Association between monocyte-to-lymphocyte ratio and mortality in patients with acute pancreatitis requiring intensive care unit admission: a retrospective cohort study and predictive model establishment based on machine learning

Why a Simple Blood Test Matters in the ICU

When people develop sudden, severe inflammation of the pancreas—known as acute pancreatitis—they often arrive at the hospital in great pain, and some become critically ill. Doctors in intensive care units need quick, reliable ways to tell which patients are likely to recover and which are at higher risk of dying, so they can tailor monitoring and treatment. This study explores whether a very simple measure from an everyday blood test—the balance between two types of white blood cells—can help forecast both short‑term and long‑term survival in these sick patients.

A Closer Look at a Dangerous Sudden Illness

Acute pancreatitis is an abrupt inflammation of the pancreas that can damage many organs. Most people survive, but when the disease is severe, deaths can approach one in three cases. There is no single cure; instead, treatment focuses on supporting the body while the inflammation settles. Because the illness can worsen quickly, doctors rely on scores and lab values to judge how sick a patient is. However, these tools can be complex or slow to update. The authors turned to a large U.S. intensive‑care database to see whether the monocyte‑to‑lymphocyte ratio—calculated from routine blood counts—could serve as a fast, inexpensive gauge of risk.

What the Blood Cell Balance Reveals

Monocytes and lymphocytes are two kinds of white blood cells that reflect different sides of the body’s defense system: monocytes signal broad inflammation, while lymphocytes are central players in immune protection. By dividing the number of monocytes by the number of lymphocytes, doctors obtain the monocyte‑to‑lymphocyte ratio (MLR). The researchers identified 1,044 adults admitted to intensive care with acute pancreatitis and grouped them into four bands from lowest to highest MLR. They then examined how often patients in each band died in the hospital, within 28 days, and within one year, while also looking at many other clinical features such as age, vital signs, organ function tests, and existing illnesses.

A U‑Shaped Link Between Ratio and Risk

The team found that MLR was strongly tied to survival. Patients with very high ratios had the worst outcomes: they were more likely to die in the ICU, had longer stays, and showed higher death rates at 28 days and at one year. Interestingly, people with extremely low ratios also faced elevated risk over the longer term. When the ratio was graphed against the chance of death, the curve formed a U‑shape, with the lowest risk around an intermediate value. This pattern held even after accounting for age, heart and kidney disease, infections, and many other factors, suggesting that both too much inflammation and too little immune capacity can be harmful in this condition.

Teaching Computers to Spot High‑Risk Patients

To see how well MLR could contribute to prediction tools, the researchers built several computer‑based models using machine‑learning methods. They trained these models on part of the patient data and then tested them on the remaining records. Across models, a traditional survival model and a type of random forest performed particularly well in estimating who would die within 28 days. When the importance of each input was examined, MLR repeatedly emerged among the key factors, alongside age, overall illness severity scores, oxygen levels, blood clotting times, and markers of liver function. This means that modern analytical tools, which can capture complex patterns, consistently recognized the blood cell ratio as a meaningful signal.

What This Means for Patients and Doctors

For everyday care, these findings highlight that a ratio already hidden in standard blood work can help doctors rapidly sort which intensive‑care patients with acute pancreatitis are in greater danger, both in the coming weeks and over the next year. Although the study is observational and based on data from a single hospital network, it suggests that keeping the monocyte‑to‑lymphocyte ratio within a moderate range reflects a healthier balance between inflammation and immune defense. With further validation, this simple measure could be folded into bedside checklists or computer alerts, helping teams focus attention and resources on those most at risk when time matters most.

Citation: Yang, J., Dong, C., Guo, M. et al. Association between monocyte-to-lymphocyte ratio and mortality in patients with acute pancreatitis requiring intensive care unit admission: a retrospective cohort study and predictive model establishment based on machine learning. Sci Rep 16, 9157 (2026). https://doi.org/10.1038/s41598-026-37791-6

Keywords: acute pancreatitis, intensive care, blood biomarkers, immune balance, machine learning prediction