blood ammonia is a predictive biomarker “>American Journal of Emergency Medicine (2012) 30, 1395-1401

Original Contribution

Blood ammonia is a predictive biomarker of neurologic outcome in cardiac arrest patients treated with therapeutic hypothermia

Young Mo Cho MD ^a, Yong Su Lim MD, PhD ^a,?, Hyuk Jun Yang MD, PhD ^a,

Won Bin Park MD ^a, Jin Seong Cho MD ^a, Jin Joo Kim MD ^a, Sung Youl Hyun MD, PhD ^a,

Mi Jin Lee MD ^b, Young Joon Kang MD ^c, Gun Lee MD, PhD ^a

^aDepartment of Emergency Medicine, Gachon University Gil Hospital, Incheon, 405-760, South Korea ^bDepartment of Emergency Medicine, Konyang University Hospital, Nonsan, South Korea ^cDepartment of Emergency Medicine, Cheju National University Hospital, Jeju City, South Korea

Received 6 July 2011; revised 31 July 2011; accepted 10 October 2011

Abstract

Purpose: The aim of this study was to investigate the value of commonly examined laboratory measurements, including ammonia and lactate, in predicting neurologic outcome of out-of-hospital cardiac arrest (OHCA) patients treated with Therapeutic hypothermia .

Methods: This was a retrospective cohort study of patients with a return of spontaneous circulation after OHCA who were treated with TH between February 2007 and July 2010. We measured typical blood measurements on arrival at the emergency department. The subjects were classified into 2 groups: the Good neurologic outcome group (Cerebral Performance Category [CPC] 1-2 at 1 month) and the poor neurologic outcome group (Cerebral Performance Category 3-5). We compared blood biomarker levels and basal characteristics between the 2 groups. Logistic regression analyses were performed to determine independent biomarkers that predict poor neurologic outcome.

Results: A total of 117 patients were included. Between the 2 groups, significantly different levels of blood measurements included hemoglobin level, pH, PaO2, PaCO2, base excess, albumin, glucose, potassium, chloride, bilirubin, phosphorous, and ammonia. In multivariate analyses, blood ammonia level (N96 mg/dL; odds ratio [OR], 7.240; 95% confidence interval [CI], 1.718-30.512), noncardiac causes (OR, 46.215; 95% CI, 9.670-220.873), and time interval from collapse to return of spontaneous circulation (N33 min; OR, 5.943; 95% CI, 1.543-22.886) were significantly related to poor neurologic outcome.

Conclusion: Among the blood measurements on emergency department arrival, blood ammonia (N96 mg/dL) was the only independent predictive biomarker of poor neurologic outcome. Thus, higher blood ammonia level was associated with poor neurologic outcome in OHCA patients treated with TH.

(C) 2012

* Corresponding author. Tel.: +82 32 460 3015; fax: +82 32 460 3019.

E-mail addresses: [email protected] (Y.M. Cho), [email protected] (Y.S. Lim), [email protected] (H.J. Yang), [email protected] (W.B. Park), [email protected] (J.S. Cho), [email protected] (J.J. Kim), [email protected] (S.Y. Hyun), [email protected] (M.J. Lee), [email protected] (Y.J. Kang), [email protected] (G. Lee).

0735-6757/$ - see front matter (C) 2012 doi:10.1016/j.ajem.2011.10.009

Introduction

Out-of-hospital cardiac arrest (OHCA) occurs in about 1 in 2500 adults in the developed world each year [1]. However, despite advances in postcardiac arrest care, the survival rate to hospital discharge with good neurologic outcome after OHCA remains low [2,3].

Therapeutic hypothermia is a proven Standard therapy used to reduce brain injury in comatose survivors with cardiopulmonary arrest (CPA). Recently, the use of mild therapeutic hypothermia (MTH) in comatose survivors with CPA, which is commonly used worldwide, was evaluated as part of the chain for cardiopulmonary resuscitation (CPR) and emergency cardiac care [4].

Although early and accurate prognostication of neuro- logic outcome is important and an essential component of postcardiac arrest care, it is difficult and challenging. Therefore, many investigators have previously evaluated predictive biomarkers, including brain-specific biomarkers, such as S-100 protein and Neuron-specific enolase in survivors with CPA [5,6]. However, most studies were performed in patients who were not treated with TH, and most general hospitals are not able examine biomarkers.

Recently, 2 studies revealed that laboratory findings including serum ammonia and lactate levels were predictors of neurologic outcome in Postcardiac arrest patients [7,8]. However, in those studies, only a small proportion of patients received TH, which is one of standard therapy, and the number of patients who had good neurologic outcome was relatively small.

Thus, the aim of this study was to investigate the association between neurologic outcome and common laboratory measurements, including ammonia and lactate level on arrival in the emergency department (ED) and the value of these measurements as predictive factors of neurologic outcome in OHCA patients treated with TH.

Methods

Study population

The setting of the study was university hospital ED with an Annual patient volume of 90 000. We conducted a single- center and retrospective cohort study between February 2007 and July 2010 at a tertiary hospital located in Incheon city, South Korea. This study was approved by the institutional review board of our hospital.

The subjects of the present study consisted of patients aged 18 years or older who were successfully resuscitated after nontraumatic OHCA and then treated with TH. We excluded patients with preexisting renal failure and liver cirrhosis because these factors can influence laboratory results. Patients who were referred from another hospital after return of spontaneous circulation (ROSC) were also

excluded because we could not obtain an initial blood sample. In addition, patients whose family refused to treat with TH were excluded.

Demographic data were obtained from a prospective registry database involving all patients treated in the ED after cardiac arrest and medical record of our hospital and other hospitals. The collected basal characteristics of the study patients were sex, age, location of arrest, witness of arrest, performance of bystander CPR, time interval from collapse to chest compression, time interval from collapse to ROSC, initial cardiac rhythm, causes of arrest, total dose of epinephrine, Acute physiology and chronic health evaluation II score, Sequential Organ Failure Assessment score, and the elapse time related to TH.

The cause of cardiac arrest was determined by emergency physicians and cardiologists based on medical history and clinical and laboratory findings. We defined cardiac causes of arrest if they had no definite noncardiac cause of arrest and if they had the evidence of cardiac cause and physician confirmed that those are related to cardiac arrest. We defined the evidence of cardiac causes as abnormalities in cardiac enzyme, electrocardiography, echocardiography, coronary angiography, coronary computed tomography, and electro- physiologic study. If there were no definite evidences of cardiac or noncardiac causes, we have classified them as an unknown cause.

Therapeutic hypothermia

Unconscious patients with Glasgow Coma Scale less than 9 after ROSC were treated with TH and admitted to the intensive care unit and treated with standard intensive care including invasive monitoring, hemodynamic support, mechanical ventilation, and analgesia sedation. The exclu- sion criteria of MTH were a shock despite the use of inotropic agents (systolic blood pressure b90 mm Hg), active bleeding, trauma, or possible causes of coma other than cardiac arrest (head trauma or cerebrovascular accident).

The MTH was performed by surface (exovascular) or internal (endovascular) cooling techniques as soon as possible after ROSC. The Surface cooling technique was performed by the application of water-circulating pad (Blanketrol; CSZ Medical, Cincinnati, OH) or water- circulating hydrogel pad (Arctic Sun; Medivance, Inc, Louisville, CO), and all of the patients received internal cooling technique who were inserted an intravascular cooling catheter (CoolLine; Alsius Corporation, Irvine, CA) into the Right internal jugular vein connected with a cooling device (CoolGard 3000; Alsius Corporation). The core temperature was measured by rectal probe. The target temperature was 32?C to 34?C, and the duration of maintenance was 24 hours for all patients. The Rewarming rate was 0.3?C per hour with the same cooling device. MTH methods of exovascular and endovascular techniques were alternately selected.

Laboratory measurements

We retrospectively reviewed the following laboratory data of individual patients measured on ED arrival. All of the blood samples were obtained within 10 minutes after arrival at our ED: complete blood cell count , including white blood cell count , hemoglobin level, and platelet count; and arterial blood gas analysis (ABGA) data, including pH, PaO2, PaCO2, HCO₃, base excess , ionized calcium (Ca), and lactate; serum biochemistry data, including protein, albumin, glucose, total bilirubin, Lactate dehydrogenase , Creatine phosphokinase , sodium (Na), potassium (K), chloride (Cl), aspartate aminotransferase, alanine amino- trasferase, Ca, phosphorous (P), amylase, ammonia, Blood urea nitrogen , and creatinine.

The CBC was evaluated by Sysmex XE-2100 (Sysmex Corporation, Kobe, Japan), and the ABGA was conducted by a GEM Premier 3000 analyzer (Instrumentation Laboratory Company, Bedford, MA). The biochemistry tests were determined with HITACH 7600 (Hitachi High-Technologies Corporation, Tokyo, Japan). All of the blood samples were obtained upon arrival at our ED.

Assessment of neurologic outcome

At 1 month after the date of the cardiac arrest, we assessed the clinical neurologic outcome of patients using Cerebral Performance Category . The performance categories were defined as follows: CPC1, “conscious and alert with normal neurologic function or only slight cerebral disability”; CPC2, “conscious with moderate cerebral disability for part-time work in sheltered environment or independent existence”; CPC3, “conscious with severe cerebral disability precluding independent existence”; CPC4, “comatose or in a Persistent vegetative state“; and CPC5, “brain dead or death.” For the purpose of this study, the patients were classified into 2 groups based on neurologic outcomes: the “good neurologic outcome” group (CPC1 and CPC2) and the “poor neurologic outcome” group (CPC3-CPC5).

Statistical analysis

Data were analyzed using SPSS (SPSS version 19.0; SPSS Inc, Chicago, IL) and MedCalc (MedCalc 11.3 version; MedCalc Inc, Mariakerke, Belgium). Continuous data are presented as the median and interquartile range as appropriate. Univariate analysis was performed by using the Mann-Whitney U test for continuous variables or the ?2 test for categorical variables. All variables found to be significant by univariate analysis (with selected P b .10 in the factor analysis) then underwent multivariate logistic regression analysis by forward stepwise (likelihood ratio) method. The differences between the area under the curve (AUC) of age,

the time interval from collapse to ROSC, and levels of Hb level, ammonia, lactate, albumin, glucose, total bilirubin, K, pH, and Base deficit were determined by using the nonparametric method proposed by Hanley and McNeil [9]. All statistical tests were 2 sided and P b .05 were considered to be statistically significant.

Results

Basal characteristics of study patients

We have tried TH in 140 patients during the study period, and TH was discontinued because of hemodynamic instability and refractory arrhythmias despite treatment in 3 patients. So, 137 patients have reached at the target temperature and finished TH protocol. Two patients who were already diagnosed with chronic renal failure and 18 patients who were referred from another hospital after ROSC were excluded. Finally, the remaining 117 patients were enrolled in this study. Thirty-four patients (30.1%) had good neurologic outcomes with 27 patients of CPC1 and 7 patients of CPC2.

The noncardiac causes (40 patients) of cardiac arrest were asphyxia (19), hanging (12), respiratory disease (asthma or chronic obstructive pulmonary disease, 3), electrolyte imbalance (2), drug overdose (2), and pulmonary embolism (2). The cardiac causes of cardiac arrest (51 patients) were coronary artery disease (ST elevation myocardial infarct, 25; Non ST elevation myocardiac infarct, 4; variant angina, 14), heart failure (5), and primary arrhythmia (3). The results of the comparisons of the basal characteristics between the good neurologic outcome and poor neurologic outcome groups are shown in Table 1.

The variables that were statistically significant different between the 2 groups (good neurologic outcome vs poor neurologic outcome group) were male sex (85.3% vs 63.9%, P = .021), age (47 vs 53, P = .036), performance of bystander CPR (29.4% vs 12.0%, P = .023), time interval from collapse to chest compression (5.0 vs 7.0 minutes, P = .041), time interval from collapse to ROSC (24.0 vs 38.0 minutes, P =

.002), ventricular fibrillation arrest rhythm (55.9% vs 19.3%,

P = .000), cardiac cause of arrest (82.4% vs 27.7%, P =

.000), APACHE II score (22.0 vs 25.0, P = .003), SOFA

score (9.0 vs 10.0, P = .014), and the time interval from cooling initiation to reaching the target temperature (234.5 vs 114.0 minutes, P = .000).

Results of blood measurements

The univariate analysis of the blood measurements found that the following blood levels were statistically correlated with neurologic outcome: Hb level, pH, PaCO2, PaO2, BE, albumin, glucose, K, Cl, total bilirubin, P, and ammonia (P b

.05) (Table 2, 3, and 4).

Good outcome (n = 34)			Poor outcome (n = 83)	P
Sex (male)	29	(85.3)	53 (63.9)	.021
Age (y)	47	(41.75-57.25)	53 (42.00-63.00)	.036
Location of arrest (residence)	15	(44.1)	44 (53.0)	.382
Witness of arrest	30	(88.2)	61 (73.5)	.082
Bystander CPR	10	(29.4)	10 (12.0)	.023
Collapse to chest compression (min)	5.00	(2.00-10.00)	7.00 (4.00-12.00)	.041
Collapse to ROSC (min)	24.00	(15.00-37.00)	38.00 (24.00-51.00)	.002
Initial cardiac rhythm
VF/VT	19	(55.9)	16 (19.3)	.000
Non-VF/VT
Asystole	13	(38.2)	47 (56.6)
PEA	2	(5.9)	20 (24.1)
Cause of arrest
Cardiac	28	(82.4)	23 (27.7)	.000
Noncardiac	3	(8.8)	37 (44.6)
Unknown	3	(8.8)	23 (27.7)
Total dose of epinephrine (mg)	3.00	(0.75-7.00)	4.00 (2.00-8.00)	.142
APACHE II score	22.00	(18.75-24.00)	25.00 (21.00-29.00)	.003
SOFA score	9.00	(6.00-10.25)	10.00 (8.00-12.00)	.014
Collapse to TH start time (min)	127.50	(93.25-214.50)	166.00 (119.00-214.50)	.118
TH start to reaching time (min)	234.50	(133.75-315.50)	114.00 (15.00-212.00)	.000
Values are expressed as number (percentage) and median (interquartile range), as appropriate. PEA indicates pulseless electrical activity; VF/VT, ventricular fibrillation/pulseless ventricular tachycardia.

As CBC results shown in Table 2, the poor outcome group had lower Hb levels than the good outcome group (13.5 vs 15.7 g/dL, P = .000).

Table 1 Basal characteristics of study population (N = 117)

As the ABGA results shown in Table 3, the poor outcome group had higher PaCO2 (77.0 vs 51.5 mm Hg, P = .000) and BE levels (-15.80 vs -12.25 mmol/L, P = .010) than the good outcome group (P = .000). In addition, the poor outcome group had significantly lower pH (6.980 vs 7.150,

P = .000) and PaO2 levels (26.0 vs 50.0 mm Hg, P = .016) than good outcome group.

As the biochemistry results shown in Table 4, the levels of glucose (240.0 vs 198.5 mg/dL, P = .016), K (5.0 vs 4.1 mEq/L, P = .000), P (7.8 vs 5.8 mg/dL, P =

.000), and ammonia (144.0 vs 93.0 mg/dL, P = .002) in the poor outcome group were higher than that in the good outcome group. In contrast, albumin (3.9 vs 4.2 g/dL, P =

.007) and Cl levels (105 vs 106 mEq/L, P = .019) in the poor outcome group were lower than that in the good outcome group.

Multivariate logistic regression analyses

In the ROC curve analysis, the cutoff point of ammonia level that predicted poor neurologic outcome was 96 mg/dL. The cutoff points of other variables were as follows: the time interval from collapse to ROSC, 33 minutes; Hb level, 14.4 g/dL; arterial pH 7.09; BE, -17.4 mmol/L; lactate, 9.2 mg/dL; albumin, 4.1 g/dL; glucose, 237 mg/dL; total bilirubin, 0.5 mg/dL; and K, 5.6 mEq/L, respectively. Based on these cutoff points, the indicators were transformed into dichotomous variables. The variable including analysis were sex, age, witness of arrest, bystander CPR, initial arrest rhythm, causes of cardiac arrest, time interval from collapse to ROSC, Hb level, pH, BE, lactate, albumin, glucose, total bilirubin, ammonia, and K.

The multivariate logistic regressions analyses by neuro- logic outcome are shown in Table 5. In multivariate analyses, noncardiac causes (odds ratio [OR], 46.215; 95% confidence interval [CI], 9.670-220.873; P = .000), blood ammonia

Table 2 Results of CBC (N = 117)

Variables Good outcome

(n = 34)

WBC (x10³/uL) 12.42 (10.44-15.52)

Hb level (g/dL) 15.70 (14.18-16.75)

Platelet (x10⁴/uL) 215.50 (173.25-268.00)

Poor outcome (n = 83)

12.11 (9.24-15.28)

13.50 (11.40-14.90)

197.00 (157.00-252.00)

P

.508

.000

.207

Values are expressed as median (interquartile range). WBC indicates white blood cell.

Variables	Good outcome (n = 34)	Poor outcome (n = 83)	P
pH	7.150 (6.988-7.293)	6.980 (6.860-7.120)	.000
PaCO2 (mm Hg)	51.50 (35.50-75.25)	77.00 (49.00-96.00)	.000
PaO2 (mm Hg)	50.00 (22.00-81.50)	26.00 (9.00-66.00)	.016
HCO3 (mmol/L)	16.20 (14.18-20.35)	16.60 (9.80-20.90)	.433
BE (mmol/L)	-12.25 (-16.43-8.28)	-15.80 (-21.50-10.50)	.010
Ionized Ca (mmol/L)	0.89 (0.545-1.000)	0.78 (0.570-1.010)	.495
Lactate (mmol/L)	8.75 (6.60-11.30)	9.40 (7.40-11.80)	.290
Values are expressed as median (interquartile range).

levels (OR, 7.240; 95% CI, 1.718-30.512; P = .007), and the

Table 3 Results of ABGA (N = 117)

time interval from collapse to ROSC (OR, 5.943; 95% CI, 1.543-22.886; P = .010) were significantly related to poor neurologic outcome.

Discussion

The patients who achieve ROSC after OHCA had neurologic injury, cardiovascular dysfunction, and multiple organ failure because of whole body Ischemia-reperfusion injury called “Postcardiac arrest syndrome (PACS)” [10]. Whole body ischemia-reperfusion causes multiorgan damage and results in changes in various laboratory measurements. However, studies concerning the changes and mechanism of the underlying laboratory measurements and influence of these in the neurologic outcome of postcardiac arrest patients, especially those treated with TH, are rare.

The present study demonstrated that patients with poor neurologic outcome statistically tended to have higher levels of blood glucose, K, P, ammonia, and PaCO2 and lower levels

Table 4 Result of the serum biochemical analysis (N = 117)

of Hb, Cl, pH, PaO2, and BE on arrival at the ED than those with good neurologic outcome, and blood ammonia levels (N96 mg/dL) were the only laboratory measurements that were an independent predictor of poor neurologic outcome in postcardiac arrest patients who were treated with TH (OR, 7.240; 95% CI, 1.718-30.512).

There were several studies that examined the changes and value of laboratory measurements after CPA. The animal study of Geddes et al [11] demonstrated that serum PaO2, PaCO2, and K levels were changed and associated with outcome. In addition, a clinical study of Yanagawa et al [7] demonstrated significant differences of platelets, pH, PaO2, PaCO2, total protein, K, phosphorus, and ammonia between the good and poor outcome groups. Their results were similar to the results of our study.

Kramer et al [12] reviewed low Hb levels in neurocritical patients and concluded that Hb level concentrations as low as 7 g/dL are well tolerated in most of these patients. Although the patient population of that study did not consist of CPA patients, their finding may be similar to CPA patients. Although there was a difference in Hb level between the

Variables	Good outcome (n = 34)	Poor outcome (n = 83)		P
Protein (g/dL)	7.05 (6.40-7.43)	6.95	(6.58-7.50)	.891
Albumin (g/dL)	4.20 (3.80-4.30)	3.90	(3.60-4.10)	.007
Glucose (mg/dL)	198.50 (168.50-231.25)	240.00	(177.00-308.00)	.016
LDH (U/L)	619.00 (483.50-754.50)	610.00	(437.00-842.00)	.883
CPK (U/L)	166.00 (95.50-381.50)	139.00	(93.00-279.00)	.465
Na (mEq/L)	140.00 (138.00-143.00)	141.00	(137.00-143.00)	.890
K (mEq/L)	4.10 (3.50-5.00)	5.00	(4.10-6.10)	.000
Cl (mEq/L)	106.00 (104-110)	105.00	(102.00-107.00)	.019
Total bilirubin (mg/dL)	0.70 (0.5-0.9)	0.50	(0.40-0.70)	.003
AST (U/L)	49.50 (29.00-81.25)	46.00	(27.00-105.00)	.771
ALT (U/L)	41.00 (23.50-79.50)	36.00	(19.00-66.00)	.347
Ca (mg/dL)	8.75 (8.20-9.13)	8.90	(8.30-9.20)	.451
P (mg/dL)	5.80 (4.45-7.93)	7.80	(5.80-9.10)	.000
Amylase (U/L)	72.00 (54.00-94.50)	70.00	(50.00-100.00)	.728
Ammonia (ug/dL)	93.00 (58.50-143.00)	144.00	(96.25-225.00)	.002
BUN (mg/dL)	15.80 (12.93-18.55)	13.70	(11.10-19.50)	.419
Creatinine (mg/dL)	1.22 (1.00-1.40)	1.20	(1.00-1.40)	.440
Values are expressed as median (interquartile range). ALT indicates alanine aminotrasferase; AST, aspartate aminotransferase; BUN, blood urea nitrogen.

Noncardiac cause	3.833	46.215	9.670	220.873	.000
Ammonia N96	1.980	7.240	1.718	30.512	.007
mg/dL
Collapse time	1.782	5.943	1.543	22.886	.010
N33 min

good and poor outcome groups, there were no patients with extremely low or high Hb levels in our study, and thus, the influence of Hb level to neurologic outcome is clinically estimated to be low.

Table 5 Multivariate analysis of factors associated with poor neurologic outcome

Variables Coefficient OR 95% CI for OR P

Lower Upper

Odds ratios were calculated using a forward stepwise logistic regression analysis.

Variables included in analysis were age older than 65 years, male, nonwitness of cardiac arrest, no bystander CPR, non-ventricular fibrillation rhythms, noncardiac causes, time interval from collapse to ROSC more than 33 minutes, Hb level less than 14.4 (x10³/uL), albumin less than 4.1 g/dL, glucose more than 237 mg/dL, K more than

5.6 mEq/L, total bilirubin less than 0.5 mg/dL, ammonia more than 96 mg/dL, arterial pH less than 7.28, BE less than -17.4 mmol/L, and lactate more than 9.2 mmol/L.

Conventionally, cardiac arrest and the consequent interruption of blood flow to metabolically active tissues cause intense hypercarbic and metabolic acidosis because of anaerobic metabolism and the accumulation of end products, such as CO₂, lactate, and hydrogen ions, during the early postcardiac arrest phase even after ROSC [13]. Some studies have reported that acidosis and BE were well correlated with the prognosis of postcardiac arrest patients. However, blood measurements were not found to be an independent predictor of outcome in postcardiac arrest patients [14]. In addition, the correlation between standard BE and lactate level has been reported to be poor, suggesting that other factors might participate in the pathogenesis of cardiac arrest acidosis [15].

Global ischemia and oxygen deficiency during CPA and CPR result in increased Blood lactate levels because of anaerobic glycolysis and impaired liver function [16,17]. Several studies have reported that blood lactate levels on admission or serially measured values in postcardiac arrest patients were correlated with survival and neurologic outcome but not with duration of CPA [15,17].

Shinozaki et al [8] reported that blood lactate levels were an independent predictor of Favorable neurologic outcomes and that cutoff value of 12.0 mmol/L corresponded to a sensitivity of 90.0% and a specificity of 52.3%. However, in our study, blood lactate levels between the 2 groups were not significantly different and were not an independent predictor (Table 4). This presumably may be due to the examination of different study population and the influence of TH.

Ammonia, which is a component of the physiologic buffer system that maintains pH homeostasis [18], has a negative relationship with pH. Therefore, ammonia levels,

which can be increased by CPA, hypoxia, or shock state, contribute to deteriorated brain functions [19,20]. Recently, some studies have suggested that metabolic and respiratory acidosis during CPA induced the release of ammonia from blood cells and that longer duration of CPA tended to increase ammonia levels [21,22].

Ischemic Hepatic injury contribute to centrilobular necrosis, and this mechanism may be one of the causes of hyperammo- nemia in postcardiac arrest patients [23-25]. Although another source of ammonia is skeletal muscle [25], there was no definite evidence of musculoskeletal cell damage in this study; levels of LDH and CPK levels were within reference range in both good and poor outcome groups.

In a review of hyperammonemia in pediatric patients by Auron and Brophy [27] hyperammonemia accelerated brain injury because of cerebral energy deficits and metabolic abnormalities. Furthermore, once the brain has been exposed to ammonia, intracerebral endogenous protective mecha- nisms that prevent or limit brain damage are triggered.

However, until now, the exact mechanism underlying the increases and the effect of ammonia in CPA patients is unknown and has not been studied so far.

Several recent studies conducted in Japan have reported that significant differences in blood ammonia level between good and poor neurologic outcome patients [7], and only the study by Shinozaki et al [8] demonstrated that ammonia was an independent variable with prognostic value in multivar- iate analysis. In this study, the cutoff value for predicting a favorable outcome was determined as 170 mg/dL (AUC, 0.714; sensitivity, 90.0%; specificity, 58.0%), which was different with the 96 mg/dL (AUC, 0.702; sensitivity, 76.5%; specificity, 55.2%) that was found to predict poor outcome in our study. However, there were some difficulties comparing and interpreting the results of this study compared with other studies because of differences in the characteristics of their study populations with our study. In previous studies, only a small number of patients received MTH, and the number of patients with good outcomes was relatively small. Thus, it was not obvious whether ammonia and lactate levels were independent predictors in postcardiac arrest patients, especially treated with TH. In the results of our study, the poor neurologic outcome group had higher blood ammonia levels than those in the good outcome group, and ammonia was the only independent prognostic factor. To the best of our knowledge, this is first study to report that ammonia is a prognostic factor in postcardiac arrest patients treated with TH.

Higher glucose levels were associated with increased mortality or worse neurologic outcome. Preventing hyper- glycemia reduced the median length of Intensive care unit stay and mortality for the intensive care unit survivors after admission after cardiac arrest. Thus, glucose control in post cardiac arrest syndrome patients has been suggested to be important in critical care. However, it did not seem to be an independent predictor of prognosis of post cardiac arrest syndrome patients [28,29].

Limitation

Our study had several limitations. First, this study was retrospective and a singe-center study with a small study population. Second, not all postcardiac patients with MTH were included in the study population because we could not obtain blood samples on arrival patients who initially arrived at another hospital after CPA. Last, no follow-up analysis of laboratory measurements was performed in our sub- jects. Therefore, we cannot assess the serial changes of these parameters.

Conclusions

Among the blood laboratory measurements on ED arrival, blood ammonia levels greater than 96 mg/dL were only independent predictive biomarker of poor neurologic outcome for OHCA patients treated with MTH. Thus, higher blood ammonia levels on ED arrival was associated with poor neurologic outcome in OHCA patients who were treated with TH.

However, further prospective and multicenter studies are needed to evaluate the exact mechanisms and determine the cutoff values.

References

1. Zheng ZJ, Croft JB, Giles WH, Mensah GA. Sudden cardiac death in the United States, 1989 to 1998. Circulation 2001;104:2158-63.
2. Sasson C, Rogers MA, Dahl J, Kellermann AL. Predictors of survival from out-of-hospital cardiac arrest: a systematic review and meta- analysis. Circ Cardiovasc Qual Outcomes 2010;3:63-81.
3. Nolan JP, Laver SR, Welch CA, et al. Outcome following admission to UK intensive care units after cardiac arrest: a secondary analysis of the ICNARC Case Mix Programme Database. Anaesthesia 2007;62: 1207-16.
4. Peberdy MA, Callaway CW, Neumar RW, et al. Part 9: post-cardiac arrest care: 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2010;122:S768-86.
5. Rundgren M, Karlsson T, Nielsen N, et al. Neuron specific enolase and S-100B as predictors of outcome after cardiac arrest and induced hypothermia. Resuscitation 2009;80:784-9.
6. Wojtczak-Soska K, Lelonek M. S-100B protein: an early prognostic marker after cardiac arrest. Cardiol J 2010;17:532-6.
7. Yanagawa Y, Sakamoto T, Sato H. Relationship between laboratory findings and the outcome of cardiopulmonary arrest. Am J Emerg Med 2009;27:308-12.
8. Shinozaki K, Oda S, Sadahiro T, et al. Blood ammonia and lactate levels on hospital arrival as a predictive biomarker in patients with out-of-hospital cardiac arrest. Resuscitation 2011;82: 404-9.
9. Hanley JA, McNeil BJ. A method of comparing the area under receiver operating characteristic curves derived from the same cases. Radiology 1983;148(3):839-43.
10. Nolan JP, Neumar RW, Adrie C, et al. Post-cardiac arrest syndrome: epidemiology, pathophysiology, treatment, and prognostication. A Scientific Statement from the International Liaison Committee on Resuscitation; the American Heart Association Emergency Cardio- vascular Care Committee; the Council on Cardiovascular Surgery and Anesthesia; the Council on Cardiopulmonary, Perioperative, and Critical Care; the Council on Clinical Cardiology; the Council on Stroke. Resuscitation 2008;79:350-79.
11. Geddes LA, Roeder RA, Rundell AE, et al. The natural biochemical changes during ventricular fibrillation with cardiopulmonary resusci- tation and the onset of post defibrillation pulseless electrical activity. Am J Emerg Med 2006;24:577-81.
12. Kramer AH, Zygun DA. Anemia and red blood cell transfusion in Neurocritical care. Crit Care 2009;13:R89.
13. von Planta M, Weil MH, Gazmuri RJ, et al. Myocardial acidosis associated with CO₂ production during cardiac arrest and resuscitation. Circulation 1989;80:684-92.
14. Makino J, Uchino S, Morimatsu H, Bellomo R. A quantitative analysis of the acidosis of cardiac arrest: a prospective observational study. Crit Care 2005;9:R357-62.
15. Prause G, Ratzenhofer-Comenda B, Smolle-Juttner F, et al. Compar- ison of lactate or BE during out-of-hospital cardiac arrest to determine metabolic acidosis. Resuscitation 2001;51:297-300.
16. De Jonghe B, Cheval C, Misset B, et al. Relationship between blood lactate and early Hepatic dysfunction in acute circulatory failure. J Crit Care 1999;14:7-11.
17. Kliegel A, Losert H, Sterz F, et al. Serial lactate determinations for prediction of outcome after cardiac arrest. Medicine 2004;83(5):274-9.
18. Atkinson DE, Bourke E. Metabolic aspects of the regulation of systemic pH. Am J Physiol 1987;252:F947-956.
19. Nelson RM, Seligson D. Studies on blood ammonia in normal and shock state. Surgery 1953;34:1-8.
20. Sakaguchi Y, Yuge K, Yoshino M, et al. Hyperammonemia in the neonate with hypoxia. Adv Exp Med Biol 1982;153:147-52.
21. Nagamine K. Does blood ammonia level at time of initial treatment predict the outcome of patients in cardiopulmonary arrest on arrival? J Jpn Assoc Acute Med 2005;16:283-8.
22. Ishida H, Mastuoka T, Yokota J, et al. The clinical significance of an increase in the Blood concentration of ammonia in cases with cardiopulmonary arrest on arrival. J Jpn Soc Emerg Med 2002;5:490-4.
23. Bailey RW, Brengman ML, Fuh KC, et al. Hemodynamic pathogen- esis of ischemic hepatic injury following cardiogenic shock/resuscita- tion. Shock 2000;14:451-9.
24. Crowell JW, Smith EE. Oxygen deficit and irreversible hemorrhagic shock. Am J Physiol 1964;206:313-6.
25. Haussinger D. Hepatocyte hepatogeneity in glutamine and ammonia metabolism and the role of an intercellular glutamine cycle during ureogenesis in perfused rat liver. Eur J Biochem 1983;133:269-75.
26. Graham TE, MacLean DA. Ammonia and amino acid metabolism in human skeletal muscle during exercise. Can J Physiol Pharmacol 1992;70:132-41.
27. Auron A, Brophy PD. Hyperammonemia in review: pathophysiology, diagnosis and treatment. Pediatr Nephrol 2011 Mar 23.
28. Sunde K, Pytte M, Jacobsen D, et al. Implementation of a standardised treatment protocol for post resuscitation care after out-of-hospital cardiac arrest. Resuscitation 2007;73:29-39.
29. Losert H, Sterz F, Roine RO, et al. Strict normoglycaemic blood glucose levels in the therapeutic management of patients within 12h after cardiac arrest might not be necessary. Resuscitation 2008;76:214-20.