https://orcid.org/0000-0001-7805-2191
Figures
Abstract
Objective
In this study, we aimed to explore the relationship between serum phosphate and clinical outcomes in sepsis with E.Coli infection based on a public database in order to help physicians do individualized medical decisions.
Methods
We performed this retrospective study based on the Medical Information Mart for Intensive Care IV(MIMIC-IV) database (https://mimic.mit.edu/iv/). All the patients were hospitalized and serum phosphate was measured in 24 hours after hospitalization. E.Coli infection was confirmed by the positive blood culture of E.Coli in the database. Three models were utilized to investigate the relationship between serum phosphate and mortality in sepsis as follows: crude model (adjusted for none), model I (adjusted for age and gender) and model II (adjusted for all potential confounders). The smooth fitting curve was performed by the generalized additive model.
Results
421 adult sepsis patients with E.Coli infection were included. The 28-day mortality was 10.69%(n=45). The median age was 70 and the proportion of males was 47.51%(n=200). The smooth fitting curve showed that the relationship between serum phosphate and 28-day mortality in sepsis with E.Coli infection was positive. When serum phosphate >2.1mg/dl, the relationship was significantly positive (OR=1.55, 95%CI:1.01–2.36, P=0.043).
Citation: Luo J, Zhou S, Ding N (2025) Serum phosphate and 28-day mortality in adult sepsis with E.Coli infection: A critical care database study. PLoS ONE 20(4): e0321063. https://doi.org/10.1371/journal.pone.0321063
Editor: Colin Johnson, Oregon State University, UNITED STATES OF AMERICA
Received: January 4, 2024; Accepted: March 2, 2025; Published: April 24, 2025
Copyright: © 2025 Luo et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data that support the findings of this study are available from the Massachusetts Institute of Technology (MIT) and Beth Israel Deaconess Medical Center (BIDMC) but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of the Massachusetts Institute of Technology (MIT) and Beth Israel Deaconess Medical Center (BIDMC). The website of data source is https://physionet.org/content/mimiciv/2.0/. PhysioNet is maintained by researchers and engineers at the MIT Laboratory for Computational Physiology (Email: contact@physionet.org.)
Funding: This study was funded by Changsha Natural Science Foundation (kq2208445, awarded to ND), Changsha Central Hospital (YNKY202306, awarded to ND), National Key Clinical Specialty Scientific Research Project (Z2023047, awarded to ND), Hunan Provincial Natural Science Foundation(2025JJ80547, awarded to ND).The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
Escherichia coli (E. Coli), as a common gram-negative bacterium in the gastrointestinal system, has been an increasing cause for bloodstream infection and sepsis in recent years [1–4]. One recent study in United Kingdom revealed that E. Coli infections were found to be accounted for 75.65% of 1113 patients with laboratory-confirmed bloodstream infection [5]. In China, a 10-year retrospective research from 2010 to 2019 in a large tertiary hospital showed that in a total of 18,180 strains which were isolated from blood cultures, the most common was the E. coli (21.7%) [6]. Another study with 3819 patients undergoing cesarean section demonstrated that the incidence of sepsis was 0.84% and E. Coli was also the most common (46.15%) in sepsis patients [7]. One systemic review enrolled the literature from January 2007 to March 2018 and concluded that E. Coli was the main pathogen for 27% of documented bacteremia attacks [8].
Phosphate has been found to be related with various physiological functions [9,10]. It is crucial for cell metabolism, which puts an impact on contractility of muscle, transmission of neuron and transportation of electrocytes [11–13]. Phosphate, as an important element in human body, plays a vital role in almost all physiological activities [14]. Disturbance of serum phosphate was associated with adverse clinical outcomes in various diseases including cardiovascular diseases [15,16], malignant tumors [17] and infection [18].
However, the serum phosphate and prognosis in adult sepsis patients with E.Coli infection has not been investigated. In our study, we aimed to evaluate the association between serum phosphate and mortality in adult sepsis patients with E.Coli infection based on a public database in order to help physicians do individualized medical decisions.
2. Methods
2.1. Database, study cohort and definition
We performed this retrospective study based on the Medical Information Mart for Intensive Care IV(MIMIC-IV) database (https://mimic.mit.edu/iv/) [19]. The diagnosis of sepsis was based on the definition of Sepsis 3.0 [20]. E.Coli infection was confirmed by the positive blood culture of E.coli in the database. Sepsis patients with E.Coli infection were included in this study (n=535). Exclusion criteria were as follows:(1) missing data of serum phosphate within 24 hours after admission(n=1); (2) missing data >5% individual variables(n=113); (3) less than 18 years old (n=0).
2.2. Ethics approval and consent to participate
This study was conducted in accordance with Good Clinical Practice (Declaration of Helsinki 2002). MIMIC-IV was an anonymized public database. To apply for access to the database, we passed the Protecting Human Research Participants exam (No.32900964). The project was approved by the institutional review boards of the Massachusetts Institute of Technology (MIT) and Beth Israel Deaconess Medical Center (BIDMC) and was given a waiver of informed consent.
2.3. Data extraction and variables
Data in MIMIC-IV were extracted by PostgreSQL 9.6 software. Variables including age, gender, comorbidities (hypertension, renal disease, diabetes and coronary artery disease (CAD)), respiratory rate (RR), systolic blood pressure (SBP), diastolic blood pressure (DBP), heart rate (HR), phosphate, aspartate aminotransferase (AST), alanine aminotransferase(ALT), total bilirubin, total calcium, chloride, creatinine, glucose, international normalized ratio(INR), potassium, hemoglobin, platelet (PLT), prothrombin time (PT), partial thrombin time (PTT), urea nitrogen, sodium, white blood cells (WBC), scores of sequential organ failure assessment (SOFA) and acute physiology and chronic health evaluation (APACHEII), days of length of stay (LOS) in ICU and hospital, and 28-day mortality were extracted. Only the first value of each variable within 24 hours after admission was enrolled for analysis.
2.4. Statistical analysis
Firstly, based on tertiles of serum phosphate, sepsis patients with E.Coli infection were divided into three groups. Variables were showed as follows: 1) continuous variables as medians; 2) categories variables as percentages or frequencies. Mann–Whitney U-test or Chi-squared test were used for analyzing the variables between different three groups. Secondly, univariable and multivariable analyses were applied to explore the associations between different variables and 28-day mortality in sepsis with E.Coli infection. Thirdly, three models were utilized to investigate the relationship between serum phosphate and 28-mortality in sepsis as follows: crude model (adjusted for none), model I (adjusted for age and gender) and model II (adjusted for age; gender; HR; SBP; DBP; RR; ALT, AST, total bilirubin; total calcium; chloride; creatinine; glucose; hemoglobin; INR; PLT; PT; PTT; potassium; sodium; urea nitrogen; WBC; renal disease; CAD; diabetes; hypertension; APAHCEII; SOFA). Fourthly, a smooth fitting curve for indicating the relationship between serum phosphate and 28-day mortality in sepsis with E.Coli infection was performed by the generalized additive model. Two models including linear model and non-linear model were compared. If the P value <0.05, the non-linear model was selected. If not, the linear model was better. Finally, the subgroups analyses were done for evaluating the stability of results in the subgroups.
The statistical software packages R (http://www.R-project.org) and EmpowerStats (http://www.empowerstats.com) were applied to complete this study. Statistical significance was defined when the P-value <0.05.
3. Results
3.1. General characteristics of the study cohort
In the present study, 421 sepsis with E.Coli infection were finally included (S1 Fig). Based on tertiles of serum phosphate, sepsis patients with E.Coli infection were divided into three groups. (Q1:<2.7mg/dl(n=130), Q2: 2.7–3.5mg/dl(n=128), Q3:>3.5 mg/dl(n=163)) (Table 1). The 28-day mortality was 10.69%(n=45). The median age was 70 and the proportion of males was 47.51%(n=200). Based on tertiles of serum phosphate, variables in three groups showed significant differences including renal disease (P=0.021), HR(P=0.027), creatinine (P<0.001), chloride(P=0.007), INR(P=0.035), hemoglobin(P=0.026), PLT (P=0.005), potassium(P<0.001), PT(P=0.029), urea nitrogen(P<0.001), APCHEII score (P=0.007) and SOFA score(P=0.026). In Q3 group, the days of LOS in ICU and hospital were significantly longer (both P=0.005). The 28-day mortalities in Q1-Q3 groups were 4.62%, 7.81% and 17.79%, respectively(P<0.001).
3.2. Univariable and multivariable analysis for 28-day mortality in sepsis with E.Coli infection
In Table 2, univariable and multivariable analysis for 28-day mortality in sepsis with E.Coli infection were demonstrated. Multivariable analysis showed that variables including renal disease(P=0.023), HR(P=0.016), phosphate(P=0.003), total calcium (P=0.003) were significantly associated with 28-day mortality in sepsis with E.Coli infection.
3.3. Association between serum phosphate and 28-day mortality in sepsis with E.Coli infection
In Table 3, three models were constructed for exploring the association between serum phosphate and 28-day mortality. In crude model (adjusted for none), with each 1mg/dl increment in serum phosphate, the risk of 28-day mortality increased by 42% (OR=1.42, 95%CI: 1.17–1.73, P<0.001). In model I (adjusted for age and gender), the risk of 28-day mortality increased by 46% (OR=1.46, 95%CI: 1.19–1.78, P<0.001). In model II (adjusted for age; gender; HR; SBP; DBP; RR; ALT, AST, total bilirubin; total calcium; chloride; creatinine; glucose; hemoglobin; INR; PLT; PT; PTT; potassium; sodium; urea nitrogen; WBC; renal disease; CAD; diabetes; hypertension; APAHCEII; SOFA), with each 1mg/dl increment in serum phosphate, the risk of 28-day mortality increased by 79% (OR=1.79, 95%CI: 1.22–2.62, P=0.003).
Based on tertiles of serum phosphate, the associations between Q1-Q3 groups and 28-day mortality in three models were analyzed. Compared with Q1 group, the ORs of 28-day mortality in Q3 group were 4.47 (95%CI: 1.80–11.14, P=0.001, crude model), 4.46 (95%CI: 1.79–11.14, P=0.001, model I), and 6.26 (95%CI: 1.72–22.81, P=0.005, model II), respectively. With the serum phosphate increasing, the risk of mortality increased significantly and all P-value for trend in three models were less than 0.05 (P<0.001 in crude model, P<0.001 in model I, and P=0.004 in model II).
In Table 4, two models including the linear model and two-segment non-linear model were applied to fit the relationship between serum phosphate and mortality. In the two-segment non-linear model, the turning point of serum phosphate was 2.1mg/dl. When the serum phosphate >2.1mg/dl, the relationship between serum phosphate and 28-day mortality was significantly positive (P=0.043). While the serum phosphate ≤2.1mg/dl, the relationship was not significant (P=0.212).
In Fig 1, the smooth fitting curve was performed for indicating the relationship between serum phosphate and 28-day mortality in sepsis with E.Coli infection was positive.
Abbreviation: E.Coli= Escherichia coli.
3.4. Kaplan-Meier analysis for survival probability based on tertiles of serum phosphate
In Fig 2, Kaplan-Meier analysis for survival probability based on tertiles of serum phosphate showed that in the high level of serum phosphate group (Q3 group), the survival probability was significantly lowest(P<0.001).
3.5. Subgroup analysis
In S1 Table, subgroup analyses were investigated. The relationship between serum phosphate and mortality in different subgroups including gender, age, renal disease, CAD, diabetes, hypertension, WBC, hemoglobin, total bilirubin, urea nitrogen and creatinine were comparatively stable.
4. Discussion
In the study, the positive relationship between serum phosphate and clinical outcomes in adult sepsis patients with E.Coli infection was found based on MIMIC-IV database. As far as we know, this was the first study to investigate the association between serum phosphate and outcomes in adult sepsis patients with E.Coli infection in MIMIC-IV database.
Previous studies showed that serum phosphate was not only linked with various diseases, but also was an indicator for clinical outcomes. A cohort on the community based on older population revealed that disturbance of serum phosphate was associated with diabetes and metabolic syndrome [21]. In stroke patients, higher levels of serum phosphate were related with cerebral small vascular disorders [22]. One research from China demonstrated that in myocardial infarction patients, those with serum phosphate >4.5mg/dl had the highest risk of mortality (Hazard ratio(HR)=1.46 (95% CI: 1.35–1.83) compared to those with normal range of serum phosphate levels [23]. A large-scale retrospective study found that in emergency department, patients with lower serum phosphate levels had increased mortality risk of 30-day (OR=1.3, 95% CI:1.1–1.4) and 90-day (OR=1.2, 95%CI: 1.1–1.3) [24].
In our study, higher levels of serum phosphate were associated with increased risk of mortality in sepsis patients with E.Coli infection. First, sepsis leads to elevated levels of serum phosphate due to damage of cellular metabolism and the release of intracellular phosphate [25]. In addition, metabolic acidosis due to sepsis also causes the transcellular shifts of phosphate [26,27]. Second, increased level of serum phosphate disturbs the balance of phosphate and calcium. It makes vascular calcification and tubular cell damage, leading to organ dysfunction [28]. Moreover, current evidences clarified that higher levels of serum phosphate might enhance cell apoptosis, interfere cell migration and impair the function of endothelial cell, which may result in tissue ischemia and poor prognosis [29,30].
The main strength of this study was that the relationship between serum phosphate and prognosis in adult sepsis patients with E.Coli infection based on a large-scale public database was identified. However, several limitations should be considered. First, 28-day mortality was our mortality assessment and the mortality might be influenced by not only the sepsis but also other factors of mortality including exacerbation of underlying comorbidities. Hence, we couldn’t make a conclusion that the cause-effect between serum phosphate and mortality in sepsis with E.Coli infection was identified. Second, some potential confounders including clinical treatments (such as dialysis, renal replacement therapy) and socio-demographic parameters were not enrolled due to the lack of some data in the database. Third, serum phosphate might be affected by some factors including dietary intake and hormone levels of phosphate-regulating. In the further study, more variables would be enrolled for comprehensively analyzed.
5. Conclusion
The positive relationship between serum phosphate and 28-day mortality in adult sepsis patients with E.Coli infection was found based on MIMIC-IV database.
References
- 1. Jauneikaite E, Honeyford K, Blandy O, Mosavie M, Pearson M, Ramzan FA, et al. Bacterial genotypic and patient risk factors for adverse outcomes in Escherichia coli bloodstream infections: a prospective molecular epidemiological study. J Antimicrob Chemother. 2022;77(6):1753–61. pmid:35265995
- 2. Carvalho MJ, Sands K, Thomson K, Portal E, Mathias J, Milton R, et al. Antibiotic resistance genes in the gut microbiota of mothers and linked neonates with or without sepsis from low- and middle-income countries. Nat Microbiol. 2022;7(9):1337–47. pmid:35927336
- 3. Somayaji R, Hantrakun V, Teparrukkul P, Wongsuvan G, Rudd KE, Day NPJ, et al. Comparative clinical characteristics and outcomes of patients with community acquired bacteremia caused by Escherichia coli, Burkholderia pseudomallei and Staphylococcus aureus: a prospective observational study (Ubon-sepsis). PLoS Negl Trop Dis. 2021;15(9):e0009704. pmid:34478439
- 4. Song K, Guo C, Zeng Z, Li C, Ding N. Factors associated with in-hospital mortality in adult sepsis with Escherichia coli infection. BMC Infect Dis. 2022;22(1):197. pmid:35227247
- 5. Mitchell E, Pearce M, Roberts A, Newton J. Predictive factors of in-hospital mortality in patients with laboratory-confirmed Escherichia coli, Klebsiella species or Pseudomonas aeruginosa bloodstream infections. PLoS One. 2021;16(11):e0259305. pmid:34727130
- 6. Liu C, Xu M, Li X, Dong H, Ming L. Trends in antimicrobial resistance in bloodstream infections at a large tertiary-care hospital in China: a 10-year retrospective study (2010-2019). J Glob Antimicrob Resist. 2022;29:413–9. pmid:34800707
- 7. Shi M, Chen L, Ma X, Wu B. The risk factors and nursing countermeasures of sepsis after cesarean section: a retrospective analysis. BMC Pregn Childbirth. 2022;22(1):696. pmid:36085040
- 8. Bonten M, Johnson JR, van den Biggelaar AHJ, Georgalis L, Geurtsen J, de Palacios PI, et al. Epidemiology of Escherichia coli Bacteremia: a systematic literature review. Clin Infect Dis. 2021;72(7):1211–9. pmid:32406495
- 9. Mendonça L, Gonçalves F, Sampaio S, Castro-Chaves P, Pereira L. Association between serum phosphorus and mortality in NHANES 2003-2006: the effect of gender and renal function. J Nephrol. 2022;35(1):165–78. pmid:33580868
- 10. Lopes MB, Karaboyas A, Zhao J, Johnson DW, Kanjanabuch T, Wilkie M, et al. Association of single and serial measures of serum phosphorus with adverse outcomes in patients on peritoneal dialysis: results from the international PDOPPS. Nephrol Dial Transplant. 2023;38(1):193–202. pmid:36029279
- 11. Kim H, Kim JS, Cho S, Kim J-Y. Serum phosphorus levels are associated with carotid intima-media thickness in asymptomatic postmenopausal women. Menopause. 2020;27(9):1042–6. pmid:32852457
- 12. Jhuang Y-H, Kao T-W, Peng T-C, Chen W-L, Chang P-K, Wu L-W. Serum phosphorus as a risk factor of metabolic syndrome in the elderly in Taiwan: a large-population cohort study. Nutrients. 2019;11(10):2340. pmid:31581608
- 13. Guo C, Su Y, He L, Zeng Z, Ding N. A non-linear positive relationship between serum phosphate and clinical outcomes in sepsis. Heliyon. 2022;8(12):e12619. pmid:36619439
- 14. Jang SA, Kwon SJ, Kim CS, Park SW, Kim KM. Association between low serum phosphate level and risk of falls in hospitalized patients over 50 years of age: a retrospective observational cohort study. Clin Interv Aging. 2022;17:1343–51. pmid:36105916
- 15. Park J, Hong KY, Min JJ, Kwon E, Lee YT, Kim WS, et al. Clinically-defined preoperative serum phosphorus abnormalities and outcomes of coronary artery bypass grafting: retrospective analysis using inverse probability weighting adjustment. PLoS One. 2019;14(12):e0225720. pmid:31851672
- 16. Campos-Obando N, Bosman A, Kavousi M, Medina-Gomez C, van der Eerden BCJ, Bos D, et al. Genetic evidence for a causal role of serum phosphate in coronary artery calcification: The Rotterdam Study. J Am Heart Assoc. 2022;11(15):e023024. pmid:35904204
- 17. Amaechi P, Jenssen F, Krishnasami Z, Achanti A, Fülöp T. Excessive elevation of serum phosphate during tumor lysis syndrome: lessons from a particularly challenging case. Clinical nephrology. Case studies. 2021;9:39-44().
- 18. Al Harbi SA, Al-Dorzi HM, Al Meshari AM, Tamim H, Abdukahil SAI, Sadat M, et al. Association between phosphate disturbances and mortality among critically ill patients with sepsis or septic shock. BMC Pharmacol Toxicol. 2021;22(1):30. pmid:34049590
- 19. Giesa N, Heeren P, Klopfenstein S, Flint A, Agha-Mir-Salim L, Poncette A, et al. MIMIC-IV as a Clinical Data Schema. Stud Health Technol Inform. 2022;294:559–60. pmid:35612143
- 20. Gotts JE, Matthay MA. Sepsis: pathophysiology and clinical management. BMJ. 2016;353:i1585. pmid:27217054
- 21. Lan Q, Zhang Y, Lin F, Meng Q, Buys N, Fan H, et al. Sex-specific associations between serum phosphate concentration and cardiometabolic disease: a cohort study on the community-based older Chinese population. Diabetes Metab Syndr Obes. 2022;15:813–26. pmid:35313679
- 22. Lv W, Cui C, Wang Z, Jiang J, Deng B. A high serum phosphate and calcium-phosphate product is associated with cerebral small vascular disease in patients with stroke: a real-world study. Front Nutr. 2022;9:801667. pmid:35445062
- 23. Zhu G-H, Sun X-P, Liu Z, Fan Z-X, Wang Y-L, Tan J, et al. The relation between serum phosphorus levels and long-term mortality in Chinese patients with ST-segment elevation myocardial infarction. J Geriatr Cardiol. 2019;16(10):775–81. pmid:31700517
- 24. Barash Y, Klang E, Soffer S, Zimlichman E, Leibowitz A, Grossman E, et al. Normal-range emergency department serum phosphorus levels and all-cause mortality. Postgrad Med J. 2021;97(1144):83–8. pmid:31932356
- 25. Thongprayoon C, Radhakrishnan Y, Cheungpasitporn W, Petnak T, Qureshi F, Mao MA, et al. Association of serum phosphate derangement with mortality in patients on continuous renal replacement therapy. Can J Kidney Health Dis. 2022;9:20543581221114697. pmid:35923184
- 26. Huang N, Li H, Fan L, Zhou Q, Fu D, Guo L, et al. Serum phosphorus and albumin in patients undergoing peritoneal dialysis: interaction and association with mortality. Front Med (Lausanne). 2021;8:760394. pmid:34926505
- 27. Robert P, Alina M, Sylwia D, Jolanta M-B, Marta B, Anna G-G, et al. Higher serum phosphorus is not an independent risk factor of mortality in heart failure with reduced ejection fraction. Nutrients. 2021;13(11):4004. pmid:34836258
- 28. Park KS, Lee Y, Park G-M, Park J-H, Kim Y-G, Yang DH, et al. Association between serum phosphorus and subclinical coronary atherosclerosis in asymptomatic Korean individuals without kidney dysfunction. Am J Clin Nutr. 2020;112(1):66–73. pmid:32453399
- 29. Yang KH, Cho S, Kim SR, Lee Y-J. Serum phosphorus levels are associated with intradialytic hypotension in hemodialysis patients. Nephron. 2021;145(3):238–44. pmid:33662953
- 30. Cao W, Li Y, Wen Y, Fang S, Zhao B, Zhang X, et al. Higher serum phosphorus and calcium levels provide prognostic value in patients with acute myocardial infarction. Front Cardiovasc Med. 2022;9:929634. pmid:36158790