Neutrophil-to-Lymphocyte Ratio and Platelet-to-Lymphocyte Ratio Combined with MELD 3.0 as a Prognostic Predictor for Patients with Liver Cirrhosis Undergoing Transjugular Intrahepatic Portosystemic Shunt (TIPS) Creation

Monday, March 31, 2025

4:03 PM – 4:12 PM CT

Location: Music City Center, 204

Presenting Author(s)

TS

Takeshi Suzuki, MD, PhD

Research Fellow
Dotter Department of Interventional Radiology, Oregon Health & Science University, United States

Author/Co-author(s)

KM

Kenkichi Michimoto, MD, PhD (he/him/his)

Postdoctoral Scholar
Dotter Department of Interventional Radiology, Oregon Health & Science University, United States
Kentaro Yamada, MD, PhD (he/him/his)

Research Assistant Professor
Oregon Health & Science University, United States
LC

Leonardo Campos, MD

Interventional Radiologist
Oregon Health & Science University, United States
Kristey Nguyen, B.S

Medical Student
Oregon Health & Science University, United States
NL

Norah Languyen, None

Researcher
Dotter Department of Interventional Radiology, United States
James Palmer, MD

Resident Physician
Oregon Health and Science University, United States
BT

Brandon Tran, BS

Medical Student
Oregon Health and Science University, United States
HS

Harneet Sangha, BS

Medical Student
University of Washington, United States
Khashayar Farsad, MD, PhD (he/him/his)

Professor
Oregon Health and Science University, United States

Purpose:

To evaluate the use of Neutrophil-to-Lymphocyte Ratio (NLR) and Platelet-to-Lymphocyte Ratio (PLR) combined with MELD 3.0 as prognostic predictors for patients with cirrhosis undergoing TIPS creation.

Materials and Methods:

From September 2005 to December 2023, a total of 64 patients with cirrhosis (aged 25-80 years, mean age 55.0 years) who underwent TIPS creation and had available blood fractionation data within one week prior to the procedure were reviewed. Two logistic regression models were constructed to predict 30-day, 3-month, and 1-year mortality. The first model used an independent factor: MELD 3.0. The second model included three independent factors: MELD 3.0, NLR, and PLR. The predictive capability of both models was compared using receiver operating characteristic (ROC) curves to examine the area under the curve (AUC).

Results: The model that included the NLR and PLR showed greater accuracy in predicting 30-day mortality, with an AUC of 0.980 (95% confidence interval [CI] 0.946-1.014). In comparison, the model without NLR and PLR had an AUC of 0.857 (95% CI 0.737-0.977). For 3-month mortality, the AUC was 0.858 (95% CI 0.695-1.021) with NLR and PLR, compared to 0.785 (95% CI 0.645-0.925) without. For 1-year mortality, the AUC was 0.732 (95% CI 0.577-0.887) with NLR and PLR, while the model without had an AUC of 0.712 (95% CI 0.571-0.854). The bootstrap test suggested that the statistical differences between the AUCs for models with and without NLR and PLR were as follows: p = 0.0486 for 3-month mortality, p = 0.250 for 6-month mortality, and p = 0.951 for 1-year mortality.

Conclusion: Combining NLR and PLR with MELD 3.0 showed higher performance in predicting short-term mortality after TIPS creation compared to MELD 3.0 alone, suggesting biomarkers of inflammation may be additionally predictive for acute TIPS outcomes.