Your privacy, your choice

We use essential cookies to make sure the site can function. We also use optional cookies for advertising, personalisation of content, usage analysis, and social media.

By accepting optional cookies, you consent to the processing of your personal data - including transfers to third parties. Some third parties are outside of the European Economic Area, with varying standards of data protection.

See our privacy policy for more information on the use of your personal data.

for further information and to change your choices.

You are viewing the site in preview mode

Skip to main content
Download PDF

Nonlinear association between hematocrit levels and short-term all-cause mortality in ICU patients with acute pancreatitis: insights from a retrospective cohort study

BMC Gastroenterology volume 25, Article number: 186 (2025) Cite this article

Abstract

Objectives

The purpose of this study was to investigate the relationship between hematocrit levels and the mortality of patients with acute pancreatitis (AP), since limited research has examined this association in intensive care unit (ICU).

Methods

In this study, clinical data were retrieved from Medical Information Mart for Intensive Care database for patients diagnosed with AP. Nonlinear relationships between hematocrit and prognosis were examined through Locally Estimated Scatterplot Smoothing (LOESS) regression, restricted cubic splines (RCS), and U-test analyses. The impact of hematocrit on prognosis was further explored using with a binomial generalized linear model with a logit link, while adjusting for potential confounding factors.

Results

The study encompassed 1,914 patients with AP, revealing a significant difference in hematocrit levels between survivors and non-survivors (33.6 (29.5, 38.1) vs. 32.1 (28.1, 37.4), P < 0.001). Hematocrit emerged as an independent prognostic indicator for mortality in both univariate and multivariate logistic regression analyses (all P < 0.05). Findings from LOESS regression, RCS regression, and the U-test indicated a U-shaped correlation between hematocrit levels and 28-day mortality, with both elevated and decreased hematocrit levels leading to increased mortality risk (P for overall < 0.001). Tertile grouping revealed that lower hematocrit levels (< 30.8%) were associated with heightened 28-day mortality risk (Crude model: Odds ratio (OR) (95%Confidence Interval (CI)) = 1.665 (1.198–2.314); fully adjusted model: adjusted OR = 1.474 (1.005–2.161), all P < 0.05). Survival analyses further supported the adverse prognosis associated with low hematocrit levels.

Conclusions

The findings of this study indicate that in AP patients in the intensive care unit, only low HCT levels were identified as a risk factor for 28-day mortality, despite the presence of a U-shaped correlation between HCT levels and 28-day all-cause mortality.

Peer Review reports

Introduction

Acute pancreatitis (AP) stands out as one of the prevalent gastrointestinal conditions necessitating immediate hospitalization [1,2,3]. Despite advancements in the therapeutic strategies and intensive care of pancreatitis in recent times, AP continues to manifest a notable mortality rate [4]. Research findings suggest that the overall mortality rate for pancreatitis hovers around 1–4% [5,6,7,8], while severe cases exhibit mortality rates as high as 16-40% [7,8,9], with nearly 50% of fatalities occurring within 14 days of symptom onset [6]. Timely identification of risk factors and protective elements crucially shapes the prognosis of patients with pancreatitis, accentuating the necessity to swiftly discern these influences to optimize patient outcomes.

Hematocrit (HCT), denoting the proportion of settled red blood cells in anticoagulated whole blood upon centrifugation, offers a straightforward means to indirectly assess the size and volume of red blood cell count. HCT offers distinct advantages over markers like white blood cell count or C-reactive protein (CRP) as it not only indicates inflammatory response but also provides essential information on oxygen delivery capacity, hemoconcentration, and potential bleeding or anemia. In the context of AP, where hypovolemia, third spacing, and pancreatic necrosis are prevalent, HCT levels may offer a more direct reflection of the severity of these processes [3, 10]. Additionally, HCT is less influenced by confounding factors such as infection or corticosteroid use, which can impact white blood cell count and CRP levels. Hence, HCT was selected for its unique capability to capture both the inflammatory and circulatory abnormalities characteristic of AP, providing a comprehensive evaluation of disease severity and prognosis.

Several research studies have explored the clinical relevance of HCT in the diagnosis and prognostication of pancreatitis. Investigations utilizing extensive prospective databases have identified HCT as a predictor of persistent organ failure and pancreatic necrosis in AP [11]. In a retrospective study utilizing the comprehensive clinical database Medical Information Mart for Intensive Care (MIMIC) IV, low HCT levels (≤ 42% in men and ≤ 37% in women) were identified as an independent risk factor linked to increased 30-day mortality in septic patients within the study cohort [12]. Additionally, another study revealed that patients with elevated HCT levels (HCT > 0.50) exhibited a higher 30-day mortality rate post-surgical procedures compared to those with normal HCT levels (0.41–0.50) (Odds Ratio [OR]: 2.23, 95% Confidence Intervals [CI]: 1.77–2.80) [13]. On the contrary, a separate prospective study failed to establish a significant correlation between HCT levels and critical clinical outcomes in AP, encompassing organ failure and mortality [14]. This discrepancy may stem from differences in study design, patient populations, or the timing of HCT measurement. Despite these mixed findings, insights into the impact of HCT on intensive care unit (ICU) patients with AP remain scarce. In ICU patients, rapid changes in medical conditions, unstable metabolic states, and fluid imbalances often result in significant fluctuations in HCT levels in the bloodstream. Monitoring HCT levels in ICU patients can provide insights into their circulatory status, hemodynamic conditions, and hemorheological properties.

This study compiled clinical chemistry parameters, severity of illness scores, and prognosis data from the MIMIC database for patients diagnosed with AP, aiming to explore the correlation between HCT levels and 28-day all-cause mortality in this patient population. The study postulated that both excessively high and excessively low HCT values in patients with AP could potentially elevate the likelihood of mortality.

Materials and methods

Study design

The researchers conducted this study following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) framework guidelines for reporting studies in epidemiology [2, 15].

Data sources

The researchers obtained data for this study from the MIMIC-IV v2.0 (2008–2019) and MIMIC-III v1.4 databases (2001–2012) [16]. Ruan completed the Collaborative Institutional Training Initiative (CITI) online training course, securing authorization to access and utilize data from the MIMIC databases (Project Approval Number: 10520411) [3]. Figure 1 depicts the research methodology.

Fig. 1
figure 1

Flow chart of the study design

Full size image

Study population

The study cohort comprised adult patients diagnosed with AP who were experiencing their initial admission to the ICU in the Beth Israel Deaconess Medical Center. Criteria for inclusion in the AP group were defined based on the International Classification of Diseases (ICD), with references to the ICD9 and the ICD10 (Supplementary Table 1) [17]. The following exclusion criteria were applied: (1) presence of repeat admission records, and (2) involvement of pediatric patients.

Grouping and study variables

The primary objective of this study was to investigate the impact of HCT on 28-day mortality among patients with AP in the ICU, with HCT representing the initial measurement of packed red blood cell volume relative to whole blood following admission to the ICU for treatment. In addition, data pertaining to patient demographics including gender, age, race, Sequential Organ Failure Assessment (SOFA) score, Age-adjusted Charlson Comorbidity Index (aCCI), Glasgow Coma Scale (GCS), Simplified Acute Physiology Score II (SAPS II), Oxford Acute Severity of Illness Score (OASIS), Logistic Organ Dysfunction System (LODS), Acute Physiology Score III (APS III), special treatments, drugs, comorbidities were gathered [18,19,20,21,22,23,24]. Furthermore, we collected data on medications administered during the patient’s ICU stay, including drug usage and Renal Replacement Therapy (RRT) (Supplementary Tables 2, and Supplementary code 1–2). Supplementary Table 3 presents the outcomes of the multicollinearity test among the variables (all Variance Inflation Factor < 10).

Categorical variables were delineated as follows: gender was stratified into male and female categories based on genetic sex. Race was segmented into white and non-white groups. Age was dichotomized into elderly and non-elderly cohorts using a threshold of 60 years. The influence of comorbidities on patient outcomes was evaluated utilizing aCCI. The aCCI serves as an indicative tool for assessing the overall disease burden and prognostic outlook of patients, considering their underlying health conditions and comorbidities [23]. Additionally, SOFA, SAPS II, OASIS, LODS, and APS III score was utilized to gauge disease severity [18,19,20,21,22]. The GCS score was used to assess the patient’s level of consciousness [24, 25]. The E-value method is a statistical tool used to assess the impact of potential confounding variables on study outcomes, helping researchers evaluate the “resilience to perturbation” of study conclusions in the presence of an unobserved confounding factor [26].

Clinical outcome

The primary clinical outcome of the study was defined as the 28-day all-cause mortality in patients with AP. Secondary clinical outcome encompassed assessments of 90-day all-cause mortality, in-hospital mortality, and the length of stay in the ICU.

Data cleaning

In clinical medical datasets derived from real-world scenarios, the occurrence of missing data is an anticipated challenge. Within this dataset, instances of missing values have been identified (Supplementary Fig. 1a). The missing values are classified as missing completely at random (Little’s MCAR test P > 0.05). In this study, missing covariates are imputed using a combination of the K-Nearest Neighbors (KNN) algorithm and the Random Forest algorithm in Anaconda 3.0 software. KNN interpolation involves identifying the most similar data to the missing values and replacing them with the values from the closest matches. However, it is essential to note that the KNN interpolation method has limitations, including sensitivity to noise in the data and the need to choose the appropriate number of neighbors (K value). The original dataset is split into training and test sets in a 7:3 ratio. The data is interpolated at different K-values, and the training set is used to train the Random Forest model. Subsequently, the model is used to predict the test set, and the Root Mean Square Error (RMSE) is calculated to assess the prediction accuracy. RMSE is a statistical metric that quantifies the difference between predicted and actual values. A smaller RMSE indicates more accurate predictions and is commonly employed to evaluate predictive model performance. Supplementary Fig. 1b display the RMSE values for varying K values, confirming that the optimal performance is achieved when K = 9. Supplementary Fig. 2 illustrates the data distribution before and after data interpolation. Moreover, to reinforce the reliability of the results, sensitivity analyses were conducted in this study incorporating the original raw data. Supplementary Fig. 2 illustrates the data distribution before and after data interpolation. Moreover, to reinforce the reliability of the results, sensitivity analyses were conducted in this study incorporating the original raw data.

Tests for non-linear associations

Nonlinear associations denote scenarios where the correlation between multiple variables deviates from a linear model. In clinical research, numerous clinical parameters exhibit nonlinear connections with clinical outcomes. Within this study, diverse statistical techniques were employed to investigate nonlinear relationships between HCT and mortality. To commence, nonlinear associations were initially examined through Locally Weighted Scatterplot Smoothing (LOWESS), a prevalent nonparametric smoothing approach adept at addressing intricate nonlinear structures in datasets [27]. LOWESS accommodates the nuances of data by integrating local structures and employing relevant weighting functions to mitigate noise and nonlinear characteristics effectively. Subsequently, Restricted Cubic Spline (RCS) regression was utilized in the study. RCS serves as a specialized smoothing tool that segments the independent variable’s value range, employing cubic polynomials in each segment to smooth and fit the data. For the regression in this study, nodes were positioned at the 5th, 35th, 65th, and 95th percentiles [28, 29]. A trend analysis was also carried out by categorizing HCT into three subsets via the tertiles method and incorporating them as dummy variables in the logistic regression model [30, 31]. Tertiles divide data into three equal parts: lower (first third), middle (second third), and upper (top third). Cut-offs are determined by the 33rd and 66th percentiles of the dataset. Lastly, the study examined the U-shaped association by integrating squared term coefficients and utilizing the “utest” method in the model [32, 33].

Developing a 28-day mortality risk score for patients with AP based on HCT levels

To create a mortality risk assessment tool based on HCT levels, this study utilize machine learning techniques, including Least Absolute Shrinkage and Selection Operator (LASSO) regression for variable selection, and develop predictive models using four machine learning models: Gaussian Naive Bayes (GNB), Decision Tree Classifier (DTC), Random Forest (RF), and Gradient Boosting Classifier (GBC) to evaluate the mortality risk [3, 34]. The dataset was randomly split into training and validation sets in a 7:3 ratio. The training set was used for model optimization, the validation set for model selection, and the testing set to evaluate the performance of the chosen model. To assess the effectiveness of each model, this research utilized key evaluation metrics such as the area under the Receiver Operating Characteristic (ROC) curve and Decision Curve Analysis (DCA) curve. Additionally, to gain insights into the impact of HCT levels on the prognosis of AP patients, the SHapley Additive exPlanations (SHAP) theory was incorporated to analyze the importance and contribution of HCT levels in the model predictions.

Statistical analysis

The statistical analysis was conducted with Stata 17.0 (StataCorp, College Station, TX). Continuous variables conforming to a normal distribution or approximating normality were presented as mean ± standard deviation (SD), while non-normally distributed variables were depicted as median (interquartile range 1 - interquartile range 3) [35]. Categorical variables were reported as count (n) and percentage (%) [36]. In instances of non-normally distributed data, the Wilcoxon rank-sum test was utilized to compare two groups, and the Kruskal-Wallis test was employed for comparisons involving three groups [37]. Associations between categorical variables were evaluated using Chi-squared tests. OR alongside their corresponding 95% CI were computed to assess the strength of associations [38]. Subgroup analyses with multiplicative interactions tests were conducted based on sex, race, comorbidity, treatment, and age. A significance threshold was set at a P-value < 0.05 to denote statistical significance.

Results

Baseline characteristics of study population

Table 1 summarizes the inclusion of 1,914 patients diagnosed with AP in this study. In the current dataset, the median time between hospital admission and ICU admission was approximately 0.72 h, with a range from 0.02 to 7.58 h. The researchers segregated the study cohort into survivor and non-survivor subsets based on the patients’ survival outcomes at 28 days post-admission. Notably, the non-survivor group exhibited a notable decline in HCT levels compared to the survivor group (median (interquartile range 1, interquartile range 3): 33.6 (29.5, 38.1) vs. 32.1 (28.1, 37.4), P < 0.01). Furthermore, significant disparities were observed between the two groups concerning age distribution, various critical illness scores, treatment modalities, and the presence of cerebrovascular disease (all P < 0.05).

Table 1 Baseline characteristics of patients with acute pancreatitis stratified by 28-Day mortality
Full size table

HCT as independent prognostic indicator for 28-day mortality

In the univariate logistic regression analysis, the HCT was found to be significantly associated with the outcome variable (OR (95%CI) = 0.967 (0.947–0.987), P < 0.05, Table 2), indicating that HCT is a significant independent prognostic indicator of the outcome. This relationship remained statistically significant after controlling for potential confounders. Using multivariable logistic regression analysis, it was found that after adjusting for covariates such as age, race, and critical care score, variable HCT remained significantly associated with the outcome variable (adjusted OR (95%CI) = 0.970 (0.948–0.993), P < 0.05, Table 2). The subgroup analyses did not reveal significant interactions (all P for interaction > 0.05, Supplementary Table 4).

Table 2 Univariate and multivariate logistic regression analysis of risk factors for 28-Day mortality in patients with acute pancreatitis
Full size table

HCT exhibited a correlation with the severity of the disease

Spearman correlation analysis revealed a weak negative correlation between HCT and the critical care scores (Fig. 2, Spearman’s|r| < 0.3), hinting at a probable link between HCT levels and disease severity. Notably, a significant negative correlation was observed between HCT and aCCI, SOFA, SAPS II, LODS, and APS III scores (all P < 0.05). These results imply that a decrease in HCT values may correspond to an increase in disease severity.

Fig. 2
figure 2

Correlation analysis between HCT and critical care score in patients with acute pancreatitis

Full size image

Non-linear association between HCT and 28-day mortality

The LOWESS regression indicated a nonlinear relationship between HCT and 28-day mortality (Fig. 3a). Subsequently, an RCS regression model was constructed with HCT set at 30 as the reference point, confirming a nonlinear association (P for overall < 0.001, Fig. 3b). Additionally, HCT values were divided into three subgroups using the tertile method. In these groups, 28-day mortality rates initially decreased gradually with higher HCT levels; however, after reaching a minimum point, a slight upward trend was observed with further increases in HCT levels (Fig. 3c). The U-shaped association between HCT and 28-day mortality was demonstrated by the U-test (P < 0.05, Supplementary Table 5).

Fig. 3
figure 3

Association between HCT and 28-days mortality in Patients with Acute Pancreatitis (a) LOWESS regression; (b) RCS regression; (C) Tertile grouping. Note: the study cohort is stratified into thirds according to the tertile values of HCT, and the 28-day mortality rates of patients within each interval are tallied. Logistic regression analysis was conducted to compute the P-value for trend using the HCT tertiles as categorical variables

Full size image

Low HCT levels as risk factor for 28-Day mortality

Due to the U-shaped relationship observed between HCT levels and 28-day mortality, participants in this study were stratified into three subgroups (Q1-Q3) representing low, medium, and high HCT levels. Utilizing Q2 as the baseline group, the impact of HCT on 28-day mortality in individuals with AP was evaluated through the creation of primary logistic regression models, as well as partially and fully adjusted models. Table 3 exhibits the OR with corresponding 95%CI for the first and third quartiles of HCT in the original model as 1.665 (1.198–2.314) and 1.055 (0.780–1.507), respectively.

Table 3 Effect size of HCT levels on 28-day mortality in patients with acute pancreatitis
Full size table

The E-value (point estimate) and confidence interval (CI) were 2.749 and 1.762 for the first quartile of HCT, and 1.279 and 1.000 for the first quartile of HCT. Upon adjusting for age and sex factors in Adjusted model 1, the quartile-corrected OR (95%CI) for HCT were 1.672 (1.199–2.331) and 1.081 (0.753–1.553) for the first and third quartiles, respectively. For Adjusted model 2, after further adjustment for age, gender, and race, the quartile-corrected OR for HCT were 1.695 (1.214–2.365) and 1.079 (0.751–1.551) for the first and third quartiles. Similarly, in Adjusted model 3, following adjustment for race, age, GCS, SOFA, aCCI, SAPSII, LODS, APSIII, cerebrovascular disease, vasopressin, dopamine and RRT, the quartile-corrected OR for HCT were 1.474 (1.005–2.161) and 0.989 (0.652–1.501) for the first and third quartiles. The trend test for 28-day mortality yielded significant results across all four models (all P for trend < 0.05).

Additionally, Kaplan-Meier survival analysis reveals that a lower level of HCT could better predict higher 28-day mortality in AP patients compared to medium (log-rank test P < 0.01, Fig. 4a) and higher HCT levels (log-rank test, P < 0.05). Using LASSO regression, 16 variables, including HCT, Sex, Age, Race, SAPSII, APSIII, GCS, aCCI, LODS, Diabetes, Hypertension, Pulmonary disease, Cerebrovascular disease, Vasopressin, RRT, and Dopamine, were evaluated for inclusion in machine learning models (Supplementary Fig. 3). Among these models, the RF model demonstrated the most effective predictive performance, as evidenced by the ROC curve and Decision Curve Analysis (DCA) curve (AUC (95% CI) = 0.878 (0.840–0.917), Fig. 4b-4). Further insights from feature importance ranking and SHAP analysis elucidated a correlation between elevated HCT levels and increased mortality risk in AP patients, with lower HCT levels associated with an increased risk of 28-day mortality (Fig. 4d-4).

Fig. 4
figure 4

The risk score based on HCT levels effectively predicts the 28-day mortality in patients with acute pancreatitis. (a) Kaplan-Meier Survival Analysis of 28-Day Mortality in Patients with Acute Pancreatitis Stratified by HCT Tertiles; (b) ROC curves predicting 90-day mortality rates for various machine learning models: GBC model AUC (95% CI) = 0.852 (0.809–0.895); DTC model AUC (95% CI) = 0.812 (0.756–0.868); RF model AUC (95% CI) = 0.878 (0.840–0.917); GNB model AUC (95% CI) = 0.821 (0.772–0.869); (c) Decision Curve Analysis curves for various machine learning models indicate performance metrics, where curves above represent better model performance; (d) Feature variable importance ranking in the RF model; (e) Visualization of SHapley Additive Explanations illustrating the impact of individual features on RF prediction model outcomes. The color spectrum denotes the degree of influence, with warmer hues indicating higher relevance to 28-day mortality risk and cooler tones indicating lower relevance to 28-day mortality risk

Full size image

The association between HCT and secondary clinical outcome

In the analysis of the impact of variables on secondary clinical outcomes, low level HCT was found to have a significant effect on the In-hospital mortality and 90-day mortality (P < 0.05, Table 4). However, no significant effect of different levels of HCT on the length of stay of patients in the ICU was found.

Table 4 Effect size of HCT on secondary clinical outcome in patients with acute pancreatitis
Full size table

Sensitivity analysis

To assess the robustness of the findings, a sensitivity analysis was conducted. The analysis was repeated with raw data, and the results remained consistent across all scenarios (Supplementary Tables 6–9). This indicates that the study findings are robust and reliable. Interaction and stratified analyses were not detected in terms of age, sex, ethnicity, drug, treatment, and co-morbidities (Supplementary Table 10).

Discussion

The study aimed to investigate the relationship between HCT levels and mortality in patients with AP in the ICU. Both univariate and multivariate logistic regression models identified HCT as an independent predictive factor for mortality in AP patients. Non-linear correlation analyses revealed a U-shaped association, indicating higher mortality risk with extremely high or low HCT levels. However, within the high HCT group, we observed a lack of significant correlation (P > 0.05) between the adjusted OR in the model and the 28-day mortality of AP patients. This suggests that within the high HCT range, blood parameters may have a lower impact on prognosis for patients with AP. Furthermore, survival analyses highlighted the significance of low HCT levels on patient prognosis. The study’s survival analyses indicated a notable increase in the risk of all-cause mortality within 28 days for individuals with lower HCT levels compared to those with intermediate HCT levels. While there was an observed rise in the risk of death within 28 days with higher HCT levels, statistical significance was not evident. In conclusion, our study findings indicate that within this specific subset of acute pancreatitis patients, only low HCT levels were identified as a risk factor for 28-day mortality, despite the presence of a U-shaped correlation between HCT levels and 28-day all-cause mortality.

The U-shaped association between HCT and mortality revealed in this study suggests its potential decision value in fluid management for AP. A prospective cohort study in the general population found that both low and high levels of HCT were linked to increased overall mortality and cardiovascular disease mortality [39]. Current guidelines by the International Association of Pancreatology (IAP) recommend early aggressive fluid resuscitation targeting 5–10 mL/kg/h of fluid therapy, but do not specify adjustment for HCT, potentially overlooking individualized needs [40]. Our study offers guidance for fluid resuscitation strategies. For high HCT (> 36.3%), it may indicate hemoconcentration (such as capillary leakage due to early systemic inflammatory response in AP, shifting fluid to the third space) or dehydration. Excessive restriction of fluid may worsen organ hypoperfusion (e.g., prerenal kidney injury). Low HCT (< 30.8%) may reflect anemia (from chronic diseases, gastrointestinal bleeding) or over-dilution of blood (from excessive crystalloid resuscitation). Furthermore, our study highlights the significant role of HCT in predicting short-term mortality in AP patients. It has been reported that HCT predicts long-term mortality in hypertensive adults in a non-linear and gender-specific manner [41]. Thus, future studies can further explore the impact of HCT on the long-term prognosis of patients with AP.

The findings of this study are consistent with previous clinical research that underscores the practical significance of HCT levels in predicting clinical outcomes in patients with AP. Previous studies exploring the association between HCT and AP prognosis have emphasized the predictive value of hematocrit levels [42]. For instance, a prospective study involving 120 AP patients identified HCT as a significant predictor of severe pancreatitis [43]. Similarly, a retrospective study highlighted the relevance of increased hematocrit in predicting severe pancreatitis, prolonged hospitalization, pancreatic necrosis, and the need for intensive care (all P < 0.05) [44]. Additionally, in a study focusing on edematous AP subtype, HCT emerged as a substantial predictor of pancreatic necrosis development [45]. In the context of elderly AP patients, HCT was independently associated with more severe disease [46]. In contrast to these existing studies, our current research introduces novel insights into pancreatitis predictors in the ICU setting through a more comprehensive analysis of this association within a larger and diverse ICU database.

Past clinical research has elucidated the significance of HCT as a prognostic indicator in AP. However, the debate persists regarding whether elevated HCT levels in AP serve as a risk factor or a protective element. Several studies have put forth the viewpoint that heightened HCT is linked to adverse clinical outcomes in AP patients [43,44,45,46]. These investigations suggest that increased HCT levels signify hemoconcentration resulting from substantial exudation in AP, a pivotal factor in the progression to severe AP and an ominous sign for patient prognosis. For example, Jinno et al. [47] reported that HCT levels ≥ 40% were associated with an increased risk of death in AP patients (OR (95%CI) = 1.07 (1.01–1.13), P < 0.05). Conversely, another study underscored that HCT levels < 35% in pancreatitis correlated with heightened mortality within 28 days [48]. The study aims to shed light on this conflicting scenario. The outcomes of our investigation propose a U-shaped correlation between HCT levels and mortality, where both excessively high and low HCT values are tied to elevated mortality risks.

In the correlation analysis section of our study, we found that excessively high or low HCT levels in AP patients were similarly linked to a poorer prognosis. Elevated HCT levels may indicate physiological abnormalities such as dehydration, blood loss, extracellular fluid depletion, or stress responses, serving as significant indicators of adverse prognosis in AP [49]. Furthermore, heightened HCT levels could lead to reduced tissue perfusion, resulting in tissue hypoxia and organ dysfunction [49]. Additionally, increased blood viscosity associated with high HCT levels may contribute to microcirculatory disturbances, excessive erythrocyte clumping, and microthrombosis, thereby exacerbating local tissue ischemia and inflammatory responses [50]. Consequently, high HCT levels are considered an unfavorable prognostic factor in AP. Conversely, certain studies suggest that low HCT levels may pose challenges in AP [48]. Low HCT levels are often indicative of conditions like heart failure, anemia, or severe inflammatory responses, which can directly impact the balance of oxygen supply and demand in tissues, exacerbating systemic metabolic stress [51,52,53]. Therefore, low HCT levels also reflect, to a certain extent, the intricate pathophysiological processes in patients with AP.

Patients with AP in the ICU often encounter various severe physiological challenges, including hemodynamic instability, fluid management complexities, and multi-organ failure. Firstly, patients with AP are susceptible to alterations in their hemodynamics, where instability can result in tissue hypoxia, microcirculatory disturbances, inadequate organ perfusion, and subsequent multi-organ dysfunction and deterioration. Secondly, fluid management is paramount in ICU care for pancreatitis patients who require intricate fluid therapy to stabilize circulation and prevent complications. Monitoring HCT levels can gauge the effectiveness of fluid management and guide healthcare providers in implementing suitable fluid adjustments to enhance patient outcomes. Additionally, e7multi-organ failure is a leading cause of mortality in ICU patients with AP. The onset of multi-organ failure is closely associated with metabolic imbalances, inflammatory responses, and systemic deterioration. Therefore, HCT plays a crucial role in assessing mortality risks and guiding the management of AP patients in the ICU, considering dynamic factors like hemodynamic instability, fluid management complexities, and multi-organ dysfunction.

Limitations of utilizing scoring systems such as the Bedside Index of Severity in Acute Pancreatitis (BISAP), APS III, RANSON, and LODS for assessing mortality in ICU patients with AP [54, 55]. While the Ranson Criteria assesses elements like age, serum albumin level, and blood urea nitrogen within the first 48 h in patients with pancreatitis, it fails to account for the dynamic physiological challenges and systemic changes encountered in the ICU setting [56]. Although scoring tools like BISAP, APS III, and LODS offer a more comprehensive evaluation of patient severity and systemic status, they have limitations, particularly in dynamically monitoring and guiding clinical decisions for ICU patients with pancreatitis. Additionally, our study did not find a significant correlation between HCT levels and GCS or OASIS scores. HCT, GCS, and OASIS serve as distinct physiological indicators and tools for clinical assessment. The GCS score evaluates immediate neurological impairments, such as brain hypoxia or elevated intracranial pressure, while the OASIS score combines various parameters to predict overall mortality in acutely hospitalized patients, focusing on holistic outcomes rather than individual organ function [20, 24]. As these metrics assess different aspects of physiology and clinical conditions, a direct correlation may not be evident. As a crucial hematological parameter, HCT serves as a valuable inclusion in various critical care scoring systems, dynamically reflecting blood volume, fluid status, and oxygen delivery. This aids in accurately evaluating disease severity and prognosis promptly and effectively.

The study’s strengths include the utilization of a large sample size from a publicly available dataset, providing a comprehensive understanding of the correlation between HCT and prognosis in AP patients. Furthermore, while prior research has primarily targeted general ward patients, this study focuses on AP patients in the ICU, thereby enhancing the literature on this specific subgroup. However, there are several potential limitations in this study. This study utilized retrospective cohort data from the MIMIC database, which may introduce selection bias, as the inclusion criteria could have excluded patients with severe disease or premature death, potentially leading to biased estimates of the association between HCT and prognosis. Furthermore, the absence of different subtypes of AP in the MIMIC database might impact prognostic assessments. To mitigate potential biases, strict inclusion and exclusion criteria were implemented to minimize duplicate cases. Missing data were addressed through KNN Imputation to reduce information bias. Multivariate logistic regression models were used to adjust for known confounders. Despite these efforts, the presence of unmeasured variables could still introduce bias into the study results. To address this, rigorous sensitivity analyses and the use of the E-Value were employed to account for confounding variables and mitigate the influence of unmeasured factors. However, the inability of the MIMIC database to incorporate patient-specific cause-of-death data may limit in-depth analyses of cause-specific mortality. This underscores the importance of more comprehensive datasets in future research endeavors. Additionally, as the study’s results were derived from a single-center database, they may be influenced by institution-specific treatment patterns, such as fluid resuscitation strategies. External validation in diverse populations, such as an Asian AP cohort, is necessary to evaluate the generalizability of the findings.

Future prospective studies could explore the implementation of HCT-guided fluid management protocols in AP patients to evaluate their effects on clinical outcomes. These studies could also delve deeper into the role of HCT in different critical illnesses. Prospective intervention studies might entail monitoring HCT levels and adapting fluid therapy according to established guidelines or thresholds. By assessing the effects of personalized fluid management strategies on outcomes such as organ dysfunction, length of hospital stays, and mortality rates, researchers can provide valuable insights into the benefits of HCT-guided therapy for improving patient care and outcomes in AP.

Conclusions

In conclusion, this study underscores the significance of HCT as a prognostic marker for AP. The findings reveal a U-shaped relationship between HCT levels and mortality in AP, where both excessively high and low HCT values are linked to increased mortality risks.

Strengths and limitations of the study

  • The study benefitted from a large sample size, enabling a robust modeling approach.

  • A U-shaped relationship between Hematocrit Levels and mortality were identified using multiple non-linear regressions.

  • The retrospective nature of the study limited the ability to infer causal relationships.

  • Further prospective studies are needed to elucidate the predictive role of HCT in patients with pancreatitis in the ICU.

Data availability

Availability of data and materialsThe datasets and codes employed in this study can be accessed from the corresponding author upon reasonable request.

Abbreviations

aCCI:

Age-adjusted charlson comorbidity index

AP:

Acute pancreatitis

APS III:

Acute physiology score III

CRP:

C-reactive protein

CI:

Confidence intervals

CITI:

Collaborative institutional training initiative

DCA:

Decision curve analysis

DTC:

Decision tree classifier

GCS:

Glasgow coma scale

GBC:

Gradient boosting classifier

GNB:

Gaussian naive bayes

HCT:

Hematocrit

IAP:

International association of pancreatology

ICD:

International classification of diseases

ICU:

Intensive care unit

KNN:

K-nearest neighbors

LASSO:

Least absolute shrinkage and selection operator

LODS:

Logistic organ dysfunction system

LOWESS:

Locally weighted scatterplot smoothing

MIMIC:

Medical information mart for intensive care

OASIS:

Oxford acute severity of illness score

OR:

Odds ratios

OR:

Odds ratio

RCS:

Restricted cubic spline

RMSE:

Root mean square error

RF:

Random forest

RCS:

Restricted cubic splines

RRT:

Renal replacement therapy

SD:

Standard deviation

SAPS II:

Simplified acute physiology score II

SOFA:

Sequential organ failure assessment

STROBE:

Strengthening the reporting of observational studies in epidemiology

References

  1. Boxhoorn L, Voermans RP, Bouwense SA, Bruno MJ, Verdonk RC, Boermeester MA, van Santvoort HC, Besselink MG. Acute pancreatitis. Lancet. 2020;396(10252):726–34.

    Article  PubMed  Google Scholar 

  2. Dancu GM, Popescu A, Sirli R, Danila M, Bende F, Tarta C, Sporea I. The BISAP score, NLR, CRP, or BUN: Which marker best predicts the outcome of acute pancreatitis? Medicine. 2021;100(51).

  3. Liu HQ, Wang GQ, Zhang CS, Wang X, Shi JK, Qu F, Ruan H. Nucleated red blood cell distribution in critically ill patients with acute pancreatitis: a retrospective cohort study. BMC Gastroenterol. 2024;24(1):353.

    Article  PubMed Central  PubMed  Google Scholar 

  4. Yu J, Liu C, Zhang J, Wang X, Song K, Wu P, Liu F. Global, regional, and National burden of pancreatitis in older adults, 1990–2019: A systematic analysis for the global burden of disease study 2019. Prev Med Rep. 2024;41:102722.

    Article  PubMed Central  PubMed  Google Scholar 

  5. Forsmark CE, Vege SS, Wilcox CM. Acute pancreatitis. N Engl J Med. 2016;375(20):1972–81.

    Article  CAS  PubMed  Google Scholar 

  6. Mutinga M, Rosenbluth A, Tenner SM, Odze RR, Sica GT, Banks PA. Does mortality occur early or late in acute pancreatitis? Int J Pancreatol. 2000;28(2):91–5.

    Article  CAS  PubMed  Google Scholar 

  7. Iannuzzi JP, King JA, Leong JH, Quan J, Windsor JW, Tanyingoh D, Coward S, Forbes N, Heitman SJ, Shaheen A-A, et al. Global incidence of acute pancreatitis is increasing over time: A systematic review and Meta-Analysis. Gastroenterology. 2022;162(1):122–34.

    Article  PubMed  Google Scholar 

  8. Fu CY, Yeh CN, Hsu JT, Jan YY, Hwang TL. Timing of mortality in severe acute pancreatitis: experience from 643 patients. World J Gastroenterol. 2007;13(13):1966–9.

    Article  PubMed Central  PubMed  Google Scholar 

  9. Nawacki L, Gluszek S. Hospital mortality rate and predictors in acute pancreatitis in Poland: A single-center experience. Asian J Surg. 2024;47(1):208–15.

    Article  PubMed  Google Scholar 

  10. Remes-Troche JM, Duarte-Rojo A, Morales G, Robles-Diaz G. Hemoconcentration is a poor predictor of severity in acute pancreatitis. World J Gastroenterol. 2005;11(44):7018–23.

    Article  PubMed Central  PubMed  Google Scholar 

  11. Koutroumpakis E, Wu BU, Bakker OJ, Dudekula A, Singh VK, Besselink MG, Yadav D, van Santvoort HC, Whitcomb DC, Gooszen HG, et al. Admission hematocrit and rise in blood Urea nitrogen at 24 h outperform other laboratory markers in predicting persistent organ failure and pancreatic necrosis in acute pancreatitis: A post hoc analysis of three large prospective databases. Am J Gastroenterol. 2015;110(12):1707–16.

    Article  PubMed  Google Scholar 

  12. Luo M, Chen Y, Cheng Y, Li N, Qing H. Association between hematocrit and the 30-day mortality of patients with sepsis: A retrospective analysis based on the large-scale clinical database MIMIC-IV. PLoS ONE. 2022;17(3):e0265758.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  13. Musallam KM, Porter JB, Soweid A, Hoballah JJ, Sfeir PM, Khalife M, Zeineldine S, Tamim HM, Richards T, Lotta LA, et al. Elevated hematocrit concentration and the risk of mortality and vascular events in patients undergoing major surgery. Blood. 2012;120(21):2088–2088.

    Article  Google Scholar 

  14. Lankisch PG, Mahlke R, Blum T, Bruns A, Bruns D, Maisonneuve P, Lowenfels AB. Hemoconcentration: an early marker of severe and/or necrotizing pancreatitis? A critical appraisal. Am J Gastroenterol. 2001;96(7):2081–5.

    Article  CAS  PubMed  Google Scholar 

  15. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP, Initiative S. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol. 2008;61(4):344–9.

    Article  Google Scholar 

  16. Johnson AEW, Pollard TJ, Shen L, Lehman L-wH, Feng M, Ghassemi M, Moody B, Szolovits P, Celi LA, Mark RG. MIMIC-III, a freely accessible critical care database. Sci Data. 2016;24(3):160035. https://doi.org/10.1038/sdata.2016.35

  17. Wang D, Lu J, Zhang P, Hu Z, Shi Y. Relationship between blood glucose levels and length of hospital stay in patients with acute pancreatitis: an analysis of MIMIC-III database. Cts-Clinical Translational Sci. 2023;16(2):246–57.

    Article  CAS  Google Scholar 

  18. Vincent JL, de Mendonca A, Cantraine F, Moreno R, Takala J, Suter PM, Sprung CL, Colardyn F, Blecher S. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Crit Care Med. 1998;26(11):1793–800.

    Article  CAS  PubMed  Google Scholar 

  19. Le Gall JR, Lemeshow S, Saulnier F. A new simplified acute physiology score (SAPS II) based on a European/North American multicenter study. JAMA. 1993;270(24):2957–63.

    Article  PubMed  Google Scholar 

  20. Chen QG, Zhang LS, Ge SH, He WM, Zeng M. Prognosis predictive value of the Oxford acute severity of illness score for sepsis: a retrospective cohort study. Peerj. 2019;10(7):e7083. https://doi.org/10.7717/peerj.7083

  21. Marshall JC, Cook DJ, Christou NV, Bernard GR, Sprung CL, Sibbald WJ. Multiple organ dysfunction score: A reliable descriptor of a complex clinical outcome. Crit Care Med. 1995;23(10):1638–52.

    Article  CAS  PubMed  Google Scholar 

  22. Knaus WA, Wagner DP, Draper EA, Zimmerman JE, Bergner M, Bastos PG, Sirio CA, Murphy DJ, Lotring T, Damiano A. The APACHE III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults. Chest. 1991;100(6):1619–36.

    Article  CAS  PubMed  Google Scholar 

  23. Koppie TM, Serio AM, Vickers AJ, Vora K, Dalbagni G, Donat SM, Herr HW, Bochner BH. Age-adjusted Charlson comorbidity score is associated with treatment decisions and clinical outcomes for patients undergoing radical cystectomy for bladder cancer. Cancer. 2008;112(11):2384–92.

    Article  PubMed  Google Scholar 

  24. Teasdale G, Maas A, Lecky F, Manley G, Stocchetti N, Murray G. The Glasgow coma scale at 40 years: standing the test of time. Lancet Neurol. 2014;13(8):844–54.

    Article  PubMed  Google Scholar 

  25. Wang X, Xia J, Shan Y, Yang Y, Li Y, Sun H. Predictive value of the Oxford acute severity of illness score in acute stroke patients with stroke-associated pneumonia. Front Neurol. 2023;4(14):1251944. https://doi.org/10.3389/fneur.2023.1251944

  26. VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E-Value. Ann Intern Med. 2017;167(4):268–.

    Article  PubMed  Google Scholar 

  27. Cleveland WS. A Reader’s Guide to Smoothing Scatterplots and Graphical Methods for Regression. In:Modern Data Analysis. Edited by Launer RL, Siegel AF: Academic Press; 1982:37–43.

  28. Ruan H, Li S-S, Zhang Q, Ran X. Elevated MMP-8 levels, inversely associated with BMI, predict mortality in mechanically ventilated patients: an observational multicenter study. Crit Care (London England). 2023;27(1):290–290.

    Article  Google Scholar 

  29. Wang M, Zhu B, Jiang L, Wen Y, Du B, Li W, Liu G, Li W, Wen J, He Y et al. Dose-response association between fluid overload and in-hospital mortality in critically ill patients: a multicentre, prospective, observational cohort study. Bmj Open 2020;10(12).

  30. Zhou H, Xu T, Huang Y, Zhan Q, Huang X, Zeng Q, Xu D. The top tertile of hematocrit change during hospitalization is associated with lower risk of mortality in acute heart failure patients. BMC Cardiovasc Disord. 2017;17(1):235.

    Article  PubMed Central  PubMed  Google Scholar 

  31. Brown DW, Giles WH, Croft JB. Hematocrit and the risk of coronary heart disease mortality. Am Heart J. 2001;142(4):657–63.

    Article  CAS  PubMed  Google Scholar 

  32. Lind JT, Mehlum H. With or without U? The appropriate test for a U-Shaped relationship. Oxf Bull Econ Stat. 2010;72(1):109–18.

    Article  Google Scholar 

  33. Ruan H, Li Y-z, Zhang Q, Wang B-r, Wu R, Li S-s, Ran X. IDENTIFICATION AND CLINICAL VALIDATION OF HYPOXIA-INDUCIBLE FACTOR 1a PROTEIN AS THE POTENTIAL BIOMARKER IN PATIENTS WITH SEPSIS. Shock. 2023;59(6):855–63.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  34. Ruan H, Ran X, Li SS, Zhang Q. Dyslipidemia versus obesity as predictors of ischemic stroke prognosis: a multi-center study in China. Lipids Health Dis 2024;23(1).

  35. Ismaiel A, Spinu M, Leucuta D-C, Popa S-L, Chis BA, Stanculete MF, Olinic DM, Dumitrascu DL. Anxiety and depression in Metabolic-Dysfunction-Associated fatty liver disease and cardiovascular risk. J Clin Med 2022;11(9).

  36. Zhang D, Wang T, Dong X, Sun L, Wu Q, Liu J, Sun X. Systemic Immune-Inflammation index for predicting the prognosis of critically ill patients with acute pancreatitis. Int J Gen Med. 2021;14:4491–8.

    Article  PubMed Central  PubMed  Google Scholar 

  37. Zhong Y, Zhang Y, Wei S, Chen J, Zhong C, Cai W, Jin W, Peng H. Dissecting the effect of sphingolipid metabolism gene in progression and microenvironment of osteosarcoma to develop a prognostic signature. Front Endocrinol. 2022;14(13):1030655. https://doi.org/10.3389/fendo.2022.1030655

  38. Yao T, Ong M, Lee A, Jiang Y, Mao Z. Smoking knowledge, attitudes, behavior, and associated factors among Chinese male surgeons. World J Surg. 2009;33(5):910–7.

    Article  PubMed  Google Scholar 

  39. Boffetta P, Islami F, Vedanthan R, Pourshams A, Kamangar F, Khademi H, Etemadi A, Salahi R, Semnani S, Emadi A, et al. A U-shaped relationship between haematocrit and mortality in a large prospective cohort study. Int J Epidemiol. 2013;42(2):601–15.

    Article  PubMed Central  PubMed  Google Scholar 

  40. IAP/APA evidence-based guidelines for the management of acute pancreatitis. Pancreatology. 2013, 13(4, Supplement 2):e1-e15.

  41. Paul L, Jeemon P, Hewitt J, McCallum L, Higgins P, Walters M, McClure J, Dawson J, Meredith P, Jones GC et al. Hematocrit predicts Long-Term mortality in a nonlinear and Sex-Specific manner in hypertensive adults. 2012;60(3):631–8.

  42. Zhou HJ, Mei X, He XH, Lan TF, Guo SB. Severity stratification and prognostic prediction of patients with acute pancreatitis at early phase A retrospective study. Medicine 2019;98(16).

  43. Patel ML, Shyam R, Atam V, Bharti H, Sachan R, Parihar A. Clinical profile, etiology, and outcome of acute pancreatitis: experience at a tertiary care center. Ann Afr Med. 2022;21(2):118–23.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  44. Gan SI, Romagnuolo J. Admission hematocrit: A simple, useful and early predictor of severe pancreatitis. Dig Dis Sci. 2004;49(11–12):1946–52.

    Article  PubMed  Google Scholar 

  45. Acehan F, Comoglu M, Kayserili FM, Hayat B, Ates I. Factors predicting the development of necrosis in patients presenting with edematous acute pancreatitis. Pancreas. 2022;51(10):1300–7.

    Article  CAS  PubMed  Google Scholar 

  46. He F, Zhu HM, Li BY, Li XC, Yang S, Wang Z, Zhang M. Factors predicting the severity of acute pancreatitis in elderly patients. Aging Clin Exp Res. 2021;33(1):183–92.

    Article  PubMed  Google Scholar 

  47. Jinno N, Hori Y, Naitoh I, Miyabe K, Yoshida M, Natsume M, Kato A, Asano G, Sano H, Hayashi K. Predictive factors for the mortality of acute pancreatitis on admission. PLoS ONE 2019;14(8).

  48. Mao EQ, Fei JA, Peng YB, Huang J, Tang YQ, Zhang SD. Rapid hemodilution is associated with increased sepsis and mortality among patients with severe acute pancreatitis. Chin Med J. 2010;123(13):1639–44.

    Article  CAS  PubMed  Google Scholar 

  49. Allen MT, Patterson SM. Hemoconcentration and stress: a review of physiological mechanisms and relevance for cardiovascular disease risk. Biol Psychol. 1995;41(1):1–27.

    Article  CAS  PubMed  Google Scholar 

  50. Pittman RN. Elevated haematocrit - when too much of a good thing wreaks havoc on the endothelial surface layer. J Physiol. 2011;589(Pt 22):5339.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  51. Northrop-Clewes CA, Thurnham DI. Biomarkers for the differentiation of anemia and their clinical usefulness. J Blood Med. 2013;4:11–22.

    PubMed Central  PubMed  Google Scholar 

  52. Anand IS, Gupta P. Anemia and Iron deficiency in heart failure. Circulation. 2018;138(1):80–98.

    Article  CAS  PubMed  Google Scholar 

  53. Yang M, Ma N, Fu H, Wei T, Tang Q, Qin B, Yang Z, Zhong R. Hematocrit level could reflect inflammatory response and disease activity in patients with systemic lupus erythematosus. Clin Lab. 2015;61(7):801–7.

    CAS  PubMed  Google Scholar 

  54. Capurso G, Ponz de Leon Pisani R, Lauri G, Archibugi L, Hegyi P, Papachristou GI, Pandanaboyana S, Maisonneuve P, Arcidiacono PG, de-Madaria E. Clinical usefulness of scoring systems to predict severe acute pancreatitis: A systematic review and meta-analysis with pre and post-test probability assessment. United Eur Gastroenterol J. 2023;11(9):825–36.

    Article  Google Scholar 

  55. Juneja D, Gopal PB, Ravula M. Scoring systems in acute pancreatitis: which one to use in intensive care units? J Crit Care 2010;25(2):358.e359-358.e315.

  56. Inal V, Yamanel L, Comert B. Predicting outcome in the ICU: comparison of Ranson criteria and Ranson + CRP levels in acute pancreatitis. Crit Care. 2012;16(Suppl 1):P386. https://doi.org/10.1186/cc10993. Epub 2012 Mar 20.

    Article  PubMed Central  Google Scholar 

Download references

Acknowledgements

We express our deepest appreciation to all individuals involved in the development of the Medical Information Marketplace in Intensive Care database, which has provided a valuable data platform for advancing clinical big data research initiatives.

Funding

This work was supported by funding from the Beijing Medical and Health Foundation (YWJKJJHKYJJ-BXS5-22001).

Author information

Authors and Affiliations

  1. Department of Rehabilitation, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China

    Li-Juan Zou

  2. Department of Emergency Medicine, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430030, Hubei Province, China

    Hang Ruan & Yong-Sheng Li

  3. Department of Critical-care Medicine, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, 430030, Hubei Province, China

    Hang Ruan & Yong-Sheng Li

Authors
  1. Li-Juan Zou
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Hang Ruan
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Yong-Sheng Li
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

L-J. Z. was responsible for the study design, statistical analyses, and the initial draft of the manuscript. Data collection was undertaken by H.R. L-J. Z. and Y-S. L. contributed to revising the manuscript critically for important intellectual content and approved the final version for publication. All authors reviewed the manuscript critically for important intellectual content and have read and approved the final manuscript.

Corresponding author

Correspondence to Yong-Sheng Li.

Ethics declarations

Ethics approval and consent to participate

Ethical approval was secured from the Ethics Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (approval number: TJ-IRB20230763). Given the retrospective nature of this study, the requirement for informed consent was deemed waived.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

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

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions

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/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Zou, LJ., Ruan, H. & Li, YS. Nonlinear association between hematocrit levels and short-term all-cause mortality in ICU patients with acute pancreatitis: insights from a retrospective cohort study. BMC Gastroenterol 25, 186 (2025). https://doi.org/10.1186/s12876-025-03764-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12876-025-03764-8

Keywords

BMC Gastroenterology

ISSN: 1471-230X

Contact us