Association Between Lactic Dehydrogenase-to-Albumin Ratio and Short-Time Mortality in Patients with Chronic Obstructive Pulmonary Disease

Introduction

Chronic obstructive pulmonary disease (COPD) is a major global health concern, associated with considerable morbidity and mortality.¹ Pulmonary function testing (PFT), including the measurement of forced expiratory volume in one second (FEV₁) and forced vital capacity (FVC), remains the gold standard for the diagnosis of COPD. In 2020, the global prevalence of COPD was estimated at 10.6%, affecting approximately 480 million individuals, and it ranks as the fourth leading cause of death worldwide.^2,3 The pathogenesis of COPD is multifactorial, with prolonged exposure to inhaled pollutants—such as tobacco smoke and airborne toxins—playing a key role.^4,5 The incidence of acute exacerbations of COPD requiring hospitalization varies between 2% and 19% among intensive care unit (ICU) admissions.⁶ Patients with COPD exhibit higher death rate in the ICU compared to those without the disease.⁷ As a result, identifying specific biomarkers for early prognosis is essential for providing timely interventions.

Patients with COPD often exhibit elevated lactic dehydrogenase (LDH) levels.^8,9 Increased LDH levels have been linked to compromised pulmonary function, particularly reductions in FVC and FEV1.¹⁰ A recent study from the US National Health and Nutrition Examination Surveys (NHANES) found a U-shaped relationship between serum LDH levels and all-cause mortality in patients with COPD.¹¹ However, this study focused on general COPD patients, which may limit the applicability of its findings to those specifically admitted to the ICU. Additionally, study indicated that serum albumin concentrations are significantly reduced in individuals with stable COPD compared to healthy controls, suggesting a substantial impairment in systemic anti-inflammatory and antioxidant mechanisms in these patients. A negative relationship has been observed between serum albumin levels and in-hospital mortality among critically ill COPD patients.¹² Since LDH reflects tissue damage and systemic inflammation, while albumin serves as a marker of nutritional and inflammatory status, the combination of these two biomarkers as the LDH-to-albumin ratio (LAR) may offer a more comprehensive prognostic insight into disease severity in COPD patients.

Recent studies suggest that the LAR serves as a valuable prognostic marker for critically ill patients with conditions like stroke-associated pneumonia, lower respiratory tract infection, acute respiratory distress syndrome, and acute kidney injury.^13–16 However, the link between LAR and mortality in COPD patients remains poorly understood.

This study aims to explore the relationship between LAR and 28-day mortality in COPD patients, using data from the Medical Information Mart for Intensive Care IV (MIMIC-IV, version 3.1) database.

Materials and Methods

Study Design

This study is a retrospective cohort analysis of 1048 COPD patients, with data obtained from the MIMIC-IV (v3.1) database. Developed and maintained by the Computational Physiology Laboratory at the Massachusetts Institute of Technology, the database contains comprehensive clinical information from 65,366 ICU patients admitted between 2008 and 2022. The data include a variety of clinical variables, such as vital signs, laboratory results, and other relevant medical information. Data extraction was performed by Xiqing Yao, who completed the required online training and exams. All patient data within MIMIC have been rendered fully anonymous in accordance with HIPAA Safe Harbor standards. Under the Measures for Ethical Review of Life Science and Medical Research Involving Human Subjects (February 18, 2023, China), Article 32(1) permits exemption from formal ethics review for research using “lawfully obtained public data or data generated through unobtrusive observation of public behavior”, and Article 32(2) further allows exemption for studies based solely on “anonymized information or biological samples”, Consistent with these provisions, this study was exempt from institutional ethical approval.

Selection of Study Population

Patients with a diagnosis of COPD were identified according to the Tenth Revision of the International Classification of Diseases (ICD-10) codes (J440, J441, J449). The study inclusion criteria were as follows: (1) patients aged over 18 years, (2) ICU admission for a duration exceeding 24 hours, and (3) measurement of LDH and albumin levels within 24 hours of admission. Following these criteria, the cohort was divided into two subgroups: the 28-day survivors and the 28-day non-survivors.

Variable Extraction

Data extraction was conducted by using Structured Query Language (SQL) within Navicat Premium (Version 16.3). A range of variables, either directly obtained or derived, were incorporated for subsequent analysis. These variables encompassed demographic characteristics (age, sex, race), comorbid conditions (sepsis-3, myocardial infarction, peripheral vascular disease, cerebrovascular disease, diabetes, renal disease, and malignancy), indicators of illness severity (Charlson Comorbidity Index, Simplified Acute Physiology Score II [SAPSII], Sequential Organ Failure Assessment [SOFA]), vital signs (oxygen saturation [SpO₂], respiratory rate, systolic and diastolic blood pressure, heart rate [HR]), initial laboratory values (blood urea nitrogen [BUN], sodium, creatinine, potassium, platelets, red blood cells [RBC], white blood cells [WBC], LDH, albumin), and therapeutic interventions (antibiotics, mechanical ventilation, continuous renal replacement therapy [CRRT]).

Variables with missing data exceeding 15% were excluded from the statistical analyses. Since the variable for height had missing data beyond this threshold, leading to the exclusion of body mass index (BMI) from the analysis. No variables exhibited missing data in the range of 5% to 15%. For those variables with missing values below 5%, such as HR, SpO₂, potassium, RBC, and WBC, imputation was performed using the mean values. Outliers were managed through the application of the winsorize method.

The primary endpoint of this study was the 28-day all-cause mortality rate.

Statistical Analysis

The normality of continuous data was evaluated through the Skewness-Kurtosis test. Continuous variables exhibiting a normal distribution were expressed as means with standard deviations, while those with a non-normal distribution were described using medians and interquartile ranges (IQR). The Student’s t-test was used to assess differences in continuous variables with a normal distribution. In contrast, for variables that were either non-continuous or did not adhere to a normal distribution, the Mann–Whitney U-test was used. Categorical variables were examined using the chi-squared test or Fisher’s exact test.

Univariate Cox regression was initially conducted to identify potential risk factors. Variables with P-values less than 0.1 were subsequently analyzed in a multivariate Cox regression to assess their independent significance. The predictive accuracy of LAR, lactate dehydrogenase, albumin, SOFA, and SAPSII for 28-day mortality in COPD patients was evaluated using receiver operating characteristic (ROC) curve analysis. The area under the curve (AUC) was calculated to measure prediction accuracy for 28-day mortality. The optimal cut-off value for LAR was determined via the Youden index, and patients were categorized into high and low LAR groups based on this threshold. Kaplan-Meier survival curves were plotted for each group, and the Log Rank test was used to compare the survival outcomes. All statistical tests were two-tailed, with a significance threshold set at P < 0.05. The analysis was performed using R statistical software (version 4.4.2).

Results

Study Population

A total of 1048 patients with COPD extracted from the MMIC-IV database from 2008 to 2022 This group comprised 710 survivors and 388 non-survivors who died within 28 days of hospital admission. The selection procedure for the study cohort is depicted in Figure 1.

Figure 1 Flow chart of the current study.

Table 1 presents the demographic and clinical characteristics of the study cohort. The overall 28-day all-cause mortality rate was 37%. On average, non-surviving patients were older than survivors (P < 0.001). No significant differences in comorbidity were identified between the two groups, except for a higher prevalence of malignancy in the non-surviving group (P = 0.002). Individuals in the non-survival cohort had elevated severity scores, as assessed by the Charlson Comorbidity Index, SAPSII, and SOFA scores (P < 0.001). Additionally, they presented with higher respiratory rates (P = 0.01) and heart rates (P < 0.001). Laboratory data indicated that the LAR at admission was notably greater in the non-surviving group compared to the survivors (321.6 [82.6, 242.0] vs 148.1 [60.3, 117.0], P < 0.001). LDH levels were significantly higher (P < 0.001) in the non-survival group, while albumin was remarkably lower in the non-survival group (P < 0.001). The RRT was used in 18.4% of the non-survival patients, a proportion significantly greater than that in the survival group (P < 0.001). No significant differences were found between the two groups regarding the administration of antibiotics or vasopressors.

Table 1 Baseline Characteristics of Total Cohort, 28-Day Survivors, and 28-Day Non- Survivors

Correlation Between LAR and 28-Day Mortality

Following the results of the univariate Cox regression analysis, several variables were selected for subsequent examination, including LAR, age, race, malignancy, Charlson Comorbidity Index, SAPS-II, SOFA, CRRT, SpO2, respiratory rate, DBP, SBP, heart rate, BUN, creatinine, potassium, platelets, RBC, and WBC. In the Model I, LAR, along with basic demographic characteristics such as age, sex, and race, were included. The univariate Cox regression analysis demonstrated that LAR was significantly correlated with 28-day all-cause mortality (HR = 1.01; 95% CI, 1.01–1.01; P < 0.001). In the Model II, after adjusting for all included variables, LAR remained a significant independent predictor of 28-day all-cause mortality (HR = 1.01, 95% CI = 1.01–1.01, P = 0.03). Although the HR for LAR per unit increase is 1.01, which is numerically small, interpreting the HR per 100-unit increase provides more clinical relevance. Specifically, a 100-unit increase in LAR corresponds to an HR of approximately 2.7, indicating a 170% increase in hazard. Furthermore, the Charlson Comorbidity Index, SAPS-II, and WBC count were identified as additional independent risk factors for COPD prognosis (Table 2).

Table 2 Multivariate COX Analysis of Risk Factors for 28-Day Mortality

Survival Analysis

The SAPSII and SOFA scoring systems have been developed and validated as tools for predicting mortality risk in critically ill individuals. In the multivariable Cox regression analysis, SAPSII was demonstrated to be an independent predictor of 28-day all-cause mortality among patients with COPD. To further investigate 28-day mortality in this cohort, ROC curves were generated for variables including LAR, lactic dehydrogenase, albumin, SOFA, and SAPSII (Figure 2).

Figure 2 ROC curve of 28-day mortality in patients with COPD.

The AUC of LAR was 69.48% (95% CI: 65.67–73.29%), demonstrating superior performance compared to LDH [66.69% (95% CI: 62.72–70.66%)], albumin [36.88% (95% CI: 32.77–40.99%)], and SOFA [61.91% (95% CI: 57.91–65.9%)]. Additionally, LAR’s diagnostic accuracy was found to be comparable to that of SAPSII, which had an AUC of 76.8% (95% CI: 73.54–80.06%). Following this, COPD patients were stratified into two groups based on a LAR cut-off value of 120: high LAR and low LAR. Kaplan-Meier survival analysis indicated that individuals in the high LAR group experienced a significantly higher 28-day mortality rate compared to those in the low LAR group (P < 0.001, Figure 3). This finding was consistent when analyzing in-hospital mortality rates (P < 0.001). Additionally, patients with higher LAR values had prolonged ICU stays (P < 0.001, Table 3).

Table 3 Relationship Between LAR-Level and Clinical Outcomes

Figure 3 Survival curves for the relationship between LAR and risk of 28-day mortality.

Discussion

The present study identifies the lactic dehydrogenase-to-albumin ratio (LAR) as an independent prognostic factor for 28-day overall mortality in patients with COPD. In comparative analyses, LAR demonstrates superior predictive accuracy (AUC = 69.48%) relative to either LDH (AUC = 66.69%) or albumin (AUC = 36.88%) alone. Additionally, the performance of LAR is found to be on par with that of the SAPSII (AUC = 76.8%). Kaplan-Meier survival analysis further highlights that patients with COPD in high LAR group exhibit a markedly high rate of both 28-day overall mortality and in-hospital mortality.

LAR has increasingly been recognized as a valuable prognostic tool in critically ill patient populations, including those with conditions such as cardiac arrest,¹⁷ acute respiratory distress syndrome,¹³ acute heart failure,¹⁸ and pulmonary embolism.¹⁹ LDH and albumin are routinely measured in these patients, not only due to their established roles as biomarkers but also because they provide additional insights into the patient’s clinical condition. LDH is a well-established marker of tissue injury and the progression of various pathological conditions,^20–22 while albumin is commonly used to assess nutritional status and inflammatory processes.^23,24 A retrospective cohort study conducted in Iran found that LAR could predict mortality in COVID-19 patients, with an AUC of 0.768.²⁵ However, no studies have yet used LAR to predict the prognosis in COPD patients. The prognostic value of LAR likely reflects the combined impact of systemic inflammation, oxidative stress, and malnutrition—all of which are prevalent in ICU-admitted COPD patients. Elevated LDH may reflect ongoing tissue damage and inflammatory activation, while low albumin levels can indicate poor nutritional reserves and impaired hepatic function. Together, LAR provides a composite index of physiological stress that is clinically relevant in acute care settings. Our study indicates that LAR is an independent predictor of 28-day mortality in COPD patients. The diagnostic accuracy of LAR (AUC = 69.48%) exceeded that of LDH (AUC = 66.69%) and albumin (AUC = 36.88%) individually. Furthermore, the performance of LAR was comparable to that of SAPSII (AUC = 76.8%). Kaplan-Meier survival analysis revealed that COPD patients with elevated LAR levels experienced significantly higher rates of both 28-day and in-hospital mortality, as well as prolonged hospital stays.

In this study, we compared the LAR, SOFA and SAPSII for predicting 28-day mortality in patients with COPD. The SOFA score is well-established as an effective prognostic tool for mortality in critically ill patients, including those with COPD.^26,27 Both SAPSII and SOFA are routinely used in clinical settings to assess the severity of illness in critically ill patients.^28,29 In particular, the SOFA score is noted for its utility in predicting the duration of hospital stay and the likelihood of mortality.³⁰ The SAPSII scoring system incorporates a broad range of physiological and clinical parameters, such as respiratory rate (RR), oxygenation index, and additional indicators. Previous research has affirmed the prognostic value of SAPSII in COPD patients.³¹ Consistent with these findings, multivariate Cox regression analyses have demonstrated that SAPSII is an independent predictor of 28-day mortality in patients with COPD. Our findings indicate that LAR exhibited predictive accuracy comparable to SAPSII. This suggests that LAR, which is calculated using only LDH and albumin, provides similar prognostic insights as SAPSII, despite the latter incorporating a broader array of clinical parameters. Consequently, LAR may offer a more streamlined approach for clinical risk assessment.

This study has several limitations. First, we employed 28-day mortality as the principal prognostic endpoint, without incorporating an evaluation of patient survival status through the MIMIC-IV database. To enhance the precision in assessing the prognostic value of the LAR, a multicenter, prospective study with comprehensive survival data for COPD patients is recommended. Second, the absence of lung function measurements for a substantial proportion of the patient cohort in this investigation limited the capacity to categorize the severity of airflow limitation in all cases of COPD. Third, despite adjusting for multiple clinical covariates, several important potential confounders—such as detailed nutritional status (eg, body mass index, prealbumin), corticosteroid use, and liver dysfunction—were not fully captured in the database and thus not included in our models.

Conclusion

The current study indicates that LAR acts as an independent prognostic factor for 28-day overall mortality in patients with COPD. These findings suggest that LAR could serve as a more convenient tool for anesthesiologists and critical care specialists in comparison to SAPSII, while offering similar accuracy in the prompt identification of COPD patients at high risk for poor outcomes.