Sepsis

Sepsis Recognition

Sepsis is currently defined by Sepsis-3 as life-threatening organ dysfunction caused by a dysregulated host response to infection. The critical diagnostic task is not merely identifying infection, but detecting infection-associated acute organ dysfunction early enough to alter trajectory. For examination purposes, distinguish three overlapping constructs: suspected infection, systemic host response, and quantifiable organ dysfunction. Fever, tachycardia and leukocytosis alone are insufficiently specific; conversely, elderly, immunosuppressed, cirrhotic, uraemic or corticosteroid-treated patients may present with hypothermia, delirium, falls, anorexia or isolated tachypnoea.

Pathobiological basis of recognition

Sepsis recognition rests on appreciating that organ dysfunction reflects maladaptive inflammation, endothelial injury, microvascular thrombosis, mitochondrial dysfunction and neurohumoral dysregulation. Pathogen-associated and damage-associated molecular patterns activate innate immunity via Toll-like receptors and inflammasomes, generating cytokines such as TNF-α, IL-1β and IL-6, complement activation, inducible nitric oxide synthase activity and tissue factor expression. Clinically this manifests as vasodilatation, capillary leak, impaired oxygen extraction, encephalopathy, acute kidney injury, cholestasis, coagulopathy, hypoxaemia and hyperlactataemia. Lactate is therefore not simply a marker of anaerobic metabolism; β-adrenergic-driven glycolysis and impaired hepatic clearance contribute, which explains raised lactate despite apparently adequate macro-haemodynamics.

Definitions and scoring systems

Tool/definition	Threshold	Use in exams and practice	Limitations
Sepsis-3 SOFA	Acute rise in SOFA score ≥2 from baseline in suspected infection	Formal definition; associated with approximately ≥10% in-hospital mortality in derivation cohorts	Requires laboratory and physiological data; baseline organ dysfunction complicates interpretation
qSOFA	At least 2 of: respiratory rate ≥22/min, systolic BP ≤100 mmHg, altered mentation	Bedside prompt for poor outcome outside ICU	Poor sensitivity for early sepsis; should not be used as a screening tool in isolation
SIRS	At least 2 of: temperature >38 or <36°C, HR >90/min, RR >20/min or PaCO2 <4.3 kPa, WCC >12 or <4 ×109/L or bands >10%	Historically important; sensitive inflammatory screen	Low specificity; may occur after trauma, pancreatitis, surgery, pulmonary embolism
NEWS2	Aggregate score; urgent escalation commonly triggered by ≥5, emergency response by ≥7	UK ward detection of acute deterioration; endorsed operationally by RCP/NHS systems	Detects deterioration, not sepsis specifically; oxygen scale 2 required for hypercapnic respiratory failure risk

The SOFA score grades six systems from 0–4: respiratory function by PaO₂/FiO₂, coagulation by platelet count, liver by bilirubin, cardiovascular status by hypotension/vasopressor requirement, central nervous system by Glasgow Coma Scale, and renal function by creatinine or urine output. In examinations, an acute creatinine rise, oliguria, thrombocytopenia, bilirubin elevation, hypoxaemia, new confusion or metabolic acidosis in an infected patient should be treated as sepsis even before the full SOFA can be calculated.

Clinical recognition: high-yield features

• Infection likelihood: focus from respiratory, urinary, intra-abdominal, biliary, skin/soft tissue, CNS, line-related, endocardial, pelvic or postoperative sources; absence of fever does not exclude infection.
• Circulatory stress: tachycardia, cool peripheries, prolonged capillary refill, mottling, narrow pulse pressure or bounding pulses in distributive physiology.
• Respiratory dysfunction: tachypnoea is often the earliest sign; oxygen saturation targets must be interpreted in context, particularly COPD or obesity hypoventilation.
• Neurological dysfunction: new delirium, reduced GCS or agitation may be the presenting organ dysfunction, especially in older adults.
• Renal/metabolic dysfunction: oliguria <0.5 mL/kg/hour, rising creatinine, bicarbonate fall, base deficit and lactate elevation.
• Haematological/hepatic dysfunction: platelets <150 ×10⁹/L, INR prolongation, bilirubin rise or transaminitis may indicate systemic involvement.

Lactate and laboratory markers

A venous lactate is acceptable for screening; arterial sampling is not mandatory for recognition. Normal lactate is typically 0.5–2.0 mmol/L. Values >2 mmol/L indicate increased risk and warrant close reassessment; ≥4 mmol/L is a major high-risk marker in sepsis pathways and was used in early goal-directed therapy trials to define severe hypoperfusion. Persistent elevation or failure to clear lactate is prognostically adverse. However, lactate rises in seizures, β₂-agonist therapy, adrenaline infusion, hepatic failure, metformin-associated lactic acidosis, mesenteric ischaemia and profound anaemia, so interpretation must be contextual.

Inflammatory biomarkers support but do not define sepsis. C-reactive protein rises after 6–12 hours and is nonspecific. Procalcitonin is induced by bacterial toxins and cytokines, suppressed by interferon-γ in viral infection, and has a half-life of approximately 24 hours; cut-offs such as <0.25 μg/L for lower respiratory tract infection or <0.5 μg/L in ICU algorithms may support de-escalation decisions, but sensitivity is imperfect and false elevations occur after major surgery, trauma, burns and renal dysfunction. It should not delay recognition or escalation.

Evidence and guideline framing

The Surviving Sepsis Campaign 2021 emphasises early identification of infection-associated organ dysfunction and immediate risk stratification, while cautioning against qSOFA as a sole screening instrument because meta-analyses show higher specificity but lower sensitivity than SIRS or NEWS-based tools. NICE sepsis guidance similarly prioritises structured assessment of temperature, heart rate, respiratory rate, oxygen saturation, blood pressure, conscious level, urine output and skin perfusion, with particular attention to extremes of age, pregnancy, immunocompromise, indwelling devices and recent surgery. For MRCP, the safest formulation is: suspected infection plus new organ dysfunction equals sepsis until proven otherwise, and normal temperature or modest inflammatory markers do not exclude it.

Septic Shock

Definition and haemodynamic phenotype

Septic shock is the subset of sepsis with profound circulatory, cellular and metabolic abnormalities associated with substantially increased mortality. Under Sepsis-3, it is defined clinically as sepsis requiring vasopressors to maintain mean arterial pressure (MAP) ≥65 mmHg and having serum lactate >2 mmol/L despite adequate fluid resuscitation. This definition identifies a cohort with hospital mortality typically >40%, compared with lower mortality in sepsis without shock.

The dominant haemodynamic pattern is distributive shock: reduced systemic vascular resistance from inducible nitric oxide synthase activation, prostacyclin, adrenomedullin and vasoplegic cytokine signalling. However, septic shock is not purely vasodilatory. Myocardial depression, capillary leak, venodilatation with reduced stressed volume, microvascular shunting, mitochondrial dysfunction and relative adrenal insufficiency may coexist. Consequently, cardiac output may be high, normal or low depending on timing, preload, ventricular function and catecholamine exposure.

Phenotype	Typical findings	Clinical implication
Warm hyperdynamic shock	Bounding pulse, warm peripheries, wide pulse pressure, high/normal cardiac output, low SVR	Early vasopressor requirement is common despite apparently “good” peripheral perfusion
Cold hypodynamic shock	Cool mottled skin, narrow pulse pressure, low cardiac output, raised lactate, oliguria	Consider myocardial depression, severe hypovolaemia, obstructive mimics, or need for inotropy
Cryptic shock	Normal blood pressure with lactate ≥4 mmol/L	High-risk state; do not be reassured by preserved MAP

Assessment of severity and resuscitation targets

Shock severity should be judged dynamically rather than from blood pressure alone. Key markers include MAP, lactate trend, capillary refill time, mottling, mental state, urine output and vasopressor dose. A commonly used renal perfusion target is urine output ≥0.5 mL/kg/hour. Lactate is not synonymous with anaerobic metabolism: accelerated glycolysis from β-adrenergic stimulation, impaired hepatic clearance and mitochondrial dysfunction contribute. Nevertheless, persistent hyperlactataemia is strongly prognostic; many protocols target lactate reduction by ≥10–20% over 2–4 hours.

The Surviving Sepsis Campaign 2021 recommends an initial MAP target of 65 mmHg rather than higher targets in most patients. The SEPSISPAM trial found no overall mortality benefit from targeting MAP 80–85 mmHg versus 65–70 mmHg, although chronic hypertensive patients had less renal replacement therapy at higher MAP but more atrial fibrillation. Individualisation is required in severe chronic hypertension, raised intracranial pressure, advanced vasculopathy or ongoing hypoperfusion despite MAP 65 mmHg.

Fluid and vasopressor strategy

Initial resuscitation usually uses intravenous crystalloid; the SSC suggests 30 mL/kg within 3 hours for sepsis-induced hypoperfusion or septic shock, but this is a weak recommendation and must be tempered by heart failure, renal failure and pulmonary oedema risk. Balanced crystalloids are generally preferred over 0.9% saline to reduce hyperchloraemic acidosis and possible renal injury. Further fluid should be guided by responsiveness rather than static filling pressures: passive leg raise with stroke volume change, bedside echocardiography, pulse pressure variation in suitable ventilated patients, or mini-fluid challenge. Central venous pressure is a poor predictor of responsiveness.

Drug	Usual adult dose	Key points
Noradrenaline	Start 0.05–0.1 microgram/kg/min; titrate commonly up to 0.5–1 microgram/kg/min or higher in refractory shock	First-line vasopressor; α1 vasoconstriction with modest β1 activity; short half-life approximately 2–3 minutes
Vasopressin	0.03 units/min fixed dose	Add to reduce noradrenaline requirement; V1-mediated vasoconstriction; avoid high doses causing ischaemia
Adrenaline	0.05–0.5 microgram/kg/min	Second-line/additional agent; increases lactate via β2-mediated glycolysis, complicating lactate interpretation
Dobutamine	2.5–20 microgram/kg/min	Consider if myocardial dysfunction and persistent hypoperfusion despite adequate MAP and preload
Hydrocortisone	200 mg/day IV, e.g. 50 mg 6-hourly or infusion	For shock with ongoing vasopressor requirement; hastens shock reversal, mortality effect variable

Noradrenaline is the vasopressor of choice. The SOAP II trial showed more arrhythmias with dopamine and a mortality signal against dopamine in cardiogenic shock; dopamine is therefore reserved for highly selected patients with bradycardia and low arrhythmia risk. Early noradrenaline through a well-sited peripheral cannula may be appropriate while central access is arranged, with frequent extravasation checks. The CENSER trial suggested earlier shock control with early low-dose noradrenaline. Vasopressin is supported by VASST and VANISH as a catecholamine-sparing adjunct rather than a proven universal mortality-reducing therapy.

Evidence base and refractory shock

Rivers’ 2001 early goal-directed therapy trial reported mortality reduction using protocolised central venous oxygen saturation, transfusion and dobutamine targets. Subsequent multicentre trials—ProCESS, ARISE and ProMISe—showed no mortality advantage over contemporary usual care, largely because early antibiotics, fluids and vasopressors had become standard. Current practice emphasises early recognition, rapid haemodynamic optimisation and repeated reassessment rather than rigid CVP/ScvO₂ protocols. ANDROMEDA-SHOCK suggested capillary refill-targeted resuscitation may be at least as safe as lactate-targeted strategies and may reduce treatment intensity.

Refractory septic shock generally denotes persistent hypotension or hypoperfusion despite adequate fluids and high-dose vasopressors, often noradrenaline-equivalent doses >0.5–1 microgram/kg/min. Check for under-resuscitation, occult bleeding, tamponade, pulmonary embolism, tension pneumothorax, severe acidosis, hypocalcaemia and abdominal compartment syndrome. Severe acidaemia reduces catecholamine responsiveness; bicarbonate is not routine but may be considered in severe metabolic acidaemia, particularly pH ≤7.2 with acute kidney injury. Corticosteroids are reasonable when shock remains vasopressor-dependent; APROCCHSS showed mortality benefit with hydrocortisone plus fludrocortisone, whereas ADRENAL showed faster shock resolution without significant 90-day mortality reduction.

Source Control

Source control is the anatomical elimination or containment of the nidus driving sepsis: drainage of infected collections, debridement of necrotic tissue, relief of obstruction, removal of infected prosthetic material or devices, and definitive operative management of perforation or ischaemia. In sepsis it is not an optional adjunct to antimicrobials; persistent microbial burden, biofilm, devitalised tissue, high inoculum effects and impaired antibiotic penetration can make pharmacotherapy alone inadequate. The Surviving Sepsis Campaign 2021 recommends that an anatomically defined focus requiring source control be identified or excluded rapidly, and that the intervention be implemented as soon as medically and logistically practical, ideally within 6–12 hours for most urgent foci.

Principles and prioritisation

The exam-relevant principle is that timely adequate source control predicts survival independently of antibiotic choice. Delay is particularly hazardous in necrotising soft tissue infection, perforated viscus, cholangitis with obstruction, obstructed infected urinary tract, empyema, septic arthritis, infected vascular grafts and catheter-related bloodstream infection due to Staphylococcus aureus, Gram-negative bacilli or Candida. The intervention should be the least invasive procedure that achieves adequate control, but “minimally invasive” must not mean incomplete. Haemodynamic instability is not usually a reason to defer source control; rather, resuscitation, vasopressor support and anaesthetic planning should proceed in parallel.

Septic focus	Preferred source control	Critical exam points
Intra-abdominal abscess	Image-guided percutaneous drainage if anatomically feasible; surgery if multiloculated, inaccessible or associated with perforation/ischaemia	Drainage plus microbiology; antibiotics alone usually fail for collections >3 cm or ongoing contamination
Perforated viscus/faecal peritonitis	Emergency laparotomy/laparoscopy, lavage, resection/repair, diversion where required	Definitive control of leak is essential; persistent contamination causes refractory shock
Ascending cholangitis	Urgent biliary decompression: ERCP with sphincterotomy/stent; percutaneous transhepatic drainage if ERCP not possible	Tokyo Guidelines classify severe disease by organ dysfunction; septic shock mandates urgent drainage
Obstructed infected kidney	Ureteric stent or percutaneous nephrostomy	Pyonephrosis is a urological emergency; antibiotics do not sterilise an obstructed system reliably
Necrotising fasciitis/myonecrosis	Immediate radical surgical debridement, repeated at 24–48 h as required	Do not await imaging if clinically obvious; clindamycin added for toxin suppression in streptococcal/clostridial disease
Infected intravascular catheter	Prompt catheter removal, culture tip if appropriate	Mandatory in shock, tunnel infection, suppurative thrombophlebitis, endocarditis, S. aureus, Pseudomonas or Candida bacteraemia
Septic arthritis/empyema	Joint washout/aspiration; intercostal drainage ± VATS decortication	Closed-space infection requires mechanical evacuation; delay increases irreversible damage

Assessment before intervention

Source control requires rapid anatomical localisation. Contrast-enhanced CT is often the highest-yield test for intra-abdominal, retroperitoneal and deep soft tissue sepsis; ultrasound is useful for biliary, pelvic and renal obstruction; echocardiography is used when endocarditis or infected intracardiac material is suspected. However, imaging should not delay treatment when the diagnosis is clinically evident, especially necrotising fasciitis or peritonitis. In candidates with acute kidney injury, the risk of contrast nephropathy is usually outweighed by the mortality risk of missed uncontrolled sepsis; modern low-osmolar iodinated contrast has a lower nephrotoxic signal than historically assumed, and post-contrast AKI is often confounded by sepsis itself.

Timing, adequacy and antibiotic duration after control

The operative question is not merely “was a procedure performed?” but “was source control adequate?” Indicators of failure include persistent vasopressor requirement, ongoing fever or hypothermia, rising lactate, non-clearing bacteraemia, worsening organ dysfunction, persistent drain output of enteric contents, or repeat imaging showing residual collection. Re-intervention should be considered early rather than escalating antibiotics indefinitely.

After adequate source control, prolonged antimicrobial courses are usually unnecessary. The STOP-IT trial in complicated intra-abdominal infection showed that approximately 4 days of antibiotics after adequate source control was non-inferior to longer therapy guided by physiological resolution, with similar rates of surgical site infection, recurrent intra-abdominal infection or death. This principle is frequently tested: ongoing fever shortly after surgery may reflect inflammation, atelectasis or inadequate control; it is not in itself an indication for a 14-day course. Exceptions include endocarditis, osteomyelitis, undrained abscess, retained infected prosthesis, S. aureus bacteraemia, candidemia and immunocompromised hosts.

Device and prosthesis-associated infection

Biofilm is central to device-related sepsis. Organisms embedded in extracellular polymeric matrix exhibit reduced metabolic activity and markedly increased minimum biofilm eradication concentrations compared with planktonic MICs. Thus, bloodstream infection may persist despite apparently susceptible isolates. Infected central venous catheters, urinary catheters, biliary stents, nephrostomy tubes, prosthetic valves, vascular grafts, orthopaedic implants and cardiac implantable electronic devices should be assessed explicitly. For MRCP purposes, remember that uncomplicated coagulase-negative staphylococcal catheter infection may occasionally be managed with line removal and short-course therapy, whereas S. aureus, Candida and Gram-negative line sepsis generally require removal and evaluation for metastatic foci.

Practical hazards

• Do not drain sterile-looking chronic collections indiscriminately: instrumentation can seed infection; correlate with clinical sepsis and imaging features.
• Do not rely on inflammatory markers alone: C-reactive protein and procalcitonin may lag; physiological trajectory and anatomical control are more important.
• Do not miss obstruction: pus under pressure in the biliary or urinary tract can cause profound bacteraemia and shock until decompressed.
• Do not delay debridement for “optimisation”: in necrotising infection, mortality rises with time; broad-spectrum antibiotics are bridging therapy, not definitive treatment.

Blood Culture Interpretation

Blood cultures are a cornerstone investigation in suspected sepsis, but their value depends as much on pre-analytical quality and clinical interpretation as on microbiological yield. In adults, bacteraemia is often low-density, typically <1 CFU/mL; diagnostic sensitivity therefore depends strongly on blood volume. A standard adult set comprises one aerobic and one anaerobic bottle, ideally inoculated with 8–10 mL per bottle. Two sets from separate venepuncture sites before antibiotics are recommended where this does not delay therapy; cumulative sensitivity rises from approximately 60–70% with one set to 80–90% with two sets and >95% with three sets in continuous bacteraemia.

Key interpretive principles

• Clinical syndrome governs significance: the same isolate may represent contamination, colonisation, transient bacteraemia, or invasive infection depending on host factors, source, prosthetic material, and systemic features.
• Number of positive bottles/sets matters: growth of the same organism from multiple sets, especially from separate venepuncture sites, strongly supports true bacteraemia.
• Time to positivity matters: shorter time to positivity correlates with higher inoculum and true infection. Most clinically significant bacteraemias signal within 12–24 h; laboratories usually incubate for 5 days, longer selectively for fastidious organisms.
• Antibiotics before sampling reduce yield: cultures should be taken before antimicrobials if feasible, but treatment of sepsis must not be delayed beyond the clinically safe window; Surviving Sepsis Campaign guidance supports antimicrobials within 1 h for septic shock or high likelihood sepsis.

Finding	Interpretation	Exam-relevant action
Same organism in ≥2 separate sets	True bacteraemia likely	Identify source, tailor therapy, assess need for echocardiography or imaging
Single positive bottle with coagulase-negative staphylococcus	Often contamination, but may be significant with central line/prosthesis/immunosuppression	Repeat cultures; assess line infection criteria and device involvement
Growth within <12 h	High bacterial burden; usually clinically significant	Escalate source search and antimicrobial optimisation
Polymicrobial growth	Suggests intra-abdominal, biliary, diabetic foot, necrotising soft-tissue infection, or line infection; contamination possible if skin flora only	Urgent source control assessment

Contaminants versus true pathogens

Blood culture contamination rates should be <3%; higher rates undermine antimicrobial stewardship and may cause unnecessary vancomycin use, line removal, prolonged admission, and echocardiography. Typical contaminants are skin commensals introduced at venepuncture: coagulase-negative staphylococci, Corynebacterium spp., Bacillus spp. other than B. anthracis, Cutibacterium acnes, and viridans streptococci in some contexts. However, none is automatically dismissible: Staphylococcus epidermidis is important in prosthetic valve endocarditis, vascular graft infection, cerebrospinal fluid shunts, and central venous catheter-related bloodstream infection.

Organisms that should usually be regarded as clinically significant even from a single positive bottle include Staphylococcus aureus, Streptococcus pneumoniae, beta-haemolytic streptococci, Enterobacterales, Pseudomonas aeruginosa, anaerobes such as Bacteroides fragilis, Candida spp., and Neisseria meningitidis. S. aureus bacteraemia is never a contaminant in routine practice: it mandates repeat cultures until clearance, source identification, assessment for metastatic foci, and echocardiography where indicated.

Catheter-related bloodstream infection

Central venous catheters complicate interpretation. Diagnostic support for catheter-related bloodstream infection includes the same organism from peripheral and catheter-drawn cultures, purulence at the exit site, or differential time to positivity: a catheter-drawn culture becoming positive at least 2 h earlier than a peripheral culture suggests the catheter as the source. Quantitative culture criteria, where used, include a catheter sample colony count at least 3-fold higher than peripheral blood. Management depends on organism and clinical stability: remove the line for S. aureus, Pseudomonas, fungi, mycobacteria, tunnel infection, septic thrombosis, endocarditis, persistent bacteraemia, or severe sepsis.

Susceptibility results and antimicrobial decisions

The Gram stain is the first actionable result and should prompt early rationalisation: Gram-positive cocci in clusters suggest staphylococci; chains suggest streptococci or enterococci; Gram-negative rods suggest Enterobacterales or non-fermenters. Modern laboratories may use MALDI-TOF or multiplex PCR directly from positive bottles, reducing organism identification time by many hours. Susceptibility interpretation follows EUCAST or CLSI breakpoints and must be integrated with infection site, drug exposure, renal function, allergy, and source control. For example, “susceptible” does not imply adequate treatment if tissue penetration is poor or a nidus remains undrained.

Result	Likely implication	Important consequence
MSSA	Beta-lactam susceptible S. aureus	Use anti-staphylococcal beta-lactam rather than vancomycin where possible
MRSA	mecA/mecC-mediated methicillin resistance	Use vancomycin, daptomycin, or linezolid according to syndrome; avoid daptomycin in pneumonia
ESBL-producing Enterobacterales	Hydrolysis of many cephalosporins	Carbapenem often preferred in severe bacteraemia; MERINO trial showed higher mortality with piperacillin-tazobactam versus meropenem in ceftriaxone-resistant E. coli/Klebsiella bacteraemia
Enterococcus faecium	Often ampicillin resistant; may be vancomycin resistant	Consider linezolid or daptomycin if VRE; search for intra-abdominal or line source
Candida spp.	True fungaemia until proven otherwise	Remove central line where feasible, start echinocandin, repeat cultures, ophthalmic assessment per local guidance

Repeat cultures and clearance

Routine repeat cultures are not required for uncomplicated Gram-negative bacteraemia that improves clinically. They are essential in S. aureus bacteraemia, candidemia, endovascular infection, persistent fever, multidrug-resistant organisms, inadequate source control, or suspected relapse. Persistent positivity beyond 48–72 h despite appropriate therapy should trigger reassessment for abscess, infected thrombus, endocarditis, retained prosthetic material, or subtherapeutic drug exposure. In exams, the safest interpretive approach is to combine organism identity, number of positive sets, time to positivity, host risk, and source plausibility rather than labelling isolates simplistically as “contaminants”.