Development and external evaluation of a routine laboratory-based three-ratio score for in- hospital mortality in older critically ill patients with atherosclerotic cardiovascular disease: a MIMIC-IV and eICU study

When an older adult with heart disease arrives at an intensive care unit, their body often enters a state of high alert. The heart may struggle to pump blood effectively, causing the kidneys to slow down their work, the blood's acid-base balance to shift, and blood sugar levels to rise in response to stress. Doctors have long known that these changes are dangerous, but they are also complex. A single blood test, such as one measuring kidney function or sugar levels, often tells only a small part of the story. It is like trying to understand the health of a car engine by looking at just the oil pressure while ignoring the temperature gauge and the fuel mixture. To truly grasp the severity of a patient's condition, clinicians need a way to see how these different systems are failing together.

Researchers set out to create a simple tool that could capture this combined picture using only four routine blood tests that are already taken from almost every patient in the intensive care unit. They focused on older patients with atherosclerotic cardiovascular disease, a condition where the arteries become hardened and narrowed, which is a leading cause of death and disability. The team developed a new scoring method that blends measurements of nitrogen waste in the blood, kidney function, acid-base balance, and blood sugar. By looking at how these four values relate to one another, rather than just their individual numbers, the researchers hoped to find a clearer signal of who was at the highest risk of dying during their hospital stay.

The study analyzed data from two massive, de-identified databases containing records from thousands of intensive care patients across the United States. The first group included over 8,000 patients from a single major hospital, while the second group included nearly 10,000 patients from 190 different hospitals. All the patients were at least 65 years old, had been in the intensive care unit for at least 24 hours, and had the necessary blood test results available. The researchers calculated their new score for every patient and then tracked who survived and who did not. They found a clear and consistent pattern: patients with higher scores were significantly more likely to die in the hospital. The relationship was not a simple jump from safe to dangerous; instead, the risk of death rose steadily as the score increased, creating a continuous gradient of danger.

To see if this new score offered anything beyond what doctors already know, the researchers compared it against standard clinical information, such as a patient's age, sex, history of heart failure, and whether they were on a breathing machine or receiving drugs to support blood pressure. When they added their new score to this standard clinical picture, the ability to predict who would die improved noticeably. In the first group of patients, the prediction accuracy rose from a moderate level to a strong one. In the second, larger group of patients from many different hospitals, the improvement was just as clear. This suggests that the new score captures something vital that the standard clinical notes alone miss. It acts as a summary of the body's metabolic stress, combining the burden of waste products, the strain on the kidneys, and the disruption of acid and sugar balance into a single number.

The researchers also tested whether using the four raw blood numbers separately would work better than their combined score. They found that using the four numbers individually provided a tiny bit more information, but the difference was very small. This is an important finding because it means the simplified score is nearly as powerful as the full set of data, making it much easier to use in a busy hospital setting. The score does not require complex calculations or special equipment; it can be derived from standard lab results that are already available. However, the study also noted that while the score's ability to rank patients by risk was consistent across different hospitals, the exact probability of death associated with a specific score varied slightly between the two groups. This indicates that while the score is excellent at identifying who is at higher risk relative to others, the specific numbers used to predict the exact chance of death may need to be adjusted for different hospital populations.

The study concludes that this new method is a promising way to assess risk in older, critically ill patients with heart disease. It successfully combines information about kidney function, acid-base balance, and glucose metabolism into a single, easy-to-calculate metric. The findings show that a higher score, reflecting a greater imbalance in these systems, is strongly linked to a higher chance of in-hospital death. While the score is not yet ready to be used as a standalone rule for treatment decisions, it provides doctors with a valuable new lens through which to view a patient's condition. It highlights the importance of looking at the body as an interconnected system, where a problem in one area, like the kidneys, often echoes through others, such as the balance of acids and sugars in the blood. Before this tool can be widely adopted, further studies will be needed to fine-tune the numbers for different settings and to see if using the score actually helps doctors save more lives.