About the CLIF-C ACLF Score (CLIF Consortium Acute-on-Chronic Liver Failure Score)
Three inputs drive it: the CLIF-C OF organ failure score, age, and white cell count, combined as CLIF-C ACLF = 10 × (0.33 × CLIF-C OF + 0.04 × age + 0.63 × ln(white cell count ×10⁹/L) − 2). It applies only to patients who already have acute-on-chronic liver failure, and it predicts mortality more accurately than MELD, MELD-Na or Child-Pugh, cutting prediction error by 19–28% at every time point tested. A score of 70 or above at intensive care admission carried over 80% 28-day mortality, which is the threshold at which futility of organ support is reasonably discussed when transplantation is not available. The score is most useful repeated — it predicted 28-day mortality better at 48 hours and at 3–7 days than at diagnosis.
Formula
CLIF-C ACLF = 10 × (0.33 × CLIF-C OF + 0.04 × age [years] + 0.63 × ln(white cell count [×10⁹/L]) − 2)- CLIF-C OF
- The CLIF Consortium Organ Failure score: six systems (liver, kidney, brain, coagulation, circulation, respiration), each scored 1, 2 or 3, summed to a total between 6 and 18.
- age
- Age in years, entered directly and weighted 0.04 per year.
- ln(white cell count)
- Natural logarithm of the white cell count expressed as ×10⁹/L. A count of 10 contributes 0.63 × ln(10) = 1.45 to the bracket, or 14.5 points to the final score.
- The CLIF-C OF table scores each organ 1–3. This is NOT the CLIF-SOFA table, which scores each organ 0–4, and the thresholds are not the same table collapsed — the kidney bands in particular split at 2.0 and 3.5 mg/dL here.
- Renal replacement therapy forces the kidney subscore to 3, and vasopressor use forces the circulation subscore to 3, regardless of the underlying creatinine or blood pressure.
- Encephalopathy grades I and II are banded together at a brain subscore of 2. CLIF-SOFA scores the grade directly, so the same patient produces different subscores in the two systems.
- The white cell term is logarithmic, not linear, so it matters far more at the low end of the range than the high end.
- The score has no upper cap. Values above 70 are meaningful rather than an indication of a data-entry error.
Interpreting the result
Read the score as a position within an ACLF population, never as an absolute risk for a general cirrhotic patient — everyone this score applies to is seriously ill. The number that carries the most weight in practice is 70: in an intensive care validation cohort, over 80% of patients scoring 70 or above on admission were dead within 28 days, and that threshold has been proposed as the point at which the futility of continued organ support is reasonably discussed where transplantation is not feasible. Treat it as a prompt for that conversation rather than a decision rule, particularly as the score does not distinguish a reversible precipitant from an irreversible one. The more important habit is serial measurement: the derivation study found the score computed at 48 hours and at 3–7 days predicted 28-day mortality significantly better than the score at diagnosis, so a single admission value is the least informative way to use it.
| Score | Band | What it means | Action |
|---|---|---|---|
| < 45 | Lower risk within ACLF | Lower mortality relative to other ACLF patients, which is not the same as low mortality — every patient in scope has organ failure on a cirrhotic background | Treat the precipitant, support failing organs, recompute at 48 hours |
| 45–64 | Intermediate risk | Substantial short-term mortality within an ACLF population | Assess transplant candidacy early; recompute at 48 hours and again at 3–7 days, since the serial score discriminates better |
| 65–69 | High risk | High short-term mortality; approaching the threshold studied for futility | Urgent transplant assessment if the patient is a candidate; recompute at 48 hours before drawing conclusions |
| ≥ 70 | Very high risk | Over 80% 28-day mortality among patients at this score on intensive care admission | Emergency transplant assessment if a candidate; where transplantation is not feasible, discuss goals and the ceiling of organ support |
What the CLIF-C ACLF needs (8 inputs)
- Total bilirubin (liver)
- Scores 1 below 6 mg/dL, 2 from 6 to under 12, and 3 at 12 and above. Note the CLIF-C OF liver bands are wider than the CLIF-SOFA ones.
- Serum creatinine (kidney)
- Scores 1 below 2.0 mg/dL, 2 from 2.0 to under 3.5, and 3 at 3.5 and above. Renal replacement therapy scores 3 whatever the creatinine reads.
- Hepatic encephalopathy grade (brain)
- Scores 1 for grade 0, 2 for grades I–II banded together, and 3 for grades III–IV. This banding is a real difference from CLIF-SOFA, which scores the West Haven grade directly.
- INR (coagulation)
- Scores 1 below 2.0, 2 from 2.0 to under 2.5, and 3 at 2.5 and above.
- Mean arterial pressure (circulation)
- Scores 1 at 70 mmHg and above, 2 below 70. Vasopressor requirement scores 3 regardless of the recorded pressure — a patient whose pressure is normal only because of noradrenaline has circulatory failure.
- PaO₂/FiO₂ or SpO₂/FiO₂ (respiratory)
- PaO₂/FiO₂ scores 1 above 300, 2 from above 200 to 300, and 3 at 200 and below. SpO₂/FiO₂ scores 1 above 357, 2 from above 214 to 357, and 3 at 214 and below. The arterial ratio is used in preference when both are available.
- Age (years)
- Enters the equation directly, weighted 0.04 per year. Twenty years of age difference moves the final score by exactly 8 points, which is often more than a whole organ subscore contributes.
- White cell count (×10⁹/L)
- Entered as ×10⁹/L, numerically the same as ×10³/µL. It enters as a natural logarithm, so it discriminates most at the low end: halving the count from 10 to 5 removes about 4.4 points, while doubling it from 20 to 40 adds the same 4.4.
Units. Bilirubin and creatinine are entered in mg/dL with a µmol/L toggle; divide µmol/L by 17.1 for bilirubin and by 88.4 for creatinine if converting by hand. The white cell count is the one to watch: it must be entered as ×10⁹/L (for example 9.5), which is numerically identical to ×10³/µL and to the 'K/µL' many US laboratories report. Entering an absolute count such as 9,500 will inflate the score by roughly 43 points.
What it returns
- CLIF-C OF score
- The organ failure component, 6 to 18, from six systems scored 1 to 3 each. Reported alongside the final score because it is the part that changes fastest with treatment.
- CLIF-C ACLF score
- A continuous value, in practice roughly 25 to 100. Higher is worse. There is no ceiling built into the formula.
- Risk band
- Anchored on the 70-point threshold from the intensive care validation study, above which 28-day mortality exceeded 80%.
How it is calculated
The CANONIC study established that acute-on-chronic liver failure is a distinct syndrome rather than simply advanced cirrhosis, characterised by organ failure and a high short-term mortality driven by systemic inflammation. Having defined the syndrome, the consortium needed a prognostic score for it, because the instruments in routine use had all been derived elsewhere. Jalan and colleagues developed CLIF-C ACLF in 275 patients with ACLF drawn from the 1,349-patient CANONIC cohort, and validated it externally. They first simplified the organ-failure assessment into the CLIF-C OF score, scoring each of six systems 1 to 3, then found that adding two variables to it improved discrimination substantially: age, which reflects the reserve available for recovery, and the white cell count, which stands in for the systemic inflammatory response that drives the syndrome's progression. The logarithmic treatment of the white cell count reflects that the prognostic information is concentrated at lower counts rather than scaling linearly with leucocytosis.
Facts & figures
| System | Score 1 | Score 2 | Score 3 |
|---|---|---|---|
| Liver — bilirubin (mg/dL) | < 6 | 6 to < 12 | ≥ 12 |
| Kidney — creatinine (mg/dL) | < 2.0 | 2.0 to < 3.5 | ≥ 3.5 or renal replacement |
| Brain — West Haven grade | Grade 0 | Grade I–II | Grade III–IV |
| Coagulation — INR | < 2.0 | 2.0 to < 2.5 | ≥ 2.5 |
| Circulation — MAP (mmHg) | ≥ 70 | < 70 | Vasopressors |
| Respiratory — PaO₂/FiO₂ | > 300 | > 200 to 300 | ≤ 200 |
| Respiratory — SpO₂/FiO₂ | > 357 | > 214 to 357 | ≤ 214 |
Total ranges from 6 to 18. This is the CLIF-C OF table and it is not interchangeable with the 0–4 CLIF-SOFA table — the same patient produces different subscores in each, most obviously for encephalopathy, where CLIF-SOFA scores the grade directly and CLIF-C OF bands grades I and II together.
| Change | Effect on CLIF-C ACLF |
|---|---|
| One point of CLIF-C OF | +3.3 points |
| One year of age | +0.4 points |
| Twenty years of age | +8.0 points |
| White cell count 5 → 10 ×10⁹/L | +4.4 points |
| White cell count 10 → 20 ×10⁹/L | +4.4 points |
| A full organ system failing (1 → 3) | +6.6 points |
Useful for sanity-checking a result. The age term surprises people: a 70-year-old and a 50-year-old with identical organ failures and white cell counts differ by 8 points, more than a whole organ system moving from normal to failing.
Evidence
Derivation — Jalan et al., CANONIC
2014 · n = 275275 patients with acute-on-chronic liver failure drawn from the 1,349-patient CANONIC study of the EASL-CLIF Consortium. The CLIF-C OF score was first derived as a simplified six-system organ failure score, then age and white cell count were added to produce the CLIF-C ACLF score.
Significantly higher predictive accuracy than MELD, MELD-Na and Child-Pugh, reducing the corresponding prediction error rates by 19–28% at 28, 90, 180 and 365 days after ACLF diagnosis, in both the CANONIC and the external validation cohorts. The score computed at 48 hours, 3–7 days and 8–15 days after diagnosis predicted 28-day mortality significantly better than the score at diagnosis.
Futility threshold in intensive care — Engelmann et al.
2018Validation study examining whether the CLIF-C ACLF score can define a threshold for futility of intensive care support in patients with acute-on-chronic liver failure.
Over 80% of patients with a CLIF-C ACLF score of 70 or above at intensive care admission died within 28 days. The authors proposed this as a threshold at which futility may reasonably be discussed where transplantation is not an option.
Independent ICU validation
2023Cirrhotic patients with ACLF requiring intensive care, comparing the CLIF-C OF and CLIF-C ACLF scores for transplant-free survival.
For 28-day mortality, CLIF-C ACLF achieved an AUROC of 0.717 (95% CI 0.626–0.809) against 0.652 (95% CI 0.554–0.750) for CLIF-C OF alone, confirming that the age and white cell count terms add discrimination over the organ failure score by itself.
CANONIC — the study that defined the syndrome
2013 · n = 1,343Prospective observational study of 1,343 patients admitted with acutely decompensated cirrhosis across 29 European units, which established acute-on-chronic liver failure as a distinct syndrome with its own organ failure definitions and mortality profile.
Defined the organ failure criteria and ACLF grades on which the CLIF-C family is built, and established the high short-term mortality that motivated a dedicated prognostic score.
How it compares
CLIF-C ACLF vs CLIF-SOFA
Sequential, not alternative — CLIF-SOFA (with the CANONIC definitions) establishes whether ACLF is present and its grade; CLIF-C ACLF then quantifies mortality risk in the patients who have it.
CLIF-SOFA scores each of six organ systems 0–4 and is used to identify organ failures and assign an ACLF grade. CLIF-C OF is a deliberately simplified 1–3 version of the same six systems, built as the organ-failure component of this prognostic score. The tables are not interchangeable: the thresholds differ, and encephalopathy in particular is banded I–II here but scored by grade in CLIF-SOFA. Use the first to define the syndrome, the second to prognosticate within it.
CLIF-C ACLF vs MELD / MELD-Na
CLIF-C ACLF is the more accurate score in established ACLF, reducing prediction error by 19–28% against MELD and MELD-Na at every time point tested.
MELD was derived to predict survival after TIPS and now governs elective transplant allocation; it contains no measure of circulatory, respiratory or cerebral failure and no measure of inflammation, so it systematically underestimates risk in a patient with several organs failing at once. CLIF-C ACLF was derived in ACLF patients specifically and outperformed MELD, MELD-Na and Child-Pugh at 28, 90, 180 and 365 days. MELD retains its role in allocation; it is simply the wrong instrument for prognosis in this syndrome.
CLIF-C ACLF vs AARC-ACLF (APASL)
They describe different patient populations because APASL and EASL-CLIF define ACLF differently — pick the one matching the definition your unit uses.
The APASL definition requires acute hepatic insult with jaundice and coagulopathy in chronic liver disease and excludes patients with prior decompensation, so it centres on liver failure. The EASL-CLIF definition behind CLIF-C ACLF is built on extrahepatic organ failure and includes previously decompensated patients. AARC therefore uses liver-centred variables (bilirubin, INR, lactate, creatinine, encephalopathy) while CLIF-C ACLF weighs six organ systems plus age and inflammation. Neither is wrong; they answer the same question about different cohorts.
CLIF-C ACLF vs Child-Pugh score
Not suitable here — Child-Pugh grades chronic disease severity and was outperformed by CLIF-C ACLF at every time point in the derivation study.
Child-Pugh includes albumin, whose 20-day half-life makes it nearly uninformative over the days across which ACLF evolves, and it captures no renal, circulatory or respiratory failure at all. It remains useful for staging chronic liver disease and for drug-dosing decisions, but it is not a prognostic instrument for a patient with multi-organ failure.
Pearls & pitfalls
- CLIF-C OF is not CLIF-SOFA. One scores each organ 1–3, the other 0–4, and they are not the same table collapsed. Using CLIF-SOFA subscores in this formula produces a meaningless number.
- Encephalopathy grades I and II are banded together at a brain subscore of 2 here, whereas CLIF-SOFA scores the West Haven grade directly. This is the single easiest subscore to carry across incorrectly.
- The white cell count goes in as ×10⁹/L, which is the same number as ×10³/µL. Entering an absolute count such as 9,500 instead of 9.5 will inflate the score by about 43 points.
- Vasopressors force circulation to 3 and renal replacement forces kidney to 3. A normal-looking blood pressure on noradrenaline is circulatory failure, not a normal circulation.
- This score does not diagnose ACLF and does not assign a grade. Establish both first using the CANONIC definitions; applying this score to a decompensated cirrhotic without organ failure is outside what it was derived for.
- A single value on admission is the weakest way to use it. The derivation study showed the score at 48 hours and at 3–7 days predicted 28-day mortality significantly better than the score at diagnosis.
- The 70-point futility threshold comes from an intensive care cohort at the point of ICU admission. It is a prompt for a conversation, not a rule, and it does not distinguish a treatable precipitant from an untreatable one.
- Age carries more weight than clinicians expect — 8 points across a 20-year difference, more than a whole organ system moving from normal to failing.
Critical actions
- Confirm ACLF is actually present, using the CANONIC organ-failure definitions, before applying this score at all.
- Identify and treat the precipitant immediately — bacterial infection above all, then alcohol-related hepatitis, variceal bleeding and drug injury. The score describes prognosis; it does not tell you what will change it.
- Recompute at 48 hours and again at 3–7 days, and act on the trajectory rather than the admission value.
- Assess transplant candidacy early rather than after further organs fail, since the window narrows as the score rises.
- At a score of 70 or above without a transplant option, open an explicit conversation with the family and the wider team about goals and the ceiling of organ support.
- Recheck the white cell count entry if the score looks implausible — a unit error there is the commonest cause of a wildly wrong result.
Why this score exists
The score exists because the CANONIC investigators had just demonstrated that acute-on-chronic liver failure was a distinct syndrome, and then found themselves without an instrument to prognosticate in it. Every score being applied to these patients had been derived somewhere else for something else — MELD for post-TIPS survival, Child-Pugh for shunt surgery risk — and all of them underestimated mortality in a patient with several organs failing at once. The two additions beyond organ failure are the interesting editorial choice: age, and the white cell count. Neither appears in MELD or Child-Pugh, and the white cell count in particular reflects the consortium's position that ACLF is driven by systemic inflammation rather than simply by the failure of the liver. Its logarithmic form encodes that the prognostic signal sits at the lower end of the range rather than scaling with the height of a leucocytosis.
About the creator
First author, 2014 derivation study; Institute for Liver and Digestive Health, University College London
Led the development and validation of the CLIF-C OF and CLIF-C ACLF scores within the EASL-CLIF Consortium.
Co-author of the 2014 derivation study.
Data Management Centre, EASL-CLIF Consortium
Statistical lead for the CANONIC study and the CLIF-C score development.
Limitations
- Applies only to patients who already meet a definition of ACLF. It neither diagnoses the syndrome nor assigns a grade, and it has not been validated in decompensated cirrhosis without organ failure.
- Derived and validated predominantly in European cohorts under the EASL-CLIF definition, which differs from the APASL definition used across much of Asia — the two select different patients, so performance does not transfer automatically.
- The white cell count is a crude proxy for systemic inflammation and is confounded by corticosteroids, active infection, gastrointestinal bleeding and recent transfusion.
- Prognostic only. It identifies who is likely to die but says nothing about which intervention would change that, and it does not separate reversible precipitants from irreversible ones.
- The 70-point futility threshold derives from a specific intensive care population measured at ICU admission. Applying it at a different point in the illness, or to a patient with a treatable precipitant, is not what the study supports.
- Discrimination is good but not excellent — AUROC around 0.72 for 28-day mortality in independent ICU validation. It should inform discussion, never settle it alone.
- Performance depends on organ-failure data being complete; a missing oxygenation ratio defaults the respiratory system to its lowest subscore and will understate the score.
If you are the patient
Acute-on-chronic liver failure means that someone with long-standing liver disease has become suddenly and severely unwell, with several organs — not just the liver — starting to fail at once. The CLIF-C ACLF score is a number the medical team calculates to judge how serious the situation is over the coming weeks. It combines how many organs are struggling and how badly, the person's age, and a blood test that reflects how much inflammation is present. A higher number means a higher risk. The team will usually recalculate it after a couple of days, because how the number moves with treatment tells them more than the first value did — a score that falls as an infection is treated means something quite different from one that keeps climbing. If the score is very high and a liver transplant is not possible, the team may start a conversation about what treatment can realistically achieve. That conversation is guided by the number, not decided by it.
Frequently asked questions
What is the CLIF-C ACLF score?#
A prognostic score for patients with acute-on-chronic liver failure, calculated as 10 × (0.33 × CLIF-C OF score + 0.04 × age + 0.63 × ln(white cell count ×10⁹/L) − 2). It predicts short-term mortality more accurately than MELD, MELD-Na or Child-Pugh in this population.
Is CLIF-C OF the same as CLIF-SOFA?#
No, and confusing them is the commonest error with this score. CLIF-SOFA scores each of six organ systems 0–4; CLIF-C OF scores the same six systems 1–3 with different thresholds, giving a total of 6 to 18. Encephalopathy is the clearest divergence — CLIF-C OF bands grades I and II together, while CLIF-SOFA scores the grade directly.
What CLIF-C ACLF score indicates futility?#
A score of 70 or above at intensive care admission was associated with over 80% 28-day mortality, and has been proposed as a threshold at which futility of continued organ support may reasonably be discussed when transplantation is not feasible. It is a prompt for that discussion rather than a decision rule, and it does not account for a reversible precipitant.
How often should the score be repeated?#
At least at 48 hours and again at 3–7 days. The derivation study found the score computed at those points predicted 28-day mortality significantly better than the score at diagnosis, which makes the trajectory more informative than any single value.
What units does the white cell count use?#
×10⁹/L, for example 9.5 — the same number as ×10³/µL or K/µL. Entering an absolute count such as 9,500 will add roughly 43 points to the score, so it is worth a second look whenever a result seems implausibly high.
Can I use CLIF-C ACLF instead of MELD for transplant listing?#
No. MELD and its variants govern elective transplant allocation and that role is unchanged. CLIF-C ACLF is a prognostic score for patients who already have ACLF, where it is the more accurate predictor of mortality, but it is not an allocation instrument.
Does this score diagnose acute-on-chronic liver failure?#
No. Establish the diagnosis and the ACLF grade first using the CANONIC organ-failure definitions, then apply this score to quantify risk in patients who meet them.
Why does age carry so much weight?#
Age is weighted 0.04 per year, so twenty years adds exactly 8 points to the final score — more than a whole organ system moving from normal to failing, which adds 6.6. It reflects the physiological reserve available for recovery, something MELD and Child-Pugh omit entirely.
References
Original / primary reference
- Jalan R, Saliba F, Pavesi M, et al. Development and validation of a prognostic score to predict mortality in patients with acute-on-chronic liver failure. J Hepatol. 2014;61(5):1038-1047.
- Moreau R, Jalan R, Gines P, et al. Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis. Gastroenterology. 2013;144(7):1426-1437.
Validation and thresholds
Guidelines
- European Association for the Study of the Liver. EASL Clinical Practice Guidelines for the management of patients with decompensated cirrhosis. J Hepatol. 2018;69(2):406-460.
- Sarin SK, Choudhury A, Sharma MK, et al. Acute-on-chronic liver failure: consensus recommendations of the Asian Pacific Association for the Study of the Liver (APASL): an update. Hepatol Int. 2019;13(4):353-390.