Splanchnic circulation
Three unpaired arteries off the front of the aorta feed the entire gut in developmental blocks joined end to end by collaterals, and take about a quarter of the resting cardiac output: bowel dies where those territories meet, when one of them is suddenly plugged, or when the body decides the gut is the organ it can most afford to lose.
What it normally does
Three unpaired arteries leave the front of the aorta and divide the gut between them in developmental order. The coeliac trunk (about T12) supplies the foregut: lower oesophagus to the second part of the duodenum, plus liver, gallbladder, spleen and most of the pancreas. The superior mesenteric artery (SMA, about L1) supplies the midgut: duodenum beyond the bile duct opening, the whole small bowel, caecum, ascending colon and the proximal two thirds of the transverse colon. The inferior mesenteric artery (IMA, about L3) supplies the hindgut: distal transverse colon, descending and sigmoid colon and the upper rectum. The territories are joined end to end by collateral channels - pancreaticoduodenal arcades between coeliac and SMA, and along the colon the marginal artery of Drummond with the more central arc of Riolan between SMA and IMA. The joins are at their thinnest and most variable at the splenic flexure (Griffiths point) and the rectosigmoid junction (Sudeck point). The lower rectum has a second supply altogether, middle and inferior rectal arteries from the internal iliac.
a narrowing that builds over years can be silent, because collaterals enlarge to cover it, so arterial gut ischaemia is a disease of sudden blockage or of the thin joins - the splenic flexure and left colon are where a fall in perfusion pressure shows up first, and the lower rectum, with two separate supplies, almost never infarcts.
The gut takes roughly 25 per cent of resting cardiac output, and SMA flow doubles or more within half an hour of a meal. Inside each villus the arteriole and venule run side by side as a hairpin, so oxygen short-circuits from artery to vein at the base and the oxygen tension falls progressively towards the tip (countercurrent exchange), leaving the tip the least well oxygenated tissue in the villus. The mucosa is also the hungriest layer of the bowel wall - it does the absorbing and turns its cells over every few days.
the villous tip is the first part of the bowel wall to die when flow falls, so mucosal injury - bleeding, sloughing, bacteria crossing the wall - happens well before the muscle layer or the serosa are involved; and a gut that is coping at rest can be ischaemic only after eating, which is what makes chronic mesenteric ischaemia a post-meal disease.
The splanchnic bed is also the body's blood reservoir and its first sacrifice. It holds about 20 to 25 per cent of the circulating volume in its veins, and its arterioles are packed with alpha-1 adrenoceptors and are unusually sensitive to angiotensin II and vasopressin. When blood pressure falls, sympathetic outflow clamps the gut down hard to defend the brain and heart.
shock, sepsis, severe heart failure, dialysis and infused vasopressors can infarct bowel with completely normal arteries on the scan (non-occlusive mesenteric ischaemia); and, turned the other way, drugs that constrict this bed on purpose are the way to cut blood flow into a bleeding portal system.
All of that blood leaves through one door. The splenic vein joins the superior mesenteric vein behind the neck of the pancreas to form the portal vein, with the inferior mesenteric vein usually draining into the splenic vein on its way. The portal vein delivers about 75 per cent of liver blood flow before anything reaches the systemic circulation. The gut's sensory nerves run back alongside the sympathetics and are few and poorly mapped, so gut pain is felt in the midline at the level of the embryological segment - foregut in the epigastrium, midgut around the umbilicus, hindgut suprapubically - and cannot be localised. The parietal peritoneum is different: somatic nerves, sharp and exactly localised.
early bowel ischaemia produces agonising, vague central pain over a soft, barely tender abdomen ("pain out of proportion to the findings"), and only becomes a localised, rigid, guarded abdomen once the bowel is dead through its full thickness and inflaming the parietal peritoneum - by which time the operation is a resection, not a rescue; and because everything drains to the liver first, a scarred liver pushes pressure back into this whole bed.
What goes wrong
- Acute mesenteric ischaemia (SMA embolus or thrombosis)← from “All of that blood leaves through one door. The…”
A clot lands in or forms on the superior mesenteric artery. Emboli - usually from a fibrillating left atrium, sometimes from a recent infarct or a valve - travel the aorta and take the SMA because it leaves at a shallow downstream angle, typically lodging just past the middle colic branch, so the first part of the jejunum survives and the rest of the small bowel and right colon do not. Thrombosis happens instead on an existing atherosclerotic plaque at the SMA origin, often in someone who has had months of post-meal pain. An embolus stops flow in minutes and the onset is abrupt; thrombosis on a chronic plaque occludes a vessel whose collaterals are already recruited, so that presentation can be less dramatic and more insidious. Either way the mucosa dies first, then the wall, then bacteria and endotoxin cross into the portal blood and the peritoneum. Pain is severe from the outset but visceral, so the abdomen feels almost normal for the first hours.
Pain out of proportion to a soft abdomen, in atrial fibrillation, with a normal lactate. Peritonism and a high lactate mean the bowel is already dead. CT angiogram now.
You would find: Sudden, severe, poorly localised central abdominal pain in an older patient, often with atrial fibrillation, and an abdomen that is soft and unimpressive on palpation. Vomiting and diarrhoea early, sometimes blood. Lactate and metabolic acidosis are normal at first and rise late - a normal lactate does not exclude it; the white cell count is often raised early but is far too non-specific to rule anything in or out. Diagnosis is CT angiography of the mesenteric vessels, done early, not a plain film. Mortality is around 50 to 80 per cent and much of that is delay to diagnosis.
- Chronic mesenteric ischaemia (intestinal angina)← from “The gut takes roughly 25 per cent of resting c…”
Atherosclerosis narrows the origins of the mesenteric arteries. Because collaterals grow, symptoms usually need two of the three vessels significantly narrowed (occasionally a single tight SMA stenosis will do it). Resting flow is enough. After a meal the gut demands double, the stenosed vessels cannot deliver it, and the bowel becomes ischaemic for an hour or two - the same arithmetic as angina in the heart. The patient learns that eating hurts and stops eating.
Post-meal pain, food fear, weight loss, smoker with vascular disease. Usually two of three vessels stenosed. The definitive treatment is revascularisation - stenting or bypass - not tablets.
You would find: Dull central abdominal pain 15 to 60 minutes after meals, in a long-term smoker in their 60s or 70s with vascular disease elsewhere. Fear of food and real weight loss are the giveaway - they have often been investigated for cancer first. An epigastric bruit may be present. CT angiography shows the stenoses.
- Ischaemic colitis← from “Three unpaired arteries leave the front of the…”
Perfusion pressure to the colon falls - from hypotension, dehydration, a low-output heart, an aortic operation that sacrifices the IMA, or vasoconstricting drugs - and the mucosa fails first at the thin joins between arterial territories, classically the splenic flexure and the rectosigmoid. This is usually a low-flow mucosal injury, not a blocked artery, so most cases stay superficial and heal. A minority go on to full-thickness necrosis and stricture. The lower rectum is nearly always spared because the internal iliac artery also feeds it.
Left-sided pain then bloody diarrhoea, splenic flexure segment, rectum spared, usually self-limiting. Small bowel ischaemia kills; colonic ischaemia usually does not.
You would find: The commonest form of gut ischaemia, usually over 60. Sudden crampy left-sided or lower abdominal pain, then bright red rectal bleeding or bloody diarrhoea within a day - and unlike acute mesenteric ischaemia, the abdomen is genuinely tender over the affected segment, so the pain and the findings match. CT shows a thickened segment; colonoscopy shows segmental mucosal change with a normal rectum. Most settle on fluids and bowel rest. In a younger patient look for a cause of splanchnic vasoconstriction - cocaine, amphetamines, pseudoephedrine - and for long-distance running.
- Non-occlusive mesenteric ischaemia← from “The splanchnic bed is also the body's blood re…”
No clot anywhere. Cardiac output falls or systemic vasoconstriction rises, and the sympathetic and angiotensin-driven clamp on the splanchnic arterioles outlives the insult, so the gut stays shut down even after the blood pressure has been restored. Classic settings are cardiogenic shock, sepsis on infused vasopressors, dialysis with a big pressure drop, and digoxin, which constricts mesenteric smooth muscle directly. It accounts for roughly one in five cases of acute mesenteric ischaemia and the arteries look patent on the scan.
Dead bowel with open arteries. Suspect it in the shocked patient on noradrenaline whose lactate will not come down.
You would find: An unwell inpatient - intensive care, post-cardiac-surgery, dialysis - with a distended, quiet abdomen, a creeping lactate and a rising vasopressor requirement that nobody can explain. The treatment is mostly about undoing causes: restore cardiac output and volume, and withdraw or minimise the drugs that are constricting the bed, while surgery is reserved for bowel that is already infarcted.
- Portal hypertension and variceal bleeding← from “All of that blood leaves through one door. The…”
A scarred liver (in Australia most often alcohol-related liver disease, metabolic/fatty liver disease, hepatitis B or C, or haemochromatosis) resists portal flow. At the same time the splanchnic arterioles dilate under locally produced nitric oxide, so more blood is pushed into a system that cannot drain, and the portal pressure gradient climbs. Above a hepatic venous pressure gradient of about 10 mmHg blood is forced through portosystemic connections, which balloon into thin-walled varices at the lower oesophagus and gastric fundus; above about 12 mmHg they bleed. Note where the problem actually is: the pressure is generated by splanchnic INFLOW as much as by hepatic outflow resistance, which is why the drugs work on the arterial side of the gut, not on the varix.
Portal pressure = flow in x resistance out. Varices bleed above a gradient of about 12 mmHg. Drugs lower the inflow; banding fixes the vein.
You would find: Large-volume haematemesis or melaena in someone with stigmata of chronic liver disease - spider naevi, ascites, splenomegaly, jaundice. Chronic liver disease sits disproportionately on Aboriginal and Torres Strait Islander communities, driven by higher rates of chronic hepatitis B and alcohol-related liver disease, and presents at younger ages, so do not discount a young patient's varices.
What we give, and how it works
Open a drug to see what it binds, what that does to the cell, and what you then see in the patient.
- Binds
- antithrombin, a circulating plasma protease inhibitor - heparin binds it, rather than binding any clotting factor directly
- Which does
- binding changes the shape of antithrombin and accelerates its inactivation of thrombin (factor IIa) and factor Xa by around a thousandfold, so new fibrin cannot be laid down
- So you see
- the existing clot stops propagating up and down the vessel and into the mesenteric veins, keeping alive the bowel that is still marginally perfused; the patient's own fibrinolysis is left to do the rest
- And the same mechanism causes
- bleeding, and specifically bleeding into gut whose mucosa is already sloughing - the same block on fibrin formation that stops the clot growing stops haemostasis anywhere the wall has been breached
- Handling
- unfractionated heparin is preferred here rather than enoxaparin because it wears off in a couple of hours and is reversible with protamine, and this patient is probably going to theatre; it is monitored by APTT or anti-Xa
Catches people out: Heparin is also a strongly negative molecule that binds platelet factor 4; in a small number of patients the complex becomes an antigen, and the resulting antibodies activate platelets, so the platelet count falls and the patient paradoxically clots (heparin-induced thrombocytopenia). Falling platelets on heparin means stop all heparin and switch to a non-heparin anticoagulant - not transfuse platelets.
- Binds
- cyclo-oxygenase-1 inside the platelet
- Which does
- aspirin acetylates a serine residue in the enzyme's active channel and blocks it irreversibly, so thromboxane A2 can no longer be made; the platelet has no nucleus and cannot make new enzyme, so it is disabled for its whole 7 to 10 day life
- So you see
- platelets recruit and aggregate far less readily on a ruptured plaque, cutting the chance of the stenosis converting into an acute occlusion
- And the same mechanism causes
- peptic ulceration and upper gastrointestinal bleeding - the same cyclo-oxygenase-1 sits in gastric mucosa making the prostaglandins that drive mucus, bicarbonate and mucosal blood flow, so blocking it removes the stomach's own defences, and because the platelet block is irreversible any bleeding that starts keeps going for days
Catches people out: Risk rises steeply with age, previous ulcer, Helicobacter pylori, and combination with an anticoagulant, another antiplatelet, a corticosteroid or an NSAID. A proton pump inhibitor is the usual co-prescription when that risk is high.
- Binds
- V1 vasopressin receptors on splanchnic arteriolar smooth muscle (terlipressin, a prodrug slowly converted to lysine-vasopressin); octreotide instead acts on somatostatin receptors and reduces splanchnic vasodilator peptides such as glucagon
- Which does
- the V1 receptor is Gq-coupled: phospholipase C makes IP3, stored calcium is released, the myosin light chains are phosphorylated and the arteriole constricts
- So you see
- mesenteric arterial inflow falls, so portal venous flow and portal pressure fall, and the bleeding varix is perfused at lower pressure - bleeding slows and rebleeding before endoscopy is less likely
- And the same mechanism causes
- ischaemia from the same constriction in the wrong place: angina and myocardial infarction, mottled or necrotic fingers, toes and skin, and - the irony worth remembering - it can itself cause the bowel ischaemia described on this page
- Handling
- terlipressin is given as intermittent intravenous boluses (or an infusion) rather than titrated to blood pressure, and is stopped once bleeding is controlled and definitive endoscopic therapy has been done
Catches people out: Avoid or use with great care in known ischaemic heart disease or peripheral vascular disease, and look at the fingers and toes daily. Terlipressin also has some V2 activity in the kidney, so it retains free water and the sodium can fall fast - check it.
- Binds
- beta-1 and beta-2 adrenoceptors (carvedilol additionally blocks alpha-1 receptors, which lowers intrahepatic resistance as well)
- Which does
- blocking cardiac beta-1 receptors lowers cyclic AMP, so heart rate and cardiac output fall; blocking splanchnic beta-2 receptors removes the vasodilator arm and leaves alpha-1 tone unopposed, so the mesenteric arterioles constrict
- So you see
- less blood is delivered into the portal system from both ends, the portal pressure gradient drops, and varices are less likely to grow and bleed
- And the same mechanism causes
- bronchoconstriction, because beta-2 receptors relaxing airway smooth muscle are blocked as well, leaving cholinergic tone unopposed; and fatigue and exercise intolerance from the capped cardiac output
- Handling
- propranolol is titrated upwards by resting heart rate rather than by blood pressure; carvedilol is titrated by fixed dose steps with an eye on blood pressure, because its alpha-1 blockade drops it further. Either way the drug is taken indefinitely - it is prevention, not treatment of active bleeding
Catches people out: Avoid in asthma. In advanced cirrhosis with refractory ascites or a low blood pressure the patient is living on their cardiac output to perfuse the kidneys; taking that reserve away can precipitate renal failure, so the dose is reduced or the drug stopped in that window.
One word, two completely different diseases, and the artery tells you which. SMA territory - small bowel and right colon - has little usable spare supply once the clot sits beyond where the collaterals feed in, so an embolus there is a catastrophe: agonising central pain, a soft abdomen, a normal lactate, an older patient in atrial fibrillation, and a mortality of 50 to 80 per cent driven largely by how long it takes someone to order the CT angiogram. Do not wait for the lactate; it rises when the bowel is already dead. IMA territory and the splenic flexure watershed is the opposite story: a low-flow mucosal injury that announces itself politely with cramping left-sided pain and bloody diarrhoea, is tender exactly where it hurts, spares the lower rectum because the internal iliac also feeds it, and usually settles with fluids and time. So: severe pain with a normal-feeling belly means think small bowel and act in minutes; blood per rectum with a tender left side means think colon and think about why the perfusion pressure dropped - shock, dialysis, an aortic operation, or a vasoconstricting drug.
Now test whether it stuck
Reading this through is not the same as being able to reconstruct it. Every question in the bank is free, with a full debrief on each option.