CLIF-SOFA Score in Acute-on-Chronic Liver Failure

The CLIF-SOFA score is a bedside tool designed to gauge how severely a person with chronic liver disease is deteriorating when a sudden crisis pushes multiple organs toward failure. Developed by the European Association for the Study of the Liver’s CLIF Consortium, it evaluates six organ systems on a numerical scale and produces a composite number that helps doctors estimate the risk of death in the coming days and weeks. The score has become central to diagnosing and grading a condition called acute-on-chronic liver failure (ACLF), and in several head-to-head comparisons it outperforms older liver-specific scoring systems at predicting short-term mortality.

What the Score Actually Measures

The CLIF-SOFA score is an adaptation of the Sequential Organ Failure Assessment (SOFA) score that intensive-care doctors have long used for critically ill patients of all kinds. The original SOFA was not built with liver disease in mind, so the CLIF Consortium modified several of its components to better reflect what goes wrong when cirrhosis destabilizes. The resulting score assigns a subscore of 0 to 4 for each of six organ systems: the liver, kidneys, brain, coagulation (blood clotting), circulation, and lungs.¹ The total can range from 0 to 24, with higher numbers indicating more widespread and severe organ dysfunction.²

Each subscore is driven by a specific lab value or clinical measurement. Liver function is scored using bilirubin levels. Kidney function uses serum creatinine, with higher subscores assigned when creatinine climbs above 2 mg/dL or when a patient needs dialysis.³ Coagulation is assessed through the international normalized ratio (INR), which reflects how well the blood clots. Circulation is judged by mean arterial pressure and the need for drugs that raise blood pressure. Lung function is measured by the ratio of oxygen in the blood to the concentration of oxygen being delivered through a breathing device (the PaO2/FiO2 ratio).⁴ One key departure from the standard SOFA is the brain component: instead of the Glasgow Coma Scale, which is a general consciousness assessment, the CLIF-SOFA uses hepatic encephalopathy grading, because the confused or drowsy state caused by liver failure looks different from other causes of altered consciousness and is harder to capture on the standard scale.⁵

How It Came About

The score was born out of the CANONIC study, a large European investigation that enrolled 1,349 hospitalized patients with cirrhosis to better define what ACLF actually is and how to predict who would survive it.⁶ Before the CANONIC study, there was no consensus scoring tool tailored to cirrhosis patients in acute crisis. The MELD score, used to prioritize patients on the transplant waiting list, focuses on liver and kidney labs but ignores the lungs, brain, and circulation entirely. The Child-Pugh score, another long-standing liver assessment, captures ascites and encephalopathy but was never designed for the kind of rapid, multi-organ deterioration that defines ACLF. The CLIF-SOFA filled that gap by borrowing the multi-organ logic of the ICU’s SOFA score and reshaping it for liver disease.

How Well It Predicts Mortality

The strongest evidence for the CLIF-SOFA’s usefulness comes from studies comparing it head-to-head against other scoring systems. In a retrospective study of patients with alcoholic cirrhosis, the CLIF-SOFA achieved an area under the receiver operating characteristic curve (a measure of how accurately a test distinguishes people who will die from those who will survive) of 0.943 for four-week mortality, significantly outperforming five other scoring methods. In a subset of patients receiving only supportive care, the score’s accuracy was even higher, reaching 0.952 at one week and 0.959 at four weeks.⁷ For context, a perfect score on that scale would be 1.0, and anything above 0.90 is considered excellent discrimination.

The score also holds up over longer time horizons, though with less precision. One study found it predicted 28-day, 90-day, and 365-day mortality with values of 0.71, 0.75, and 0.66, respectively. A cutoff score above 11 was able to identify patients at higher risk across all three intervals, with specificities running from 72% to 89%.⁸ A systematic review found that, across multiple studies, the CLIF-SOFA and its successor scores consistently outperformed other models at predicting 28-day mortality and remained useful at 90, 180, and 365 days.⁹

CLIF-SOFA Versus MELD, Child-Pugh, and Other Scores

The MELD score is probably the most familiar liver disease scoring tool to anyone who has followed transplant medicine. It uses three lab values (bilirubin, creatinine, and INR) to estimate 90-day mortality and allocate donor organs. The CLIF-SOFA has been shown to beat MELD and MELD-Na (a sodium-adjusted variant) in predicting 90-day mortality among ACLF patients.¹⁰ The reason is straightforward: MELD does not account for hemodynamic collapse, respiratory failure, or brain dysfunction. When a patient’s liver labs are middling but they are on a ventilator and requiring blood-pressure medications, MELD underestimates how sick they are.

Compared to the standard SOFA used in general ICU settings, the CLIF-SOFA’s liver-specific modifications give it an edge for cirrhosis populations. A multicenter study of cirrhosis patients admitted with infections found the CLIF-SOFA predicted in-hospital mortality significantly better than the Sepsis-3 criteria (0.80 versus 0.75) and the quick SOFA bedside screen (0.80 versus 0.67).¹¹ This makes intuitive sense: a tool designed for cirrhosis patients captures their specific failure patterns better than a general-purpose ICU tool.

Because of this track record, some researchers have suggested that CLIF-SOFA should be considered alongside MELD when making transplant decisions for ACLF patients, since the MELD-based allocation system can undervalue patients who are rapidly deteriorating through multi-organ failure.¹²

The Evolution Into CLIF-C Scores

The original CLIF-SOFA was something of a first draft. After the CANONIC study established the basic framework, the CLIF Consortium refined it into a family of scores. The most important offspring is the CLIF Consortium Organ Failure score (CLIF-C OFs), which simplifies the six subscores into a narrower 6-to-18 range and is used specifically to define ACLF grades.¹³ From there, two prognostic formulas were built on top of it:

CLIF-C ACLF: For patients who meet the ACLF criteria, this score folds in the CLIF-C OFs result along with the patient’s age and white blood cell count to generate a mortality prediction.
CLIF-C AD: For patients with acute decompensation of cirrhosis who do not meet the ACLF threshold, this score uses age, creatinine, INR, white blood cell count, and sodium to estimate prognosis.

Both formulas were validated against MELD and other scoring systems and generally performed better at predicting death at various time points.¹⁴ The inclusion of age and white blood cell count reflects the recognition that how old a patient is and how intense their inflammatory response is both matter independent of how many organs have failed.

The ACLF Grading System It Supports

The CLIF-SOFA’s practical impact is not just a number on a chart. It underpins a grading system that sorts patients into ACLF grades 1, 2, and 3, depending on how many organ systems have failed. A patient with single kidney failure, or single non-kidney organ failure combined with milder kidney or brain dysfunction, falls into grade 1. Two failing organs means grade 2, and three or more means grade 3. As the grades climb, survival drops sharply. In a validation study of 1,470 patients, survival rates fell significantly with each increase in ACLF grade.¹⁵ This grading helps clinicians make concrete decisions: a grade 3 patient who shows no improvement after several days of intensive care may be a candidate for urgent transplant evaluation, or, if transplant is not feasible, for a goals-of-care conversation about comfort measures.

Competing Definitions of ACLF Around the World

One complication with the CLIF-SOFA approach is that not everyone agrees on what ACLF actually is. The EASL-CLIF definition, which the score was built for, centers on extrahepatic organ failures, particularly the kidneys. Under this framework, a person with cirrhosis can be diagnosed with ACLF even without liver failure, as long as kidney failure is present. The Asia-Pacific Association for the Study of the Liver (APASL) takes a different view, requiring liver failure itself to be the primary driver.¹⁶

This difference matters. A head-to-head comparison in a smaller study found that the CLIF-SOFA criteria classified 38 out of 50 patients as having ACLF, while the APASL criteria identified only 19. The 28-day mortality gap between ACLF and non-ACLF groups was stark under CLIF-SOFA (roughly 47% versus 8%), but virtually nonexistent under APASL criteria (39% versus 37%).¹⁷ In other words, the CLIF-SOFA-based definition was far better at separating high-risk from low-risk patients. However, a recent analysis has raised questions about the reliance on renal failure in the EASL-CLIF framework, noting that patients who met the ACLF definition solely through kidney failure had significantly higher survival rates than those with more classic multi-organ presentations. Some of these patients may have had transient kidney injuries from infection or bleeding that resolved on their own.¹⁸

What Drives the Organ Failures the Score Captures

The CLIF-SOFA is essentially a snapshot of damage, but understanding what causes that damage helps explain why the score changes over time. Research into the biology of ACLF has shown that the process is driven by intense systemic inflammation. Molecules released by bacteria (in infection-triggered ACLF) and by damaged liver and gut tissues flood the bloodstream and set off an immune response that can spiral out of control. The resulting organ failures appear to stem from a combination of inadequate blood flow to tissues, direct immune-mediated damage to organs, and dysfunction in the mitochondria that supply energy to cells.¹⁹ This is why infections, heavy alcohol use, and gastrointestinal bleeding are the most common triggers of ACLF: all of them pour inflammatory signals into an already fragile system.

The Score in Children

Pediatric liver disease is uncommon enough that validated scoring tools for children with ACLF barely exist. A few groups have tried using the CLIF-SOFA in children by modifying the thresholds for blood pressure, creatinine, and hepatic encephalopathy grading to account for age-related differences. In one pediatric study, the modified CLIF-SOFA was better at predicting poor outcomes than scores designed specifically for children, including the Pediatric End-Stage Liver Disease score, the Child-Pugh score, and the Pediatric Risk of Mortality-III.²⁰ These findings are preliminary and based on small numbers, but they suggest the multi-organ-failure logic of the CLIF-SOFA may translate across age groups with appropriate calibration.

Automating the Score From Hospital Records

One practical barrier to using the CLIF-SOFA routinely is that it requires someone to look up six different lab values and clinical observations, grade each on a subscale, and add them up. In a busy ICU, scores can go stale within hours as a patient’s condition shifts. Researchers have described a method for pulling the necessary data automatically from electronic health records, calculating rolling CLIF-SOFA and CLIF-C scores that update as new labs and vitals are entered. The trickiest piece was hepatic encephalopathy, which is a bedside clinical judgment rather than a lab value. The team solved this by mapping nursing flowsheet entries (documentation of mental status, orientation, and level of consciousness) to encephalopathy grades, then tested the approach across 239 patient admissions over five years.²¹ If this kind of automation becomes standard, it could make real-time CLIF-SOFA tracking feasible in any hospital with modern electronic charting, rather than limiting its use to research centers where a dedicated team calculates it manually.

The Financial Weight of ACLF Admissions

Behind the scoring and grading lies an enormous economic burden. Patients with ACLF require intensive monitoring, often need mechanical ventilation and dialysis, and spend extended periods in ICU beds. A nationwide study in Thailand found that the average cost of an ACLF hospitalization was roughly three and a half times higher than a hospitalization for cirrhosis without ACLF.²² In settings where reimbursement systems do not account for this intensity gap, hospitals absorb large losses caring for these patients. Accurate prognostic scoring is partly a clinical tool and partly a resource-allocation tool: identifying early which patients are likely to recover with intensive support, which need urgent transplant evaluation, and which are unlikely to benefit from continued aggressive care helps allocate ICU beds and transplant organs where they can do the most good.