12.1 GI Physiology, Acute Abdomen, Bleed, Ileus, and Perforation (01.S.1–5)

A board-like abdomen is a luxury the neuro ICU often does not have. The same patient who needs an external ventricular drain (EVD), propofol, and a neuromuscular blocker cannot report pain, cannot guard, and may be on dexamethasone that further erases peritoneal signs. Gastrointestinal failure still kills through hemorrhage, ischemia, perforation, and abdominal hypertension that then raises intracranial pressure (ICP). This section covers the gut physiology that makes those failures predictable, then the five examination topics grouped under gastrointestinal disorders: acute abdomen, bleeding, ileus, and perforation. Independent OpenExamPrep material here follows topics listed in the ABPN Content Specifications. It is not an ABIM or ABPN product.

Splanchnic physiology the examination actually uses

At rest the splanchnic circulation takes about 25% of cardiac output. The celiac axis supplies the foregut (stomach, duodenum, liver, pancreas, spleen). The superior mesenteric artery (SMA) supplies the midgut from the distal duodenum through the proximal two-thirds of the colon. The inferior mesenteric artery (IMA) supplies the hindgut. Mucosal perfusion is the most oxygen-hungry layer and the first to fail in shock, which is why occult gut ischemia presents as feeding intolerance, ileus, and a rising lactate long before rigidity.

Gastric mucosal defense is mucus, bicarbonate, intact microcirculation, and prostaglandins. Intracranial hypertension increases vagal drive and gastric acid — the physiology behind Cushing ulcers after traumatic brain injury (TBI) and other high-ICP states. Stress-related mucosal injury is also common in mechanically ventilated, coagulopathic patients; who needs pharmacologic prophylaxis is covered with ICU bundles elsewhere. Here the point is mechanism: acid plus ischemia plus coagulopathy produces bleeding, and acid plus steroids plus an unexaminable belly produces silent perforation.

The gut is also a capacitance bed. Venous pooling and ileus expand intra-abdominal volume. Opioids, clonidine, anticholinergic drugs, hypokalemia, hypomagnesemia, hyperglycemia, and retroperitoneal or spinal injury all slow transit. In a patient with aneurysmal subarachnoid hemorrhage (SAH) who is kept euvolemic and on nicardipine, the combination of opioids, immobility, and potassium shifts from nimodipine or diuresis is a classic setup for adynamic ileus — not a reason to assume obstruction.

Acute abdomen: examination versus imaging

Peritonism means involuntary guarding, rebound, and rigidity from inflammation of the parietal peritoneum. Those signs require an intact sensorium and intact abdominal wall motor function. They are unreliable after sedation, paralysis, high spinal cord injury, recent laparotomy, and high-dose glucocorticoids. In a conversant patient, pain out of proportion to a soft abdomen still matters — that pattern is mesenteric ischemia until the workup says otherwise, not “reassurance because there is no rebound.”

Do not send an unstable patient to computed tomography (CT). Hypotension, ongoing hematemesis, a tense surgical abdomen with septic shock, or abdominal catastrophe with hemorrhage is a resuscitation-and-operating-room problem. Two large-bore intravenous catheters, blood products if bleeding, airway protection, and a surgeon at the bedside happen before a trip down the hall. Upright or reverse-Trendelenburg chest radiography can show subdiaphragmatic free air in seconds; a negative film does not exclude perforation.

When the patient can be made safe for transport, CT of the abdomen and pelvis with intravenous contrast is the workhorse for the medical acute abdomen. Oral contrast delays diagnosis and is not required to see free air, vascular occlusion, or most perforations. Look for pneumoperitoneum, transition points, closed-loop anatomy, bowel-wall hypoenhancement, pneumatosis, portal venous gas, drainable collections, and pancreatitis. Serial abdominal examinations are a plan only in a patient who can actually be examined.

Worked example: a 64-year-old on dexamethasone after resection of a brain metastasis becomes tachycardic, with a lactate of 3.1 mmol/L and a soft, distended abdomen. There is no rebound. CT shows extraluminal air and fat stranding at the sigmoid. The examination did not “clear” the abdomen; steroids and encephalopathy did. That patient needs source control, not another six hours of observation because peritoneal signs never appeared.

When laparotomy cannot wait

Imaging is not a moral good. The following situations are operative (or interventional) now, in parallel with resuscitation:

1. Diffuse peritonitis with shock in a patient whose examination can be trusted.
2. Pneumoperitoneum plus instability, or free air in an unexaminable patient.
3. Dead or frankly ischemic bowel, closed-loop obstruction, or abdominal catastrophe with ongoing hemorrhage.
4. Abdominal compartment syndrome that fails medical decompression (next section).
5. Source control for perforated viscus: antibiotics without an operation are not definitive if the hole is still open.

Delaying the operating room to “optimize INR to 1.5,” to finish an oral-contrast protocol, or to wait for a lactate to “declare” is a standard examination trap. Reverse coagulopathy on the way to source control. Lactate is a late marker of gut infarction. A rising vasopressor dose after steroids or after an EVD is placed is sometimes abdominal sepsis, not only brainstem failure.

Gastrointestinal bleeding: resuscitation first

Upper gastrointestinal bleeding (UGIB) is proximal to the ligament of Treitz: peptic ulcer, varices, Mallory–Weiss tears, erosive gastritis, Dieulafoy lesions. Lower gastrointestinal bleeding (LGIB) is distal: diverticular bleeding, angioectasia, ischemic colitis, neoplasm, hemorrhoids. Melena is usually upper. Hematochezia is usually lower unless the patient is in shock — brisk duodenal ulcer bleeding can dump red blood per rectum. In that vignette, place a nasogastric tube if it will change the next test, but do not let a nonbloody aspirate talk you out of an upper source; the tube can sit distal to a duodenal bleed.

Resuscitate before you argue about PPI versus octreotide. Protect the airway if hematemesis threatens it — a common neuro ICU collision when Glasgow Coma Scale is already 8. Use two large-bore peripheral catheters; a 22-gauge in the hand is not a resuscitation line. The Villanueva trial (NEJM 2013) randomized acute UGIB to a restrictive hemoglobin threshold of 7 g/dL versus 9 g/dL. The restrictive strategy reduced death and rebleeding, with the largest benefit in Child–Pugh A and B cirrhosis. Do not overtransfuse variceal bleeders: extra volume raises portal pressure. Patients with active acute coronary ischemia may need a higher threshold; that is a cardiac exception, not a reason to transfuse everyone to 10 g/dL.

Ulcer versus varices

Until endoscopy, treat the sick cirrhotic hematemesis as variceal. That bundle is a splanchnic vasoconstrictor (octreotide 50 mcg intravenous bolus then 50 mcg/h, or terlipressin where it is used for bleeding), antibiotics (commonly ceftriaxone 1 g daily) because infection and spontaneous bacterial peritonitis drive mortality, and endoscopy as soon as resuscitation allows — often within 12 hours for varices, and within 24 hours for most UGIB admitted to hospital. Balloon tamponade (Sengstaken–Blakemore or similar) is a bridge to endoscopy or transjugular intrahepatic portosystemic shunt (TIPS), not a destination. TIPS is rescue for failure of endoscopic hemostasis or early rebleeding in selected Child–Pugh A/B patients, not the first procedure in an unresuscitated shock patient.

Ulcer bleeding is an acid and vessel problem. Erythromycin 250 mg intravenous as a single dose before endoscopy improves visualization by emptying the stomach; it is not a hemostatic drug. Pre-endoscopic PPI can downstage high-risk stigmata but has not been shown to cut rebleeding or death and must not delay endoscopy. After successful endoscopic therapy of high-risk ulcers (active spurting or oozing, nonbleeding visible vessel), high-dose proton-pump inhibitor (PPI) for 3 days reduces further bleeding and mortality compared with no acid suppression. Conceptually that is the classic 80 mg bolus then 8 mg/h continuous infusion. The 2021 American College of Gastroenterology ulcer-bleeding guidance also accepts intermittent high-dose PPI (about 80–160 mg/day, often an 80 mg bolus then 40 mg two to four times daily). After those 3 days, twice-daily oral PPI through about 2 weeks is used for high-risk lesions. Clean-based ulcers do not need an infusion.

Coagulopathy reversal versus endoscopy

Endoscopy is the hemostatic procedure. Reversal is supportive. For warfarin-associated life-threatening bleed, give 4-factor prothrombin complex concentrate plus vitamin K — the same logic used for intracranial hemorrhage. For dabigatran, idarucizumab; for factor Xa inhibitors, andexanet or 4-factor PCC depending on availability and the bleed. Platelets below about 50 × 10^9/L during active bleeding or before endoscopic therapy are commonly repleted. In cirrhosis, the international normalized ratio (INR) does not measure the same bleeding risk it does on warfarin: both procoagulant and anticoagulant factors fall. Flooding the patient with fresh frozen plasma to chase an INR of 1.5 delays endoscopy, adds volume, and can raise portal pressure and ICP. Correct life-threatening coagulopathy in parallel; do not hold the endoscope for a perfect laboratory panel. Viscoelastic testing can guide product choice when it is available; it is not a reason to postpone a crashing bleed.

Ileus versus obstruction

Adynamic (paralytic) ileus is failure of peristalsis without a mechanical block. Plain films and CT show gas in small bowel and colon, including the rectum, without a transition point. Bowel sounds may be quiet. Causes in this unit: opioids, hypokalemia, hypomagnesemia, anticholinergics, retroperitoneal hematoma, spine surgery, pneumonia, and abdominal hypertension itself. Treatment is nasogastric decompression if vomiting or severely distended, nothing by mouth, electrolyte repletion, opioid-sparing analgesia, and time. Surgery for ileus without ischemia or perforation is the wrong answer.

Mechanical small-bowel obstruction has a transition point, decompressed distal bowel, and often a history of adhesions, hernia, or cancer. Partial adhesive obstruction can have a brief nonoperative trial with nasogastric suction if there is no fever, peritonitis, lactate rise, or closed-loop anatomy. Complete obstruction, closed loop, hernia, ischemia, or perforation is operative. Oral water-soluble contrast can be diagnostic and sometimes therapeutic in adhesive partial obstruction; it is not a plan for a patient with peritonitis.

Acute colonic pseudo-obstruction (Ogilvie syndrome) is massive colonic dilation without a mechanical lesion, often after spine surgery, opioid exposure, or electrolyte failure. If the cecum approaches 12 cm or the patient is failing conservative care, neostigmine 2 mg intravenous over several minutes with atropine at the bedside for bradycardia is the medical decompression. Colonoscopy decompresses failures and contraindications to neostigmine. Ischemia or perforation is an operation. Do not give neostigmine to a patient with mechanical obstruction, recent anastomosis, or active bronchospasm.

Free air, steroids, and the EVD patient who cannot declare peritonism

Pneumoperitoneum is perforation until a surgeon agrees it is not. Residual air after recent laparotomy can persist for days, and alveolar rupture can theoretically track, but those are diagnoses of exclusion. In the neuro ICU the dangerous perforations are peptic ulcer and diverticular perforations in patients on high-dose dexamethasone for tumor edema or, less often, other glucocorticoid courses. Steroids thin mucosa, impair healing, and ablate peritoneal signs. The vignette is unexplained tachycardia, a new fever, a vasopressor bump, pneumoperitoneum on a chest film obtained for “atelectasis,” and a still-soft abdomen.

Patients with an EVD add coma, sedation, and often neuromuscular blockade. They cannot report pain. Guarding is pharmacologically deleted. Clues that the abdomen — not only the brain — is failing:

• New fever or leukocytosis without a pulmonary or urine source
• Rising vasopressor requirement after the ICP has been controlled
• Feeding intolerance, abdominal distention, or high gastric residuals
• Unexplained lactate or metabolic acidosis
• Sudden high peak airway pressures from a tight abdomen
• Free air or unexplained fluid on imaging obtained for another reason

The correct move is a low threshold for IV-contrast CT and surgical consultation, not waiting for rigidity that paralysis will never allow. Antibiotics cover secondary peritonitis once perforation is likely, but they do not close a hole. If the patient is too unstable for CT and free air is already visible, the surgeon and the operating room are the test.

Exam traps for this cluster: treating every distended SAH patient as obstruction; holding endoscopy for an INR of 1.4 in cirrhosis; using PPI infusion as the sole therapy for variceal shock; assuming a soft abdomen excludes perforation in a steroid- or paralytic-treated patient; and sending a hypotensive bleed to CT instead of the resuscitation bay and the endoscopist.