Postoperative Sepsis — Fever Workup, Scores, Hour-1 Resuscitation, Source Control and the Device/Leak Sources

Day 2 after anterior resection: 37.8 C, chest clear, urine flowing, wound dry. Day 5: 38.9 C with tachycardia, a rigid lower abdomen and a rising CRP. The first fever was physiology; the second is a leak until proven otherwise. This page teaches the exact line between them — the day cutoffs, the scores by location, the hour-1 evidence with its honest counterweight, and the source-control clock — with every number taken from the papers named beside it.^[12][15][23]

Overview & Definition — Sepsis-3, shock, and the postoperative taxonomy

Sepsis is life-threatening organ dysfunction caused by a dysregulated host response to infection.^[1] Operationally, organ dysfunction is an increase in the Sequential Organ Failure Assessment (SOFA) score of 2 points or more, which is associated with an in-hospital mortality greater than 10%.^[1] Septic shock is the subset in which profound circulatory, cellular and metabolic abnormalities carry greater mortality than sepsis alone — identified at the bedside by a vasopressor requirement holding mean arterial pressure at 65 mmHg or greater plus serum lactate above 2 mmol/L in the absence of hypovolaemia, a combination with hospital mortality above 40%.^[1] The derivation behind that pair is explicit: the vasopressor-plus-lactate group died at 42.3% against the alternatives, and the systematic review behind the definition found crude septic-shock mortality of 46.5%.^[3]

Outside intensive care, the prompt is the quick SOFA: respiratory rate 22 per minute or greater, altered mentation, or systolic blood pressure 100 mmHg or less — two or more flags a ward patient as likely to suffer poor outcomes typical of sepsis.^[1] Each element scores one point on a 0-to-3 scale built from 1.3 million encounters.^[2] The surgical footnote belongs in the same breath: the Global Alliance for Infections in Surgery welcomed Sepsis-3 as an important step forward yet recorded concerns about its use in surgical patients — quote the definitions, then show you know their limits.^[9]

Classify the source the surgical way. Intra-abdominal infections divide into uncomplicated versus complicated disease causing localized or diffuse peritonitis — and complicated infection is the second most common sepsis-related death, at 23 to 38% hospital mortality.^[46] Beyond the abdomen, the US surveillance taxonomy weights your pre-test probability: among healthcare-associated infections, pneumonia and surgical-site infections each account for 21.8%, gastrointestinal infections 17.1%, and Clostridioides difficile is the single commonest pathogen at 12.1%.^[11]

Epidemiology & Risk Factors — the denominators that anchor timing arguments

Sepsis surveillance gives the scale: 173,690 cases identified by clinical criteria among 2,901,019 adult admissions in 2014 — a 6.0% incidence — of whom 15.0% died in hospital and 6.2% were discharged to hospice.^[8] The trend behind that number is the viva trap: clinical-criteria incidence was stable at plus 0.6% per year while claims-based incidence rose 10.3% per year — coding inflates, physiology does not.^[8]

Point prevalence on any given day: 452 of 11,282 inpatients (4.0%) carried at least one healthcare-associated infection across 183 hospitals.^[11] After abdominal surgery specifically, the Italian ICU registry enrolled 784 complicated intra-abdominal infections across 23 hospitals: in-hospital mortality 23.9%, with septic shock (36.2%) and sepsis (35.9%) the leading reasons for ICU admission, adequate source control achieved in only 61.5%, re-operation required in 21%, and Escherichia coli (23.1%) followed by Enterococcus species (15.4%) as the commonest isolates.^[48] The French postoperative cohort sharpens the post-laparotomy picture: 271 healthcare-associated intra-abdominal infections after initial abdominal surgery, 57% in septic shock, 90-day mortality 14.4%.^[47]

Risk concentrates, not spreads. Independent predictors of septic shock after postoperative intra-abdominal infection are ASA class above II (odds ratio 2.48), antibiotic exposure within 48 hours (1.85), fungal colonization (4.37) and E. coli bacteraemia (3.25) — with absence of cancer paradoxically predictive (2.03), a finding to reason through rather than recite.^[47] And the rescue arithmetic every fellow must carry: among 312,862 emergency general surgery cases, 16,306 required readmission, mortality after readmission was 2.4%, and 90.6% of those deaths were attributable to failure-to-rescue — with frailty, high-risk procedures, pulmonary complications, AKI, sepsis and reoperation marking the patients rescue fails.^[10]

Pathophysiology — dysregulation, lactate kinetics, and why anastomoses leak

The definition names the mechanism: a host response to infection that escapes containment and produces organ dysfunction — which is why inflammation alone was abandoned as the model and why anti-inflammatory logic never became therapy.^[1] In the postoperative abdomen that escape runs contamination, mediator release, third-spacing and progressive multi-organ failure — a loop antibiotics slow but only source control breaks.

Read lactate as two different signals across time. Clearance follows two phases: an initial 6-to-12-hour predominantly flow-dependent phase where elevation reflects hypoperfusion and answers fluids, pressors and oxygen delivery — then a later phase where persistent hyperlactataemia more often reflects metabolic reprogramming, immune activation and mitochondrial dysfunction.^[55] The bedside consequence is blunt: persistent hyperlactataemia without hypoperfusion may be misleading and should not drive further resuscitation.^[55] Excessive beta-stimulation, particularly epinephrine, can raise lactate without any metabolic shift at all.^[55]

The leak has its own biology, and the examiner will ask why technique has not fixed it. Anastomotic healing fails through a complex interplay of host genetics, gut microbiome, inflammation and immunity — and despite technical improvements and identified risk factors, incidence has not significantly changed over time.^[25] Dogma that failed this biology is now explicit: routine intra-abdominal drains and mechanical bowel preparation alone are useless against leak and should be abandoned, while fluorescence angiography earns its place by changing the operative strategy mid-case.^[25]

Clinical Presentation — fever timing, the POD rules, and the leak spectrum

Fever is one of the most common postoperative complications — and only about 40% of inpatient fever episodes are infectious.^[13][12] That single fraction reframes the whole workup: most early fever is physiology, atelectasis, haematoma, drugs or inflammation — but every fever still obliges you to exclude the life threat hiding underneath.

Two day-cutoffs govern the bedside. New or continuing fever more than three days after surgery should raise strong suspicion of persistent illness or a new complication.^[12] The liver-resection denominator proves the rule quantitatively: 50.3% of 797 patients developed postoperative fever, yet only 14.6% suffered an infectious complication — 21.9% of febrile patients versus 7.1% of afebrile ones — with three independent predictors of a febrile infectious complication: temperature above 38.6 C (odds ratio 2.24), first fever after postoperative day 2 (odds ratio 6.00), and multiple fever spikes (odds ratio 2.04).^[15] First fever after day 2 is the single strongest flag — a sixfold signal.

The leak declares itself on a slower clock. Its clinical spectrum runs from a radiological-only finding through peritonitis to sepsis with multi-organ failure.^[24] Pooled diagnosis lands at a mean of 7.7 days — which is why the POD-5 fever after colorectal anastomosis is a leak hunt, not a reassurance exercise.^[23]

^{[12] [13] [1] [11] [10]}

Differential Diagnosis — infectious versus non-infectious, and the six-source hunt

Separate physiology from pathology first, then infectious from non-infectious causes — the systematic approach that limits costly workups and improves outcomes.^[13] Non-infective fever is common and expected: inflammatory response to tissue injury, atelectasis, haematoma resorption, drug fever, venous thromboembolism, transfusion reaction. Testing follows clinical findings — subsequent testing should be based on the clinical findings, full stop.^[12]

When fever does mean infection, hunt six sources in order of the surveillance weights: wound and surgical-site infection, anastomotic leak or undrained collection, pneumonia, urinary infection, line infection, and C. difficile colitis.^[11] Distinguish leak from routine ileus by trajectory plus biomarkers: CRP and procalcitonin rising from postoperative day 2 to 3 with CT signs means leak until proven otherwise; a self-limited course with falling markers means ileus.^[24][26]

^{[12] [15] [14]}

Clinical & Bedside Assessment — the right score in the right place

Match the score to the location — the surgical cohort proved they are not interchangeable. In 13,780 surgical intermediate and intensive care patients, suspected infection touched 18.3% of intermediate-care, 35.5% of ICU and 62.0% of combined encounters, with 3.3% overall mortality — and qSOFA predicted mortality best on the intermediate unit (area under the curve 0.82) while SOFA won once ICU care was involved (0.73).^[5] So: qSOFA on the ward, SOFA in the unit.

The ICU validation behind SOFA is absolute: among 184,875 infected ICU admissions, 18.7% died; SOFA rise of 2 or more was present in 90.1% and discriminated mortality (AUROC 0.753) far above SIRS (0.589) or qSOFA (0.607) — both of which have limited mortality utility once the patient is in ICU.^[4] The 57-study synthesis completes the operating characteristics: SIRS is sensitive (0.85) but nonspecific (0.41), qSOFA specific (0.98) but insensitive (0.42), NEWS balanced (0.71 sensitivity, 0.85 specificity), and SOFA best for inpatient mortality (0.89 and 0.69).^[6] Biomarkers alone lose the face-off: across 29 studies and 41,469 patients, SOFA (AUROC 0.819) significantly outperformed procalcitonin and lactate — though lactate-adjusted qSOFA matched SOFA at 0.823.^[7]

Investigations — CRP and procalcitonin timing, CT signs, lactate, cultures

Time the leak markers, do not just order them. C-reactive protein and procalcitonin are early leak predictors starting from postoperative day 2 to 3 — but contrast-enhanced abdominal-pelvic CT remains the gold standard for diagnosis.^[24] The meta-analytic cutoff is quotable: pooled diagnosis at 7.7 days, and a POD-3 CRP of 148 mg/L carries 95% sensitivity and 95% specificity for colorectal leak.^[23] On CT, hunt fluid collections, pneumoperitoneum, extraluminal contrast extravasation and abscess formation — then integrate markers with imaging into a patient-specific call rather than trusting either alone.^[26]

Use lactate as a context, not a number. Initial and 2-hour lactate predict mortality alongside SOFA — but clearance must be read against perfusion: hyperlactataemia with abnormal capillary refill, low central venous saturation or wide veno-arterial CO2 gap means ongoing hypoperfusion, while isolated persistent elevation means metabolism, not flow.^[55] As a resuscitation target, lactate clearance of at least 10% proved non-inferior to central venous oxygen saturation of 70% or more (23% versus 17% mortality, no significant difference, 300 patients) — yet the pooled signal across 7 trials favours clearance-guided therapy (mortality relative ratio 0.68, ICU stay shorter by 1.64 days, ventilation shorter by 10.22 hours).^[52][53]

Culture before antibiotics, but respect the yield fence. Blood cultures in blind fever workups are positive in 0 to 13.3% of cases, and only two culture-positive patients across the orthopaedic systematic review developed clinical sepsis.^[14] So: cultures first, antibiotics inside the hour regardless, and a negative panel never excludes surgical sepsis — the abdomen, not the laboratory, rules out a leak.^[39][24]

Management — Resuscitation: the hour-1 bundle and its honest counterweight

Give antibiotics inside the first hour. Across 17,990 analysed severe-sepsis and septic-shock patients with 29.7% mortality, hospital mortality adjusted for severity, admission source and region increased steadily after 1 hour of time to antibiotic administration.^[39] The per-hour price is pooled: each hour of door-to-antibiotic delay raises mortality odds by 1.07 across 15 studies and 106,845 patients — a number needed to treat of 91 per hour saved.^[41]

Stratify by shock, because the benefit is not uniform. In a 3,035-patient prospective multicentre cohort, antibiotics within 1 hour carried adjusted mortality odds of 0.78 overall — 0.66 and significant in septic shock, 0.85 and non-significant without shock — and shock patients paid 35% more mortality for every hour delayed inside three hours.^[40] Systems still fail this standard: in one shock cohort only 26.4% received antibiotics within an hour of vasopressor start, while SOFA of 11 or more, qSOFA of 3 and lactate of 4 or more marked the patients who died.^[43]

Now carry the counterweight, because the examiner will. A 482-patient Sepsis-3 prospective cohort with median time to antibiotics of 115 minutes found neither timing nor appropriateness associated with 7-, 14- or 28-day mortality in multivariate analysis.^[42] The fellowship reconciliation: a single-centre cohort cannot refute the multi-thousand-patient gradients — but it proves antibiotics without source control change nothing, which is why the surgical hunt runs inside the same hour.^[39][42][44]

Resuscitate to perfusion with norepinephrine first. Against dopamine, norepinephrine cut all-cause mortality (relative risk 0.89, absolute reduction 11%, number needed to treat 9) with fewer major adverse events and arrhythmias across 32 trials and 3,544 patients.^[57] The dopamine side is explicit: increased death in randomized trials (relative risk 1.12) and arrhythmias more than doubled where reported (relative risk 2.34).^[56] Hold mean arterial pressure at 65 mmHg or more, then stop chasing pressure and chase perfusion — mentation, urine, lactate kinetics in context, capillary refill — because the ANDROMEDA Bayesian reanalysis gives above-90% posterior probability that peripheral-perfusion-targeted resuscitation beats lactate-targeted resuscitation at 28 days (odds ratio 0.61).^[1][54][55]

Management — Definitive & Stepwise: source control, short courses, leaks, wounds, lungs, gut, lines

Run the GAIS cornerstones in order: rapid accurate diagnostics; timely adequate source control; appropriate short-duration antimicrobials by pharmacokinetic and stewardship principles; and fluids plus vasopressors for critical illness.^[45] Each has its number below.

Source control: urgent beats crash. In 1,077 secondary-peritonitis ICU patients with 29.7% mortality, urgent control at 2 to 6 hours from diagnosis carried half the mortality odds of emergency control inside 2 hours (odds ratio 0.50) — resuscitate, image, plan, then operate — while failed control with persistent inflammation at day 7 multiplied death odds 5.71-fold and late-onset hospital-acquired peritonitis raised them 1.71-fold.^[44] The Italian registry concurs from the other side: adequate control was achieved in only 61.5%, re-operation ran at 21%, and effective control protected strongly in the French cohort (odds ratio 0.22).^[48][47]

Antibiotics: four days after adequate control. STOP-IT randomized 518 complicated intra-abdominal infections: fixed 4-day courses matched 8-day courses to resolution of physiological abnormality (21.8% versus 22.3% for infection, recurrence or death) while halving exposure (median 4 versus 8 days).^[49] The planned re-laparotomy extension holds the line in its fenced single-centre form: short (5 days or fewer) versus long courses recurred equally (41.2% versus 44.4%).^[50] Real-world adoption lags the evidence — antibiotic days fell after STOP-IT with no change in stay, abscess or death, yet short-course compliance reached only 30 to 52%.^[51] Non-resolution means re-look, never another week of drugs.^[44]

Leak: detect early, individualize management. Screen with CRP and procalcitonin from day 2 to 3, confirm with contrast CT, and manage by condition, defect size and location, index indication and proximal diversion — with a diverting stoma in place, reoperation for sepsis control is rarely necessary, especially for extraperitoneal anastomoses.^[24] Score risk preoperatively with the named systems (Colon Leakage Score and its modified form, REAL, PROCOLE) rather than gestalt.^[28]

Wound: debride established infection, stop prophylaxis at 24 hours. Established surgical-site infection needs aggressive debridement plus effective antimicrobials.^[18] Prevention rests on strong evidence only: bowel preparation with oral antibiotics in colorectal surgery, preoperative smoking cessation, timely prophylactic antibiotics, chlorhexidine skin antisepsis and perioperative normothermia — against a background where 2 to 5% of operations still infect.^[17] The WHO position is 13 preoperative recommendations as the global backbone.^[16] Everything beyond that is contested: eight guidelines agree homeostasis matters but diverge on oxygenation, volume, temperature and glucose targets.^[19] And the prophylaxis stop-rule is absolute: benefit ends at 24 hours, longer courses do not cut SSI, and each extra day raises kidney injury (numbers needed to harm 9, 6 and 4) and C. difficile (2000, 90 and 50).^[20]

Lungs: bundle the ward, ventilate honestly, rescue with NIV. The ward bundle that works is unglamorous — oral care, incentive spirometry, coughing and deep breathing, activity, head-up positioning, family education — as implemented for hepatectomy and Whipple patients.^[31] What does not work is routine escalation: prophylactic CPAP, NIV or high-flow oxygen did not prevent pneumonia (4.9% versus 5.5% across 38 trials and 9,782 patients), the PRIME-AIR lung-expansion bundle did not reduce pulmonary-complication severity (1.60 versus 1.53), and low tidal volumes (6 versus 10 mL/kg at equal PEEP of 5) changed nothing (38% versus 39%).^[29][30][33] Reserve non-invasive ventilation for treatment, not prevention: for established post-extubation failure NIV halves re-intubation (odds ratio 0.49) and cuts nosocomial pneumonia, ICU stay and mortality — in high-risk patients, not low-risk ones.^[32]

Gut: expect C. difficile late, after discharge. Postoperative C. difficile strikes at a median of 17 days with nearly half the cases after discharge, adds 12 days of stay, and triples readmission and quintuples mortality — driven by prophylaxis beyond 2 days, ICU stay beyond 2 days and hyperglycaemia.^[34] After cystectomy the rate reaches 4%, predicted by female sex (1.46), neobladder diversion (1.57) and low albumin (1.45), with readmission at 31.1% versus 19.2%.^[35] Safety-net every post-antibiotic discharge for late diarrhoea.

Lines and catheters: remove, review daily, bundle what remains. CAUTI prevention on surgical units is daily multidisciplinary review and early removal, sustained by huddle-level communication.^[38] For central lines the global gap is embarrassing: across 22 countries and 85 ICUs, 17.4% of hospitals had no bloodstream-infection surveillance, 7.1% of ICUs had no prevention bundle and 23.5% had no checklist.^[36] A febrile postoperative patient with a line and no abdominal source is a line-and-urine-and-chest hunt first.^[11]

^{[1] [39] [57] [24] [44] [49]}

Specific Subtypes & Scenarios — the six sources worked through

Anastomotic leak after colorectal surgery. Incidence static despite technique; biology is host-microbiome-immune interplay.^[25] Predict with POD-3 CRP 148 (95/95) and POD 2-to-3 procalcitonin; confirm with CT extraluminal contrast or abscess; manage by anatomy and diversion status; score preoperatively; and keep NSAIDs — especially diclofenac at odds ratio 2.79 — out of the chart, noting ketorolac's neutral 1.36 without declaring it safe.^{[23][24][26][28][27]}

Surgical-site infection. Two to five percent of operations; WHO's 13 preoperative measures; the five strong-evidence interventions; debridement when established; prophylaxis stopped at 24 hours; dual vancomycin plus beta-lactam only where it pays — cardiac surgery (adjusted risk ratio 0.61) with number needed to treat 53 in MRSA carriers versus 176 otherwise, priced against 23.8% kidney-injury rates.^{[17][16][18][20][21]} In colorectal prophylaxis specifically, ertapenem beat cefotetan on failure (40.2% versus 50.9%) and SSI (17.1% versus 26.2%) with numerically more C. difficile (1.7% versus 0.6%, non-significant).^[22]

Postoperative pneumonia. Routine support fails; ward bundles, honest ventilation and rescue NIV are the three true moves.^{[29][31][32][33]}

C. difficile colitis. Late, post-discharge, lethal — the 17-day median with 5-fold mortality is the consent and safety-net number.^[34][35]

Line and urinary sources. Remove early, review daily, bundle insertions — the international survey shows the bundle itself is still missing in too many units.^[38][36]

Postoperative healthcare-associated intra-abdominal infection. The Sepsi-Pop phenotype — older, comorbid, recently antibioticked, fungally colonized, E. coli bacteraemic — with 57% shock and 14.4% 90-day mortality, rescued only by effective source control.^[47] For the full intra-abdominal-infection framework including diverticular staging and open-abdomen indications, see the sibling topic on surgical sepsis and source control.

Complications & Pitfalls — the errors examiners reward you for naming

Failure-to-rescue kills more than complications do. Nine in ten post-readmission emergency-surgery deaths are rescue failures — so the complication is forgivable, the un-escalated observations are not.^[10]

Prophylaxis harms are dose-dependent. Every extra day past 24 hours buys kidney injury and C. difficile without preventing a single SSI — numbers needed to harm 9, 6, 4 and 2000, 90, 50.^[20] Dual prophylaxis trades a cardiac SSI gain for kidney injury everywhere else.^[21]

Negative trials deserve humility, not nihilism. Routine respiratory support, lung-expansion bundles and low tidal volumes all failed their primaries — escalate the patient, not the protocol.^[29][30][33] Likewise the antibiotic-timing null cohort tempers, but does not topple, the hour-1 doctrine in shock.^[42][40]

The two classic traps. Operating in unresuscitated shock inside 2 hours doubles the odds against urgent 2-to-6-hour control — resuscitate, then operate.^[44] And prolonging antibiotics instead of re-looking a failing abdomen ignores the 5.71-fold price of failed control.^[44]

Prognosis & Disposition — the mortality ladder and where the patient goes

Stack the ladder: ward sepsis 15.0% dead plus 6.2% to hospice; ICU intra-abdominal infection 23.9%; secondary peritonitis 29.7%; the hour-1 bundle cohort 29.7%; septic-shock criteria 42.3 to 46.5%; postoperative healthcare-associated intra-abdominal infection 14.4% at 90 days.^{[8][48][44][39][3][47]} Modifiers climb both ways: SOFA gradients, ASA above II, age, immunosuppression, re-operation, inadequate control and healthcare-associated origin worsen it; effective source control protects at odds ratio 0.22.^[47]

^{[8] [48] [44] [3] [47] [10]}

Disposition follows the score: qSOFA 2 or SOFA rise 2 means high-dependency or ICU care, never the routine ward; readmitted emergency-surgery patients need rescue pathways with senior review, not routine observation.^[1][10] Discharge with a net: warn every post-antibiotic patient that C. difficile peaks at day 17, often at home.^[34]

Special Populations — elderly, immunocompromised, colonised, metabolic

• Elderly and frail: age accrues intra-abdominal-infection mortality yearly and frailty drives rescue failure — escalate early, set ceilings explicitly, but keep the source-control clock.^[48][10]
• Immunocompromised and cancer: immunosuppression nearly doubles IAI mortality odds, and three-quarters of postoperative healthcare-associated infections strike cancer patients — image sooner, cover broader, drain earlier, because signs under-call severity.^[48][47]
• Colonised (MRSA, MDR Gram-negatives): dual prophylaxis pays only where MRSA colonisation or cardiac surgery concentrates risk (numbers needed to treat 53 versus 176); multidrug-resistant Gram-negative carriers fall under dedicated ESCMID/EUCIC culture-based perioperative prophylaxis guidance.^[21][37]
• Metabolic risk: postoperative hyperglycaemia predicts C. difficile and low albumin predicts it after cystectomy (odds ratio 1.45) — optimise glucose and nutrition before elective surgery and safety-net after it.^[34][35]

Evidence, Guidelines & Regional Differences — who says what, and what is still fought over

• Sepsis-3 (SCCM/ESICM task force, global): the dysregulated-response model with SOFA operationalization and retired severe sepsis — endorsed as a step forward, with the surgeons' alliance concerns on record for surgical patients.^[1][9]
• WHO (global SSI prevention): 13 preoperative recommendations balancing benefit, harm, cost and values — the backbone every local bundle hangs from.^[16]
• STOP-IT and its aftermath (US surgical): 4-day fixed courses non-inferior with only half of practice converted — the gap between evidence and habit is itself examinable.^[49][51]
• GAIS intra-abdominal position (surgeon-led, international): rapid diagnostics, timely adequate control, short PK/PD-guided therapy with stewardship, fluids plus pressors — the cornerstones quote.^[45]
• Contested ground: hour-1 absolutism (Ferrer/Im/Huang gradients versus the Seok null — shock sharpens, non-shock divides); resuscitation targets (lactate-clearance mortality signal versus ANDROMEDA perfusion-first Bayesian win, reconciled by biphasic kinetics); routine respiratory escalation (three negative trial families); and crash versus urgent source control (the AbSeS paradox favouring resuscitated urgency).^{[39][40][41][42][53][54][55][29][30][33][44]}
• Sibling-topic boundary: the companion on surgical sepsis and source control owns diverticular staging, the 6-hour open-abdomen clock and open-versus-closed decisions in depth — this topic owns the postoperative arc from fever to rescue. Cite each where it lives.

Exam Pearls — the triads that score

• Recognition triad: SOFA rise 2 (mortality above 10%), qSOFA 2 of 3 (rate 22+, mentation, systolic 100 or less), shock as pressor plus lactate above 2 (mortality above 40%).^[1]
• Timing triad: antibiotics inside the hour (91 treated per life saved per hour), source control urgent at 2 to 6 hours (half the odds of crashing inside 2), antibiotics stopped at 4 days after adequate control.^[41][44][49]
• Fever triad: 40% of fevers infectious, suspicion after day 3, first fever after day 2 a sixfold flag.^[12][15]
• Leak triad: CRP 148 on day 3 (95 and 95), CT the gold standard, diclofenac nearly triples leak risk.^[23][24][27]
• Stewardship triad: prophylaxis benefit ends at 24 hours (kidney-harm numbers 9, 6, 4), dual prophylaxis only for cardiac or MRSA-colonised patients (53 versus 176), C. difficile strikes at day 17 after discharge.^[20][21][34]
• Rescue line: nine in ten post-readmission emergency-surgery deaths are failure-to-rescue.^[10]

Revision summary

Define postoperative sepsis by Sepsis-3 — dysregulated response, SOFA rise 2, shock as pressor-plus-lactate — with the surgeons' caveat noted.^[1][9] Suspect infection with fever after day 2 to 3 (sixfold flag; 40% base rate) and hunt six sources, testing to findings because blind panels yield 0 to 13% and cost hundreds.^{[15][12][14][11]} Score by location — qSOFA 0.82 on the intermediate unit, SOFA 0.75 in ICU — and move qSOFA-2 patients now.^[5][4] Give antibiotics inside the hour (each hour odds 1.07; shock subgroup odds 0.66), carry the Seok null honestly, resuscitate norepinephrine-first (number needed to treat 9) toward perfusion not numbers.^{[41][40][42][57][54]} Control the source urgently at 2 to 6 hours, stop antibiotics at 4 days, detect leaks with day 2-to-3 CRP/PCT plus CT, prevent SSI with the five strong-evidence moves and 24-hour prophylaxis limit, bundle lungs on the ward while reserving NIV for failure, expect C. difficile at day 17, pull the lines — and rescue every readmission, because 90.6% of post-readmission deaths are failure-to-rescue.^{[44][49][24][17][20][31][32][34][36][10]}

Shock patients gain 35% mortality per hour of antibiotic delay inside 3 h.^[11] C. difficile strikes 0.97% after cardiac surgery, median onset day 17.^[34] 48% of cases present post-discharge.^[21] AKI ran 23.8% versus 13.9% as prophylaxis lengthened.^[21] The SOFA mortality gradient runs 19.6% to 55.4%.^[44]

Shield the immunocompromised: OR 1.99 for IAI death; cancer dominates post-op HC-IAI (75.6%) yet absence of cancer paradoxically predicted shock — reason it through, don't memorise it (PMIDs 41555473, 41949841).^[47]

References57ShowHide

1. [1]Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, et al. The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA, 2016.PMID 26903338
2. [2]Seymour CW, Liu VX, Iwashyna TJ, Brunkhorst FM, et al. Assessment of Clinical Criteria for Sepsis: For the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA, 2016.PMID 26903335
3. [3]Shankar-Hari M, Phillips GS, Levy ML, Seymour CW, et al. Developing a New Definition and Assessing New Clinical Criteria for Septic Shock: For the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA, 2016.PMID 26903336
4. [4]Raith EP, Udy AA, Bailey M, McGloughlin S, et al. Prognostic Accuracy of the SOFA Score, SIRS Criteria, and qSOFA Score for In-Hospital Mortality Among Adults With Suspected Infection Admitted to the Intensive Care Unit. JAMA, 2017.PMID 28114553
5. [5]Koch C, Edinger F, Fischer T, Brenck F, et al. Comparison of qSOFA score, SOFA score, and SIRS criteria for the prediction of infection and mortality among surgical intermediate and intensive care patients. World J Emerg Surg, 2020.PMID 33239088
6. [6]Qiu X, Lei YP, Zhou RX SIRS, SOFA, qSOFA, and NEWS in the diagnosis of sepsis and prediction of adverse outcomes: a systematic review and meta-analysis. Expert Rev Anti Infect Ther, 2023.PMID 37450490
7. [7]Lu J, Dong Z, Ye L, Gao Y, et al. Predictive value of SOFA, PCT, Lactate, qSOFA and their combinations for mortality in patients with sepsis: A systematic review and meta-analysis. PLoS One, 2025.PMID 40961067
8. [8]Rhee C, Dantes R, Epstein L, Murphy DJ, et al. Incidence and Trends of Sepsis in US Hospitals Using Clinical vs Claims Data, 2009-2014. JAMA, 2017.PMID 28903154
9. [9]Sartelli M, Kluger Y, Ansaloni L, Hardcastle TC, et al. Raising concerns about the Sepsis-3 definitions. World J Emerg Surg, 2018.PMID 29416555
10. [10]Coimbra R, Kim M, Allison-Aipa T, Zakhary B, et al. Deaths After Readmissions are Mostly Attributable to Failure-to-Rescue in EGS Patients. Am Surg, 2024.PMID 38656140
11. [11]Magill SS, Edwards JR, Bamberg W, Beldavs ZG, et al. Multistate point-prevalence survey of health care-associated infections. N Engl J Med, 2014.PMID 24670166
12. [12]Dionigi R, Dionigi G, Rovera F, Boni L Postoperative fever. Surg Infect (Larchmt), 2006.PMID 16895496
13. [13]Maday KR, Hurt JB, Harrelson P, Porterfield J Evaluating postoperative fever. JAAPA, 2016.PMID 27623291
14. [14]Ashley B, Spiegel DA, Cahill P, Talwar D, et al. Post-operative fever in orthopaedic surgery: How effective is the 'fever workup?'. J Orthop Surg (Hong Kong), 2017.PMID 28851265
15. [15]Lai HF, Chau IY, Lei HJ, Chou SC, et al. Postoperative fever after liver resection: Incidence, risk factors, and characteristics associated with febrile infectious complication. PLoS One, 2022.PMID 35025947
16. [16]Allegranzi B, Bischoff P, de Jonge S, Kubilay NZ, et al. New WHO recommendations on preoperative measures for surgical site infection prevention: an evidence-based global perspective. Lancet Infect Dis, 2016.PMID 27816413
17. [17]Fuglestad MA, Tracey EL, Leinicke JA Evidence-based Prevention of Surgical Site Infection. Surg Clin North Am, 2021.PMID 34774274
18. [18]Seidelman J, Anderson DJ Surgical Site Infections. Infect Dis Clin North Am, 2021.PMID 34752225
19. [19]Le Couteur J, Druce P, Myles PS, Peel T Systematic Review of Surgical Site Infection Prevention Guideline Recommendations for Maintenance of Homeostasis in the Perioperative Period. Anesthesiology, 2025.PMID 40358339
20. [20]Branch-Elliman W, O'Brien W, Strymish J, Itani K, et al. Association of Duration and Type of Surgical Prophylaxis With Antimicrobial-Associated Adverse Events. JAMA Surg, 2019.PMID 31017647
21. [21]Branch-Elliman W, Ripollone JE, O'Brien WJ, Itani KMF, et al. Risk of surgical site infection, acute kidney injury, and Clostridium difficile infection following antibiotic prophylaxis with vancomycin plus a beta-lactam versus either drug alone: A national propensity-score-adjusted retrospective cohort study. PLoS Med, 2017.PMID 28692690
22. [22]Itani KM, Wilson SE, Awad SS, Jensen EH, et al. Ertapenem versus cefotetan prophylaxis in elective colorectal surgery. N Engl J Med, 2006.PMID 17182989
23. [23]Yeung DE, Peterknecht E, Hajibandeh S, Hajibandeh S, et al. C-reactive protein can predict anastomotic leak in colorectal surgery: a systematic review and meta-analysis. Int J Colorectal Dis, 2021.PMID 33555423
24. [24]Chiarello MM, Fransvea P, Cariati M, Adams NJ, et al. Anastomotic leakage in colorectal cancer surgery. Surg Oncol, 2022.PMID 35092916
25. [25]Meyer J, Naiken S, Christou N, Liot E, et al. Reducing anastomotic leak in colorectal surgery: The old dogmas and the new challenges. World J Gastroenterol, 2019.PMID 31558854
26. [26]Yung HC, Daroch AK, Parikh R, Mathur DV, et al. Diagnostic Modalities for Early Detection of Anastomotic Leak After Colorectal Surgery. J Surg Res, 2024.PMID 39047384
27. [27]Modasi A, Pace D, Godwin M, Smith C, et al. NSAID administration post colorectal surgery increases anastomotic leak rate: systematic review/meta-analysis. Surg Endosc, 2019.PMID 29998389
28. [28]Litchinko A, Buchs N, Balaphas A, Toso C, et al. Score prediction of anastomotic leak in colorectal surgery: a systematic review. Surg Endosc, 2024.PMID 38418633
29. [29]Hui S, Fowler AJ, Cashmore RMJ, Fisher TJ, et al. Routine postoperative noninvasive respiratory support and pneumonia after elective surgery: a systematic review and meta-analysis of randomised trials. Br J Anaesth, 2022.PMID 34916050
30. [30]Fernandez-Bustamante A, Parker RA, Frendl G, Lee JW, et al. Perioperative lung expansion and pulmonary outcomes after open abdominal surgery versus usual care in the USA (PRIME-AIR): a multicentre, randomised, controlled, phase 3 trial. Lancet Respir Med, 2025.PMID 40020692
31. [31]Mahama G, Vigneswaran L, Silva A, Maeda A, et al. A bundled approach to care: reducing the incidence of postoperative pneumonia in patients undergoing hepatectomy and Whipple procedures. Can J Surg, 2021.PMID 33411999
32. [32]Pettenuzzo T, Boscolo A, Pistollato E, Pretto C, et al. Effects of non-invasive respiratory support in post-operative patients: a systematic review and network meta-analysis. Crit Care, 2024.PMID 38720332
33. [33]Karalapillai D, Weinberg L, Peyton P, Ellard L, et al. Effect of Intraoperative Low Tidal Volume vs Conventional Tidal Volume on Postoperative Pulmonary Complications in Patients Undergoing Major Surgery: A Randomized Clinical Trial. JAMA, 2020.PMID 32870298
34. [34]Kirkwood KA, Gulack BC, Iribarne A, Bowdish ME, et al. A multi-institutional cohort study confirming the risks of Clostridium difficile infection associated with prolonged antibiotic prophylaxis. J Thorac Cardiovasc Surg, 2018.PMID 29102205
35. [35]Prunty M, Bukavina L, Mahran A, Mishra K, et al. Risk factors for postoperative Clostridium difficile infection after radical cystectomy for bladder cancer: a NSQIP database analysis. Can J Urol, 2022.PMID 35691039
36. [36]Devrim I, Erdem H, El-Kholy A, Almohaizeie A, et al. Analyzing central-line associated bloodstream infection prevention bundles in 22 countries: The results of ID-IRI survey. Am J Infect Control, 2022.PMID 35263612
37. [37]Righi E, Mutters NT, Guirao X, Del Toro MD, et al. ESCMID/EUCIC clinical practice guidelines on perioperative antibiotic prophylaxis in patients colonized by multidrug-resistant Gram-negative bacteria before surgery. Clin Microbiol Infect, 2023.PMID 36566836
38. [38]Gauron G, Bigand T Implementation of evidence-based strategies to reduce catheter-associated urinary tract infections among hospitalized, post-surgical adults. Am J Infect Control, 2021.PMID 33221473
39. [39]Ferrer R, Martin-Loeches I, Phillips G, Osborn TM, et al. Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: results from a guideline-based performance improvement program. Crit Care Med, 2014.PMID 24717459
40. [40]Im Y, Kang D, Ko RE, Lee YJ, et al. Time-to-antibiotics and clinical outcomes in patients with sepsis and septic shock: a prospective nationwide multicenter cohort study. Crit Care, 2022.PMID 35027073
41. [41]Huang J, Yang JT, Liu JC The association between mortality and door-to-antibiotic time: a systematic review and meta-analysis. Postgrad Med J, 2023.PMID 36917816
42. [42]Seok H, Song J, Jeon JH, Choi HK, et al. Timing of antibiotics in septic patients: a prospective cohort study. Clin Microbiol Infect, 2020.PMID 32062049
43. [43]Akyol D, Çankayalı İ, Ersel M, Demirağ K, et al. Impact of the empirical therapy timing on the clinical progress of septic shock patients. Diagn Microbiol Infect Dis, 2024.PMID 38142580
44. [44]De Pascale G, Antonelli M, Deschepper M, Arvaniti K, et al. Poor timing and failure of source control are risk factors for mortality in critically ill patients with secondary peritonitis. Intensive Care Med, 2022.PMID 36151335
45. [45]Sartelli M, Barie P, Agnoletti V, Al-Hasan MN, et al. Intra-abdominal infections survival guide: a position statement by the Global Alliance For Infections In Surgery. World J Emerg Surg, 2024.PMID 38851700
46. [46]Bova R, Griggio G, Vallicelli C, Santandrea G, et al. Source Control and Antibiotics in Intra-Abdominal Infections. Antibiotics (Basel), 2024.PMID 39200076
47. [47]Mokart D, Bianchi A, Ronflé R, Boutaba M, et al. Determinants of Septic Shock and 90-Day Mortality After Postoperative Healthcare-Associated Intra-Abdominal Infection: The Sepsi-Pop Study, a Multicenter Retrospective Study. Shock, 2026.PMID 41949841
48. [48]Brogi E, Cremonini C, Ceresoli M, Catena F, et al. Epidemiology and outcome of intra-abdominal infections in intensive care unit in Italy from the Italian Register of complicated Intra-abdominal InfectionS-the IRIS study: a prospective observational nationwide study. J Anesth Analg Crit Care, 2026.PMID 41555473
49. [49]Sawyer RG, Claridge JA, Nathens AB, Rotstein OD, et al. Trial of short-course antimicrobial therapy for intraabdominal infection. N Engl J Med, 2015.PMID 25992746
50. [50]Hochstetler LJ, Olney WJ, Bishop JM, Warriner ZD, et al. Antibiotics for Patients With a Planned Re-Laparotomy for Intra-Abdominal Infection. Surg Infect (Larchmt), 2024.PMID 38407831
51. [51]Posillico SE, Young BT, Ladhani HA, Zosa BM, et al. Current Evaluation of Antibiotic Usage in Complicated Intra-Abdominal Infection after the STOP IT Trial: Did We STOP IT? Surg Infect (Larchmt), 2019.PMID 30676237
52. [52]Jones AE, Shapiro NI, Trzeciak S, Arnold RC, et al. Lactate clearance vs central venous oxygen saturation as goals of early sepsis therapy: a randomized clinical trial. JAMA, 2010.PMID 20179283
53. [53]Pan J, Peng M, Liao C, Hu X, et al. Relative efficacy and safety of early lactate clearance-guided therapy resuscitation in patients with sepsis: A meta-analysis. Medicine (Baltimore), 2019.PMID 30813144
54. [54]Zampieri FG, Damiani LP, Bakker J, Ospina-Tascón GA, et al. Effects of a Resuscitation Strategy Targeting Peripheral Perfusion Status versus Serum Lactate Levels among Patients with Septic Shock. A Bayesian Reanalysis of the ANDROMEDA-SHOCK Trial. Am J Respir Crit Care Med, 2020.PMID 31574228
55. [55]Castro R, Hernández G, Kattan E, Bakker J What every intensivist should know about the biphasic kinetics of lactate in septic shock. J Crit Care, 2026.PMID 41101140
56. [56]De Backer D, Aldecoa C, Njimi H, Vincent JL Dopamine versus norepinephrine in the treatment of septic shock: a meta-analysis*. Crit Care Med, 2012.PMID 22036860
57. [57]Avni T, Lador A, Lev S, Leibovici L, et al. Vasopressors for the Treatment of Septic Shock: Systematic Review and Meta-Analysis. PLoS One, 2015.PMID 26237037