Overview

The Revised Natural History Model for Primary Sclerosing Cholangitis, published by Kim, Therneau, Wiesner, and colleagues at the Mayo Clinic in 2000, is the most widely validated survival prediction tool specifically designed for primary sclerosing cholangitis (PSC). It was developed to address the central limitation of the original Mayo PSC model (Wiesner et al., 1989): its dependence on histological staging from liver biopsy, which is invasive, subject to sampling error, and infrequently repeated in clinical practice.

The revised model eliminates liver biopsy entirely, replacing histological stage with two non-invasive surrogates of portal hypertension, hepatocellular injury, and cholestatic progression. The resulting five-variable model uses age, serum bilirubin, serum aspartate aminotransferase (AST), history of variceal bleeding, and splenomegaly to generate a continuous risk score. This risk score is then used with a baseline survival function to estimate the probability of transplant-free survival at defined time horizons, typically 1, 2, and 4 years.

The Revised Mayo PSC Risk Score occupies a central role in the clinical management of PSC: it guides the timing of liver transplantation evaluation and listing, provides a validated prognostic benchmark for clinical trial stratification, enables serial monitoring of disease progression during follow-up, and facilitates evidence-based discussions about prognosis with patients. Its non-invasive design makes it practical for routine outpatient use.

Primary Sclerosing Cholangitis: Disease Biology and Epidemiology

Primary sclerosing cholangitis is a chronic, progressive, immune-mediated fibro- inflammatory disease of the bile ducts characterised by diffuse intrahepatic and extrahepatic biliary stricturing, obliterative fibrosis, and progressive cholestasis leading to cirrhosis, hepatic failure, and portal hypertension over years to decades. It is one of the leading indications for liver transplantation in Western countries.

PSC has an estimated prevalence of 6 to 16 cases per 100,000 population in Northern European and North American registries, with a global prevalence increase attributable to rising awareness and improved diagnostic imaging. The disease predominantly affects young to middle-aged men (male-to-female ratio approximately 2:1), with a median age at diagnosis of 35 to 40 years, though presentation across all age groups is well documented. Approximately 70 to 80% of PSC patients in Western populations have coexisting inflammatory bowel disease (IBD), most commonly ulcerative colitis, and this association is a defining epidemiological feature that distinguishes PSC from other cholestatic liver diseases.

The pathogenesis of PSC remains incompletely understood, involving dysregulated immune activation targeting biliary epithelia, an abnormal gut-liver axis (the “leaky gut” hypothesis), and genetic susceptibility across multiple HLA haplotypes and non-HLA loci. Genome-wide association studies have identified more than 20 susceptibility loci including HLA-B*08, HLA-DRB1*03:01, FUT2, CARD9, and IL2, implicating both adaptive immune dysregulation and intestinal barrier dysfunction.

Biliary inflammation and periductal fibrosis produce progressive obliterative cholangiopathy of both large and small bile ducts. Large-duct PSC, characterised by the classic beaded appearance of the biliary tree on cholangiography (MRCP or ERCP), is the predominant subtype. Small-duct PSC (a variant with normal cholangiography but PSC histology) has a more indolent course and different natural history. The revised Mayo model was derived in patients with large-duct PSC, which is the clinically dominant and prognostically more severe subtype.

A defining complication distinguishing PSC from most other chronic liver diseases is its association with malignancy. PSC carries a lifetime risk of cholangiocarcinoma (CCA) of 10 to 15%, with a cumulative 10-year incidence of approximately 7 to 9% in large cohort studies. Gallbladder carcinoma occurs in approximately 3 to 14% of PSC patients with gallbladder polyps or masses, and colorectal carcinoma risk is substantially elevated in PSC-IBD compared with IBD alone (relative risk approximately 4-fold). These malignancy risks complicate prognostication because death or transplant may be preceded by de novo CCA that is not predicted by the natural history model.

Natural History and Clinical Course

The natural history of PSC is characterised by considerable inter-individual variability and frequent periods of apparent stability punctuated by acute complications. The median transplant-free survival from diagnosis has been estimated at approximately 12 to 18 years in population-based cohorts, though this figure reflects the current era of earlier diagnosis through improved MRCP availability. Pre-transplant era Mayo Clinic series reported shorter median survivals of 9 to 12 years, likely reflecting ascertainment bias toward more advanced disease.

Approximately 20 to 30% of patients have an asymptomatic presentation detected incidentally through elevated liver biochemistry (predominantly alkaline phosphatase and gamma-glutamyl transferase) or during IBD workup. These asymptomatic patients nonetheless have demonstrable cholangiographic abnormalities and progress over time, though more slowly than symptomatic patients.

The clinical trajectory follows a broadly predictable pattern:

• Asymptomatic or minimally symptomatic phase (years to decades): Elevated cholestatic enzymes, normal or mildly elevated bilirubin, and detectable biliary stricturing on MRCP without clinical symptoms. Fatigue and pruritus may be present intermittently. Disease progression is occult by clinical examination but demonstrable histologically and biochemically.
• Symptomatic cholestasis phase: Progressive jaundice, worsening pruritus, fatigue, fat-soluble vitamin deficiencies (A, D, E, K), and metabolic bone disease. Serum bilirubin rises progressively, initially fluctuating before plateauing at elevated levels.
• Portal hypertension phase: Splenomegaly, thrombocytopenia, variceal formation, and ascites as hepatic fibrosis advances to cirrhosis. At this stage, the pace of clinical deterioration accelerates significantly.
• Hepatic decompensation and pre-transplant phase: Variceal haemorrhage, spontaneous bacterial peritonitis, hepatic encephalopathy, refractory ascites, and progressive jaundice heralding end-stage liver disease.

Dominant biliary strictures (focal, high-grade narrowing superimposed on diffuse disease) complicate approximately 15 to 20% of PSC patients and carry an important differential from cholangiocarcinoma. Bacterial cholangitis episodes arise from biliary stasis and instrumentation, contributing to accelerated deterioration in a subset of patients. The unpredictable timing and severity of these complicating events accounts for much of the individual variability in the PSC natural history trajectory.

Historical Context: From Original to Revised Mayo Model

The Original Mayo PSC Model (1989)

The original Mayo Clinic natural history model for PSC, published by Wiesner and colleagues in Hepatology in 1989, was derived from a cohort of 174 PSC patients using Cox proportional hazards regression with time-to-liver-transplantation or death as the primary endpoint. It identified five independent prognostic variables:

• Age (years)
• Serum bilirubin (μmol/L or mg/dL, log-transformed)
• Histological stage (I to IV, Larusso/Ludwig classification)
• Splenomegaly (present or absent on physical examination or imaging)
• Serum haemoglobin (g/dL)

The model demonstrated strong prognostic discrimination and became the reference standard for PSC prognostication throughout the 1990s. It was used to guide transplant listing decisions, enrol patients in clinical trials at comparable risk levels, and communicate prognosis with patients and families.

However, the requirement for liver biopsy histological staging was a fundamental practical limitation. Liver biopsy in PSC is associated with particular risks compared with other chronic liver diseases: the biliary anatomy abnormalities may distort the peri-hepatic space, post-biopsy bile leak is a potential complication, and the characteristic periductal fibrosis of PSC produces significant sampling variability. In clinical practice, many PSC patients undergo liver biopsy once at diagnosis but rarely thereafter, making repeated serial risk scoring impractical. Additionally, biopsy is increasingly avoided as MRCP has become the diagnostic standard for PSC, reducing the clinical frequency of histological staging in routine practice.

Derivation of the Revised Model (2000)

Kim and colleagues revisited the Mayo PSC cohort with the specific objective of developing a clinically practical model that eliminated biopsy dependence. The revised model was derived from 405 PSC patients with definite or probable PSC followed prospectively at the Mayo Clinic between 1976 and 1997, with a total of 3,379 patient-years of follow-up. The composite endpoint was time to liver transplantation or death (whichever occurred first).

The analysis tested whether histological stage could be replaced or rendered redundant by other, non-invasive clinical and laboratory variables. Using time- dependent Cox regression, the study identified that serum AST and the presence of variceal bleeding, when combined with the non-histological variables from the original model (age, bilirubin, splenomegaly), provided prognostic information equivalent to or exceeding that of the original five- variable model including histological stage.

The resulting revised model uses five entirely non-invasive variables, generates a continuous risk score, and enables calculation of predicted survival at any time horizon using a Weibull baseline survival function. The model demonstrated superior or equivalent prognostic concordance (c-statistic) to the original model in validation analyses, confirming that liver biopsy-derived histological stage could be replaced without meaningful loss of prognostic accuracy.

The Five Prognostic Variables

1. Age

Patient age at the time of risk score calculation contributes a linear positive effect on risk. Older age at presentation is associated with shorter survival to transplantation or death, consistent with the well-established relationship between age and transplant-free survival across most chronic liver diseases. In PSC, this age effect likely reflects both the accumulated burden of co-morbidities in older patients, the reduced physiological reserve available to tolerate progressive cholestasis and portal hypertension, and the higher probability that older patients have been living with subclinical disease for longer before diagnosis.

Because PSC commonly presents in young adults, the age contribution to the risk score is often modest in the early disease phase but grows as a patient ages through sequential risk assessments over the years of clinical follow-up. Serial risk scores calculated annually or at clinically significant events will show a steady upward age contribution independent of disease activity.

2. Serum Bilirubin

Serum total bilirubin is the single most powerful prognostic variable in PSC, consistent across the original model, the revised model, and subsequent validation studies. It is entered into the revised model as a natural logarithm transformation, reflecting the non-linear, multiplicative nature of the bilirubin-survival relationship: doubling the bilirubin from 1 to 2 mg/dL carries different absolute prognostic weight than doubling from 10 to 20 mg/dL, but the log-linear model captures both appropriately on a continuous scale.

Bilirubin in PSC rises as a consequence of progressive biliary obstruction from diffuse stricturing, impaired biliary secretion from bile duct destruction, and secondarily from hepatocyte injury as cholestasis advances. Its elevation is therefore a marker of both biliary structural damage and hepatocellular reserve loss. When total bilirubin remains below approximately 2 mg/dL, the log-bilirubin term contributes a near-zero or negative value to the risk score (since ln(1) = 0 and ln(values below 1) is negative), indicating low risk from this variable.

Clinically, a rapidly rising bilirubin in a PSC patient should always prompt evaluation for a dominant biliary stricture (a superimposed focal high-grade narrowing that may be amenable to endoscopic or percutaneous dilatation) and for cholangiocarcinoma. Dominant strictures may cause abrupt bilirubin elevation disproportionate to the underlying diffuse disease and their detection and treatment can produce meaningful bilirubin reduction, which will be reflected in a lower revised Mayo risk score on reassessment.

3. Serum AST (Aspartate Aminotransferase)

Serum AST, entered into the model after normalisation to the upper limit of normal (ULN) of the reporting laboratory and log-transformation, serves as a non-invasive surrogate for the degree of hepatocellular injury and necroinflammatory activity that, in the original model, was partially captured by histological stage.

In PSC, hepatocellular injury arises from multiple mechanisms: bile acid-mediated toxicity to hepatocytes exposed to retained hydrophobic bile acids, ischaemia from peribiliary vascular changes, and direct immune-mediated hepatocyte damage. The degree of AST elevation therefore integrates information about the severity and activity of these processes. In PSC, AST is typically elevated to a lesser degree than in autoimmune hepatitis or acute hepatocellular injury, often 1.5 to 5 times the ULN; the revised model therefore uses normalised AST (AST/ULN) to adjust for inter-laboratory ULN variation, ensuring that the coefficient applies consistently across different clinical settings.

It is important to note that in the revised Mayo model, AST captures hepatocellular injury rather than biliary obstruction per se (which is more specifically reflected by alkaline phosphatase, GGT, and bilirubin). The inclusion of AST alongside bilirubin therefore adds independent prognostic information about the hepatocellular compartment that complements the biliary obstruction signal from bilirubin alone.

4. Variceal Bleeding

A history of oesophageal or gastric variceal bleeding is a binary variable (present or absent) and contributes a substantial positive increment to the risk score. Variceal bleeding in PSC indicates established portal hypertension severe enough to generate clinically significant varices, reflecting advanced hepatic fibrosis and cirrhosis that has progressed to a haemodynamically meaningful degree (hepatic venous pressure gradient above 12 mmHg for variceal formation, above 20 mmHg for haemorrhagic risk).

The strong prognostic weight of variceal bleeding in the revised model reflects its role as a threshold marker: once variceal haemorrhage has occurred, the patient has crossed a critical threshold of disease severity that carries a markedly different survival trajectory from those who have not reached this stage. In liver transplant organ allocation systems, a history of variceal bleeding combined with other markers of decompensation triggers specific listing priority adjustments beyond the MELD score alone.

In PSC patients without a prior variceal haemorrhage, the revised model still captures portal hypertension risk through the splenomegaly variable (below). The two variables together provide a two-tier representation of portal hypertension severity: splenomegaly (subclinical portal hypertension) and variceal bleeding (clinically significant portal hypertension that has crossed the haemorrhagic threshold).

5. Splenomegaly

Splenomegaly, defined as an enlarged spleen detectable by physical examination or imaging (typically a spleen longitudinal diameter exceeding 12 to 13 cm on ultrasound or cross-sectional imaging), is a binary variable that serves as a non-invasive marker of portal hypertension and advanced portal fibrosis or cirrhosis.

The spleen enlarges in portal hypertension through two mechanisms: passive congestion from elevated portal pressure and reticuloendothelial hyperplasia driven by the chronic inflammatory state. Its clinical correlates include thrombocytopenia (platelet sequestration), leucopenia, and anaemia from hypersplenism, though splenomegaly alone is not a sensitive marker of platelet count below any specific threshold.

In the context of the revised Mayo model, splenomegaly functions as a simple, reproducible, and clinically accessible indicator that a patient has developed significant portal hypertension. Its binary encoding in the model is an approximation of a continuous hepatic venous pressure gradient, but its reproducibility across clinical settings without additional blood tests or invasive procedures makes it a practical choice for a model designed for routine outpatient use.

The Risk Score Formula and Survival Estimation

The revised Mayo PSC risk score is calculated as a linear combination of the five variable terms with their regression coefficients:

R = (0.03 × age) + (0.54 × ln[bilirubin]) + (0.54 × ln[AST/ULN]) + (1.24 × variceal bleeding) + (0.84 × splenomegaly)

Where:

• Age is in years
• Bilirubin is the total serum bilirubin in mg/dL (or converted from μmol/L by dividing by 17.1)
• AST/ULN is serum AST divided by the institutional upper limit of normal for AST
• Variceal bleeding = 1 if present, 0 if absent
• Splenomegaly = 1 if present, 0 if absent
• ln denotes the natural logarithm

The resulting risk score R is a continuous value, typically ranging from approximately −2 to +8 in the derivation cohort. Higher values indicate higher risk of transplantation or death.

Survival Probability Estimation

Predicted survival probability at a given time t is calculated using the Weibull baseline survival function derived from the derivation cohort:

P(survival at t) = S₀(t)^exp(R − R̄)

Where S₀(t) is the baseline survival function at time t (the predicted survival for a patient with mean risk score R̄), exp(R − R̄) is the relative hazard compared to a patient with the mean risk score, and R̄ is the mean risk score from the derivation cohort.

Typical estimated survival probabilities from the original derivation cohort for representative risk score values:

These are approximate ranges for illustrative communication. The calculator generates patient-specific estimates using the exact baseline survival function parameters from the derivation cohort. Survival probabilities are most meaningful for communicating the magnitude of risk at a population level and for monitoring individual patient trajectory through serial measurements rather than as precise individual predictions.

Clinical Applications of the Revised Mayo PSC Risk Score

Timing of Liver Transplantation Evaluation

The primary clinical application of the revised Mayo PSC risk score is informing the optimal timing of referral for liver transplantation evaluation and listing. Liver transplantation is the only definitive treatment for end-stage PSC and produces excellent long-term outcomes, with 5-year post-transplant survival exceeding 85% in most series. However, transplantation carries perioperative risk, requires lifelong immunosuppression with its attendant complications, and the organ supply is finite. Optimal timing therefore requires balancing the risk of remaining on the waiting list against the risk of dying before receiving a transplant.

AASLD (American Association for the Study of Liver Diseases) and EASL (European Association for the Study of the Liver) guidelines historically recommended transplantation evaluation when the revised Mayo PSC risk score predicts a 2-year transplant-free survival below approximately 50%, corresponding roughly to a risk score above 4. However, in modern practice, most centres initiate transplantation evaluation earlier, particularly when MELD score is rising, when quality of life is significantly impaired by refractory pruritus or recurrent cholangitis, or when CCA or other complications emerge.

The PSC-specific MELD exception criteria in the United States UNOS allocation system provide additional MELD points to PSC patients with recurrent bacterial cholangitis or dominant strictures, acknowledging that the standard MELD may underestimate PSC-specific waiting list mortality. The revised Mayo PSC score complements MELD in this context by providing a disease-specific prognostic estimate that captures PSC pathophysiology more directly than the generic liver-failure metrics of MELD.

Serial Monitoring and Disease Progression Surveillance

Because the revised Mayo model uses entirely routine clinical and laboratory variables, it can be calculated at every clinical encounter. Serial risk scoring at 6 to 12-month intervals enables:

• Detection of disease acceleration: A rapid rise in the risk score over a short interval (e.g. increase of more than 1 point per year) signals accelerated disease progression and warrants prompt evaluation for a dominant stricture, CCA, or other complicating event driving the deterioration.
• Objective documentation of clinical trajectory: Serial risk scores provide a quantitative record that supplements subjective clinical assessment, useful for communicating progression to patients, subspecialty colleagues, and insurance or disability assessors.
• Timing for interventional procedures: While dominant strictures are managed based on direct cholangiographic findings, the overall risk score context informs the urgency and intensity of investigation when biochemistry deteriorates.
• Quality of life and symptom management planning: An escalating risk score in a patient with intractable pruritus or fatigue strengthens the case for early transplant listing rather than indefinite symptomatic management.

Clinical Trial Stratification and Endpoint Surrogacy

The revised Mayo PSC risk score has been extensively used as a stratification variable and efficacy endpoint in clinical trials testing potential disease-modifying therapies for PSC. Because no medical therapy has yet been convincingly shown to alter the natural history of PSC (the most extensively tested agents include ursodeoxycholic acid at various doses, norursodeoxycholic acid, fibrates, and multiple immunosuppressive agents), validated surrogate endpoints that can be measured within trial timelines are critical.

The risk score provides a continuous, interval-scale outcome measure that is sensitive to the biochemical changes (bilirubin, AST) that current trial endpoints target. Its use as a pre-specified secondary endpoint in PSC trials allows assessment of whether biochemical improvements translate into predicted survival benefit using validated natural history relationships.

Patient Counselling and Shared Decision-Making

Communicating prognosis in PSC is challenging because the disease trajectory is highly variable and the landmark events (transplantation, CCA development, decompensation) occur at unpredictable intervals. The revised Mayo model provides a structured framework for converting objective clinical and laboratory findings into time-framed survival probability estimates that can support evidence-based patient counselling.

When presenting risk score findings to patients, key principles include:

• Framing survival probabilities as population-level estimates rather than individual predictions: a 2-year survival estimate of 70% means that of 100 patients with this risk score, approximately 70 will be alive and transplant-free at 2 years, not that the individual patient has a 70% chance.
• Acknowledging that the estimates reflect a pre-2000 cohort that predates modern interventions (ERCP for dominant strictures, current antibiotic prophylaxis protocols, contemporary transplantation techniques) and likely underestimate survival with modern management.
• Contextualising the risk score alongside quality-of-life factors (pruritus, fatigue, IBD activity) that may justify transplantation consideration independently of the mortality risk estimate.
• Using serial scores to show patients objective evidence of stability or progression, supporting engagement with follow-up and treatment planning.

Comparison with MELD and Other Prognostic Tools

MELD Score

The Model for End-Stage Liver Disease (MELD) score, and its sodium-adjusted variant MELD-Na, is the primary organ allocation metric used by UNOS and Eurotransplant for liver transplant prioritisation across all liver disease aetiologies. MELD uses serum bilirubin, serum creatinine, and INR (or sodium for MELD-Na) to estimate 90-day waiting list mortality.

In PSC, MELD has a known limitation: it tends to underestimate disease-specific mortality risk compared with other chronic liver diseases of equivalent MELD score. This is because PSC patients often maintain relatively preserved synthetic function (reflected by normal or near-normal INR and creatinine) for prolonged periods despite advanced biliary disease, portal hypertension, and severely impaired quality of life. The cholestatic biology of PSC produces a disproportionate elevation of bilirubin relative to INR compared with hepatocellular diseases such as alcoholic hepatitis or acute liver failure.

The revised Mayo PSC risk score captures disease-specific risk factors (bilirubin, AST, variceal bleeding, splenomegaly, age) that reflect PSC pathophysiology more directly than MELD. It therefore provides complementary information that may better identify PSC patients at high risk of waiting list mortality or disease progression who are not adequately represented by their MELD score alone.

UK-PSC Score

The UK-PSC score, published by Goode et al. in 2021, was developed from a large multicentre UK cohort of 1,001 PSC patients to predict 5-year and 10-year liver transplantation-free survival. It uses alkaline phosphatase, albumin, bilirubin, platelet count, AST, and age, incorporating variables not in the revised Mayo model (albumin, alkaline phosphatase, platelets) that reflect nutritional status, biliary injury, and portal hypertension respectively.

The UK-PSC score demonstrated superior discrimination for long-term outcomes compared with the revised Mayo model in the UK derivation and validation cohorts, suggesting that additional variables beyond those in the revised Mayo model carry independent prognostic information. However, the UK-PSC score is less widely validated in non-UK populations, and the revised Mayo model retains its status as the primary internationally accepted standard due to its longer validation history.

PSC Risk Estimate Tool (PREsTo)

PREsTo is a dynamic prediction tool for PSC that models individual patient prognosis using longitudinal biomarker data rather than single-timepoint values. It uses repeat alkaline phosphatase measurements alongside baseline clinical variables to model disease trajectory, capturing the information content of biomarker trends rather than single readings. PREsTo has shown strong predictive performance in European PSC cohorts and is an emerging tool for research and specialised clinical applications. It is more computationally intensive than the revised Mayo model and has not yet been widely implemented in routine clinical calculators.

Cholangiocarcinoma Risk in PSC: A Critical Limitation of Natural History Models

All natural history models for PSC, including the revised Mayo model, were derived primarily from patients who followed the expected progressive cholestatic trajectory without developing CCA during observation. Patients who develop CCA during follow-up represent a distinct competing risk that is not captured by the standard survival model framework.

CCA occurring in a PSC patient who would otherwise have an excellent revised Mayo risk score can produce rapid, catastrophic deterioration that is entirely unpredictable from the model. Conversely, a patient with an elevated revised Mayo risk score who develops a dominant stricture that turns out to be CCA will have a fundamentally different prognosis than predicted by the natural history model alone.

This limitation means that the revised Mayo PSC risk score should always be interpreted alongside surveillance for biliary malignancy. Current EASL and AASLD guidelines recommend annual surveillance with serum CA 19-9, MRCP with gadolinium (or liver ultrasound in centres without MRCP access), and evaluation of any new dominant stricture with brush cytology and/or cholangioscopy for malignancy. The risk score captures cholestatic progression and portal hypertension; CCA surveillance addresses the superimposed malignancy risk that natural history models cannot predict.

Limitations of the Revised Mayo PSC Model

• Derivation cohort era and treatment context: The model was derived from patients followed between 1976 and 1997 at a single tertiary referral centre. Management practices for PSC complications (endoscopic treatment of dominant strictures, antibiotic prophylaxis, management of portal hypertension) have evolved substantially since then. Absolute survival estimates from the model likely underestimate current transplant-free survival with modern management. The relative risk stratification (higher score = worse prognosis) retains validity, but absolute survival probabilities should be interpreted with appropriate uncertainty.
• Single-centre derivation: All patients were managed at the Mayo Clinic, a tertiary referral centre with a case mix that may not represent the full spectrum of PSC severity encountered in community and general hepatology practice. Patients referred to tertiary centres may have more advanced or complicated disease than population-based cohorts.
• Cholangiocarcinoma as a competing risk: As discussed above, CCA is not incorporated into the survival model and represents a critical competing risk that can dominate individual patient outcomes independently of the cholestatic natural history.
• No IBD severity variable: Concurrent IBD, which is present in the majority of PSC patients, is not included in the revised model. Severe IBD requiring colectomy or complicated by colorectal dysplasia may modify transplant timing decisions and post-transplant outcomes. The model does not capture IBD disease activity as a prognostic modifier.
• AST normalisation variability: The AST/ULN ratio requires knowledge of the institutional ULN for AST. ULN values vary between laboratories and between patient populations (e.g., gender-specific ULNs). Inconsistent AST normalisation may produce different risk score values for the same absolute AST level depending on the ULN applied, introducing a source of inter-centre variability.
• Variceal bleeding documentation: The binary variceal bleeding variable requires reliable documentation of prior bleeding events. In patients with limited medical records or unreliable history, this variable may be misclassified. Additionally, patients who have undergone secondary prophylaxis (band ligation, TIPS, beta-blockers) may have different current risk than their prior bleeding history implies.
• Does not replace MELD for organ allocation: The revised Mayo PSC score is a prognostic tool and research instrument, not an organ allocation metric. Transplant listing and MELD exception applications are governed by local and national organ allocation protocols independent of the Mayo PSC score.
• Not applicable to small-duct PSC: Small-duct PSC has a substantially different and more favourable natural history than large-duct PSC. The revised model was derived in large-duct PSC patients and should not be applied to the small-duct variant.
• Not validated post-transplantation: The model predicts transplant-free survival; it has no validated role in predicting post-transplant outcomes or PSC recurrence after transplantation.

Dominant Strictures and Their Impact on Risk Scores

A dominant biliary stricture in PSC is defined by cholangiographic criteria as a stenosis with a diameter of 1.5 mm or less in the common bile duct or 1.0 mm or less in the hepatic duct. Dominant strictures occur in approximately 15 to 20% of PSC patients and produce disproportionate bilirubin elevation, worsening cholangitis, and pain that may temporarily elevate the revised Mayo risk score beyond what the underlying diffuse disease burden would predict.

Endoscopic management (balloon dilatation with or without short-term stenting) of dominant strictures can produce significant reductions in serum bilirubin and AST, with corresponding improvement in the revised Mayo risk score. This score reduction after dominant stricture treatment should be interpreted cautiously: it represents reversal of a superimposed obstructive component rather than improvement in the underlying diffuse PSC, and the underlying risk level may be better represented by pre-stricture baseline values.

Every dominant stricture must be evaluated for CCA with brush cytology, fluorescence in situ hybridisation (FISH), and if available, single-operator cholangioscopy with targeted biopsy before attributing bilirubin elevation to benign disease. A serum CA 19-9 above 130 U/mL in the context of a dominant stricture substantially raises the probability of CCA and warrants cross-sectional imaging review and multidisciplinary discussion before proceeding with endoscopic dilatation alone.

PSC and Liver Transplantation

Liver transplantation for PSC produces excellent long-term outcomes, with 5-year patient survival of 85 to 90% and 10-year survival of 70 to 80% in major transplant programmes. PSC accounts for approximately 5 to 10% of all adult liver transplants in Western countries. Key transplantation considerations specific to PSC include:

• Living donor liver transplantation (LDLT): PSC patients tend to have a relatively intact hepatic arterial supply and bile duct length, making them technically suitable candidates for LDLT in experienced centres. LDLT may allow more timely access to transplantation independent of deceased donor waiting times.
• Biliary reconstruction: PSC recipients typically undergo Roux-en-Y hepaticojejunostomy rather than duct-to-duct biliary anastomosis, due to the recipient’s diseased extrahepatic bile duct. This affects post-transplant biliary complication patterns and ERCP accessibility.
• IBD management post-transplantation: IBD may flare post- transplant, and calcineurin inhibitor-based immunosuppression provides limited protection against IBD activity. Ongoing colorectal surveillance is mandatory post-transplantation given the elevated colorectal carcinoma risk in PSC-IBD.
• PSC recurrence: PSC recurs in the allograft in approximately 20 to 25% of recipients at 5 to 10 years post-transplant. Recurrence is more common with living donors (particularly first-degree relatives), younger recipients, male sex, and possibly intact colon status (intact colon with active IBD being a risk factor). Recurrent PSC is typically managed conservatively; retransplantation may be required in severe cases.

How to Use This Calculator

The CalcMD Revised Mayo PSC Risk Score calculator requires five inputs from the patient’s current clinical assessment:

1. Age (years): Patient age in whole years at the time of risk score calculation. For serial scoring, use the current age at each assessment.
2. Serum total bilirubin (mg/dL): Total bilirubin from the most recent laboratory assessment, in mg/dL. If your laboratory reports in μmol/L, divide by 17.1 to convert. Use the pre-intervention value where possible; if the patient has recently undergone ERCP for dominant stricture treatment, consider whether the current bilirubin reflects the underlying disease status or an acute post-procedural change.
3. Serum AST (as a ratio to ULN): Enter the serum AST divided by the institutional upper limit of normal. For example, if AST is 75 U/L and your laboratory ULN is 40 U/L, enter 1.875. If the ULN is not available, consult the laboratory reference range on the report; ULN for AST is typically 35 to 40 U/L for men and 25 to 35 U/L for women, though institution-specific values should be used.
4. History of variceal bleeding (Yes/No): Indicate whether the patient has a documented history of oesophageal or gastric variceal haemorrhage. This includes acute bleeding episodes managed endoscopically, radiologically, or surgically. Bleeding from portal hypertensive gastropathy or gastric antral vascular ectasia (GAVE) should not be counted unless accompanied by confirmed variceal haemorrhage.
5. Splenomegaly (Yes/No): Indicate whether splenomegaly is present on physical examination (palpable spleen below the left costal margin) or on cross-sectional imaging (spleen longitudinal diameter above 12 to 13 cm on ultrasound, CT, or MRI). Use the most recent imaging report; intermittent splenomegaly should be classified as present if documented on any recent study.

The calculator outputs the continuous risk score R and the estimated transplant-free survival probabilities at 1, 2, and 4 years. These estimates are most useful as: (1) a current prognostic snapshot to contextualise the clinical discussion about transplantation timing; (2) a baseline for serial comparison at subsequent assessments; and (3) a communication aid for discussions with patients, caregivers, and multidisciplinary team members. They should not be used as the sole determinant of transplant listing decisions, which require multidisciplinary evaluation and consideration of MELD score, quality of life, IBD status, CCA surveillance findings, and institutional transplant programme criteria.