Abstract
Hypertensive heart disease (HHD) complicated by heart failure is a significant cause of morbidity and mortality. The glucose-to-lymphocyte ratio (GLR) has been suggested as a potential marker of inflammation and stress. This study aimed to investigate the relationship between GLR and 28-day all-cause mortality in patients with HHD complicated by heart failure using the Medical Information Mart for Intensive Care, version IV database. We conducted a retrospective analysis of 6203 patients admitted to the intensive care unit (ICU) with HHD complicated by heart failure. Patients were excluded if they were not admitted to the ICU, had glucose deficiency, lymphocytopenia, or were identified as outliers. Patients were divided into quartiles based on GLR values. Cox regression models were used to evaluate the association between GLR and 28-day all-cause mortality. Subgroup analyses were performed to assess the impact of various clinical factors on this relationship. The study population had a mean age of 73.2 years, with 42.0% being male. The median GLR was 11.2 (interquartile range 6.4–20.0). In univariate Cox regression analysis, GLR was significantly associated with 28-day all-cause mortality (hazard ratio [HR] 1.0049, 95% confidence interval [CI] 1.0037–1.006, P < .001). Multivariate analysis confirmed this association, with the highest GLR quartile (Q4) showing a significantly higher risk of mortality compared to the lowest (Q1) (HR 2.57, 95% CI 2.18–3.03, P < .001). Threshold effect analysis identified a turning point at GLR 31.879, where the hazard ratio was 1.04 (95% CI 1.032–1.048, P < .001) for GLR below this value and 1.0023 (95% CI 0.9949–1.0097, P = .5481) for GLR above it. Kaplan–Meier survival analysis and forest plots further supported these findings. GLR is significantly associated with 28-day all-cause mortality in patients with HHD complicated by heart failure. Higher GLR values are associated with an increased risk of mortality. Our findings suggest that GLR could serve as a useful prognostic marker in this patient population.
Keywords: glucose-to-lymphocyte ratio, heart failure, hypertensive heart disease, MIMIC-IV database, prognosis
1. Introduction
Heart failure represents a critical stage in the progression of various cardiovascular diseases and poses a significant challenge to global health.[1] Hypertensive heart disease (HHD), as one of the common causes of heart failure, has seen a continuous rise in its incidence and mortality rates over the past few decades.[2] The prognosis of heart failure patients is influenced by a multitude of factors, including age, gender, ethnicity, comorbidities, and markers of inflammation and stress.[3,4] In recent years, the role of inflammation in cardiovascular diseases has gained increasing attention, being recognized as one of the crucial mechanisms in the development and prognosis of cardiovascular diseases.[5,6]
The glucose-to-lymphocyte ratio (GLR) is an emerging marker of inflammation and stress, easily calculated from readily available blood glucose and lymphocyte counts.[7] GLR has been shown to correlate with the prognosis of various cardiovascular diseases, including coronary artery bypass grafting, ST-Elevation Myocardial Infarction, and diabetes.[8–12] However, the role of GLR in patients with HHD complicated by heart failure has not been thoroughly investigated.
The MIMIC-IV database (Medical Information Mart for Intensive Care, version IV) is a large, publicly accessible database containing detailed clinical data from intensive care units (ICUs).[13] This database provides a valuable resource for studying the relationship between GLR and the prognosis of patients with HHD complicated by heart failure.
The aim of this study is to use the MIMIC-IV database to conduct a retrospective analysis exploring the association between GLR and 28-day all-cause mortality in patients with HHD complicated by heart failure. We hypothesize that GLR may serve as a useful prognostic marker, assisting clinicians in better assessing and managing the treatment and prognosis of these patients. Through this study, we hope to provide new insights into the clinical management of patients with HHD complicated by heart failure and lay the groundwork for future research.
2. Methods
2.1. Study design and data source
This retrospective cohort study utilized data from the MIMIC-IV database, which includes de-identified health data from ICU patients.[13] The study was conducted in accordance with the ethical standards of the Declaration of Helsinki and was exempt from institutional review board approval due to the use of publicly available de-identified data. Author Guang Tu completed the CITI Data or Specimens Only Research course, obtained approval for database access, and assumed responsibility for data extraction (certification number 65828445).
2.2. Study population
We identified patients with a primary diagnosis of hypertensive heart disease complicated by heart failure using relevant International Classification of Diseases, Tenth Revision codes (I110, I132, 40291, 40201, I130, 40491, 40493, 40401, 40411, 40211, 40403, 40413) from the MIMIC-IV database. The inclusion criteria were as follows:
Patients admitted to the ICU with a diagnosis of hypertensive heart disease and heart failure.
Patients with complete medical records available in the database.
The exclusion criteria were:
Patients not admitted to the ICU.
Patients with glucose deficiency or lymphocytopenia.
Outliers identified through statistical analysis.
2.3. Data collection
Data collected included demographic information (age, gender, and race), clinical variables (heart rate, blood pressure, temperature, and blood tests), and outcomes (28-day all-cause mortality). The GLR was calculated using the formula: GLR = blood glucose (mg/L)/lymphocyte count (×109/L).
2.4. Handling of missing data
To address missing data, we employed multiple imputation by chained equations. This method involves creating multiple imputed datasets by iteratively imputing missing values using a series of regression models. The imputed datasets were then combined to provide robust estimates of the parameters of interest. This approach helps to reduce bias and improve the efficiency of the estimates compared to complete-case analysis.
2.5. Statistical analysis
We performed all statistical analyses using R (version 4.2.2) and the Free Statistics Analysis Platform (version 2.1.1),[14] with a 2-sided P < .05 considered significant. Univariate and multivariate Cox proportional-hazards models were constructed to estimate the association between the GLR and 28-day all-cause mortality; multivariable adjustment was performed in a stepwise fashion for demographics (age, sex, and race) and then for additional clinical covariates including myocardial infarction, cerebrovascular disease, rheumatic disease, diabetes, paraplegia, and renal disease.
To explore nonlinear relationships, we applied restricted cubic spline (RCS) regression to identify potential threshold effects of GLR on mortality. Kaplan–Meier survival curves were generated for GLR quartiles and compared with the log-rank test. Finally, subgroup analyses stratified by age, sex, and comorbidities were conducted, with the results visualized in forest plots; interaction terms were included in the Cox models to assess effect modification across subgroups.
3. Results
3.1. Patient inclusion and baseline characteristics
A total of 6203 patients with hypertensive heart disease complicated by heart failure were included in the final analysis after excluding 22,182 patients not admitted to the ICU, 115 patients with glucose deficiency, 3882 patients with lymphocyte deficiency, and 29 outliers (Fig. 1). The patients were divided into 4 quartiles based on their GLR values: Q1 (GLR < 6.37), Q2 (6.37 ≤ GLR < 11.21), Q3 (11.21 ≤ GLR < 20.00), and Q4 (GLR ≥ 20.00).
Figure 1.
Flowchart of patient inclusion.
The baseline characteristics of the study population are presented in Table 1. Significant differences were observed across the quartiles for age, race, heart rate, temperature, peripheral capillary oxygen saturation, platelets, white blood cell count, albumin, anion gap, bicarbonate, blood urea nitrogen, calcium, chloride, creatinine, sodium, potassium, basophils, eosinophils, lymphocytes, monocytes, neutrophils, international normalized ratio, prothrombin time (activated partial thromboplastin time, alanine aminotransferase, alkaline phosphatase, aspartate aminotransferase, total bilirubin, lactate dehydrogenase, lactate, pH, PO2, PCO2, diabetes, paraplegia, renal disease, and GLR (all P < .001).
Table 1.
Baseline characteristics of study population.
| Variables | Total (n = 6203) | Q1 (n = 1551) | Q2 (n = 1550) | Q3 (n = 1535) | Q4 (n = 1567) | P-value |
|---|---|---|---|---|---|---|
| Gender, n (%) | .354 | |||||
| Male | 2607 (42.0) | 652 (42) | 680 (43.9) | 631 (41.1) | 644 (41.1) | |
| Female | 3596 (58.0) | 899 (58) | 870 (56.1) | 904 (58.9) | 923 (58.9) | |
| Age, years, mean ± SD | 73.2 ± 12.6 | 71.0 ± 13.4 | 73.4 ± 12.8 | 74.2 ± 12.1 | 74.3 ± 11.9 | <.001 |
| Race, n (%) | <.001 | |||||
| Non-White | 2267 (36.5) | 601 (38.7) | 627 (40.5) | 527 (34.3) | 512 (32.7) | |
| White | 3936 (63.5) | 950 (61.3) | 923 (59.5) | 1008 (65.7) | 1055 (67.3) | |
| Heart rate, beats/min, mean ± SD | 70.2 ± 15.4 | 69.5 ± 14.4 | 69.0 ± 15.2 | 70.4 ± 16.1 | 71.8 ± 15.8 | <.001 |
| Systolic blood pressure, mm Hg, mean ± SD | 89.8 ± 17.8 | 89.6 ± 16.9 | 90.4 ± 17.5 | 90.2 ± 18.5 | 89.1 ± 18.4 | .185 |
| Diastolic blood pressure, mm Hg, mean ± SD | 46.6 ± 11.6 | 46.4 ± 10.9 | 47.0 ± 11.5 | 46.8 ± 11.9 | 46.4 ± 11.9 | .476 |
| Temperature, °C, mean ± SD | 36.3 ± 0.7 | 36.2 ± 0.8 | 36.3 ± 0.7 | 36.4 ± 0.7 | 36.4 ± 0.6 | <.001 |
| SpO2, %, mean ± SD | 90.2 ± 7.5 | 91.1 ± 7.0 | 90.5 ± 7.2 | 89.8 ± 7.9 | 89.4 ± 7.9 | <.001 |
| Hemoglobin, g/dL, mean ± SD | 9.4 ± 2.3 | 9.3 ± 2.2 | 9.4 ± 2.4 | 9.4 ± 2.3 | 9.4 ± 2.3 | .523 |
| Platelets, ×109/L, mean ± SD | 184.3 ± 95.8 | 185.8 ± 98.7 | 188.7 ± 90.5 | 183.9 ± 91.1 | 178.8 ± 102.3 | .032 |
| White blood cell count, ×109/L, median (IQR) | 9.5 (6.8, 13.0) | 10.2 (7.7, 13.5) | 9.5 (7.0, 12.5) | 9.2 (6.7, 12.7) | 9.0 (6.0, 13.1) | <.001 |
| Albumin, g/dL, mean ± SD | 3.2 ± 0.6 | 3.3 ± 0.7 | 3.3 ± 0.6 | 3.2 ± 0.6 | 3.1 ± 0.6 | <.001 |
| Anion gap, mmol/L, mean ± SD | 13.1 ± 4.2 | 12.3 ± 3.9 | 12.9 ± 4.0 | 13.3 ± 4.3 | 13.8 ± 4.4 | <.001 |
| Bicarbonate, mmol/L, mean ± SD | 21.1 ± 5.5 | 21.2 ± 4.9 | 21.2 ± 5.4 | 21.1 ± 5.6 | 21.0 ± 6.1 | .817 |
| Blood urea nitrogen, mg/dL, median (IQR) | 27.0 (17.0, 45.0) | 21.0 (15.0, 37.0) | 25.0 (16.0, 42.0) | 29.0 (18.0, 47.0) | 33.0 (20.0, 55.0) | <.001 |
| Calcium, mmol/L, mean ± SD | 8.3 ± 0.8 | 8.4 ± 0.8 | 8.3 ± 0.8 | 8.3 ± 0.8 | 8.2 ± 0.8 | <.001 |
| Chloride, mmol/L, mean ± SD | 99.0 ± 6.9 | 100.6 ± 6.5 | 99.4 ± 7.1 | 98.5 ± 6.8 | 97.6 ± 6.8 | <.001 |
| Creatinine, mg/dL, median (IQR) | 1.3 (0.9, 2.2) | 1.1 (0.9, 1.9) | 1.2 (0.9, 2.1) | 1.4 (0.9, 2.3) | 1.5 (1.0, 2.6) | <.001 |
| Sodium, mmol/L, mean ± SD | 136.4 ± 5.6 | 136.9 ± 5.1 | 136.6 ± 5.6 | 136.2 ± 5.8 | 135.8 ± 5.7 | <.001 |
| Potassium, mmol/L, mean ± SD | 4.1 ± 0.6 | 4.2 ± 0.6 | 4.1 ± 0.7 | 4.1 ± 0.6 | 4.1 ± 0.7 | <.001 |
| Basophils, ×109/L, median (IQR) | 0.0 (0.0, 0.0) | 0.0 (0.0, 0.1) | 0.0 (0.0, 0.0) | 0.0 (0.0, 0.0) | 0.0 (0.0, 0.0) | <.001 |
| Eosinophils, ×109/L, median (IQR) | 0.0 (0.0, 0.1) | 0.1 (0.0, 0.3) | 0.1 (0.0, 0.1) | 0.0 (0.0, 0.1) | 0.0 (0.0, 0.0) | <.001 |
| Lymphocytes, ×109/L, median (IQR) | 1.0 (0.6, 1.7) | 2.4 (1.9, 3.5) | 1.3 (1.0, 1.6) | 0.8 (0.7, 1.0) | 0.4 (0.3, 0.6) | <.001 |
| Monocytes, ×109/L, median (IQR) | 0.7 (0.4, 1.0) | 0.8 (0.5, 1.2) | 0.7 (0.5, 1.0) | 0.7 (0.4, 1.0) | 0.5 (0.2, 0.8) | <.001 |
| Neutrophils, ×109/L, median (IQR) | 8.8 (5.9, 13.1) | 9.9 (6.5, 15.7) | 8.2 (5.7, 11.8) | 8.6 (5.8, 12.5) | 8.7 (5.6, 12.9) | <.001 |
| INR, mean ± SD | 1.5 ± 0.8 | 1.4 ± 0.8 | 1.5 ± 0.8 | 1.5 ± 0.7 | 1.6 ± 0.8 | <.001 |
| PT, seconds, mean ± SD | 16.3 ± 8.3 | 15.6 ± 8.4 | 15.9 ± 8.5 | 16.6 ± 7.3 | 17.2 ± 8.8 | <.001 |
| APTT, seconds, mean ± SD | 32.4 ± 13.3 | 30.6 ± 10.1 | 32.5 ± 13.8 | 32.9 ± 13.6 | 33.9 ± 14.9 | <.001 |
| ALT, U/L, median (IQR) | 21.0 (13.0, 40.0) | 21.0 (13.0, 40.0) | 20.0 (12.0, 37.0) | 20.0 (13.0, 39.0) | 22.0 (13.0, 46.0) | .022 |
| ALP, U/L, median (IQR) | 89.0 (67.0, 125.0) | 85.0 (64.0, 115.8) | 89.0 (68.0, 123.0) | 90.0 (67.0, 128.0) | 91.0 (68.0, 131.0) | <.001 |
| AST, U/L, median (IQR) | 30.0 (20.0, 60.0) | 33.0 (21.0, 63.0) | 29.0 (19.0, 52.0) | 29.5 (19.0, 60.0) | 31.0 (19.2, 65.0) | .046 |
| Total bilirubin, mg/dL, median (IQR) | 0.6 (0.3, 0.9) | 0.5 (0.3, 0.9) | 0.5 (0.4, 0.8) | 0.6 (0.4, 1.0) | 0.6 (0.4, 1.0) | <.001 |
| LDH, U/L, median (IQR) | 292.0 (220.0, 431.0) | 294.0 (223.0, 437.8) | 289.0 (216.0, 448.0) | 284.5 (214.2, 411.0) | 302.5 (226.0, 444.8) | .111 |
| Lactate, mmol/L, median (IQR) | 1.4 (1.1, 2.0) | 1.3 (1.0, 1.8) | 1.3 (1.0, 1.8) | 1.5 (1.1, 2.2) | 1.6 (1.2, 2.3) | <.001 |
| Ph, mean ± SD | 7.3 ± 0.1 | 7.3 ± 0.1 | 7.3 ± 0.1 | 7.3 ± 0.1 | 7.3 ± 0.1 | <.001 |
| PO2, mm Hg, mean ± SD | 85.4 ± 47.1 | 93.8 ± 46.3 | 87.4 ± 45.2 | 82.3 ± 49.4 | 78.0 ± 45.9 | <.001 |
| PCO2, mm Hg, mean ± SD | 39.0 ± 11.5 | 37.2 ± 10.0 | 39.0 ± 11.4 | 39.2 ± 11.7 | 40.7 ± 12.7 | <.001 |
| Myocardial infarction, n (%) | .321 | |||||
| No | 3886 (62.6) | 941 (60.7) | 978 (63.1) | 974 (63.5) | 993 (63.4) | |
| Yes | 2317 (37.4) | 610 (39.3) | 572 (36.9) | 561 (36.5) | 574 (36.6) | |
| Cerebrovascular disease, n (%) | .522 | |||||
| No | 5236 (84.4) | 1315 (84.8) | 1295 (83.5) | 1289 (84) | 1337 (85.3) | |
| Yes | 967 (15.6) | 236 (15.2) | 255 (16.5) | 246 (16) | 230 (14.7) | |
| Rheumatic disease, n (%) | .022 | |||||
| No | 5917 (95.4) | 1483 (95.6) | 1498 (96.6) | 1454 (94.7) | 1482 (94.6) | |
| Yes | 286 (4.6) | 68 (4.4) | 52 (3.4) | 81 (5.3) | 85 (5.4) | |
| Diabetes, n (%) | <.001 | |||||
| No | 4009 (64.6) | 1093 (70.5) | 1032 (66.6) | 964 (62.8) | 920 (58.7) | |
| Yes | 2194 (35.4) | 458 (29.5) | 518 (33.4) | 571 (37.2) | 647 (41.3) | |
| Paraplegia, n (%) | .023 | |||||
| No | 5904 (95.2) | 1471 (94.8) | 1466 (94.6) | 1453 (94.7) | 1514 (96.6) | |
| Yes | 299 (4.8) | 80 (5.2) | 84 (5.4) | 82 (5.3) | 53 (3.4) | |
| Renal disease, n (%) | <.001 | |||||
| No | 3022 (48.7) | 828 (53.4) | 804 (51.9) | 709 (46.2) | 681 (43.5) | |
| Yes | 3181 (51.3) | 723 (46.6) | 746 (48.1) | 826 (53.8) | 886 (56.5) | |
| Apsiii, mean ± SD | 54.4 ± 22.9 | 50.4 ± 22.6 | 51.4 ± 21.6 | 55.6 ± 23.4 | 60.0 ± 22.8 | <.001 |
| Sapsii, mean ± SD | 41.8 ± 13.3 | 40.5 ± 13.5 | 40.4 ± 12.0 | 42.3 ± 13.6 | 43.9 ± 13.7 | <.001 |
| Oasis, mean ± SD | 33.8 ± 9.1 | 33.2 ± 8.5 | 33.1 ± 8.9 | 34.0 ± 9.2 | 34.9 ± 9.6 | <.001 |
| Sofa score, mean ± SD | 4.1 ± 2.3 | 4.2 ± 2.3 | 4.0 ± 2.1 | 4.0 ± 2.3 | 4.2 ± 2.4 | .065 |
| Glucose, mg/L, mean ± SD | 12.5 ± 4.9 | 10.7 ± 3.4 | 11.4 ± 3.4 | 12.9 ± 4.8 | 15.1 ± 6.3 | <.001 |
| GLR, median (IQR) | 11.2 (6.4, 20.0) | 4.4 (3.0, 5.4) | 8.6 (7.4, 9.9) | 14.9 (12.8, 17.0) | 30.6 (24.0, 46.2) | <.001 |
ALP = alkaline phosphatase, ALT = alanine aminotransferase, APTT = activated partial thromboplastin time, AST = aspartate aminotransferase, GLR = glucose-to-lymphocyte ratio, INR = international normalized ratio, IQR = interquartile range, LDH = lactate dehydrogenase, PT = prothrombin time, SpO2 = peripheral capillary oxygen saturation.
3.2. Univariate Cox regression analysis
Univariate Cox regression analysis showed that GLR was significantly associated with 28-day all-cause mortality (hazard ratio [HR] 1.0049, 95% confidence interval [CI] 1.0037–1.006, P < .001) (Table 2).
Table 2.
A univariate Cox regression model evaluated the association between glucose-to-lymphocyte ratio and 28-day all-cause mortality in patients with hypertensive heart disease complicated by heart failure.
| Item | HR (95% CI) | P (Wald test) |
|---|---|---|
| Gender (male vs female) | 0.95 (0.85, 1.06) | .373 |
| Age (per year) | 1.03 (1.03, 1.04) | <.001 |
| Race (nonwhite vs white) | 1.002 (0.897, 1.121) | .959 |
| Heart rate (per beat/min) | 1.008 (1.004, 1.012) | <.001 |
| Systolic blood pressure (per mm Hg) | 0.98 (0.97, 0.98) | <.001 |
| Diastolic blood pressure (per mm Hg) | 0.97 (0.97, 0.98) | <.001 |
| Temperature (per °C) | 0.82 (0.79, 0.86) | <.001 |
| SpO2 (per %) | 0.97 (0.96, 0.97) | <.001 |
| Hemoglobin (per g/dL) | 0.96 (0.94, 0.99) | .002 |
| Platelets (per × 109/L) | 0.9996 (0.999, 1.0001) | .141 |
| White blood cell count (per × 10⁹/L) | 1.02 (1.01, 1.02) | <.001 |
| Albumin (per g/dL) | 0.46 (0.42, 0.51) | <.001 |
| Anion gap (per mmol/L) | 1.11 (1.1, 1.13) | <.001 |
| Bicarbonate (per mmol/L) | 0.94 (0.93, 0.95) | <.001 |
| Blood urea nitrogen (per mg/dL) | 1.01 (1.01, 1.01) | <.001 |
| Calcium (per mmol/L) | 0.82 (0.77, 0.88) | <.001 |
| Chloride (per mmol/L) | 0.98 (0.97, 0.99) | <.001 |
| Creatinine (per mg/dL) | 1.08 (1.05, 1.11) | <.001 |
| Sodium (per mmol/L) | 1 (0.9904, 1.0098) | .997 |
| Potassium (per mmol/L) | 1.16 (1.07, 1.27) | <.001 |
| Basophils (per × 109/L) | 0.86 (0.75, 0.97) | .015 |
| Eosinophils (per × 109/L) | 0.96 (0.92, 0.99) | .015 |
| Lymphocytes (per × 109/L) | 0.999 (0.997, 1.0009) | .304 |
| Monocytes (per × 109/L) | 0.995 (0.987, 1.002) | .160 |
| Neutrophils (per × 109/L) | 0.9997 (0.9993, 1.0002) | .264 |
| INR (per unit) | 1.28 (1.24, 1.33) | <.001 |
| PT (per second) | 1.03 (1.02, 1.03) | <.001 |
| APTT (per second) | 1.01 (1.01, 1.02) | <.001 |
| ALT (per U/L) | 1.0003 (1.0002, 1.0004) | <.001 |
| ALP (per U/L) | 1.0017 (1.0014, 1.002) | <.001 |
| AST (per U/L) | 1.0001 (1.0001, 1.0002) | <.001 |
| Total bilirubin (per mg/dL) | 1.08 (1.06, 1.1) | <.001 |
| LDH (per U/L) | 1.0002 (1.0001, 1.0002) | <.001 |
| Lactate (per mmol/L) | 1.29 (1.26, 1.31) | <.001 |
| pH (per unit) | 0.02 (0.01, 0.04) | <.001 |
| PO2 (per mm Hg) | 0.997 (0.996, 0.999) | <.001 |
| PCO2 (per mm Hg) | 0.9988 (0.9932, 1.0044) | .669 |
| Myocardial infarction (yes vs no) | 1.16 (1.04, 1.29) | .008 |
| Cerebrovascular disease (yes vs no) | 1.43 (1.25, 1.63) | <.001 |
| Rheumatic disease (yes vs no) | 1.11 (0.87, 1.42) | .387 |
| Diabetes (yes vs no) | 1.04 (0.93, 1.16) | .525 |
| Paraplegia (yes vs no) | 1.4 (1.13, 1.74) | .002 |
| Renal disease (yes vs no) | 1.31 (1.17, 1.45) | <.001 |
| APACHE III (per point) | 1.03 (1.03, 1.03) | <.001 |
| SAPS II (per point) | 1.05 (1.05, 1.06) | <.001 |
| OASIS (per point) | 1.07 (1.07, 1.08) | <.001 |
| SOFA score (per point) | 1.09 (1.06, 1.11) | <.001 |
| Glucose (per mg/L) | 1.03 (1.02, 1.04) | <.001 |
| GLR (per unit) | 1.0049 (1.0037, 1.006) | <.001 |
ALP = alkaline phosphatase, ALT = alanine aminotransferase, APTT = activated partial thromboplastin time, AST = aspartate aminotransferase, CI = confidence interval, GLR = glucose-to-lymphocyte ratio, HR = hazard ratio, INR = international normalized ratio, LDH = lactate dehydrogenase, PT = prothrombin time, SpO2 = peripheral capillary oxygen saturation.
3.3. Multivariate Cox regression analysis
To evaluate the relationship between GLR and 28-day all-cause mortality while accounting for potential confounders, a multivariate Cox regression analysis was conducted across 3 progressively adjusted models. The initial model (Model 1) featured GLR as the sole predictor, yielding a HR of 1.01 (95% CI 1.00–1.01, P < .001). Model 2 incorporated demographic variables such as gender, age, and race, with GLR maintaining an HR of 1.01 (95% CI 1.00–1.01, P < .001). The most comprehensive model (Model 3) further considered clinical factors like myocardial infarction, cerebrovascular disease, rheumatic disease, diabetes, paraplegia, and renal disease, and GLR still exhibited an HR of 1.01 (95% CI 1.00–1.01, P < .001). Across all models, the trend test consistently indicated a significant association between GLR and 28-day all-cause mortality (all P < .001) (Table 3).
Table 3.
A multivariate Cox regression model assessed the relationship between glucose-to-lymphocyte ratio and 28-day all-cause mortality in patients with hypertensive heart disease complicated by heart failure.
| Variable | No | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|---|
| HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | ||
| GLR | 6203 | 1 (1–1.01) | <.001 | 1.01 (1–1.01) | <.001 | 1.01 (1–1.01) | <.001 |
| GLR | |||||||
| Q1 | 1551 | 1 (Ref) | 1 (Ref) | 1 (Ref) | |||
| Q2 | 1550 | 1.45 (1.21–1.74) | <.001 | 1.34 (1.12–1.61) | .001 | 1.35 (1.13–1.62) | .001 |
| Q3 | 1535 | 1.89 (1.59–2.24) | <.001 | 1.73 (1.46–2.06) | <.001 | 1.73 (1.45–2.06) | <.001 |
| Q4 | 1567 | 2.76 (2.34–3.24) | <.001 | 2.57 (2.18–3.02) | <.001 | 2.57 (2.18–3.03) | <.001 |
| Trend test | 6203 | 1.39 (1.32–1.46) | <.001 | 1.37 (1.3–1.44) | <.001 | 1.37 (1.3–1.44) | <.001 |
Model 1 is the unadjusted model. Model 2 is adjusted for gender, age, and race. Model 3 is adjusted for gender, age, race, myocardial infarction, cerebrovascular disease, rheumatic disease, diabetes, paraplegia, and renal disease.
CI = confidence interval, GLR = glucose-to-lymphocyte ratio, HR = hazard ratio.
3.4. Threshold effect of GLR
The threshold effect of GLR on 28-day all-cause mortality was analyzed using a Cox regression model. The turning point was identified at 31.879. Below this threshold, the HR was 1.04 (95% CI 1.032–1.048, P < .001), and above this threshold, the HR was 1.0023 (95% CI 0.9949–1.0097, P = .5481) (Table 4). The likelihood ratio test showed a significant threshold effect (P < .001).
Table 4.
Cox regression model was used to analyze the threshold effect of glucose-to-lymphocyte ratio on 28-day all-cause mortality.
| Item | HR (95% CI) | P-value |
|---|---|---|
| Turning point | 31.879 (30.103, 33.654) | |
| GLR < 31.879 | 1.04 (1.032, 1.048) | <.001 |
| GLR ≧ 31.879 | 1.0023 (0.9949, 1.0097) | .5481 |
| Likelihood ratio test | <.001 |
CI = confidence interval, GLR = glucose-to-lymphocyte ratio, HR = hazard ratio.
3.5. RCS curve and Kaplan–Meier survival analysis
The RCS analysis indicated a notable nonlinear correlation with 28-day all-cause mortality (P for nonlinearity < .001) (Fig. 2). Meanwhile, the Kaplan–Meier survival curves highlighted a significant variation in 28-day all-cause mortality among the quartiles (P < .0001) (Fig. 3).
Figure 2.
RCS curve for the glucose-to-lymphocyte ratio GLR (quartile): Q1 (<6.37), Q2 (6.37–11.21), Q3 (11.21–20.00), Q4 (≥20.00). RCS = restricted cubic spline, GLR = glucose-to-lymphocyte ratio.
Figure 3.
Kaplan–Meier survival analysis curves for 28-day all-cause mortality.
3.6. Subgroup and forest plot analysis
Subgroup analyses were conducted to assess the correlation between GLR and 28-day all-cause mortality among various subgroups. The findings are depicted in Figure 4. Notable interactions were detected between GLR and crude mortality (P < .001), myocardial infarction (P = .675), and diabetes (P = .108). Trend analysis revealed a significant link between GLR and 28-day all-cause mortality across all subgroups (all P < .050).
Figure 4.
Forest plot for the subgroup analysis of the relationship between 28-day all-cause mortality and glucose-to-lymphocyte ratio.
4. Discussion
Our study has demonstrated a significant association between the GLR and 28-day all-cause mortality in patients with hypertensive heart disease complicated by heart failure. The findings underscore the potential of GLR as a clinically relevant prognostic marker, offering a simple yet effective tool for risk assessment and guiding therapeutic strategies. The consistent association across various subgroups and the identification of a threshold effect further enhance the clinical applicability of GLR in managing patients with heart failure.
This study extends and deepens the existing literature in several aspects. Firstly, we utilized a large and comprehensive dataset from the MIMIC-IV database, which enhances the generalizability and credibility of our findings. In contrast to previous studies that may have been limited by smaller sample sizes or less diverse patient populations, our analysis reduces potential biases and strengthens the robustness of the results. For instance, a study based on the MIMIC-IV database explored the association between GLR and the risk of all-cause mortality in patients with nontraumatic intracranial hemorrhage in the ICU and hospital setting, indicating a significant linear relationship between GLR and mortality.[15] Another study assessed the relationship between GLR and long-term mortality after coronary artery bypass grafting in women, also using data from the MIMIC database, suggesting that GLR has predictive value for acute mortality and prognosis in cardiovascular diseases.[8] These studies, while providing valuable insights, focused on specific patient groups or clinical outcomes, whereas our study offers a broader perspective, encompassing a more diverse patient population and considering the relationship between GLR and 28-day all-cause mortality.
Furthermore, by incorporating a threshold effect analysis, we provided a more nuanced understanding of the relationship between GLR and mortality, which has often been overlooked in previous research. Our results indicate that GLR is not only associated with mortality but also that this relationship may manifest differently across various ranges of GLR values, which is significant for risk assessment and therapeutic decision-making in clinical practice.
The association between GLR and mortality may be biologically plausible. GLR is considered an indicator of systemic inflammation and metabolic status.[16] In patients with heart failure, enhanced inflammatory responses are closely associated with adverse outcomes.[17] Research has highlighted that the mortality rate within 5 years following the diagnosis of heart failure can reach up to 50%, which is a significant risk to human health,[18] Additionally, GLR, being a novel inflammatory biomarker, is regarded as an independent prognostic indicator for numerous diseases.[19–21] Additionally, elevated blood glucose levels may indicate metabolic dysfunction, which is common in heart failure patients and may exacerbate cardiac dysfunction.[22] GLR, by integrating these factors, may provide a more comprehensive perspective on the patient’s inflammatory and metabolic status, thereby influencing their prognosis. The biological mechanisms underlying this association may involve the direct impact of inflammatory mediators on cardiac function and the indirect effects of metabolic disorders on cardiac energy supply and myocardial cell survival.[8–11] Therefore, GLR may serve as a comprehensive indicator reflecting the overall pathophysiological state of heart failure patients. Inflammatory mediators such as tumor necrosis factor-alpha and interleukin-6 have been shown to promote cardiac remodeling and functional decline,[23] and an elevated GLR may indicate an increase in these inflammatory mediators, thereby affecting cardiac function and prognosis. Additionally, elevated blood glucose levels may induce chronic oxidative stress, which in turn can impair endothelial function and foster the advancement of atherosclerosis.[24] Under hyperglycemic conditions, the inflammatory response intensifies, characterized by elevated levels of inflammatory cytokines, which promote smooth muscle cell proliferation and collagen deposition, leading to atherosclerosis.[25] Additionally, diabetes also promotes the production of plasminogen activator inhibitor-1, further promoting thrombosis.[26] These biological mechanisms provide a possible explanation for the association between GLR and mortality in patients with heart failure.
One of the major strengths of our study is the use of the MIMIC-IV database, which includes a rich and diverse set of clinical data. This allowed us to conduct a robust analysis with a large sample size, reducing the risk of bias and increasing the confidence in our results. Moreover, our study is clinically relevant as it focuses on a patient population that is frequently encountered in clinical practice. By linking GLR to mortality outcomes, we have provided clinicians with a potentially useful tool for assessing patient risk and guiding treatment strategies.
5. Limitations and efforts to mitigate them
While our study presents significant findings, it is not without limitations. As a retrospective analysis, we were unable to establish causality, and the potential for unmeasured confounding factors exists. To address these limitations, we performed multivariate analyses to adjust for known confounders and conducted subgroup analyses to explore potential effect modifiers. Additionally, we used statistical methods such as the threshold effect analysis to better understand the relationship between GLR and mortality. Despite these efforts, further prospective studies are needed to validate our findings and explore the causal mechanisms underlying the association between GLR and mortality in patients with hypertensive heart disease complicated by heart failure.
6. Conclusions
In summary, our study provides robust evidence that the GLR is significantly associated with 28-day all-cause mortality in patients with hypertensive heart disease complicated by heart failure. The identification of a threshold effect and the consistency of this association across various subgroups further highlight the potential of GLR as a valuable prognostic marker. This easily calculated ratio could serve as a simple yet effective tool for clinicians to assess patient risk and guide therapeutic strategies. Future prospective studies are warranted to validate our findings and explore the underlying mechanisms linking GLR to mortality in this patient population.
Author contributions
Conceptualization: Guang Tu.
Data curation: Guang Tu, Chunyan Wang, Zhonglan Cai, Shengnan Xue.
Formal analysis: Zhonglan Cai, Shengnan Xue.
Methodology: Guang Tu, Chunyan Wang, Zhonglan Cai.
Project administration: Guang Tu.
Resources: Chunyan Wang.
Supervision: Guang Tu.
Writing – original draft: Guang Tu, Chunyan Wang.
Writing – review & editing: Guang Tu, Chunyan Wang.
Abbreviations:
- CI
- confidence interval
- GLR
- glucose-to-lymphocyte ratio
- HHD
- hypertensive heart disease
- HR
- hazard ratio
- ICU
- intensive care unit
- MIMIC-IV
- medical information mart for intensive care, version IV
- RCS
- restricted cubic spline
The research adhered to the principles outlined in the 1964 Helsinki Declaration and subsequent revisions. MIMIC-IV serves as a de-identified public dataset. Approval for this endeavor was granted by the review boards of both the Massachusetts Institute of Technology (MIT) and Beth Israel Deaconess Medical Center (BIDMC), with an exemption from obtaining informed consent.
The authors have no funding and conflicts of interest to disclose.
The datasets generated during and/or analyzed during the current study are publicly available.
How to cite this article: Wang C, Cai Z, Xue S, Tu G. Relationship between GLR and 28-day all-cause mortality in patients with hypertensive heart disease complicated by heart failure: A retrospective analysis of the MIMIC-IV database. Medicine 2025;104:40(e44967).
Contributor Information
Chunyan Wang, Email: wangchunyan@shaphc.org.
Zhonglan Cai, Email: 294361658@qq.com.
Shengnan Xue, Email: xueshengnan1203@163.com.
References
- [1].Zhang L, He J, Wang J, et al. Knockout RAGE alleviates cardiac fibrosis through repressing endothelial-to-mesenchymal transition (EndMT) mediated by autophagy. Cell Death Dis. 2021;12:470. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Díez J, Butler J. Growing heart failure burden of hypertensive heart disease: a call to action. Hypertension. 2023;80:13–21. [DOI] [PubMed] [Google Scholar]
- [3].Méndez-Bailón M, Lorenzo-Villalba N, Jiménez-García R, et al. Clinical characteristics, management, and in-hospital mortality in patients with heart failure with reduced ejection fraction according to sex and the presence of type 2 diabetes mellitus. J Clin Med. 2022;11:1030. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Kheirbek RE, Alemi F, Fletcher R. Heart failure prognosis: comorbidities matter. J Palliat Med. 2015;18:447–52. [DOI] [PubMed] [Google Scholar]
- [5].Wang J, Deng B, Liu Q, et al. Pyroptosis and ferroptosis induced by mixed lineage kinase 3 (MLK3) signaling in cardiomyocytes are essential for myocardial fibrosis in response to pressure overload. Cell Death Dis. 2020;11:574. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Hu B, Sai W, Yuan J, et al. PGF2α-FP receptor ameliorates senescence of VSMCs in vascular remodeling by Src/PAI-1 signal pathway. Oxid Med Cell Longev. 2022;29:2908261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Tuncay A, Yilmaz Y, Baran O, Kelesoglu S. Inflammatory markers and saphenous vein graft stenosis: insights into the use of glucose-to-lymphocyte ratio as a prognostic marker. J Clin Med. 2025;14:2634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Zhang K, Zhang R, Huang M, Li B. Association between glucose-to-lymphocyte ratio and long-term mortality among women after coronary artery bypass grafting: filling gaps in female. BMC Cardiovasc Disord. 2025;25:336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Pu J, Zhang H, Wang F, et al. The glucose-to-lymphocyte ratio predicts all-cause mortality and cardiovascular mortality in ST-elevation myocardial infarction patients: a retrospective study. Rev Cardiovasc Med. 2025;26:26065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Lei J, Duan W, Yao X, et al. Association of glucose to lymphocyte ratio with the risk of death in patients with atherosclerotic cardiovascular disease. Sci Rep. 2025;15:3861. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Liu J, Hu X. Association between glucose-to-lymphocyte ratio and in-hospital mortality in acute myocardial infarction patients. PLoS One. 2023;18:e0295602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Jin L, Zheng D, Mo D, et al. Glucose-to-Lymphocyte Ratio (GLR) as a predictor of preoperative central lymph node metastasis in papillary thyroid cancer patients with type 2 diabetes mellitus and construction of the nomogram. Front Endocrinol (Lausanne). 2022;13:829009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Johnson AEW, Bulgarelli L, Shen L, et al. MIMIC-IV, a freely accessible electronic health record dataset. Sci Data. 2023;10:1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Han K, Gu J, Wang Z, et al. Association between METS-IR and prehypertension or hypertension among normoglycemia subjects in Japan: a retrospective study. Front Endocrinol (Lausanne). 2022;13:851338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Yang S, Liu Y, Wang S, Cai Z, Yang A, Hui X. Association between high serum blood glucose lymphocyte ratio and all-cause mortality in non-traumatic cerebral hemorrhage: a retrospective analysis of the MIMIC-IV database. Front Endocrinol (Lausanne). 2023;14:1290176. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Yang M, Zhang Q, Ge Y, et al. Prognostic roles of glucose to lymphocyte ratio and modified glasgow prognosis score in patients with non-small cell lung cancer. Front Nutr. 2022;9:871301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Ćorović A, Nus M, Mallat Z, Rudd JHF, Tarkin JM. PET imaging of post-infarct myocardial inflammation. Curr Cardiol Rep. 2021;23:99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Lee CJ, Ryu HY, Chun K, et al. Association of muscular fitness with rehospitalization for heart failure with reduced ejection fraction. Clin Cardiol. 2021;44:244–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Liu L, Zhang B, Li Y, et al. Preoperative glucose-to-lymphocyte ratio predicts survival in cancer. Front Endocrinol (Lausanne). 2024;15:1284152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Park S, Kang I, Hong S, et al. Glucose-to-Lymphocyte Ratio (GLR) as an independent prognostic factor in patients with resected pancreatic ductal adenocarcinoma-cohort study. Cancers (Basel). 2024;16:1844. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Chen J, Tang R, Zhan X, et al. Clinical significance of serum glucose to lymphocyte ratio as a prognostic marker in peritoneal dialysis patients. Ren Fail. 2023;45:2224893. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Hua Y, Qian J, Cao J, Wang X, Zhang W, Zhang J. Ca(2+)/Calmodulin-dependent protein kinase II regulation by inhibitor of receptor interacting protein kinase 3 alleviates necroptosis in glycation end products-induced cardiomyocytes injury. Int J Mol Sci . 2022;23:6988. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Sinha A, Bavishi A, Hibler EA, et al. Interconnected clinical and social risk factors in breast cancer and heart failure. Front Cardiovasc Med. 2022;9:847975. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Chen L, Cheng H, Kuan Y, Liang T-J, Chao Y-Y, Lin H-C. Therapeutic potential of luteolin on impaired wound healing in streptozotocin-induced rats. Biomedicines. 2021;9:761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Rahadian A, Fukuda D, Salim HM, et al. Canagliflozin prevents diabetes-induced vascular dysfunction in ApoE-deficient mice. J Atheroscler Thromb. 2020;27:1141–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Muradi A, Jasirwan COM, Simanjuntak CD, et al. The correlation of short-chain fatty acids with peripheral arterial disease in diabetes mellitus patients. Life (Basel). 2022;12:1464. [DOI] [PMC free article] [PubMed] [Google Scholar]