- Email: [email protected]
Prognostic value of serum phosphate level in adult patients resuscitated from cardiac arrest
Accepted Manuscript Title: Prognostic value of serum phosphate level in adult patients resuscitated from cardiac arrest Authors: Yong Hun Jung, Byung Kook Lee, Kyung Woon Jeung, Chun Song Youn, Dong Hun Lee, Sung Min Lee, Tag Heo, Yong Il Min PII: DOI: Reference:
S0300-9572(18)30183-7 https://doi.org/10.1016/j.resuscitation.2018.04.026 RESUS 7583
To appear in:
Resuscitation
Received date: Revised date: Accepted date:
18-1-2018 29-3-2018 17-4-2018
Please cite this article as: Jung Yong Hun, Lee Byung Kook, Jeung Kyung Woon, Youn Chun Song, Lee Dong Hun, Lee Sung Min, Heo Tag, Min Yong Il.Prognostic value of serum phosphate level in adult patients resuscitated from cardiac arrest.Resuscitation https://doi.org/10.1016/j.resuscitation.2018.04.026 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Prognostic value of serum phosphate level in adult patients resuscitated from cardiac arrest
IP T
Yong Hun Junga, Byung Kook Leea, Kyung Woon Jeunga,*, Chun Song Younb, Dong Hun Leea, Sung Min Leea, Tag Heoa, Yong Il Mina
Department of Emergency Medicine, Chonnam National University Hospital, 42 Jebong-ro, Donggu,
SC R
a
Gwangju, Republic of Korea
Department of Emergency Medicine, Seoul St. Mary’s Hospital, The Catholic University of Korea,
Corresponding author at: Department of Emergency Medicine, Chonnam National University
A
*
N
222 Banpo-daero, Seocho-gu, Seoul, Republic of Korea
U
b
M
Hospital, 42 Jebong-ro, Donggu, Gwangju, Republic of Korea.
ED
E-mail address: [email protected] (KW Jeung).
CC E
PT
Word counts: 250 words for abstract; 2, 979 words for the text
ABSTRACT
Background: Several studies have reported increased levels of phosphate after cardiac arrest. Given
A
the relationship between phosphate level and the severity of ischaemic injury reported in previous studies, higher phosphate levels may be associated with worse outcomes. We investigated the prognostic value of phosphate level after the restoration of spontaneous circulation (ROSC) in adult cardiac arrest patients. Methods: This study was a retrospective observational study including adult cardiac arrest survivors treated at the Chonnam National University Hospital between January 2014 and June 2017. From
medical records, data regarding clinical characteristics, outcome at hospital discharge, and laboratory parameters including phosphate levels after ROSC were collected. The primary outcome was poor outcome at hospital discharge, defined as Cerebral Performance Categories 3–5. Results: Of the 674 included patients, 465 had poor outcome at hospital discharge. Serum phosphate level was significantly higher in patients with poor outcome than in those with good outcome (p
IP T
<0.001). Phosphate level was correlated with time to ROSC (r = 0.350, p <0.001). Receiver operating characteristic curve analysis revealed an area under the curve of 0.805 (95% confidence interval [CI],
SC R
0.777–0.838) for phosphate level. In multivariate analysis, a higher phosphate level was
independently associated with poor outcome at hospital discharge (odds ratio, 1.432; 95% CI, 1.245– 1.626; p <0.001).
U
Conclusions: A higher phosphate level after ROSC was independently associated with poor outcome
N
at hospital discharge in adult cardiac arrest patients. However, given its modest prognostic
M
A
performance, phosphate level should be used in combination with other prognostic indicators.
Keywords: Heart Arrest; Phosphorus; Prognosis
ED
Chonnam National University Hospital Institutional Review Board Protocol No. CNUH-2018-004
Introduction
PT
Abbreviations1
Cardiac arrest is a life-threatening emergency, with an incidence of about 1 per 1,000 person year [1].
CC E
Despite advances in cardiopulmonary resuscitation (CPR) and post-resuscitation care, outcomes after cardiac arrest remain poor [2]. Accurate outcome prediction, particularly early in the treatment process, would help in planning monitoring and treatment strategies. Several modalities including
A
brain-specific biomarkers and neurophysiologic tests have been shown to accurately predict outcomes of cardiac arrest [3–5], but the results of these tests are not available during the early hours after 1
CPR, cardiopulmonary resuscitation; ROSC, restoration of spontaneous circulation; CPC, Cerebral Performance Category; TTM, targeted temperature management; GCS, Glasgow Coma Scale; SOFA, sequential organ failure assessment; CVA, cerebrovascular accident; CKD, chronic kidney disease; ROC, receiver operating characteristic; AUC, area under the receiver operating characteristic curve; CI, confidence interval; OR, odds ratio
restoration of spontaneous circulation (ROSC). During this early post-resuscitation phase, physicians frequently rely on arrest and CPR characteristics, such as initial rhythm and bystander CPR status, in order to estimate injury severity and thus predict patient outcome. Although these parameters are wellknown predictors of cardiac arrest outcomes [6–8], none of these parameters directly reflects the magnitude of ischaemic insult or reliably predicts outcomes after cardiac arrest.
IP T
Phosphate is an essential element that plays a crucial role in multiple physiological processes
including cellular signal transduction, mineral metabolism, and energy exchange. Several studies have
SC R
demonstrated that serum phosphate level increases after cardiac arrest [9–11], though this is not well recognised. Makino et al. [10] compared the acid-base and electrolyte variables of 105 cardiac arrest
patients with those of 28 mildly traumatised patients, and reported higher phosphate levels among the
U
cardiac arrest patients. Experimental and clinical studies have suggested that the extent of increase in
N
the phosphate level after ischaemia is directly proportional to the severity of ischaemic injury [11,12].
A
Given the relationship between phosphate level and the severity of ischaemic injury shown in
M
previous studies [11,12], it is plausible that higher phosphate levels may be associated with worse outcomes in cardiac arrest patients. So far, a few studies have evaluated the relationship between
ED
initial serum phosphate levels and outcomes in cardiac arrest patients [13,14]. However, these studies included only small numbers of patients or did not look into the association between phosphate level
PT
and outcome using multivariate analysis.
In the present study, we therefore sought to investigate the prognostic value of serum phosphate level
CC E
in a larger cohort of adult cardiac arrest patients. We hypothesised that serum phosphate level obtained after ROSC would be associated with outcome in cardiac arrest patients.
A
Methods
Study design, population, and setting This study was a retrospective observational cohort study of adult cardiac arrest patients treated at Chonnam National University Hospital, a university-affiliated, 1005-bed hospital located in Gwangju, Korea, from January 2008 to June 2017. The Institutional Review Board of the Chonnam National
University Hospital approved the study protocol (CNUH-2018-004). The need for informed consent was waived because of the retrospective nature of the study. We included patients aged 18 years or older who had experienced cardiac arrest and achieved sustained ROSC. Patients were excluded if: (1) the cause of cardiac arrest was trauma, (2) they had a poor pre-arrest neurologic status (Cerebral Performance Category [CPC] 3 or 4), (3) they had an
IP T
intracranial haemorrhage or acute ischaemic stroke on brain imaging obtained after ROSC because
these brain pathologies may themselves contribute to a poor neurologic outcome, (4) serum phosphate
SC R
level measured within 6 hours after ROSC was not available, or (5) information on arrest and CPR
was not available. We also excluded patients with pre-existing parathyroid disease because it could influence the phosphate level. All patients exhibiting sustained ROSC underwent post-cardiac arrest
U
care as per recent resuscitation guidelines [15,16]. According to our protocol, targeted temperature
N
management (TTM) with a target core body temperature of 32–36°C was considered for all adult non-
A
traumatic cardiac arrest survivors who were unable to obey commands. Patients were considered
M
ineligible for TTM if they had other causes of coma other than cardiac arrest, active bleeding, circulatory shock unresponsive to conventional medical therapy, known malignancy in the end stages,
ED
or a poor pre-arrest neurologic status (CPC 3 or 4). Data collection
PT
The following data were obtained from electronic medical records: age, sex, comorbidities, location of arrest (out-of-hospital versus in-hospital), presence of a witness on collapse, bystander CPR, first
CC E
monitored rhythm, dose of adrenaline (epinephrine) administered during CPR, time to ROSC (defined as the time interval from recognition of cardiac arrest to ROSC), aetiology of arrest, Glasgow Coma Scale (GCS) score on admission, sequential organ failure assessment (SOFA) score within the first 24
A
hours, TTM, and outcome assessed using the CPC scale at hospital discharge [17]. An investigator blinded to the study hypothesis determined the CPC score by medical record review. Among the first available laboratory data obtained after ROSC, we also collected levels of serum phosphate and other electrolytes (sodium, potassium, chloride, calcium, and magnesium), as well as parameters reported as outcome predictors (glucose and lactate) [8,18,19], and parameters that might affect phosphate level
(creatinine, albumin, and haemoglobin) [20]. Comorbidities included coronary artery disease, arrhythmia, heart failure, cerebrovascular accident (CVA), neurological disease other than CVA, hypertension, diabetes, pulmonary disease, malignancy, chronic kidney disease (CKD), and liver cirrhosis. At our institution, serum phosphate level was included in the routine laboratory panel, and was determined using UniCel DxC 800 Synchron Clinical Systems (Beckman Coulter, Fullerton,
IP T
USA). The primary outcome of this study was poor outcome at hospital discharge, defined as CPC 3– 5.
SC R
Statistical analysis
Continuous variables were presented as median values with interquartile ranges because all continuous variables showed non-normal distributions. Mann–Whitney U tests were conducted for
U
comparisons of continuous variables. Correlation between continuous variables was assessed by using
N
Spearman’s rank correlation coefficient. Categorical variables were presented as frequencies and
A
percentages. Comparisons of categorical variables were performed using χ2 or Fisher’s exact tests, as
M
appropriate. Receiver operating characteristic (ROC) analyses were performed to examine the performances of serum phosphate and lactate levels in predicting poor outcome. Multivariate logistic
ED
regression analysis was conducted to determine whether serum phosphate level was independently associated with poor outcome. All variables with p <0.2 in the univariate analyses were included in
PT
the multivariate regression model. Backward selection was used to obtain the final model. We subsequently constructed another multivariate prediction model by removing serum phosphate level
CC E
from the final model. ROC analyses were performed to assess the prognostic performances of the multivariate models. Comparisons of the areas under the ROC curves (AUC) were performed as recommended by DeLong et al. [21]. Data were analysed using MedCalc version 16.1 (MedCalc
A
Software, Ostend, Belgium) and SPSS for Windows version 18.0 (SPSS, Inc., Chicago, IL, USA). A two-tailed significance level of 0.05 was used for statistical significance.
Results Characteristics of included patients
A total of 923 adult cardiac arrest patients who achieved ROSC were treated during the study period. Of these patients, 249 patients were excluded as shown in Fig. 1. Thus, 674 patients were included in this study and were divided into two groups according to their outcome at hospital discharge; 209 with good outcomes and 465 with poor outcomes. The clinical characteristics stratified by outcome at hospital discharge are shown in Table 1. Patients with poor outcome were older. They
IP T
had a lower incidence of pre-existing coronary artery disease and high incidences of diabetes and liver
cirrhosis. They had a lower incidence of witnessed collapse, were less likely to receive bystander CPR,
SC R
were more likely to have a non-shockable rhythm, were more likely to have a non-cardiac aetiology, received higher doses of adrenaline during CPR, had a longer time to ROSC, had lower GCS scores, and had higher SOFA scores. TTM was performed in 332 patients (49.3%). The proportion of patients
U
who underwent TTM did not differ between the good and poor outcome groups (98 [46.9%] versus
N
234 [50.3%]; p = 0.459). The distributions of laboratory parameters after ROSC are shown in Table 1.
A
Glucose, lactate, phosphate, potassium, magnesium, and creatinine levels were higher in patients with
M
poor outcome than in those with a good outcome, while calcium, albumin, and haemoglobin levels were lower in patients with poor outcome than those with a good outcome. Sodium and chloride
ED
levels did not differ between the two groups. Prognostic performances of phosphate and lactate levels
PT
Fig. 2 shows the proportion of patients with poor outcome in relation to the phosphate level. Supplemental Fig. 1 shows the correlation between serum phosphate and lactate levels. Serum
CC E
phosphate level was significantly correlated with lactate level (r = 0.726; p <0.001), as well as time to ROSC (r = 0.350; p <0.001), and creatinine level (r = 0.398; p <0.001). The correlation between phosphate and creatinine levels was also significant in patients without pre-existing CKD (r = 0.516;
A
p <0.001). In the ROC analyses, the AUC of phosphate was 0.805 (95% confidence interval [CI], 0.777–0.838). The optimal cut-off value of 5.8 mg dl-1 corresponded to a sensitivity of 74.4% and a specificity of 76.1%. The AUC of phosphate was higher than that of lactate (AUC, 0.778; 95% CI, 0.745–0.809), but this difference did not reach statistical significance. Association between phosphate level and outcome
Table 2 shows the results of the multivariate logistic regression analysis with outcome at hospital discharge as the dependent variable. In the multivariate analysis, higher phosphate level was independently associated with poor outcome at hospital discharge (odds ratio [OR], 1.432; 95% CI, 1.245–1.626; p <0.001). Meanwhile, lactate level was no longer associated with outcome at hospital discharge. Serum lactate level showed an independent association with outcome at hospital discharge
IP T
(OR, 1.100; 95% CI, 1.034–1.169) only after excluding phosphate level from the multivariate model.
Given the impact of the first monitored rhythm [22], arrest aetiology [23] and TTM on outcome
SC R
[24,25], we stratified the overall cohort (shockable or non-shockable, cardiac or non-cardiac, TTM or non-TTM) and performed subgroup analyses. In these subgroup analyses, the association between
serum phosphate level and outcome remained significant across the subgroups (Table 3).
U
Subsequently, subgroup analysis was performed to assess patients with or without diabetes separately
N
as patients with diabetes frequently develop disturbances in their phosphate metabolism [26,27]. This
A
analysis revealed a significant association between serum phosphate level and outcome for both
M
subgroups. Finally, an additional subgroup analysis was performed to assess CKD and non-CKD patients separately. The association between phosphate level and outcome remained significant in
ED
patients without CKD, but it was no longer significant in patients with CKD. Prognostic performance of phosphate in conjunction with other predictors
PT
The AUC for the multivariate logistic regression model including phosphate level was 0.930 (95% CI, 0.908–0.948) (Fig. 3). This was significantly higher than the AUC for phosphate alone (p <0.001);
CC E
however, it did not differ from the AUC for the multivariate logistic regression model excluding phosphate level (AUC, 0.925; 95% CI, 0.902–0.944).
A
Discussion
The present study investigated the prognostic value of serum phosphate level in 674 adult patients
who were resuscitated following cardiac arrest. Phosphate levels were significantly higher among the patients with poor outcome. The ROC curve of phosphate level yielded an AUC of 0.805 for the prediction of poor outcome. In the multivariate regression model, phosphate level showed a
significant association with poor outcome after adjustment for confounders. To the best of our knowledge, this study is the largest study investigating the association between phosphate level and outcome in cardiac arrest patients, and is also the first study to report phosphate level as an independent predictor of outcome after cardiac arrest. Among the laboratory parameters obtainable in the early hours after ROSC, lactate level has
IP T
traditionally been used to estimate the magnitude of ischaemic insult and thus to predict outcome after
cardiac arrest. Although it has been suggested as a promising predictor of outcomes after cardiac
SC R
arrest, a number of studies have failed to demonstrate independent associations between the lactate
levels on admission and outcomes after cardiac arrest [13,28,29]. Several studies have reported that lactate clearance, rather than the lactate level measured on admission, can be used for the prediction
U
of outcomes after cardiac arrest [28,30]. However, this parameter cannot be obtained in the early
N
hours after ROSC because serial lactate measurements are essential to calculate the lactate clearance.
A
In our analyses using multivariate regression models, phosphate level showed a significant association
M
with outcome, while lactate level did only after excluding phosphate level from the model. The AUC of phosphate was higher than that of lactate, although the difference did not reach statistical
ED
significance. Furthermore, serum phosphate levels can be rapidly measured in many hospital laboratories. Thus, our findings suggest that serum phosphate level may be valuable as a prognostic
PT
biomarker during the early post-resuscitation phase. The potential mechanism underlying the association between serum phosphate level and outcome
CC E
shown in our study remains unclear. Experimental studies have suggested that serum phosphate level increases after ischaemic injury, the magnitude of which depends on the duration of ischaemic injury [11,12]. In a study involving serial measurements of serum phosphate levels after superior mesenteric
A
artery ligation in rabbits [12], serum phosphate levels continued to increase as the duration of ischaemia increased. Consistent with these studies, phosphate levels were significantly correlated with time to ROSC in the present study. Therefore, it can be postulated that phosphate is released from damaged cells under ischaemic conditions at rates related to the severity of ischaemic insult. Meanwhile, high phosphate level by itself may mediate harm by inducing endothelial dysfunction and
oxidative stress [31,32]. In an in vitro study using rat mitochondria [32], Oliveira et al. reported that inorganic phosphate stimulated the release of mitochondrial reactive oxygen species. In our study, individual multivariate regression models stratified by the first monitored rhythm, arrest aetiology, TTM, and diabetes revealed higher phosphate levels to be consistently associated with poor outcome across these subgroups. These findings suggest that the first monitored rhythm, arrest
IP T
aetiology, TTM, and diabetes had little influence on the association between the phosphate level and the outcome and that phosphate level can be used for prognostication regardless of these factors.
SC R
Meanwhile, the association between phosphate level and outcome was not significant in patients with
CKD. Phosphate elimination depends on renal function, and the phosphate level is commonly increased in patients with CKD [33]. Hence, the phosphate levels before cardiac arrest might have
U
already been high, and this might have influenced the association between phosphate level and
N
outcome in this subgroup. Interestingly, phosphate level was significantly correlated with creatinine
A
level even in patients without CKD. However, in the patients without CKD, phosphate level was
M
independently associated with poor outcome after adjustment for serum creatinine. Nonetheless, it is still possible that renal function confounded the association between serum phosphate level and
ED
outcome in patients without CKD since serum creatinine level does not reflect renal function precisely [34].
PT
Although phosphate level was identified as an independent predictor of poor outcome in the present study, it demonstrated only a modest predictive performance (AUC of 0.805), precluding the use of
CC E
phosphate level alone for prognostication. The multivariate logistic regression model including phosphate and other significant predictors showed superior prognostic performance compared to phosphate alone. However, no significant difference in the prognostic performance was found
A
between the multivariate model including phosphate and the multivariate model after removing phosphate. These findings suggest that phosphate level should be used in combination with other prognostic indicators in clinical prognostication. Interestingly, albumin and potassium levels also showed a significant association with outcome in our multivariate regression models. With regard to albumin level, several studies have reported
consistent findings [35,36]. In a prospective observational study including 1,269 adult out-of-hospital cardiac arrest patients, Matsuyama et al. [35] reported that serum albumin level on hospital arrival was independently associated with outcome at 1 month. Geddes et al. [37] evaluated the biochemical changes during ventricular fibrillation cardiac arrest in 14 anesthetised pigs, and reported that potassium levels increased with increasing duration of arrest, suggesting potassium level as a potential
IP T
biomarker for the assessment of the severity of ischaemic injury. However, to the best of our knowledge, no clinical study has specifically assessed the relationship between potassium level on
SC R
admission and outcomes after cardiac arrest. Yanagawa et al. [14] investigated the relationship
between laboratory parameters on admission and outcome in 135 adult out-of-hospital cardiac arrest patients. In their study, potassium levels were higher in patients with poor outcome than in those with
U
good outcome, but the association was not significant after adjusting for confounders. Given the small
N
sample size of the study conducted by Yanagawa et al. [14], further studies are required to confirm the
A
association between potassium level on admission and outcomes after cardiac arrest.
M
The present study has several limitations. First, it was a retrospective study of patients treated at a single centre; thus, further studies are required to enhance the generalisability of the findings. Second,
ED
the decision to perform laboratory studies was left to the discretion of treating physician. Serum phosphate level was not available in 88 (9.5%) among the 923 patients reviewed for inclusion, which
PT
might have introduced selection bias. Third, treating physicians were not blinded to the phosphate levels. However, they were unaware of the association between serum phosphate level and outcome,
CC E
and the optimal cut-off value. Therefore, the treatments were unlikely to have been influenced by the phosphate levels. Fourth, we did not include the data on haemodynamic status or vasopressor therapy during the early post-resuscitation phase. Hypotension, as well as vasopressor therapy to reverse it,
A
may have affected the phosphate level by increasing ischaemia in various organs. Thus, these factors could be confounders in the analyses. Fifth, we could not obtain data regarding parathyroid hormone and vitamin D levels, as well as information on the use of phosphate-lowering medication, all of which are potential sources of bias. Sixth, despite our efforts, unmeasured confounders might have existed.
Conclusions A higher serum phosphate level after successful resuscitation was independently associated with poor outcome at hospital discharge in adult cardiac arrest patients. However, given the modest
prognostic indicators in clinical prognostication of cardiac arrest patients.
A
CC E
PT
ED
M
A
N
U
SC R
Conflicts of interest: All authors have no potential conflicts of interest to disclose.
IP T
prognostic performance of serum phosphate level, it should be used in combination with other
References [1]. Smith TW, Cain ME. Sudden cardiac death: epidemiologic and financial worldwide perspective. J Interv Card Electrophysiol 2006;17:199–203.
[2]. Sasson C, Rogers MA, Dahl J, Kellermann AL. Predictors of survival from out-of-hospital cardiac
IP T
arrest: a systematic review and meta-analysis. Circ Cardiovasc Qual Outcomes 2010;3:63–81.
SC R
[3]. Zandbergen EG, Hijdra A, Koelman JH, Hart AA, Vos PE, Verbeek MM, et al. Prediction of poor outcome within the first 3 days of postanoxic coma. Neurology 2006;66:62–8.
U
[4]. Rosén H, Sunnerhagen KS, Herlitz J, Blomstrand C, Rosengren L. Serum levels of the brain-
N
derived proteins S-100 and NSE predict long-term outcome after cardiac arrest. Resuscitation
M
A
2001;49:183–91.
[5]. Oh SH, Park KN, Shon YM, Kim YM, Kim HJ, Youn CS, et al. Continuous Amplitude-Integrated
ED
Electroencephalographic Monitoring Is a Useful Prognostic Tool for Hypothermia-Treated Cardiac
PT
Arrest Patients. Circulation 2015;132:1094–103.
[6]. Holmberg M, Holmberg S, Herlitz J. Effect of bystander cardiopulmonary resuscitation in out-of-
CC E
hospital cardiac arrest patients in Sweden. Resuscitation 2000;47:59–70.
[7]. Fairbanks RJ, Shah MN, Lerner EB, Ilangovan K, Pennington EC, Schneider SM. Epidemiology
A
and outcomes of out-of-hospital cardiac arrest in Rochester, New York. Resuscitation 2007;72:415–24.
[8]. Lee DH, Cho IS, Lee SH, Min YI, Min JH, Kim SH, et al. Correlation between initial serum levels of lactate after return of spontaneous circulation and survival and neurological outcomes in patients who undergo therapeutic hypothermia after cardiac arrest. Resuscitation 2015;88:143–9.
[9]. Bleske BE, Song J, Chow MS, Kluger J, White CM. Hematologic and chemical changes observed during and after cardiac arrest in a canine model--a pilot study. Pharmacotherapy 2001;21:1187–91.
[10]. Makino J, Uchino S, Morimatsu H, Bellomo R. A quantitative analysis of the acidosis of cardiac
IP T
arrest: a prospective observational study. Crit Care 2005;9:R357–62.
SC R
[11]. Smolenski RT, Swierczyński J, Narkiewicz M, Zydowo MM. Purines, lactate and phosphate release from child and adult heart during cardioplegic arrest. Clin Chim Acta 1990;192:155–63.
U
[12]. Dundar ZD, Cander B, Gul M, Karabulut KU, Kocak S, Girisgin S, et al. Serum intestinal fatty
N
acid binding protein and phosphate levels in the diagnosis of acute intestinal ischemia: an
M
A
experimental study in rabbits. J Emerg Med 2012;42:741–7.
[13]. Cho YM, Lim YS, Yang HJ, Park WB, Cho JS, Kim JJ, et al. Blood ammonia is a predictive
ED
biomarker of neurologic outcome in cardiac arrest patients treated with therapeutic hypothermia. Am J
PT
Emerg Med 2012;30:1395–401.
[14]. Yanagawa Y, Sakamoto T, Sato H. Relationship between laboratory findings and the outcome of
CC E
cardiopulmonary arrest. Am J Emerg Med 2009;27:308–12.
[15]. Morrison LJ, Deakin CD, Morley PT, Callaway CW, Kerber RE, Kronick SL, et al. Part 8:
A
Advanced life support: 2010 International Consensus on Cardiopulmonary Resuscitation and Emergency
Cardiovascular
Care
Science
With
Treatment
Recommendations.
Circulation
2010;122:S345–421.
[16]. Callaway CW, Donnino MW, Fink EL, Geocadin RG, Golan E, Kern KB, et al. Part 8: Post-
Cardiac Arrest Care: 2015 American Heart Association Guidelines Update for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2015;132:S465–82.
[17]. Booth CM, Boone RH, Tomlinson G, Detsky AS. Is this patient dead, vegetative, or severely neurologically impaired? Assessing outcome for comatose survivors of cardiac arrest. JAMA
IP T
2004;291:870–9.
SC R
[18]. Shinozaki K, Oda S, Sadahiro T, Nakamura M, Hirayama Y, Watanabe E, et al. Blood ammonia
and lactate levels on hospital arrival as a predictive biomarker in patients with out-of-hospital cardiac
U
arrest. Resuscitation 2011;82:404–9.
N
[19]. Kim SH, Choi SP, Park KN, Lee SJ, Lee KW, Jeong TO, et al. Association of blood glucose at
A
admission with outcomes in patients treated with therapeutic hypothermia after cardiac arrest. Am J
M
Emerg Med 2014;32:900–4.
ED
[20]. Tonelli M, Sacks F, Pfeffer M, Gao Z, Curhan G; Cholesterol And Recurrent Events Trial Investigators. Relation between serum phosphate level and cardiovascular event rate in people with
PT
coronary disease. Circulation 2005;112:2627–33.
CC E
[21]. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics
A
1988;44:837–45.
[22]. Terman SW, Hume B, Meurer WJ, Silbergleit R. Impact of presenting rhythm on short- and long-term neurologic outcome in comatose survivors of cardiac arrest treated with therapeutic hypothermia. Crit Care Med 2014;42:2225–34.
[23]. Wallmuller C, Meron G, Kurkciyan I, Schober A, Stratil P, Sterz F. Causes of in-hospital cardiac arrest and influence on outcome. Resuscitation 2012;83:1206–11.
[24]. Hypothermia after Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve the
IP T
neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549–56.
[25]. Bernard SA, Gray TW, Buist MD, Jones BM, Silvester W, Gutteridge G, et al. Treatment of
SC R
comatose survivors of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med 2002;346:557–63.
U
[26]. Ditzel J, Brøchner-Mortensen J, Rødbro P. Elevated glomerular filtration rate in early diabetes
N
may be explained by increased sodium reabsorption secondary to an impairment in renal tubular
M
A
handling of phosphate. Horm Metab Res Suppl 1981;11:87–9.
[27]. Ditzel J, Brøchner-Mortensen J, Kawahara R. Dysfunction of tubular phosphate reabsorption
ED
related to glomerular filtration and blood glucose control in diabetic children. Diabetologia
PT
1982;23:406–10.
[28]. Lee TR, Kang MJ, Cha WC, Shin TG, Sim MS, Jo IJ, et al. Better lactate clearance associated
CC E
with good neurologic outcome in survivors who treated with therapeutic hypothermia after out-ofhospital cardiac arrest. Crit Care 2013;17:R260.
A
[29]. Starodub R, Abella BS, Grossestreuer AV, Shofer FS, Perman SM, Leary M, et al. Association of serum lactate and survival outcomes in patients undergoing therapeutic hypothermia after cardiac arrest. Resuscitation 2013;84:1078–82.
[30]. Donnino MW, Miller J, Goyal N, Loomba M, Sankey SS, Dolcourt B, et al. Effective lactate
clearance is associated with improved outcome in post-cardiac arrest patients. Resuscitation 2007;75:229–34.
[31]. Shuto E, Taketani Y, Tanaka R, Harada N, Isshiki M, Sato M, et al. Dietary phosphorus acutely
IP T
impairs endothelial function. J Am Soc Nephrol 2009;20:1504–12.
[32]. Oliveira GA, Kowaltowski AJ. Phosphate increases mitochondrial reactive oxygen species
SC R
release. Free Radic Res 2004;38:1113–8.
[33]. Block GA, Hulbert-Shearon TE, Levin NW, Port FK. Association of serum phosphorus and
U
calcium x phosphate product with mortality risk in chronic hemodialysis patients: a national study.
A
N
Am J Kidney Dis 1998;31:607–17.
M
[34]. Shemesh O, Golbetz H, Kriss JP, Myers BD. Limitations of creatinine as a filtration marker in
ED
glomerulopathic patients. Kidney Int 1985;28:830–8.
[35]. Matsuyama T, Iwami T, Yamada T, Hayakawa K, Yoshiya K, Irisawa T, et al. Effect of Serum
PT
Albumin Concentration on Neurological Outcome After Out-of-Hospital Cardiac Arrest (from the CRITICAL [Comprehensive Registry of Intensive Cares for OHCA Survival] Study in Osaka, Japan).
CC E
Am J Cardiol 2018;121:156–61.
[36]. Park I, Lee JH, Kim K, Jo YH, Kim J, Kim T, et al. Prognostic Significance of Initial Serum
A
Albumin on Mortality in Out-of-hospital Cardiac Arrest. J Korean Soc Emerg Med 2013;24:500–7.
[37]. Geddes LA, Roeder RA, Rundell AE, Otlewski MP, Kemeny AE, Lottes AE. The natural biochemical changes during ventricular fibrillation with cardiopulmonary resuscitation and the onset of postdefibrillation pulseless electrical activity. Am J Emerg Med 2006;24:577–81.
Legends to figures Fig. 1. Flow diagram of patient inclusion in the present study.
Fig. 2. Proportion of patients with poor outcome at hospital discharge in relation to phosphate level. The overall cohort was divided into four categories based on the distribution of phosphate levels,
IP T
using quartiles as cut-off values between categories.
SC R
Fig. 3. Receiver operating characteristic curves of phosphate alone, multivariate logistic regression
A
CC E
PT
ED
M
A
N
U
model including phosphate, and multivariate logistic regression model after excluding phosphate.
A ED
PT
CC E
IP T
SC R
U
N
A
M
Figures:
A ED
PT
CC E
IP T
SC R
U
N
A
M
Tables:
Table 1 Clinical and laboratory characteristics stratified by outcome at hospital discharge. Good outcome
Poor outcome
p-
(n = 674)
(n = 209)
(n = 465)
value
64.0 (52.0–
57.0 (49.0–
Age, years
IP T
Total
<0.0
67.0 (55.0–76.0) 69.0)
452 (67.1)
150 (71.8)
302 (64.9)
0.098
Coronary artery disease
83 (12.3)
35 (16.7)
48 (10.3)
0.026
Arrhythmia
28 (4.2)
10 (4.8)
18 (3.9)
0.733
Heart failure
34 (5.0)
14 (6.7)
20 (4.3)
0.261
CVA
38 (5.6)
12 (5.7)
26 (5.6)
1.000
Male sex
01
SC R
74.0)
M
A
N
U
Comorbidities
265 (39.3)
74 (35.4)
191 (41.1)
0.191
Diabetes
178 (26.4)
40 (19.1)
138 (29.7)
0.006
48 (7.1)
11 (5.3)
37 (8.0)
0.273
23 (3.4)
3 (1.4)
20 (4.3)
0.096
Malignancy
56 (8.3)
13 (6.2)
43 (9.2)
0.244
Chronic kidney disease
72 (10.7)
16 (7.7)
56 (12.0)
0.116
Liver cirrhosis
18 (2.7)
1 (0.5)
17 (3.7)
0.035
Witnessed collapse
456 (67.7)
174 (83.3)
282 (60.6)
PT
Pulmonary disease
ED
Hypertension
Neurological disease other
A
CC E
than CVA
<0.0 01
Bystander CPR
443 (65.7)
163 (78.0)
280 (60.2)
<0.0
01 <0.0 First monitored rhythm 01 193 (28.6)
121 (57.9)
72 (15.5)
Non-shockable
481 (71.4)
88 (42.1)
393 (84.5)
Location of arrest
IP T
Shockable
0.067
488 (72.4)
141 (67.5)
In-hospital
186 (27.6)
68 (32.5)
118 (25.4)
3.0 (1.0–5.0)
1.0 (0.0–2.0)
3.0 (2.0–6.0)
Adrenaline administered during
U
CPR, mg 25.0 (13.0–
15.0 (9.0–23.0)
N
Time to ROSC, minute
01 <0.0
PT
Non-cardiac
CC E
Glasgow Coma Scale
01
342 (50.7)
170 (81.3)
172 (37.0)
332 (49.3)
39 (18.7)
293 (63.0)
3.0 (3.0–6.0)
7.0 (3.0–14.0)
3.0 (3.0–4.0)
ED
Cardiac
01 <0.0
M
Aetiology
<0.0
31.0 (18.0–45.0)
A
40.0)
347 (74.6)
SC R
Out-of-hospital
<0.0 01 11.0 (8.0–
SOFA score
12.0 (10.0–
<0.0
14.0)a
01
7.0 (5.0–10.0) 13.0)a 231.0 (166.0–
206.0 (152.0–
247.0 (176.0–
<0.0
311.5)b
272.0)c
325.0)d
01
8.9 (5.6–13.3)e
5.3 (3.2–8.2)f
10.5 (7.6–14.6)g
A
Glucose, mg dl-1
<0.0 Lactate, mmol l-1
01 Phosphate, mg dl-1
6.6 (4.6–8.5)
4.5 (3.5–5.8)
7.7 (5.8–8.8)
<0.0
01 141.0 (137.0–
141.0 (138.0–
141.0 (137.0–
144.0)
143.0)
144.0)
4.2 (3.7–5.2)
3.8 (3.5–4.2)
4.6 (3.8–5.6)
Sodium, mEq l-1
0.919
<0.0 Potassium, mEq l-1
107.0 (103.0–
107.0 (104.0–
Chloride, mEq l-1
IP T
01 106.0 (102.0–
0.283
8.0 (7.4–8.5)
Creatinine, mg dl-1
2.2 (2.0–2.5)
1.1 (0.9–1.3)
M
Albumin, g dl-1
ED
3.0 (2.4–3.5)
12.3 (10.0–
Hemoglobin, g dl-1
01 <0.0
2.5 (2.2–2.9) 01 <0.0 1.4 (1.1–2.0) 01 <0.0
3.5 (3.1–3.8)
2.8 (2.3–3.3) 01
13.8 (11.6–
<0.0 11.7 (9.5–13.4)
15.1)
PT
14.2)
<0.0
7.6 (6.9–8.1)
A
1.3 (1.0–1.7)
SC R
7.7 (7.0–8.3)
2.4 (2.1–2.8)
110.0)
U
Magnesium, mg dl-1
110.0)
N
Calcium, mg dl-1
110.0)
01
Data are presented as n (%) or as medians with interquartile ranges. CVA, cerebrovascular accident;
CC E
CPR, cardiopulmonary resuscitation; ROSC, restoration of spontaneous circulation; SOFA, sequential organ failure assessment.
A
Missing data; a n = 2; b n = 7; c n = 4; d n = 3; e n = 14; f n = 9; g n = 5.
Table 2 Multivariate logistic regression analyses of poor outcome at hospital discharge. OR (95% CI)
p-value
Age
1.032 (1.014 – 1.050)
0.001
Bystander CPR
0.617 (0.345 – 1.103)
0.103
Shockable rhythm
0.430 (0.233 – 0.796)
0.007
Time to ROSC
1.035 (1.016 – 1.055)
<0.001
Cardiac aetiology
0.316 (0.170 – 0.589)
<0.001
Glasgow Coma Scale
0.827 (0.760 – 0.901)
<0.001
SOFA score
1.170 (1.064 – 1.286)
0.001
Phosphate
1.423 (1.245 – 1.626)
<0.001
Albumin
0.608 (0.399 – 0.928)
SC R
U 0.021
N
A
Model 2
1.023 (1.006 – 1.041)
0.008
Shockable rhythm
0.450 (0.248 – 0.815)
0.008
Time to ROSC
1.037 (1.018 – 1.056)
<0.001
0.341 (0.187 – 0.621)
<0.001
0.812 (0.746 – 0.885)
<0.001
1.180 (1.071 – 1.300)
0.001
Lactate
1.100 (1.034 – 1.169)
0.002
Potassium
1.375 (1.068 – 1.770)
0.013
Albumin
0.576 (0.375 – 0.883)
0.011
ED
A
CC E
SOFA score
PT
Glasgow Coma Scale
M
Age
Cardiac aetiology
IP T
Model 1
The multivariate model 1 included all variables with p <0.2 in univariate analyses, whereas model 2 included the variables included in model 1 except phosphate level. From the result of the HosmerLemeshow test, the p-value of Model 1 was 0.190 and that of Model 2 was 0.347. OR, odds ratio; CI, confidence interval; CPR, cardiopulmonary resuscitation; ROSC, restoration of spontaneous circulation; SOFA, sequential organ failure assessment.
Table 3 Adjusted odds ratio and 95% confidence interval for poor outcome associated with each 1 mg dl-1 increase in serum phosphate level by subgroup. Subgroup
OR (95% CI)
p-value
Shockable
1.284 (1.014 – 1.626)
0.038
Non-shockable
1.570 (1.326 – 1.858)
<0.001
Cardiac
1.459 (1.210 – 1.758)
<0.001
Non-cardiac
1.501 (1.171 – 1.925)
0.001
TTM
1.493 (1.226 – 1.817)
<0.001
Non-TTM
1.697 (1.287 – 2.237)
<0.001
IP T
First monitored rhythm
N
1.746 (1.297 – 2.349)
<0.001
Non-diabetes
1.438 (1.229 – 1.682)
<0.001
Pre-existing CKD
Non-CKD
1.939 (0.719 – 5.228)
0.191
1.448 (1.249 – 1.680)
<0.001
PT
CKD
ED
Diabetes
M
A
Diabetes
U
TTM
SC R
Aetiology of arrest
CC E
OR, odds ratio; CI, confidence interval; TTM, targeted temperature management; CKD, chronic
A
kidney disease.
S0300-9572(18)30183-7 https://doi.org/10.1016/j.resuscitation.2018.04.026 RESUS 7583
To appear in:
Resuscitation
Received date: Revised date: Accepted date:
18-1-2018 29-3-2018 17-4-2018
Please cite this article as: Jung Yong Hun, Lee Byung Kook, Jeung Kyung Woon, Youn Chun Song, Lee Dong Hun, Lee Sung Min, Heo Tag, Min Yong Il.Prognostic value of serum phosphate level in adult patients resuscitated from cardiac arrest.Resuscitation https://doi.org/10.1016/j.resuscitation.2018.04.026 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Prognostic value of serum phosphate level in adult patients resuscitated from cardiac arrest
IP T
Yong Hun Junga, Byung Kook Leea, Kyung Woon Jeunga,*, Chun Song Younb, Dong Hun Leea, Sung Min Leea, Tag Heoa, Yong Il Mina
Department of Emergency Medicine, Chonnam National University Hospital, 42 Jebong-ro, Donggu,
SC R
a
Gwangju, Republic of Korea
Department of Emergency Medicine, Seoul St. Mary’s Hospital, The Catholic University of Korea,
Corresponding author at: Department of Emergency Medicine, Chonnam National University
A
*
N
222 Banpo-daero, Seocho-gu, Seoul, Republic of Korea
U
b
M
Hospital, 42 Jebong-ro, Donggu, Gwangju, Republic of Korea.
ED
E-mail address: [email protected] (KW Jeung).
CC E
PT
Word counts: 250 words for abstract; 2, 979 words for the text
ABSTRACT
Background: Several studies have reported increased levels of phosphate after cardiac arrest. Given
A
the relationship between phosphate level and the severity of ischaemic injury reported in previous studies, higher phosphate levels may be associated with worse outcomes. We investigated the prognostic value of phosphate level after the restoration of spontaneous circulation (ROSC) in adult cardiac arrest patients. Methods: This study was a retrospective observational study including adult cardiac arrest survivors treated at the Chonnam National University Hospital between January 2014 and June 2017. From
medical records, data regarding clinical characteristics, outcome at hospital discharge, and laboratory parameters including phosphate levels after ROSC were collected. The primary outcome was poor outcome at hospital discharge, defined as Cerebral Performance Categories 3–5. Results: Of the 674 included patients, 465 had poor outcome at hospital discharge. Serum phosphate level was significantly higher in patients with poor outcome than in those with good outcome (p
IP T
<0.001). Phosphate level was correlated with time to ROSC (r = 0.350, p <0.001). Receiver operating characteristic curve analysis revealed an area under the curve of 0.805 (95% confidence interval [CI],
SC R
0.777–0.838) for phosphate level. In multivariate analysis, a higher phosphate level was
independently associated with poor outcome at hospital discharge (odds ratio, 1.432; 95% CI, 1.245– 1.626; p <0.001).
U
Conclusions: A higher phosphate level after ROSC was independently associated with poor outcome
N
at hospital discharge in adult cardiac arrest patients. However, given its modest prognostic
M
A
performance, phosphate level should be used in combination with other prognostic indicators.
Keywords: Heart Arrest; Phosphorus; Prognosis
ED
Chonnam National University Hospital Institutional Review Board Protocol No. CNUH-2018-004
Introduction
PT
Abbreviations1
Cardiac arrest is a life-threatening emergency, with an incidence of about 1 per 1,000 person year [1].
CC E
Despite advances in cardiopulmonary resuscitation (CPR) and post-resuscitation care, outcomes after cardiac arrest remain poor [2]. Accurate outcome prediction, particularly early in the treatment process, would help in planning monitoring and treatment strategies. Several modalities including
A
brain-specific biomarkers and neurophysiologic tests have been shown to accurately predict outcomes of cardiac arrest [3–5], but the results of these tests are not available during the early hours after 1
CPR, cardiopulmonary resuscitation; ROSC, restoration of spontaneous circulation; CPC, Cerebral Performance Category; TTM, targeted temperature management; GCS, Glasgow Coma Scale; SOFA, sequential organ failure assessment; CVA, cerebrovascular accident; CKD, chronic kidney disease; ROC, receiver operating characteristic; AUC, area under the receiver operating characteristic curve; CI, confidence interval; OR, odds ratio
restoration of spontaneous circulation (ROSC). During this early post-resuscitation phase, physicians frequently rely on arrest and CPR characteristics, such as initial rhythm and bystander CPR status, in order to estimate injury severity and thus predict patient outcome. Although these parameters are wellknown predictors of cardiac arrest outcomes [6–8], none of these parameters directly reflects the magnitude of ischaemic insult or reliably predicts outcomes after cardiac arrest.
IP T
Phosphate is an essential element that plays a crucial role in multiple physiological processes
including cellular signal transduction, mineral metabolism, and energy exchange. Several studies have
SC R
demonstrated that serum phosphate level increases after cardiac arrest [9–11], though this is not well recognised. Makino et al. [10] compared the acid-base and electrolyte variables of 105 cardiac arrest
patients with those of 28 mildly traumatised patients, and reported higher phosphate levels among the
U
cardiac arrest patients. Experimental and clinical studies have suggested that the extent of increase in
N
the phosphate level after ischaemia is directly proportional to the severity of ischaemic injury [11,12].
A
Given the relationship between phosphate level and the severity of ischaemic injury shown in
M
previous studies [11,12], it is plausible that higher phosphate levels may be associated with worse outcomes in cardiac arrest patients. So far, a few studies have evaluated the relationship between
ED
initial serum phosphate levels and outcomes in cardiac arrest patients [13,14]. However, these studies included only small numbers of patients or did not look into the association between phosphate level
PT
and outcome using multivariate analysis.
In the present study, we therefore sought to investigate the prognostic value of serum phosphate level
CC E
in a larger cohort of adult cardiac arrest patients. We hypothesised that serum phosphate level obtained after ROSC would be associated with outcome in cardiac arrest patients.
A
Methods
Study design, population, and setting This study was a retrospective observational cohort study of adult cardiac arrest patients treated at Chonnam National University Hospital, a university-affiliated, 1005-bed hospital located in Gwangju, Korea, from January 2008 to June 2017. The Institutional Review Board of the Chonnam National
University Hospital approved the study protocol (CNUH-2018-004). The need for informed consent was waived because of the retrospective nature of the study. We included patients aged 18 years or older who had experienced cardiac arrest and achieved sustained ROSC. Patients were excluded if: (1) the cause of cardiac arrest was trauma, (2) they had a poor pre-arrest neurologic status (Cerebral Performance Category [CPC] 3 or 4), (3) they had an
IP T
intracranial haemorrhage or acute ischaemic stroke on brain imaging obtained after ROSC because
these brain pathologies may themselves contribute to a poor neurologic outcome, (4) serum phosphate
SC R
level measured within 6 hours after ROSC was not available, or (5) information on arrest and CPR
was not available. We also excluded patients with pre-existing parathyroid disease because it could influence the phosphate level. All patients exhibiting sustained ROSC underwent post-cardiac arrest
U
care as per recent resuscitation guidelines [15,16]. According to our protocol, targeted temperature
N
management (TTM) with a target core body temperature of 32–36°C was considered for all adult non-
A
traumatic cardiac arrest survivors who were unable to obey commands. Patients were considered
M
ineligible for TTM if they had other causes of coma other than cardiac arrest, active bleeding, circulatory shock unresponsive to conventional medical therapy, known malignancy in the end stages,
ED
or a poor pre-arrest neurologic status (CPC 3 or 4). Data collection
PT
The following data were obtained from electronic medical records: age, sex, comorbidities, location of arrest (out-of-hospital versus in-hospital), presence of a witness on collapse, bystander CPR, first
CC E
monitored rhythm, dose of adrenaline (epinephrine) administered during CPR, time to ROSC (defined as the time interval from recognition of cardiac arrest to ROSC), aetiology of arrest, Glasgow Coma Scale (GCS) score on admission, sequential organ failure assessment (SOFA) score within the first 24
A
hours, TTM, and outcome assessed using the CPC scale at hospital discharge [17]. An investigator blinded to the study hypothesis determined the CPC score by medical record review. Among the first available laboratory data obtained after ROSC, we also collected levels of serum phosphate and other electrolytes (sodium, potassium, chloride, calcium, and magnesium), as well as parameters reported as outcome predictors (glucose and lactate) [8,18,19], and parameters that might affect phosphate level
(creatinine, albumin, and haemoglobin) [20]. Comorbidities included coronary artery disease, arrhythmia, heart failure, cerebrovascular accident (CVA), neurological disease other than CVA, hypertension, diabetes, pulmonary disease, malignancy, chronic kidney disease (CKD), and liver cirrhosis. At our institution, serum phosphate level was included in the routine laboratory panel, and was determined using UniCel DxC 800 Synchron Clinical Systems (Beckman Coulter, Fullerton,
IP T
USA). The primary outcome of this study was poor outcome at hospital discharge, defined as CPC 3– 5.
SC R
Statistical analysis
Continuous variables were presented as median values with interquartile ranges because all continuous variables showed non-normal distributions. Mann–Whitney U tests were conducted for
U
comparisons of continuous variables. Correlation between continuous variables was assessed by using
N
Spearman’s rank correlation coefficient. Categorical variables were presented as frequencies and
A
percentages. Comparisons of categorical variables were performed using χ2 or Fisher’s exact tests, as
M
appropriate. Receiver operating characteristic (ROC) analyses were performed to examine the performances of serum phosphate and lactate levels in predicting poor outcome. Multivariate logistic
ED
regression analysis was conducted to determine whether serum phosphate level was independently associated with poor outcome. All variables with p <0.2 in the univariate analyses were included in
PT
the multivariate regression model. Backward selection was used to obtain the final model. We subsequently constructed another multivariate prediction model by removing serum phosphate level
CC E
from the final model. ROC analyses were performed to assess the prognostic performances of the multivariate models. Comparisons of the areas under the ROC curves (AUC) were performed as recommended by DeLong et al. [21]. Data were analysed using MedCalc version 16.1 (MedCalc
A
Software, Ostend, Belgium) and SPSS for Windows version 18.0 (SPSS, Inc., Chicago, IL, USA). A two-tailed significance level of 0.05 was used for statistical significance.
Results Characteristics of included patients
A total of 923 adult cardiac arrest patients who achieved ROSC were treated during the study period. Of these patients, 249 patients were excluded as shown in Fig. 1. Thus, 674 patients were included in this study and were divided into two groups according to their outcome at hospital discharge; 209 with good outcomes and 465 with poor outcomes. The clinical characteristics stratified by outcome at hospital discharge are shown in Table 1. Patients with poor outcome were older. They
IP T
had a lower incidence of pre-existing coronary artery disease and high incidences of diabetes and liver
cirrhosis. They had a lower incidence of witnessed collapse, were less likely to receive bystander CPR,
SC R
were more likely to have a non-shockable rhythm, were more likely to have a non-cardiac aetiology, received higher doses of adrenaline during CPR, had a longer time to ROSC, had lower GCS scores, and had higher SOFA scores. TTM was performed in 332 patients (49.3%). The proportion of patients
U
who underwent TTM did not differ between the good and poor outcome groups (98 [46.9%] versus
N
234 [50.3%]; p = 0.459). The distributions of laboratory parameters after ROSC are shown in Table 1.
A
Glucose, lactate, phosphate, potassium, magnesium, and creatinine levels were higher in patients with
M
poor outcome than in those with a good outcome, while calcium, albumin, and haemoglobin levels were lower in patients with poor outcome than those with a good outcome. Sodium and chloride
ED
levels did not differ between the two groups. Prognostic performances of phosphate and lactate levels
PT
Fig. 2 shows the proportion of patients with poor outcome in relation to the phosphate level. Supplemental Fig. 1 shows the correlation between serum phosphate and lactate levels. Serum
CC E
phosphate level was significantly correlated with lactate level (r = 0.726; p <0.001), as well as time to ROSC (r = 0.350; p <0.001), and creatinine level (r = 0.398; p <0.001). The correlation between phosphate and creatinine levels was also significant in patients without pre-existing CKD (r = 0.516;
A
p <0.001). In the ROC analyses, the AUC of phosphate was 0.805 (95% confidence interval [CI], 0.777–0.838). The optimal cut-off value of 5.8 mg dl-1 corresponded to a sensitivity of 74.4% and a specificity of 76.1%. The AUC of phosphate was higher than that of lactate (AUC, 0.778; 95% CI, 0.745–0.809), but this difference did not reach statistical significance. Association between phosphate level and outcome
Table 2 shows the results of the multivariate logistic regression analysis with outcome at hospital discharge as the dependent variable. In the multivariate analysis, higher phosphate level was independently associated with poor outcome at hospital discharge (odds ratio [OR], 1.432; 95% CI, 1.245–1.626; p <0.001). Meanwhile, lactate level was no longer associated with outcome at hospital discharge. Serum lactate level showed an independent association with outcome at hospital discharge
IP T
(OR, 1.100; 95% CI, 1.034–1.169) only after excluding phosphate level from the multivariate model.
Given the impact of the first monitored rhythm [22], arrest aetiology [23] and TTM on outcome
SC R
[24,25], we stratified the overall cohort (shockable or non-shockable, cardiac or non-cardiac, TTM or non-TTM) and performed subgroup analyses. In these subgroup analyses, the association between
serum phosphate level and outcome remained significant across the subgroups (Table 3).
U
Subsequently, subgroup analysis was performed to assess patients with or without diabetes separately
N
as patients with diabetes frequently develop disturbances in their phosphate metabolism [26,27]. This
A
analysis revealed a significant association between serum phosphate level and outcome for both
M
subgroups. Finally, an additional subgroup analysis was performed to assess CKD and non-CKD patients separately. The association between phosphate level and outcome remained significant in
ED
patients without CKD, but it was no longer significant in patients with CKD. Prognostic performance of phosphate in conjunction with other predictors
PT
The AUC for the multivariate logistic regression model including phosphate level was 0.930 (95% CI, 0.908–0.948) (Fig. 3). This was significantly higher than the AUC for phosphate alone (p <0.001);
CC E
however, it did not differ from the AUC for the multivariate logistic regression model excluding phosphate level (AUC, 0.925; 95% CI, 0.902–0.944).
A
Discussion
The present study investigated the prognostic value of serum phosphate level in 674 adult patients
who were resuscitated following cardiac arrest. Phosphate levels were significantly higher among the patients with poor outcome. The ROC curve of phosphate level yielded an AUC of 0.805 for the prediction of poor outcome. In the multivariate regression model, phosphate level showed a
significant association with poor outcome after adjustment for confounders. To the best of our knowledge, this study is the largest study investigating the association between phosphate level and outcome in cardiac arrest patients, and is also the first study to report phosphate level as an independent predictor of outcome after cardiac arrest. Among the laboratory parameters obtainable in the early hours after ROSC, lactate level has
IP T
traditionally been used to estimate the magnitude of ischaemic insult and thus to predict outcome after
cardiac arrest. Although it has been suggested as a promising predictor of outcomes after cardiac
SC R
arrest, a number of studies have failed to demonstrate independent associations between the lactate
levels on admission and outcomes after cardiac arrest [13,28,29]. Several studies have reported that lactate clearance, rather than the lactate level measured on admission, can be used for the prediction
U
of outcomes after cardiac arrest [28,30]. However, this parameter cannot be obtained in the early
N
hours after ROSC because serial lactate measurements are essential to calculate the lactate clearance.
A
In our analyses using multivariate regression models, phosphate level showed a significant association
M
with outcome, while lactate level did only after excluding phosphate level from the model. The AUC of phosphate was higher than that of lactate, although the difference did not reach statistical
ED
significance. Furthermore, serum phosphate levels can be rapidly measured in many hospital laboratories. Thus, our findings suggest that serum phosphate level may be valuable as a prognostic
PT
biomarker during the early post-resuscitation phase. The potential mechanism underlying the association between serum phosphate level and outcome
CC E
shown in our study remains unclear. Experimental studies have suggested that serum phosphate level increases after ischaemic injury, the magnitude of which depends on the duration of ischaemic injury [11,12]. In a study involving serial measurements of serum phosphate levels after superior mesenteric
A
artery ligation in rabbits [12], serum phosphate levels continued to increase as the duration of ischaemia increased. Consistent with these studies, phosphate levels were significantly correlated with time to ROSC in the present study. Therefore, it can be postulated that phosphate is released from damaged cells under ischaemic conditions at rates related to the severity of ischaemic insult. Meanwhile, high phosphate level by itself may mediate harm by inducing endothelial dysfunction and
oxidative stress [31,32]. In an in vitro study using rat mitochondria [32], Oliveira et al. reported that inorganic phosphate stimulated the release of mitochondrial reactive oxygen species. In our study, individual multivariate regression models stratified by the first monitored rhythm, arrest aetiology, TTM, and diabetes revealed higher phosphate levels to be consistently associated with poor outcome across these subgroups. These findings suggest that the first monitored rhythm, arrest
IP T
aetiology, TTM, and diabetes had little influence on the association between the phosphate level and the outcome and that phosphate level can be used for prognostication regardless of these factors.
SC R
Meanwhile, the association between phosphate level and outcome was not significant in patients with
CKD. Phosphate elimination depends on renal function, and the phosphate level is commonly increased in patients with CKD [33]. Hence, the phosphate levels before cardiac arrest might have
U
already been high, and this might have influenced the association between phosphate level and
N
outcome in this subgroup. Interestingly, phosphate level was significantly correlated with creatinine
A
level even in patients without CKD. However, in the patients without CKD, phosphate level was
M
independently associated with poor outcome after adjustment for serum creatinine. Nonetheless, it is still possible that renal function confounded the association between serum phosphate level and
ED
outcome in patients without CKD since serum creatinine level does not reflect renal function precisely [34].
PT
Although phosphate level was identified as an independent predictor of poor outcome in the present study, it demonstrated only a modest predictive performance (AUC of 0.805), precluding the use of
CC E
phosphate level alone for prognostication. The multivariate logistic regression model including phosphate and other significant predictors showed superior prognostic performance compared to phosphate alone. However, no significant difference in the prognostic performance was found
A
between the multivariate model including phosphate and the multivariate model after removing phosphate. These findings suggest that phosphate level should be used in combination with other prognostic indicators in clinical prognostication. Interestingly, albumin and potassium levels also showed a significant association with outcome in our multivariate regression models. With regard to albumin level, several studies have reported
consistent findings [35,36]. In a prospective observational study including 1,269 adult out-of-hospital cardiac arrest patients, Matsuyama et al. [35] reported that serum albumin level on hospital arrival was independently associated with outcome at 1 month. Geddes et al. [37] evaluated the biochemical changes during ventricular fibrillation cardiac arrest in 14 anesthetised pigs, and reported that potassium levels increased with increasing duration of arrest, suggesting potassium level as a potential
IP T
biomarker for the assessment of the severity of ischaemic injury. However, to the best of our knowledge, no clinical study has specifically assessed the relationship between potassium level on
SC R
admission and outcomes after cardiac arrest. Yanagawa et al. [14] investigated the relationship
between laboratory parameters on admission and outcome in 135 adult out-of-hospital cardiac arrest patients. In their study, potassium levels were higher in patients with poor outcome than in those with
U
good outcome, but the association was not significant after adjusting for confounders. Given the small
N
sample size of the study conducted by Yanagawa et al. [14], further studies are required to confirm the
A
association between potassium level on admission and outcomes after cardiac arrest.
M
The present study has several limitations. First, it was a retrospective study of patients treated at a single centre; thus, further studies are required to enhance the generalisability of the findings. Second,
ED
the decision to perform laboratory studies was left to the discretion of treating physician. Serum phosphate level was not available in 88 (9.5%) among the 923 patients reviewed for inclusion, which
PT
might have introduced selection bias. Third, treating physicians were not blinded to the phosphate levels. However, they were unaware of the association between serum phosphate level and outcome,
CC E
and the optimal cut-off value. Therefore, the treatments were unlikely to have been influenced by the phosphate levels. Fourth, we did not include the data on haemodynamic status or vasopressor therapy during the early post-resuscitation phase. Hypotension, as well as vasopressor therapy to reverse it,
A
may have affected the phosphate level by increasing ischaemia in various organs. Thus, these factors could be confounders in the analyses. Fifth, we could not obtain data regarding parathyroid hormone and vitamin D levels, as well as information on the use of phosphate-lowering medication, all of which are potential sources of bias. Sixth, despite our efforts, unmeasured confounders might have existed.
Conclusions A higher serum phosphate level after successful resuscitation was independently associated with poor outcome at hospital discharge in adult cardiac arrest patients. However, given the modest
prognostic indicators in clinical prognostication of cardiac arrest patients.
A
CC E
PT
ED
M
A
N
U
SC R
Conflicts of interest: All authors have no potential conflicts of interest to disclose.
IP T
prognostic performance of serum phosphate level, it should be used in combination with other
References [1]. Smith TW, Cain ME. Sudden cardiac death: epidemiologic and financial worldwide perspective. J Interv Card Electrophysiol 2006;17:199–203.
[2]. Sasson C, Rogers MA, Dahl J, Kellermann AL. Predictors of survival from out-of-hospital cardiac
IP T
arrest: a systematic review and meta-analysis. Circ Cardiovasc Qual Outcomes 2010;3:63–81.
SC R
[3]. Zandbergen EG, Hijdra A, Koelman JH, Hart AA, Vos PE, Verbeek MM, et al. Prediction of poor outcome within the first 3 days of postanoxic coma. Neurology 2006;66:62–8.
U
[4]. Rosén H, Sunnerhagen KS, Herlitz J, Blomstrand C, Rosengren L. Serum levels of the brain-
N
derived proteins S-100 and NSE predict long-term outcome after cardiac arrest. Resuscitation
M
A
2001;49:183–91.
[5]. Oh SH, Park KN, Shon YM, Kim YM, Kim HJ, Youn CS, et al. Continuous Amplitude-Integrated
ED
Electroencephalographic Monitoring Is a Useful Prognostic Tool for Hypothermia-Treated Cardiac
PT
Arrest Patients. Circulation 2015;132:1094–103.
[6]. Holmberg M, Holmberg S, Herlitz J. Effect of bystander cardiopulmonary resuscitation in out-of-
CC E
hospital cardiac arrest patients in Sweden. Resuscitation 2000;47:59–70.
[7]. Fairbanks RJ, Shah MN, Lerner EB, Ilangovan K, Pennington EC, Schneider SM. Epidemiology
A
and outcomes of out-of-hospital cardiac arrest in Rochester, New York. Resuscitation 2007;72:415–24.
[8]. Lee DH, Cho IS, Lee SH, Min YI, Min JH, Kim SH, et al. Correlation between initial serum levels of lactate after return of spontaneous circulation and survival and neurological outcomes in patients who undergo therapeutic hypothermia after cardiac arrest. Resuscitation 2015;88:143–9.
[9]. Bleske BE, Song J, Chow MS, Kluger J, White CM. Hematologic and chemical changes observed during and after cardiac arrest in a canine model--a pilot study. Pharmacotherapy 2001;21:1187–91.
[10]. Makino J, Uchino S, Morimatsu H, Bellomo R. A quantitative analysis of the acidosis of cardiac
IP T
arrest: a prospective observational study. Crit Care 2005;9:R357–62.
SC R
[11]. Smolenski RT, Swierczyński J, Narkiewicz M, Zydowo MM. Purines, lactate and phosphate release from child and adult heart during cardioplegic arrest. Clin Chim Acta 1990;192:155–63.
U
[12]. Dundar ZD, Cander B, Gul M, Karabulut KU, Kocak S, Girisgin S, et al. Serum intestinal fatty
N
acid binding protein and phosphate levels in the diagnosis of acute intestinal ischemia: an
M
A
experimental study in rabbits. J Emerg Med 2012;42:741–7.
[13]. Cho YM, Lim YS, Yang HJ, Park WB, Cho JS, Kim JJ, et al. Blood ammonia is a predictive
ED
biomarker of neurologic outcome in cardiac arrest patients treated with therapeutic hypothermia. Am J
PT
Emerg Med 2012;30:1395–401.
[14]. Yanagawa Y, Sakamoto T, Sato H. Relationship between laboratory findings and the outcome of
CC E
cardiopulmonary arrest. Am J Emerg Med 2009;27:308–12.
[15]. Morrison LJ, Deakin CD, Morley PT, Callaway CW, Kerber RE, Kronick SL, et al. Part 8:
A
Advanced life support: 2010 International Consensus on Cardiopulmonary Resuscitation and Emergency
Cardiovascular
Care
Science
With
Treatment
Recommendations.
Circulation
2010;122:S345–421.
[16]. Callaway CW, Donnino MW, Fink EL, Geocadin RG, Golan E, Kern KB, et al. Part 8: Post-
Cardiac Arrest Care: 2015 American Heart Association Guidelines Update for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2015;132:S465–82.
[17]. Booth CM, Boone RH, Tomlinson G, Detsky AS. Is this patient dead, vegetative, or severely neurologically impaired? Assessing outcome for comatose survivors of cardiac arrest. JAMA
IP T
2004;291:870–9.
SC R
[18]. Shinozaki K, Oda S, Sadahiro T, Nakamura M, Hirayama Y, Watanabe E, et al. Blood ammonia
and lactate levels on hospital arrival as a predictive biomarker in patients with out-of-hospital cardiac
U
arrest. Resuscitation 2011;82:404–9.
N
[19]. Kim SH, Choi SP, Park KN, Lee SJ, Lee KW, Jeong TO, et al. Association of blood glucose at
A
admission with outcomes in patients treated with therapeutic hypothermia after cardiac arrest. Am J
M
Emerg Med 2014;32:900–4.
ED
[20]. Tonelli M, Sacks F, Pfeffer M, Gao Z, Curhan G; Cholesterol And Recurrent Events Trial Investigators. Relation between serum phosphate level and cardiovascular event rate in people with
PT
coronary disease. Circulation 2005;112:2627–33.
CC E
[21]. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics
A
1988;44:837–45.
[22]. Terman SW, Hume B, Meurer WJ, Silbergleit R. Impact of presenting rhythm on short- and long-term neurologic outcome in comatose survivors of cardiac arrest treated with therapeutic hypothermia. Crit Care Med 2014;42:2225–34.
[23]. Wallmuller C, Meron G, Kurkciyan I, Schober A, Stratil P, Sterz F. Causes of in-hospital cardiac arrest and influence on outcome. Resuscitation 2012;83:1206–11.
[24]. Hypothermia after Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve the
IP T
neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549–56.
[25]. Bernard SA, Gray TW, Buist MD, Jones BM, Silvester W, Gutteridge G, et al. Treatment of
SC R
comatose survivors of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med 2002;346:557–63.
U
[26]. Ditzel J, Brøchner-Mortensen J, Rødbro P. Elevated glomerular filtration rate in early diabetes
N
may be explained by increased sodium reabsorption secondary to an impairment in renal tubular
M
A
handling of phosphate. Horm Metab Res Suppl 1981;11:87–9.
[27]. Ditzel J, Brøchner-Mortensen J, Kawahara R. Dysfunction of tubular phosphate reabsorption
ED
related to glomerular filtration and blood glucose control in diabetic children. Diabetologia
PT
1982;23:406–10.
[28]. Lee TR, Kang MJ, Cha WC, Shin TG, Sim MS, Jo IJ, et al. Better lactate clearance associated
CC E
with good neurologic outcome in survivors who treated with therapeutic hypothermia after out-ofhospital cardiac arrest. Crit Care 2013;17:R260.
A
[29]. Starodub R, Abella BS, Grossestreuer AV, Shofer FS, Perman SM, Leary M, et al. Association of serum lactate and survival outcomes in patients undergoing therapeutic hypothermia after cardiac arrest. Resuscitation 2013;84:1078–82.
[30]. Donnino MW, Miller J, Goyal N, Loomba M, Sankey SS, Dolcourt B, et al. Effective lactate
clearance is associated with improved outcome in post-cardiac arrest patients. Resuscitation 2007;75:229–34.
[31]. Shuto E, Taketani Y, Tanaka R, Harada N, Isshiki M, Sato M, et al. Dietary phosphorus acutely
IP T
impairs endothelial function. J Am Soc Nephrol 2009;20:1504–12.
[32]. Oliveira GA, Kowaltowski AJ. Phosphate increases mitochondrial reactive oxygen species
SC R
release. Free Radic Res 2004;38:1113–8.
[33]. Block GA, Hulbert-Shearon TE, Levin NW, Port FK. Association of serum phosphorus and
U
calcium x phosphate product with mortality risk in chronic hemodialysis patients: a national study.
A
N
Am J Kidney Dis 1998;31:607–17.
M
[34]. Shemesh O, Golbetz H, Kriss JP, Myers BD. Limitations of creatinine as a filtration marker in
ED
glomerulopathic patients. Kidney Int 1985;28:830–8.
[35]. Matsuyama T, Iwami T, Yamada T, Hayakawa K, Yoshiya K, Irisawa T, et al. Effect of Serum
PT
Albumin Concentration on Neurological Outcome After Out-of-Hospital Cardiac Arrest (from the CRITICAL [Comprehensive Registry of Intensive Cares for OHCA Survival] Study in Osaka, Japan).
CC E
Am J Cardiol 2018;121:156–61.
[36]. Park I, Lee JH, Kim K, Jo YH, Kim J, Kim T, et al. Prognostic Significance of Initial Serum
A
Albumin on Mortality in Out-of-hospital Cardiac Arrest. J Korean Soc Emerg Med 2013;24:500–7.
[37]. Geddes LA, Roeder RA, Rundell AE, Otlewski MP, Kemeny AE, Lottes AE. The natural biochemical changes during ventricular fibrillation with cardiopulmonary resuscitation and the onset of postdefibrillation pulseless electrical activity. Am J Emerg Med 2006;24:577–81.
Legends to figures Fig. 1. Flow diagram of patient inclusion in the present study.
Fig. 2. Proportion of patients with poor outcome at hospital discharge in relation to phosphate level. The overall cohort was divided into four categories based on the distribution of phosphate levels,
IP T
using quartiles as cut-off values between categories.
SC R
Fig. 3. Receiver operating characteristic curves of phosphate alone, multivariate logistic regression
A
CC E
PT
ED
M
A
N
U
model including phosphate, and multivariate logistic regression model after excluding phosphate.
A ED
PT
CC E
IP T
SC R
U
N
A
M
Figures:
A ED
PT
CC E
IP T
SC R
U
N
A
M
Tables:
Table 1 Clinical and laboratory characteristics stratified by outcome at hospital discharge. Good outcome
Poor outcome
p-
(n = 674)
(n = 209)
(n = 465)
value
64.0 (52.0–
57.0 (49.0–
Age, years
IP T
Total
<0.0
67.0 (55.0–76.0) 69.0)
452 (67.1)
150 (71.8)
302 (64.9)
0.098
Coronary artery disease
83 (12.3)
35 (16.7)
48 (10.3)
0.026
Arrhythmia
28 (4.2)
10 (4.8)
18 (3.9)
0.733
Heart failure
34 (5.0)
14 (6.7)
20 (4.3)
0.261
CVA
38 (5.6)
12 (5.7)
26 (5.6)
1.000
Male sex
01
SC R
74.0)
M
A
N
U
Comorbidities
265 (39.3)
74 (35.4)
191 (41.1)
0.191
Diabetes
178 (26.4)
40 (19.1)
138 (29.7)
0.006
48 (7.1)
11 (5.3)
37 (8.0)
0.273
23 (3.4)
3 (1.4)
20 (4.3)
0.096
Malignancy
56 (8.3)
13 (6.2)
43 (9.2)
0.244
Chronic kidney disease
72 (10.7)
16 (7.7)
56 (12.0)
0.116
Liver cirrhosis
18 (2.7)
1 (0.5)
17 (3.7)
0.035
Witnessed collapse
456 (67.7)
174 (83.3)
282 (60.6)
PT
Pulmonary disease
ED
Hypertension
Neurological disease other
A
CC E
than CVA
<0.0 01
Bystander CPR
443 (65.7)
163 (78.0)
280 (60.2)
<0.0
01 <0.0 First monitored rhythm 01 193 (28.6)
121 (57.9)
72 (15.5)
Non-shockable
481 (71.4)
88 (42.1)
393 (84.5)
Location of arrest
IP T
Shockable
0.067
488 (72.4)
141 (67.5)
In-hospital
186 (27.6)
68 (32.5)
118 (25.4)
3.0 (1.0–5.0)
1.0 (0.0–2.0)
3.0 (2.0–6.0)
Adrenaline administered during
U
CPR, mg 25.0 (13.0–
15.0 (9.0–23.0)
N
Time to ROSC, minute
01 <0.0
PT
Non-cardiac
CC E
Glasgow Coma Scale
01
342 (50.7)
170 (81.3)
172 (37.0)
332 (49.3)
39 (18.7)
293 (63.0)
3.0 (3.0–6.0)
7.0 (3.0–14.0)
3.0 (3.0–4.0)
ED
Cardiac
01 <0.0
M
Aetiology
<0.0
31.0 (18.0–45.0)
A
40.0)
347 (74.6)
SC R
Out-of-hospital
<0.0 01 11.0 (8.0–
SOFA score
12.0 (10.0–
<0.0
14.0)a
01
7.0 (5.0–10.0) 13.0)a 231.0 (166.0–
206.0 (152.0–
247.0 (176.0–
<0.0
311.5)b
272.0)c
325.0)d
01
8.9 (5.6–13.3)e
5.3 (3.2–8.2)f
10.5 (7.6–14.6)g
A
Glucose, mg dl-1
<0.0 Lactate, mmol l-1
01 Phosphate, mg dl-1
6.6 (4.6–8.5)
4.5 (3.5–5.8)
7.7 (5.8–8.8)
<0.0
01 141.0 (137.0–
141.0 (138.0–
141.0 (137.0–
144.0)
143.0)
144.0)
4.2 (3.7–5.2)
3.8 (3.5–4.2)
4.6 (3.8–5.6)
Sodium, mEq l-1
0.919
<0.0 Potassium, mEq l-1
107.0 (103.0–
107.0 (104.0–
Chloride, mEq l-1
IP T
01 106.0 (102.0–
0.283
8.0 (7.4–8.5)
Creatinine, mg dl-1
2.2 (2.0–2.5)
1.1 (0.9–1.3)
M
Albumin, g dl-1
ED
3.0 (2.4–3.5)
12.3 (10.0–
Hemoglobin, g dl-1
01 <0.0
2.5 (2.2–2.9) 01 <0.0 1.4 (1.1–2.0) 01 <0.0
3.5 (3.1–3.8)
2.8 (2.3–3.3) 01
13.8 (11.6–
<0.0 11.7 (9.5–13.4)
15.1)
PT
14.2)
<0.0
7.6 (6.9–8.1)
A
1.3 (1.0–1.7)
SC R
7.7 (7.0–8.3)
2.4 (2.1–2.8)
110.0)
U
Magnesium, mg dl-1
110.0)
N
Calcium, mg dl-1
110.0)
01
Data are presented as n (%) or as medians with interquartile ranges. CVA, cerebrovascular accident;
CC E
CPR, cardiopulmonary resuscitation; ROSC, restoration of spontaneous circulation; SOFA, sequential organ failure assessment.
A
Missing data; a n = 2; b n = 7; c n = 4; d n = 3; e n = 14; f n = 9; g n = 5.
Table 2 Multivariate logistic regression analyses of poor outcome at hospital discharge. OR (95% CI)
p-value
Age
1.032 (1.014 – 1.050)
0.001
Bystander CPR
0.617 (0.345 – 1.103)
0.103
Shockable rhythm
0.430 (0.233 – 0.796)
0.007
Time to ROSC
1.035 (1.016 – 1.055)
<0.001
Cardiac aetiology
0.316 (0.170 – 0.589)
<0.001
Glasgow Coma Scale
0.827 (0.760 – 0.901)
<0.001
SOFA score
1.170 (1.064 – 1.286)
0.001
Phosphate
1.423 (1.245 – 1.626)
<0.001
Albumin
0.608 (0.399 – 0.928)
SC R
U 0.021
N
A
Model 2
1.023 (1.006 – 1.041)
0.008
Shockable rhythm
0.450 (0.248 – 0.815)
0.008
Time to ROSC
1.037 (1.018 – 1.056)
<0.001
0.341 (0.187 – 0.621)
<0.001
0.812 (0.746 – 0.885)
<0.001
1.180 (1.071 – 1.300)
0.001
Lactate
1.100 (1.034 – 1.169)
0.002
Potassium
1.375 (1.068 – 1.770)
0.013
Albumin
0.576 (0.375 – 0.883)
0.011
ED
A
CC E
SOFA score
PT
Glasgow Coma Scale
M
Age
Cardiac aetiology
IP T
Model 1
The multivariate model 1 included all variables with p <0.2 in univariate analyses, whereas model 2 included the variables included in model 1 except phosphate level. From the result of the HosmerLemeshow test, the p-value of Model 1 was 0.190 and that of Model 2 was 0.347. OR, odds ratio; CI, confidence interval; CPR, cardiopulmonary resuscitation; ROSC, restoration of spontaneous circulation; SOFA, sequential organ failure assessment.
Table 3 Adjusted odds ratio and 95% confidence interval for poor outcome associated with each 1 mg dl-1 increase in serum phosphate level by subgroup. Subgroup
OR (95% CI)
p-value
Shockable
1.284 (1.014 – 1.626)
0.038
Non-shockable
1.570 (1.326 – 1.858)
<0.001
Cardiac
1.459 (1.210 – 1.758)
<0.001
Non-cardiac
1.501 (1.171 – 1.925)
0.001
TTM
1.493 (1.226 – 1.817)
<0.001
Non-TTM
1.697 (1.287 – 2.237)
<0.001
IP T
First monitored rhythm
N
1.746 (1.297 – 2.349)
<0.001
Non-diabetes
1.438 (1.229 – 1.682)
<0.001
Pre-existing CKD
Non-CKD
1.939 (0.719 – 5.228)
0.191
1.448 (1.249 – 1.680)
<0.001
PT
CKD
ED
Diabetes
M
A
Diabetes
U
TTM
SC R
Aetiology of arrest
CC E
OR, odds ratio; CI, confidence interval; TTM, targeted temperature management; CKD, chronic
A
kidney disease.