Association Between Fluid Balance Trajectories and Prognosis in Patients with Heart Failure and Preserved Ejection Fraction

Association Between Fluid Balance Trajectories and Prognosis in Patients with Heart Failure and Preserved Ejection Fraction | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Association Between Fluid Balance Trajectories and Prognosis in Patients with Heart Failure and Preserved Ejection Fraction Chunmei Zhang, Guangyu Lin, Fengzhen Chen, Qitian Zhang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6956502/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Background Fluid balance (FB) is a critical prognostic factor in heart failure, yet its longitudinal impact on critically ill HFpEF patients remains unclear. This study evaluates how FB trajectory changes influence 30-day mortality in HFpEF patients. Methods Using MIMIC-IV data, we conducted a retrospective cohort study of HFpEF patients. Group-based trajectory modeling (GBTM) identified distinct FB trajectory subgroups. Survival differences were assessed via Kaplan-Meier analysis, and Cox regression models evaluated associations between FB trajectories and mortality. Results Among 1,089 HFpEF patients, four FB trajectories emerged: T1 (Negative balance stability), T2 (Rapid transition to negative balance), T3 (Positive balance gradual decline), and T4 (High-level decline). K-M analysis revealed significantly higher mortality in T3 and T4. Fluid-overloaded patients had worse survival than non-overloaded. Adjusted Cox models showed lower mortality in T1 (HR = 0.67, 95%CI 0.53–0.85) and T2 (HR = 0.60, 95%CI 0.45–0.80) vs. T3, with no T3–T4 difference (p = 0.35). Subgroup and sensitivity analyses supported these findings. Conclusions HFpEF patients with positive balance gradual decline (T3) or high-level decline FB (T4) had the poorest prognosis, while those maintaining negative balance (T1/T2) exhibited better survival. GBTM effectively stratifies risk, aiding clinical subgroup identification. Health sciences/Cardiology Health sciences/Medical research Heart Failure with Preserved Ejection Fraction GBTM Fluid Balance MIMIC Database Prognosis Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Heart failure (HF) is a complex clinical syndrome affecting over 60 million patients worldwide 1 . Heart failure with preserved ejection fraction (HFpEF) is a type of heart failure characterized by the presence of signs and symptoms of heart failure with a normal or near-normal left ventricular ejection fraction (≥ 50%) 2 . The global prevalence of HFpEF is approximately 2% 3 , accounting for nearly half of all heart failure cases, and its prevalence is expected to continue rising with the aging of the population 4 . HFpEF is associated with high rates of hospitalization and mortality, with hospitalization and mortality rates as high as 80% and 50% within 5 years, respectively 5 . Furthermore, due to the frequent readmissions of HFpEF patients, the overall hospitalization burden remains high 6 . Patients with HFpEF often exhibit symptoms of fluid overload(FO), which is an independent adverse prognostic indicator in HFpEF patients 7 . FO is associated with cardiac remodeling and subtle contractile dysfunction, and renal function and fluid status may jointly influence the progression and prognosis of HFpEF 8 . The interaction between the heart and lungs may affect fluid balance(FB) and cardiac function, further complicating the management of HFpEF 9 . Current research on the relationship between FB status and the prognosis of HFpEF patients is very limited. Although studies have shown that HFpEF patients with FO have a poor prognosis 7 , these studies use a single baseline fluid status for assessment. However, defining FO at a single arbitrary time point may not provide sufficient information to explore the impact of FO on clinical outcomes 10 . A recent study of patients undergoing cardiac or aortic surgery showed that changes in FB trends are related to clinical outcomes 11 . Fluid management is an essential component of critical patient care, and although various strategies for assessing fluid responsiveness have been developed 12 , the optimal fluid management plan remains controversial. Group-based trajectory modeling (GBTM) can more accurately describe and understand the heterogeneity and similarities among individuals 13 , and is used to explore dynamic changes among individuals 14 . GBTM has been widely applied in medical and psychological fields 14 . This study aims to use the GBTM method to analyze the FB trajectories of patients with HFpEF, revealing the association between different trajectory patterns and patient prognosis. This study can help optimize fluid management strategies for HFpEF patients, provide theoretical support for early intervention and personalized treatment by accurately identifying high-risk patient groups, thereby improving patient outcomes. Results Baseline Characteristics A total of 19,969 heart failure patients were screened from the MIMIC-IV database. After applying the exclusion criteria, 1,089 patients with HFpEF were ultimately included to form the study cohort (Fig. 1 ). Variables with excessive missing data (Supplementary Fig. S1 ), high correlation (Supplementary Fig. S2), and strong collinearity (Supplementary Fig. S3) were excluded. Table 1 presents the baseline characteristics of the patients. The median age of all HFpEF patients was 75.0 years, with 592 (54.4%) being male. A total of 1,062 (97.5%) patients received mechanical ventilation, 101 (9.3%) underwent continuous renal replacement therapy (CRRT), and 617 (56.7%) had atrial fibrillation. The prevalence of comorbidities included hypertension in 816 (75.0%), diabetes mellitus in 447 (41.1%), and hyperlipidemia in 521 (47.8%). The admission scores to the ICU were a SOFA score of 6.0 and a Charlson score of 7.00. FB Trajectory Description Prior to grouping the FB trajectories, we pre-analyzed the FB status from day 1 to day 7. However, due to a significant proportion of missing values, particularly on days 4 to 7 (6.9%, 28.4%, 45.5%, and 57.0%, respectively), we only included data from days 1 to 4 in the analysis. Table 2 presents the model fit statistics and AvePP used to determine the optimal number of FB trajectory groups. Although the AIC and BIC were smallest for the 5-group model, the proportion of patients in some groups was less than 5%, which was not acceptable. The 4-group model had AIC and BIC values comparable to the 5-group model, with all group proportions exceeding 5%. The AvePP for each group was greater than 0.7, and the minimum OCC performed best among all groupings. Considering clinical relevance and interpretability, the 4-group trajectory model was ultimately selected. Table 2. Performance of the group-based trajectory model for fluid balance trajectories Trajectories BIC AIC AvePP Minimum OCC Class proportion 1 group 40699.98 40668.08 1.00 NaN 100.0% 2 groups 39894.87 39831.08 0.89/0.90 5.60 62.0%/38.0% 3 groups 39760.49 39664.80 0.89/0.87/0.78 6.39 60.1%/24.0%/15.9% 4 groups 39521.51 39381.16 0.87/0.79/0.78/0.90 11.57 32.0%/23.0%/23.5%/21.4% 5 groups 39512.43 39333.81 0.79/1.00/0.83/0.89/1.00 4.72 47.2%/0.09%/24.0%/0.19%/28.5% Abbreviation: BIC: Bayesian information criterion, AIC: Akaike information criterion, AvePP: average posterior probability, Occ: odds of correct classification The FB trajectory groups are shown in Fig. 2 . Trajectory 1 (Negative balance stability) included 344 (31.6%) patients, with FB approximately − 10 mL/kg on day 1 and fluctuating slightly over the subsequent days but generally remaining between − 10 mL/kg and − 5 mL/kg. Trajectory 2 (Rapid transition to negative balance) included 256 (23.5%) patients, with FB approximately 40 mL/kg on day 1 and rapidly decreasing to near 0 mL/kg by day 2 and further dropping to around − 5 mL/kg on days 3 and 4. Trajectory 3 (Positive balance gradual decline) included 261 (24.0%) patients, with FB approximately 15 mL/kg on day 1 and gradually decreasing at a relatively slow rate, remaining above 0 mL/kg throughout the period. Trajectory 4 (High-level decline) included 228 (20.9%) patients, with FB the highest on day 1 at approximately 60 mL/kg, rapidly decreasing to around 10 mL/kg by day 2, and further dropping to near 0 mL/kg on days 3 and 4. The baseline characteristics of each trajectory group are shown in Table 1. There were no significant differences among the groups in terms of age, gender, and Charlson score. Notably, Trajectory 2 (Rapid transition to negative balance) had significantly lower anion gap (AG), creatinine (Cr) values, the highest proportion of ACEI/ARB and beta-blocker use, and significantly lower proportions of CRRT and sepsis. Relationship Between FB Trajectories, FO Status, and Survival Kaplan-Meier survival analysis revealed the survival status of different FB trajectory groups and different fluid load states, as shown in Fig. 3 . Significant differences in survival rates were observed among the trajectory groups (P < 0.001), with Trajectory 1 (Negative balance stability) and Trajectory 2 (Rapid transition to negative balance) having lower mortality rates compared to Trajectory 3 (Positive balance gradual decline) and Trajectory 4 (High-level decline). Additionally, patients with FO had a significantly higher risk of death compared to those without FO (P < 0.001). The results of univariate and multivariate Cox regression analyses are presented in Supplementary Table S1 . Univariate analysis indicated that FB trajectory subtypes were significantly associated with 30-day mortality. Multivariate Cox regression analysis showed that compared to Trajectory 1, FB trajectory subtypes were significantly associated with 30-day mortality (with Trajectory 3 as the reference group, Trajectory 1: HR = 0.71, 95% CI 0.56–0.91, p = 0.006; Trajectory 2: HR = 0.65, 95% CI 0.48–0.87, p = 0.004; Trajectory 4: HR = 0.91, 95% CI 0.70–1.18, p = 0.472). Other significant factors affecting 30-day mortality included the presence of hyperlipidemia (HLP), pneumonia (PNA), chronic kidney disease (CKD), use of ACEI/ARB, epinephrine, norepinephrine, as well as age, weight, red cell distribution width (RDW), anion gap (AG), creatinine (Cr), prothrombin time (PT), partial pressure of carbon dioxide (PaCO₂), partial pressure of oxygen (PaO₂), and Charlson Comorbidity Index (all P < 0.05). Further adjustment for various covariates was conducted using multiple Cox regression models to assess the association between FB trajectories and 30-day survival, as shown in Table 3 . Model 1 did not adjust for any covariates, and compared to Trajectory 3 (Positive balance gradual decline), both Trajectory 1 (Negative balance stability) and Trajectory 2 (Rapid transition to negative balance) had significantly lower 30-day all-cause mortality (Trajectory 1: HR = 0.68, 95% CI 0.54–0.85, p < 0.001; Trajectory 2: HR = 0.42, 95% CI 0.32–0.56, p < 0.001). Model 2 adjusted for age and weight, Model 3 added adjustment for RDW, AG, Cr, PT, PaCO₂, PaO₂, and Charlson Comorbidity Index, and Model 4 further added adjustment for HLP, PNA, CKD, ACEI/ARB, epinephrine, and norepinephrine. In all four Cox regression models with covariate adjustment, Trajectory 1 and Trajectory 2 consistently showed significantly higher survival compared to Trajectory 3, while no significant difference was observed between Trajectory 3 and Trajectory 4. The results of Model 4, which adjusted for multiple confounding factors, showed that Trajectory 1 and Trajectory 2 had significantly higher survival compared to Trajectory 3 (Trajectory 1: HR = 0.67, 95% CI 0.53–0.85, p < 0.001; Trajectory 2: HR = 0.60, 95% CI 0.45–0.80, p < 0.001), while no significant difference was observed between Trajectory 3 and Trajectory 4 (p = 0.35). Subgroup Analysis We conducted stratified subgroup analyses based on the following variables: age, gender, presence of hypotension, SOFA score, presence of CKD, and use of norepinephrine, with results shown in Fig. 4 . No statistically significant interactions were observed between FB trajectory categories and any stratification variables, indicating no interaction effects (P for interaction > 0.05). These results suggest a consistent association between FB trajectory groups and 30-day mortality risk across HFpEF patients with different baseline characteristics. Sensitivity Analysis In terms of sensitivity analysis, we excluded 101 (9.3%) HFpEF patients who underwent CRRT to eliminate the potential impact of CRRT on FB effects. Nevertheless, similar FB trajectory patterns were still observed (Supplementary Fig. S4). Among them, 236 (23.9%) had a positive balance gradual decline, 334 (33.5%) had a negative balance stability, 234 (23.4%) had a rapid transition to negative balance, and 184 (18.4%) had a high-level decline. The 30-day all-cause mortality was lower in the negative balance stability group and rapid transition to negative balance group compared to the positive balance gradual decline and high-level decline groups (p < 0.001) (Supplementary Fig. S5), which was consistent with the results before sensitivity analysis. Discussion This study employed the GBTM method to investigate the FB trends in patients with HfpEF and established a correlation between FB trajectory levels and patient mortality risk. We identified four distinct FB trajectories: Trajectory 1 (Negative balance stability), Trajectory 2 (Rapid transition to negative balance), Trajectory 3 (Positive balance gradual decline), and Trajectory 4 (High-level decline). After adjusting for all confounding factors, it was observed that the 30-day mortality risk was significantly lower in Trajectory 1 and Trajectory 2 compared to Trajectory 3 and Trajectory 4. Similar results were observed in both subgroup analyses and sensitivity analyses. Additionally, we also found that the FO group was associated with adverse clinical outcomes. Fluid management is a fundamental treatment for critically ill patients, aiming to maintain hemodynamic stability, balance electrolytes and acid-base levels, and improve tissue perfusion 15 . In ICU patients with HFpEF, we observed that positive FB is associated with adverse outcomes. Our study results showed that Trajectory 1 and Trajectory 2, which achieved negative FB more rapidly, had better prognoses compared to Trajectory 3 and Trajectory 4. Additionally, patients with FO had significantly worse outcomes. Numerous studies have confirmed that positive FB or FO is associated with an increased risk of adverse clinical events in critically ill patients 16 – 20 . In patients with heart failure, FO is also associated with increased hospitalizations for heart failure and all-cause mortality 21 . Moreover, previous studies have indicated that the severity and rate of fluid accumulation are independent risk factors for mortality in critically ill patients, and FO is not merely a numerical value exceeding a critical threshold; any degree of positive FB is associated with adverse outcomes 16 . Several potential mechanisms may explain the association between positive FB or FO and increased mortality risk in HFpEF. Research suggests that FO is not solely due to excessive fluid and salt intake or inadequate diuretic therapy but may also be related to fluid redistribution, neurohormonal activation, and inflammation 22 . In patients with acute heart failure, pleural fluid overload can affect hemodynamics and is associated with adverse outcomes after discharge 23 . The pathophysiology of FB in heart failure is complex, with the heart, kidneys, and lungs deeply involved in volume regulation and management. Therefore, when addressing FO in heart failure patients, the interactions between these organs should be emphasized and considered 24 . Positive FB in heart failure patients leads to fluid retention, which increases the cardiac burden and subsequently impairs cardiac function 24 . In acute heart failure patients, FO not only causes pulmonary congestion but also leads to systemic congestion, thereby exacerbating the condition 25 . Additionally, FO and positive balance may worsen the condition of heart failure patients by affecting renal function. Fluid retention can lead to renal hypoperfusion, resulting in acute kidney injury, which is highly prevalent in heart failure patients 26 . Studies have shown that FO is closely related to the occurrence of acute kidney injury and often precedes it 27 . Bioimpedance-measured FO indices are associated with significant cardiorenal outcome risks in patients with heart failure and chronic kidney disease. Research indicates a positive correlation between FO and the progression of heart failure events and chronic kidney disease 28 . In heart failure patients, monitoring FO can reduce healthcare costs by decreasing hospital readmission rates and length of stay 29 . Notably, previous studies often relied on single FB values, which may not accurately reflect the fluctuations in FB 7 . In clinical practice, FB status is susceptible to various factors, such as urine output, oral and intravenous fluid intake, and the use of relevant diuretics or hemodynamic agents. FB is dynamic and rapidly evolving; a static FB status alone cannot fully capture the complexity of a patient's fluid state. Our study, by dynamically measuring FB data, can more accurately reflect the FB status and trends in HFpEF patients. The calculation of FB in our study was based on the statistical analysis of fluid input and output in HFpEF patients recorded in the MIMIC database. In addition to our method, there are multiple approaches to assess FB in critically ill patients: digital FB monitoring technology 30 , bioimpedance spectroscopy 31 , and smartphone applications 32 . Compared with these methods, ours has the advantages of being convenient to operate and cost-effective, making it suitable for large-scale clinical application. However, with the continuous advancement of technology, it is anticipated that better methods will emerge to measure patients' FB status more accurately. To the best of our knowledge, this represents the inaugural retrospective cohort study elucidating the impact of longitudinal FB patterns in patients with heart failure with preserved ejection fraction (HFpEF). Utilizing GBTM, we categorized the dynamic FB trajectories among HFpEF patients, thereby facilitating targeted care and early intervention for high-risk cohorts in clinical practice. Moreover, the consistency of results across multiple subgroup and sensitivity analyses further solidifies the robustness of our findings. However, several limitations merit consideration. Firstly, the MIMIC database, being a single-center repository, inherently introduces selection bias and precludes external validation from alternative databases, potentially limiting the generalizability of our conclusions. Secondly, despite comprehensive efforts to adjust for confounding variables, residual confounding may persist, particularly due to the absence of detailed diuretic usage records and the omission of insensible water loss in our analyses. Thirdly, our exclusion of patients with ICU stays shorter than 3 days, aimed at capturing meaningful FB changes, inadvertently reduced the sample size and may have attenuated the observed impact of trajectory indicators on mortality. Lastly, as an observational study, our design permits the establishment of associations between FB trajectories and survival outcomes in HFpEF but is inherently incapable of inferring causality between FB and clinical outcomes. Conclusion This study investigated the relationship between FB trajectories and 30-day mortality in patients with HFpEF. Utilizing GBTM, four distinct FB trajectories were identified, with the negative balance stability group and the moderate-level rapid decline group being associated with higher survival rates. Furthermore, FO was found to increase mortality risk in HFpEF patients. Monitoring FB trajectories may aid in identifying high-risk individuals with HFpEF, and regularly assessing daily fluid status while limiting fluid overload is crucial for the recovery of critically ill patients. Materials and methods Data source The MIMIC-IV database is a publicly available electronic health record database derived from the intensive care units (ICUs) at Beth Israel Deaconess Medical Center 33 . All data in MIMIC-IV are de-identified and cannot be used to identify individual patients. Therefore, this project does not require obtaining written informed consent from patients, nor does it require approval from an ethics committee or institutional review board. The research report strictly adheres to the guidelines of the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement 34 . Study population We included patients with HFpEF. Patients with heart failure in the MIMIC database were identified using ICD9 and ICD10 codes (prefixed with 428, I50), and patients with unknown or 18 years and < 100 years; (2) only first-time ICU admissions; (3) ICU stay of three days or more. The exclusion criteria were: (1) fewer than three FB measurements within seven days; (2) more than 10% of sample variables missing; (3) patients with unknown admission weight. Variables and outcome measures We extracted FB data from days 1 to 7 after HFpEF patients were admitted to the ICU, as well as prognostic indicators. Additionally, the following indicators were included: (1) Demographic data: age, sex, and body weight; (2) Vital signs: body temperature, heart rate, respiratory rate, and blood pressure; (3) Laboratory test indicators: complete blood count, biochemical profile, coagulation function, and blood gas analysis; (4) Comorbidities: acute myocardial infarction, atrial fibrillation, hypertension, diabetes mellitus, hyperlipidemia, chronic obstructive pulmonary disease, pneumonia, and chronic kidney disease; (5) Medication data: Angiotensin-Converting Enzyme Inhibitors/Angiotensin Receptor Blockers(ACEI/ARB), β-blockers, diuretics, and vasopressors; (6) Other indicators: such as whether continuous renal replacement therapy (CRRT) was performed, whether mechanical ventilation was used, whether sepsis or acute kidney injury was present, and severity scores such as the SOFA and Charlson scores. FB was calculated using the following formula: Fluid balance(FB) = (total fluid intake - total fluid output) /initial body weight. Fluid overload (FO) was defined as cumulative FB exceeding 10% of the initial body weight [19]. The primary endpoint was 30-day in-hospital mortality, defined as the survival status of patients within 30 days after admission to the ICU. We used Navicat Premium 17.0 software and Structured Query Language (SQL) to extract indicators from the MIMIC-IV database. There are several points to note regarding data extraction: (1) Continuous fluid intake and output data were recorded from day 1 to day 7 to calculate FB and FO. (2) The left ventricular ejection fraction was based on the minimum value measured at admission. (3) Laboratory indicators were extracted as the mean values within 24 hours of ICU admission. Group-based trajectory model (GBTM) GBTM is a semi-parametric model specifically designed for longitudinal data analysis, capable of effectively identifying populations with similar trajectories of FB changes 35 . In this study, the GBTM method was employed. Initially, trajectories were fitted using cubic polynomials, and the optimal number of groups was determined based on relevant parameters. Subsequently, the significance of linear, quadratic, and cubic polynomials was assessed within the selected number of groups to optimize the trajectory shapes. The criteria for determining the optimal trajectory are as follows: (1) Bayesian Information Criterion (BIC) and Akaike Information Criterion (AIC): The closer the values are to 0, the better the fit; (2) Average Posterior Probability of Group Membership (Avepp): A value of > 0.7 indicates reliable subgroup classification; (3) The proportion of patients in each trajectory group should be > 5%; (4) Odds of Correct Classification (OCC): The minimum OCC value for each group should be greater than 5.0, and higher values are preferred; (5) Additionally, the model's simplicity and clinical interpretability must be taken into account. Statistical analysis In the baseline description, the Shapiro-Wilk test was used to assess the normality of continuous variables. Data that followed a normal distribution were presented as mean ± standard deviation, and comparisons between groups were conducted using one-way analysis of variance (ANOVA). Non-normally distributed data were expressed as median (interquartile range), and comparisons between groups were performed using the Kruskal-Wallis test. Categorical variables were described using frequencies and percentages (%), and comparisons were made using the chi-square (χ²) test or Fisher's exact test. Samples with missing data exceeding 10% were excluded, and missing values in the remaining samples were imputed using the K-Nearest Neighbors (KNN) method. Kaplan-Meier curves were used to compare survival differences between different FB trajectory groups and different FO statuses. Univariate and multivariate forward stepwise Cox regression analyses were employed to evaluate the association between potential trajectory groups and the study outcomes, with entry and removal criteria for variables set at 0.1 and 0.05, respectively. Correlation heatmaps were used to compare the correlations between variables, and the variance inflation factor (VIF) was calculated to assess multicollinearity among covariates. Variables with correlation coefficients > 0.5 or VIF > 5 were subjected to further selection. Cox proportional hazards models were constructed to calculate hazard ratios (HR) and 95% confidence intervals (CI) to evaluate the association between FB trajectories and 30-day mortality in HFpEF patients. Model 1 was unadjusted, Model 2 adjusted for age and weight, Model 3 added covariates including red cell distribution width (RDW), anion gap (AG), creatinine (Cr), prothrombin time (PT), partial pressure of carbon dioxide (PaCO₂), partial pressure of oxygen (PaO₂), and Charlson score; Model 4 further included covariates such as hyperlipidemia (HLP), pneumonia (PNA), chronic kidney disease (CKD), ACEI/ARB, epinephrine, and norepinephrine. We conducted several subgroup analyses to explore potential factors influencing the relationship between FB groups and prognosis in HFpEF patients, including age (≤ 65 years vs. >65 years), sex (male vs. female), blood pressure (< 90 mmHg vs. ≥90 mmHg), SOFA score (< 6 vs. ≥6), presence of CKD (no vs. yes), and use of norepinephrine (no vs. yes). Additionally, we performed sensitivity analyses excluding patients receiving CRRT to verify the robustness of the results. Statistical analyses in this study were conducted using R software (version 4.1.2, www.r-project.org ), and a two-sided P-value < 0.05 was considered statistically significant. Declarations Acknowledgments We thank our colleagues for their valuable contributions to this study. We also appreciate the assistance of medical writers, proofreaders, and editors. Author Contributions Q.Z., C.Z., and G.L. conceived and designed the study. Data collection and analysis were performed by C.Z., G.L., and F.C. Figures and tables were prepared by C.Z., G.L., and F.C. The initial draft of the manuscript was written by C.Z., with contributions to manuscript writing and revision from G.L. All authors reviewed and approved the final manuscript. Funding This work received no financial support. Data Availability Statement The original contributions presented in this study are included in the article/supplementary material, further inquiries can be directed to the corresponding author. Competing Interests Statement The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Ethics declarations All data in MIMIC-IV have been de-identified and cannot be used to identify specific patients. Therefore, this project does not require written informed consent from patients, nor does it need approval from a research ethics committee or institutional review board. References Savarese, G. et al. Global burden of heart failure: a comprehensive and updated review of epidemiology. Cardiovasc Res 118 , 3272-3287 (2023). https://doi.org/10.1093/cvr/cvac013 Ma, C., Luo, H., Fan, L., Liu, X. & Gao, C. Heart failure with preserved ejection fraction: an update on pathophysiology, diagnosis, treatment, and prognosis. Braz J Med Biol Res 53 , e9646 (2020). https://doi.org/10.1590/1414-431x20209646 Cho, D. H. & Yoo, B. S. Current Prevalence, Incidence, and Outcomes of Heart Failure with Preserved Ejection Fraction. Heart Fail Clin 17 , 315-326 (2021). https://doi.org/10.1016/j.hfc.2021.03.002 Hamo, C. E. et al. Heart failure with preserved ejection fraction. Nat Rev Dis Primers 10 , 55 (2024). https://doi.org/10.1038/s41572-024-00540-y Redfield, M. M. & Borlaug, B. A. Heart Failure With Preserved Ejection Fraction: A Review. Jama 329 , 827-838 (2023). https://doi.org/10.1001/jama.2023.2020 Rogers, J. K. et al. Analysing recurrent hospitalizations in heart failure: a review of statistical methodology, with application to CHARM-Preserved. Eur J Heart Fail 16 , 33-40 (2014). https://doi.org/10.1002/ejhf.29 Koell, B. et al. Fluid status and outcome in patients with heart failure and preserved ejection fraction. Int J Cardiol 230 , 476-481 (2017). https://doi.org/10.1016/j.ijcard.2016.12.080 Gori, M. et al. Association between renal function and cardiovascular structure and function in heart failure with preserved ejection fraction. Eur Heart J 35 , 3442-3451 (2014). https://doi.org/10.1093/eurheartj/ehu254 Takeda, M. & Shiba, N. Heart-Lung Interaction and Its Prognostic Significance in Heart Failure Patients With Preserved Ejection Fraction. Circ J 85 , 1435-1437 (2021). https://doi.org/10.1253/circj.CJ-21-0358 Wang, M. et al. Association between Latent Trajectories of Fluid Balance and Clinical Outcomes in Critically Ill Patients with Acute Kidney Injury: A Prospective Multicenter Observational Study. Kidney Dis (Basel) 8 , 82-92 (2022). https://doi.org/10.1159/000515533 Kuo, G. et al. Latent Trajectories of Fluid Balance Are Associated With Outcomes in Cardiac and Aortic Surgery. Ann Thorac Surg 109 , 1343-1349 (2020). https://doi.org/10.1016/j.athoracsur.2019.09.068 Monnet, X., Marik, P. E. & Teboul, J. L. Prediction of fluid responsiveness: an update. Ann Intensive Care 6 , 111 (2016). https://doi.org/10.1186/s13613-016-0216-7 Nagin, D. S., Jones, B. L. & Elmer, J. Recent Advances in Group-Based Trajectory Modeling for Clinical Research. Annu Rev Clin Psychol 20 , 285-305 (2024). https://doi.org/10.1146/annurev-clinpsy-081122-012416 Tie, X. et al. Associations between serum albumin level trajectories and clinical outcomes in sepsis patients in ICU: insights from longitudinal group trajectory modeling. Front Nutr 11 , 1433544 (2024). https://doi.org/10.3389/fnut.2024.1433544 Messmer, A. S. et al. Fluid Overload and Mortality in Adult Critical Care Patients-A Systematic Review and Meta-Analysis of Observational Studies. Crit Care Med 48 , 1862-1870 (2020). https://doi.org/10.1097/ccm.0000000000004617 Garzotto, F. et al. The Dose Response Multicentre Investigation on Fluid Assessment (DoReMIFA) in critically ill patients. Crit Care 20 , 196 (2016). https://doi.org/10.1186/s13054-016-1355-9 Zinter, M. S. et al. Positive Cumulative Fluid Balance Is Associated With Mortality in Pediatric Acute Respiratory Distress Syndrome in the Setting of Acute Kidney Injury. Pediatr Crit Care Med 20 , 323-331 (2019). https://doi.org/10.1097/pcc.0000000000001845 Codes, L., de Souza, Y. G., D'Oliveira, R. A. C., Bastos, J. L. A. & Bittencourt, P. L. Cumulative positive fluid balance is a risk factor for acute kidney injury and requirement for renal replacement therapy after liver transplantation. World J Transplant 8 , 44-51 (2018). https://doi.org/10.5500/wjt.v8.i2.44 Chao, W. C. et al. Association of day 4 cumulative fluid balance with mortality in critically ill patients with influenza: A multicenter retrospective cohort study in Taiwan. PLoS One 13 , e0190952 (2018). https://doi.org/10.1371/journal.pone.0190952 Tang, J., Wu, C. & Zhong, Z. Group-Based Trajectory Modeling of Fluid Balance in Elderly Patients with Acute Ischemic Stroke: Analysis from Multicenter ICUs. Neurol Ther 13 , 749-761 (2024). https://doi.org/10.1007/s40120-024-00612-x Cooper, L. B. et al. The Burden of Congestion in Patients Hospitalized With Acute Decompensated Heart Failure. Am J Cardiol 124 , 545-553 (2019). https://doi.org/10.1016/j.amjcard.2019.05.030 Cotter, G., Metra, M., Milo-Cotter, O., Dittrich, H. C. & Gheorghiade, M. Fluid overload in acute heart failure--re-distribution and other mechanisms beyond fluid accumulation. Eur J Heart Fail 10 , 165-169 (2008). https://doi.org/10.1016/j.ejheart.2008.01.007 Chang, H. C. et al. Nocturnal thoracic volume overload and post-discharge outcomes in patients hospitalized for acute heart failure. ESC Heart Fail 7 , 2807-2817 (2020). https://doi.org/10.1002/ehf2.12881 Cosentino, N. et al. Fluid balance in heart failure. Eur J Prev Cardiol 30 , ii9-ii15 (2023). https://doi.org/10.1093/eurjpc/zwad166 Fiaccadori, E. et al. Ultrafiltration in heart failure. Am Heart J 161 , 439-449 (2011). https://doi.org/10.1016/j.ahj.2010.09.014 Hassinger, A. B., Wald, E. L. & Goodman, D. M. Early postoperative fluid overload precedes acute kidney injury and is associated with higher morbidity in pediatric cardiac surgery patients. Pediatr Crit Care Med 15 , 131-138 (2014). https://doi.org/10.1097/pcc.0000000000000043 Sethi, S. K. et al. Fluid Overload and Renal Angina Index at Admission Are Associated With Worse Outcomes in Critically Ill Children. Front Pediatr 6 , 118 (2018). https://doi.org/10.3389/fped.2018.00118 Mayne, K. J. et al. Bioimpedance Indices of Fluid Overload and Cardiorenal Outcomes in Heart Failure and Chronic Kidney Disease: a Systematic Review. J Card Fail 28 , 1628-1641 (2022). https://doi.org/10.1016/j.cardfail.2022.08.005 Costanzo, M. R., Fonarow, G. C. & Rizzo, J. A. Ultrafiltration versus diuretics for the treatment of fluid overload in patients with heart failure: a hospital cost analysis. J Med Econ 22 , 577-583 (2019). https://doi.org/10.1080/13696998.2019.1584109 Leinum, L. R., Baandrup, A. O., Gögenur, I., Krogsgaard, M. & Azawi, N. Evaluation of a real-life experience with a digital fluid balance monitoring technology. Technol Health Care 32 , 3913-3924 (2024). https://doi.org/10.3233/thc-231303 Dewitte, A. et al. Bioelectrical impedance spectroscopy to estimate fluid balance in critically ill patients. J Clin Monit Comput 30 , 227-233 (2016). https://doi.org/10.1007/s10877-015-9706-7 Shen, Z. et al. A Smart-Phone App for Fluid Balance Monitoring in Patients with Heart Failure: A Usability Study. Patient Prefer Adherence 16 , 1843-1853 (2022). https://doi.org/10.2147/ppa.S373393 Johnson, A. E. W. et al. MIMIC-IV, a freely accessible electronic health record dataset. Sci Data 10 , 1 (2023). https://doi.org/10.1038/s41597-022-01899-x von Elm, E. et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 370 , 1453-1457 (2007). https://doi.org/10.1016/s0140-6736(07)61602-x Nguena Nguefack, H. L. et al. Trajectory Modelling Techniques Useful to Epidemiological Research: A Comparative Narrative Review of Approaches. Clin Epidemiol 12 , 1205-1222 (2020). https://doi.org/10.2147/clep.S265287 Table 1 Table 1 is available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Table1BaselineCharacteristicsofHFpEFPatientswithDifferentFluidBalanceTrajectoryGroups.xlsx Supplementaryinformationfiles.pdf Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 17 Jul, 2025 Editor assigned by journal 17 Jul, 2025 Editor invited by journal 10 Jul, 2025 Submission checks completed at journal 04 Jul, 2025 First submitted to journal 25 Jun, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6956502","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":486726426,"identity":"40a88ec6-ec61-4481-9936-28b218128244","order_by":0,"name":"Chunmei Zhang","email":"","orcid":"","institution":"Zhangzhou Affiliated Hospital of Fujian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Chunmei","middleName":"","lastName":"Zhang","suffix":""},{"id":486726427,"identity":"af61c8fa-98fd-4bdf-a57b-d4831ada7003","order_by":1,"name":"Guangyu Lin","email":"","orcid":"","institution":"Zhangzhou Affiliated Hospital of Fujian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Guangyu","middleName":"","lastName":"Lin","suffix":""},{"id":486726432,"identity":"0a59a653-65fe-4d07-930a-a422dc995b99","order_by":2,"name":"Fengzhen Chen","email":"","orcid":"","institution":"Zhangpu County Traditional Chinese Medicine Hospital","correspondingAuthor":false,"prefix":"","firstName":"Fengzhen","middleName":"","lastName":"Chen","suffix":""},{"id":486726434,"identity":"2d101ca4-b51c-46a3-a923-463c5d3226a2","order_by":3,"name":"Qitian Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4UlEQVRIiWNgGAWjYBACNvb2AwYSFTZybMz8Hx8kVNQQ1sLHcyahwOJMmjE/O4OxwYMzxwhrkZNIMPhQ2XY4cWY/g5nkwxZmIhzGcyBxw822tMQNhxnSKhIb2Bj427sTCPil8bDhjHM2xkAtx24k7pBhkDhzdgMhW9KMJcrSZDccZmy7kXiGjcFAIpeAFokE899/2A4zbjjMzFaQ2MZMlBYDA4m2w4ozm9nYGIjTAgxkAwlQIDPzMEsknDnGQ9Av8u2wqOQ/w/jxR0WNHH97L34tGICHNOWjYBSMglEwCrACAOjASxa6epPPAAAAAElFTkSuQmCC","orcid":"","institution":"Zhangzhou Affiliated Hospital of Fujian Medical University","correspondingAuthor":true,"prefix":"","firstName":"Qitian","middleName":"","lastName":"Zhang","suffix":""}],"badges":[],"createdAt":"2025-06-23 11:53:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6956502/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6956502/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":87378149,"identity":"3db274e9-b876-46f7-afd8-a3cf6b891975","added_by":"auto","created_at":"2025-07-23 08:22:13","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":534059,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFlowchart of HFpEF patient selection process\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Figure1FlowchartofHFpEFpatientselectionprocess.png","url":"https://assets-eu.researchsquare.com/files/rs-6956502/v1/366690b7ed2c39f542308b97.png"},{"id":87378156,"identity":"e1d47e66-9395-457a-8219-777903e19de5","added_by":"auto","created_at":"2025-07-23 08:22:13","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":142362,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFluid balance trajectories in patients with HFpEF\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Figure2FluidbalancetrajectoriesinpatientswithHFpEF.png","url":"https://assets-eu.researchsquare.com/files/rs-6956502/v1/5f851079ac0e40ea7b819538.png"},{"id":87381557,"identity":"93225a8a-de4b-4357-bab1-83fd09f4717b","added_by":"auto","created_at":"2025-07-23 08:38:13","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":598199,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eK–M survival curves showing the relationship between different HFpEF groups and 30-day mortality A. By FB trajectory groups; B. By FO status\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Figure3KMsurvivalcurvesshowingtherelationshipbetweendifferentHFpEFgroupsand30daymortalityaByFBtrajectorygroupsbByFOstatus.png","url":"https://assets-eu.researchsquare.com/files/rs-6956502/v1/78bf1e3e0d2e738408a67f44.png"},{"id":87378150,"identity":"76a8bc91-92b1-475b-a489-c671ef22808f","added_by":"auto","created_at":"2025-07-23 08:22:13","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":876153,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eForest Subgroup Analysis FB Trajectories and 30-Day ICU Mortality in HFpEF Patients\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Figure4ForestSubgroupAnalysisFBTrajectoriesand30DayICUMortalityinHFpEFPatients.png","url":"https://assets-eu.researchsquare.com/files/rs-6956502/v1/2563cd2b979fc2646d26f8ad.png"},{"id":87386100,"identity":"f13711bd-01e5-416f-b716-7a385acd9098","added_by":"auto","created_at":"2025-07-23 08:54:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2877788,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6956502/v1/c9e8eb19-151c-47e9-b722-c52fd82549b5.pdf"},{"id":87378148,"identity":"0bdf8fc9-418d-4c56-a4f0-dedf548a10c0","added_by":"auto","created_at":"2025-07-23 08:22:13","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":19458,"visible":true,"origin":"","legend":"","description":"","filename":"Table1BaselineCharacteristicsofHFpEFPatientswithDifferentFluidBalanceTrajectoryGroups.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6956502/v1/c9e3fed4724571fe6d327440.xlsx"},{"id":87378151,"identity":"a80015ab-f342-48de-a9d0-1be06b573621","added_by":"auto","created_at":"2025-07-23 08:22:13","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":770006,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementaryinformationfiles.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6956502/v1/452ae2f6e28748354aa73d2d.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Association Between Fluid Balance Trajectories and Prognosis in Patients with Heart Failure and Preserved Ejection Fraction","fulltext":[{"header":"Introduction","content":"\u003cp\u003eHeart failure (HF) is a complex clinical syndrome affecting over 60\u0026nbsp;million patients worldwide\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Heart failure with preserved ejection fraction (HFpEF) is a type of heart failure characterized by the presence of signs and symptoms of heart failure with a normal or near-normal left ventricular ejection fraction (\u0026ge;\u0026thinsp;50%)\u003csup\u003e2\u003c/sup\u003e. The global prevalence of HFpEF is approximately 2%\u003csup\u003e3\u003c/sup\u003e, accounting for nearly half of all heart failure cases, and its prevalence is expected to continue rising with the aging of the population\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. HFpEF is associated with high rates of hospitalization and mortality, with hospitalization and mortality rates as high as 80% and 50% within 5 years, respectively\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. Furthermore, due to the frequent readmissions of HFpEF patients, the overall hospitalization burden remains high\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003ePatients with HFpEF often exhibit symptoms of fluid overload(FO), which is an independent adverse prognostic indicator in HFpEF patients\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. FO is associated with cardiac remodeling and subtle contractile dysfunction, and renal function and fluid status may jointly influence the progression and prognosis of HFpEF\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. The interaction between the heart and lungs may affect fluid balance(FB) and cardiac function, further complicating the management of HFpEF\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. Current research on the relationship between FB status and the prognosis of HFpEF patients is very limited. Although studies have shown that HFpEF patients with FO have a poor prognosis\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e, these studies use a single baseline fluid status for assessment. However, defining FO at a single arbitrary time point may not provide sufficient information to explore the impact of FO on clinical outcomes\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. A recent study of patients undergoing cardiac or aortic surgery showed that changes in FB trends are related to clinical outcomes\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Fluid management is an essential component of critical patient care, and although various strategies for assessing fluid responsiveness have been developed\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e, the optimal fluid management plan remains controversial.\u003c/p\u003e\u003cp\u003eGroup-based trajectory modeling (GBTM) can more accurately describe and understand the heterogeneity and similarities among individuals\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e, and is used to explore dynamic changes among individuals\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. GBTM has been widely applied in medical and psychological fields\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. This study aims to use the GBTM method to analyze the FB trajectories of patients with HFpEF, revealing the association between different trajectory patterns and patient prognosis. This study can help optimize fluid management strategies for HFpEF patients, provide theoretical support for early intervention and personalized treatment by accurately identifying high-risk patient groups, thereby improving patient outcomes.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\n \u003ch2\u003eBaseline Characteristics\u003c/h2\u003e\n \u003cp\u003eA total of 19,969 heart failure patients were screened from the MIMIC-IV database. After applying the exclusion criteria, 1,089 patients with HFpEF were ultimately included to form the study cohort (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e). Variables with excessive missing data (Supplementary Fig. \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003e), high correlation (Supplementary Fig. S2), and strong collinearity (Supplementary Fig. S3) were excluded. Table 1 presents the baseline characteristics of the patients. The median age of all HFpEF patients was 75.0 years, with 592 (54.4%) being male. A total of 1,062 (97.5%) patients received mechanical ventilation, 101 (9.3%) underwent continuous renal replacement therapy (CRRT), and 617 (56.7%) had atrial fibrillation. The prevalence of comorbidities included hypertension in 816 (75.0%), diabetes mellitus in 447 (41.1%), and hyperlipidemia in 521 (47.8%). The admission scores to the ICU were a SOFA score of 6.0 and a Charlson score of 7.00.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eFB Trajectory Description\u003c/h3\u003e\n\u003cp\u003ePrior to grouping the FB trajectories, we pre-analyzed the FB status from day 1 to day 7. However, due to a significant proportion of missing values, particularly on days 4 to 7 (6.9%, 28.4%, 45.5%, and 57.0%, respectively), we only included data from days 1 to 4 in the analysis. Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e presents the model fit statistics and AvePP used to determine the optimal number of FB trajectory groups. Although the AIC and BIC were smallest for the 5-group model, the proportion of patients in some groups was less than 5%, which was not acceptable. The 4-group model had AIC and BIC values comparable to the 5-group model, with all group proportions exceeding 5%. The AvePP for each group was greater than 0.7, and the minimum OCC performed best among all groupings. Considering clinical relevance and interpretability, the 4-group trajectory model was ultimately selected.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"709\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\" style=\"width: 709px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 2. Performance of the group-based trajectory model for fluid balance trajectories\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTrajectories\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBIC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAIC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAvePP\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 111px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMinimum OCC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eClass proportion\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e1 group\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e40699.98\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e40668.08\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 161px;\"\u003e\n \u003cp\u003e1.00\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 111px;\"\u003e\n \u003cp\u003eNaN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003e\n \u003cp\u003e100.0%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e2 groups\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e39894.87\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e39831.08\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 161px;\"\u003e\n \u003cp\u003e0.89/0.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 111px;\"\u003e\n \u003cp\u003e5.60\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003e\n \u003cp\u003e62.0%/38.0%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e3 groups\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e39760.49\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e39664.80\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 161px;\"\u003e\n \u003cp\u003e0.89/0.87/0.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 111px;\"\u003e\n \u003cp\u003e6.39\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003e\n \u003cp\u003e60.1%/24.0%/15.9%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e4 groups\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e39521.51\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e39381.16\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 161px;\"\u003e\n \u003cp\u003e0.87/0.79/0.78/0.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 111px;\"\u003e\n \u003cp\u003e11.57\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003e\n \u003cp\u003e32.0%/23.0%/23.5%/21.4%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 86px;\"\u003e\n \u003cp\u003e5 groups\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e39512.43\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 62px;\"\u003e\n \u003cp\u003e39333.81\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 161px;\"\u003e\n \u003cp\u003e0.79/1.00/0.83/0.89/1.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 111px;\"\u003e\n \u003cp\u003e4.72\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003e\n \u003cp\u003e47.2%/0.09%/24.0%/0.19%/28.5%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\" style=\"width: 709px;\"\u003e\n \u003cp\u003eAbbreviation: BIC: Bayesian information criterion, AIC: Akaike information criterion, AvePP: average posterior probability, Occ: odds of correct classification\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003cdiv align=\"left\" class=\"colspec\"\u003e\u003cbr\u003e\u003c/div\u003e\n \u003cdiv align=\"left\" class=\"colspec\"\u003e\u003cimg src=\"https://myfiles.space/user_files/122228_c8a1650c59388082/122228_custom_files/img1753116021.png\"\u003e\u003c/div\u003e\n \u003cdiv align=\"left\" class=\"colspec\"\u003eThe FB trajectory groups are shown in Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e. Trajectory 1 (Negative balance stability) included 344 (31.6%) patients, with FB approximately \u0026minus;\u0026thinsp;10 mL/kg on day 1 and fluctuating slightly over the subsequent days but generally remaining between \u0026minus;\u0026thinsp;10 mL/kg and \u0026minus;\u0026thinsp;5 mL/kg. Trajectory 2 (Rapid transition to negative balance) included 256 (23.5%) patients, with FB approximately 40 mL/kg on day 1 and rapidly decreasing to near 0 mL/kg by day 2 and further dropping to around \u0026minus;\u0026thinsp;5 mL/kg on days 3 and 4. Trajectory 3 (Positive balance gradual decline) included 261 (24.0%) patients, with FB approximately 15 mL/kg on day 1 and gradually decreasing at a relatively slow rate, remaining above 0 mL/kg throughout the period. Trajectory 4 (High-level decline) included 228 (20.9%) patients, with FB the highest on day 1 at approximately 60 mL/kg, rapidly decreasing to around 10 mL/kg by day 2, and further dropping to near 0 mL/kg on days 3 and 4. The baseline characteristics of each trajectory group are shown in Table 1. There were no significant differences among the groups in terms of age, gender, and Charlson score. Notably, Trajectory 2 (Rapid transition to negative balance) had significantly lower anion gap (AG), creatinine (Cr) values, the highest proportion of ACEI/ARB and beta-blocker use, and significantly lower proportions of CRRT and sepsis.\u003c/div\u003e\n\u003c/div\u003e\n\u003ch3\u003eRelationship Between FB Trajectories, FO Status, and Survival\u003c/h3\u003e\n\u003cp\u003eKaplan-Meier survival analysis revealed the survival status of different FB trajectory groups and different fluid load states, as shown in Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e. Significant differences in survival rates were observed among the trajectory groups (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), with Trajectory 1 (Negative balance stability) and Trajectory 2 (Rapid transition to negative balance) having lower mortality rates compared to Trajectory 3 (Positive balance gradual decline) and Trajectory 4 (High-level decline). Additionally, patients with FO had a significantly higher risk of death compared to those without FO (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e\n\u003cp\u003eThe results of univariate and multivariate Cox regression analyses are presented in Supplementary Table \u003cspan class=\"InternalRef\"\u003eS1\u003c/span\u003e. Univariate analysis indicated that FB trajectory subtypes were significantly associated with 30-day mortality. Multivariate Cox regression analysis showed that compared to Trajectory 1, FB trajectory subtypes were significantly associated with 30-day mortality (with Trajectory 3 as the reference group, Trajectory 1: HR\u0026thinsp;=\u0026thinsp;0.71, 95% CI 0.56\u0026ndash;0.91, p\u0026thinsp;=\u0026thinsp;0.006; Trajectory 2: HR\u0026thinsp;=\u0026thinsp;0.65, 95% CI 0.48\u0026ndash;0.87, p\u0026thinsp;=\u0026thinsp;0.004; Trajectory 4: HR\u0026thinsp;=\u0026thinsp;0.91, 95% CI 0.70\u0026ndash;1.18, p\u0026thinsp;=\u0026thinsp;0.472). Other significant factors affecting 30-day mortality included the presence of hyperlipidemia (HLP), pneumonia (PNA), chronic kidney disease (CKD), use of ACEI/ARB, epinephrine, norepinephrine, as well as age, weight, red cell distribution width (RDW), anion gap (AG), creatinine (Cr), prothrombin time (PT), partial pressure of carbon dioxide (PaCO₂), partial pressure of oxygen (PaO₂), and Charlson Comorbidity Index (all P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Further adjustment for various covariates was conducted using multiple Cox regression models to assess the association between FB trajectories and 30-day survival, as shown in Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e. Model 1 did not adjust for any covariates, and compared to Trajectory 3 (Positive balance gradual decline), both Trajectory 1 (Negative balance stability) and Trajectory 2 (Rapid transition to negative balance) had significantly lower 30-day all-cause mortality (Trajectory 1: HR\u0026thinsp;=\u0026thinsp;0.68, 95% CI 0.54\u0026ndash;0.85, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; Trajectory 2: HR\u0026thinsp;=\u0026thinsp;0.42, 95% CI 0.32\u0026ndash;0.56, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Model 2 adjusted for age and weight, Model 3 added adjustment for RDW, AG, Cr, PT, PaCO₂, PaO₂, and Charlson Comorbidity Index, and Model 4 further added adjustment for HLP, PNA, CKD, ACEI/ARB, epinephrine, and norepinephrine. In all four Cox regression models with covariate adjustment, Trajectory 1 and Trajectory 2 consistently showed significantly higher survival compared to Trajectory 3, while no significant difference was observed between Trajectory 3 and Trajectory 4. The results of Model 4, which adjusted for multiple confounding factors, showed that Trajectory 1 and Trajectory 2 had significantly higher survival compared to Trajectory 3 (Trajectory 1: HR\u0026thinsp;=\u0026thinsp;0.67, 95% CI 0.53\u0026ndash;0.85, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; Trajectory 2: HR\u0026thinsp;=\u0026thinsp;0.60, 95% CI 0.45\u0026ndash;0.80, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), while no significant difference was observed between Trajectory 3 and Trajectory 4 (p\u0026thinsp;=\u0026thinsp;0.35).\u003c/p\u003e\n\u003ch3\u003eSubgroup Analysis\u003c/h3\u003e\n\u003cp\u003eWe conducted stratified subgroup analyses based on the following variables: age, gender, presence of hypotension, SOFA score, presence of CKD, and use of norepinephrine, with results shown in Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e. No statistically significant interactions were observed between FB trajectory categories and any stratification variables, indicating no interaction effects (P for interaction\u0026thinsp;\u0026gt;\u0026thinsp;0.05). These results suggest a consistent association between FB trajectory groups and 30-day mortality risk across HFpEF patients with different baseline characteristics.\u003c/p\u003e\n\u003cp\u003eSensitivity Analysis\u003c/p\u003e\n\u003cp\u003eIn terms of sensitivity analysis, we excluded 101 (9.3%) HFpEF patients who underwent CRRT to eliminate the potential impact of CRRT on FB effects. Nevertheless, similar FB trajectory patterns were still observed (Supplementary Fig. S4). Among them, 236 (23.9%) had a positive balance gradual decline, 334 (33.5%) had a negative balance stability, 234 (23.4%) had a rapid transition to negative balance, and 184 (18.4%) had a high-level decline. The 30-day all-cause mortality was lower in the negative balance stability group and rapid transition to negative balance group compared to the positive balance gradual decline and high-level decline groups (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Supplementary Fig. S5), which was consistent with the results before sensitivity analysis.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study employed the GBTM method to investigate the FB trends in patients with HfpEF and established a correlation between FB trajectory levels and patient mortality risk. We identified four distinct FB trajectories: Trajectory 1 (Negative balance stability), Trajectory 2 (Rapid transition to negative balance), Trajectory 3 (Positive balance gradual decline), and Trajectory 4 (High-level decline). After adjusting for all confounding factors, it was observed that the 30-day mortality risk was significantly lower in Trajectory 1 and Trajectory 2 compared to Trajectory 3 and Trajectory 4. Similar results were observed in both subgroup analyses and sensitivity analyses. Additionally, we also found that the FO group was associated with adverse clinical outcomes.\u003c/p\u003e\u003cp\u003eFluid management is a fundamental treatment for critically ill patients, aiming to maintain hemodynamic stability, balance electrolytes and acid-base levels, and improve tissue perfusion\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. In ICU patients with HFpEF, we observed that positive FB is associated with adverse outcomes. Our study results showed that Trajectory 1 and Trajectory 2, which achieved negative FB more rapidly, had better prognoses compared to Trajectory 3 and Trajectory 4. Additionally, patients with FO had significantly worse outcomes. Numerous studies have confirmed that positive FB or FO is associated with an increased risk of adverse clinical events in critically ill patients \u003csup\u003e\u003cspan additionalcitationids=\"CR17 CR18 CR19\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. In patients with heart failure, FO is also associated with increased hospitalizations for heart failure and all-cause mortality\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. Moreover, previous studies have indicated that the severity and rate of fluid accumulation are independent risk factors for mortality in critically ill patients, and FO is not merely a numerical value exceeding a critical threshold; any degree of positive FB is associated with adverse outcomes\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eSeveral potential mechanisms may explain the association between positive FB or FO and increased mortality risk in HFpEF. Research suggests that FO is not solely due to excessive fluid and salt intake or inadequate diuretic therapy but may also be related to fluid redistribution, neurohormonal activation, and inflammation\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. In patients with acute heart failure, pleural fluid overload can affect hemodynamics and is associated with adverse outcomes after discharge\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. The pathophysiology of FB in heart failure is complex, with the heart, kidneys, and lungs deeply involved in volume regulation and management. Therefore, when addressing FO in heart failure patients, the interactions between these organs should be emphasized and considered\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. Positive FB in heart failure patients leads to fluid retention, which increases the cardiac burden and subsequently impairs cardiac function\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. In acute heart failure patients, FO not only causes pulmonary congestion but also leads to systemic congestion, thereby exacerbating the condition\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. Additionally, FO and positive balance may worsen the condition of heart failure patients by affecting renal function. Fluid retention can lead to renal hypoperfusion, resulting in acute kidney injury, which is highly prevalent in heart failure patients\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. Studies have shown that FO is closely related to the occurrence of acute kidney injury and often precedes it\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. Bioimpedance-measured FO indices are associated with significant cardiorenal outcome risks in patients with heart failure and chronic kidney disease. Research indicates a positive correlation between FO and the progression of heart failure events and chronic kidney disease\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. In heart failure patients, monitoring FO can reduce healthcare costs by decreasing hospital readmission rates and length of stay\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eNotably, previous studies often relied on single FB values, which may not accurately reflect the fluctuations in FB\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. In clinical practice, FB status is susceptible to various factors, such as urine output, oral and intravenous fluid intake, and the use of relevant diuretics or hemodynamic agents. FB is dynamic and rapidly evolving; a static FB status alone cannot fully capture the complexity of a patient's fluid state. Our study, by dynamically measuring FB data, can more accurately reflect the FB status and trends in HFpEF patients. The calculation of FB in our study was based on the statistical analysis of fluid input and output in HFpEF patients recorded in the MIMIC database. In addition to our method, there are multiple approaches to assess FB in critically ill patients: digital FB monitoring technology\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e, bioimpedance spectroscopy\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e, and smartphone applications\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. Compared with these methods, ours has the advantages of being convenient to operate and cost-effective, making it suitable for large-scale clinical application. However, with the continuous advancement of technology, it is anticipated that better methods will emerge to measure patients' FB status more accurately.\u003c/p\u003e\u003cp\u003eTo the best of our knowledge, this represents the inaugural retrospective cohort study elucidating the impact of longitudinal FB patterns in patients with heart failure with preserved ejection fraction (HFpEF). Utilizing GBTM, we categorized the dynamic FB trajectories among HFpEF patients, thereby facilitating targeted care and early intervention for high-risk cohorts in clinical practice. Moreover, the consistency of results across multiple subgroup and sensitivity analyses further solidifies the robustness of our findings. However, several limitations merit consideration. Firstly, the MIMIC database, being a single-center repository, inherently introduces selection bias and precludes external validation from alternative databases, potentially limiting the generalizability of our conclusions. Secondly, despite comprehensive efforts to adjust for confounding variables, residual confounding may persist, particularly due to the absence of detailed diuretic usage records and the omission of insensible water loss in our analyses. Thirdly, our exclusion of patients with ICU stays shorter than 3 days, aimed at capturing meaningful FB changes, inadvertently reduced the sample size and may have attenuated the observed impact of trajectory indicators on mortality. Lastly, as an observational study, our design permits the establishment of associations between FB trajectories and survival outcomes in HFpEF but is inherently incapable of inferring causality between FB and clinical outcomes.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study investigated the relationship between FB trajectories and 30-day mortality in patients with HFpEF. Utilizing GBTM, four distinct FB trajectories were identified, with the negative balance stability group and the moderate-level rapid decline group being associated with higher survival rates. Furthermore, FO was found to increase mortality risk in HFpEF patients. Monitoring FB trajectories may aid in identifying high-risk individuals with HFpEF, and regularly assessing daily fluid status while limiting fluid overload is crucial for the recovery of critically ill patients.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eData source\u003c/h2\u003e\u003cp\u003eThe MIMIC-IV database is a publicly available electronic health record database derived from the intensive care units (ICUs) at Beth Israel Deaconess Medical Center\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. All data in MIMIC-IV are de-identified and cannot be used to identify individual patients. Therefore, this project does not require obtaining written informed consent from patients, nor does it require approval from an ethics committee or institutional review board. The research report strictly adheres to the guidelines of the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eStudy population\u003c/h2\u003e\u003cp\u003eWe included patients with HFpEF. Patients with heart failure in the MIMIC database were identified using ICD9 and ICD10 codes (prefixed with 428, I50), and patients with unknown or \u0026lt;\u0026thinsp;50% left ventricular ejection fraction (LVEF) were excluded using the note module in the dataset. In addition, our inclusion criteria were: (1) age\u0026thinsp;\u0026gt;\u0026thinsp;18 years and \u0026lt;\u0026thinsp;100 years; (2) only first-time ICU admissions; (3) ICU stay of three days or more. The exclusion criteria were: (1) fewer than three FB measurements within seven days; (2) more than 10% of sample variables missing; (3) patients with unknown admission weight.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003eVariables and outcome measures\u003c/h2\u003e\u003cp\u003eWe extracted FB data from days 1 to 7 after HFpEF patients were admitted to the ICU, as well as prognostic indicators. Additionally, the following indicators were included: (1) Demographic data: age, sex, and body weight; (2) Vital signs: body temperature, heart rate, respiratory rate, and blood pressure; (3) Laboratory test indicators: complete blood count, biochemical profile, coagulation function, and blood gas analysis; (4) Comorbidities: acute myocardial infarction, atrial fibrillation, hypertension, diabetes mellitus, hyperlipidemia, chronic obstructive pulmonary disease, pneumonia, and chronic kidney disease; (5) Medication data: Angiotensin-Converting Enzyme Inhibitors/Angiotensin Receptor Blockers(ACEI/ARB), β-blockers, diuretics, and vasopressors; (6) Other indicators: such as whether continuous renal replacement therapy (CRRT) was performed, whether mechanical ventilation was used, whether sepsis or acute kidney injury was present, and severity scores such as the SOFA and Charlson scores. FB was calculated using the following formula: Fluid balance(FB) = (total fluid intake - total fluid output) /initial body weight. Fluid overload (FO) was defined as cumulative FB exceeding 10% of the initial body weight [19]. The primary endpoint was 30-day in-hospital mortality, defined as the survival status of patients within 30 days after admission to the ICU.\u003c/p\u003e\u003cp\u003eWe used Navicat Premium 17.0 software and Structured Query Language (SQL) to extract indicators from the MIMIC-IV database. There are several points to note regarding data extraction: (1) Continuous fluid intake and output data were recorded from day 1 to day 7 to calculate FB and FO. (2) The left ventricular ejection fraction was based on the minimum value measured at admission. (3) Laboratory indicators were extracted as the mean values within 24 hours of ICU admission.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eGroup-based trajectory model (GBTM)\u003c/h2\u003e\u003cp\u003eGBTM is a semi-parametric model specifically designed for longitudinal data analysis, capable of effectively identifying populations with similar trajectories of FB changes\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. In this study, the GBTM method was employed. Initially, trajectories were fitted using cubic polynomials, and the optimal number of groups was determined based on relevant parameters. Subsequently, the significance of linear, quadratic, and cubic polynomials was assessed within the selected number of groups to optimize the trajectory shapes.\u003c/p\u003e\u003cp\u003eThe criteria for determining the optimal trajectory are as follows: (1) Bayesian Information Criterion (BIC) and Akaike Information Criterion (AIC): The closer the values are to 0, the better the fit; (2) Average Posterior Probability of Group Membership (Avepp): A value of \u0026gt;\u0026thinsp;0.7 indicates reliable subgroup classification; (3) The proportion of patients in each trajectory group should be \u0026gt;\u0026thinsp;5%; (4) Odds of Correct Classification (OCC): The minimum OCC value for each group should be greater than 5.0, and higher values are preferred; (5) Additionally, the model's simplicity and clinical interpretability must be taken into account.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eStatistical analysis\u003c/h2\u003e\u003cp\u003eIn the baseline description, the Shapiro-Wilk test was used to assess the normality of continuous variables. Data that followed a normal distribution were presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation, and comparisons between groups were conducted using one-way analysis of variance (ANOVA). Non-normally distributed data were expressed as median (interquartile range), and comparisons between groups were performed using the Kruskal-Wallis test. Categorical variables were described using frequencies and percentages (%), and comparisons were made using the chi-square (χ\u0026sup2;) test or Fisher's exact test. Samples with missing data exceeding 10% were excluded, and missing values in the remaining samples were imputed using the K-Nearest Neighbors (KNN) method. Kaplan-Meier curves were used to compare survival differences between different FB trajectory groups and different FO statuses. Univariate and multivariate forward stepwise Cox regression analyses were employed to evaluate the association between potential trajectory groups and the study outcomes, with entry and removal criteria for variables set at 0.1 and 0.05, respectively. Correlation heatmaps were used to compare the correlations between variables, and the variance inflation factor (VIF) was calculated to assess multicollinearity among covariates. Variables with correlation coefficients\u0026thinsp;\u0026gt;\u0026thinsp;0.5 or VIF\u0026thinsp;\u0026gt;\u0026thinsp;5 were subjected to further selection. Cox proportional hazards models were constructed to calculate hazard ratios (HR) and 95% confidence intervals (CI) to evaluate the association between FB trajectories and 30-day mortality in HFpEF patients. Model 1 was unadjusted, Model 2 adjusted for age and weight, Model 3 added covariates including red cell distribution width (RDW), anion gap (AG), creatinine (Cr), prothrombin time (PT), partial pressure of carbon dioxide (PaCO₂), partial pressure of oxygen (PaO₂), and Charlson score; Model 4 further included covariates such as hyperlipidemia (HLP), pneumonia (PNA), chronic kidney disease (CKD), ACEI/ARB, epinephrine, and norepinephrine. We conducted several subgroup analyses to explore potential factors influencing the relationship between FB groups and prognosis in HFpEF patients, including age (\u0026le;\u0026thinsp;65 years vs. \u0026gt;65 years), sex (male vs. female), blood pressure (\u0026lt;\u0026thinsp;90 mmHg vs. \u0026ge;90 mmHg), SOFA score (\u0026lt;\u0026thinsp;6 vs. \u0026ge;6), presence of CKD (no vs. yes), and use of norepinephrine (no vs. yes). Additionally, we performed sensitivity analyses excluding patients receiving CRRT to verify the robustness of the results. Statistical analyses in this study were conducted using R software (version 4.1.2, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ewww.r-project.org\u003c/span\u003e\u003c/span\u003e), and a two-sided P-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant.\u003c/p\u003e\u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank our colleagues for their valuable contributions to this study. We also appreciate the assistance of medical writers, proofreaders, and editors.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eQ.Z., C.Z., and G.L. conceived and designed the study. Data collection and analysis were performed by C.Z., G.L., and F.C. Figures and tables were prepared by C.Z., G.L., and F.C. The initial draft of the manuscript was written by C.Z., with contributions to manuscript writing and revision from G.L. All authors reviewed and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work received no financial support.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe original contributions presented in this study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eCompeting Interests Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eEthics declarations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data in MIMIC-IV have been de-identified and cannot be used to identify specific patients. Therefore, this project does not require written informed consent from patients, nor does it need approval from a research ethics committee or institutional review board.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eSavarese, G.\u003cem\u003e et al.\u003c/em\u003e Global burden of heart failure: a comprehensive and updated review of epidemiology. \u003cem\u003eCardiovasc Res\u003c/em\u003e \u003cstrong\u003e118\u003c/strong\u003e, 3272-3287 (2023). https://doi.org/10.1093/cvr/cvac013\u003c/li\u003e\n\u003cli\u003eMa, C., Luo, H., Fan, L., Liu, X. \u0026amp; Gao, C. Heart failure with preserved ejection fraction: an update on pathophysiology, diagnosis, treatment, and prognosis. \u003cem\u003eBraz J Med Biol Res\u003c/em\u003e \u003cstrong\u003e53\u003c/strong\u003e, e9646 (2020). https://doi.org/10.1590/1414-431x20209646\u003c/li\u003e\n\u003cli\u003eCho, D. H. \u0026amp; Yoo, B. S. Current Prevalence, Incidence, and Outcomes of Heart Failure with Preserved Ejection Fraction. \u003cem\u003eHeart Fail Clin\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 315-326 (2021). https://doi.org/10.1016/j.hfc.2021.03.002\u003c/li\u003e\n\u003cli\u003eHamo, C. E.\u003cem\u003e et al.\u003c/em\u003e Heart failure with preserved ejection fraction. \u003cem\u003eNat Rev Dis Primers\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 55 (2024). https://doi.org/10.1038/s41572-024-00540-y\u003c/li\u003e\n\u003cli\u003eRedfield, M. M. \u0026amp; Borlaug, B. A. Heart Failure With Preserved Ejection Fraction: A Review. \u003cem\u003eJama\u003c/em\u003e \u003cstrong\u003e329\u003c/strong\u003e, 827-838 (2023). https://doi.org/10.1001/jama.2023.2020\u003c/li\u003e\n\u003cli\u003eRogers, J. K.\u003cem\u003e et al.\u003c/em\u003e Analysing recurrent hospitalizations in heart failure: a review of statistical methodology, with application to CHARM-Preserved. \u003cem\u003eEur J Heart Fail\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 33-40 (2014). https://doi.org/10.1002/ejhf.29\u003c/li\u003e\n\u003cli\u003eKoell, B.\u003cem\u003e et al.\u003c/em\u003e Fluid status and outcome in patients with heart failure and preserved ejection fraction. \u003cem\u003eInt J Cardiol\u003c/em\u003e \u003cstrong\u003e230\u003c/strong\u003e, 476-481 (2017). https://doi.org/10.1016/j.ijcard.2016.12.080\u003c/li\u003e\n\u003cli\u003eGori, M.\u003cem\u003e et al.\u003c/em\u003e Association between renal function and cardiovascular structure and function in heart failure with preserved ejection fraction. \u003cem\u003eEur Heart J\u003c/em\u003e \u003cstrong\u003e35\u003c/strong\u003e, 3442-3451 (2014). https://doi.org/10.1093/eurheartj/ehu254\u003c/li\u003e\n\u003cli\u003eTakeda, M. \u0026amp; Shiba, N. Heart-Lung Interaction and Its Prognostic Significance in Heart Failure Patients With Preserved Ejection Fraction. \u003cem\u003eCirc J\u003c/em\u003e \u003cstrong\u003e85\u003c/strong\u003e, 1435-1437 (2021). https://doi.org/10.1253/circj.CJ-21-0358\u003c/li\u003e\n\u003cli\u003eWang, M.\u003cem\u003e et al.\u003c/em\u003e Association between Latent Trajectories of Fluid Balance and Clinical Outcomes in Critically Ill Patients with Acute Kidney Injury: A Prospective Multicenter Observational Study. \u003cem\u003eKidney Dis (Basel)\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 82-92 (2022). https://doi.org/10.1159/000515533\u003c/li\u003e\n\u003cli\u003eKuo, G.\u003cem\u003e et al.\u003c/em\u003e Latent Trajectories of Fluid Balance Are Associated With Outcomes in Cardiac and Aortic Surgery. \u003cem\u003eAnn Thorac Surg\u003c/em\u003e \u003cstrong\u003e109\u003c/strong\u003e, 1343-1349 (2020). https://doi.org/10.1016/j.athoracsur.2019.09.068\u003c/li\u003e\n\u003cli\u003eMonnet, X., Marik, P. E. \u0026amp; Teboul, J. L. Prediction of fluid responsiveness: an update. \u003cem\u003eAnn Intensive Care\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 111 (2016). https://doi.org/10.1186/s13613-016-0216-7\u003c/li\u003e\n\u003cli\u003eNagin, D. S., Jones, B. L. \u0026amp; Elmer, J. Recent Advances in Group-Based Trajectory Modeling for Clinical Research. \u003cem\u003eAnnu Rev Clin Psychol\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 285-305 (2024). https://doi.org/10.1146/annurev-clinpsy-081122-012416\u003c/li\u003e\n\u003cli\u003eTie, X.\u003cem\u003e et al.\u003c/em\u003e Associations between serum albumin level trajectories and clinical outcomes in sepsis patients in ICU: insights from longitudinal group trajectory modeling. \u003cem\u003eFront Nutr\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 1433544 (2024). https://doi.org/10.3389/fnut.2024.1433544\u003c/li\u003e\n\u003cli\u003eMessmer, A. S.\u003cem\u003e et al.\u003c/em\u003e Fluid Overload and Mortality in Adult Critical Care Patients-A Systematic Review and Meta-Analysis of Observational Studies. \u003cem\u003eCrit Care Med\u003c/em\u003e \u003cstrong\u003e48\u003c/strong\u003e, 1862-1870 (2020). https://doi.org/10.1097/ccm.0000000000004617\u003c/li\u003e\n\u003cli\u003eGarzotto, F.\u003cem\u003e et al.\u003c/em\u003e The Dose Response Multicentre Investigation on Fluid Assessment (DoReMIFA) in critically ill patients. \u003cem\u003eCrit Care\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 196 (2016). https://doi.org/10.1186/s13054-016-1355-9\u003c/li\u003e\n\u003cli\u003eZinter, M. S.\u003cem\u003e et al.\u003c/em\u003e Positive Cumulative Fluid Balance Is Associated With Mortality in Pediatric Acute Respiratory Distress Syndrome in the Setting of Acute Kidney Injury. \u003cem\u003ePediatr Crit Care Med\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 323-331 (2019). https://doi.org/10.1097/pcc.0000000000001845\u003c/li\u003e\n\u003cli\u003eCodes, L., de Souza, Y. G., D\u0026apos;Oliveira, R. A. C., Bastos, J. L. A. \u0026amp; Bittencourt, P. L. Cumulative positive fluid balance is a risk factor for acute kidney injury and requirement for renal replacement therapy after liver transplantation. \u003cem\u003eWorld J Transplant\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 44-51 (2018). https://doi.org/10.5500/wjt.v8.i2.44\u003c/li\u003e\n\u003cli\u003eChao, W. C.\u003cem\u003e et al.\u003c/em\u003e Association of day 4 cumulative fluid balance with mortality in critically ill patients with influenza: A multicenter retrospective cohort study in Taiwan. \u003cem\u003ePLoS One\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, e0190952 (2018). https://doi.org/10.1371/journal.pone.0190952\u003c/li\u003e\n\u003cli\u003eTang, J., Wu, C. \u0026amp; Zhong, Z. Group-Based Trajectory Modeling of Fluid Balance in Elderly Patients with Acute Ischemic Stroke: Analysis from Multicenter ICUs. \u003cem\u003eNeurol Ther\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 749-761 (2024). https://doi.org/10.1007/s40120-024-00612-x\u003c/li\u003e\n\u003cli\u003eCooper, L. B.\u003cem\u003e et al.\u003c/em\u003e The Burden of Congestion in Patients Hospitalized With Acute Decompensated Heart Failure. \u003cem\u003eAm J Cardiol\u003c/em\u003e \u003cstrong\u003e124\u003c/strong\u003e, 545-553 (2019). https://doi.org/10.1016/j.amjcard.2019.05.030\u003c/li\u003e\n\u003cli\u003eCotter, G., Metra, M., Milo-Cotter, O., Dittrich, H. C. \u0026amp; Gheorghiade, M. Fluid overload in acute heart failure--re-distribution and other mechanisms beyond fluid accumulation. \u003cem\u003eEur J Heart Fail\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 165-169 (2008). https://doi.org/10.1016/j.ejheart.2008.01.007\u003c/li\u003e\n\u003cli\u003eChang, H. C.\u003cem\u003e et al.\u003c/em\u003e Nocturnal thoracic volume overload and post-discharge outcomes in patients hospitalized for acute heart failure. \u003cem\u003eESC Heart Fail\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 2807-2817 (2020). https://doi.org/10.1002/ehf2.12881\u003c/li\u003e\n\u003cli\u003eCosentino, N.\u003cem\u003e et al.\u003c/em\u003e Fluid balance in heart failure. \u003cem\u003eEur J Prev Cardiol\u003c/em\u003e \u003cstrong\u003e30\u003c/strong\u003e, ii9-ii15 (2023). https://doi.org/10.1093/eurjpc/zwad166\u003c/li\u003e\n\u003cli\u003eFiaccadori, E.\u003cem\u003e et al.\u003c/em\u003e Ultrafiltration in heart failure. \u003cem\u003eAm Heart J\u003c/em\u003e \u003cstrong\u003e161\u003c/strong\u003e, 439-449 (2011). https://doi.org/10.1016/j.ahj.2010.09.014\u003c/li\u003e\n\u003cli\u003eHassinger, A. B., Wald, E. L. \u0026amp; Goodman, D. M. Early postoperative fluid overload precedes acute kidney injury and is associated with higher morbidity in pediatric cardiac surgery patients. \u003cem\u003ePediatr Crit Care Med\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 131-138 (2014). https://doi.org/10.1097/pcc.0000000000000043\u003c/li\u003e\n\u003cli\u003eSethi, S. K.\u003cem\u003e et al.\u003c/em\u003e Fluid Overload and Renal Angina Index at Admission Are Associated With Worse Outcomes in Critically Ill Children. \u003cem\u003eFront Pediatr\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 118 (2018). https://doi.org/10.3389/fped.2018.00118\u003c/li\u003e\n\u003cli\u003eMayne, K. J.\u003cem\u003e et al.\u003c/em\u003e Bioimpedance Indices of Fluid Overload and Cardiorenal Outcomes in Heart Failure and Chronic Kidney Disease: a Systematic Review. \u003cem\u003eJ Card Fail\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 1628-1641 (2022). https://doi.org/10.1016/j.cardfail.2022.08.005\u003c/li\u003e\n\u003cli\u003eCostanzo, M. R., Fonarow, G. C. \u0026amp; Rizzo, J. A. Ultrafiltration versus diuretics for the treatment of fluid overload in patients with heart failure: a hospital cost analysis. \u003cem\u003eJ Med Econ\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 577-583 (2019). https://doi.org/10.1080/13696998.2019.1584109\u003c/li\u003e\n\u003cli\u003eLeinum, L. R., Baandrup, A. O., G\u0026ouml;genur, I., Krogsgaard, M. \u0026amp; Azawi, N. Evaluation of a real-life experience with a digital fluid balance monitoring technology. \u003cem\u003eTechnol Health Care\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e, 3913-3924 (2024). https://doi.org/10.3233/thc-231303\u003c/li\u003e\n\u003cli\u003eDewitte, A.\u003cem\u003e et al.\u003c/em\u003e Bioelectrical impedance spectroscopy to estimate fluid balance in critically ill patients. \u003cem\u003eJ Clin Monit Comput\u003c/em\u003e \u003cstrong\u003e30\u003c/strong\u003e, 227-233 (2016). https://doi.org/10.1007/s10877-015-9706-7\u003c/li\u003e\n\u003cli\u003eShen, Z.\u003cem\u003e et al.\u003c/em\u003e A Smart-Phone App for Fluid Balance Monitoring in Patients with Heart Failure: A Usability Study. \u003cem\u003ePatient Prefer Adherence\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 1843-1853 (2022). https://doi.org/10.2147/ppa.S373393\u003c/li\u003e\n\u003cli\u003eJohnson, A. E. W.\u003cem\u003e et al.\u003c/em\u003e MIMIC-IV, a freely accessible electronic health record dataset. \u003cem\u003eSci Data\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 1 (2023). https://doi.org/10.1038/s41597-022-01899-x\u003c/li\u003e\n\u003cli\u003evon Elm, E.\u003cem\u003e et al.\u003c/em\u003e The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. \u003cem\u003eLancet\u003c/em\u003e \u003cstrong\u003e370\u003c/strong\u003e, 1453-1457 (2007). https://doi.org/10.1016/s0140-6736(07)61602-x\u003c/li\u003e\n\u003cli\u003eNguena Nguefack, H. L.\u003cem\u003e et al.\u003c/em\u003e Trajectory Modelling Techniques Useful to Epidemiological Research: A Comparative Narrative Review of Approaches. \u003cem\u003eClin Epidemiol\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 1205-1222 (2020). https://doi.org/10.2147/clep.S265287\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Table 1","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Heart Failure with Preserved Ejection Fraction, GBTM, Fluid Balance, MIMIC Database, Prognosis","lastPublishedDoi":"10.21203/rs.3.rs-6956502/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6956502/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003eFluid balance (FB) is a critical prognostic factor in heart failure, yet its longitudinal impact on critically ill HFpEF patients remains unclear. This study evaluates how FB trajectory changes influence 30-day mortality in HFpEF patients.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eUsing MIMIC-IV data, we conducted a retrospective cohort study of HFpEF patients. Group-based trajectory modeling (GBTM) identified distinct FB trajectory subgroups. Survival differences were assessed via Kaplan-Meier analysis, and Cox regression models evaluated associations between FB trajectories and mortality.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eAmong 1,089 HFpEF patients, four FB trajectories emerged: T1 (Negative balance stability), T2 (Rapid transition to negative balance), T3 (Positive balance gradual decline), and T4 (High-level decline). K-M analysis revealed significantly higher mortality in T3 and T4. Fluid-overloaded patients had worse survival than non-overloaded. Adjusted Cox models showed lower mortality in T1 (HR\u0026thinsp;=\u0026thinsp;0.67, 95%CI 0.53\u0026ndash;0.85) and T2 (HR\u0026thinsp;=\u0026thinsp;0.60, 95%CI 0.45\u0026ndash;0.80) vs. T3, with no T3\u0026ndash;T4 difference (p\u0026thinsp;=\u0026thinsp;0.35). Subgroup and sensitivity analyses supported these findings.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e\u003cp\u003eHFpEF patients with positive balance gradual decline (T3) or high-level decline FB (T4) had the poorest prognosis, while those maintaining negative balance (T1/T2) exhibited better survival. GBTM effectively stratifies risk, aiding clinical subgroup identification.\u003c/p\u003e","manuscriptTitle":"Association Between Fluid Balance Trajectories and Prognosis in Patients with Heart Failure and Preserved Ejection Fraction","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-23 08:22:08","doi":"10.21203/rs.3.rs-6956502/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2025-07-17T07:23:56+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-17T07:17:37+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-07-10T09:09:47+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-07-04T05:03:21+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-06-25T14:10:14+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"df9a2817-ce11-4edb-a915-b538192188be","owner":[],"postedDate":"July 23rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":51672097,"name":"Health sciences/Cardiology"},{"id":51672098,"name":"Health sciences/Medical research"}],"tags":[],"updatedAt":"2025-07-23T08:22:08+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-23 08:22:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6956502","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6956502","identity":"rs-6956502","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}