ICU Nutrition in Surgical Patients — Enteral Dose, Parenteral Timing, Shock Gut, Protein, Immunonutrition, Refeeding and Glycaemic Targets

Day-two post-laparotomy, noradrenaline running, lactate 3.1, abdomen distended, nasogastric tube draining 400 mL, no stool, phosphate falling since feeds started and glucose 14 on the morning gas. Do you push full enteral calories, add parenteral nutrition now, hold everything, restrict for refeeding, or tighten the insulin — and how much protein does the oliguric kidney tolerate? The examiner will watch you dose by EDEN and PermiT, time PN by EPaNIC, respect shock gut by NUTRIREA-2, ration protein by EFFORT, restrict by the refeeding trial, and cap glucose by NICE-SUGAR. This page teaches all six moves with every number taken from the papers named beside it.^[11][6][1]

Overview & Definition — the fed, the underfed and the refed surgical patient

The ICU nutrition patient is defined by route, dose and timing, not by a calorie number alone: enteral nutrition started early but delivered trophically (about 400 kcal per day) versus fully (about 1300 kcal per day) for the first 6 days is the EDEN comparison.^[1] Permissive underfeeding means 40 to 60 percent of calculated requirements versus 70 to 100 percent standard, with protein held equal between arms.^[3] Early parenteral nutrition means started within 48 hours of ICU admission versus late PN held until day 8, in patients whose targets cannot be met enterally.^[6] Refeeding hypophosphatemia in the trials that changed practice means a phosphate fall of more than 0.16 with levels below 0.65.^[9] No single universal weight-based calorie formula is quoted here because the verified trials randomise strategies and report delivered calories rather than prescribing equations — state the gap rather than inventing a number.

The strategic arc of the whole topic fits one sentence: randomised trials show early full nutrition support induces dose-dependent harm through hyperglycaemia, metabolic damage and suppressed cellular repair — so early full support, including early PN, should be avoided.^[33]

Classification — four feeding strategies, one recovery goal

• Early trophic-to-moderate enteral: start EN early, accept 400 kcal-range or 40-to-60-percent delivery in the acute phase — the EDEN and PermiT arms that matched full feeding on every hard outcome with less intolerance.^[1][3]
• Early full enteral: pushing 1300 kcal-range or 70-to-100-percent delivery from day one — more vomiting, high residuals and constipation in EDEN with no mortality, ventilator or infection gain.^[1]
• Early supplemental parenteral: topping up failing EN within 48 hours — slower recovery, more ICU infections, more cholestasis, longer ventilation and more renal-replacement days in EPaNIC.^[6]
• Late parenteral: tolerating the deficit and starting PN on day 8 only if EN still fails — faster ICU and hospital discharge alive with fewer complications, converging to equal 2-year survival.^[6][7]

Each strategy serves recovery, not a calorie ledger — calories buy nothing until perfusion, phosphate and glucose allow their use.^[33][9]

Epidemiology & Risk Factors — the denominators that frame every decision

EDEN randomised 1000 adults within 48 hours of acute lung injury across 44 hospitals of the ARDS Clinical Trials Network.^[1] PermiT randomised 894 medical, surgical and trauma adults across 7 centres, 96.8 percent ventilated.^[3] EPaNIC randomised 2312 patients to early PN against 2328 to late PN.^[6] CALORIES randomised 2400 unplanned ICU admissions across 33 English units with 99.5 percent analysed.^[10] NUTRIREA-2 randomised 2410 ventilated adults on vasopressor support across 44 French ICUs.^[11] EFFORT Protein randomised 1301 nutritionally high-risk ventilated adults across 85 ICUs in 16 countries.^[13] NICE-SUGAR randomised 6104 patients, 3054 to intensive and 3050 to conventional glucose control.^[25] REDOXS randomised 1223 ventilated multiorgan-failure adults across 40 ICUs in three continents, and MetaPlus randomised 301 prolonged-ventilation patients across 14 European ICUs.^[16][17] The refeeding trial enrolled 339 ICU adults who developed refeeding syndrome across 13 Australasian units, and the surgical refeeding cohort followed 200 at-risk surgical ICU patients.^[21][23]

Risk concentrates where the surgeon already worries: starvation beyond 5 days dominated the surgical refeeding cohort at 55 percent, with alcohol, insulin, chemotherapy, diuretic or antacid exposure at 18 percent.^[23] High nutritional risk described 42.3 percent of PermiT patients — yet feeding strategy showed no mortality association in either risk stratum, so risk scores describe the patient without prescribing the dose.^[5] ICU admission itself with initial feeding above 20 kcal per kg per day marked the highest refeeding incidence band at 44 percent syndrome and 27 percent hypophosphatemia across 35 studies.^[22]

Pathophysiology — hyperglycaemia, suppressed repair and phosphate first

Dose-dependent harm is the mechanism anchor: early nutrition support worsens hyperglycaemia with metabolic damage and suppresses cellular repair mechanisms such as autophagy — which is why full early support injures while trophic support suffices.^[33] The autophagy link is measured, not theoretical: late PN raised the LC3II-to-LC3I autophagosome-formation ratio, and a higher ratio independently tracked less weakness.^[8] With late PN, weakness on first assessment fell from 43 to 34 percent and recovered faster.^[8]

Phosphate is the second mechanism every surgical intensivist must narrate: a relative phosphate fall over the first 2 ICU days marks unreadiness for artificial feeding and interacts with early-PN harm, while an absolute low value alone does not carry the same interaction.^[9] In relative-fall patients, early PN independently lowered the chance of earlier live ICU discharge; in patients without the fall it did not.^[9] The fall is poorly predicted by admission characteristics, so it must be measured serially, not assumed from history.^[9]

Hypoglycaemia is the third mechanism and the intensivist's own error: intensive glucose control raised 90-day death while multiplying severe hypoglycaemia thirteen-fold, and the operative subgroup trended worse than the medical one.^[25] The de-adoption cohort proves the direction: abandoning intensive control raised mean glucose yet cut adjusted 90-day death odds nearly in half as hypoglycaemia collapsed.^[26]

Clinical Presentation — the surgical patient facing nutrition decisions

The EPaNIC candidate declares through EN failure: a critically ill adult whose caloric targets cannot be met by enteral nutrition alone — the exact population where PN timing is contested, and where waiting until day 8 wins.^[6] The NUTRIREA-2 candidate declares through shock: ventilated, on vasopressors, randomised within 24 hours of intubation — the closest trial model to the shocked postoperative abdomen, where enteral-first feeding raised digestive complications without cutting death.^[11] The refeeding candidate declares through starvation plus restart: little or no intake beyond 5 days, then feeds commenced with a falling phosphate — restrict calories and watch, per the trial that improved survival by restriction.^[23][21] The EFFORT candidate declares through high risk plus organ failure: ventilated, nutritionally high-risk, with acute kidney injury or high organ-failure scores marking the subgroup where extra protein harms most.^[13][14]

Practical assessment at the bedside: combine perfusion (vasopressor trajectory, lactate, abdominal examination), tolerance (vomiting, residuals pattern, stool, distension), phosphate trajectory, and glucose behaviour — then choose trophic EN, PN bridge, or temporary restriction rather than defaulting to full feeds.^[11][9][25]

Differential Diagnosis — intolerance versus ileus versus ischaemia versus refeeding

Split feed intolerance (vomiting, high residuals, diarrhoea — prokinetic-responsive with moderate-certainty evidence) from expected postoperative ileus (self-limited, managed by patience and mobilisation) from bowel ischaemia (the NUTRIREA-2 enteral excess: 19 versus 5 cases — examine, image and stop EN rather than prokinet through it).^[29][11] Split EN failure needing a timed PN bridge (the EPaNIC day-8 and CALORIES route-equivalence logic) from refeeding physiology demanding temporary restriction (phosphate fall with feeding start — the refeeding-trial population).^[6][10][21] Split surgical-stress hyperglycaemia (cap at 180 or less) from over-treatment hypoglycaemia (the NICE-SUGAR killer gradient — treat the low glucose as the emergency).^[25] Split ICU pharmaconutrition harm (glutamine and antioxidant cocktails in multiorgan failure — REDOXS and MetaPlus) from perioperative GI-cancer immunonutrition benefit (arginine, nucleotide and omega-3 formulas — the GI meta-analyses): different patients, different formulas, opposite answers, and the viva rewards the candidate who never crosses them.^[16][17][18]

Name the category aloud before ordering feeds — the viva rewards the candidate who separates what nutrition fixes from what only perfusion, source control or the operation fixes.^[11][33]

Clinical & Bedside Assessment — risk tools, phosphate and glucose

Use nutritional-risk scores honestly: 378 of 894 PermiT patients scored high risk and 516 low risk, yet permissive underfeeding matched standard feeding on 90-day death in both strata with no interaction — so the score describes vulnerability without dictating dose.^[5] Monitor phosphate as a ready-to-feed signal: relative hypophosphatemia developed in 23.7 percent of EPaNIC patients with phosphate data, absolute in 9.1 percent and combined in 5.3 percent — and only the relative fall interacted with early-PN harm.^[9] Monitor glucose by time-weighted behaviour: the transition cohort raised mean glucose from 6.94 to 8.2 on abandoning intensive control while severe hypoglycaemia fell from 1.2 to 0.4 percent and moderate from 23.3 to 5.9 percent.^[26] State the gastric-residual position honestly, because the examiner will probe it: abandoning routine residual monitoring does not worsen outcome and no gastric-versus-postpyloric consensus exists — so never quote an adult residual cutoff from this set, because none is verified here.^[32][27]

Investigations — phosphate, glucose, triglycerides unquoted and the honest calorimeter gap

Draw phosphate serially over the first ICU days (the relative fall is the actionable signal, poorly predicted by admission characteristics), glucose repeatedly with a hypoglycaemia-avoiding protocol, and electrolytes with refeeding vigilance — but never let the laboratory gate the strategy decision already made on perfusion and tolerance.^[9][25][33] State the viscoelastic-equivalent honesty for nutrition: no verified indirect-calorimetry equation, weight-based formula or triglyceride threshold sits in this pack — the trials report delivered calories (EDEN 400 versus 1300 per day; PermiT 835 versus 1299 per day; CALORIES targets unmet in most patients either way), so quote delivered doses and name the formula gap.^[1][3][10] The French guideline reality check belongs here too: of 34 adult recommendations only 3 rest on high-level evidence with 12 moderate and 19 expert opinion — individualise rather than protocolise.^[31]

Management — Enteral dose: trophic, full and permissive underfeeding

Give EDEN exactly: 1000 lung-injury patients, trophic versus full EN for 6 days at about 400 versus 1300 kcal per day — ventilator-free days 14.9 versus 15.0 and 60-day death 23.2 versus 22.2 percent, neither different, with more vomiting, high residuals and constipation on full feeding and higher glucose and insulin needs.^[1] Add the long-term honesty: survivors walk at 64 to 66 percent of predicted with a quarter cognitively impaired at 12 months, and initial dose changed none of it.^[2] Give PermiT exactly: 894 patients, 40-to-60 versus 70-to-100 percent of requirements for up to 14 days with equal protein — delivered 835 versus 1299 kcal per day (46 versus 71 percent of requirements) with 90-day death 27.2 versus 28.9 percent and no differences in intolerance, diarrhoea, ICU infection or stay.^[3] The conclusion is the viva sentence: moderate nonprotein calories matched full calories on mortality.^[3] Nutritional risk does not rescue full feeding: high-risk and low-risk strata both showed no strategy-mortality association.^[5]

The surgical resolution: start enteral nutrition early but run it trophic-to-moderate through the acute phase — the gut open at low volume beats the gut forced at full volume.^[1][3][33]

Management — Parenteral timing and route: EPaNIC, CALORIES and the 2-year tail

Give EPaNIC exactly: 2312 early-PN (within 48 hours) versus 2328 late-PN (day 8) patients with early EN protocolised in both arms — late initiation brought a 6.3 percent relative gain in earlier live discharge from ICU and hospital, fewer ICU infections (22.8 versus 26.2 percent), less cholestasis, 9.7 percent fewer prolonged-ventilation cases, 3 fewer renal-replacement days and EUR 1,110 saved per patient.^[6] Rates of ICU, hospital and 90-day death were similar — faster recovery, not fewer deaths, is the claim to own.^[6] The weakness mechanism explains the speed: late PN cut weakness from 43 to 34 percent with faster recovery, via preserved autophagic quality control of myofibres.^[8] And the 2-year tail keeps the claim calibrated: mortality (20.5 versus 19.8 percent) and physical function were identical at 2 years in every nutritional-risk subgroup — late PN buys the early months, not extra years.^[7]

Give CALORIES exactly: 2400 unplanned admissions across 33 units, PN versus EN within 36 hours for up to 5 days — 30-day death 33.1 versus 34.2 percent (RR 0.97), with less hypoglycaemia and vomiting on PN but equal treated infections, equal 90-day death and equal adverse events across 14 further secondaries.^[10] Caloric intake was similar with targets unmet in most patients — route changes tolerance, not survival, when EN is feasible.^[10] Enteral stays preferable as the default with parenteral non-inferior in recent studies — but preferable means attempted first, never forced through shock gut.^[32][11]

Management — The shock gut: NUTRIREA-2 and early EN in circulatory shock

Give NUTRIREA-2 exactly, because this is the surgeon's trial: 2410 ventilated adults on vasopressors, isocaloric EN versus PN at 20 to 25 kcal per kg per day within 24 hours of intubation — day-28 death 37 versus 35 percent with equal ICU-acquired infections (14 versus 16 percent), but enteral carried more vomiting (34 versus 20 percent), diarrhoea, bowel ischaemia (19 versus fewer than 1 percent of cases) and acute colonic pseudo-obstruction.^[11] The interpretation is the viva sentence: early isocaloric EN did not cut death or infection but raised digestive complications versus PN in shock.^[11]

Add the cohort nuance so the answer stays calibrated: in 626 ventilated shock patients, EN started under 48 hours looked better unadjusted — more ICU-free and vasopressor-free days with shorter ventilation — but every signal vanished after severity adjustment, with no mortality or intolerance difference.^[12] The authors call for randomised trials of early versus delayed EN in shock — quote the uncertainty, not a timing rule.^[12] The bedside resolution: in escalating vasopressor-dependent shock with a distended, unpassed postoperative abdomen, bridge with PN and advance EN as perfusion returns; in resolving shock with a soft abdomen and passing flatus, early EN is the defensible default.^[11][12]

Management — Protein: EFFORT, AKI harm and the refeeding-protein warning

Give EFFORT Protein exactly: 1301 high-risk ventilated adults, prescribed 2.2-plus versus 1.2-or-less g per kg per day within 96 hours for up to 28 days — discharge alive by day 60 at 46.1 versus 50.2 percent (HR 0.91) with 60-day death 34.6 versus 32.1 percent (RR 1.08), and harm concentrated in acute kidney injury and high organ-failure scores.^[13] The interpretation is the viva sentence: higher protein did not improve discharge-alive time and might have worsened AKI and high-failure outcomes.^[13] Give the AKI post-hoc exactly: 312 EFFORT patients developing AKI — high protein slowed discharge alive (HR 0.5) and raised 60-day death (RR 1.4) across all AKI stages, with harm disappearing only once kidney replacement therapy started.^[14] Give the Bayesian read: 72 percent posterior probability of mortality harm and 92 percent for delayed discharge, worsening with creatinine and SOFA.^[15]

Set the PermiT-protein null beside it so the dose-response reads correctly: above versus below 0.80 g per kg per day changed nothing on 90-day death (24.2 versus 25.9 percent) — extra protein raised urinary nitrogen without moving prealbumin, transferrin or nitrogen balance.^[4] Then add the refeeding-protein warning, because the surgeon meets it weekly: in refeeding hypophosphatemia, protein above 0.71 g per kg per day in days 1 to 3 more than doubled 6-month death hazard — carbohydrate and lipid did not.^[24] Resolve for the surgeon: prescribe usual-dose protein, go lower and slower with AKI, high SOFA or falling phosphate, and never chase high protein targets in multiorgan failure.^[13][14][24]

Management — Pharmaconutrition in ICU: glutamine, antioxidants and the era-closure

Give REDOXS exactly: 1223 ventilated multiorgan-failure adults, glutamine and antioxidants each versus placebo in factorial design — glutamine trended to higher 28-day death (32.4 versus 27.2 percent, OR 1.28) with significantly higher hospital and 6-month death and no organ-failure or infection effect, while antioxidants changed nothing on 28-day death (30.8 versus 28.8 percent) or any secondary.^[16] The conclusion is the viva sentence: early glutamine or antioxidants did not improve outcomes, and glutamine tracked increased death in multiorgan failure.^[16] Give MetaPlus: 301 prolonged-ventilation patients, immune-modulating high-protein EN versus standard high-protein EN — new infections 53 versus 52 percent with no other endpoint differences except higher adjusted 6-month death in the medical subgroup (HR 1.57).^[17] The review line agrees: pharmaconutrition confers no superior outcomes and may even confer harm — which is exactly why ESPEN still singles out glutamine and omega-3 for special attention, and why the examiner probes whether you know the difference between attention and endorsement.^[32][30]

Management — Perioperative immunonutrition in GI cancer surgery: where benefit lives

Give the GI-cancer meta-analysis exactly: 90 randomised trials with 55 pooled and 7462 patients — perioperative arginine, nucleotide and omega-3 formulas cut anastomotic leak (OR 0.62), respiratory, urinary, wound and sepsis infections, and shortened stay by a mean of 2.47 days — while omega-3 alone improved nothing.^[18] Give the umbrella exactly: 16 reviews with 41 072 GI-cancer surgery patients — infections RR 0.62, leak RR 0.68, stay shorter by 1.92 days.^[19] State the boundary trial so the answer never overreaches: SWOG S1600 randomised 203 cystectomy patients to arginine, omega-3 and nucleotide drinks versus standard nutrition — 30-day complications 62.2 versus 58.0 percent (OR 1.18) with high-grade events 11.1 versus 12.5 percent — no reduction.^[20] Resolve for the surgeon: offer combined immunonutrition perioperatively for gastrointestinal cancer resections; do not extrapolate to cystectomy, to ICU pharmaconutrition, or to single-nutrient omega-3.^[18][20][16]

Management — Refeeding syndrome: restriction trial, incidence and the surgical cohort

Give the Refeeding RCT exactly: 339 ICU adults developing refeeding syndrome within 72 hours of starting nutrition, standard support versus protocolised caloric restriction — days alive after ICU discharge 39.9 versus 44.8 (difference 4.9, nonsignificant), but day-60 survival 78 versus 91 percent with longer overall survival, and no safety concerns with restriction.^[21] The interpretation is the viva sentence: protocolised caloric restriction is a suitable therapeutic option for critically ill adults who develop refeeding syndrome.^[21] Give the incidence frame: across 35 studies the syndrome ranged from 0 to 62 percent with no universal definition — ICU patients fed above 20 kcal per kg per day approached 44 percent syndrome and 27 percent hypophosphatemia.^[22] Give the surgical cohort: 200 at-risk surgical ICU patients, mostly enterally fed — starvation beyond 5 days the dominant risk factor, NICE screening criteria at 33 percent sensitivity and 70 percent specificity, and syndrome confirmation diverging between King and ASPEN criteria (3 versus 25 cases with 1 versus 3 deaths).^[23] The screening lesson: use risk criteria as a guideline to identify at-risk patients, never as a diagnosis — and act on the phosphate fall, which marks restriction-responsive physiology and early-PN harm.^[23][9]

Management — Glycaemic control: NICE-SUGAR, de-adoption and the safe-target close

Give NICE-SUGAR exactly: 6104 ICU adults expected to stay 3 or more days, intensive 81-to-108 versus conventional 180-or-less targets — 90-day death 27.5 versus 24.9 percent (OR 1.14), severe hypoglycaemia 6.8 versus 0.5 percent, with the operative subgroup at OR 1.31 against 1.07 medical (interaction nonsignificant).^[25] The conclusion is the viva sentence: a target of 180 or less beat 81 to 108 on mortality.^[25] Give the de-adoption proof: a 5202-patient ICU switching from modified-intensive to conventional control raised time-weighted glucose yet cut adjusted 90-day death odds by 47 percent, with severe hypoglycaemia falling from 1.2 to 0.4 percent and moderate from 23.3 to 5.9 percent.^[26] State the conditional line for the examiner who cites the Leuven era: tight control may be superior only inside a protocol that avoids hypoglycaemia with regular accurate measurement and no insulin boluses — without such a protocol, avoid severe hyperglycaemia and all hypoglycaemia.^[33] For the steroid-treated, septic postoperative patient this means: cap at 180 or less, measure often, and treat every low reading as the complication that kills.^[25][33]

Management — Tubes, tolerance and protocols: access, prokinetics and the negative protocol

Give the postpyloric meta-analysis exactly: 17 randomised trials — postpyloric delivery raised delivered energy by 12 percent and cut residuals by 169 mL, yet changed nothing on mortality, new pneumonia or aspiration.^[27] Give the jejunal RCT: 110 patients across 11 hospitals, gastric versus jejunal feeding — GI complications 57 versus 24 percent favouring jejunal via fewer high residuals, but pneumonia (40 versus 32 percent), stay and death all unchanged, with better day-7 delivery only in units already experienced in jejunal feeding.^[28] The conclusion pair for viva: jejunal feeding brings fewer GI complications with a learning curve, but nasojejunal routing is not an efficacious pneumonia-prevention measure.^[28] Give the prokinetics meta-analysis: 13 trials with 1341 patients — feeding intolerance down (RR 0.73, 17.3 percent absolute), high residuals down (RR 0.69), post-pyloric placement success up (RR 1.60), with no change in vomiting, diarrhoea, stay or death.^[29] Quote the negative protocol truthfully, because the surgeon runs the unit: a 170-patient surgical-ICU nutrition protocol delivered near-identical calories to clinician discretion (about 775 versus 773 kcal per day) with no feeding or outcome gain — protocols organise care, they do not substitute for the six decisions above.^[34]

Complications & Pitfalls — the six traps

The shock-gut forcing trap — early isocaloric EN in shock raised bowel ischaemia nearly four-fold with more vomiting, diarrhoea and pseudo-obstruction and no mortality or infection gain.^[11] Examine the distended vasopressor-dependent abdomen, image ischaemia early, and bridge with PN — never prokinet through a dead gut.^[11][29]

The early-PN topping trap — adding PN within 48 hours to failing EN slowed recovery, raised ICU infections and cholestasis, prolonged ventilation and added renal-replacement days; the deficit of week one is tolerated physiology, not a debt to repay.^[6]

The high-protein trap — 2.2-plus g per kg per day did not hasten discharge and raised death overall, slowed discharge and raised death in AKI at every stage, and carried 72 percent posterior harm probability overall and 92 percent for delayed discharge.^[13][14][15]

The glutamine-cocktail trap — glutamine in multiorgan failure tracked higher hospital and 6-month death with no organ or infection benefit, and immune-modulating EN suggested late harm; keep immune formulas in the GI-cancer perioperative lane where meta-analyses favour them.^[16][17][18]

The refeeding-blindness trap — full calories into the starved patient without phosphate monitoring invites the 44-percent-incidence physiology; the restriction trial, the phosphate secondary and the protein-in-RH cohort all point the same way — restrict, replete, go slow.^{[22][21][9][24]}

The tight-glucose trap — 81-to-108 targeting raised death 27.5 versus 24.9 percent through thirteen-fold severe hypoglycaemia, and the operative subgroup fared numerically worse; cap at 180 or less unless a hypoglycaemia-proof tight protocol truly runs on your unit.^[25][33]

Prognosis & Disposition — the numbers that set expectations

^{[1] [3] [6] [11] [13] [25]}

Every high-risk fed ICU patient needs intensivist, surgical, dietetic and pharmacy input with daily review of perfusion, tolerance, phosphate and glucose — set-and-forget feeding after laparotomy is a failure-to-rescue setup. Modifiers: vasopressor trajectory (shock gut), AKI and SOFA (protein harm), phosphate fall (restriction), starvation history (refeeding), and diabetes with steroids or sepsis (glucose volatility).^{[11][13][9][25]}

Special Populations — surgical contexts that change the emphasis

• Vasopressor-dependent shock: the NUTRIREA-2 population — isocaloric EN raised digestive harm without survival gain, so PN bridging with EN advanced on perfusion recovery is the defensible sequence.^[11]
• Acute kidney injury: the EFFORT harm subgroup — high protein slowed discharge and raised death at every AKI stage except on replacement therapy, so dose protein down and recheck.^[14]
• GI cancer resection: the immunonutrition population — perioperative arginine, nucleotide and omega-3 formulas cut leak and infection and shortened stay, while single-nutrient omega-3 and non-GI operations show no such benefit.^[18][20]
• Starved and obstructed presentations: the refeeding population — starvation beyond 5 days dominates risk, restriction improved survival, and early protein above 0.71 g per kg per day in hypophosphatemia more than doubled late death hazard.^[23][21][24]
• High organ-failure burden: the EFFORT-Bayes interaction — higher creatinine and SOFA scores steepen protein harm, so the sicker the patient the more conservative the protein.^[15]
• Post-laparotomy open or tenuous abdomen: no verified feeding-timing numbers sit in this pack for this anatomy — apply shock-gut and EPaNIC principles, protect the fascia, feed distally where feasible, and name the evidence gap rather than inventing numbers.
• Diabetes with sepsis or steroids: the NICE-SUGAR operative signal plus transition-cohort proof — cap at 180 or less, measure often, and let hypoglycaemia avoidance dominate every insulin decision.^[25][26]

Evidence, Guidelines & Regional Differences — the six stories and who led them

• The enteral-dose story (US-led): EDEN 1000-patient trophic-versus-full tie with less intolerance → PermiT 894-patient permissive-versus-standard tie with equal protein → risk-stratified post-hocs confirming no subgroup to force-feed → postoperative reviews synthesising dose-dependent harm.^{[1][3][5][33]}
• The parenteral-timing story (Belgian-led): EPaNIC 4640-patient late-PN win on speed and complications → weakness subanalysis supplying the autophagy mechanism → phosphate secondary supplying the ready-to-feed signal → 2-year follow-up bounding the claim to early recovery.^[6][8][9][7]
• The route story (UK/French): CALORIES 2400-patient PN-versus-EN equivalence on death with tolerance differences → NUTRIREA-2 2410-patient shock trial reversing the default toward PN in shock gut → shock-timing cohort leaving early-versus-delayed EN unrandomised.^[10][11][12]
• The protein story (international): PermiT-protein null at moderate doses → EFFORT high-dose negative with AKI harm → Bayesian confirmation with creatinine and SOFA interactions → refeeding-hypophosphatemia cohort warning against early protein specifically.^{[4][13][15][24]}
• The pharmaconutrition split (North America/Europe versus surgical oncology): REDOXS glutamine harm and MetaPlus immune-EN harm in ICU multiorgan failure → GI-cancer meta-analyses favouring combined immunonutrition → SWOG cystectomy negative bounding the benefit to GI surgery.^{[16][17][18][20]}
• The glucose story (ANZ-led): NICE-SUGAR 6104-patient intensive-control harm → transition-cohort de-adoption halving adjusted death odds → conditional tight-control line requiring a hypoglycaemia-proof protocol.^[25][26][33]
• Sibling-topic boundary: shock owns pressors and perfusion targets; ARDS owns ventilation and proning; postoperative sepsis owns source control and antimicrobials — this topic owns dose, timing, route, protein, refeeding and glucose. Cite each where it lives.
• Guideline honesty: ESPEN defines at-risk patients, assessment, energy, route and special conditions including abdominal surgery and sepsis, and the French guideline fields 34 adult recommendations mostly on moderate evidence or expert opinion — both direct individualisation, and neither supplies the numeric formulas this topic declines to invent.^[30][31]

Exam Pearls — the one-liners that score

• EDEN in one breath: 1000 patients, 400 vs 1300 kcal for 6 days — vent-free 14.9 vs 15.0, death 23.2 vs 22.2, intolerance worse on full.^[1]
• PermiT in one breath: 894 patients, 46 vs 71 percent of requirements with equal protein — 90-day death 27.2 vs 28.9; risk scores do not prescribe.^[3][5]
• EPaNIC in one breath: 2312 early vs 2328 late PN — late wins speed, infections 22.8 vs 26.2, cholestasis, ventilation, 3 renal days, EUR 1,110; 2-year equal.^[6][7]
• CALORIES in one breath: 2400 patients, PN vs EN — 30-day death 33.1 vs 34.2; less vomiting and hypoglycaemia on PN, targets unmet either way.^[10]
• NUTRIREA-2 in one breath: 2410 shock patients, isocaloric EN vs PN — 37 vs 35 percent dead, infections equal, bowel ischaemia excess on EN.^[11]
• EFFORT in one breath: 1301 high-risk patients, 2.2-plus vs 1.2-or-less protein — discharge 46.1 vs 50.2, death 34.6 vs 32.1; AKI RR 1.4, Bayes 72 percent harm.^[13][14][15]
• REDOXS and MetaPlus in one breath: glutamine trends death up (32.4 vs 27.2) with higher late death; immune EN 53 vs 52 percent infections with medical-subgroup late harm.^[16][17]
• GI immunonutrition in one breath: leak OR 0.62, infections down, stay shorter by about 2 days — omega-3 alone nil, cystectomy nil.^[18][20]
• Refeeding in one breath: restriction beats standard (day-60 survival 91 vs 78 percent); ICU high-feed incidence near 44 percent; phosphate fall means restrict.^[21][22][9]
• NICE-SUGAR in one breath: 6104 patients, 81-to-108 vs 180-or-less — death 27.5 vs 24.9, severe hypo 6.8 vs 0.5; switch cohorts halve adjusted death.^[25][26]
• Tubes and tolerance in one breath: postpyloric gains delivery not survival; jejunal cuts GI hassle not pneumonia; prokinetics cut intolerance (RR 0.73) not death.^[27][28][29]

Revision summary

Start EN early but run it trophic-to-moderate (EDEN 400 vs 1300 with tied vent-free days and death; PermiT 46 vs 71 percent with tied 90-day death; risk scores describe without prescribing).^[1][3][5] Withhold PN top-ups until day 8 (EPaNIC faster live discharge with fewer infections and less ventilation; weakness 34 vs 43 percent via autophagy; 2-year convergence), and treat routes as tolerance choices outside shock (CALORIES 33.1 vs 34.2 percent).^{[6][8][7][10]} In vasopressor-dependent shock, bridge with PN and advance EN on perfusion (NUTRIREA-2 isocaloric EN raised bowel ischaemia with equal death; shock-timing cohort unadjusted only).^[11][12] Dose protein at usual levels — high dose missed discharge benefit overall, harmed AKI at RR 1.4 with 72 percent Bayesian harm probability, and early protein in refeeding hypophosphatemia more than doubled late death hazard.^{[13][14][15][24]} Keep glutamine and antioxidant cocktails out of multiorgan failure (REDOXS late-death excess; MetaPlus null with subgroup harm) while offering combined immunonutrition for GI cancer resections (leak OR 0.62, shorter stay; cystectomy and single-nutrient exceptions noted).^{[16][17][18][20]} Restrict calories when refeeding declares itself (trial survival 91 vs 78 percent; ICU high-feed incidence 44 percent; phosphate-fall action rule) and cap glucose at 180 or less (NICE-SUGAR death excess with thirteen-fold severe hypoglycaemia; transition cohorts confirm de-adoption benefit).^{[21][22][9][25][26]}

Refeeding hypophosphataemia meant a phosphate fall of more than 0.16 mmol/L to below 0.65 mmol/L in the trials that moved practice.^[9] EDEN: 60-day death 23.2% versus 22.2% on trophic versus full feeding.^[1] EPaNIC cut ICU infections 22.8% versus 26.2% with late PN.^[6] CALORIES: 30-day death 33.1% versus 34.2%, parenteral versus enteral.^[10]

isocaloric EN versus PN at 20 to 25 kcal/kg/d within 24 hours of intubation in shock — day-28 death 37% versus 35% with equal ICU infections, but more vomiting, diarrhoea, bowel ischaemia and pseudo-obstruction on EN (PMID 29128300).^[11] 2.2-plus versus 1.2-or-less g/kg/d in high-risk ventilated adults — discharge-alive 46.1% versus 50.2% and 60-day death 34.6% versus 32.1%, with harm concentrated in AKI and high organ-failure scores (PMID 36708732).^[13] glutamine trended to higher 28-day death (32.4% versus 27.2%, OR 1.28) with significantly higher hospital and 6-month death, and antioxidants changed nothing (PMID 23594003).^[16] immune-modulating high-protein EN changed nothing on new infections (53% versus 52%) and raised adjusted 6-month death in the medical subgroup (HR 1.57) (PMID 25096691).^[17] SWOG S1600 cystectomy IMN changed nothing on 30-day complications (62.2% versus 58.0%, OR 1.18) — benefit evidence is GI-cancer-specific, not universal to all cancer surgery (PMID 42475098).^[20] 200 at-risk surgical ICU patients, starvation beyond 5 days the dominant risk factor (55%), enteral feeds in 84.5%, with NICE criteria at 33% sensitivity and 70% specificity — a screening aid, not a diagnosis (PMID 34857199).^[23] intensive 81 to 108 versus conventional 180-or-less targets — 90-day death 27.5% versus 24.9% (OR 1.14), severe hypoglycaemia 6.8% versus 0.5%, with the surgical subgroup at OR 1.31 (PMID 19318384).^[25] 110 patients across 11 hospitals — jejunal feeding cut GI complications (57% versus 24%) via fewer high residuals but left pneumonia (40% versus 32%), stay and death unchanged, with a learning-curve qualifier (PMID 11940748).^[28]

in refeeding hypophosphatemia, early protein above 0.71 g/kg/d more than doubled 6-month death hazard — carbohydrate and lipid did not (PMID 37202045).^[24]

References34ShowHide

1. [1]Rice TW, Wheeler AP, Thompson BT, et al. Initial trophic vs full enteral feeding in patients with acute lung injury: the EDEN randomized trial. JAMA, 2012.PMID 22307571
2. [2]Needham DM, Dinglas VD, Morris PE, et al. Physical and cognitive performance of patients with acute lung injury 1 year after initial trophic versus full enteral feeding. EDEN trial follow-up. Am J Respir Crit Care Med, 2013.PMID 23805899
3. [3]Arabi YM, Aldawood AS, Haddad SH, et al. Permissive Underfeeding or Standard Enteral Feeding in Critically Ill Adults. N Engl J Med, 2015.PMID 25992505
4. [4]Arabi YM, Al-Dorzi HM, Mehta S, et al. Association of protein intake with the outcomes of critically ill patients: a post hoc analysis of the PermiT trial. Am J Clin Nutr, 2018.PMID 30475959
5. [5]Arabi YM, Aldawood AS, Al-Dorzi HM, et al. Permissive Underfeeding or Standard Enteral Feeding in High- and Low-Nutritional-Risk Critically Ill Adults. Post Hoc Analysis of the PermiT Trial. Am J Respir Crit Care Med, 2017.PMID 27589411
6. [6]Casaer MP, Mesotten D, Hermans G, et al. Early versus late parenteral nutrition in critically ill adults. N Engl J Med, 2011.PMID 21714640
7. [7]Casaer MP, Stragier H, Hermans G, et al. Impact of withholding early parenteral nutrition on 2-year mortality and functional outcome in critically ill adults. Intensive Care Med, 2024.PMID 39017697
8. [8]Hermans G, Casaer MP, Clerckx B, et al. Effect of tolerating macronutrient deficit on the development of intensive-care unit acquired weakness: a subanalysis of the EPaNIC trial. Lancet Respir Med, 2013.PMID 24461665
9. [9]Lauwers C, Langouche L, Wouters PJ, et al. Early phosphate changes as potential indicator of unreadiness for artificial feeding: a secondary analysis of the EPaNIC RCT. Crit Care, 2025.PMID 39875953
10. [10]Harvey SE, Parrott F, Harrison DA, et al. Trial of the route of early nutritional support in critically ill adults. N Engl J Med, 2014.PMID 25271389
11. [11]Reignier J, Boisramé-Helms J, Brisard L, et al. Enteral versus parenteral early nutrition in ventilated adults with shock: a randomised, controlled, multicentre, open-label, parallel-group study (NUTRIREA-2). Lancet, 2018.PMID 29128300
12. [12]Ortiz-Reyes L, Patel JJ, Jiang X, et al. Early versus delayed enteral nutrition in mechanically ventilated patients with circulatory shock: a nested cohort analysis of an international multicenter, pragmatic clinical trial. Crit Care, 2022.PMID 35681220
13. [13]Heyland DK, Patel J, Compher C, et al. The effect of higher protein dosing in critically ill patients with high nutritional risk (EFFORT Protein): an international, multicentre, pragmatic, registry-based randomised trial. Lancet, 2023.PMID 36708732
14. [14]Stoppe C, Patel JJ, Zarbock A, et al. The impact of higher protein dosing on outcomes in critically ill patients with acute kidney injury: a post hoc analysis of the EFFORT protein trial. Crit Care, 2023.PMID 37853490
15. [15]Haines RW, Granholm A, Puthucheary Z, et al. The effect of high protein dosing in critically ill patients: an exploratory, secondary Bayesian analyses of the EFFORT Protein trial. Br J Anaesth, 2024.PMID 39455305
16. [16]Heyland D, Muscedere J, Wischmeyer PE, et al. A randomized trial of glutamine and antioxidants in critically ill patients. N Engl J Med, 2013.PMID 23594003
17. [17]van Zanten AR, Sztark F, Kaisers UX, et al. High-protein enteral nutrition enriched with immune-modulating nutrients vs standard high-protein enteral nutrition and nosocomial infections in the ICU: a randomized clinical trial. JAMA, 2014.PMID 25096691
18. [18]Budai BC, Panait R, Laczkó B, et al. Immunonutrition Decreases Postoperative Complications in Gastrointestinal Cancer-A Systematic Review and Meta-analysis of Randomized Controlled Trials. Adv Nutr, 2026.PMID 42331289
19. [19]Goyal A, Macias CA, Corzo MP, et al. Perioperative Immunonutrition in Gastrointestinal Oncology: A Comprehensive Umbrella Review and Meta-Analysis on Behalf of TROGSS-The Robotic Global Surgical Society. Nutrients, 2025.PMID 40732929
20. [20]Hamilton-Reeves JM, Unger JM, Holzbeierlein JM, et al. Immune-Enhancing Nutrition and Outcomes After Radical Cystectomy: A Randomized Clinical Trial. JAMA Netw Open, 2026.PMID 42475098
21. [21]Doig GS, Simpson F, Heighes PT, et al. Restricted versus continued standard caloric intake during the management of refeeding syndrome in critically ill adults: a randomised, parallel-group, multicentre, single-blind controlled trial. Lancet Respir Med, 2015.PMID 26597128
22. [22]Cioffi I, Ponzo V, Pellegrini M, et al. The incidence of the refeeding syndrome. A systematic review and meta-analyses of literature. Clin Nutr, 2021.PMID 34134001
23. [23]Buitendag J, Variawa S, Davids R, et al. Refeeding syndrome in surgical patients post initiation of artificial feeding, a prospective cohort study in a low-income country. Clin Nutr ESPEN, 2021.PMID 34857199
24. [24]Slingerland-Boot R, Rooijakkers E, Koekkoek K, et al. Macronutrient intake and outcomes of ICU patients with refeeding hypophosphatemia. Clin Nutr ESPEN, 2023.PMID 37202045
25. [25]Finfer S, Chittock DR, Su SY, et al. Intensive versus conventional glucose control in critically ill patients. N Engl J Med, 2009.PMID 19318384
26. [26]Orford NR, Bailey M, Kaukonen K, et al. Glycaemic control and long-term outcomes following transition from modified intensive insulin therapy to conventional glycaemic control. Anaesth Intensive Care, 2014.PMID 24580391
27. [27]Zhang Z, Xu X, Ding J, et al. Comparison of postpyloric tube feeding and gastric tube feeding in intensive care unit patients: a meta-analysis. Nutr Clin Pract, 2013.PMID 23614960
28. [28]Montejo JC, Grau T, Acosta J, et al. Multicenter, prospective, randomized, single-blind study comparing the efficacy and gastrointestinal complications of early jejunal feeding with early gastric feeding in critically ill patients. Crit Care Med, 2002.PMID 11940748
29. [29]Lewis K, Alqahtani Z, Mcintyre L, et al. The efficacy and safety of prokinetic agents in critically ill patients receiving enteral nutrition: a systematic review and meta-analysis of randomized trials. Crit Care, 2016.PMID 27527069
30. [30]Singer P, Blaser AR, Berger MM, et al. ESPEN guideline on clinical nutrition in the intensive care unit. Clin Nutr, 2019.PMID 30348463
31. [31]Reignier J, Gaillard-Le Roux B, Dequin PF, et al. Expert consensus‑based clinical practice guidelines for nutritional support in the intensive care unit: the French Intensive Care Society (SRLF) and the French-Speaking Group of Pediatric Emergency Physicians and Intensivists (GFRUP). Ann Intensive Care, 2025.PMID 40665004
32. [32]Koekkoek KW, van Zanten AR. Nutrition in the critically ill patient. Curr Opin Anaesthesiol, 2017.PMID 28151828
33. [33]Vanderwegen E, Vanhorebeek I, Gunst J. Postoperative metabolic support in the ICU. Best Pract Res Clin Anaesthesiol, 2026.PMID 42552018
34. [34]Chinda P, Poomthong P, Toadithep P, et al. The implementation of a nutrition protocol in a surgical intensive care unit; a randomized controlled trial at a tertiary care hospital. PLoS One, 2020.PMID 32298381