Prognostic Value of Stress Hyperglycemia Ratio, Glycemic Gap, and Hemoglobin Glycation Index for Mortality in ICU Patients with Pneumonia: A Retrospective Cohort Study Based on the MIMIC-IV Database
This retrospective cohort study of 2,740 ICU pneumonia patients using the MIMIC-IV database demonstrates that the stress hyperglycemia ratio (SHR) is a more robust and independent prognostic indicator for 28-day and 90-day mortality compared to the glycemic gap and hemoglobin glycation index (HGI).
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
When a person is critically ill, their body often reacts with a surge of stress that causes blood sugar levels to rise sharply. This phenomenon, known as stress hyperglycemia, is common in intensive care units and is frequently linked to worse outcomes, including death. However, doctors face a tricky problem when they look at a patient's blood sugar upon arrival. A high number could mean the patient is experiencing a severe, acute stress reaction, or it could simply reflect a long-term condition like diabetes where blood sugar is chronically elevated. Distinguishing between a temporary spike caused by the illness and a permanent baseline is difficult using a single blood test, yet this distinction is vital for understanding how sick a patient truly is and predicting their chances of survival.
To solve this puzzle, researchers have developed ways to compare a patient's current blood sugar against what their sugar levels usually are. They use a standard blood test called hemoglobin A1c, which acts like a historical record of blood sugar control over the past few months, to estimate what a patient's normal sugar level should be. By comparing the immediate, stress-induced number against this historical average, doctors can calculate a ratio that reveals the true intensity of the stress response. This approach helps separate the noise of chronic disease from the signal of acute crisis, offering a clearer picture of the patient's immediate physiological state.
A new study published in 2026 takes this concept and applies it to a specific, high-risk group: adults admitted to intensive care units with pneumonia. The researchers, working with a massive database of medical records from a major hospital in the United States, set out to compare three different methods of measuring this stress response. They wanted to know which of these mathematical tools was the most reliable at predicting who would survive and who would not. The study focused on 2,740 adult patients, tracking their outcomes over 28 days and again at 90 days to see if these sugar-based metrics could accurately forecast mortality.
The team calculated three distinct values for every patient. The first was a ratio that compared the admission blood sugar to the estimated average sugar derived from the A1c test. The second was the simple difference between the two numbers, showing the absolute gap. The third was a measure of how much a patient's blood sugar history deviated from what would be expected for their current level. The researchers then ran these numbers through complex statistical models that accounted for age, sex, other diseases, and the severity of the pneumonia and sepsis. They were looking for a consistent link between these sugar metrics and the likelihood of death.
The results pointed clearly to one method as the superior predictor. The ratio that compared current sugar to the historical average proved to be a robust indicator of mortality risk. Patients with a higher ratio, meaning their blood sugar was significantly elevated above their personal baseline, faced a substantially higher risk of dying within 28 and 90 days. This relationship held true even after the researchers adjusted for every other known factor that could influence survival, such as kidney failure or heart disease. The study found that for every standard increase in this ratio, the risk of death rose significantly, suggesting that the degree of stress hyperglycemia is a powerful signal of how severely the body is struggling.
In contrast, the other two methods performed less effectively. The simple difference between the current sugar and the average, while showing some statistical association, offered very little practical value when looking at small changes in the numbers. Its effect was so small that it was difficult to use for predicting individual outcomes. The third method, which measured the deviation in sugar history, initially appeared to be linked to survival, but this connection disappeared entirely once the researchers accounted for the severity of the patient's other illnesses. This suggested that the third method was not measuring the stress of the pneumonia itself, but rather reflecting the general burden of disease and other complications the patient was already facing.
The study also explored whether this relationship changed depending on the type of patient. The researchers found that the predictive power of the sugar ratio was strongest in patients who also had sepsis, a life-threatening reaction to infection. In patients with pneumonia but without sepsis, the link was not statistically significant, though the researchers noted that this group was smaller and might have required more data to draw a firm conclusion. The analysis showed that the risk increased steadily as the ratio went up, without any sudden jumps or drops, indicating a linear relationship where higher stress levels consistently meant higher danger.
These findings offer a practical tool for the intensive care unit. The most effective metric requires only two tests that are already standard in almost every hospital: a blood sugar test taken when the patient arrives and a hemoglobin A1c test. Because it relies on existing data, it costs nothing extra and provides immediate insight. The study concludes that for doctors treating pneumonia in the ICU, this simple ratio is a more reliable way to assess risk than looking at blood sugar levels alone or using the other calculated differences. It highlights that the body's reaction to acute stress, when viewed against its own baseline, is a critical factor in determining the outcome for these critically ill patients.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.