The occurrence of gastrointestinal complications in the critically ill patient is common.31 A multicenter study of 400 patients conducted in Spain in 1999 found that diarrhea complicated 15% of patients admitted to the ICU.32 A similar study of over 1300 ICU patients published 10 years later reported a 14% incidence of diarrhea.33 The occurrence of diarrhea continues to complicate the care of ICU patients and its management is an ongoing challenge, especially in the face of recommendations for earlier and more aggressive enteral feeding.34,35 This section will discuss the approach and management to diarrhea that develops in the critical care patient, followed by a separate discussion of Clostridium difficile.
CAUSES OF DIARRHEA IN CRITICAL CARE
While diarrhea can be classified into osmotic, secretory, infectious, or noninfectious, the causes of diarrhea in the critically ill patient can be simplified to those due to infection, medication, oral or enteral feeds, or preexisting intestinal disorders of absorption or motility. Most diarrhea in the ICU is acute in onset (<14 days) as opposed to persistent ≥14 days or chronic ≥4 weeks.36 Infection must always be considered and ruled out in any new-onset diarrhea (see Fig. 104-4 and Tables 104-3 and 104-4). A careful history including any collateral information from the patient’s family or close associates may reveal a preexisting condition such as lactose intolerance, celiac disease, inflammatory bowel disease, or irritable bowel syndrome. Investigation generally involves fecal specimen analysis for common bacterial and viral pathogens, and assessment for the presence of fecal toxins when infection with C difficile or enterotoxigenic/enterohemorrhagic bacteria is suspected. Routine testing for ova and parasites is not cost-effective for the majority of patients,37 but should be considered in the setting of persistent diarrhea, patients with a history of travel to a high risk area, and immunocompromised patients.36
Algorithm for evaluation of acute diarrhea. FODMAPs, fermentable, oligo-, di-, monosaccharides, and polyols.
Common Causes of Acute Infectious Diarrhea in Developed Countries
|Favorite Table|Download (.pdf) TABLE 104-3
Common Causes of Acute Infectious Diarrhea in Developed Countries
|Bacteria ||Salmonella, Campylobacter, Shigella, E coli 0157, Clostridium difficile |
|Viruses ||Caliciviruses (eg, norovirus), rotavirus, adenovirus, astrovirus |
|Protozoa ||Cryptosporidium, Giardia, Cyclospora, Entamoeba histolytica |
Major Causes of Chronic Diarrhea
|Favorite Table|Download (.pdf) TABLE 104-4
Major Causes of Chronic Diarrhea
|Osmotic || |
|Fatty || |
Malabsorption (mucosal disease, short bowel, bacterial overgrowth, ischemia)
Maldigestion (pancreatic exocrine insufficiency, inadequate bile)
|Inflammatory || |
Inflammatory bowel disease (ulcerative colitis, Crohn, diverticulitis)
Infectious (C difficile colitis, invasive bacteria, or ulcerating viruses)
|Secretory || |
Bile salt malabsorption
Drugs and poisons
Dysmotility (postvagotomy, postsympathectomy, diabetic autoimmune, hyperthyroidism, irritable bowel)
Inflammatory bowel disease (ulcerative colitis, Crohn, lymphocytic, collagenous colitis, diverticulitis)
Neoplasia (villous adenoma, colon cancer, lymphoma)
Medications are another cause of diarrhea by a variety of mechanisms. Antibiotics are frequently associated with diarrhea by altering the colonic flora, and laxatives and prokinetics increase intestinal motility. Acid-suppressive medications also have an inherent propensity to cause diarrhea (up to 7% of proton-pump inhibitors) and many oral electrolyte formulations or antacids are known irritants to the gastrointestinal mucosa (magnesium, phosphates). A study of 27 ICU patients treated for constipation showed 70% of them subsequently developed diarrhea,38 and another study showed diarrhea resolved in over 25% of patients following the discontinuation of laxative therapy.39 Many oral medications are hyperosmolar and/or contain sorbitol which can cause GI intolerance especially when given in large volumes. Sorbitol is a sugar alcohol that is used as a sweetener in many oral liquid medications and is known to cause osmotic diarrhea and cramping when ingested in amounts over 10 to 20 g in healthy volunteers.40 The amount of sorbitol is often not specified on medication labels as it is an inactive ingredient, and thus the amount of sorbitol being delivered to a patient is often difficult to determine.
MANAGEMENT OF DIARRHEA IN CRITICAL CARE
Fluid and electrolyte repletion is an important initial therapy as large-volume diarrhea can quickly lead to significant fluid, electrolyte, and acid-base disturbances; repletion should be accomplished via the intravenous route until the etiology of the diarrhea is determined.
The patient’s medication list should be examined for causative agents and these should be discontinued or substituted with alternative medications or routes of administration when appropriate. Sorbitol-containing liquid medications should be discontinued and sorbitol-free formulations or crushed tablets used when available. Dilution of any necessary hyperosmolar medications should be considered.41,42
Antimotility and antidiarrheal agents should be reserved for those patients whose diarrhea persists despite the identification and treatment of the underlying cause. C difficile infection should specifically be ruled out as antimotility agents in this setting can precipitate the development of a toxic megacolon.43 Antimotility agents include loperamide, diphenoxylate/atropine, and oral narcotic derivatives. Loperamide is advocated as the medication of choice as it has the lowest risk of central nervous system adverse effects.42 Bismuth subsalicylate is less effective than loperamide and lacks supporting data to recommend its use.44 Cholestyramine is effective in the treatment of diarrhea caused by bile acid malabsorption (eg, patients with short bowel syndrome, terminal ileum resection, postcholecystectomy) but given concerns about binding to other medications—most notably oral vancomycin45—and the lack of data outside of these specific patient populations, its general application is not recommended.
The composition of enteral nutrition formulas (ENF) can also be responsible for diarrhea and modification of these components can bring resolution.31 ENF with high osmolality may cause diarrhea, especially when being fed directly into the small bowel, and changing to a lower osmolality formula may alleviate diarrhea.46,47 Some ENF may contain poorly absorbed and rapidly fermentable short-chain carbohydrates collectively termed FODMAPs (fermentable, oligo-, di-, monosaccharides, and polyols).41,48 These act similarly to undigested lactose and include fructooligosaccharides (FOS), galactooligosaccharides (GOS), and fructose. FODMAPs significantly increase output from the small bowel due to osmotic effects and present rapidly fermentable substrates to colonic bacteria with subsequent excessive and ongoing gas production. Dietary FODMAPs have even been shown to induce symptoms in healthy volunteers,49,50 and their role in intestinal dysmotility in the ICU is an area of active research.
Fiber is also often incorporated into the ENF or be added as a supplement. Fiber can be classified as soluble or insoluble, and each has different effects. Soluble fibers, typically found in fruits and vegetables51 (eg, partially hydrolyzed guar gum, fructooligosaccharides, pectin, inulin, psyllium), are fermented by colonic anaerobic bacteria to short-chain fatty acids.42 These fatty acids are a preferred fuel for colonocytes and may mitigate diarrhea by improving sodium and water reabsorption in the colon.52 Insoluble fibers, typically found in whole grains (eg, cellulose and hemicellulose) may decrease diarrhea by increasing stool bulk and absorbing water.42,51 Many types of fiber have been studied in the prevention and treatment of diarrhea in patients. Conflicting meta-analyses have been published53,54 although the more recent analysis including over 1700 patients in 51 studies showed a reduction in the incidence of diarrhea with fiber supplementation of enteral feeds.53 The SCCM and ASPEN guidelines state that the use of soluble (but not insoluble fiber) may be useful in patients who develop diarrhea while receiving enteral nutrition.34
Modification of the microflora of the gastrointestinal tract is another area of active investigation and its role in the ICU has yet to be determined. Probiotics are a preparation or product containing viable defined microorganisms in sufficient numbers which alter the microflora by implantation or colonization in a compartment of the host and that exert beneficial effects in the host. Prebiotics are nondigestible food ingredients that beneficially affect the host by selectively stimulating the growth or activity of one or a limited number of bacteria in the colon, and thus improve the health of the host. Synbiotics are a combination of prebiotics and probiotics able to modulate gut immunity and facilitate nutrient/factor interaction necessary for gut recovery.55 A review of the literature by Isakow in 2007 found no evidence for the use of probiotics in critically ill patients.56 The SCCM/ASPEN Clinical Practice Guidelines published in 2009 state there exist insufficient data to make recommendations for general usage in the ICU population.34 There are currently a number of active trials investigating their use in the ICU population.
Finally, the method of delivering enteral feeds can also cause diarrhea with continuous feeding potentially having a lower incidence compared with intermittent feeding.57,58 Bacterial contamination of enteral feeds has also been postulated as a cause of diarrhea although data for this are lacking.34
The Clostridium difficile bacillus was first described in 1935, although its association with disease was not identified until 1978.59Clostridium difficile infection (CDI) has become an increasingly common cause of health care–associated diarrhea and an increasingly common reason for admission to the intensive care unit (ICU).60-63
The prevalence of asymptomatic colonization with C difficile is 7% to 26% among adult inpatients in acute care facilities.43 The primary reservoirs of C difficile are colonized or infected patients and contaminated environments and surfaces within hospitals.64 Recently, there have been major outbreaks of CDI in North America, England, Europe, and Asia, and the emergence of more virulent strains is leading to greater numbers of infected patients with greater disease severity and mortality.63,65-67
Pathogenesis of Clostridium Difficile: C difficile is an anaerobic gram-positive spore-forming bacillus that most commonly exists in a vegetative form that is readily killed by even brief exposures to oxygen.68 When in its spore form, it is heat stable and resistant to gastric acid and many ethanol-based disinfectants. It is transmitted via the fecal-oral route, from person to person and fomite to person.
In healthy adults, the colon contains more than 500 species of bacteria,69 some of which antagonize the adherence and proliferation of C difficile in the colonic crypts. Creation of a suitable local environment allows C difficile reproduction and the generation of toxins, thereby establishing CDI. The pathogenesis of CDI is dependent on the three events: alteration of normal fecal flora, colonic colonization with toxigenic C difficile, and growth of the organism and elaboration of its toxins.64
C difficile toxin A (enterotoxin) and B (cytotoxin) are both exotoxins. Toxin B is more potent (up to 10×) than toxin A and demonstrates cytotoxic effects,68 but both cause increased vascular permeability by opening tight junctions between cells. They both induce the production of TNF-α and proinflammatory cytokines which contribute to the associated inflammatory response and formation of colonic pseudomembranes. The majority of toxigenic C difficile strains produce both toxins, but approximately 1% to 2% of strains in the USA are negative for toxin A.43
In 2002, hospitals in Quebec, Canada, began experiencing an epidemic of CDI with over 14,000 nosocomial cases reported between 2003 and 2004 with a mortality rate of almost 14% (historic controls had 2% mortality).43,66 Similar outbreaks were also reported in a handful of US states. In 2005, the NAP1/B1/027 epidemic strain of C difficile was reported as the causative strain in these outbreaks with its increased virulence traced to its ability to produce 16 × more toxin A and 23 × more toxin B than control strains, as well as uniformly carrying the gene for binary toxin.66 Further studies demonstrated this strain to have resistance to fluoroquinolone antibiotics that was new when compared to historic isolates.65 A recent assessment shows this strain has now spread to over 40 US states and 7 Canadian provinces, and caused outbreaks in England, Europe, and Asia.43,62,63 Infection with this strain may mean a reduced time from the development of symptoms to severe disease and thus require more aggressive therapy and monitoring.70,71
Identified risk factors for the development of CDI include age ≥65 years, increased duration of hospitalization, exposure to antimicrobial agents (with longer exposure and exposure to multiple antimicrobials increasing this risk), cancer chemotherapy, gastrointestinal surgery, and gastric acid suppression.43,68 Prognostic factors for poor outcome following CDI include advanced age, comorbidities, decreased antibody response to the infection, gastric acid suppressants, the need to prolong inciting antibiotic therapy, immunodeficiency, and ICU admission.62,71
Clinical Presentation of Clostridium Difficile: CDI symptoms present at a median of 2 to 3 days after colonization. Typical clinical features consist of watery, grossly nonbloody diarrhea and abdominal pain. Systemic features can include fever, anorexia, nausea, malaise, and a leukocytosis. Severe disease can develop a colonic ileus and toxic dilation with minimal or no diarrhea. Complications of CDI include dehydration, electrolyte imbalances, hypoalbuminemia, renal failure, toxic megacolon, colonic perforation, SIRS, sepsis, shock, and death. Fulminant colitis occurs in 1% to 3% of patients and the hospital mortality associated with toxic megacolon is reported as 24% to 38%.64,72
Diagnosis of Clostridium Difficile: Diagnosis is made based on a combination of clinical and laboratory information. Recently published US and European guidelines have better defined this disease and its treatment.43,71,73 CDI is defined by the presence of symptoms (usually diarrhea) and either a stool test positive for C difficile toxins or toxigenic C difficile, or colonoscopic or histopathologic findings revealing pseudomembranous colitis. Laboratory testing should only be performed on diarrheal stool unless ileus due to CDI is suspected, and it is not recommended to test asymptomatic patients or use as a test of cure.68 Although stool culture is the most sensitive test, it requires 2 to 3 days for results. Enzyme immunoassays (EIA) for toxin are rapid but have historically had poor sensitivity. However, newer EIA and polymerase chain reaction tests for toxin have high sensitivity and specificity rates.74 Since no testing strategy is 100% sensitive and specific, clinical suspicion and consideration of patient risk factors remain important in making clinical decisions about treatment.43,71 This is particularly important for the up to 37% of fulminant CDI that present with ileus rather than diarrhea.
In CDI patients with ileus, and some with diarrhea, the diagnosis can sometimes be made endoscopically. Pseudomembranes are islands of neutrophils, fibrin, mucin, and cellular debris that can be appreciated histologically or on direct visualization via endoscopy.64 Of the patients who are diagnosed with combined laboratory and clinical criteria, approximately 50% of these cases will have pseudomembranes on direct visualization or on histopathologic examination, but the percentage with pseudomembranes in severe disease can be as high as 85%. Thus, although endoscopy is not a sensitive test, the visualization of pseudomembranes is sufficiently specific to confirm the diagnosis of CDI.
Radiographic imaging may also provide information suggestive of CDI. Abdominal x-ray may show “thumbprinting” which is representative of colonic haustral thickening that can be seen in any form of colitis. Abdominal computed tomography (CT) is a more sensitive test to assess for radiographic features of colitis (see Fig. 104-5).61 In the authors’ experience, an advantage of CT lies in its ability to help identify the etiology of undifferentiated severe abdominal sepsis while confirmatory laboratory testing is pending. A patient hospitalized with new onset abdominal sepsis, diarrhea, and pancolitis on CT scan has CDI until proven otherwise and requires immediate empiric therapy.
Abdominal CT with classic findings of C difficile including diffusely thickened colon and ascites. (Used with permission of LN Tremblay, MD.)
Treatment of Clostridium Difficile
Medical Medical therapy for CDI largely revolves around antibiotic therapy and supportive care. The various society guidelines vary depending on the severity of disease, with most recommending for severe cases the use of oral vancomycin alone or in combination with intravenous metronidazole. There are also a number of other important aspects for CDI management. First, the inciting antibiotics should be discontinued if possible. In severe cases, empiric therapy should be initiated as soon as the diagnosis is suspected and not delayed for laboratory confirmation. Antiperistaltic agents should be avoided as they may obscure symptoms and precipitate toxic megacolon. Also, appropriate infection control measures including glove and gown contact precautions for the duration of the diarrhea, hand hygiene compliance, and private rooms for CDI patients should be instituted.
In addition to these standard therapies, multiple adjunctive therapies have been proposed but generally lack sufficient evidence to recommend. For example, both the US and European guidelines state there is no role for the use of probiotics in the treatment or prevention of CDI.43,71 Similarly the data on the effectiveness of IVIG are not particularly compelling and until sufficient evidence is available its use is not recommended. In vitro and animal data showed that both cholestyramine and colestipol bind the toxins produced by C difficile.45 Unfortunately, these agents also bind vancomycin, thereby decreasing the amount of active drug available. Given the lack of evidence and the current recommendation for vancomycin as first-line therapy for severe CDI, anion-exchange resins are not recommended. Fecal transplantation is another treatment that has recently shown promise for recurrent CDI, but it similarly at present has no role in acute severe CDI.
There are a number of antibiotics whose role in CDI is evolving.75 Nitazoxanide, an antiprotozoal drug, has been shown to be at least as effective as both metronidazole and vancomycin in small randomized controlled trials.73,76 Although not available in the United States, the use of teicoplanin is endorsed by both US and European guidelines as being equally efficacious to vancomycin and metronidazole. Rifaximin is an antibiotic with minimal absorption in the gastrointestinal tract after oral administration that has been reported to be effective against C difficile.77 Fidaxomicin is a new nonabsorbed macrocyclic antibiotic that has recently has been approved by the US Food and Drug Administration for treatment of C difficile, with studies suggest it is noninferior to oral vancomycin in mild to moderate CDI cases.78
Surgical Surgical therapy for CDI is indicated for toxic megacolon, perforation, failure of medical therapy, and fulminant disease. While the former two indications are relatively easy to identify, the latter two can be more difficult. Clinical response to medical therapy may take 48 to 72 hours to manifest.79 Multiple retrospective series have sought to identify predictive factors that could be applied to decide when such patients would benefit from an early colectomy.61,80-82 Certain factors such as lactate level of 5.0 mmol/L or greater,61 WBC >3780 or >50,61 use of preoperative vasopressors, preoperative single or multiorgan failure, and immunosuppression have been identified as predictive of mortality. While there are no prospectively validated data to confirm the clinical utility of these factors, it is the opinion of the authors that after the initial resuscitation and institution of antibiotic therapy of patients with CDI admitted to the ICU, any new onset or worsening of existing organ failure, increasing lactic acidosis, increasing WBC, or the requirement for vasoconstrictor medications for support should prompt consideration of surgical therapy, with a lower threshold applied to those who are immunocompromised.
The gold standard surgical treatment for severe CDI currently is a total abdominal colectomy with sparing of the rectum and creation of an end ileostomy.83 Unfortunately, this operation is associated with high mortality and morbidity. In a retrospective analysis of the Quebec 2003-2005 CDI outbreak series, following adjustment for several confounders, the group undergoing this surgery had lower mortality (adjusted odds ratio of 0.22; 95% CI, 0.07-0.67; p = 0.008) versus those managed nonoperatively61; but the 30-day mortality for the surgical group was 34%! A recent long-term follow-up study of patients treated with colectomy for CDI found 1-year, 2-year, 5–year, and 7-year mortality of 68.5%, 79.6%, 88.9%, and 90.7%.72 In an effort to improve outcomes, another surgical technique—a diverting loop ileostomy and colonic lavage—has recently gained favor. A preliminary study using this new approach suggested a much lower early mortality of only 19%.84 Further studies are ongoing to better define the optimal surgical approach.83