I now have comprehensive content to write a detailed answer. Let me compile everything.
Sepsis: Diagnosis & Investigations
Definitions (Sepsis-3, 2016)
There is no gold-standard diagnostic test for sepsis. The diagnosis integrates clinical, laboratory, radiologic, and microbiologic data.
| Condition | Definition | Diagnostic Criteria |
|---|
| Sepsis | Life-threatening organ dysfunction caused by a dysregulated host response to infection | Suspected/proven infection + SOFA score increase ≥2 from baseline |
| Septic shock | A subset of sepsis with profound circulatory and cellular metabolic abnormalities | Sepsis + vasopressor required to maintain MAP ≥65 mmHg + lactate >2 mmol/L despite adequate fluids |
- Tintinalli's Emergency Medicine, p. 1039
- Harrison's Principles of Internal Medicine 22e, p. 2363
Screening Tools
qSOFA (Quick SOFA)
A rapid bedside screening tool - score ≥2 out of 3 indicates high risk for poor outcome:
| Criterion | Threshold |
|---|
| Respiratory rate | ≥22 breaths/min |
| Glasgow Coma Scale | <15 (altered mental status) |
| Systolic blood pressure | ≤100 mmHg |
Important limitations: qSOFA has higher specificity but lower sensitivity than SIRS criteria for identifying end-organ dysfunction. It was not designed, nor is it recommended, as the sole screening tool. The 2021 Surviving Sepsis Campaign guidelines do not endorse any single screening tool over others. - Harrison's 22e, p. 2282; Tintinalli's, p. 1038
SOFA Score
Used to formally confirm sepsis. Composed of 6 organ systems, each scored 0-4:
| System | Measure | Score 1 | Score 2 | Score 3 | Score 4 |
|---|
| Respiration | PaO2/FiO2 (mmHg) | <400 | <300 | <200 + resp. support | <100 + resp. support |
| Coagulation | Platelets (x10³/µL) | <150 | <100 | <50 | <20 |
| Liver | Bilirubin (mg/dL) | 1.2-1.9 | 2.0-5.9 | 6.0-11.9 | >12.0 |
| Cardiovascular | MAP/vasopressors | MAP <70 | Dopa <5 or any dobutamine | Dopa 5-15, epi/norepi <0.1 | Dopa >15, epi/norepi >0.1 |
| CNS | Glasgow Coma Scale | 13-14 | 10-12 | 6-9 | <6 |
| Renal | Creatinine (mg/dL) | 1.2-1.9 | 2.0-3.4 | 3.5-4.9 or UO <500 mL/d | >5.0 or UO <200 mL/d |
A rise ≥2 points from baseline in the context of suspected infection = sepsis. Rising serial scores correlate with increasing mortality. - Harrison's 22e (Table 311-1)
SIRS Criteria (older, now less favored)
Two or more of: fever >38°C or hypothermia <36°C, heart rate >90/min, respiratory rate >20/min or PaCO2 <32 mmHg, WBC >12,000 or <4,000/µL or >10% bands. SIRS is no longer sufficient for a sepsis diagnosis under Sepsis-3 because it is too nonspecific (trauma, pancreatitis, and burns also meet criteria). - Tintinalli's, p. 1039
Other screening tools include NEWS (National Early Warning Score) and MEWS (Modified Early Warning Score), as well as AI-based tools like the Targeted Real-Time Early Warning System.
Laboratory Investigations
Hematology
- CBC with differential: Leukocytosis (WBC >12,000/µL) or leukopenia (<4,000/µL) support infection. A bandemia ≥5-10% (immature forms) suggests active marrow response. Neither finding is sensitive or specific enough alone. Neutropenia (<500 cells/mm³) in a chemotherapy patient warrants empirical IV antibiotics.
- Platelets: Thrombocytopenia (<100,000/µL) is an organ dysfunction marker. Elevated platelets can also be an acute-phase reactant.
- Coagulation panel (PT, aPTT, fibrinogen, D-dimer/fibrin split products): INR >1.5 or aPTT >60 s points to coagulopathy; combined with thrombocytopenia and elevated D-dimer, raises concern for DIC.
Chemistry
- Serum lactate: The single most important prognostic and diagnostic lab value. Levels reflect tissue hypoperfusion:
- Lactate 0-2.5 mmol/L ~ 5% mortality
- Lactate 2.5-4.0 mmol/L ~ 9% mortality
- Lactate >4 mmol/L ~ 28% mortality
- Lactate >2 mmol/L with vasopressors = septic shock criterion
- Serial lactate measurement (repeat within 2-6 hours if initially elevated) guides resuscitation.
- Basic metabolic panel: Elevated creatinine (organ dysfunction), low bicarbonate (metabolic acidosis), elevated anion gap (usually lactic acidosis or DKA).
- Blood gas (ABG/VBG): Classifies acid-base disturbances; metabolic acidosis suggests inadequate perfusion.
- Liver function tests (LFTs): Elevated bilirubin (SOFA criterion), elevated transaminases. Hyperbilirubinemia >4 mg/dL is an organ dysfunction marker; may also suggest biliary source.
- Lipase: Elevation may reveal pancreatitis as the underlying cause of the SIRS response.
- Calcium, magnesium, phosphorus: Should be checked and corrected.
- Blood glucose: Hyperglycemia (>120 mg/dL in non-diabetics) is a general sepsis variable.
Biomarkers
- Procalcitonin (PCT): Rises within hours of bacterial infection. Its primary validated role is serial measurement for antibiotic stewardship (to guide de-escalation/discontinuation) rather than diagnosis. PCT- or CRP-guided antibiotic discontinuation in critically ill septic patients is supported by recent meta-analysis (PMID 38949476, Crit Care Med 2024).
- C-reactive protein (CRP): Rises more slowly (peaks at 24-48 hrs); a value >2 SD above normal is one of the Sepsis-2 inflammatory variables. Less useful for acute diagnosis.
Microbiology
- Blood cultures (x2 sets, aerobic and anaerobic): Should be drawn before antibiotics, but antibiotic initiation must not be delayed waiting for cultures. Positive in only ~30-40% of clinical sepsis cases. Rapid diagnostic tests (RDTs) - such as PCR-based panels or MALDI-TOF - combined with antimicrobial stewardship programs improve outcomes (PMID 38676943, Clin Infect Dis 2024).
- Urine culture and urinalysis: Essential, especially in elderly patients who often lack localizing urinary symptoms.
- Sputum/BAL culture: If pneumonia is suspected.
- Wound swabs, CSF, pleural/peritoneal fluid: As clinically directed.
- Gram stains of any obtained fluid should be reviewed immediately to help narrow empirical therapy.
Imaging
Imaging is used to identify the source of infection, not to diagnose sepsis per se:
| Imaging | Indication |
|---|
| Chest X-ray | All patients with suspected sepsis - focal infiltrate (pneumonia), bilateral infiltrates (ARDS), free air under diaphragm (perforation), pneumomediastinum (esophageal perforation) |
| CT abdomen/pelvis | Suspected intra-abdominal source: appendicitis, diverticulitis, necrotizing pancreatitis, bowel perforation, abscess |
| CT head | Before LP if signs of raised ICP; septic emboli from endocarditis |
| Ultrasound (abdominal) | Suspected cholecystitis, hydronephrosis, biliary obstruction |
| Pelvic ultrasound | Tubo-ovarian abscess, endometritis |
| Transoesophageal echo (TOE) | Suspected endocarditis with valvular vegetations |
| MRI | Necrotizing fasciitis, epidural abscess, osteomyelitis |
| Bone scan | Osteomyelitis (when plain X-ray inconclusive) |
| Soft tissue X-ray | Necrotizing/gas-forming infections (low sensitivity) |
Point-of-care ultrasound (POCUS) is commonly used to assess volume status and cardiac function, though its role in guiding resuscitation remains under study. - Rosen's Emergency Medicine, pp. 2757-2758
Differential Diagnosis
Sepsis mimics must be excluded - in one ICU study, ~25% of admitted "septic" patients retrospectively had non-infectious causes:
- Cardiovascular: Heart failure, cardiogenic shock, cardiac tamponade, myocardial infarction
- Toxicologic: Drug overdose, neuroleptic malignant syndrome
- Metabolic: DKA, adrenal insufficiency, hyperthyroidism
- Inflammatory: Pancreatitis, vasculitis, connective tissue disease, inflammatory bowel disease
- Environmental: Heat stroke, burns, major trauma
- Neurologic: Hypothalamic injury
Diagnostic Approach Summary
Suspected infection (fever, tachycardia, altered mental status, localizing source)
|
↓
qSOFA ≥2 OR clinical concern
|
↓
Immediate workup:
- Serum lactate
- Blood cultures x2 (before antibiotics)
- CBC + differential
- BMP (electrolytes, creatinine, glucose)
- LFTs, coagulation panel (PT/aPTT/fibrinogen)
- Urinalysis + urine culture
- ABG/VBG
- CRP/Procalcitonin
- Chest X-ray
- Source-directed imaging (CT/US)
|
↓
Calculate SOFA score from results
SOFA rise ≥2 from baseline = SEPSIS confirmed
|
↓
If MAP <65 mmHg despite fluids + lactate >2 mmol/L = SEPTIC SHOCK
Recent evidence note: A 2024 systematic review (PMID 39037814, JAMA Network Open) found that AI-based sepsis alert systems in emergency departments were associated with improved adherence to sepsis bundles and reduced mortality, though the evidence base remains heterogeneous.
Sources: Harrison's Principles of Internal Medicine 22e | Tintinalli's Emergency Medicine | Rosen's Emergency Medicine 10e | Morgan & Mikhail's Clinical Anesthesiology 7e | Schwartz's Principles of Surgery 11e