📄 medicine

Association Between the Triglyceride–Glucose Index and Short-Term Mortality in Critically Ill Patients: A Retrospective Cohort Study Using the MIMIC-IV Database

This retrospective cohort study of 6,342 critically ill patients using the MIMIC-IV database reveals that while an elevated Triglyceride–Glucose (TyG) index is significantly associated with increased 28-day mortality, its independent prognostic value is modest and attenuated when adjusted for organ dysfunction scores like SOFA, suggesting it may serve best as a complementary metabolic biomarker rather than a standalone predictor.

Original authors: Xiaoqing Gan, Jianying Dai, Chao Ding, Qingjiang Lyu

Published 2026-06-29

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Xiaoqing Gan, Jianying Dai, Chao Ding, Qingjiang Lyu

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

Imagine the Intensive Care Unit (ICU) as a high-stakes emergency room where patients are fighting for their lives against severe illness. In this chaotic environment, doctors need quick, reliable clues to guess who might survive and who might not.

This study is like a detective story where the researchers looked at a massive digital archive of patient records (called MIMIC-IV) to find a hidden clue. That clue is something called the TyG Index.

What is the TyG Index?

Think of the TyG Index as a "Metabolic Stress Score." It's a simple math formula that combines two numbers doctors already check on almost every patient:

Blood Sugar (Glucose)
Blood Fat (Triglycerides)

Normally, these numbers tell us about diabetes or diet. But the researchers wondered: In a critically ill patient, do these numbers act like a "canary in a coal mine," warning us about how hard the body is struggling?

The Main Discovery: The Higher the Score, The Riskier the Ride

The researchers looked at over 6,300 patients. They found a clear pattern:

The "Low Score" Group: Patients with lower TyG numbers had a better chance of surviving the first 28 days.
The "High Score" Group: Patients with the highest TyG numbers were significantly more likely to die within 28 days.

It's like a speedometer on a car. If the needle is in the green zone, the engine is running okay. If the needle is screaming in the red zone (a high TyG score), the engine is overheating, and the car is more likely to break down completely.

The "U-Shape" Surprise

Here is where it gets interesting. The relationship wasn't just a straight line going up. The researchers found a U-shape.

Too High: Bad. The body is overwhelmed by sugar and fat, causing inflammation and stress.
Too Low: Also bad. If the score is extremely low, it might mean the patient is so sick they are starving or their liver has stopped working properly.
The Sweet Spot: There is a "Goldilocks zone" in the middle (around a score of 7.92) where the risk of death is lowest.

The Big Comparison: The TyG Index vs. The "Master Score"

This is the most important part of the story. In the ICU, doctors already use a famous tool called the SOFA Score.

The SOFA Score is like a comprehensive health report card. It checks six different organs (heart, lungs, kidneys, brain, etc.) to see how badly they are failing. It's heavy, detailed, and very accurate.
The TyG Index is like a quick snapshot of just the body's fuel system.

The researchers asked: Is the TyG Index better than the SOFA Score?
The answer was no.

The SOFA Score was a much better predictor of who would die (like a master detective).
When they tried to use the TyG Index alongside the SOFA Score, the TyG Index didn't add much new information. It was like trying to add a single piece of a puzzle when you already have the whole picture.

Who Does This Clue Help Most?

The study found that the TyG Index was a stronger warning signal for specific groups:

Older patients: The "fuel system" warning was more critical for seniors.
Non-diabetic patients: For people who didn't already have diabetes, a sudden spike in this score was a huge red flag.
Non-septic patients: For patients who weren't fighting a massive infection, this metabolic clue was more useful.

For patients who already had diabetes or severe infections, the TyG Index wasn't as helpful because their bodies were already in such a chaotic state that the score didn't tell them anything new.

The Bottom Line

Think of the TyG Index as a cheap, easy-to-get "check engine" light.

It's simple to calculate (just two blood tests).
It's free.
It does tell you something important: if the light is on (high score), the patient is at higher risk.

However, it is not a replacement for the full engine diagnostic (the SOFA Score). It's a helpful extra tool, especially if you don't have time to run the full diagnostic, or if you want a quick second opinion on how the patient's metabolism is handling the stress of critical illness.

In short: A high TyG score in the ICU is a warning sign that the patient's body is under severe metabolic stress, which increases the risk of death. But doctors should still rely on the full organ-check (SOFA) for the most accurate prediction.

Technical Summary: Association Between the Triglyceride–Glucose Index and Short-Term Mortality in Critically Ill Patients

Problem Statement
Insulin resistance (IR) and metabolic dysregulation, including hyperglycemia and dyslipidemia, are prevalent in critically ill patients and are linked to cardiovascular dysfunction and increased mortality. While the Triglyceride–Glucose (TyG) index has been established as a reliable, cost-effective surrogate marker for IR in general and cardiovascular populations, its prognostic value in unselected, heterogeneous intensive care unit (ICU) cohorts remains poorly characterized. Existing critical care literature often focuses on specific subgroups (e.g., sepsis or acute kidney injury) or fails to contextualize the TyG index against established severity-of-illness scores like the Sequential Organ Failure Assessment (SOFA) score. This study addresses the gap in understanding the relationship between the TyG index and short-term mortality in a broad ICU population and evaluates its incremental prognostic value relative to organ dysfunction metrics.

Methodology
This retrospective cohort study utilized the MIMIC-IV (version 3.1) database, encompassing data from the Beth Israel Deaconess Medical Center (2008–2022).

Study Population: 6,342 adult patients (≥18 years) admitted to the ICU for the first time, with available serum triglyceride and glucose measurements within 24 hours of admission. Patients with missing demographics or those discharged/dying within the first 24 hours were excluded.
Exposure Variable: The TyG index was calculated as $\ln[\text{fasting triglycerides (mg/dL)} \times \text{fasting glucose (mg/dL)} / 2]$ . Values were treated as both continuous variables and quartiles (Q1–Q4).
Outcomes: The primary outcome was 28-day all-cause mortality. Secondary outcomes included hospital mortality and ICU mortality.
Statistical Analysis: Multivariable logistic regression models adjusted for age, sex, diabetes, hypertension, heart failure, chronic kidney disease (CKD), COPD, liver disease, cancer, sepsis, and the Charlson Comorbidity Index (CCI). The study employed restricted cubic spline (RCS) analysis to assess non-linear relationships, receiver operating characteristic (ROC) curves for discriminatory performance, and comparative analyses against the SOFA score. Subgroup analyses were conducted based on age, sex, diabetes status, sepsis, and comorbidities.

Key Results

Baseline Characteristics: The median TyG index was 8.88. Higher TyG quartiles were associated with younger age, higher male prevalence, and increased rates of diabetes, liver disease, CKD, and sepsis.
Mortality Association: There was a significant positive association between higher TyG values and mortality. The 28-day mortality rate increased from 13.4% in the lowest quartile (Q1) to 20.4% in the highest quartile (Q4).
Regression Analysis: After multivariable adjustment, each 1-unit increase in the TyG index was associated with a 26.6% increase in the odds of 28-day mortality (Adjusted OR 1.266, 95% CI 1.157–1.385). Similar significant associations were observed for hospital and ICU mortality.
Non-Linear Relationship: RCS analysis revealed a U-shaped association, where mortality risk decreased to a nadir around a TyG index of 7.92 before rising progressively with higher values.
Discriminatory Performance: As a standalone predictor, the TyG index demonstrated modest discriminatory ability for 28-day mortality (AUC = 0.554), outperforming albumin but underperforming compared to RDW and glucose.
Comparison with SOFA Score: The SOFA score alone showed superior discriminatory performance (AUC = 0.732). When both TyG and SOFA were included in a multivariable model, the SOFA score remained a strong independent predictor, while the TyG index lost statistical significance (aOR 1.049, p = 0.338).
Subgroup Findings: The association between TyG and mortality was significantly stronger in patients aged ≥65 years and in non-diabetic, non-septic individuals. No significant association was found in diabetic or septic subgroups.

Key Contributions and Significance
The study provides a large-scale evaluation of the TyG index in a heterogeneous ICU population, confirming that elevated TyG values are associated with increased short-term mortality. However, the authors emphasize a crucial nuance regarding the clinical utility of the index:

Contextualization of Prognostic Value: The paper explicitly states that the TyG index does not provide substantial prognostic information beyond what is captured by the SOFA score. The loss of independent significance when adjusted for SOFA suggests that the TyG-mortality association is partially mediated by organ dysfunction rather than representing a purely independent metabolic risk factor.
Complementary Role: The authors conclude that the TyG index should not be viewed as a replacement for established severity scores or a standalone biomarker. Instead, its significance lies in its role as a complementary metabolic biomarker. Due to its simplicity, low cost, and reliance on routine laboratory panels, it may serve as a useful adjunctive tool for cardiovascular risk assessment, particularly in resource-limited settings where comprehensive SOFA calculation is not immediately feasible, or for tracking metabolic trajectories during the ICU course.
Specific Populations: The findings suggest the TyG index may offer the most incremental prognostic value in non-diabetic and non-septic patients, where metabolic risk stratification might otherwise be less defined by acute infection or chronic diabetic complications.

The study concludes that while metabolic dysregulation is a relevant factor in critical illness outcomes, the TyG index's utility is best realized as a supplementary indicator rather than a primary predictor of mortality in critically ill patients.