Phosphate Supplementation and Target Serum Levels in Sepsis: Impact on Patient Outcomes

doi:10.21203/rs.3.rs-4719577/v1

Download PDF

Article

Phosphate Supplementation and Target Serum Levels in Sepsis: Impact on Patient Outcomes

https://doi.org/10.21203/rs.3.rs-4719577/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Objectives

There is no clinical protocol for the management of serum inorganic phosphate (Pi) in critically ill patients until now. We aimed to assess the effect of phosphate supplementation on the in-hospital mortality of septic patients with hypophosphatemia and explore the optimal threshold level of serum Pi in patients with sepsis.

Methods

A retrospective analysis of septic patients selected from the Medical Information Mart for Intensive Care (MIMIC)-IV database was carried out. Adjusted OR was used to examine the impact of phosphate supplementation on the hospital mortality. Serum Pi levels were divided into 6 bands. Multivariable logistic regression model was established for each band to investigate the optimal threshold Pi level.

Results

8067 admissions were enrolled and 88967 recorded Pi measurements were collected. The incidence of hypophosphatemia was 26.4% in septic patients. After adjusting for confounding factors, it was found that with phosphate supplemented, mortality tended to increase in septic patients with hypophosphatemia, though not significantly (OR 1.28, 95%CI 0.98–1.66, p = 0.33 in Model 2). The regression models indicated that the hospital mortality benefit accrues within the serum Pi level of 1.5–3.5 mg/dl.

Conclusions

Our findings suggest that phosphate supplementation without a standard protocol did not improve the prognosis of septic patients with hypophosphatemia. Target serum Pi level appears to be the dominant factor influencing the risk of hospital mortality. The optimal level was lower than the normal range in septic population in our study. Randomized studies are needed to improve the knowledge of management for serum Pi in septic patients.

Health sciences/Medical research/Outcomes research

Health sciences/Diseases/Infectious diseases

Sepsis

inorganic phosphate

supplementation

prognosis

database

Sepsis refers to a life-threatening organ dysfunction resulting from a dysregulated host response to infection according to Sepsis-3 definition [1]. It remains a major cause of death in critically ill patients globally. Underpinned by a better understanding of the pathogenesis of sepsis, researches on the treatment of sepsis have also developed rapidly over the past decades [2]. Current clinical efforts are mainly focused on correcting underlying pathological conditions, reducing organ dysfunction and restoring aberrant parameters to physiologic levels [2–3].

Serum inorganic phosphate (Pi) disturbances, which are common in critically ill patients, can be attributed to several factors, including gastrointestinal dysfunction, acute kidney dysfunction, refeeding syndrome and the redistribution of Pi from the extracellular space into the cells [4]. In addition, hypophosphatemia is often underestimated in severely ill patients, especially those who receive renal replacement therapy [5, 6]. A recent survey illustrated that at least 15.4% of ICU patients had hypophosphatemia, and 60% of participants lacked phosphate repletion protocols [7].

Several studies found that hypophosphatemia was associated with weaning failure and prolonged duration of mechanical ventilation [8, 9]. However, the association between hypophosphatemia and the risk of mortality in ICU patients remains controversial [10, 11]. Furthermore, there is a lack of comparative studies assessing the impact of phosphate repletion on clinical outcomes [6]. While clinical chemistry laboratories provide well-defined reference ranges for serum phosphate levels, the management of hypophosphatemia lacks clarity. Specifically, it lacks a standardized definition and a clear treatment goal [11–13]. Moreover, several studies have reported that the higher level of serum phosphate was associated with worse outcomes in critically ill patients, even in normal range [14–16]. According to the latest international guidelines for the management of sepsis and septic shock from Surviving Sepsis Campaign (SSC), there was no recommendation on the management of hypophosphatemia [17].

In our previous study, we observed a significant association between higher levels of serum phosphorus (Pi), even within the normal range, and an increased risk of in-hospital mortality among patients with sepsis [18]. In the current study, our objective was to investigate the potential benefits of phosphate repletion in septic patients with hypophosphatemia. Additionally, we sought evidence for an optimal target range of serum phosphate associated with the lowest risk of mortality. This study includes a large database that consists of de-identified health-related data from ICU patients.

2.1 Database and Study Population

The Medical Information Mart for Intensive Care (MIMIC) is a freely-available database that comprises health-related data collected from patients who were admitted to the critical care units of the Beth Israel Deaconess Medical Center. More importantly, patient identifiers were removed according to the Health Insurance Portability and Accountability Act (HIPAA) Safe Harbor provision. MIMIC-IV contains data from 2008 to 2019, which was collected from Metavision bedside monitors, a clinical information system archiving and displaying data for patients cared in intensive care units. The project was approved by the Institutional Review Boards of Beth Israel Deaconess Medical Center (Boston, MA) and the Massachusetts Institute of Technology (Cambridge, MA). All methods in this study were performed in accordance with the Declaration of Helsinki. MIMIC-IV was available on the PhysioNet, an on-line forum for the dissemination and exchange of recorded biomedical signals and open-source software for analyzing them (Johnson, Alistair, et al. "MIMIC-IV" (version 1.0), PhysioNet (2021), https://doi.org/10.13026/s6n6-xd98) [19].

The database is accessible to researchers who have completed ‘protecting human subjects’ training. Data presented in this study were extracted by author Li, who completed the online training course of Data or Specimens Only Research (certification number: 38455531).

Patients meeting diagnostic criteria of sepsis, older than 18 years old were enrolled into the study. Sepsis was diagnosed according to the Sepsis-3 definition. For clinical operationalization, organ dysfunction can be represented by an increase in the Sequential Organ Failure Assessment (SOFA) score of 2 points or more [1]. Patients with parathyroid diseases or lacking serum Pi measurements during ICU stay were excluded.

2.2 Data Collection and Cleansing

The following information was extracted using pgAdmin4 PostgreSQL: age, gender, weight, comorbidities (hypertension, diabetes, coronary artery disease (CAD)), acute kidney injury (AKI) during ICU stay, highest SOFA scores and vasopressor use within ICU stay, refeeding syndrome(RFS), serum Pi, white blood cell count (WBC count), serum creatinine, hospital and ICU length of stay (HLOS and ILOS), main cause of sepsis. The vasopressor use is defined as any use of the following vasopressors, including norepinephrine, epinephrine, dopamine and dobutamine. All Pi measurements during ICU stay were extracted. Besides, the highest white blood cell count and serum creatinine within the 48 hours of sepsis were extracted. The diagnosis and staging of AKI were accomplished based on the criteria of the Kidney Disease: Improving Global Outcomes (KDIGO) AKI Guideline Work Group [20]. Refeeding syndrome was considered by the definition in Diog et al’s study, which was a decrease in serum phosphorus levels to below 0.65 mmol/L within 72 h after starting nutritional support and the change needed to be greater than a 0.16 mmol/L decrease from any concentration previously recorded during the patient’s ICU stay [21].

The phosphate repletion protocols, including drugs for phosphate supplementation, administration routes (intravenous or oral administration), time and the frequency of administration were extracted. The drugs for phosphate supplementation in MIMIC IV consist of potassium phosphate, phosphate-sodium-potassium oral suspension, potassium phosphate solution, sodium glycerophosphate and sodium phosphate.

The primary endpoint of the study was in-hospital mortality, which was defined as any death during hospitalization.

Data cleaning was performed using R version 4.1.2. The time weighted mean (TWM) value, generated by having the area under the time-value curve divided by the time between the first and last measurements, assuming linear trend between measurements, was calculated to represent serum Pi levels over the course [3, 22]. And the TWM value of serum Pi measurements within 24 hours of the first documented serum Pi measurement was calculated to represent the initial Pi level. The standard reference range for serum phosphate (Pi) in MIMIC IV is 2.7–4.5 mg/dl. Hypophosphatemia was then defined as an initial Pi level below 2.7 mg/dl. To evaluate the effect of phosphate supplementation on the prognosis of septic patients with varying degrees of initial Pi levels, hypophosphatemic patients were divided into four subgroups based on their initial Pi levels: <1.5mg/dl, 1.5-2.0mg/dl, 2.0-2.5mg/dl, > 2.5mg/dl. TWM values of serum Pi during ICU stay were calculated separately for each subgroup.

In order to explore the association between serum Pi level and the primary outcome, 6 bands of serum Pi level were further defined: low (< 1.5 mg/dl), slightly low (1.5–2.5 mg/dl), low-normal (2.5–3.5 mg/dl), high-normal (3.5–4.5 mg/dl), slightly high (4.5–5.5 mg/dl), high (> 5.5mg/dl) bands. Each band represents a range of serum Pi levels. During each single admission, patients could have serum Pi levels falling in several bands. To eliminate the bias arising from varying observation frequencies, time-weighting was implemented. Assuming a linear trend between individual measurements, the time that each band occupied was then divided into the measuring duration, i.e. the time between the first and last measurements. This method has been utilized in the exploration of the optimal glucose threshold in critically ill patients [3]. In order to evaluate the impact of Pi level over the whole ICU stay on the primary outcome, the admissions whose measuring durations were less than the 50% of the ICU stays were excluded from this part of analysis.

The data contained the missing data in weight, WBC count, creatinine level. The proportion of complete cases was 95.2% in this study. We assumed the missing values were missing at random, and replaced them with multiple imputations (MI) with mice R package [23].

2.3 Data Analysis

The data were analyzed using the R version 4.1.2. Continuous variables were presented as mean ± standard deviation (SD) and they were compared with Student’s t test. Differences between variables were assessed by Wilcoxon rank-sum test for nonparametric data and the data were presented as median with interquartile range (IQR). Binomial variables were reported as number and percentage, and they were analyzed using Chi-squared or Fisher’s exact test as appropriate.

Logistic regression was employed to determine the impact of phosphate supplementation on in-hospital mortality in septic patients with hypophosphatemia.

Using time weighting of each Pi band and possible confounding factors as the exposure variables, the multivariable logistic regression model was set up to verify the impact of the Pi band on hospital mortality. A separate model was employed for each of 6 Pi bands.

Moreover, we used two different models to adjust potential confounders. Model 1 was adjusted for age, gender, weight, vasopressor use, SOFA score, pre-existing diseases (hypertension, diabetes, CAD). Model 2 was further adjusted for Acute Complications (AKI, RFS), laboratory results (creatinine level, WBC count), main cause of sepsis (respiratory infection, blood system infection, digestive system infection, urinary system infection). Considering RFS, AKI play important roles in the phosphorus metabolism disorders, we analyzed whether the association would change in the subgroups without these comorbidities when adjusted by covariates in Model 2.

In order to investigate the optimal lower limit of serum Pi, a post hoc analysis was carried out. In the post hoc analysis, the low band was further divided into severely low (< 1.0 mg/dl) and moderately low (1.0-1.5 mg/dl) bands.

A P value of 0.05 or less was considered statistically significant for all comparisons.

3.1 Demographic Data and General Clinical Characteristics

A flow chart of the study protocol is presented in Fig. 1. In brief, 69619 admissions in the MIMIC-IV database were admitted to ICUs, of which 12529 admissions were considered to be diagnosed as sepsis according to the definition of Sepsis-3. The 471 repeated admissions were excluded since the second admissions were dependent on those of the first. In addition, 147 admissions were excluded because no serum Pi measurement was recorded during ICU stay. 2864 patients were excluded due to the history of CKD, 1594 patients were excluded for the treatment of RRT. And 188 patients were excluded for the diagnosis of parathyroid diseases. Finally, 8067 admissions were enrolled and 88967 recorded Pi measurements were collected in the study. Admissions had missing data in weight (3.80%), WBC count (0.88%), creatinine level (0.21%). The overall incidence of RFS was 19.3% (1458/7566).

The clinical characteristics of survivors and non-survivors are shown in Table 1. The overall in-hospital mortality was 17.7%. The non-survivors were with older age (P < 0.01), lower weight (P = 0.01), higher SOFA scores (P < 0.01), more vasopressor use (P < 0.01), more AKI (P < 0.01), less RFS (P < 0.01), higher WBC count (P < 0.01), higher serum creatinine level (P < 0.01), higher serum Pi level (P < 0.01), shorter hospital stay (P < 0.01) and longer ICU stay (P < 0.01), higher incidence of respiratory infection (P < 0.01), lower incidence of urinary system infection(P < 0.01) and blood system infection (P < 0.01).

Table 1

Comparisons of demographics between survivors and non–survivors
	Survivors (n = 6643)	Non-survivors (n = 1424)	P value
Age(years)	65.6 ± 16.6	70.3 ± 14.9	< 0.01
Female(n%)	3332(50.1%)	713(50.2%)	0.98
Weight(kilograms)	76.8 (64.0-93.03)	73.0 (61.0-88.7)	< 0.01
SOFA score	5(4–8)	9(7–12)	< 0.01
Vasopressor use (n (%))	2302(34.7%)	905(63.6%)	< 0.01
Hypertension (n (%))	1930(29.1%)	401(28.2%)	0.52
Diabetes (n (%))	1512(22.8%)	290(20.4%)	0.05
CAD (n (%))	757(11.4%)	179(12.6%)	0.23
AKI (n (%))	2054(30.9%)	842(59.1%)	< 0.01
RFS (n (%))	1248(20.0%)	210(15.7%)	< 0.01
WBC(*10^9/l)	9.60(8.0–17.0)	12.4(8.50–20.5)	< 0.01
Creatinine(mg/dl)	1.0(0.70–1.40)	1.30(0.8-2.00)	< 0.01
Serum Pi (mg/dl)	3.17 ± 0.78	3.88 + 1.48	< 0.01
HLOS (days)	10.6(6.18–19.3)	9.08(3.70–16.8)	< 0.01
ILOS (days)	3.95(1.93-10.0)	6.42(2.56–13.9)	< 0.01
Infection site (n (%))
Respiratory system	1675(25.2%)	506(35.5%)	< 0.01
Digestive system	529(7.96%)	91(6.39%)	0.05
Urinary system	1523(22.9%)	213(15.0%)	< 0.01
Blood system	258(3.88%)	28(1.97%)	< 0.01
Other system	2795(42.1%)	605(42.5%)	0.80
Definition of abbreviations: SOFA score: the Sequential Organ Failure Assessment; CAD: coronary artery disease; AKI: acute kidney disease; WBC: white blood cell; Pi: inorganic phosphate; HLOS: hospital length of stay; ILOS: ICU length of stay; RFS: refeeding syndrome.

3.2 Phosphate Supplementation and In-Hospital Mortality

The t-test and Chi-square test were applied in the comparison of serum Pi level and in-hospital mortality between groups with phosphate supplementation or not, according to the different initial Pi levels in patients with hypophosphatemia respectively (Table 2). A total of 2132 admissions with initial serum Pi level less than 2.7 mg/dl were defined as hypophosphatemia, of which 1170 were given phosphate supplementation, while the other 962 were not. The overall mortality of patients with hypophosphatemia was 13.0%. The Table 2 depicted that patients with phosphate supplementation have higher serum Pi levels over ICU stay, despite that the difference was only significant in the subgroup with initial Pi < 1.5mg/dl (P = 0.02). In the cohort with varying degrees of hypophosphatemia, the in-hospital mortality was higher in the group receiving phosphate repletion compared to the group without. However, it was not significant in the subgroups (Table 2). After adjusting for confounding factors in Model 1, the logistic regression model showed that phosphate supplementation was associated with the higher risk of in-hospital mortality, albeit not reaching statistical significance (OR 1.29, 95%CI 0.99–1.68, p = 0.05. Table 3). A comparable phenomenon was observed in the Model 2 (OR 1.28, 95%CI 0.98–1.66, p = 0.07. Table 3).

Table 2

Differences of serum Pi level and in-hospital mortality between groups with or without phosphate supplementation
Subgroups of Patients With Hypophosphatemia	Without Phosphate Supplementation group (n = 1170)	Phosphate Supplementation group (n = 962)	P value
	Serum Pi levels over ICU stay^*
< 1.5 mg/dl	2.10 ± 0.68(35)	2.43 ± 0.63(70)	0.02
1.5-2.0 mg/dl	2.39 ± 0.58(205)	2.48 ± 0.61(253)	0.10
2.0-2.5 mg/dl	2.61 ± 0.50(631)	2.67 ± 0.56(473)	0.05
> 2.5 mg/dl	2.80 ± 0.47(299)	2.84 ± 0.55(166)	0.41
	In-hospital mortality
< 1.5 mg/dl	3/35(0.09)	11/70(0.16)	0.48
1.5-2.0 mg/dl	20/205(0.10)	27/253(0.11)	0.87
2.0-2.5 mg/dl	70/631(0.11)	71/473(0.15)	0.07
> 2.5 mg/dl	41/299(0.14)	35/166(0.21)	0.05
Definition of abbreviations: *: The time weighted mean of serum Pi values during ICU stay was used to represent the overall level for each patient throughout their ICU stay.
Pi: inorganic phosphate

Table 3

Results of logistic regression analysis
Models	B	SE	Z	Adjusted OR	P value
Model 1(n = 2132)
Phosphate Supplementation	0.26	0.13	1.92	1.29(0.99–1.68)	0.05
Model 2(n = 2132)
Phosphate Supplementation	0.24	0.13	1.80	1.28(0.98–1.66)	0.07
Model 2 in patients without RFS(n = 1427)
Phosphate Supplementation	0.04	0.16	0.26	1.05(0.75–1.46)	0.80
Model 2 in patients without AKI(n = 1602)
Phosphate Supplementation	0.24	0.13	1.79	1.27(0.98–1.66)	0.07
Model 2 in patients without RFS and AKI(n = 1082)
Phosphate Supplementation	0.43	0.22	1.98	1.54(1.01–2.38)	0.04
Model 1 was adjusted by age, gender, weight, vasopressor use, SOFA score, hypertension, diabetes, CAD.
Model 2 was adjusted by age, gender, weight, vasopressor use, SOFA score, hypertension, diabetes, CAD, creatinine level, WBC count, AKI, refeeding syndrome, main cause for sepsis (respiratory infection, blood system infection, digestive system infection, urinary system infection).
Definition of abbreviations: RFS: refeeding syndrome; SOFA score: the Sequential Organ Failure Assessment;

Subgroup analyses, excluding participants without RFS, and AKI, yielded results consistent with the main finding (Table 3). This suggests that, in the context of septic patients receiving phosphate repletion, the decisive factor influencing in-hospital mortality may be the target serum phosphate level rather than the timing of phosphate supplementation.

3.3 Serum Pi Level and In-Hospital Mortality

A multivariable logistic regression model was utilized to investigate whether there is a specific range of serum phosphate levels associated with the lowest risk of hospital mortality. 6853 admissions with measuring durations greater than the 50% of ICU stays were involved in this model. The results are presented in Table 4.

Table 4

Relationship between serum pi level and in-hospital mortality
Models(time-weighting of each band^*)	Model 1(n = 6853)		Model 2(n = 6853)
Models(time-weighting of each band^*)	Adjusted OR (95%CI)	P value	Adjusted OR (95%CI)	P value
Serum Pi < 1.5 mg/dl	0.46(0.13–1.72)	0.25	1.02(0.27–3.8)	0.98
Serum Pi 1.5–2.5 mg/dl	0.25(0.19–0.34)	< 0.01	0.32(0.23–0.45)	< 0.01
Serum Pi 2.5–3.5 mg/dl	0.29(0.23–0.38)	< 0.01	0.32(0.25–0.43)	< 0.01
Serum Pi 3.5–4.5 mg/dl	1.78(1.36–2.33)	< 0.01	1.37(1.02–1.84)	0.03
Serum Pi 4.5–5.5 mg/dl	9.78(6.67–14.3)	< 0.01	5.63(3.69–8.61)	< 0.01
Serum Pi > 5.5 mg/dl	27.3(17.4–42.6)	< 0.01	26.6(16.0-44.1)	< 0.01
Serum Pi < 1.0 mg/dl	1.32(0.11–126.6)	0.46	7.08(0.19–270.9)	0.29
Serum Pi 1.0-1.5 mg/dl	0.32(0.08–1.35)	0.12	0.76(0.17–3.28)	0.71
Definition of abbreviations: Pi: inorganic phosphate.
*: time-weighting of each band was computed by assuming a linear trend between individual measurements, the time that each band occupied was then divided into the measuring duration, i.e. the time between the first and last measurements.
Model 1 was adjusted by age, gender, weight, vasopressor use, SOFA score, hypertension, diabetes, CAD.
Model 2 was adjusted by age, gender, weight, vasopressor use, SOFA score, hypertension, diabetes, CAD, creatinine level, WBC count, AKI, refeeding syndrome, main cause for sepsis (respiratory infection, blood system infection, digestive system infection, urinary system infection).
Definition of abbreviations: SOFA score: sequential organ failure assessment score; Pi: inorganic phosphate.
Definition of abbreviations: Pi: inorganic phosphate; RRT: renal replacement therapy; CKD: chronic kidney disease; RFS: refeeding syndrome.

After adjusting for covariates in Model 1, the logistic regression results revealed that longer exposure duration in the high-normal (3.5–4.5 mg/dl) and varying degrees of hyperphosphatemia bands was associated with predicted odds ratios (ORs) for in-hospital mortality greater than 1.0, whereas in low (< 1.5 mg/dl), slightly low (1.5–2.5 mg/dl) and low-normal (2.5–3.5 mg/dl) bands, it was associated with predicted ORs of less than 1.0, though the results were not significant for low band (Table 4). The results of the Model 2 were similar to the findings in Model 1 (Table 4).

The results of post hoc analysis showed that more time spent in severely low (< 1.0 mg/dl) band was associated with increased hospital mortality though not significantly in Model 1 and Model 2 (Table 4).

It can be inferred that maintaining serum phosphate levels above low (1.5 mg/dl) and below low-normal (2.5–3.5 mg/dl) during ICU stay in this study likely reduced the risk of in-hospital mortality.

The same model (Model 2) was built for the subgroup analyses. As shown in Fig. 2, in different subgroups, the multivariable logistic regression analyses showed that the estimates of the association between the time weighting of different serum Pi bands and in-hospital mortality were similar to the main findings (Fig. 2).

This investigation contributes to the existing understanding of the impact of phosphate supplementation on in-hospital mortality among septic patients with hypophosphatemia, and the optimal threshold serum phosphorus (Pi) level in septic cases [6, 14–15, 24].

Hypophosphatemia is frequently encountered in the intensive care units out of several risk factors, while serum Pi is not monitored routinely among these patients [6]. Meanwhile, the question whether the patients can benefit from phosphate repletion has not been well answered. To our knowledge, currently there is no comparative study evaluating whether the correction of hypophosphatemia can improve the outcome in critically ill patients [4, 6]. Published studies have predominantly focused either on evaluating the safety and efficacy of phosphate repletion or on analyzing the association between specific clinical benefits and phosphate repletion [24–26]. In a published RCT, the authors only compared the differences of short-term adverse outcomes rather than prognostic indicators between rapid (2–3 h) and gradual (4–6 h) hypophosphatemia correction groups [27].

In current study, a retrospective comparative analysis was made on 2132 septic patients with hypophosphatemia. Phosphate supplementation dramatically seems to be associated with a trend towards increased risk of in-hospital mortality while elevating the serum Pi level in septic patients with hypophosphatemia irrespective of initial Pi levels. Currently, there is a paucity of evidence regarding the optimal target and there is no standardized protocol for the phosphate repletion in critically ill patients. The indiscriminate and non-standardized phosphorus supplementation did not improve patient prognosis in this study. Varied target values in different phosphate repletion schemes may be the primary factor influencing clinical outcomes.

We also tried to explore the optimal threshold of Pi level. Our data suggested that low level (< 1.5mg/dl), high-normal Pi level (3.5–4.5 mg/dl), and hyperphosphatemia (> 4.5 mg/dl) elevate the risk of in-hospital mortality. Therefore, compared with hyperphosphatemia, the risk of hypophosphatemia seems to be moderate. This is in line with the findings of other studies [28–30]. Conclusions regarding the impact of hypophosphatemia on the prognosis of critically ill patients were controversial. A recent meta-analysis showed that hypophosphatemia in ICU was associated with increased hospital and ICU length of stay but not all-cause mortality [11]. Studies focused on the correlation between severe hypophosphatemia and prognosis of critically ill patients are limited. Though the sample size was small, one study reported that severe hypophosphatemia (< 1 mg/dl) increased the risk of death by nearly 8-fold in septic patients [28]. Several studies have reported that higher serum phosphate levels, even within the normal range, are associated with a higher risk of adverse outcomes in critically and non-critically ill patients [29, 31]. This phenomenon may be attributed to factors such as diminished muscle strength, subclinical vascular disease, vascular calcification, and cardiovascular disorders [22, 31–33]. These findings prompt consideration of the clinical applicability of laboratory normal reference ranges, particularly in critically ill patients. Unfortunately, until now, there was no evidence showing the exact threshold level of serum Pi associated with the lowest risk of death. We tried to explore an optimal target value of serum Pi in this study. Furthermore, we employed time-weighting for all Pi measurements, capturing the dynamic serum Pi levels throughout the entire ICU duration, instead of relying on a single measurement for representing the Pi level of each admission. This algorithm provides a more accurate reflection of the dynamic changes in serum Pi during the ICU stay.

Our data indicates a shift in mortality predictions from less than 1.0 in slightly low, and low-normal Pi levels to greater than 1.0 in lower or higher Pi bands. This suggests that patients spending more time in these two bands have a lower likelihood of mortality compared to those spending less time there. Similar conclusions were also obtained in the subgroup analysis. Hyperphosphatemia, even within the higher range of normal levels, increases the risk of patient mortality. Conversely, lowering phosphate levels, not below the low threshold (1.5 mg/dL), is associated with improved patient survival benefits. Based on this hypothesis, the most suitable target Pi range seems to be the slightly low band (1.5–2.5 mg/dl).

However, in clinical practice, physicians often administer phosphate supplements to manage low phosphate levels, aiming to maintain serum phosphate within the normal reference range. This study indicates that the optimal serum phosphate threshold is actually lower than the normal reference range. Elevated phosphate levels, even within the normal range, increase the risk of in-hospital mortality. This may explain why phosphate supplementation, particularly without established management consensus, has not improved patient outcomes in this study. Clinicians lack awareness of the risks associated with elevated phosphate levels (even within the normal range). For sepsis patients, the appropriate timing for phosphate supplementation, the method of administration, and the target phosphate levels urgently need to be addressed.

The present investigation had several limitations. Firstly, this paper did not evaluate the safety and efficacy of different phosphate supplementation drugs for correcting hypophosphatemia. Secondly, the occurrence of hypophosphatemia during ICU stay was limited in this study. Thirdly, the study was a retrospective analysis rather than a randomized controlled trial. Lastly, the baseline SOFA score was assumed to be zero, as we did not know if the patient suffered organ dysfunction before the onset of infection. Regarding the limitations of this study, an optimal target for phosphate supplementation cannot be precisely recommended.

More randomized trials are needed to assess the impact of phosphate supplementation on the clinical outcomes and explore the optimal target level of phosphate repletion.

Phosphate supplementation without a standard protocol did not decrease the risk of hospital mortality in septic patients with hypophosphatemia clinically. In the subgroup without AKI and RFS, phosphate supplementation even significantly increased the risk of hospital mortality. The target serum Pi level rather than the timing of phosphate supplementation influencing the outcome predominantly. The optimal level, in this study, was lower than the normal range for individuals with sepsis. Randomized studies are needed to improve the knowledge of management for serum Pi in septic patients.

Authors' Contributions

Conceptualization: XFC, QZ; methodology: ZYL, JHL; software: ZYL; formal analysis: ZYL, ZY, JHL; investigation: TWS, ZY; writing—original draft preparation: TWS; writing—review and editing: ZYL; visualization: ZYL; supervision: XFC, QZ; project administration and Funding acquisition: XFC; All authors have read and agreed to the published version of the manuscript.

Funding

This study was financed by National Natural Science Foundation of China (NNSFC), NO.82072159. The corresponding author Dr. Chen received the funding in 2021 and initiated sequential researches regarding to acute methamphetamine poisoning and organ dysfunction.

Data Availability Statement: The dataset used and/or analyzed during the current study is available at https://doi.org/10.13026/s6n6-xd98.

Conflicts of interest: The authors declare that there are no potential conflicts of interest.

Patient and Public Involvement: It was not appropriate or possible to involve patients or the public in the design, or conduct, or reporting, or dissemination plans of our research.

Singer, M., et al., The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA, 2016. 315(8): p. 801 – 10.
Huang, M., S. Cai, and J. Su, The Pathogenesis of Sepsis and Potential Therapeutic Targets. Int J Mol Sci, 2019. 20(21).
Finney, S.J., et al., Glucose control and mortality in critically ill patients. JAMA, 2003. 290(15): p. 2041–7.
Geerse, D.A., et al., Treatment of hypophosphatemia in the intensive care unit: a review. Crit Care, 2010. 14(4): p. R147.
Pistolesi, V., et al., Hypophosphatemia in critically ill patients with acute kidney injury on renal replacement therapies. J Nephrol, 2019. 32(6): p. 895–908.
Reintam Blaser, A., et al., Hypophosphatemia in critically ill adults and children - A systematic review. Clin Nutr, 2021. 40(4): p. 1744–1754.
Berger, M.M., et al., Prevalence of hypophosphatemia in the ICU - Results of an international one-day point prevalence survey. Clin Nutr, 2021. 40(5): p. 3615–3621.
Federspiel, C.K., et al., Hypophosphatemia and duration of respiratory failure and mortality in critically ill patients. Acta Anaesthesiol Scand, 2018.
Demirjian, S., et al., Hypophosphatemia during continuous hemodialysis is associated with prolonged respiratory failure in patients with acute kidney injury. Nephrol Dial Transplant, 2011. 26(11): p. 3508–14.
Broman, M., et al., Analysis of Hypo- and Hyperphosphatemia in an Intensive Care Unit Cohort. Anesth Analg, 2017. 124(6): p. 1897–1905.
Sin, J.C.K., et al., Hypophosphatemia and Outcomes in ICU: A Systematic Review and Meta-Analysis. J Intensive Care Med, 2021. 36(9): p. 1025–1035.
Du, H., et al., Derivation of Outcome-Based Pediatric Critical Values. Am J Clin Pathol, 2018. 149(4): p. 324–331.
Florenzano, P., et al., Approach to patients with hypophosphataemia. Lancet Diabetes Endocrinol, 2020. 8(2): p. 163–174.
Chen, Y., et al., Association between serum phosphate and mortality in critically ill patients: a large retrospective cohort study. BMJ Open, 2021. 11(9): p. e044473.
Wang, H., et al., Hyperphosphatemia rather than hypophosphatemia indicates a poor prognosis in patients with sepsis. Clin Biochem, 2021. 91: p. 9–15.
Al Harbi, S.A., et al., Association between phosphate disturbances and mortality among critically ill patients with sepsis or septic shock. BMC Pharmacol Toxicol, 2021. 22(1): p. 30.
Evans, L., et al., Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Intensive Care Med, 2021. 47(11): p. 1181–1247.
Li, Z., T. Shen, and Y. Han, Effect of Serum Phosphate on the Prognosis of Septic Patients: A Retrospective Study Based on MIMIC-IV Database. Front Med (Lausanne), 2022. 9: p. 728887.
Goldberger, A.L., et al., PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals. Circulation, 2000. 101(23): p. E215-20.
Kellum, J.A. and N. Lameire, Diagnosis, evaluation, and management of acute kidney injury: a KDIGO summary (Part 1). Crit Care, 2013. 17(1): p. 204.
da Silva, J.S.V., D.S. Seres, and K. Sabino, ASPEN Consensus Recommendations for Refeeding Syndrome. 2020. 35(2): p. 178–195.
Miller, C.J., et al., Impact of Serum Phosphate in Mechanically Ventilated Patients With Severe Sepsis and Septic Shock. J Intensive Care Med, 2020. 35(5): p. 485–493.
Pishgar, F., et al., MatchThem:: Matching and Weighting after Multiple Imputation. 2020.
Charron, T., et al., Intravenous phosphate in the intensive care unit: more aggressive repletion regimens for moderate and severe hypophosphatemia. Intensive Care Med, 2003. 29(8): p. 1273–8.
Vasiliu, I., et al., Phosphate supplementation in critical long term mechanically ventilated patients increases ventilator free days. Clinical Nutrition Supplements, 2012. 7(1): p. 251.
Engwerda, et al., Efficacy and safety of a phosphate replacement strategy for severe hypophosphatemia in the ICU. The Netherlands journal of medicine, 2018. 76(10): p. 437–441.
Zazzo, J.F., et al., High incidence of hypophosphatemia in surgical intensive care patients: efficacy of phosphorus therapy on myocardial function. Intensive Care Med, 1995. 21(10): p. 826–31.
Shor, R., et al., Severe hypophosphatemia in sepsis as a mortality predictor. Ann Clin Lab Sci, 2006. 36(1): p. 67–72.
Chen, Y., et al., Association between serum phosphate and mortality in critically ill patients: a large retrospective cohort study. 2021. 11(9): p. e044473.
Sin, J.C.K., et al., Phosphate abnormalities and outcomes among admissions to the intensive care unit: A retrospective multicentre cohort study. J Crit Care, 2021. 64: p. 154–159.
Ginsberg, C., et al., Serum Phosphate and Microvascular Function in a Population-Based Cohort. Clin J Am Soc Nephrol, 2019. 14(11): p. 1626–1633.
Poudel, K., et al., Association of serum calcium and phosphorus with measures of left ventricular structure and function: The ARIC study. Nutr Metab Cardiovasc Dis, 2020. 30(5): p. 758–767.
Gutiérrez, O.M., Increased serum phosphate and adverse clinical outcomes: unraveling mechanisms of disease. Curr Opin Nephrol Hypertens, 2011. 20(3): p. 224–8.

No competing interests reported.

Download PDF

Reviewers agreed at journal
09 Sep, 2024
Reviewers agreed at journal
09 Sep, 2024
Reviewers invited by journal
07 Sep, 2024
Editor assigned by journal
07 Sep, 2024
Editor invited by journal
15 Jul, 2024
Submission checks completed at journal
12 Jul, 2024
First submitted to journal
10 Jul, 2024

You are reading this latest preprint version

Phosphate Supplementation and Target Serum Levels in Sepsis: Impact on Patient Outcomes

Status:

Version 1

Abstract

Objectives

Methods

Results

Conclusions

Figures

1. Introduction

2. Materials and Methods

2.1 Database and Study Population

2.2 Data Collection and Cleansing

2.3 Data Analysis

3. Results

3.1 Demographic Data and General Clinical Characteristics

3.2 Phosphate Supplementation and In-Hospital Mortality

3.3 Serum Pi Level and In-Hospital Mortality

4. Discussion

5. Conclusion

Declarations

References

Additional Declarations

Status:

Version 1