
==== Front
BMC Cardiovasc Disord
BMC Cardiovasc Disord
BMC Cardiovascular Disorders
1471-2261
BioMed Central London

4138
10.1186/s12872-024-04138-w
Research
The glycemic gap as a prognostic indicator in cardiogenic shock: a retrospective cohort study
Xu Qianqian
Wang Jinsheng
Lin Zhihui
Song Dongyan
http://orcid.org/0000-0003-3043-9746
Ji Kangting jikt@wmu.edu.cn

http://orcid.org/0000-0002-1911-4089
Xiang Huaqiang 1422309773@qq.com

grid.417384.d 0000 0004 1764 2632 Department of Cardiology, The Second Affiliated Hospital and Yuying Children’s Hospital, Wenzhou Medical University, Xueyuanxi Road, No 109, Wenzhou , Zhejiang, 325027 China
2 9 2024
2 9 2024
2024
24 46814 3 2024
21 8 2024
© The Author(s) 2024
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Background

Stress-induced hyperglycemia (SIH) is associated with poor outcomes in cardiogenic shock (CS), and there have been inconsistent results among patients with or without diabetes mellitus (DM). The glycemic gap (GG) is derived by subtracting A1c-derived average glucose from blood glucose levels; it is a superior indicator of SIH. We aimed to explore the role of GG in the outcomes of patients with CS.

Methods

Data on patients diagnosed with CS were extracted from the MIMIC-IV v2.0 database to investigate the relationship between GG and 30-day mortality (Number of absolute GG subjects = 359; Number of relative GG subjects = 357). CS patients from the Second Affiliated Hospital of Wenzhou Medical University were enrolled to explore the correlation between GG and lactic acid (Number of absolute GG subjects = 252; Number of relative GG subjects = 251). Multivariate analysis, propensity score-matched (PSM) analysis, inverse probability treatment weighting (IPTW), and Pearson correlation analysis were applied.

Results

Absolute GG was associated with 30-day all-cause mortality in CS patients (HRadjusted: 1.779 95% CI: 1.137–2.783; HRPSM: 1.954 95% CI: 1.186–3.220; HRIPTW: 1.634 95% CI: 1.213–2.202). The higher the absolute GG level, the higher the lactic acid level (βadjusted: 1.448 95% CI: 0.474–2.423). A similar trend existed in relative GG (HRadjusted: 1.562 95% CI: 1.003–2.432; HRPSM: 1.790 95% CI: 1.127–2.845; HRIPTW: 1.740 95% CI: 1.287–2.352; βadjusted:1.294 95% CI: 0.369–2.219). Subgroup analysis showed that the relationship existed irrespective of DM. The area under the curve of GG combined with the Glasgow Coma Scale (GCS) for 30-day all-cause mortality was higher than that of GCS (absolute GG: 0.689 vs. 0.637; relative GG: 0.688 vs. 0.633). GG was positively related to the triglyceride-glucose index. Kaplan–Meier curves revealed that groups of higher GG with DM had the worst outcomes. The outcomes differed among races and GG levels (all P < 0.05).

Conclusions

Among patients with CS, absolute and relative GGs were associated with increased 30-day all-cause mortality, regardless of DM. The relationship was stable after multivariate Cox regression analysis, PSM, and IPTW analysis. Furthermore, they reflect the severity of CS to some extent. Hyperlactatemia and insulin resistance may underlie the relationship between stress-induced hyperglycemia and poor outcomes in CS patients. They both improve the predictive efficacy of the GCS.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12872-024-04138-w.

Keywords

Glycemic gap
Cardiogenic shock
Prognosis
Propensity score-matched analysis
Inverse probability treatment weighting
issue-copyright-statement© BioMed Central Ltd., part of Springer Nature 2024
==== Body
pmcBackground

Cardiogenic shock (CS) is common in critically ill patients with cardiovascular disease; it is caused by severe impairment of myocardial performance, resulting in diminished cardiac output, life-threatening end‐organ hypoperfusion, and hypoxia [1, 2]. Although reperfusion therapy has advanced, 30-day mortality following acute myocardial infarction remains high, at about 50% from 2010 to 2017 [3]. For these reasons, risk stratification of CS patients is essential.

Stress-induced hyperglycemia (SIH) is the relative transient increase in glucose during a critical illness such as CS [4]. Numerous studies found that SIH during critical illness is related to an increased risk of morbidity and mortality [5]. For example, Shen et al. found that elevated levels of stress hyperglycemia ratio, a biomarker for SIH, are associated with increased in-hospital mortality and prolonged length of stay in stroke patients [6]. Interestingly, this conclusion has inconsistent results among patients with or without diabetes mellitus (DM). Some studies report that high glucose levels on admission were associated with increased mortality in DM and non-DM CS patients [7, 8]. However, a study showed that serum glucose levels were not associated with mortality for CS patients with DM [9]. There may be several reasons behind this phenomenon. First, the blood glucose level of patients with DM on admission may be affected by hypoglycemic drugs and cannot reflect the proper level of SIH. Second, long-term exposure to hyperglycemia downregulates glucose transportation capacity for patients with DM and protects cells from glucose toxicity [10]. HbA1c reflects the glucose level in the preceding three months. A1c-derived average glucose (ADAG, mg/dl) was calculated by the formula (28.7 × HbA1c [%])–46.7), representing the chronic glycemia level [11]. In recent years, the glycemic gap (GG), derived by ADAG from blood glucose levels, has been proven to be a superior indicator of SIH as it improves the accuracy of assessment by removing the impact of chronic hyperglycemia on the evaluation of disease severity [12, 13].

To date, the predictive value of GG has not been determined in patients with CS. Therefore, the present study aimed to investigate whether GG is associated with 30-day all-cause death in a population with CS, and whether the relationship is the same in people with and without DM. Moreover, we intended to investigate the potential mechanisms linking GG to CS. The absolute GG represents the absolute increase in glycemia. Considering the magnitude of the absolute GG might represent a different risk depending on the average plasma glucose,we also assessed the impact of the relative GG on mortality risk [14]. We hypothesize that the relationship between absolute or relative GG and CS outcomes is irrespective of DM status.

Herein, we conducted a retrospective analysis using data from the Medical Information Mart for Intensive Care-IV database version 2.0 (MIMIC-IV v2.0) to examine the prognostic significance of GG in CS patients. Hyperlacticaemia is commonly seen in shock patients; it is an independent adverse prognostic factor for CS patients [15, 16]. We intended to explore the relationship between GG and lactic acid by collecting data from the Second Affiliated Hospital of Wenzhou Medical University (Fey), which may be a linking underlie the relationship between GG and prognosis in cardiogenic shock.

Methods

Data source

In this retrospective study, we used data from MIMIC-IV v2·0, a publicly available relational database provided by the Laboratory for Computational Physiology at the Massachusetts Institute of Technology, which includes information on critical care patients who were admitted to the intensive care unit (ICU) at the Beth Israel Deaconess Medical Center from 2008 to 2019. Patient identifiers were removed according to the Health Insurance Portability and Accountability Act Safe Harbor provision [17]. Details of the MIMIC-IV database have been described elsewhere [18]. One of the authors (Xu) completed the Collaborative Institutional Training Initiative examination and had access to the database (Certificate number 40053337).

The Institutional Research and Ethics Institute of Fey approved this study involving human subjects data. The data collected was anonymous, so no informed consent was necessary. The study was conducted per the Declaration of Helsinki, and an ethical lot number of 2023-K-136–01 was assigned.

Population selection criteria

We included all patients diagnosed with CS using the International Classification of Diseases (ICD)-9/10 diagnosis codes in the MIMIC-IV v2.0 database from 2008 to 2019. The ICD-9 codes for CS are as follows: 78,551 and 99,801. The ICD-10 codes for CS are T8111, T8111XA, T8111XD, T8111XS and R570. To rule out the effects of multiple previous ICU stays, we only included patients admitted to the hospital and ICU for the first time. The target population is the adult population and thus we removed patients with age < 16. Considering that there are human factors beyond our control, we only collected patients who stayed longer than 24 h. HbA1c and blood glucose are necessary items for calculating the GG and thus we excluded patients who missed either one. To avoid the influence of outliers of GG, we have eliminated them. The method of defining outliers (GG exceeds the mean ± three standard deviations [SD]) applies to any distributed data.

For data from Fey, we included adult patients with discharge diagnoses from February 2012 to May 2023: cardiogenic shock or Killip IV. We excluded patients whose HbA1c, blood glucose, or lactic acid values were missing. We also excluded patients with outliers whose absolute/relative GG values exceeded the mean ± three SD. Patients with baseline information missing by more than 10% were also excluded.

Data extraction

First, we extracted patients with CS according to ICD-9/10 codes from MIMIC-IV using Structured Query Language (inclusion number = 2547). Then, we only included patients admitted to the hospital and ICU for the first time (exclusion number = 790). Next, we excluded patients younger than 16 (N = 0). And we excluded patients who stayed in the ICU for less than 24 h (N = 212). We also collected the baseline data on admission to the ICU on the study population, including demographic data, primary vital signs, comorbidities, basic laboratory parameters, Glasgow coma score (GCS), and some therapeutic measures. The initial measured value was used when there were multiple records for an indicator. Then, we excluded patients who lacked values of HbA1c (N = 1178) or blood glucose (N = 4). The absolute GG had four outliers. The relative GG had six outliers. Finally, 359 patients were included in the group of absolute GG (Fig. 1a), and 357 patients were included in the group of relative GG (Fig. S1a). If the missing variables were more significant than 10%, the variable would not be included. We analyzed the variables included in the MIMIC-IV v2.0 cohort. Only two albumin values were missing in the absolute GG cohort and one in the relative GG cohort. None of the other variables were missing (Table S1). Given the small amount missing, we set the missing part to null for a missing variable of less than 10%. Demographic information included age, gender, and ethnicity. Vital signs on admission included systolic blood pressure, diastolic blood pressure (DBP), mean arterial pressure, heart rate, respiratory rate, temperature, and pulse oximetry-derived oxygen saturation (SpO2). Complications included myocardial infarction, peripheral vascular disease, cerebrovascular disease, chronic pulmonary disease, DM, chronic kidney disease (CKD), and cancer. Laboratory parameters included HbA1c, blood glucose, hemoglobin, platelet, white blood cell count (WBC), international normalized ratio, prothrombin time (PT), activated partial thromboplastin time, albumin, serum urea nitrogen, and creatinine. We also collected whether insulin, vasopressin, and renal replacement treatment (RRT) were used during hospitalization. The baseline data obtained from Fey were nearly identical to the collected data from MIMIC-IV cohort.Besides,we collected variables, including ejection fraction, N-terminal pro-brain natriuretic peptide, cardiac troponin I, triglyceride, total cholesterol, high-density lipoprotein, low-density lipoprotein, lactic acid, intra-aortic balloon pump (IABP), mechanical ventilation (MV), which missed greater than 10% in MIMIC-IV database. ADAG (mg/dl) was calculated by the formula ((28.7 × HbA1c [%])–46.7) [11]. Absolute GG (mg/dl) was calculated as the first admittance of blood glucose – ADAG. Relative GG was calculated as absolute glycemic gap/ADAG × 100%.Fig. 1 Flow chart of study population of absolute glycemic gap. Notes: (a) Flow chart of study population of absolute glycemic gap from MIMIC-IV v2.0; (b) Flow chart of study population of absolute glycemic gap from the Second Affiliated Hospital of Wenzhou Medical University. Abbreviations: ICD, International Classification of Diseases; ICU, intensive care unit; IPTW, inverse probability treatment weight; MIMIC-IV v2.0, medical information mart for intensive care-IV database version 2.0; PSM, propensity score match

Endpoint and follow-up

The outcome of MIMIC-IV cohort was 30-day all-cause mortality, followed by the patient’s admission. The date of death in MIMIC-IV was obtained from Social Security Death Index records from the U.S. government. And the outcome of Fey cohort was serum lactic acid, which generated for the first time during hospitalization.

Statistical analysis

First, we conducted a smooth curve fitting to examine the relationship between CS patients’ GG and 30-day all-cause mortality. Four knots were used in the restricted cubic spline analysis (Fig. S2). We can see an evident inflection point between CS patients’ relative GG and 30-day all-cause mortality (Fig. S2). Then, we apply segmented (piece-wise) regression that uses a separate line segment to fit each interval. A log-likelihood ratio test comparing a one-line (non-segmented) model to a segmented regression model was used to determine whether a threshold exists. The inflection point that connecting the segments was based on the model gives maximum likelihood, and it was determined using the two-step recursive method. Table S2 shows that the P-value of the logarithmic likelihood ratio tests is both > 0.05, indicating a linear relationship between the GG and 30-day all-cause mortality of CS patients. The GG values were categorized into dichotomous groups, and the maximum value of the lower absolute/relative GG group was equal to 32.59 mg/dl and 24.77, respectively, in the MIMIC-IV v2.0 cohort. In the Fey cohort, the lower absolute/relative GG group’s maximum value was 50.00 mg/dl and 43.23, respectively. The following analysis set the lower absolute/relative GG level groups as the reference groups. Categorical variables were expressed by frequency (percentage), and the Chi-square test or Fisher’s exact test was used to compare the different groups. Continuous variables were represented by the mean (SD) and compared with the Kruskal–Wallis H-test or variance analysis. We established models 1–3 to investigate the association between GG and 30-day all-cause mortality among CS patients using COX’s proportional hazards regression analysis.The results were expressed as hazard ratios [HR] (95% confidence interval [95% CI]). And models 4–6 were established to evaluate the relationship between GG and lactic acid among CS patients using linear regression analysis, with the results expressed as β (95% CI).The selection of covariates was based on the estimated value of impact > 10% [19] or were considered meaningful by clinicians in our model.Model1 and model4 were adjusted for nothing. Model 2 was adjusted for age, gender, and ethnicity. For absolute GG, model 3 was adjusted for model2 + hemoglobin BUN, DM, vasopressin, SpO2, albumin, creatinine, CKD, RRT, and GCS. For relative GG, model 3 was adjusted for model2 + BUN, DM, CKD, SpO2, hemoglobin, albumin, creatinine, vasopressin, and RRT. Model 5 was adjusted for age and gender. For absolute GG, model6 was adjusted for model5 + IABP, MV, vasoactive agent, CRRT, white blood count, PT, glutamic-pyruvic transaminase (ALT), glutamic oxaloacetic transaminase (AST), uric acid, cholesterol, kidney dysfunction, hyperlipemia, temperature, and SpO2. For relative GG, model6 was adjusted for model5 + IABP, MV, vasoactive agent, continuous renal replacement therapy(CRRT), white blood count, neutrophil percentage, PT, ALT, AST, serum urea nitrogen, creatine, uric acid, cholesterol, kidney dysfunction, chronic obstructive pulmonary disease, hypertension, heart rate, DBP, SpO2. Table S3 summarized variables used in modeling. Propensity score matching (PSM) and inverse probability treatment weighting (IPTW) were also used to adjust the covariates to ensure the robustness of our findings [20, 21]. PSM was performed at a ratio of 1:1 using a caliper width of 0.05 of the SD of the logit of the propensity score. Then, we established two PSM-adjusted and IPTW-adjusted Cox proportional-hazards models separately to explore whether the relationship between GG and 30-day all-cause mortality still existed stably. Risks of 30-day all-cause mortality in different subgroups (low GG, high GG) (non-DM/low GG, non-DM/high GG, DM/low GG, DM/high GG) (White/low GG, White/high GG, Black/low GG, Black/high GG,Others/low GG, Others/high GG) were presented by Kaplan–Meier survival curves and compared by log-rank test. The subgroup analysis was performed to explore if the association differed among subgroups classified using different parameters, including age, gender, ethnicity, DM, and insulin usage. A receiver operating characteristic (ROC) curve was used to estimate the predictive value of the GCS and GCS plus absolute/relative GG for 30-day all-cause mortality in patients with cardiogenic shock. The areas under the ROC curves (AUCs) were compared using the DeLong test. The predictive capabilities of the models were further assessed using net reclassification improvement (NRI) and integrated discrimination improvement (IDI). Pearson analysis was adopted to analyze the relationship between absolute or relative GG and triglyceride-glucose index (TyG). All the analyses were conducted using R software (Version 3.6.1, http://www.r-project.org). A two-tailed P-value of < 0.05 indicated a significant difference.

Results

Characteristics of the study population

According to inclusion and exclusion criteria, 359 patients were included for analysis in the group of the absolute glycemic gap(Fig. 1a), and 357 patients were included for analysis in the group of the relative glycemic gap in MIMIC-IV v2.0 (Fig.S1a). The baseline characteristics of the study population of the absolute GG were shown in Table 1. The baseline characteristics of the study population of the relative GG are shown in Table S4. In the absolute glycemic gap cohort, 119 women and 240 men have a mean age of 66.29 ± 14.13 years and a mean GG level of 52.02 ± 87.57mg/dl. In relative glycemic gap cohort, there are 116 women and 241 men with a mean age of 66.31 ± 14.10 years and a mean GG level of 38.86 ± 57.73. According to the GG level, the eligible participants were divided into two groups (≤ 32.59 mg/dl, > 32.59 mg/dl; Table 1) (≤ 24.77, > 24.77; Table S4). The platelet and WBC levels were significantly higher in the high GG group. Moreover, they are more likely to suffer from myocardial infarction. Insulin use was also higher. The study subjects also comprised patients with CS between February 2012 and May 2023 from the Fey cohort (Fig. 1b and Fig.S1b). In the Fey cohort, 94 women and 158 men with a mean age of 70.36 ± 14.39 years and a mean GG level of 70.44 ± 91.11mg/dl were included for analysis in group of absolute glycemic gap (Table 2), 95 women and 156 men with a mean age of 70.51 ± 14.23 years, and a mean GG level of 67.50 ± 157.34 were included for analysis in group of relative glycemic gap (Table S5). According to the GG level, the eligible participants were divided into two groups (absolute GG: ≤ 50.00 mg/dl, > 50.00 mg/dl; Table 2) (relative GG: ≤ 43.23, > 43.23; Table S5). Similar to the cohort from MIMIC-IV v2.0, the level of WBC was significantly higher in the high GG group. The level of lactic acid was also higher in the high GG group. And they were likelier to use IABP and MV (all P < 0.05). Table 1 Baseline characteristics of participants based on absolute glycemic gap in MIMIC-IV v2.0 before and after PSM and IPTW

		Unadjusted(N = 359)			PSM(N = 266)			IPTW a(N = 300)			
Variables	All patients	Low (≤ 32.59)	High(> 32.59)	P-value	Low (≤ 32.59)	High(> 32.59)	P	Low (≤ 32.59)	High (> 32.59)	P-value	
N	359	179	180		133	133		147	153		
Absolute glycemic gap, mg/dl	52.02 ± 87.57	-11.18 ± 39.29	114.87 ± 76.38	 < 0.01	-11.01 ± 41.66	116.81 ± 96.65	 < 0.01	-12.53 ± 38.89	103.56 ± 74.63	 < 0.01	
Clinical parameters	
Age, years	66.29 ± 14.13	66.11 ± 15.59	66.47 ± 12.54	0.81	65.74 ± 14.60	66.63 ± 12.62	0.59	66.73 ± 15.33	66.48 ± 12.23	0.88	
Gender, n (%)				0.94			0.69			0.90	
Female	119 (33.15)	59 (32.96)	60 (33.33)		43 (32.33)	39 (29.32)		31.97%	32.64%		
Male	240 (66.85)	120 (67.04)	120(66.67)		90 (67.67)	94 (70.68)		68.03%	67.36%		
Ethnicity, n (%)				0.06			0.13			0.08	
White	218 (60.72)	109 (60.89)	109(60.56)		84 (63.15)	87 (65.41)		59.27%	63.68%		
Black	22 (6.13)	16 (8.94)	6 (3.33)		10 (7.52)	3 (2.26)		8.67%	2.65%		
Others	119 (33.15)	54 (30.17)	65 (36.11)		39 (29.32)	43 (32.33)		32.06%	33.67%		
Vital signs											
SBP, mmHg	114.16 ± 21.11	114.52 ± 20.62	113.81 ± 21.64	0.75	112.21 ± 19.74	116.16 ± 22.63	0.13	115.53 ± 20.54	114.74 ± 21.93	0.75	
DBP, mmHg	67.70 ± 18.66	67.84 ± 17.23	67.56 ± 20.02	0.88	67.25 ± 17.46	68.92 ± 20.67	0.48	69.09 ± 16.77	69.08 ± 20.01	0.99	
MBP, mmHg	83.46 ± 20.52	83.28 ± 21.49	83.63 ± 19.55	0.87	83.11 ± 22.46	84.95 ± 20.95	0.49	84.06 ± 20.55	83.76 ± 19.68	0.90	
Heart rate, beats/minute	92.00 ± 21.13	91.91 ± 22.06	92.10 ± 20.22	0.93	92.98 ± 21.54	92.50 ± 20.69	0.85	91.33 ± 19.84	92.98 ± 20.66	0.48	
Respiratory rate, times/minute	21.20 ± 6.89	20.93 ± 6.81	21.47 ± 6.97	0.46	20.93 ± 6.98	21.47 ± 6.36	0.52	21.18 ± 7.33	20.99 ± 6.64	0.81	
Temperature, ℃	36.63 ± 0.84	36.74 ± 0.65	36.53 ± 0.98	0.02	36.70 ± 0.65	36.75 ± 0.74	0.55	36.65 ± 0.76	36.62 ± 0.90	0.73	
SpO2, %	95.55 ± 6.08	96.07 ± 5.23	95.04 ± 6.79	0.11	95.98 ± 4.94	95.14 ± 5.97	0.22	95.84 ± 5.12	95.49 ± 6.07	0.59	
Lab parameters	
HbA1C, %	6.45 ± 1.67	6.46 ± 1.84	6.44 ± 1.49	0.94	6.50 ± 1.88	6.56 ± 1.58	0.78	6.60 ± 2.02	6.37 ± 1.50	0.26	
ADAG, mg/dl	138.36 ± 47.93	138.57 ± 52.83	138.16 ± 42.65	0.94	139.94 ± 53.85	141.64 ± 45.41	0.78	142.61 ± 58.06	136.00 ± 43.17	0.26	
Glucose, mg/dl	190.38 ± 95.76	127.39 ± 38.68	253.03 ± 94.55	 < 0.01	128.92 ± 40.70	258.45 ± 113.81	 < 0.01	130.08 ± 44.28	239.56 ± 94.96	 < 0.01	
Hemoglobin, g/dl	12.23 ± 2.48	11.99 ± 2.46	12.48 ± 2.48	0.06	11.98 ± 2.51	12.35 ± 2.45	0.22	12.03 ± 2.50	12.10 ± 2.54	0.82	
Platelet,109/l	236.20 ± 103.41	223.34 ± 114.06	248.99 ± 90.12	0.02	230.62 ± 121.85	249.79 ± 82.70	0.13	236.96 ± 119.03	236.58 ± 85.49	0.98	
WBC,109/l	14.06 ± 6.44	12.59 ± 6.17	15.52 ± 6.39	 < 0.01	13.44 ± 6.47	14.55 ± 5.68	0.14	13.86 ± 7.35	14.22 ± 6.20	0.65	
INR	1.51 ± 0.77	1.53 ± 0.67	1.48 ± 0.86	0.03	1.54 ± 0.69	1.52 ± 0.94	0.80	1.52 ± 0.62	1.51 ± 0.92	0.92	
PT, second	16.44 ± 8.23	16.68 ± 7.04	16.20 ± 9.28	0.03	16.88 ± 7.40	16.56 ± 10.13	0.77	16.54 ± 6.60	16.55 ± 9.85	0.99	
APTT, second	50.77 ± 34.75	44.13 ± 29.94	57.37 ± 37.90	 < 0.01	47.67 ± 33.28	49.30 ± 29.82	0.68	48.53 ± 34.02	50.30 ± 33.41	0.65	
Albumin, g/dL	3.27 ± 0.57	3.26 ± 0.53	3.28 ± 0.61	0.77	3.27 ± 0.53	3.34 ± 0.58	0.36	3.27 ± 0.54	3.26 ± 0.61	0.89	
Serum urea nitrogen, mg/dl	33.14 ± 23.22	35.07 ± 25.57	31.22 ± 20.50	0.12	32.16 ± 20.14	32.76 ± 21.61	0.82	34.0 ± 21.75	32.53 ± 20.81	0.54	
Creatinine, mg/dL	1.75 ± 1.62	1.69 ± 1.22	1.80 ± 1.95	0.54	1.62 ± 1.20	1.75 ± 1.50	0.43	1.74 ± 1.27	1.75 ± 1.69	0.95	
Scoring system	
GCS	11.49 ± 4.24	11.79 ± 3.83	11.20 ± 4.60	0.19	11.99 ± 3.76	11.57 ± 4.38	0.40	11.30 ± 4.04	11.4 ± 4.43	0.78	
Comorbidities, n (%)	
Myocardial infarction	231 (64.35)	104 (58.10)	127 (70.56)	0.01	84 (63.16)	88 (66.17)	0.70	62.04%	61.40%	0.91	
Peripheral vascular disease	56 (15.60)	36 (20.11)	20 (11.11)	0.02	22 (16.54)	18 (13.53)	0.61	16.93%	17.60%	0.88	
Cerebrovascular disease	66 (18.38)	43 (24.02)	23 (12.78)	0.01	21 (15.79)	18 (13.53)	0.73	17.30%	14.77%	0.55	
Chronic pulmonary disease	90 (25.07)	49 (27.37)	41 (22.78)	0.32	36 (27.07)	26 (19.55)	0.19	23.16%	21.12%	0.67	
DM	126 (35.10)	55 (30.73)	71 (39.44)	0.08	44 (33.08)	58 (43.61)	0.10	34.38%	34.40%	1.00	
CKD	97 (27.02)	53 (29.61)	44 (24.44)	0.27	42 (31.58)	34 (25.56)	0.34	27.31%	25.91%	0.78	
Malignant cancer	18 (5.01)	14 (7.82)	4 (2.22)	0.02	8 (6.02)	4 (3.01)	0.38	5.07%	4.71%	0.89	
Therapies, n(%)	
Vasopressin	108 (30.08)	49 (27.37)	59 (32.78)	0.26	41 (30.83)	39 (29.32)	0.89	32.58%	32.63%	0.99	
Insulin	223 (62.12)	100 (55.87)	123 (68.33)	0.02	80 (60.15)	85 (63.91)	0.61	61.42%	64.07%	0.64	
RRT	20 (5.57)	9 (5.03)	11 (6.11)	0.66	9 (6.77)	8 (6.02)	1.00	5.78%	5.88%	0.97	
Outcome, n (%)	
30-day mortality	102 (28.41)	39 (21.79)	63 (35.00)	 < 0.01	24 (18.05)	43 (32.33)	0.01	24.81%	36.10%	0.03	
aThe number of subjects changed after IPTW in the calculation; however, the actual number of subjects did not change. Continuous variables were expressed as mean ± standard deviation. Categorical variables were expressed as number(percentage). Propensity score match analysis was performed at a ratio of 1: 1 using a caliper width of 0.05 of the SD of the logit of the propensity score. The matching variables are as follows: insulin, malignant cancer, myocardial infarction, peripheral vascular disease, cerebrovascular disease, PT, APTT, INR, WBC, platelet and temperature

Abbreviations: ADAG A1C-derived average glucose, APTT activated partial thromboplastin time, CKD chronic kidney disease, DBP diastolic blood pressure, DM diabetes mellitus, GCS Glasgow coma scale, HbA1C Glycosylated Hemoglobin, Type A1CINR international normalized ratio, IPTW inverse probability treatment weight, MAP mean arterial pressure, PSM propensity score match, PT prothrombin time, RRT renal replacement treatment, SBP systolic blood pressure, SpO2 pulse oximetry-derived oxygen saturation, WBC white blood cell count

Table 2 Baseline characteristics of participants based on absolute glycemic gap in the Second Affiliated Hospital of Wenzhou Medical University

Variables	All patients	Absolute glycemic gap ≤ 50.00	Absolute glycemic gap > 50.00	P value	
N	252	126	126		
glycemic gap, mg/dl	70.44 ± 91.11	3.27 ± 36.18	137.61 ± 79.15	 < 0.01	
Clinical parameters	
Age, years	70.36 ± 14.39	70.36 ± 15.34	70.37 ± 13.45	1.00	
Gender, n (%)				0.19	
Female	94 (37.30)	42 (33.33)	52 (41.27)		
Male	158 (62.70)	84 (66.67)	74 (58.73)		
Vital signs	
Temperature, ℃	36.70 ± 0.87	36.69 ± 0.90	36.71 ± 0.85	0.83	
Heart rate, beats/minute	94.92 ± 21.42	90.92 ± 18.11	98.92 ± 23.69	 < 0.01	
Respiratory rate, times/minute	21.59 ± 6.95	21.14 ± 5.33	22.04 ± 8.25	0.31	
SBP, mmHg	120.71 ± 26.51	118.70 ± 25.38	122.72 ± 27.55	0.23	
DBP, mmHg	71.88 ± 17.94	70.92 ± 17.01	72.83 ± 18.85	0.40	
MBP, mmHg	88.17 ± 19.21	86.85 ± 18.46	89.48 ± 19.92	0.28	
SpO2, %	93.83 ± 9.59	93.40 ± 11.33	94.26 ± 7.45	0.48	
Lab parameters	
HbA1C, %	6.77 ± 1.53	6.70 ± 1.61	6.84 ± 1.44	0.47	
ADAG, mg/dL	147.71 ± 43.78	145.71 ± 46.08	149.71 ± 41.45	0.47	
Glucose, mg/dL	218.15 ± 98.64	148.97 ± 45.19	287.32 ± 88.59	 < 0.01	
Ejection fraction, %	44.94 ± 13.27	46.16 ± 13.53	43.69 ± 12.93	0.15	
WBC, 109/L	12.79 ± 5.19	11.59 ± 4.74	14.01 ± 5.35	 < 0.01	
Neutrophil percentage	0.87 ± 0.85	0.80 ± 0.11	0.93 ± 1.20	0.23	
Hemoglobin, g/L	121.09 ± 26.19	120.35 ± 26.20	121.84 ± 26.27	0.65	
Platelet, 109/L	218.94 ± 87.57	208.91 ± 75.12	229.06 ± 97.81	0.07	
PT, second	16.04 ± 4.71	15.86 ± 4.86	16.21 ± 4.58	0.56	
INR	1.49 ± 2.99	1.28 ± 0.42	1.70 ± 4.20	0.27	
APTT, second	50.45 ± 32.10	44.22 ± 20.66	57.29 ± 40.16	0.24	
ALT, U/L	198.95 ± 536.85	169.71 ± 409.32	228.19 ± 639.88	0.39	
AST, U/L	366.93 ± 865.70	353.06 ± 850.99	381.26 ± 883.94	0.80	
Albumin, g/L	35.36 ± 4.84	35.84 ± 4.65	34.89 ± 5.00	0.12	
Serum urea nitrogen, mmol/L	11.57 ± 7.81	12.82 ± 9.03	10.32 ± 6.14	0.07	
Creatine, μmol/L	157.75 ± 153.05	174.55 ± 177.70	140.95 ± 122.04	0.08	
Uric acid, μmol/L	474.54 ± 207.20	491.14 ± 224.05	458.08 ± 188.48	0.21	
Triglyceride, mmol/L	1.39 ± 0.74	1.35 ± 0.67	1.44 ± 0.79	0.35	
Cholesterol, mmol/L	4.08 ± 1.34	4.21 ± 1.31	3.96 ± 1.36	0.16	
LDL, mmol/L	2.72 ± 1.23	2.81 ± 1.21	2.63 ± 1.24	0.26	
NT-proBNP, pg /ml	8,166.82 ± 8,653.41	7,343.37 ± 8,265.60	8,875.90 ± 8,951.31	0.21	
cTnI, ng/ml	23.98 ± 92.86	22.25 ± 103.76	25.73 ± 80.86	0.01	
Etiology, n(%)				0.05	
AMI	217 (86.11)	103(81.75)	114 (90.48)		
Others	35 (13.89)	23 (18.25)	12 (9.52)		
Therapies, n(%)	
IABP	57 (22.62)	12 (9.52)	45 (35.71)	 < 0.01	
MV	171 (67.86)	73 (57.94)	98 (77.78)	 < 0.01	
CRRT	55 (21.83)	26 (20.63)	29 (23.02)	0.65	
Vasoactive agent	226 (89.68)	109 (86.51)	117 (92.86)	0.10	
Hypoglycemic agent	61 (24.21)	28 (22.22)	33 (26.19)	0.46	
Insulin	153 (60.71)	68 (53.97)	85 (67.46)	0.03	
Comorbidities, n (%)	
Kidney disfunction	156 (61.90)	80 (63.49)	76 (60.32)	0.60	
Hyperuricemia	152 (61.04)	78 (62.90)	74 (59.20)	0.55	
Hyperlipemia	67 (28.51)	40 (34.78)	27 (22.50)	0.04	
Coronary heart disease	223 (88.49)	108 (85.71)	115 (91.27)	0.17	
Peripheral vascular disease	42 (16.67)	30 (23.81)	12 (9.52)	 < 0.01	
Pulmonary infection	170 (67.46)	87 (69.05)	83 (65.87)	0.59	
COPD	11 (4.37)	6 (4.76)	5 (3.97)	0.76	
Hypertension	148 (58.73)	68 (53.97)	80 (63.49)	0.13	
Diabetes mellitus	118 (47.01)	49 (38.89)	69 (55.20)	0.01	
Outcome					
Lactic acid, mmol/L	4.28 ± 4.08	3.49 ± 3.62	5.18 ± 4.51	 < 0.01	
Abbreviations: ADAG A1c-derived average glucose, ALT glutamic-pyruvic transaminase, AMI acute myocardial infarction, APTT activated partial thromboplastin time, AST glutamic oxaloacetic transaminase, cTn-I cardiac troponin I, COPD chronic obstructive pulmonary disease, CRRT continuous renal replacement therapy, DBP diastolic blood pressure, HbA1C glycosylated hemoglobin, type A1C, IABP intra-aortic balloon pump, INR international normalized ratio, LDL low density lipoprotein, MBP mean blood pressure, MV mechanical ventilation, NT-proBNP N-terminal pro brain natriuretic peptide, PT prothrombin time, SBP systolic blood pressure,SpO2 pulse oximetry-derived oxygen saturation, WBC white blood cell count

The association between GG and mortality

During the 30-day follow-up period, 102 deaths were recorded (Table 1 and S4). Three regression models were presented after adjusting for possible confounding variables (Table 3 and S6). When the absolute glycemic gap was analyzed as a categorical variable, the HR (95% CI) of the high absolute GG level group (> 32.59 mg/dl) was 1.803 (1.209, 2.688) in model 1, compared with the low GG level group (≤ 32.59 mg/dl). After adjusting for age, gender, and ethnicity in model 2, HR (95% CI) for 30-day all-cause mortality was 1.796 (1.196, 2.697). After adjusting for age, gender, ethnicity, hemoglobin, BUN, DM, vasopressin, SpO2, albumin, creatinine, and CKD, RRT, and GCS in model 3, HR (95% CI) for 30-day all-cause mortality was 1.779 (1.137, 2.783), which was still statistically significant. When the absolute GG was analyzed as a continuous variable, the HR was more than one in all three models (Table 3). A similar correlation was observed between 30-day all-cause mortality and relative GG (HR (95% CI) were 1.623 (1.092, 2.411) in model, 1.581 (1.059, 2.362) in model 2 and 1.562 (1.003, 2.432) in model 3) (Table S6). In addition, we divided the absolute/relative GG into three groups and four groups, respectively, and three models were also established to evaluate the relationship between the GG and the outcomes of CS patients (Table 3 and S6). For 30-day all-cause mortality, the HR of the group with the highest GG level was more than one in all three models, compared with the group with the lowest GG level (P < 0.05), irrespective of GG being divided into three or four groups. Above all, both absolute and relative GG were risk prognostic factors for 30-day all-cause mortality in CS patients. Table 3 Association between absolute glycemic gap and 30-day all-cause mortality among patients with cardiogenic shock using COX’s proportional hazards regression model

Exposure	Model1		Model2		Model3		
	HR (95%CI)	P value	HR (95%CI)	P value	HR (95%CI)	P value	
Absolute glycemic gap (continuous variable, mg/dl)	1.004(1.002,1.006)	 < 0.001	1.003(1.001,1.005)	0.002	1.003(1.001,1.005)	0.008	
Absolute glycemic gap (categorical variable, mg/dl)	
1st Dichotomize	Reference		Reference		Reference		
2nd Dichotomize	1.803(1.209,2.688)	0.004	1.796(1.196,2.697)	0.005	1.779(1.137,2.783)	0.012	
Absolute glycemic gap (categorical variable, mg/dl)	
1st Tertile	Reference		Reference		Reference		
2nd Tertile	1.184(0.713,1.967)	0.513	1.212(0.726,2.023)	0.463	1.334(0.765,2.327)	0.310	
3rd Tertile	1.656(1.026,2.671)	0.039	1.678(1.029,2.736)	0.038	1.746(1.018,2.995)	0.043	
P for trend		0.036		0.004		0.043	
Absolute glycemic gap (categorical variable, mg/dl)	
1st Quartile	Reference		Reference		Reference		
2nd Quartile	0.760(0.404,1.431)	0.395	0.829(0.439,1.566)	0.563	0.934(0.461,1.890)	0.849	
3rd Quartile	1.256(0.712,2.216)	0.432	1.355(0.764,2.404)	0.299	1.479(0.796,2.747)	0.216	
4th Quartile	1.942(1.146,3.293)	0.014	1.960(1.146,3.353)	0.014	2.067(1.144,3.735)	0.016	
P for trend		0.003		0.004		0.011	
Model1: adjusted for none

Model2: adjusted for age, gender and ethnicity

Model3: adjusted for model2 + hemoglobin, BUN, DM, vasopressin, SpO2, albumin, creatinine, CKD, RRT and GCS

Abbreviations: BUN urea nitrogen, CI confidence interval, CKD chronic kidney disease, DM diabetes mellitus, GCS Glasgow coma scale, HR hazard ratio, RRT renal replacement treatment, SpO2 pulse oximetry-derived oxygen saturation

PSM and IPTW

After PSM, 133 pairs of patients are included in the absolute glycemic gap cohort, and 137 pairs of patients are included in the relative glycemic gap cohort (Fig. 1a and S1a). The baseline characteristics of patients after PSM and IPTW are shown in Table 1 and S4, respectively. There were no significant differences between the two groups. After PSM and IPTW, Cox proportional hazard regression analysis revealed that high absolute GG level was independently related to worse outcomes (Table 4)(PSM: HR: 1.954; 95% CI: 1.186–3.220; P = 0.009) (IPTW: HR: 1.634; 95% CI: 1.213–2.202; P = 0.001). A similar correlation was observed between 30-day all-cause mortality and relative GG (Table S7) (PSM: HR: 1.790; 95% CI: 1.127–2.845; P = 0.014) (IPTW: HR: 1.740; 95% CI: 1.287–2.352; P < 0.001). From the above we can know that the relationship between GG and 30-day all-cause mortality among CS patients remained stable after PSM and IPTW. Table 4 Association between absolute glycemic gap and 30-day all-cause mortality among patients with cardiogenic shock using COX’s proportional hazards regression model after PSM and IPTW

Exposure	HR	Lower limit of 95% CI	Upper limit of 95% CI	P value	
After PSM	
Absolute glycemic gap (continuous variable, mg/dl)	1.002	1.001	1.004	0.012	
Absolute glycemic gap (categorical variable, mg/dl)					
Group of low level	Reference				
Group of high level	1.954	1.186	3.220	0.009	
After IPTW	
Absolute glycemic gap (continuous variable, mg/dl)	1.003	1.001	1.005	 < 0.001	
Absolute glycemic gap (categorical variable, mg/dl)					
Group of low level	Reference				
Group of high level	1.634	1.213	2.202	0.001	
Abbreviations: CI confidence interval, HR hazard ratio, IPTW inverse probability treatment weighting, PSM propensity score matching

Kaplan–Meier analysis

The GG was significantly associated with the incidence of 30-day all-cause mortality in all participants, as shown in Fig. 2(a) and S3(a). The DM/high GG combination group had the highest risk of 30-day all-cause death compared with the other groups no matter the absolute GG (P = 0.0130) (Fig. 2b) or relative GG (P = 0.0292) (Fig. S3b). There existed statistic difference of survival analysis across populations among different level of the GG with other ethnicities (log-rank P for absolute GG = 0.0278 in Fig. 2(c); log-rank P for relative GG = 0.0025 in Fig.S3c).Fig. 2 Survival analysis. Notes: (a) Survival analysis across populations among different level of absolute glycemic gap; (b) Survival analysis across populations among different level of absolute glycemic gap with different glucose status; (c) Survival analysis across populations among different level of absolute glycemic gap with different ethnics. Abbreviations: DM, diabetes mellitus; GG, glycemic gap

Subgroup analyses

Upon analyzing subgroups stratified by various factors, including age, sex, race, glucose status (DM or non-DM) and usage of insulin or not, all interaction P-values exceeded 0.05, underscoring the stability and reliability of our findings. And the results remained stable either GG analyzed as a continuous variable (Fig. 3a and S4a) or categorical variable (Fig. 3b and S4b).Fig. 3 Subgroup analysis. Notes: (a) Subgroup analysis of the associations between absolute glycemic gap as a continuous variable and 30-day all-cause mortality among patients with cardiogenic shock; (b) Subgroup analysis of the associations between absolute glycemic gap as a categorical variable and 30-day all-cause mortality among patients with cardiogenic shock. Abbreviations: CI, confidence interval; DM, diabetes mellitus; HR, hazard ratio

ROC curve analysis

ROC curve analysis was conducted to examine further the possible predictive effect of the GG in predicting 30-day all-cause mortality of CS patients. The GCS incorporated with GG could increase the predictive capacity for 30-day all-cause mortality and increase the area under the ROC from 0.637 to 0.689 (absolute GG, Fig. 4), 0.633 to 0.688 (relative GG, Fig. S5) (P all < 0.05). The inclusion of GG into the GCS model also increased the values of NRI and IDI, all of which were statistically significant (absolute GG, Table 5) (relative GG, Table S8).Fig. 4 ROC curve analysis. Notes: Model1 only for GCS.Model2 was for GCS plus absolute glycemic gap. The area under the curve of GCS + glycemic gap was higher than that of GCS (0.689 vs. 0.637,P = 0.0221). Abbreviations: AUC, area under curve; GCS, Glasgow coma scale; ROC, receiver operating characteristic

Table 5 The discrimination ability of absolute glycemic gap plus GCS versus GCS for 30-day all-cause mortality in cardiogenic shock patients

	NRI	P value	IDI	P value	
GCS	Reference		Reference		
GCS + absolute glycemic gap	0.1299(0.0337,0.2262)	0.0082	0.1299(0.0332,0.2266)	0.0084	
Abbreviations: GCS Glasgow coma scale, IDI integrated discrimination improvement, NRI net reclassification improvement

Analysis of the relationship between the absolute or relative GG and TyG

Scatter plots revealed a positive correlation between absolute or relative GG and TyG (Fig. 5: Rab = 0.5900 in the total study population, Rab = 0.4817 in DM, Rab = 0.6965 in non-DM; Fig.S6: Rre = 0.1689 in the total study population, Rre = 0.2597 in DM, and Rre = 0.2223 in non-DM). Kaplan–Meier analysis showed that the high GG group with DM had the worst prognosis. To explore the potential mechanisms behind it, we investigated the TyG levels among different levels of glycemic gap with varying glucose statuses. And the results showed that the group with high glycemic gap with DM mellitus had the highest level of TyG (Table 6 and Table S9).Fig. 5 Pearson analysis between absolute glycemic gap and TyG. Notes: Scatter plot showed a positive correlation between absolute glycemic gap and TyG in Fig. 5. [Rab = 0.5900 in total study population (a), Rab = 0.4817 in diabetes (b), Rab = 0.6965 in non-diabetes (c)]. Abbreviations: TyG, triglyceride-glucose index

Table 6 TyG among different level of absolute glycemic gap with different glucose status

Variables	Non-DM/Low-absolute GG	Non-DM/High-absolute GG	DM/Low-absolute GG	DM/High-absolute GG	P-value	
N	124	109	55	71		
TyG	8.66 ± 0.47	9.44 ± 0.61	9.03 ± 0.89	9.71 ± 0.54	 < 0.001	
Abbreviations: DM diabetes mellitus, GG glycemic gap, TyG triglyceride-glucose index

Association between the GG and lactic acid

In the Fey cohort, we can find that serum lactic acid level was higher in groups with a higher glycemic gap either in absolute GG (Table 2) or relative GG (Table S5). After adjusting for possible confounding variables, three regression models are displayed in Table 7 and S10. When the glycemic gap was analyzed as a continuous variable, the β was more than zero in all three models. When the absolute glycemic gap was interpreted as a categorical variable, β (95% CI) of the high absolute GG level group (> 50.00 mg/dl) was 1.689 (0.679, 2.698) in model 4, compared with the low GG level group (≤ 50.00 mg/dl). After adjusting for age and gender in model 5, β (95% CI) was 1.688 (0.671, 2.705). After adjusting for age, gender, IABP, MV, vasoactive agent, CRRT, white blood count, PT, ALT, AST, uric acid, cholesterol, kidney dysfunction, hyperlipemia, temperature, SpO2 in model6, β (95% CI) was 1.448 (0.474, 2.423), which was still statistically significant. A similar correlation was observed between lactic acid and relative GG (β [95% CI] were 1.804 [0.793, 2.815] in model 4, 1.811 [0.790, 2.832] in model 5 and 1.294 [0.369, 2.219] in model 6).Consequently,we can conclude that GG reflected severity of CS patients to some extent. Table 7 Association between absolute glycemic gap and lactic acid among patients with cardiogenic shock using linear regression model

	Model4		Model5		Model6		
	β(95%CI)	P value	β(95%CI)	P value	β(95%CI)	P value	
Absolute glycemic gap (continuous variable, mg/dl)	0.013(0.007,0.018)	 < 0.001	0.013(0.007,0.019)	 < 0.001	0.011(0.006,0.017)	 < 0.001	
Absolute glycemic gap (categorical variable, mg/dl)	
1st Dichotomize	Reference		Reference		Reference		
2nd Dichotomize	1.689(0.679,2.698)	0.001	1.688(0.671,2.705)	0.001	1.448(0.474,2.423)	0.004	
Model4: adjusted for none

Model5: adjusted for age and gender

Model6: adjusted for model5 + IABP, MV, vasoactive agent, CRRT, white blood count, PT, ALT, AST, uric acid; cholesterol, kidney disfunction, hyperlipemia, temperature, SpO2

Abbreviations: ALT glutamic-pyruvic transaminase, AST glutamic oxaloacetic transaminase, CI confidence interval, CRRT continuous renal replacement therapy, IABP intra-aortic balloon pump, MV mechanical ventilation, PT prothrombin time, SpO2 pulse oximetry-derived oxygen saturation

Discussion

First, whether analyzed as a continuous or categorical variable, absolute glycemic and relative glycemic gaps were both stable and independent risk factors for increased 30-day all-cause mortality in patients with CS before and after adjusting the covariates by multivariate Cox proportional-hazards analysis. Next, the results remained stationary after PSM and IPTW analysis; subgroup analysis indicated that this conclusion has consistent results among patients with or without DM. DM patients with higher values of glycemic gap have the worst outcomes. Moreover, ROC curve analysis revealed that absolute and relative glycemic gaps can increase the predictive value of GCS in patients with cardiogenic shock. Furthermore, GG was positively associated with TyG, irrespective of DM. Finally, linear regression analysis revealed that a higher level of lactic acid accompanies a higher absolute or relative GG.

The GG has been extensively studied as an indicator of stress-induced hyperglycemia (SIH). Zhang et al. [13] found that GG was independently associated with 30-day all-cause mortality and major cardiovascular adverse events in patients with acute ST-segment elevation myocardial infarction, which was not affected by DM status. Afshin et al. [22] demonstrated that admission GG is associated with the risk of in-hospital mortality and can potentially represent a valuable prognostic biomarker for ICH patients with DM. Our study shows that absolute glycemic gap and relative glycemic gap are independent adverse prognostic factors for CS patients, whether with DM or not. Hence, the GG may be a better indicator of SIH than glucose in CS patients with DM, which was consistent with previous studies. However, Guo et al. revealed the presence of a U-shaped relationship between the GG and mortality in critically ill patients (Number of study population = 935) [23]. Jensen, Andreas Vestergaard, et al. found that the highest and lowest GG may increase the risk of 90-day mortality in patients with community-acquired pneumonia (Number of study population = 1933) [14]. This finding suggests that the relationship between GG and outcomes in patients with CS may also be non-linear. Our statistical analysis supported a linear relationship between GG and outcomes in patients with CS. This finding may be due to the small sample size of the MIMIC cohort in our study (Number of study population = 359/357). However, it is difficult for us to expand our sample size in the MIMIC-IV database. Future research should determine whether there is a non-linear relationship by expanding the sample size.

The predominant cause of SIH is the intense counterregulatory hormone and cytokine responses of critical illness [24]. In particular, sympathetic nervous system activation [25] and cytokines such as tumor necrosis factor-α (TNFα) [26] stimulate glucagon release, together with other anti-insulin hormones, including cortisol and growth hormone, leading to hyperglycemia. Furthermore, a reduction of pancreatic β-cell function and growth of insulin resistance both contribute to SIH in non-DM [27]. Additionally, SIH makes a vicious cycle by inducing an increase in oxidative stress [28], inflammation [29], endothelial cell dysfunction [30], and amplifying gluconeogenesis and insulin resistance [31]. The TyG index has been proposed as a reliable alternative marker of insulin resistance, which is superior to HOMA-IR in assessing insulin resistance in individuals with and without DM [32]. In our study, GG was positively related to the TyG index, an evaluating indicator for insulin resistance. Consequently, GG may be partly mediated by the insulin resistance conditions.Zhang et al. found that a high level of stress hyperglycemia ratio without DM has worse outcomes in CICU patients [33]. However, we found that DM patients with higher values of GG have the worst outcomes. How does one explain the discrepancy? Our findings show that the group with a high-level glycemic gap and DM had the highest level of TyG (Table 6 and S9). In our study, it is reasonable to assume that DM patients with high GG levels have higher insulin resistance. However, we cannot get the information on insulin resistance from Q. Zhang’s research. Further research is needed to confirm our hypothesis. Besides, the level of platelet and WBC were significantly higher in the high GG group, which can be explained by the conclusion that stress exposure could be a risk factor for abnormal WBC, RBC, and platelets [34, 35].

The GCS was introduced by Teasdale and Jennett [36, 37] in 1974 and 1976 as a means for serial assessment of patients with traumatic brain injury who were admitted to ICU, with the advantage of the convention. Changes in the level of consciousness may also occur in patients with cardiogenic shock. In recent years, a lower GCS score was a significant predictor for in-hospital mortality in CS patients [38]. ROC curve analysis revealed that absolute and relative glycemic gaps can increase the predictive value of GCS in patients with cardiogenic shock. We can combine these two indicators in the future practice of predicting the short-term outcomes of CS.

The first demonstration of lactic acid in human blood in shock was described by Johann Joseph Scherer (1814–1869) in January 1843. Increased lactic acid in the blood (hyperlactatemia) reflects severe illness, in which the increased blood lactate levels may result from both anaerobic and aerobic production or a decreased clearance [39]. Our study shows that a higher level of lactic acid often accompanies a higher absolute or relative GG level. There are several explanations. Firstly, stressful stimuli will likely increase lactate production from glucose in muscles [40]. A prospective observational study also indicated that hyperglycemia and increased glucose turnover substantially influence lactate metabolism in severe septic and CS [41]. Moreover, Weekers et al. reported that lactic acidosis lasted longer in the hyperglycemic group as compared with a normoglycemic group in an animal model of severe injury [42], which may result from a deterioration of cardiac output [43], mitochondrial function [44], or microcirculation [45]. Richards et al.noted that the relationship between hyperglycemia and multiple organ failure in patients with blunt trauma was mediated by lactate level [46]. A similar correlation has been seen in other settings, such as critically ill patients [47] and sepsis [48]. Hyperlactatemia is an independent indicator of outcomes in CS patients [49]. These findings suggest that GG reflects the severity of CS, and hyperlactatemia may underlie the relationship between SIH and poor outcomes in CS patients.

Unfortunately, fixed guidelines to manage SIH have yet to be studied in detail. The SIH needs more in-depth studies to construct a more organized and direct approach to diagnosing and managing such a condition. Although there are no guidelines to manage SIH yet, a study showed that the likelihood for favorable outcome was lowest among patients with undiagnosed DM compared to acute ischemic stroke patients with true non-DM in (adjusted relative risk, 0.42 [99% CI, 0.19–0.94]) [50]. Our study shows that DM CS patients with higher values of GG have the worst outcomes (Fig. 2b and S3b). HBA1c is not routinely tested in critical patients. Consequently, it is necessary to detect HbA1c for CS patients. It can not only help diagnose DM but also be used to calculate the glycemic gap.

Strengths and limitations

The present study has several strengths. To our knowledge, we are the first study to evaluate the predictive value of the GG in patients with CS, with subgroup analysis according to the status of DM. Additionally, we used three methods (Cox proportional-hazards analysis, PSM, and IPTW) to adjust for confounding factors, which made our conclusions more credible. However, our study has some limitations. One is that the GG in the study was calculated from random blood glucose. It took a lot of work for us to eliminate the confounding effect of food in observational studies, which may affect the accuracy of the results. Consequently, future research should use fasting glucose rather than admission random glucose. Furthermore, this was an observational retrospective study. Although we have conducted PSM and IPTW, residual confounding is still possible. In addition, a reduction in the sample size after PSM analysis may reduce the study’s statistical power and lead to instability in the results. Ethnic disparities between HbA1c and mean glucose levels have been shown in several studies [51]. We found that the outcomes of GG differ across races. We’ve only done preliminary validation in Chinese populations. Further research on diverse populations should be conducted in the future. The fluctuation of blood glucose can change the GG in the short term. It is one-sided to evaluate the impact of the GG on outcomes in CS patients using a single value. And whether hypoglycemic agents will affect the relationship between GG and outcomes by controlling the blood glucose still remains unknown. Consequently, dynamic blood glucose and HbA1c monitoring and studies of hypoglycemic agents influence are also necessary in the future. Based on monitoring the GG, we can study the influence of the change of GG on the outcomes to guide the treatment. In addition, we only investigate the impact on short-term outcomes, whether it affects long-term outcomes remains unknown. Our results need to be verified by prospective studies, and more research should be done.

Conclusions

Among patients with cardiogenic shock, absolute glycemic gap and relative glycemic gap, whether analyzed as a continuous variable or a categorical variable, were both associated with increased 30-day all-cause mortality, regardless of DM. The relationship was stable after multivariate Cox regression analysis, PSM, and IPTW analysis. They reflect the severity of CS to some extent. Hyperlactatemia and insulin resistance may underlie the relationship between SIH and poor outcomes in CS patients. Additionally, they can both improve the prognostic efficacy of GCS. It can aid in risk stratification and shape future therapeutic strategies in CS patients.

Supplementary Information

Supplementary Material 1.

Supplementary Material 2.

Supplementary Material 3.

Supplementary Material 4.

Supplementary Material 5.

Supplementary Material 6.

Supplementary Material 7.

Supplementary Material 8.

Supplementary Material 9.

Supplementary Material 10.

Supplementary Material 11.

Supplementary Material 12.

Supplementary Material 13.

Supplementary Material 14.

Supplementary Material 15.

Supplementary Material 16.

Abbreviations

SIH stress-induced hyperglycemia

CS cardiogenic shock

GG glycemic gap

PSM propensity score-matched

IPTW inverse probability treatment weighting

TyG triglyceride-glucose index

DM diabetes mellitus

ADAG A1c-derived average glucose

MIMIC-IV v2.0 Medical Information Mart for Intensive Care-IV database version 2.0

Fey The Second Affiliated Hospital of Wenzhou Medical University

ICU intensive care unit

CITI Collaborative Institutional Training Initiative

ICD International Classification of Diseases

SD standard deviation

GCS Glasgow coma score

SBP systolic blood pressure

DBP diastolic blood pressure

MAP mean arterial pressure

SpO2 pulse oximetry-derived oxygen saturation

CKD chronic kidney disease

WBC white blood cell

INR international normalized ratio

PT prothrombin time

APTT activated partial thromboplastin time

RRT renal replacement treatment

EF ejection fraction

NT-proBNP N-terminal pro-brain natriuretic peptide

cTn-I cardiac troponin I

IABP intra-aortic balloon pump

MV mechanical ventilation

HR hazard ratios

CI confidence interval

ALT glutamic-pyruvic transaminase

AST glutamic oxaloacetic transaminase

CRRT continuous renal replacement therapy

COPD chronic obstructive pulmonary disease

MACEs major cardiovascular adverse events

STEMI acute ST-segment elevation myocardial infarction

TNF-α tumor necrosis factor-α

Acknowledgements

Not applicable.

Authors’ contributions

Conceptualization, QX, JW, KJ and HX; Data curation, ZL and DS; Formal analysis, QX and JW; Funding acquisition, KJ and HX; Investigation, QX and JW; Methodology, QX and JW; Project administration, KJ and HX; Resources, ZL and DS; Software, ZL and DS; Supervision, KJ and HX; Validation, QX and JW; Visualization, ZL and DS; Writing–original draft, QX and JW; Writing–review & editing, QX and HX. All authors read and approved the final manuscript.

Funding

This work was supported by the Scientific Research Foundation of the Science and Technology Department of Wenzhou City under grant (No.2022Y0566) and the Clinical Research Foundation of the 2nd Affiliated Hospital of Wenzhou Medical University under grant (SAHoWMU-CR2019-01–117).

Availability of data and materials

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Research and Ethics Institute of the Second Affiliated Hospital of Wenzhou Medical University (protocol code 2023-K-136–01). Our research used data obtained in previous clinical trials. And we didn’t use data that the patient explicitly refused to utilize. Besides, the data was anonymous. Consequently, we applied for exemption of informed consent to the Institutional Research and Ethics Institute of the Second Affiliated Hospital of Wenzhou Medical University. And the organization approved our application. The ethics committee approval letter of this study was attached below. And you can see the items reviewed included a informed consent waiver application form.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Qianqian Xu and Jinsheng Wang contributed equally to this work.
==== Refs
References

1. Reynolds H Hochman J Cardiogenic shock: current concepts and improving outcomes Circulation 2008 117 5 686 697 10.1161/CIRCULATIONAHA.106.613596 18250279
Reynolds H, Hochman J. Cardiogenic shock: current concepts and improving outcomes. Circulation. 2008;117(5):686–97.18250279 10.1161/CIRCULATIONAHA.106.613596
2. Thiele H Zeymer U Neumann F Ferenc M Olbrich H Hausleiter J Intraaortic balloon support for myocardial infarction with cardiogenic shock N Engl J Med 2012 367 14 1287 1296 10.1056/NEJMoa1208410 22920912
Thiele H, Zeymer U, Neumann F, Ferenc M, Olbrich H, Hausleiter J, et al. Intraaortic balloon support for myocardial infarction with cardiogenic shock. N Engl J Med. 2012;367(14):1287–96.22920912 10.1056/NEJMoa1208410
3. Helgestad O Josiassen J Hassager C Jensen L Holmvang L Sørensen A Temporal trends in incidence and patient characteristics in cardiogenic shock following acute myocardial infarction from 2010 to 2017: a Danish cohort study Eur J Heart Fail 2019 21 11 1370 1378 10.1002/ejhf.1566 31339222
Helgestad O, Josiassen J, Hassager C, Jensen L, Holmvang L, Sørensen A, et al. Temporal trends in incidence and patient characteristics in cardiogenic shock following acute myocardial infarction from 2010 to 2017: a Danish cohort study. Eur J Heart Fail. 2019;21(11):1370–8.31339222 10.1002/ejhf.1566
4. Ye X Cai X Nie D Chen Y Li J Xu X Stress-Induced Hyperglycemia and Remote Diffusion-Weighted Imaging Lesions in Primary Intracerebral Hemorrhage Neurocrit Care 2020 32 2 427 436 10.1007/s12028-019-00747-y 31313140
Ye X, Cai X, Nie D, Chen Y, Li J, Xu X, et al. Stress-Induced Hyperglycemia and Remote Diffusion-Weighted Imaging Lesions in Primary Intracerebral Hemorrhage. Neurocrit Care. 2020;32(2):427–36.31313140 10.1007/s12028-019-00747-y
5. Vanhorebeek I Langouche L Van den Berghe G Tight blood glucose control with insulin in the ICU: facts and controversies Chest 2007 132 1 268 278 10.1378/chest.06-3121 17625087
Vanhorebeek I, Langouche L, Van den Berghe G. Tight blood glucose control with insulin in the ICU: facts and controversies. Chest. 2007;132(1):268–78.17625087 10.1378/chest.06-3121
6. Di S Xintian C Qing Z Mulalibieke H Junli H Shuaiwei S Increased stress hyperglycemia ratio at hospital admission in stroke patients are associated with increased in-hospital mortality and length of stay Diabetol Metab Syndr. 2024 16 1 69 10.1186/s13098-024-01303-1 38491539
Di S, Xintian C, Qing Z, Mulalibieke H, Junli H, Shuaiwei S, et al. Increased stress hyperglycemia ratio at hospital admission in stroke patients are associated with increased in-hospital mortality and length of stay. Diabetol Metab Syndr. 2024;16(1):69.38491539 10.1186/s13098-024-01303-1
7. Abdin A Pöss J Fuernau G Ouarrak T Desch S Eitel I Prognostic impact of baseline glucose levels in acute myocardial infarction complicated by cardiogenic shock—a substudy of the IABP-SHOCK II-trial [corrected] Clinical research in cardiology : official journal of the German Cardiac Society 2018 107 6 517 523 10.1007/s00392-018-1213-7 29423774
Abdin A, Pöss J, Fuernau G, Ouarrak T, Desch S, Eitel I, et al. Prognostic impact of baseline glucose levels in acute myocardial infarction complicated by cardiogenic shock—a substudy of the IABP-SHOCK II-trial [corrected]. Clinical research in cardiology : official journal of the German Cardiac Society. 2018;107(6):517–23.29423774 10.1007/s00392-018-1213-7
8. Thoegersen M Josiassen J Helgestad O Berg Ravn H Schmidt H Holmvang L The association of diabetes and admission blood glucose with 30-day mortality in patients with acute myocardial infarction complicated by cardiogenic shock Eur Heart J Acute Cardiovasc Care 2020 9 6 626 635 10.1177/2048872620925265 32450719
Thoegersen M, Josiassen J, Helgestad O, Berg Ravn H, Schmidt H, Holmvang L, et al. The association of diabetes and admission blood glucose with 30-day mortality in patients with acute myocardial infarction complicated by cardiogenic shock. Eur Heart J Acute Cardiovasc Care. 2020;9(6):626–35.32450719 10.1177/2048872620925265
9. Choi S Yoon G Lee M Park S Ko Y Ahn C Prognostic Impact of Plasma Glucose on Patients With Cardiogenic Shock With or Without Diabetes Mellitus from the SMART RESCUE Trial Am J Cardiol 2022 175 145 151 10.1016/j.amjcard.2022.04.008 35550823
Choi S, Yoon G, Lee M, Park S, Ko Y, Ahn C, et al. Prognostic Impact of Plasma Glucose on Patients With Cardiogenic Shock With or Without Diabetes Mellitus from the SMART RESCUE Trial. Am J Cardiol. 2022;175:145–51.35550823 10.1016/j.amjcard.2022.04.008
10. Yang J Song P Song Y Hahn J Choi S Choi J Prognostic value of admission blood glucose level in patients with and without diabetes mellitus who sustain ST segment elevation myocardial infarction complicated by cardiogenic shock Critical care (London, England) 2013 17 5 R218 10.1186/cc13035 24090250
Yang J, Song P, Song Y, Hahn J, Choi S, Choi J, et al. Prognostic value of admission blood glucose level in patients with and without diabetes mellitus who sustain ST segment elevation myocardial infarction complicated by cardiogenic shock. Critical care (London, England). 2013;17(5):R218.24090250 10.1186/cc13035
11. Nathan D Kuenen J Borg R Zheng H Schoenfeld D Heine R Translating the A1C assay into estimated average glucose values Diabetes Care 2008 31 8 1473 1478 10.2337/dc08-0545 18540046
Nathan D, Kuenen J, Borg R, Zheng H, Schoenfeld D, Heine R. Translating the A1C assay into estimated average glucose values. Diabetes Care. 2008;31(8):1473–8.18540046 10.2337/dc08-0545
12. Sun P Poblete R Nguyen P Bulic S Kim-Tenser M Marehbian J Association Between Glycemic Gap and In-hospital Outcomes in Aneurysmal Subarachnoid Hemorrhage Front Neurol 2021 12 714341 10.3389/fneur.2021.714341 34887824
Sun P, Poblete R, Nguyen P, Bulic S, Kim-Tenser M, Marehbian J, et al. Association Between Glycemic Gap and In-hospital Outcomes in Aneurysmal Subarachnoid Hemorrhage. Front Neurol. 2021;12: 714341.34887824 10.3389/fneur.2021.714341
13. Wu S Yang Y Zhu J Xu W Wang L Lyu S Impact of glycemic gap on 30-day adverse outcomes in patients with acute ST-segment elevation myocardial infarction Atherosclerosis 2022 360 34 41 10.1016/j.atherosclerosis.2022.10.003 36265428
Wu S, Yang Y, Zhu J, Xu W, Wang L, Lyu S, et al. Impact of glycemic gap on 30-day adverse outcomes in patients with acute ST-segment elevation myocardial infarction. Atherosclerosis. 2022;360:34–41.36265428 10.1016/j.atherosclerosis.2022.10.003
14. Jensen A Baunbæk Egelund G Bang Andersen S Petersen P Benfield T Witzenrath M The Glycemic Gap and 90-Day Mortality in Community-acquired Pneumonia. A Prospective Cohort Study Ann Am Thoracic Soc. 2019 16 12 1518 26 10.1513/AnnalsATS.201901-007OC
Jensen A, Baunbæk Egelund G, Bang Andersen S, Petersen P, Benfield T, Witzenrath M, et al. The Glycemic Gap and 90-Day Mortality in Community-acquired Pneumonia. A Prospective Cohort Study. Ann Am Thoracic Soc. 2019;16(12):1518–26.10.1513/AnnalsATS.201901-007OC
15. Wei-Chieh L Chih-Yuan F Huang-Chung C Chien-Jen C Cheng-Hsu Y Chi-Ling H Associations with 30-day survival following extracorporeal membrane oxygenation in patients with acute ST segment elevation myocardial infarction and profound cardiogenic shock Heart Lung. 2016 45 6 532 537 10.1016/j.hrtlng.2016.08.006 27601212
Wei-Chieh L, Chih-Yuan F, Huang-Chung C, Chien-Jen C, Cheng-Hsu Y, Chi-Ling H, et al. Associations with 30-day survival following extracorporeal membrane oxygenation in patients with acute ST segment elevation myocardial infarction and profound cardiogenic shock. Heart Lung. 2016;45(6):532–7.27601212 10.1016/j.hrtlng.2016.08.006
16. Fernando Luís S Daniel S Débora Vacaro F Miguel G Marciane Maria R Marcely Gimenes B Association between serum lactate levels and mortality in patients with cardiogenic shock receiving mechanical circulatory support: a multicenter retrospective cohort study BMC Cardiovasc Disord. 2020 20 1 496 10.1186/s12872-020-01785-7 33234107
Fernando Luís S, Daniel S, Débora Vacaro F, Miguel G, Marciane Maria R, Marcely Gimenes B, et al. Association between serum lactate levels and mortality in patients with cardiogenic shock receiving mechanical circulatory support: a multicenter retrospective cohort study. BMC Cardiovasc Disord. 2020;20(1):496.33234107 10.1186/s12872-020-01785-7
17. Brittany K, Steven M M. Twenty Years of the Health Insurance Portability and Accountability Act Safe Harbor Provision: Unsolved Challenges and Ways Forward. JMIR Med Inform. 2022;10(8):e37756.
18. Goldberger A Amaral L Glass L Hausdorff J Ivanov P Mark R PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals Circulation 2000 101 23 E215 E220 10.1161/01.CIR.101.23.e215 10851218
Goldberger A, Amaral L, Glass L, Hausdorff J, Ivanov P, Mark R, et al. PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals. Circulation. 2000;101(23):E215–20.10851218 10.1161/01.CIR.101.23.e215
19. Jaddoe V de Jonge L Hofman A Franco O Steegers E Gaillard R First trimester fetal growth restriction and cardiovascular risk factors in school age children: population based cohort study BMJ (Clinical research ed). 2014 348 g14 24458585
Jaddoe V, de Jonge L, Hofman A, Franco O, Steegers E, Gaillard R. First trimester fetal growth restriction and cardiovascular risk factors in school age children: population based cohort study. BMJ (Clinical research ed). 2014;348:g14.24458585
20. Zhang Z Propensity score method: a non-parametric technique to reduce model dependence Annals of translational medicine 2017 5 1 7 10.21037/atm.2016.08.57 28164092
Zhang Z. Propensity score method: a non-parametric technique to reduce model dependence. Annals of translational medicine. 2017;5(1):7.28164092 10.21037/atm.2016.08.57
21. Grafféo N Latouche A Le Tourneau C Chevret S ipcwswitch: An R package for inverse probability of censoring weighting with an application to switches in clinical trials Comput Biol Med. 2019 111 103339 10.1016/j.compbiomed.2019.103339 31442762
Grafféo N, Latouche A, Le Tourneau C, Chevret S. ipcwswitch: An R package for inverse probability of censoring weighting with an application to switches in clinical trials. Comput Biol Med. 2019;111:103339.31442762 10.1016/j.compbiomed.2019.103339
22. Zarean E Lattanzi S Looha M Napoli M Chou S Jafarli A Glycemic Gap Predicts in-Hospital Mortality in Diabetic Patients with Intracerebral Hemorrhage Journal of stroke and cerebrovascular diseases : the official journal of National Stroke Association. 2021 30 5 105669 10.1016/j.jstrokecerebrovasdis.2021.105669 33636475
Zarean E, Lattanzi S, Looha M, Napoli M, Chou S, Jafarli A, et al. Glycemic Gap Predicts in-Hospital Mortality in Diabetic Patients with Intracerebral Hemorrhage. Journal of stroke and cerebrovascular diseases : the official journal of National Stroke Association. 2021;30(5):105669.33636475 10.1016/j.jstrokecerebrovasdis.2021.105669
23. Guo J Chou R Kuo C Chao T Wu C Tsai Y The paradox of the glycemic gap: Does relative hypoglycemia exist in critically ill patients? Clinical nutrition (Edinburgh, Scotland) 2021 40 7 4654 4661 10.1016/j.clnu.2021.06.004 34229272
Guo J, Chou R, Kuo C, Chao T, Wu C, Tsai Y, et al. The paradox of the glycemic gap: Does relative hypoglycemia exist in critically ill patients? Clinical nutrition (Edinburgh, Scotland). 2021;40(7):4654–61.34229272 10.1016/j.clnu.2021.06.004
24. McCowen K Malhotra A Bistrian B Stress-induced hyperglycemia Crit Care Clin 2001 17 1 107 124 10.1016/S0749-0704(05)70154-8 11219223
McCowen K, Malhotra A, Bistrian B. Stress-induced hyperglycemia. Crit Care Clin. 2001;17(1):107–24.11219223 10.1016/S0749-0704(05)70154-8
25. Dünser M Hasibeder W Sympathetic overstimulation during critical illness: adverse effects of adrenergic stress J Intensive Care Med 2009 24 5 293 316 10.1177/0885066609340519 19703817
Dünser M, Hasibeder W. Sympathetic overstimulation during critical illness: adverse effects of adrenergic stress. J Intensive Care Med. 2009;24(5):293–316.19703817 10.1177/0885066609340519
26. Blumberg D Hochwald S Burt M Donner D Brennan M Tumor necrosis factor alpha stimulates gluconeogenesis from alanine in vivo J Surg Oncol. 1995 59 4 220 4 10.1002/jso.2930590404 7630167
Blumberg D, Hochwald S, Burt M, Donner D, Brennan M. Tumor necrosis factor alpha stimulates gluconeogenesis from alanine in vivo. J Surg Oncol. 1995;59(4):220–4 discussion 4–5.7630167 10.1002/jso.2930590404
27. Kruyt N van Westerloo D DeVries J Stress-induced hyperglycemia in healthy bungee jumpers without diabetes due to decreased pancreatic β-cell function and increased insulin resistance Diabetes Technol Ther 2012 14 4 311 314 10.1089/dia.2011.0171 22283617
Kruyt N, van Westerloo D, DeVries J. Stress-induced hyperglycemia in healthy bungee jumpers without diabetes due to decreased pancreatic β-cell function and increased insulin resistance. Diabetes Technol Ther. 2012;14(4):311–4.22283617 10.1089/dia.2011.0171
28. Papachristoforou E Lambadiari V Maratou E Makrilakis K Association of Glycemic Indices (Hyperglycemia, Glucose Variability, and Hypoglycemia) with Oxidative Stress and Diabetic Complications J Diabetes Res 2020 2020 7489795 10.1155/2020/7489795 33123598
Papachristoforou E, Lambadiari V, Maratou E, Makrilakis K. Association of Glycemic Indices (Hyperglycemia, Glucose Variability, and Hypoglycemia) with Oxidative Stress and Diabetic Complications. J Diabetes Res. 2020;2020:7489795.33123598 10.1155/2020/7489795
29. Ray K Cannon C Morrow D Kirtane A Buros J Rifai N Synergistic relationship between hyperglycaemia and inflammation with respect to clinical outcomes in non-ST-elevation acute coronary syndromes: analyses from OPUS-TIMI 16 and TACTICS-TIMI 18 Eur Heart J 2007 28 7 806 813 10.1093/eurheartj/ehm010 17403721
Ray K, Cannon C, Morrow D, Kirtane A, Buros J, Rifai N, et al. Synergistic relationship between hyperglycaemia and inflammation with respect to clinical outcomes in non-ST-elevation acute coronary syndromes: analyses from OPUS-TIMI 16 and TACTICS-TIMI 18. Eur Heart J. 2007;28(7):806–13.17403721 10.1093/eurheartj/ehm010
30. Qiu C Sui J Zhang Q Wei P Wang P Fu Q Relationship of Endothelial Cell-Specific Molecule 1 Level in Stress Hyperglycemia Patients With Acute ST-Segment Elevation Myocardial Infarction: A Pilot Study Angiology 2016 67 9 829 834 10.1177/0003319715621996 26685180
Qiu C, Sui J, Zhang Q, Wei P, Wang P, Fu Q. Relationship of Endothelial Cell-Specific Molecule 1 Level in Stress Hyperglycemia Patients With Acute ST-Segment Elevation Myocardial Infarction: A Pilot Study. Angiology. 2016;67(9):829–34.26685180 10.1177/0003319715621996
31. Barth E Albuszies G Baumgart K Matejovic M Wachter U Vogt J Glucose metabolism and catecholamines Crit Care Med 2007 35 S508 S518 10.1097/01.CCM.0000278047.06965.20 17713401
Barth E, Albuszies G, Baumgart K, Matejovic M, Wachter U, Vogt J, et al. Glucose metabolism and catecholamines. Crit Care Med. 2007;35:S508–18.17713401 10.1097/01.CCM.0000278047.06965.20
32. Simental-Mendía L Rodríguez-Morán M Guerrero-Romero F The product of fasting glucose and triglycerides as surrogate for identifying insulin resistance in apparently healthy subjects Metab Syndr Relat Disord 2008 6 4 299 304 10.1089/met.2008.0034 19067533
Simental-Mendía L, Rodríguez-Morán M, Guerrero-Romero F. The product of fasting glucose and triglycerides as surrogate for identifying insulin resistance in apparently healthy subjects. Metab Syndr Relat Disord. 2008;6(4):299–304.19067533 10.1089/met.2008.0034
33. Qianqian Z, Huaqiang X, Qianqian X, Zesong H, Yangjing X, Jie W, et al. Stress hyperglycemia ratio linked to worse prognosis in Cardiac Intensive Care Unit patients: A retrospective observational study. Diabetes Res Clin Pract. 2024;209(0):111598.
34. Gharibi V Khanjani N Heidari H Ebrahimi M Hosseinabadi M The effect of heat stress on hematological parameters and oxidative stress among bakery workers Toxicol Ind Health 2020 36 1 1 10 10.1177/0748233719899824 31934822
Gharibi V, Khanjani N, Heidari H, Ebrahimi M, Hosseinabadi M. The effect of heat stress on hematological parameters and oxidative stress among bakery workers. Toxicol Ind Health. 2020;36(1):1–10.31934822 10.1177/0748233719899824
35. Kario K McEwen B Pickering T Disasters and the heart: a review of the effects of earthquake-induced stress on cardiovascular disease Hypertension research : official journal of the Japanese Society of Hypertension 2003 26 5 355 367 10.1291/hypres.26.355 12887126
Kario K, McEwen B, Pickering T. Disasters and the heart: a review of the effects of earthquake-induced stress on cardiovascular disease. Hypertension research : official journal of the Japanese Society of Hypertension. 2003;26(5):355–67.12887126 10.1291/hypres.26.355
36. Teasdale G Jennett B Assessment of coma and impaired consciousness. A practical scale Lancet (London, England). 1974 2 7872 81 4 10.1016/S0140-6736(74)91639-0 4136544
Teasdale G, Jennett B. Assessment of coma and impaired consciousness. A practical scale. Lancet (London, England). 1974;2(7872):81–4.4136544 10.1016/S0140-6736(74)91639-0
37. Teasdale G Jennett B Assessment and prognosis of coma after head injury Acta Neurochir 1976 34 45 55 10.1007/BF01405862 961490
Teasdale G, Jennett B. Assessment and prognosis of coma after head injury. Acta Neurochir. 1976;34:45–55.961490 10.1007/BF01405862
38. Kim D Na S Cho Y Chung C Jeon K Suh G Predictors of Survival to Discharge After Successful Weaning From Venoarterial Extracorporeal Membrane Oxygenation in Patients With Cardiogenic Shock Circ J 2020 84 12 2205 2211 10.1253/circj.CJ-20-0550 33041291
Kim D, Na S, Cho Y, Chung C, Jeon K, Suh G, et al. Predictors of Survival to Discharge After Successful Weaning From Venoarterial Extracorporeal Membrane Oxygenation in Patients With Cardiogenic Shock. Circ J. 2020;84(12):2205–11.33041291 10.1253/circj.CJ-20-0550
39. Kompanje E Jansen T van der Hoven B Bakker J The first demonstration of lactic acid in human blood in shock by Johann Joseph Scherer (1814–1869) in January 1843 Intensive Care Med 2007 33 11 1967 1971 10.1007/s00134-007-0788-7 17661014
Kompanje E, Jansen T, van der Hoven B, Bakker J. The first demonstration of lactic acid in human blood in shock by Johann Joseph Scherer (1814–1869) in January 1843. Intensive Care Med. 2007;33(11):1967–71.17661014 10.1007/s00134-007-0788-7
40. Tayek J Katz J Glucose production, recycling, Cori cycle, and gluconeogenesis in humans: relationship to serum cortisol Am J Physiol 1997 272 E476 E484 9124555
Tayek J, Katz J. Glucose production, recycling, Cori cycle, and gluconeogenesis in humans: relationship to serum cortisol. Am J Physiol. 1997;272:E476–84.9124555
41. Revelly J Tappy L Martinez A Bollmann M Cayeux M Berger M Lactate and glucose metabolism in severe sepsis and cardiogenic shock Crit Care Med 2005 33 10 2235 2240 10.1097/01.CCM.0000181525.99295.8F 16215376
Revelly J, Tappy L, Martinez A, Bollmann M, Cayeux M, Berger M, et al. Lactate and glucose metabolism in severe sepsis and cardiogenic shock. Crit Care Med. 2005;33(10):2235–40.16215376 10.1097/01.CCM.0000181525.99295.8F
42. Weekers F Giulietti A Michalaki M Coopmans W Van Herck E Mathieu C Metabolic, endocrine, and immune effects of stress hyperglycemia in a rabbit model of prolonged critical illness Endocrinology 2003 144 12 5329 5338 10.1210/en.2003-0697 12960028
Weekers F, Giulietti A, Michalaki M, Coopmans W, Van Herck E, Mathieu C, et al. Metabolic, endocrine, and immune effects of stress hyperglycemia in a rabbit model of prolonged critical illness. Endocrinology. 2003;144(12):5329–38.12960028 10.1210/en.2003-0697
43. Sodi-Pallares D Ponce de León J Bisteni A Medrano G Potassium, glucose, and insulin in myocardial infarction Lancet (London, England). 1969 1 7609 1315 6 10.1016/S0140-6736(69)92251-X 4182204
Sodi-Pallares D, Ponce de León J, Bisteni A, Medrano G. Potassium, glucose, and insulin in myocardial infarction. Lancet (London, England). 1969;1(7609):1315–6.4182204 10.1016/S0140-6736(69)92251-X
44. Nishikawa T Edelstein D Du X Yamagishi S Matsumura T Kaneda Y Normalizing mitochondrial superoxide production blocks three pathways of hyperglycaemic damage Nature 2000 404 6779 787 790 10.1038/35008121 10783895
Nishikawa T, Edelstein D, Du X, Yamagishi S, Matsumura T, Kaneda Y, et al. Normalizing mitochondrial superoxide production blocks three pathways of hyperglycaemic damage. Nature. 2000;404(6779):787–90.10783895 10.1038/35008121
45. Guerci B Böhme P Kearney-Schwartz A Zannad F Drouin P Endothelial dysfunction and type 2 diabetes. Part 2: altered endothelial function and the effects of treatments in type 2 diabetes mellitus Diabet metab. 2001 27 436 47
Guerci B, Böhme P, Kearney-Schwartz A, Zannad F, Drouin P. Endothelial dysfunction and type 2 diabetes. Part 2: altered endothelial function and the effects of treatments in type 2 diabetes mellitus. Diabet metab. 2001;27:436–47.
46. Richards J Scalea T Mazzeffi M Rock P Galvagno S Does Lactate Affect the Association of Early Hyperglycemia and Multiple Organ Failure in Severely Injured Blunt Trauma Patients? Anesth Analg 2018 126 3 904 910 10.1213/ANE.0000000000002626 29283920
Richards J, Scalea T, Mazzeffi M, Rock P, Galvagno S. Does Lactate Affect the Association of Early Hyperglycemia and Multiple Organ Failure in Severely Injured Blunt Trauma Patients? Anesth Analg. 2018;126(3):904–10.29283920 10.1213/ANE.0000000000002626
47. Kaukonen K Bailey M Egi M Orford N Glassford N Marik P Stress hyperlactatemia modifies the relationship between stress hyperglycemia and outcome: a retrospective observational study Crit Care Med 2014 42 6 1379 1385 10.1097/CCM.0000000000000214 24561567
Kaukonen K, Bailey M, Egi M, Orford N, Glassford N, Marik P, et al. Stress hyperlactatemia modifies the relationship between stress hyperglycemia and outcome: a retrospective observational study. Crit Care Med. 2014;42(6):1379–85.24561567 10.1097/CCM.0000000000000214
48. Green J Berger T Garg N Horeczko T Suarez A Radeos M Hyperlactatemia affects the association of hyperglycemia with mortality in nondiabetic adults with sepsis Acad Emerg Med Off J Soc Acad Emerg Med 2012 19 11 1268 1275 10.1111/acem.12015
Green J, Berger T, Garg N, Horeczko T, Suarez A, Radeos M, et al. Hyperlactatemia affects the association of hyperglycemia with mortality in nondiabetic adults with sepsis. Acad Emerg Med Off J Soc Acad Emerg Med. 2012;19(11):1268–75.10.1111/acem.12015
49. Valente S Lazzeri C Vecchio S Giglioli C Margheri M Bernardo P Predictors of in-hospital mortality after percutaneous coronary intervention for cardiogenic shock Int J Cardiol 2007 114 2 176 182 10.1016/j.ijcard.2006.01.024 16737749
Valente S, Lazzeri C, Vecchio S, Giglioli C, Margheri M, Bernardo P, et al. Predictors of in-hospital mortality after percutaneous coronary intervention for cardiogenic shock. Int J Cardiol. 2007;114(2):176–82.16737749 10.1016/j.ijcard.2006.01.024
50. Torbey M Pauls Q Gentile N Falciglia M Meurer W Pettigrew C Intensive Versus Standard Treatment of Hyperglycemia in Acute Ischemic Stroke Patient: A Randomized Clinical Trial Subgroups Analysis Stroke 2022 53 5 1510 1515 10.1161/STROKEAHA.120.033048 35331007
Torbey M, Pauls Q, Gentile N, Falciglia M, Meurer W, Pettigrew C, et al. Intensive Versus Standard Treatment of Hyperglycemia in Acute Ischemic Stroke Patient: A Randomized Clinical Trial Subgroups Analysis. Stroke. 2022;53(5):1510–5.35331007 10.1161/STROKEAHA.120.033048
51. William HH Yong M Gabriel U Steven H Steven EK Edward SH Differences in A1C by race and ethnicity among patients with impaired glucose tolerance in the Diabetes Prevention Program Diabetes Care. 2007 30 10 2453 7 10.2337/dc06-2003 17536077
William HH, Yong M, Gabriel U, Steven H, Steven EK, Edward SH, et al. Differences in A1C by race and ethnicity among patients with impaired glucose tolerance in the Diabetes Prevention Program. Diabetes Care. 2007;30(10):2453–7.17536077 10.2337/dc06-2003
