Is Monitoring for Phosphate a Necessary Measure in Critical Ill Patients ?

Background: In recent years, some studies have shown that hyperphosphatemia and hypophosphatemia are associated with high mortality in intensive care unit (ICU) patients. Based on this, we speculated that the abnormalities of serum phosphate when patients enter the ICU had an adverse effect on the mortality of ICU patients and conducted our study. Methods: For our retrospective cohort study, we collected the data from a publicly accessible critical care database. We selected adult patients admitted to ICU with some inclusion criteria. And we extracted a large number of clinical variables. Using multivariate Cox regression and other statistical methods, we assessed the association between serum phosphate and primary endpoints. Results: Of 27131 eligible patients, the patients with hyperphosphatemia or hypophosphatemia had poorer clinical outcomes. After adjustment for potential confounders, there was no signicant association between abnormalities phosphate and 28 or 90-day ICU mortality. Nevertheless, at the medical intensive care unit, hyperphosphatemia was associated with a 37% higher risk of 28-day mortality or a 25% higher risk of 90-day mortality. Conclusion: After adjustment for potential confounders, hypophosphatemia and hyperphosphatemia at ICU admission were not the independent risk factors of 28 or 90-day mortality for general ICU patients. However, hyperphosphatemia at admission increased the risk of 28 or 90-day mortality among patients admitted to the medical intensive care unit. (about score; (VIII) comorbidities: congestive heart failure, coronary artery disease, hypertension, diabetes, chronic pulmonary disease, atrial brillation, liver disease, uid and electrolyte disorder, malignant tumor, chronic kidney disease, (IX) laboratory results (within the rst day at ICU admission): serum phosphate, serum sodium, serum potassium, serum calcium, serum creatinine, blood urea nitrogen, serum bicarbonate, serum chloride, white blood cell, hemoglobin, platelet; (X) interventions (within the rst day at ICU admission): use of mechanical ventilation, renal replacement therapy, vasopressor. For some patients whose multiple laboratory measurements within the rst day at ICU admission were available, the rst measurement was used in our study. Because all variables included less than 3% of missing observations, missing values were imputed as the mean values. admission by about 35% or 25%. It indicated that timely adjustment of hyperphosphatemia after MICU admission might be a necessary treatment. However, our study had not studied whether correcting hyperphosphatemia would improve the prognosis of patients at the MICU. So, this issue would be further explored in future research. In addition, some results need to be discussed were: (I) the inverse association between hyperphosphatemia and 28 or 90-day ICU mortality for the SICU patients; (II) the inverse association between hyperphosphatemia and 28-day ICU mortality for the TSICU patients. These results seemed to be contrary to intuition, which we thought were related to the selection of participants. In our study, for patients with multiple ICU admission records, only the rst ICU admission was analyzed. As we all know, the SICU and TSICU mainly aimed at surgical critically ill patients. They stayed in the ICU for a relatively short term. Generally, they would be transferred out of the ICU within 1–2 days after the operation. So, we thought it was not accurate to dene clinical outcomes based on the rst ICU admission. We speculated that a more reasonable approach was when critically ill patients were prepared for the operation in the SICU or TSICU, and then transferred out of the ICU for the operation. In this circumstance, it was more accurate to use hospital mortality as the clinical outcome. In the future, we need to target the patients of SICU or TSICU to verify our speculation. most advantage study the large ICU sample size, which performed for subgroup analysis and adjustment for confounding factors.


Background
In the human body, phosphorus is in the form of phosphate, most of which (about 85%) is stored in bones and teeth, and the rest (about 14%) is mainly in cells and the extracellular uid. Phosphorus only accounts for 1% (of which 70% exists in organic form, and 30% exists in inorganic form). Inorganic phosphate is the value of serum phosphate measured in the clinic [1][2][3][4][5]. Although serum phosphate is very low, it plays an important role in the body [1][2][3][6][7][8][9]. For example, (I) participating in the composition of cell membranes in the form of phospholipids to maintain the integrity and function of the cell structure; (II) regulating hemoglobin release oxygen in the form of 2,3-Diphosphoglycerate; (III) directly participating in energy storage and metabolism in the form of adenosine triphosphate. At present, the more generally accepted normal range of phosphate is 0.8 mmol/L ~ 1.5 mmol/L (2.5 mg/dL ~ 4.5 mg/dL) [3,4,6]. In addition, phosphate homeostasis is a complicated process [1-3, 5, 7, 10-12], phosphate is mainly absorbed in the intestine, and then reabsorbed or excreted in the kidney, which is mainly related to the following three factors: (I) 1,25-Dihydroxy Vitamin D [1,25(OH)2D]: up-regulating the expression of sodiumdependent phosphate co-transporter 2b (NPT2b) to stimulate the absorption of phosphate by the intestine; (II) Fibroblast Growth Factor-23 (FGF23): reducing 1,25(OH)2D and NPT2b to inhibition intestinal transport of phosphate; (III) Parathyroid Hormone (PTH): indirectly increasing the absorption of phosphate in the intestine through the action of 1,25(OH)2D. Phosphate abnormalities can cause dysfunction of multiple organ systems, including the respiratory system, cardiovascular system, immune system, urinary system, hematology, or neuromuscular [1, 3-6, 11, 13-16].
The Intensive Care Unit (ICU) is distinct from the emergency department ward. It is a department that provides medical services for patients who need life support and have a very high risk of death. It is mainly divided into ve adult ICU departments [17]: coronary care unit (CCU), cardiac surgery recovery unit (CSRU), medical intensive care unit (MICU), surgical intensive care unit (SICU), trauma surgical intensive care unit (TSICU). As we all know, patients often have electrolyte disturbances during ICU, the hypophosphatemia and hyperphosphatemia are common electrolyte disturbances. Recent studies have shown that ICU patients have a higher prevalence of hypophosphatemia and hyperphosphatemia [1,3,6,8,18,19]. It has varying degrees of impact on clinical outcomes, but due to its atypical clinical symptoms, it has not received much attention.
In the past few decades, there were some researches about the associations between abnormal serum phosphate levels and clinical outcomes. For hyperphosphatemia, most studies had proved that hyperphosphatemia was an independent risk factor for mortality or morbidity [1,7,11,14,15,[20][21][22][23][24][25][26]. For hypophosphatemia, some studies showed that it was an independent risk factor for mortality or morbidity [1,8,23,27], but some studies showed that it was not signi cantly associated with mortality or morbidity [15,20,22,28]. Thus, it was not determined whether serum phosphate abnormalities are directly related to the increase in mortality in ICU patients or merely a sign of the severity of disease for ICU patients. We reviewed the previous literature and found that the relationship between serum phosphate levels and the prognosis of the ICU population at the time of initial admission to the ICU had been poorly studied in the past, and the sample size was small. Based on the above, we speculated that the abnormalities of serum phosphate when patients enter the ICU had an adverse effect on the mortality of ICU patients, and the impact of abnormal serum phosphate levels on the mortality of patients at various ICU departments was different. To con rm this hypothesis, we conducted a single-center retrospective cohort study. Our primary study endpoints were 28-day mortality and 90day mortality after ICU admission.

Sources of Data
The data for our study were collected from a publicly accessible critical care database named Multiparameter Intelligent Monitoring in Intensive Care Database III (MIMIC-III, version 1.4) [29]. MIMIC III is a large, single-centre database containing information of 61532 ICU admissions to Beth Israel Deaconess Medical Center (a teaching hospital of Harvard Medical School in Boston, Massachusetts) between 2001 and 2012. This database is approved by the institutional review boards of the Massachusetts Institute of Technology (MIT) and Beth Israel Deaconess Medical Center (BIDMC). Because all the data is deidenti ed, no informed consent is required. In recent years, some high-quality papers using this database's data have been published [30]. In this study, data presented were extracted by author Chen, who completed the National Institutes of Health (NIH) Web-based training course and the Protecting Human Research Participants examination (No. 36328122). Data extraction was performed using structure query language (SQL) with pgAdmin4 PostgreSQL.

Selection of Participants and Strati cation Method
In our study, all patients admitted to ICU in this database were included. Exclusion criteria as follows: (I) age<18 years; (II) patients who did not complete serum phosphate measurement within the rst day after admission to the ICU or had unclear records of the rst serum phosphate measured value; (III) patients who received phosphate supplementation during the ICU stay; (IV) for patients with multiple records of ICU admission, excluding records other than the rst ICU admission. Finally, the total number of 27131 patients were included. The detail was shown in Figure 1.
( Figure 1 should be placed here).

Extraction of variables
As mentioned before, using SQL with pgAdmin4 PostgreSQL, we extracted the following data: (I) age; (II) gender; (III) ethnicities; (IV) admission types; (V) ICU types; (VI) vital signs (within the rst day at ICU admission): heart rate, mean arterial pressure, respiratory rate, temperature, oxygen saturation; (VII) severity at ICU admission: Sequential Organ Failure Assessment (SOFA) score, Simpli ed Acute Physiology Score II (SAPS II), Overall Anxiety Severity and Impairment Scale (OASIS) score; (VIII) comorbidities: congestive heart failure, coronary artery disease, hypertension, diabetes, chronic pulmonary disease, atrial brillation, liver disease, uid and electrolyte disorder, malignant tumor, chronic kidney disease, (IX) laboratory results (within the rst day at ICU admission): serum phosphate, serum sodium, serum potassium, serum calcium, serum creatinine, blood urea nitrogen, serum bicarbonate, serum chloride, white blood cell, hemoglobin, platelet; (X) interventions (within the rst day at ICU admission): use of mechanical ventilation, renal replacement therapy, vasopressor. For some patients whose multiple laboratory measurements within the rst day at ICU admission were available, the rst measurement was used in our study. Because all variables included less than 3% of missing observations, missing values were imputed as the mean values.

Outcome Variables
We extracted the following outcomes variables: (I) all-cause 28-day mortality after ICU admission; (II) all-cause 90-day mortality after ICU admission; (III) allcause mortality during ICU stay; (IV) length of ICU stay. Among the four, all-cause 28-day ICU mortality and 90-day ICU mortality were the primary endpoints, all-cause mortality during ICU stay and length of ICU stay were the secondary endpoints. Moreover, the secondary endpoints were extracted only for descriptive purposes. The data for the 28 or 90-day ICU mortality was con rmed by inspection of the death records in the database.

Statistical Analysis
Patient characteristics were described using descriptive statistics. Continuous variables were examined for normality using the Shapiro-Wilk test. According to the types and distributions of variables, normally distributed continuous variables are presented in the tables as the mean with standard deviation (SD).
Skewed variables were presented as the median with interquartile ranges (IQR). Categorical variables are presented as a percentage. Analysis of variance (or the Kruskal-Wallis test) and Chi-square (or Fisher's exact) tests were used for comparisons between groups.
Kaplan-Meier survival analysis for the cumulative rate of all-cause 28 or 90-day mortality after ICU admission was used to compare the death distributions of patients among six groups of serum phosphate at admission.
We used the restricted cubic spline functions to explore nonlinear relationships between the different levels of the serum phosphate at ICU admission as a continuous variable and our primary endpoints after ICU admission.
In order to evaluate independent associations between serum phosphate levels at ICU admission and the primary endpoints, we used univariate and multivariable Cox regression models. We used two different models to adjust potential confounders. Model 1: including age, gender, ethnicities, heart rate, mean arterial pressure, respiratory rate, temperature, oxygen saturation; Model 2: including the same as Model 1, furthermore, including admission types, ICU types, SOFA score, SAPS II, OASIS score, congestive heart failure, coronary artery disease, hypertension, diabetes, chronic pulmonary disease, atrial brillation, liver disease, uid and electrolyte disorder, malignant tumor, chronic kidney disease, serum sodium, serum potassium, serum calcium, serum creatinine, blood urea nitrogen, serum bicarbonate, serum chloride, white blood cell, hemoglobin, platelet, use of mechanical ventilation on the rst day, use of renal replacement therapy on the rst day, use of vasopressor on the rst day. We chose the G2 as the model's reference group and calculated adjusted hazard ratios (HRs) for the other groups in comparison to the reference group. In addition, we used the same way to evaluate independent associations between serum phosphate levels at ICU admission and the primary endpoints at the different ICU departments.
Software Stata version 16.0 and R version 3.6.3 were used for statistical analyses. P-values of less than 0.05 was considered to indicate statistical signi cance in all analyses.
As shown in Fig. 2, the Kaplan-Meier curves of all-cause 28 or 90-day mortality after ICU admission intuitively re ected the death distributions of the six groups' patients and showed the result as before (P < 0.0001).
Cubic spline analysis demonstrated that the relationship between serum phosphate and the probability of 28 or 90-day ICU mortality had a linear relationship.
The detail was shown in Fig. 3.

Associations Between Serum Phosphate and Clinical Outcomes
The multivariable Cox regression analyses were used to indicate the relationship between serum phosphate levels and all-cause 28 or 90-day ICU mortality.
Compared with patients of G2, patients of G1, G5 and G6 had increased 28-day ICU mortality with a non-adjusted hazard ratio (G1: HR = 1.  Table 3. Model 2 was adjusted by age, gender, ethnicities, heart rate, mean arterial pressure, respiratory rate, temperature, oxygen saturation, admission types, ICU types, Sequential Organ Failure Assessment score, Simpli ed Acute Physiology Score II, Overall Anxiety Severity and Impairment Scale score, congestive heart failure, coronary artery disease, hypertension, diabetes, chronic pulmonary disease, atrial brillation, liver disease, uid and electrolyte disorder, malignant tumor, chronic kidney disease, serum sodium, serum potassium, serum calcium, serum creatinine, blood urea nitrogen, serum bicarbonate, serum chloride, white blood cell, hemoglobin, platelet, use of mechanical ventilation on the rst day, use of renal replacement therapy on the rst day, use of vasopressor on the rst day. Statistical signi cance (P < 0.05).
Model 1 was adjusted by age, gender, ethnicities, heart rate, mean arterial pressure, respiratory rate, temperature, oxygen saturation; Model 2 was adjusted by age, gender, ethnicities, heart rate, mean arterial pressure, respiratory rate, temperature, oxygen saturation, admission types, Sequent Organ Failure Assessment score, Simpli ed Acute Physiology Score II, Overall Anxiety Severity and Impairment Scale score, congestive heart failure, coronary artery disease, hypertension, diabetes, chronic pulmonary disease, atrial brillation, liver disease, uid and electrolyte disorder, malignant tumor, chronic kidne disease, serum sodium, serum potassium, serum calcium, serum creatinine, blood urea nitrogen, serum bicarbonate, serum chloride, white blood cell, hemoglobin, platelet, use of mechanical ventilation on the rst day, use of renal replacement therapy on the rst day, use of vasopressor on the rst day. Statistical signi cance (P < 0.05).

Discussion
After reviewing the literature in recent years, we found that few studies have focused on the association between serum phosphate and ICU patients' clinical outcomes. Compared with these studies, our study seemed to be the only one study currently available with a large sample for patients in ICU, and it was the advantage of our study. In addition, the results of these studies in the past were partly consistent and partly controversial. Suzuki et al.'s study [28] showed the relationship between hypophosphatemia and 28-day ICU mortality among 2730 ICU patients. They found higher ICU mortality of patients with hypophosphatemia (12% versus 7%) and a longer ICU stay [(3.6, 2.2 to 6.8) versus (1.7, 0.9 to 3.1)], but the incidence of hypophosphatemia was not an independent risk factor for ICU mortality (OR = 0.86, 95%CI 0.66 to 1.10, adjusted). Also, they found the timing or the duration of hypophosphatemia had no Our study selected patients with serum phosphate measurements (within the rst day at ICU admission) as the study population and explored the associations between serum phosphate at ICU admission and clinical outcomes to verify the hypothesis mentioned above. We retrospectively analyzed the data from a large critical care database (MIMIC-III database). Through our study, we found that the comparison of the clinical outcomes (ICU mortality,28-day ICU mortality,90-day ICU mortality, ICU stay) among the six groups was G6 > G5 > G1 > G2, G3, G4 approximately. Unlike previous results, we found that veryhigh-normal phosphate (G5) was also associated with worse clinical outcomes in critically ill patients. It seemed to remind us whether it was necessary to rede ne the normal range of ICU admission phosphate. The results of the multivariable Cox regression analysis showed that hypophosphatemia (G1) and hyperphosphatemia (G6) at ICU admission were not the independent risk factors of 28 or 90-day mortality for critically ill patients admitted to the ICU in total.
It showed that the abnormal phosphate seemed to be just a marker of illness severity. And monitoring for phosphate might not be necessary for ICU patients.
Besides, high-normal phosphate (G4) was the independent protection factor of 28-day ICU mortality. However, after being strati ed by types of ICU, this association only appeared at the SICU. It showed that controlling the phosphate value at 3.0-4.5 mg/dL might have a positive effect on reducing short-term mortality at the SICU. Moreover, we found that hyperphosphatemia was an independent risk factor of 28 or 90-day mortality at the MICU, which would increase the risk of 28 or 90-day mortality after MICU admission by about 35% or 25%. It indicated that timely adjustment of hyperphosphatemia after MICU admission might be a necessary treatment. However, our study had not studied whether correcting hyperphosphatemia would improve the prognosis of patients at the MICU. So, this issue would be further explored in future research. In addition, some results need to be discussed were: (I) the inverse association between hyperphosphatemia and 28 or 90-day ICU mortality for the SICU patients; (II) the inverse association between hyperphosphatemia and 28-day ICU mortality for the TSICU patients. These results seemed to be contrary to intuition, which we thought were related to the selection of participants. In our study, for patients with multiple ICU admission records, only the rst ICU admission was analyzed. As we all know, the SICU and TSICU mainly aimed at surgical critically ill patients. They stayed in the ICU for a relatively short term. Generally, they would be transferred out of the ICU within 1-2 days after the operation. So, we thought it was not accurate to de ne clinical outcomes based on the rst ICU admission. We speculated that a more reasonable approach was when critically ill patients were prepared for the operation in the SICU or TSICU, and then transferred out of the ICU for the operation. In this circumstance, it was more accurate to use hospital mortality as the clinical outcome. In the future, we need to target the patients of SICU or TSICU to verify our speculation.
The most obvious advantage of this study was the large ICU sample size, which performed for subgroup analysis and adjustment for confounding factors.
And another important nding was that hyperphosphatemia was not an independent risk factor for ICU clinical outcomes, which was contrary to the results of previous studies. In addition, hypophosphatemia was also not an independent risk factor for ICU clinical outcomes. This was the same as the results of some previous studies. Our study provided clinical evidence for this conclusion. However, there were several limitations to our study. Firstly, this was a single-centre, a retrospective study based on the MIMIC-III database. Wherefore, the additional investigation was required to generalize our ndings to other institutions. Secondly, because of the missing observations of albumin measurement up to 45.7%, we did not adjust our model for serum albumin. Thirdly, the main ethnicity in our study was white. We also required further studies with a more heterogeneous ethnicity population. Fourthly, we only evaluated the baseline measurements of the serum phosphate level at ICU admission, but we ignored the evaluation of serum phosphate levels over time. Finally, although we found hyperphosphatemia was an independent risk factor of 28 or 90-day mortality at the MICU, whether correcting hyperphosphatemia would improve patients' outcomes at the MICU was unknown. Therefore, prospective studies were still needed to con rm the results of our study further.

Conclusion
After adjustment for potential confounders, hypophosphatemia and hyperphosphatemia at ICU admission were not the independent risk factors of 28 or 90day mortality for general critically ill patients but were still a sign of worse clinical outcomes. It leads us to consider whether monitoring for phosphate is not a necessary measure in general ICU patients, at least we concluded that it might not be necessary. After being strati ed by types of ICU, hyperphosphatemia at admission increased the risk of 28 or 90-day mortality among patients admitted to the MICU, which emphasizes the potential importance of early monitoring for phosphate and treatments of hyperphosphatemia for the MICU patients. Physiology Score II; OASIS, Overall Anxiety Severity and Impairment Scale; SD, standard deviation; IQR, interquartile ranges; HR, hazard ratio; CI, con dence interval; OR, odds ratio.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.
Availability of data and materials The full data set generated and/or analyzed during the present study is publicly available in the MIMIC-III Database. https://physionet.org/content/mimiciii/1.4/.

Con ict of Interest
The authors declare that they have no con icts of interest.

Authors' contributions
Chen and He designed the study. Chen extracted the data. Huang, Zhao, Xu analyzed the data. Chen and Luo drafted the manuscript. Chen and He revised the text for critical content. All authors read and approved the nal manuscript.