic on refeeding, pharmacologic therapy may have to be used. At present, the drug of first choice is metoclopramide, which accelerates gastric emptying in most patients (10). Metoclopramide not only increases gastric motility, but also acts centrally in the chemoreceptor trigger zone as an anti-emetic (25). Doses of 5 to 20 mg 30 minutes before meals and at bedtime are used. The usual dose is 10 mg four times a day. When a patient first starts taking metoclopramide, the drug may have to be given parenterally for a day or two because it must be emptied from the stomach to be absorbed when given orally. Up to 10% of patients treated with metoclopramide develop neurologic symptoms ranging from nervousness or somnolence to frank extrapyramidal symptoms such as torticollis. The drug also produces elevated serum prolactin levels and rarely may cause galactorrhea. If metoclopramide is ineffective, bethanechol or a cholinesterase inhibitior such as ambenonium are sometimes helpful ( 2 ) . Surgery is a last resort. Pyloroplasty, gastrojejunostomy, and antrectomy have been attempted but usually with only limited success (4, 26). Constipation CLINICAL F E A T U R E S A N D DIAGNOSIS

In his classic study of diabetic neuropathy, Rundles (1) found that the commonest gastrointestinal symptom Table 2. Phases of Spike Activity in Stomach and Small Bowel During Fasting Phase

1 2 3 4

March 1983 • Annals of Internal Medicine • Volume 98 • Number 3

Downloaded From: http://annals.org/ by a Penn State University Hershey User on 05/17/2016

Approximate Length min 60-90 20-30 5-10 0-10

Spike Activity

None Random With every slow wave Random

Figure 3. Mean ( ± SE) acid output under basal conditions, in response to vagal stimulation induced by sham feeding, and in response to a maximally effective dose of pentagastrin in 1 1 normal subjects without diabetes (/V), 10 diabetic patients without nausea or vomiting (D), and 8 diabetic patients with chronic nausea and vomiting (DV). Significant differences in diabetic patients compared to normal subjects are shown by asterisks. Diabetics had a reduced response to sham feeding but a near normal response to pentagastrin, suggesting a selective vagal defect. Adapted f r o m data reported by Feldman and associates ( 1 3 ) .

was constipation. Severe constipation, often requiring use of regular enemas, occurred in 2 0 % of his patients with neuropathy. Subsequent studies have found that the prevalence of constipation is greater in patients with diabetes mellitus than in an age- and sex-matched control population (27). The striking prevalence of constipation in patients attending two diabetes clinics in Dallas is shown in Table 1. Men and women had a similar prevalence of constipation. The prevalence of constipation rose as patients had more symptoms of neuropathy: Only 2 9 % of patients with no symptoms of neuropathy were constipated, whereas 8 8 % of those with five symptoms of neuropathy (orthostatic symptoms, numbness or burning in hands and feet, sexual dysfunction, urinary incontinence, and sweating disorders) were constipated. Constipation was typically intermittent and alternated with diarrhea in 24 of the 82 patients ( 2 9 % ) . When a diabetic patient is constipated, the extent of the evaluation depends on several factors: severity and duration of constipation, age of the patient, and associated symptoms (weight loss, rectal bleeding). In most patients, digital rectal examination and proctosigmoidoscopy, including testing of stool for occult blood, is sufficient. In some patients, barium enema or colonoscopic examination may be necessary to exclude other pathologic findings, in particular, carcinoma of the colon. The diagnosis of irritable bowel syndrome should be made cautiously in a patient with diabetes.

activity was normal in patients with diabetes, the myoelectrical and colonic motor responses to the meal were blunted and delayed in the diabetic patients with little or no constipation (Figure 4, left) and were absent in the diabetic patients with severe constipation (Figure 4, right). This latter pattern is similar to that seen in patients with chronic idiopathic intestinal pseudo-obstruction (29). Although the diabetic patients did not respond normally to an eaten meal (Figure 4 ) , they did respond appropriately to parenteral injections of neostigmine or metoclopramide. This finding suggests that the colonic muscle of constipated diabetic patients is capable of responding to exogenous stimulation but that the ingestion of a meal does not trigger the appropriate neurohormonal signals to the colon. Whether this defect is the cause or merely a result of constipation in patients with diabetes mellitus is unclear. TREATMENT

Therapy of constipation in diabetic patients is usually symptomatic. Many patients will already be using laxatives, cathartics, or enemas before they seek medical attention. Recent anecdotal reports suggest that metoclopramide may be useful in some patients. In a controlled trial of metoclopramide in diabetic patients with gastroparesis, co-existent constipation improved while the patients received metoclopramide (10). Additional, prospective studies are needed to evaluate the effect of metoclopramide therapy for diabetic patients with constipation. Moreover, more must be known about the risks of long-term metoclopramide therapy before it can be recommended for this use. Other treatments, such as high fiber diets and anorectal myectomy, have not been extensively evaluated in diabetic patients. Diarrhea CLINICAL FEATURES AND DIAGNOSIS

According to our survey, 2 2 % of diabetic outpatients had diarrhea (loose stools or frequency) at least 2 to 3 days per year. As already mentioned, many of these patients also had intermittent constipation: 24 of 30 patients with diarrhea also had episodes of constipation. Al-

PATHOPHYSIOLOGY

Because the pathogenesis of constipation in patients with diabetes mellitus was not well understood, Battle and associates (28) recently did an extensive pathophysiologic study. The authors measured basal and postprandial myoelectrical and motor activity from the distal colon in diabetic patients with severe constipation (defined as two or less spontaneous stools per week) as well as in diabetic patients with little or no constipation. As shown in Figure 4, a meal normally leads to a rapid and large increase in electrical spike activity (and also in motor activity) in the colon. Although basal spike (and motor)

Figure 4. Colonic myoelectrical activity, expressed as spikes per 3 0 minutes, basally and after a meal in normal subjects without diabetes, diabetic patients with little or no constipation, and diabetic patients with severe constipation. The response to the meal was delayed in diabetic patients with no constipation (left) and absent in diabetic patients with constipation (right). Adapted from data reported by Battle and associates ( 2 8 ) . Feldman and Schiller

Downloaded From: http://annals.org/ by a Penn State University Hershey User on 05/17/2016

• Gastrointestinal Motility

381

though nocturnal diarrhea is supposed to be common in diabetic diarrhea, we found that diarrhea occurring only at night was distinctly unusual. For example, diarrhea sometimes occurred at night in 21 of 30 patients with diarrhea, but it occurred only at night in 2 patients. Fecal incontinence was common in patients with diarrhea in our survey: 13 of 30 such patients also had a history of incontinence. The diagnostic evaluation of diarrhea in patients with diabetes must be particularly artful because of several diagnostic traps. From the start, the physician should be certain that he or she understands exactly what the patient means by diarrhea. Of the three cardinal signs of diarrhea—frequent stools, loose consistency of stools, and increased stool volume (weight), patients are most attuned to the first two. Thus, patients may complain of diarrhea when they do not have an increased stool volume (greater than 200 g / d ) . The importance of this is that low volume diarrhea may have different causes than high volume diarrhea. For instance, a patient with the frequent passage of small volumes of semiformed stool may have anal sphincter dysfunction and fecal incontinence rather than the usual causes of diarrhea. Because a patient's impression of daily stool volume is apt to be incorrect, a 24-hour stool collection for weight and fat measurement should precede an extensive evaluation. The differential diagnosis of increased stool weight in diabetic patients includes any of the conditions that can cause acute and chronic diarrhea in patients without diabetes and, in particular, certain chronic disorders that occur with increased frequency in patients with diabetes, such as celiac sprue (gluten-induced enteropathy) (30), pancreatic insufficiency (31), and bacterial overgrowth of the small intestine (32). Only when these conditions are excluded should the diagnosis of idiopathic diabetic diarrhea be made. Celiac sprue can be excluded by examining jejunal biopsy specimens. Pancreatic insufficiency must be considered, especially when steatorrhea is found, and should be excluded by appropriate pancreatic function tests or by a trial of oral pancreatic enzyme therapy. Bacterial overgrowth can sometimes be diagnosed by quantitative bacterial culture of intestinal contents or by antibiotics. If alternative causes are excluded, the most likely explanation for chronic or intermittent diarrhea in a diabetic patient is idiopathic diabetic diarrhea. However, this diagnosis must be made by exclusion because the clinical features and radiologic findings are not specific. For example, nocturnal diarrhea can occur in patients with other disorders, and the roentgenographic appearance of the small bowel in diabetic patients with diarrhea (prolonged transit through the intestinal tract, segmentation of barium, localized segmental dilatation, abnormal fold patterns) can occur in patients with other diseases, as well as in diabetic patients with constipation rather than diarrhea. PATHOPHYSIOLOGY

Although the cause of idiopathic diabetic diarrhea is not known, the most popular explanation is that it is a 382

manifestation of autonomic neuropathy. This theory is supported by the observation that both truncal vagotomy and sympathectomy may cause diarrhea in humans. Truncal vagotomy can also cause mild steatorrhea (33). Moreover, there is histologic evidence of abnormal vagal and sympathetic nerve anatomy in some patients (21, 3438), although one study did not find these changes (39). Several studies (40-42) have found that the small bowel of diabetic patients is less sensitive to distention-induced pain, suggesting a defect in afferent nerve function. The role of bacterial overgrowth in the pathogenesis of diarrhea in diabetes is controversial. The following sequence has been proposed. Because of a motility disorder, there is gastric and small intestinal stasis; as a result, bacteria overgrow in the stomach and upper small intestine (32, 43); bacteria in the lumen of the upper small intestine deconjugate bile salts, making them unavailable for micelle formation (44); and fat malabsorption, steatorrhea, and diarrhea result. Although this schema is attractive, it is not supported by every study. For example, one study found no evidence of abnormal bacterial growth in the upper small intestine of diabetic patients with diarrhea (40). Moreover, not every diabetic patient with diarrhea has steatorrhea. Evidence used to support the importance of bacterial overgrowth in diabetic diarrhea is the observation that diarrhea sometimes responds dramatically to antibiotic therapy. For example, Malins and French (45) reported that chlortetracycline controlled diarrhea in 16 of 22 diabetic patients. However, because diabetic diarrhea is sometimes an intermittent condition, it is possible that a response in these patients would have occurred without antibiotics. Only controlled, double-blind studies can settle this issue. Motility studies of the small intestine in diabetic diarrhea are not extensive and are contradictory (40-42). Some studies have found that diabetic patients with diarrhea have slow small-bowel transit times, whereas others have reported that transit is more rapid than normal. It is possible that there are two groups of patients: one group with slow transit predisposed to bacterial overgrowth, and a second group with rapid transit leading to diarrhea due to intestinal hurry. The interdigestive migrating myoelectric complex in the small intestine has not been studied in diabetic patients with or without diarrhea. TREATMENT

At present, treatment of idiopathic diabetic diarrhea is largely symptomatic and only moderately successful. Opiates such as codeine, diphenoxylate, or loperamide can be used, but results may be disappointing. Metoclopramide, anticholinergics, and psyllium hydrophilic mucilloid (Metamucil powder; Searle Laboratories, Chicago, Illinois) have also been used in uncontrolled trials. It seems prudent to administer a 2-week course of antibiotics (for example, tetracycline, 500 to 1000 m g / d ) to any diabetic patient with unexplained chronic diarrhea, especially if steatorrhea is also present. Some patients may need intermittent or chronic antibiotic therapy to control symptoms. However, often the best one can hope for is a temporary remission.

March 1983 • Annals of Internal Medicine • Volume 98 • Number 3

Downloaded From: http://annals.org/ by a Penn State University Hershey User on 05/17/2016

Fecal Incontinence CLINICAL FEATURES AND DIAGNOSIS

Fecal incontinence (accidental passage of stool) is usually regarded as evidence of particularly severe diarrhea by both patients and physicians. However, accumulating evidence suggests that this is not necessarily so. Instead, fecal incontinence in many patients should be viewed as an independent problem with its own clinical features (46). Fecal incontinence is a common (Table 1) but often unmentioned problem in patients with diabetes. Almost half of the diabetic patients with diarrhea in our survey had incontinence when specifically asked. Interestingly, 10% of diabetic outpatients surveyed claimed to have had episodes of fecal incontinence but denied chronic diarrhea. A recent study attempted to define the clinical features of fecal incontinence in 16 diabetic patients with chronic diarrhea ( 4 7 ) . Diarrhea began approximately 10 years after the initial diagnosis of diabetes mellitus in this group. Incontinence began at about the same time as diarrhea in 12 patients and within 6 years of the onset of diarrhea in the other 4 patients. All of these patients had extragastrointestinal symptoms of autonomic neuropathy. Although nocturnal incontinence is thought to be characteristic of diabetic diarrhea, incontinence occurred exclusively at night in only 3 of these 16 patients. Incontinence occurred only with loose stools in all but one of these patients. Diagnosis of fecal incontinence rests on obtaining an accurate history. Because patients may not volunteer this potentially embarrassing complaint, the physician should ask directly about accidental passage of stool. Objective tests of anal sphincter pressures and of continence for both solids and liquids are available and can substantiate and quantify the suspected anal sphincter dysfunction (46). PATHOPHYSIOLOGY

The mechanisms controlling continence are only partially understood. Important factors seem to be the anatomic relationships of the pelvic musculature and the rectum, the strength of both internal and external anal sphincters, and the integrity of both the afferent (sensory) and efferent (motor) nerve supply to the region (48, 49). One early study (50) found that four incontinent diabetic patients had normal anal sphincter pressures, and suggested that these patients were incontinent because of abnormal rectal sensation (inability to sense the presence of stool in the rectum) rather than because of incompetent sphincter strength. A recent study (47) in 16 incontinent diabetic patients failed to substantiate this notion. Instead, mean basal anal sphincter pressure was significantly reduced compared to that of both healthy persons and continent diabetic patients (Figure 5). However, the mean increment in anal sphincter pressure that occurred during voluntary squeezing was not significantly different among these three groups. Because basal pressure is thought to be generated largely by the tonic contraction

Figure 5. Basal anal sphincter pressure in controls without diabetes, continent diabetic patients, and incontinent diabetic patients. Mean pressures are shown by horizontal lines. Incontinent diabetic patients had significantly lower mean basal pressures than controls without diabetes (p < 0 . 0 0 1 ) and than continent diabetic patients (p < 0 . 0 2 ) . Adapted from data reported by Schiller and associates (47).

of the internal anal sphincter (smooth muscle with autonomic innervation) and the squeeze increment is generated by the external anal sphincter (voluntary striated muscle with peripheral innervation) (48, 49), these results suggest that diabetic patients with fecal incontinence have a defect in internal anal sphincter function. Furthermore, these findings suggest that this abnormality is due to either defects in the autonomic innervation of the internal sphincter, or defects in the smooth muscle itself. A study such as this does not exclude the possibility that other defects contribute to incontinence in these patients. Indeed, the overlap in pressures shown in Figure 5 suggests that factors other than basal pressure must contribute to incontinence in certain persons. Several hypotheses may explain the co-existence and simultaneous development of fecal incontinence and chronic diarrhea in many of these patients. First, diabetic diarrhea may be so voluminous that it overwhelms the continence mechanisms. This was not the case in the study mentioned (47): mean ( ± SE) stool weight in incontinent diabetic patients was only 356 + 95 g/d. A second possibility is that continent diabetic patients have a latent defect in sphincter function that becomes clinically manifest only when diarrhea develops. This, too, was not confirmed in the study above: continent diabetic patients did not differ significantly from healthy subjects in their sphincter pressures (Figure 5) or in their ability to retain a solid test object, or to retain saline infused into the rectum. A third possibility is that a common defect, such as gastrointestinal autonomic neuropathy, is responsible for both diarrhea and fecal incontinence, with the combination of symptoms occurring when the neuropathy involves both the intestine and rectum. Finally, an isolated defect in anal sphincter function might cause loose stools and low-volume diarrhea by allowing premature emptying of the colon, before completion of normal colonic water absorption. TREATMENT

One approach to fecal incontinence is biofeedback training (51). However, the efficacy of this treatment has not been formally assessed in diabetic patients. PharmaFeldman and Schiller

Downloaded From: http://annals.org/ by a Penn State University Hershey User on 05/17/2016

• Gastrointestinal Motility

383

cologic therapy also has not been formally tested. Although at least one commonly used antidiarrheal agent, diphenoxylate with atropine (Lomotil; Searle and Co., San Juan, Puerto Rico), does not affect anal sphincter pressures (52), it seems reasonable to try to control diarrhea with antidiarrheal drugs in these patients in the hope that the continence mechanisms will be better able to cope with a smaller volume of more solid stool. Occasionally, diarrhea and incontinence will remit spontaneously, but permanent remission appears to be unusual. A C K N O W L E D G M E N T S : The authors thank Drs. Leonard Madison and Holbrook Seltzer for allowing them to survey diabetic clinic patients at Parkland Memorial Hospital and the Dallas V.A. Medical Center. The authors also thank Drs. Charles Richardson and John Fordtran for their advice and comments. Grant support: in part by grants AM 16816 and AM26794 from the National Institute of Arthritis, Metabolism and Digestive Diseases, and by a grant from the Abbe K. Dreyfuss Fund of the Southwestern Medical Foundation.

ic study. Gastroenterology. 21.

1980;78:757-63.

KRISTENSSON K, N O R D B O R G C, OLSSON Y, S O U R A N D E R P. Changes in

the vagus nerve in diabetes mellitus. Acta Pathol Microbiol Scand. 1971;79:684-5. 22. BEARDWOOD JT. The abdominal symptomatology of diabetic acidosis. JAMA. 1935;105:1168-72. 23.

C A R A N D A N G N, SCHUMAN BM, W H I T E H O U S E FW. The gastric muco-

sa of patients in diabetic ketoacidosis: a gastrocamera study [Abstract]. Diabetes. 1968; 17:319. 24. H O W L A N D WJ, D R I N K A R D W. Acute diabetic gastric atony. JAMA. 1963;185:214-6. 25.

P I N D E R RM, B R O G D E N RN, S A W Y E R PR, S P E I G H T TM, A V E R Y GS.

Metoclopramide: a review of its pharmacological properties and clinical use. Drugs. 1976;12:81-131. 26. ROON AJ, MASON GR. Surgical management of gastroparesis diabeticorum. Calif Med. 1972;116:58-61. 27. M A Y N E N. Neuropathy in the diabetic and non-diabetic populations. Lancet. 1965;2:1313-6. 28.

B A T T L E WM, S N A P E WJ J R , A L A V I A, C O H E N S, B R A U N S T E I N S.

Colonic dysfunction 1980;79:1217-21. 29.

• Requests for reprints should be addressed to Mark Feldman, M.D.; V.A. Medical Center (151), 4500 S. Lancaster Road; Dallas, TX 75216.

in

diabetes

mellitus.

Gastroenterology.

S N A P E WJ J R , SULLIVAN MA, C O H E N S. Abnormal gastrocolic re-

sponse in patients with intestinal pseudo-obstruction. Arch Intern Med. 1980;140:386-7. 30. THOMPSON MW. Heredity, maternal age, and birth order in the etiology of celiac disease. Am J Hum Genet. 1951;3:159-66. 31. F R I E R BM, S A U N D E R S JHB, W O R M S L E Y K G , B O U C H I E R IAD. Exoc-

rine pancreatic function 1976;17:685-91.

References 1. R U N D L E S RW. Diabetic neuropathy: general review with report of 125 cases. Medicine. 1945;24:111-60. 2. ZITOMER BR, G R A M M HF, ZOZAK GP. Gastric neuropathy in diabetes mellitus: clinical and radiologic observations. Metabolism. 1968; 17:199211. 3. CAMPBELL A, C O N W A Y H. Gastric retention and hypoglycemia in diabetes. Scott Med J. 1960;5:167-8. 4. W O O T E N RL, M E R I W E T H E R TW. Diabetic gastric atony: a clinical study. JAMA. 1961;176:1082-7. 5. G U P T A K K , H E D G E K P , L A L R. Diabetic gastric neuropathy with acute hypoglycemic attacks. J Indian Med Assoc. 1971;57:258-9. 6. KASSANDER P. Asymptomatic gastric retention in diabetics (gastroparesis diabeticorum). Ann Intern Med. 1958;48:797-812.

32.

33.

34. 35.

36.

7. H O D G E S FJ, R U N D L E S RW, H A N E L I N J. Roentgenologic study of the

small intestine: II. Dysfunction associated with neurological diseases. Radiology. 1947;49:659-73. 8. SCARPELLO JHB, BARBER DC, H A G U E RV, C U L L E N DR,

37.

SLADEN

GE. Gastric emptying of solid meals in diabetics. Br Med J. 1976;2:6713.

38.

DJ,

39.

CLARKE BF. Gastric emptying in diabetic autonomic neuropathy. Gut. 1977;18:462-7.

40.

9. C A M P B E L L IW, H E A D I N G RC, T O T H I L L P, B U I S T AS, E W I N G

10. S N A P E WJ, B A T T L E WM, S C H W A R T Z SS, B R A U N S T E I N SN, G O L D -

in juvenile-onset

diabetes mellitus. Gut.

G O L D S T E I N F, W I R T S CW, K O W L E S S A R O D . Diabetic diarrhea and

steatorrhea: microbiologic and clinical observations. Ann Intern Med. 1970;72:215-8. SHILS ME, G I L A T T. The effect of esophagectomy on absorption in man: clinical and metabolic observations. Gastroenterology. 1966;50:347-57. SMITH B. Neuropathology of the esophagus in diabetes mellitus. J Neurol Neurosurg Psychiatry. 1974;37:1151-4. APPENZELLER O, OGIN G. Myelinated fibers in human paravertebral sympathetic chain: white rami communicantes in alcoholic and diabetic patients. / Neurol Neurosurg Psychiatry. 1974;37:1155-61. W A R D JD. Pathogenesis of diabetic neuropathy. In: BAJAJ JS, ed. Proceedings of the Ninth Congress of the International Diabetes Federation. Amsterdam and Oxford: Excerpta Medica; 1977:649-55. D U C H E N LW, A N J O R I N A, W A T K I N S PJ, M A C K A Y J D . Pathology of

autonomic neuropathy in diabetes mellitus. Ann Intern Med. 1980;92:301-3. HENSLEY GT, SOERGEL KH. Neuropathology findings in diabetic diarrhea. Arch Pathol. 1968;85:587-97. BERGE KG, SPRAGUE RG, B E N N E T T WA. The intestinal tract in diabetic diarrhea: a pathologic study. Diabetes. 1956;5:289-94. W H A L E N GE, SOERGEL KH, G E E N E N JE. Diabetic diarrhea: a clinical and pathophysiological study. Gastroenterology. 1969;56:1021-32. M C N A L L Y EF, R E I N H A R D AE, SCHWARTZ PE. Small bowel motility in diabetics. Am J Dig Dis. 1969;14:163-9. D R E W E S VM. Mechanical and electrical activity in the duodenum of diabetics with and without diarrhea: pressures, differential pressures and action potential. Am J Dig Dis. 1971;16:628-34. SUMI SM, FINDAY JM. On the pathogenesis of diabetic steatorrhea. Ann Intern Med. 1961;55:994-7.

STEIN HA, A L A V I A. Metoclopramide to treat gastroparesis due to diabetes mellitus: a double-blind, controlled trial. Ann Intern Med. 1982;96:444-6. 11. K E L L Y KA. Gastric emptying of liquids and solids: roles of proximal and distal stomach. Am J Physiol. 1980;239:G71-6. 12. K E L L Y KA. Motility of the stomach and gastroduodenal junction. In: JOHNSON LR, ed. Physiology of the Gastrointestinal Tract. New York: Raven Press; 1981:393-410.

41.

13. F E L D M A N M, C O R B E T T DB, R A M S E Y EJ, W A L S H JH, R I C H A R D S O N

cocholate test in diabetic diarrhea. Br Med J. 1976;2:673-5. 45. MALINS JM, F R E N C H JM. Diabetic diarrhea. Q J Med. 1957;26:467-80.

CT. Abnormal gastric function in longstanding, insulin-dependent diabetic patients. Gastroenterology. 1979;77:12-7. 14. SZURSZEWSKI JH. A migrating electric complex of the canine small intestine. Am J Physiol. 1969;217:1757-63. 15. WEISBRODT NW. Motility of the small intestine. In: JOHNSON LR, ed. Physiology of the Gastrointestinal Tract. New York: Raven Press; 1981: 411-43. 16.

V A N T R A P P E N G, JANSSENS J, H E L L E M A N S J, G H O O S Y. The interdiges-

tive motor complex of normal subjects and patients with bacterial overgrowth of the small intestine. / Clin Invest. 1977;59:1158-66. 17. C O D E CF, SCHLEGEL JF. The gastrointestinal interdigestive housekeeper: motor correlates of the interdigestive myoelectric complex of the dog. In: D A N I E L EE, ed. Proceedings of the Fourth International Symposium on Gastrointestinal Motility. Vancouver, Canada: Mitchell Press; 1974: 631-4. 18. BRADY PG, RICHARDSON R. Gastric bezoar formation secondary to gastroparesis diabeticorum. Arch Intern Med. 1977;137:1729. 19.

42.

43. 44.

46.

384

14

C-gly-

R E A D NW, H A R F O R D WV, S C H M U L E N AC, R E A D MG, SANTA A N A

C, FORDTRAN JS. A clinical study of patients with fecal incontinence and diarrhea. Gastroenterology. 1979;76:747-56. 47.

SCHILLER LR, SANTA A N A CA, S C H M U L E N AC, H E N D L E R RS, H A R -

FORD WV, FORDTRAN JS. Pathogenesis of fecal incontinence in diabetes mellitus: evidence for internal anal sphincter dysfunction. TV Engl. J Med. 1982;307:1666-71. 48. SCHUSTER MM. Motor action of rectum and anal sphincters in continence and defecation. In: C O D E CF, ed. Handbook of Physiology. Washington, D.C.: American Physiological Society; 1968:2121-46. 49. CHRISTENSEN J. Motility of the colon. In: JOHNSON LR, ed. Physiology of the Gastrointestinal Tract. New York: Raven Press; 1981:445-71. 50. K A T Z LA, K A U F M A N HJ, SPIRO HM. Anal sphincter pressure characteristics. Gastroenterology. 1967;52:513-8. 51. C E R U L L I MA, N I K O O M A N E S H P, SCHUSTER MM. Progress in biofeed-

back conditioning 1979;76:742-6.

M A L A G E L A D A J-R, R E E S WDW, M A Z Z O T T A LJ, G O VLW. Gastric

motor abnormalities in diabetic and postvagotomy gastroparesis: effect of metoclopramide and bethanechol. Gastroenterology. 1980;78:286-93. 20. F o x S, BEHAR J. Pathogenesis of diabetic gastroparesis: a pharmacolog-

SCARPELLO JHB, H A G U E RV, C U L L E N DR, S L A D E N G E . The

52.

for

fecal

incontinence.

Gastroenterology.

H A R F O R D WV, K R E J S GJ, SANTA A N A CA, F O R D T R A N JS. Acute

effect of diphenoxylate with atropine (Lomotil) in patients with chronic diarrhea and fecal incontinence. Gastroenterology. 1980;78:440-3.

March 1983 • Annals of Internal Medicine • Volume 98 • Number 3

Downloaded From: http://annals.org/ by a Penn State University Hershey User on 05/17/2016