Clinical background
The SOFA score was developed in 1996 by a working group within the European Society of Intensive Care Medicine to quantify and follow the degree of organ dysfunction in critically ill patients over time [1]. The instrument serves two clinical purposes. First, it gives a structured, reproducible overview of the extent of organ failure in intensive care, where the change over time is more informative than a single score. Second, and more concretely, the Sepsis-3 consensus of 2016 defines sepsis as an acute increase in SOFA of ≥2 points in patients with suspected infection, which makes the score an integral part of the modern definition of sepsis [2].
The decision the instrument serves is therefore not purely prognostic. It is primarily descriptive and longitudinal: to measure whether and how far organ function deteriorates, and to set a threshold for when organ dysfunction is pronounced enough to qualify for a diagnosis of sepsis.
Calculating the SOFA score
The SOFA score is the sum of six organ subscores, each 0 to 4, for respiration, coagulation, liver, circulation, CNS and kidneys:
The total range is 0 to 24. The individual subscores:
- Respiration (PaO₂/FiO₂, mmHg): ≥400 scores 0, <400 scores 1, <300 scores 2, <200 with respiratory support scores 3, <100 with respiratory support scores 4.
- Coagulation (platelets, ×10³/µL): ≥150 scores 0, <150 scores 1, <100 scores 2, <50 scores 3, <20 scores 4.
- Liver (bilirubin, mg/dL): <1.2 scores 0, 1.2 to 1.9 scores 1, 2.0 to 5.9 scores 2, 6.0 to 11.9 scores 3, ≥12.0 scores 4.
- Circulation: MAP ≥70 scores 0, MAP <70 scores 1, dopamine ≤5 µg/kg/min or any dobutamine scores 2, dopamine >5 µg/kg/min or noradrenaline ≤0.1 µg/kg/min scores 3, dopamine >15 µg/kg/min or noradrenaline >0.1 µg/kg/min scores 4.
- CNS (Glasgow Coma Scale): 15 scores 0, 13 to 14 scores 1, 10 to 12 scores 2, 6 to 9 scores 3, <6 scores 4.
- Kidneys (creatinine, mg/dL): <1.2 scores 0, 1.2 to 1.9 scores 1, 2.0 to 3.4 scores 2, 3.5 to 4.9 or urine output <500 mL/day scores 3, ≥5.0 or urine output <200 mL/day scores 4.
The derivation cohort consisted of patients in intensive care units at several European hospitals, and the instrument was originally published under the name "Sepsis-related Organ Failure Assessment" before it was changed to "Sequential Organ Failure Assessment" to emphasise that it is not restricted to sepsis [1]. The working group chose the variables on the basis of clinical face validity and availability in intensive care units, not on statistical modelling. The score was initially validated in a prospective multicentre study of just over 1,400 intensive care patients, in which mortality was found to rise stepwise with increasing total score.
Interpretation in practice
The SOFA score has two main modes of interpretation depending on context.
As a longitudinal instrument in intensive care. The score is calculated on admission and thereafter regularly, usually daily. The trend is what is clinically relevant: a rising score indicates worsening organ dysfunction and should prompt an increased level of monitoring, investigation of the underlying cause and possible escalation of organ support. A falling score indicates recovery and can support decisions about weaning from ventilation or vasopressors.
As a definition of sepsis (Sepsis-3). In patients with suspected infection, an acute increase of ≥2 points from baseline defines sepsis [2]. In patients with no known previous organ dysfunction, the baseline is assumed to be 0. This threshold was not chosen as an optimal prognostic cut-off, but because it captures patients whose organ failure is pronounced enough to justify a diagnosis of sepsis and thereby aggressive treatment.
| Change in SOFA | Interpretation | Clinical action |
|---|---|---|
| ≥2 points from baseline with suspected infection | Sepsis by Sepsis-3 | Start the sepsis bundle: blood cultures, broad-spectrum antibiotics, fluids, lactate |
| Rising trend during intensive care | Progressive organ failure | Consider escalating organ support, search for a focal source |
| Unchanged or falling | Stable or improving | Continue the current strategy, consider weaning |
Validation and performance
The largest empirical evaluation of SOFA as a criterion for sepsis was performed by Seymour et al. within the Sepsis-3 work [2]. In a primary cohort of 148,907 encounters with suspected infection at 12 hospitals in southwestern Pennsylvania (2010 to 2012), with 4% in-hospital mortality, discrimination for in-hospital mortality was calculated. Among ICU patients (n = 7,932 with suspected infection, 16% mortality), SOFA had an AUROC of 0.74 (95% CI 0.73 to 0.76), comparable to LODS (0.75) and significantly better than both SIRS (0.64) and qSOFA (0.66) [2]. The results were confirmed in four external datasets comprising over 700,000 encounters in total.
Performance falls outside intensive care, however. In the non-ICU cohort of Seymour's material (n = 66,522, 3% mortality), SOFA had an AUROC of 0.79, which was lower than qSOFA (0.81) [2]. A Danish study of 2,045 patients with infection in an emergency department found an AUROC of 0.68 for SOFA ≥2 for 28-day mortality, formally better than both SIRS (0.52) and qSOFA (0.63), but with low clinical usefulness [3]. Sensitivity was 61% and specificity 67%, and the authors concluded that the prognostic accuracy of SOFA in the emergency department is poor [3].
A German study of 13,780 surgical patients in intermediate care units (IMCUs) and ICUs showed that SOFA performed best in the most severely ill [4]. Among patients cared for in both an IMCU and an ICU, the AUROC for mortality was 0.73 for SOFA compared with 0.59 for qSOFA. Among patients cared for in an IMCU only, SOFA had essentially no predictive ability (AUROC 0.52 to 0.63), while qSOFA performed better (AUROC 0.82) [4]. Neither score could predict suspected infection at any level of care.
An updated version, SOFA-2, was published in 2025 by Ranzani et al. with the aim of improving discrimination through revised variable weights and thresholds [5]. This is not what the present calculator implements, however, and it has not yet become established in clinical practice.
Limitations
The SOFA score was designed for intensive care patients and requires laboratory values and data on vasopressor doses that are not always available outside the ICU. In the emergency department or on a general ward, the wait for laboratory results is a practical obstacle, and the instrument's discrimination is poorer in these settings [2, 3].
The score presupposes that the baseline is known. In patients with chronic organ dysfunction, for example chronic renal failure or chronic liver disease, the baseline SOFA may already be raised on arrival. Sepsis-3 recommends assuming a baseline of 0 in patients with no known previous organ dysfunction, but in patients with established organ failure the clinician must estimate the change caused by the infection, which introduces subjectivity into the interpretation [2].
The circulation subscore uses dopamine as the reference vasopressor, reflecting the clinical practice prevailing when the score was developed in the 1990s. Dopamine is now rarely used as first-line treatment in septic shock, where noradrenaline predominates. This means that the dopamine thresholds within the subscore may seem outdated, even though noradrenaline is included in the higher point levels.
The instrument was not designed for children or pregnant women, and it has not been validated in these populations. It quantifies organ dysfunction but says nothing about aetiology. A high SOFA score may be caused by trauma, toxic injury or haemorrhage just as readily as by infection, and the score should therefore never be interpreted in isolation from the clinical context.
References
- Vincent JL et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. Intensive Care Med. 1996. PMID: 8844239
- Seymour CW 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
- Abdullah SMOB et al. Prognostic Accuracy of SOFA, qSOFA, and SIRS for Mortality Among Emergency Department Patients with Infections. Infect Drug Resist. 2021. PMID: 34321893
- Koch C 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
- Ranzani OT et al. Development and Validation of the Sequential Organ Failure Assessment (SOFA)-2 Score. JAMA. 2025. PMID: 41159833
