Ana Marenco Flores, MD, is a research fellow in medicine in the division of gastroenterology at Beth Israel Deaconess Medical Center and Harvard Medical School.

Ana Marenco Flores, MD

Validating the Amsterdam-Oxford Model for PSC Prognosis, with Ana Marenco Flores, MD

New research is confirming the validity of the Amsterdam-Oxford Model for predicting long-term 

Study findings were presented at Digestive Disease Week (DDW) 2025 by Ana Marenco Flores, MD, a research fellow in medicine in the division of gastroenterology at Beth Israel Deaconess Medical Center and Harvard Medical School, and demonstrate the Amsterdam-Oxford Model’s ability to successfully predict adverse outcomes in a US cohort of PSC patients, supporting its generalizability across diverse populations and clinical settings.

The Amsterdam-Oxford Model, which was developed from Dutch and UK cohorts, incorporates 7 clinical variables to predict long-term PSC outcomes: PSC subtype, age at diagnosis, aspartate aminotransferase, alkaline phosphatase, total bilirubin, albumin, and platelet count.

“[This is] a tool that can help you and the patient see what the possibilities of their outcomes are,” Marenco Flores explained to 

. “You as a clinician may be able to save time and resources by doing an equation with some labs or questions and see if you have to dedicate more time to this patient… that can give you a perspective of the follow ups that you have to do with the patient or if you need to refer them to another specialist. I think doing an assessment like this in your first visit is really important to see every patient as a single case, because not all PSC patients behave the same way.”

While the model has demonstrated strong predictive accuracy, Flores and colleagues recognized the need for external validation to assess its generalizability. To do this, they analyzed data from patients with PSC enrolled in an autoimmune prospective registry between 2018 and 2024. For inclusion, patients were required to ≥ 18 years of age with confirmed PSC, complete datasets for calculating both risk scores, and ≥ 4 years of follow-up.

Predictive performance was assessed using the area under the receiver operating characteristic curve. Investigators conducted Kaplan-Meier survival analysis, stratified by Mayo Risk Score classification and corresponding Amsterdam-Oxford Model values, to compare clinical outcomes, with survival defined as time to liver transplantation or death.

In total, the study included 109 patients with PSC. Their mean age was 45 ± 15 years, and the majority of patients (51%) were female. Among the cohort, 85% of patients remained alive without liver transplantation (LT), 12% underwent LT, and 3% died after a median follow-up of 4.63 years.

The median Amsterdam-Oxford Model score was 1.48 (interquartile range [IQR], 1.04 to 1.83) and the median Mayo Risk Score was -0.5 (IQR, -1.03 to -0.02). Logistic regression identified the Amsterdam-Oxford Model (odds ratio [OR], 4.95; 

 = .027) as independent predictors of LT or death, and multivariate analysis confirmed the Amsterdam-Oxford Model’s significance (OR, 4.96; 

Further survival analysis demonstrated worse outcomes in patients classified as high risk by both the Mayo Risk Score and Amsterdam-Oxford Model, with a 2-year survival rate of 66% compared to 92% for intermediate-risk and 95% for low-risk patients.

“ I think for us at every clinic or hospital, your resources and time are really valuable,” Marenco Flores said. “By applying a score with data that you have at your everyday assessment of the patient, you can get a sense of personalized medicine, and that's something that we want to achieve in the future.”

Editors’ note: Marenco Flores has no relevant disclosures.

Marenco-Flores A, Amaris NR, Kahan TF, et al. EXTERNAL VALIDATION OF THE AMSTERDAM-OXFORD MODEL FOR PREDICTING LIVER TRANSPLANTATION AND MORTALITY IN U.S. PATIENTS WITH PRIMARY SCLEROSING CHOLANGITIS: A COMPARATIVE ANALYSIS WITH THE MAYO RISK SCORE. Abstract presented at Digestive Disease Week 2025 in San Diego, CA, from May 3 - May 6, 2025.

de Vries EM, Wang J, Williamson KD, et al. A novel prognostic model for transplant-free survival in primary sclerosing cholangitis. 

Videos

The Amsterdam-Oxford Model successfully predicted adverse outcomes in a US cohort of PSC patients, validating its generalizability.

Sleep Disorders and PBC Disease Severity, with Ana Marenco Flores, MD

New research is shedding light on the prevalence of sleep disorders in patients with 

, additionally highlighting their association with worse biochemical markers, greater symptom burden, and reduced treatment response rates.

The data were presented at Digestive Disease Week (DDW) 2025 by Ana Marenco Flores, MD, a research fellow in medicine in the division of gastroenterology at Beth Israel Deaconess Medical Center and Harvard Medical School, and highlight the potential benefit of assessing sleep disorders as part of routine clinical care and implementing timely interventions for improving outcomes and quality of life in this patient population.

“In clinical practice, almost 80 to 85% of patients are going to tell you they are having fatigue, having trouble sleeping, having a lot of itching, or having difficulties at their job because they’re feeling tired a lot,” Marenco Flores explained to 

. “There’s a really strong association between quality of life or patient reported outcomes with PBC.”

To assess this association in the context of sleep disorders, she and colleagues conducted a prospective study of patients with PBC enrolled in a single-center Autoimmune Liver Registry between 2018-2024. Diagnoses were based on EASL criteria and all patients received standard ursodeoxycholic acid therapy with ≥ 1 year of follow-up. Exclusion criteria included autoimmune overlap, drug-induced liver injury, viral hepatitis, alcohol-associated liver disease, and incomplete questionnaires.

Sleep disorders and fatigue were assessed using the CLDQ and PBC-40 questionnaires, while pruritus was evaluated using the Visual Analogue Scale (VAS) and the 5-D Itch Scale at the 1-year follow-up. Treatment response was defined according to Paris-II criteria.

Among 104 patients with PBC, 33% reported sleep disorders. Investigators noted this group had significantly increased alkaline phosphatase (ALP) levels (median 203 IU/L vs 122 IU/L; 

= .003) and total bilirubin (TBIL) levels (median 0.6 mg/dL vs 0.4 mg/dL; 

= .031) compared to those without sleep disorders. They also reported more severe fatigue (

<.001), and a significantly lower likelihood of achieving treatment response (44% vs 94%; 

<.001). Of note, pruritus severity did not differ significantly between the groups (

“Before, there were studies that said that most patients have difficulty sleeping because they have itch. In our cohort, we saw that even though some patients have itching and sleep issues, a lot of them didn't report itching but still had sleep difficulties,” Marenco Flores said.

Multivariable analysis identified greater ALP (odds ratio [OR], 1.01; 

<.001) and poorer self-rated health (OR, 3.4; 

<.001) as independent predictors of sleep disorders.

“I think for us as a lab, especially in autoimmune liver diseases, we aim to also assess patient reported outcomes,” Marenco Flores said. “The objective parameters are obviously really important, but sometimes patients will do great with the labs and the treatment and still have some complaints. For them, feeling good in every aspect of their life is as good as having a great ALP or another imaging test, so that's what we aim for.”

Marenco-Flores A, Amaris NR, Barba R, et al. SLEEP DISORDERS IN PRIMARY BILIARY CHOLANGITIS: INDEPENDENT OF PRURITUS BUT PREDICTIVE OF DISEASE SEVERITY IN A U.S. SINGLE-CENTER COHORT. Abstract presented at Digestive Disease Week 2025 in San Diego, CA, from May 3 - May 6, 2025.

Findings suggest sleep disorders are common in PBC and are linked to worse biochemical markers, greater symptom burden, and reduced treatment response rates.

Ana Marenco Flores, MD

Articles