## Talks

## Academic Level (Author 1)

Resident

## Discipline/Specialty (Author 1)

Internal Medicine

## Academic Level (Author 2)

Faculty

## Discipline/Specialty (Author 2)

Internal Medicine

## Academic Level (Author 3)

Staff

## Discipline/Specialty (Author 3)

Internal Medicine

## Academic Level (Author 4)

Staff

## Discipline/Specialty (Author 4)

Internal Medicine

## Academic Level (Author 5)

Faculty

## Discipline/Specialty (Author 5)

Internal Medicine

## Academic Level (Author 6)

Faculty

## Discipline/Specialty (Author 6)

Internal Medicine

## Discipline Track

Clinical Science

## Abstract

**Introduction: **Measurement of serum lactate remains pivotal in sepsis evaluation as it is closely related to tissue hypoxia and mortality.^{1, 2} Serum lactate, however, can be elevated due to a myriad of reasons beyond hypoperfusion.^{3} A proposed alternative to control for serum lactate’s inherent variability is to measure pyruvate concomitantly to calculate the lactate/pyruvate (L/P) ratio.^{4, 5} As pyruvate transforms into lactate in anaerobic conditions, its value decreases in serum, increasing the L/P ratio, hence indicating an actual tissue-perfusion mismatch.^{6} Despite its clinical utility, there is a paucity of evidence evaluating the role of the L/P ratio in patients with sepsis without evidence of shock.

**Methods: **Using an observational design, we evaluated the clinical utility of the L/P ratio in patients with sepsis. We included patients with sepsis, regardless of the etiology, that were hemodynamically stable. Lactate and pyruvate, and consequently the L/P ratio, were measured at baseline, 4 hours, and 8 hours. Our primary outcome was to determine the prognostic utility of the L/P ratio on mortality across time. Our secondary outcome was to compare the L/P ratio prognostic utility when adding standard risk calculators such as APACHE-II and SOFA scores. Model performance and accuracy were determined using the residual mean square error (RMSE), R^{2}, C-statistic, accuracy, and a visual inspection of the receiver operating characteristic (ROC) curve. A non-parametric bootstrap regression was used for all models, given that we expect a non-linear relationship between variables and to allow for more model flexibility.

**Results: **We had a total of 80 patients, with 18 (22.5%) survivors and 62 (77.5%) non-survivors. Pneumonia was the most common source of infection (44/80; 55%) and gastrointestinal the least common (15/80; 19%) (**Table 1**). There were no differences at baseline except for the SOFA and APACHE-II scores (**Table 1**). For our primary outcome, the prognostic model, patients with a higher L/P ratio at 8 hours had an OR of 1.08 (95% CI 1.02 – 1.18), holding constant the rest of the variables and had better performance than the other two models (**Figure 1**). For our secondary outcome, we found that the prognostic model using the APACHE-II in addition to the L/P ratio and the L/P ratio alone had a very similar association to the mortality (OR 1.07, 95% CI 1.01 – 1.17 and OR 1.08, 95% CI 1.02 – 1.18) (**Figure 2**). However, there was a substantial difference in the model’s performance between all models (**Figure 2**).

**Conclusions: **While the L/P ratio appears to be an earlier predictor of mortality than hyperlactatemia alone, its comparability against other standard clinical risk scores is negligible or suboptimal. The L/P ratio’s clinical applicability seems to be more towards daily use to avoid misinterpretation of isolated hyperlactatemia.

## Presentation Type

Talk

## Recommended Citation

Dorsey-Trevino, Edgar G.; Cantu-Rodriguez, Olga G.; Hawing-Zarate, Jose A.; Gonzalez-Cantu, Grecia A.; Gomez-Almaguer, David; and Gutierrez-Aguirre, Cesar H., "Prognostic Utility of the Lactate/Pyruvate Ratio for Mortality in Patients with Sepsis: An Observational Study" (2024). *Research Colloquium*. 5.

https://scholarworks.utrgv.edu/colloquium/2024/talks/5

Prognostic Utility of the Lactate/Pyruvate Ratio for Mortality in Patients with Sepsis: An Observational Study

**Introduction: **Measurement of serum lactate remains pivotal in sepsis evaluation as it is closely related to tissue hypoxia and mortality.^{1, 2} Serum lactate, however, can be elevated due to a myriad of reasons beyond hypoperfusion.^{3} A proposed alternative to control for serum lactate’s inherent variability is to measure pyruvate concomitantly to calculate the lactate/pyruvate (L/P) ratio.^{4, 5} As pyruvate transforms into lactate in anaerobic conditions, its value decreases in serum, increasing the L/P ratio, hence indicating an actual tissue-perfusion mismatch.^{6} Despite its clinical utility, there is a paucity of evidence evaluating the role of the L/P ratio in patients with sepsis without evidence of shock.

**Methods: **Using an observational design, we evaluated the clinical utility of the L/P ratio in patients with sepsis. We included patients with sepsis, regardless of the etiology, that were hemodynamically stable. Lactate and pyruvate, and consequently the L/P ratio, were measured at baseline, 4 hours, and 8 hours. Our primary outcome was to determine the prognostic utility of the L/P ratio on mortality across time. Our secondary outcome was to compare the L/P ratio prognostic utility when adding standard risk calculators such as APACHE-II and SOFA scores. Model performance and accuracy were determined using the residual mean square error (RMSE), R^{2}, C-statistic, accuracy, and a visual inspection of the receiver operating characteristic (ROC) curve. A non-parametric bootstrap regression was used for all models, given that we expect a non-linear relationship between variables and to allow for more model flexibility.

**Results: **We had a total of 80 patients, with 18 (22.5%) survivors and 62 (77.5%) non-survivors. Pneumonia was the most common source of infection (44/80; 55%) and gastrointestinal the least common (15/80; 19%) (**Table 1**). There were no differences at baseline except for the SOFA and APACHE-II scores (**Table 1**). For our primary outcome, the prognostic model, patients with a higher L/P ratio at 8 hours had an OR of 1.08 (95% CI 1.02 – 1.18), holding constant the rest of the variables and had better performance than the other two models (**Figure 1**). For our secondary outcome, we found that the prognostic model using the APACHE-II in addition to the L/P ratio and the L/P ratio alone had a very similar association to the mortality (OR 1.07, 95% CI 1.01 – 1.17 and OR 1.08, 95% CI 1.02 – 1.18) (**Figure 2**). However, there was a substantial difference in the model’s performance between all models (**Figure 2**).

**Conclusions: **While the L/P ratio appears to be an earlier predictor of mortality than hyperlactatemia alone, its comparability against other standard clinical risk scores is negligible or suboptimal. The L/P ratio’s clinical applicability seems to be more towards daily use to avoid misinterpretation of isolated hyperlactatemia.