Medications are a frequent cause of iatrogenic diarrhea in the ICU setting. Many of the drugs commonly used in the ICU can cause diarrhea (Table 96.2). Therefore, any medication
or combination of medications should be suspected, and uncertainty on the part of the physician warrants consultation with a pharmaceutical reference. Antibiotic-associated diarrhea occurs in 3% to 29% of hospitalized patients [5]. The frequency of diarrhea varies considerably depending on the antibiotic administered. The rate of diarrhea associated with parenterally administered antibiotics is comparable to orally administered antibiotics, especially antibiotics excreted into the enterohepatic circulation. Antibiotics most commonly associated with diarrhea include ampicillin, tetracycline, clindamycin, azithromycin, clarithromycin, fluoroquinolones, and many of the cephalosporins [6]. Antibiotic agents often cause a nonspecific, noninflammatory diarrhea associated with nausea, abdominal cramping, and bloating. In these instances, diagnostic studies generally are negative. Fluid and electrolyte losses are minimal and symptoms often abate after withdrawal or change of the medication. Alterations in intestinal flora, breakdown of dietary carbohydrate products, and prokinetic effects (e.g., from erythromycin) are all postulated mechanisms of diarrhea [7].

Clostridium difficile infection (CDI) is the most common cause of infectious diarrhea in the ICU [8]. In fact, residence in the ICU has been identified as a risk factor for developing CDI [9], and some authors believe that C. difficile toxins are responsible for 50% of the cases of diarrhea in the ICU setting. CDI in the ICU is increasing in not only in incidence but also in severity [10]. It can present as a serious and sometimes life-threatening complication. CDI must always be considered in ICU patients with diarrhea who are commonly exposed to various mediations, particularly antibiotics that predispose to the development of CDI. Classically, clindamycin, penicillin, and broad-spectrum cephalosporins have been implicated. However, CDI may be caused by any antibiotic, including metronidazole and vancomycin, the agents typically used to treat CDI. The risk factors associated with CDI, besides antibiotic exposure and environmental factors, include age greater than 60, severe underlying disease, gastric acid suppression, and immunologic susceptibility [7]. C. difficile produces multiple toxins, two of which have been well characterized. Toxin-induced changes in colonocyte function, cytokine release, and alterations in intestinal motility result in the signs and symptoms characteristic of CDI [11]. One strain of C. difficile, NAP1 (North American pulsed-field electrophoresis type 1), has been associated with both an increased morbidity and mortality [12]. Prompt recognition and treatment of CDI are essential because severe cases of CDI can progress to fulminant colitis and toxic megacolon requiring urgent surgical intervention.

Agents that increase the osmotic load in the gut lumen are also frequent causes of diarrhea in the ICU patient. Magnesium-containing antacids (e.g., Maalox and Mylanta) are common examples of such agents. The gut lumen osmotic load can also be increased as a result of aggressive enteral repletion of nutrients such as magnesium and phosphorus. Lactulose, a useful agent in the treatment of hepatic encephalopathy, provides an osmotic gradient resulting in increased fluid secretion and stool output. Many medications contain inert additives, sorbitol or lactose, which may also cause an osmotic diarrhea. In one study including 29 tube-fed patients with diarrhea, 48% of the cases were attributed to sorbitol-containing elixirs [13].

Proton pump inhibitors (PPIs), another commonly used class of medication in the ICU setting, frequently cause diarrhea, particularly when administered in higher doses. In fact, in a large study of more than 40,000 patients treated with omeprazole, lansoprazole, or pantoprazole, the most common adverse event was diarrhea [14].

Immunosuppressants used in transplantation (e.g., tacrolimus, sirolimus, mycophenolate mofetil, cyclosporine, and
azathioprine) are associated with diarrhea. However, these agents may not be causative. As an example, an alternative explanation was found in 50% of kidney transplant patients who developed diarrhea while receiving mycophenolate [15]. In patients with HIV who are treated with highly active antiretroviral therapy (HAART), drug-induced diarrhea occurs in up to 75% of patients, and the protease inhibitors as well as integrase inhibitors are the most common drug-related cause of diarrhea in this population [16]. Symptoms are lessened by dose reduction or eliminated by discontinuation of therapy.

Withdrawal from medications (e.g., long-term sedatives, analgesics) may also be associated with diarrhea [17].

Other medications associated with diarrhea include colchicine, quinidine, digitalis, metoclopramide, theophylline, levothyroxine, aspirin, nonsteroidal anti-inflammatory drugs, misoprostol, cimetidine, diuretics, cholinergic agents (e.g., bethanechol), and beta-blockers.

Enteral feedings are the most common cause of diarrhea in the ICU setting, occurring in up to 63% of ICU patients [1,18]. Numerous studies have investigated the role of enteral feedings in causing diarrhea in the critically ill patient. Certain aspects, such as concurrent administration of antibiotics, osmolality of solution, type of solution, and serum albumin, have been assessed to determine their contributing roles in the occurrence and severity of diarrhea in these patients [19]. In most instances, diarrhea in enterally fed patients is associated with concurrent antibiotic administration [18,20]. However, enteral feeds also cause changes in gut function that can result in diarrhea. The osmolarity of the enteral solution may play a role when elemental-type diets are used, and especially when feedings are rapidly administered directly into the small intestine. Bolus feeding may be more physiologic, especially with regard to glucose homeostasis; however, feedings administered in this manner distal to the pylorus introduce high-osmolar contents rapidly into the small bowel, resulting in a higher incidence of diarrhea [21]. The impact of enteral nutrition-related complications, including diarrhea, was illustrated in a prospective, multicenter cohort study of 400 patients [22]. In this study, 62.8% (251 of 400) patients suffered GI complication with 14.7% of the studied patients experiencing enteral nutrition-related diarrhea. These authors found that patients with GI complications had a reduction in their tube feed volumes, longer length of stay in the ICU, and higher mortality.

Enteral formulas high in lactose or fat content may also be a factor in susceptible patients. Starved or chronically parenterally fed patients who have developed small bowel villus atrophy and a decrease in mucosal disaccharidase enzyme activity may experience diarrhea when enteral feedings are initiated.

The relationship between hypoalbuminemia and diarrhea is controversial. Hwang et al. [2] compared ICU patients with and without diarrhea and found that the albumin level was statistically different between groups (1.90 g per dL vs. 3.40 g per dL in the groups with or without diarrhea, respectively). Hypoalbuminemia with resulting lowered oncotic pressure may cause diarrhea by inducing changes in the Starling forces sufficient to inhibit intestinal fluid absorption. Some authors claim that concurrent nutritional intake and correction of the albumin deficit with intravenous salt-poor albumin may result in normalization and maintenance of albumin levels with an improved tolerance to enteral feedings and resolution of diarrhea [23]. Conversely, patients with severe hypoalbuminemia secondary to cirrhosis or nephrotic syndrome do not uniformly have diarrhea. Until further studies show efficacy, routine use of intravenous albumin repletion cannot be recommended.

Studies investigating the role of the intestinal response to tube feedings have revealed that intraduodenal infusion resulted in a normal postprandial pattern of small intestinal motility and an increase in the volume of fluid entering the colon, but did not result in diarrhea [24]. Intragastric infusion, on the contrary, resulted in small intestinal motility and colonic flow similar to fasting, and the majority of subjects developed diarrhea [25]. This has led to the conclusion that enteral feeding–related diarrhea may be secondary to a disorder in colonic function. Further supporting this hypothesis are studies that have shown that the ascending colon, normally the site of maximal absorption of water and electrolytes, actually secretes water, sodium, and chloride during intragastric and intraduodenal infusion [26]. Up to 3.2 L per day was secreted by the ascending colon in these studies. Although this is well within the estimated 5.7 L per day maximal absorptive capacity of the colon, diarrhea still occurred, suggesting that this reversal of normal colonic physiology seriously impairs the absorptive capacity of the colon [27].