a b s t r a c t

Purpose: Obesity is a well-known risk factor in various health conditions. We analyzed the association between obesity and clinical outcomes, and its effect on Targeted temperature management (TTM) practice for cardiac ar- rest survivors by calculating and classifying their body mass indexes (BMIs).

Methods: We conducted a retrospective data analysis of adult comatose cardiac arrest survivors treated with TTM from 2008 to 2015. BMI was calculated and the cohort was divided into four categories based on the cut-off values of 18.5, 23.0, and 27.5 kg m^-2. The primary outcome was six-month mortality and the secondary out- comes were neurologic outcome at hospital discharge, cooling rate, and rewarming rate.

Results: The study included 468 patients. Poor neurologic outcome at discharge and six-month mortality were re- ported in 311 (66.5%) and 271 (57.9%) patients, respectively. A multivariate logistic analysis showed that an over- weight compared to normal BMI was associated with lower probability of six-month mortality (odds ratio [OR], 0.481; 95% confidence interval [CI], 0.274-0.846; p = 0.011) and poor neurologic outcome at discharge (OR, 0.482; 95% CI, 0.258-0.903; p = 0.023). BMI correlated with cooling rate (B, - 0.073; 95% CI, -0.108 to

-0.039; p b 0.001), but had no association with rewarming rate (B, 0.003; 95% CI, -0.001-0.008; p = 0.058). Conclusion: Overweight BMI compared to normal BMI classification was found to be associated with lower six- month mortality and poor neurologic outcome at discharge in cardiac arrest survivors treated with TTM. Higher BMI correlated with a slower induction rate.

(C) 2016

Introduction

Obesity is a major Public health problem worldwide and a well- known risk factor for numerous health conditions, including cardiovas- cular diseases [1-3]. Previous studies have examined the association be- tween obesity and clinical outcomes in cardiac arrest survivors [4-7].A study including cardiac arrest survivors treated with therapeutic hypo- thermia reported that obese body mass index (BMI) was associated with mortality [4]. However, other studies demonstrated that moder- ately elevated or overweight BMI was associated with Good neurologic outcomes or survival to discharge [5,6]. This finding corresponds to the “obesity paradox”, which suggests that obesity is associated with better clinical outcome in heart failure patients [8,9]. However, the asso- ciation between obesity and the clinical outcomes in cardiac arrest sur- vivors has not yet been fully elucidated.

* Corresponding author.

E-mail address: [email protected] (B.K. Lee).

targeted temperature management is a Standard care in co- matose cardiac arrest survivors [10]. Although the optimal time for TTM initiation remains unknown, it is recommended to be initiated as soon as possible after return of spontaneous circulation (ROSC) in co- matose cardiac arrest survivors. This is because several studies have re- ported that a lag between ROSC and the initiation of TTM is associated with poor neurologic outcomes or mortality [11,12]. In TTM, the induc- tion and rewarming duration are defined as the temperature changing phases during which cooling and rewarming to target temperatures are achieved. High BMI can interfere with the time required for cooling and rewarming to achieve the target temperature. A previous study re- ported that a higher BMI was associated with prolonged induction time [7]. However, there is a lack of evidence on the association between BMI and cooling and rewarming rates.

We hypothesized that the obesity paradox would be acceptable for comatose cardiac arrest survivors and that BMI has an effect on induc- tion and rewarming durations. The present study, therefore, analyzed the association between BMI and six-month mortality or neurological outcomes at discharge, and the effect of BMI on TTM practice, including cooling and rewarming rates, for cardiac arrest survivors.

http://dx.doi.org/10.1016/j.ajem.2016.10.070

0735-6757/(C) 2016

Methods

Study Design and Population

We performed a retrospective observational cohort study including comatose cardiac arrest survivors treated with TTM (target tempera- ture, 33 ?C; maintenance of therapy duration, 24 h) at OO Hospital be- tween January 2008 and December 2015. This study was approved by the Institutional Review Board of OO Hospital (OO-2016-104).

Adult cardiac arrest survivors over 18 years of age having completed TTM were included. Patients were excluded based on the following criteria: (i) transfer during TTM; (ii) receipt of a TTM protocol with a dif- ferent target temperature of 32 ?C or duration with 72 h; (iii) extracor- poreal membrane oxygenation applied during post-cardiac arrest care; or (iv) incomplete BMI data.

TTM Protocol

TTM was applied to all comatose patients after non-traumatic cardi- ac arrest excluding the following cases: i) refusal of TTM by next-of-kin;

ii) terminal illness or significant bleeding; iii) hemorrhagic stroke. TTM was applied according to a written TTM protocol. The induction phase of TTM is defined as the time interval from the initiation of cooling to at- tainment of target temperature. TTM was induced with ice packs, intra- venous cold saline, and cooling devices as soon as possible after the decision to apply TTM. The maintenance phase is a 24-hour period dur- ing which a target temperature of 33 +- 1 ?C is maintained. During TTM, the target temperature was maintained using either feedback- controlled endovascular catheters (COOLGARD3000(R) Thermal Regula- tion System, Alsius Corporation, Irvine, USA) or Surface cooling devices (Blanketrol(R) II, Cincinnati Subzero Products, Cincinnati, USA or Artic Sun(R) Energy Transfer Pads(TM), Medivance Corp, Louisville, USA). The rewarming phase is defined as the time interval from completion of the maintenance phase to achieving a target temperature. Patients were rewarmed at a target rate of 0.25 ?C h^-1 or 0.5 ?C h^-1.

Data Collection and Outcomes

The following data were obtained for each patient: age, sex, comor- bidities, first monitored rhythm, etiology of cardiac arrest, location of cardiac arrest (out-of-hospital or in-hospital), presence of a witness on collapse, bystander cardiopulmonary resuscitation (CPR), downtime, Glasgow coma scale (GCS) after ROSC, BMI, initial temperature, induc- tion duration, cooling rate, rewarming duration, target rewarming rate (0.25 ?C h^-1 or 0.5 ?C h^-1), actual rewarming rate, cooling device (blan- ket, endovascular, or hydrogel pad), neurologic outcome at discharge, and vital status at six-month after collapse (alive or dead). BMI was cat- egorized according to World Health Organization classification for Asian populations as follows: underweight (b 18.5 kg m^-2), normal (18.5-22.9 kg m^-2), overweight (23.0-27.4 kg m^-2), or obese (>= 27.5 kg m^-2) [13]. The Sequential Organ Failure Assessment score within the first 24 h after admission was used to assess the severity of Multiple organ dysfunction [14]. Neurologic outcome was assessed using the Glasgow-Pittsburgh Cerebral Performance Categories (CPC) at discharge and recorded as CPC 1 (good performance), CPC 2 (moder- ate disability), CPC 3 (severe disability), CPC 4 (vegetative state), and CPC 5 (brain death or death) [15]. Neurologic outcomes were dichoto- mized as either good (CPC 1 and CPC 2) or poor (CPC 3 to 5). The prima- ry outcome was six-month mortality. The secondary outcomes were neurologic outcome at discharge and TTM practice.

Data Analysis

Continuous variables were given as median values with interquartile ranges according to the results of a normality test. Mann-Whitney U tests and Kruskal-Wallis tests were conducted for comparisons of

continuous variables as appropriate. Categorical variables were present- ed as frequencies and percentages. Comparisons of categorical variables were performed using ?² or Fisher exact tests, as indicated. Logistic re- gression analysis was used to examine the association between BMI and six-month mortality or neurologic outcome at discharge after adjusting for relevant covariates. Backward selection was used to obtain the final model. The goodness-of-fit of the final model was evaluated using the Hosmer-Lemeshow test. BMI and BMI classifications were modeled using different models to avoid confounding errors. Multivar- iate Linear regression analysis was used to identify the association be- tween BMI and TTM practice (cooling and rewarming rates). All variables where p b 0.2 in univariate analyses of BMI classifications were included in the Multivariate linear regression analysis. Backward selection was used to obtain the final model. Data were analyzed using PASW/SPSS(TM) software, version 18 (IBM Inc., Chicago, IL, USA). A two-sided significance level of 0.05 was used for statistical significance.

Results

Patient Population

During the study period, 636 adult cardiac arrest survivors were treated with TTM. As summarized in Fig. 1, 46 were treated with extra- corporeal membrane oxygenation, 69 were treated with a different TTM protocol, 7 were transferred, and 44 had incomplete BMI data. Finally, 468 patients were included in this study (Fig. 1).

The median age was 60.0 (47.0-70.0) years; 146 (31.2%) patients

had shockable rhythm, 262 (56.0%) had cardiac etiology, 378 (80.8%) had out-of-hospital cardiac arrest, 358 (76.5%) were witnessed on col- lapse, 234 (50.0%) received bystander CPR, and the median downtime was 28.0 min (17.0-40.0 min) (Table 1). Poor neurologic outcome at discharge and six-month mortality were reported in 311 (66.5%) and 271 (57.9%) patients, respectively.

Association of BMI With Six-Month Mortality

Table 1 compares survivor and non-survivor data six months after cardiac arrest. The six-month survivor group was younger, male domi- nant, had significantly lower incidence of comorbidities (hypertension, diabetes, pulmonary disease, and renal disease), and was more likely to have a shockable rhythm, cardiac etiology, and witnessed collapse (Table 1). Survivors also had a shorter downtime, higher GCS after

Fig. 1. A schematic diagram showing the Selection process of patients for analysis.

Table 1

Baseline characteristics of patients stratified according to outcomes.

	Total (n = 468)	Survivor (n = 197)	Non-survivor (n = 271)	p	Good (n = 157)	Poor (n = 311)	p
Male sex	315 (67.3)	144 (73.1)	171 (63.1)	0.023	115 (73.2)	200 (64.3)	0.052
Age, yr	60.0 (47.0-70.0)	52.0 (40.0-61.5)	66.0 (53.0-73.0)	b0.001	52.0 (40.0-61.0)	64.0 (51.0-72.0)	b0.001
Comorbidities
CAD	67 (14.3)	30 (15.2)	37 (13.7)	0.631	26 (16.6)	41 (13.2)	0.325
Heart failure	39 (8.3)	12 (6.1)	27 (10.0)	0.135	11 (7.0)	28 (9.0)	0.461
Hypertension	188 (40.2)	62 (31.5)	126 (46.5)	0.001	47 (29.9)	141 (45.3)	0.001
Diabetes	123 (26.3)	23 (11.7)	100 (36.9)	b0.001	19 (12.1)	104 (33.4)	b0.001
Pulmonary disease	25 (5.3)	5 (2.5)	20 (7.4)	0.021	3 (1.9)	22 (7.1)	0.019
Renal disease	53 (11.3)	9 (4.6)	44 (16.2)	b0.001	9 (5.7)	44 (14.1)	0.007
CVA	35 (7.5)	10 (5.1)	25 (9.2)	0.092	6 (3.8)	29 (9.3)	0.033
Hepatic disease	11 (2.4)	3 (1.5)	8 (3.0)	0.371	2 (1.3)	9 (2.9)	0.275
First monitored rhythm				b0.001			b0.001
VF/pulseless VT	146 (31.2)	106 (53.8)	40 (14.8)		96 (61.1)	50 (16.1)
PEA	101 (21.6)	37 (18.8)	64 (23.6)		30 (19.1)	71 (22.8)
Asystole	192 (41.0)	43 (21.8)	149 (55.0)		24 (15.3)	168 (54.0)
Unknown	29 (6.2)	11 (5.6)	18 (6.6)		7 (4.5)	22 (7.1)
Etiology				b0.001			b0.001
Cardiac	262 (56.0)	147 (74.6)	115 (42.4)		129 (82.2)	133 (42.8)
Other medical	105 (22.4)	19 (9.6)	86 (31.7)		13 (8.3)	92 (29.6)
Asphyxia	66 (14.1)	17 (8.6)	49 (18.1)		4 (2.5)	62 (19.9)
Drug overdose	31 (6.6)	13 (6.6)	18 (6.6)		11 (7.0)	20 (6.4)
Drowning	4 (0.9)	1 (0.5)	3 (1.1)		0 (0.0)	4 (1.3)
Location				0.061			0.586
OHCA	378 (80.8)	167 (84.8)	211 (77.9)		129 (82.2)	249 (80.1)
IHCA	90 (19.2)	30 (15.2)	60 (22.1)		28 (17.8)	62 (19.9)
Witnessed	358 (76.5)	166 (84.3)	192 (70.8)	0.001	136 (86.6)	222 (71.4)	b0.001
Bystander CPR	234 (50.0)	106 (53.8)	128 (47.2)	0.160	86 (54.8)	148 (47.6)	0.142
Downtime, min	28.0 (17.0-40.0)	24.0 (15.0-34.0)	30.0 (20.0-41.0)	b0.001	21.0 (14.5-30.0)	30.0 (20.0-41.0)	b0.001
Glasgow coma scale	3 (3-4)	3 (3-6)	3 (3-3)	b0.001	4 (3-5)	3 (3-3)	b0.001
Body mass index, kg m-2	22.9 (21.0-25.0)	23.4 (21.5-25.4)	22.5 (20.7-24.7)	0.004	23.4 (21.3-25.4)	22.6 (20.8-24.8)	0.073
Initial temperature, ?C	36.0 (35.2-36.8)	36.3 (35.8-37.0)	35.9 (34.7-36.4)	b0.001	36.4 (35.8-37.0)	36.0 (34.8-36.5)	b0.001
Induction duration, h	2.5 (1.5-4.0), 466a	3.0 (2.0-4.8)	2.0 (1.0-3.1), 269a	b0.001	3.0 (2.0-5.0)	2.0 (1.2-3.5), 309a	b0.001
Rewarming duration, h	12.0 (11.0-14.0), 437a	12.0 (10.0-13.0)	12.0 (11.0-14.0), 240a	0.017	12.0 (10.0-13.0)	12.0 (11.0-14.0), 280a	0.020
SOFA score	10 (7-12)	8 (6-11)	11 (8-13)	b0.001	8 (6-10)	11 (8-12)	b0.001

CAD, coronary artery disease; CVA, cerebrovascular accident; VF, ventricular fibrillation; VT, ventricular tachycardia; PEA, pulseless electrical activity; OHCA, out-of-hospital cardiac arrest; IHCA, in-hospital cardiac arrest; CPR, cardiopulmonary resuscitation; SOFA, sequential organ failure assessment.

^a Number of cases for analysis.

ROSC, higher initial temperature, longer induction duration, shorter rewarming duration, lower SOFA score, and higher BMI compared to non-survivors (Table 1). In multivariate logistic analysis, older age, diabetes, Non-shockable rhythm, non-cardiac etiology, non-witness, lower GCS, longer downtime, and higher SOFA score were associated with six-month mortality (Table 2). BMI as continuous variable was not associated with six-month mortality after adjusting covariates. However, overweight BMI classification (odds ratio [OR], 0.481; 95% confidence interval [CI], 0.274-0.846; p = 0.011) was significantly

associated with reduced six-month mortality compared to normal BMI classification (Table 2).

Association Between BMI and Neurologic Outcome at Hospital Discharge

The group of patients with good neurologic outcomes was younger in age, had significantly lower incidence of comorbidities (hyperten- sion, diabetes, pulmonary disease, renal disease, and cerebrovascular

Table 2

Multivariate logistic regression analysis for six-month mortality.

	Adjusteda OR (95% CI)	p	Adjustedb OR (95% CI)	p
Age, yr	1.042 (1.022-1.063)	b0.001	1.045 (1.025-1.066)	b0.001
Diabetes	3.421 (1.769-6.614)	b0.001	3.287 (1.687-6.402)	b0.001
Shockable rhythm	0.352 (0.179-0.693)	0.003	0.335 (0.170-0.661)	0.002
Cardiac etiology	0.388 (0.207-0.729)	0.003	0.376 (0.201-0.705)	0.002
Witness	0.443 (0.236-0.831)	0.011	0.460 (0.244-0.866)	0.016
Glasgow coma scale	0.625 (0.500-0.782)	b0.001	0.621 (0.496-0.777)	b0.001
SOFA score	1.174 (1.080-1.276)	b0.001	1.176 (1.081-1.280)	b0.001
Downtime, min	1.021 (1.005-1.038)	0.012	1.024 (1.007-1.042)	0.005
Induction duration, h	0.979 (0.874-1.096)	0.707	0.987 (0.883-1.104)	0.824
Rewarming duration, h	1.054 (0.984-1.129)	0.136	1.056 (0.984-1.132)	0.131
Body mass index, kg m-2	0.928 (0.862-1.000)	0.051	NA
Normal	NA		Reference
Underweight	NA		2.111 (0.745-5.984)	0.160
Overweight	NA		0.481 (0.274-0.846)	0.011
Obese	NA		0.679 (0.261-1.765)	0.427

OR, odds ratio; CI, confidence interval; SOFA, sequential organ failure assessment;

^a Model 1, logistic regression analysis including body mass index as continuous variable.

^b Model 2, logistic regression analysis including body mass index as nominal variable.

accident), and was more likely to have a shockable rhythm and cardiac etiology (Table 1). This group also showed a higher incidence of witnessed collapse, shorter downtime, higher GCS, higher initial tem- perature, longer induction duration, shorter rewarming duration and lower SOFA score compared to the group with poor neurologic out- comes (Table 1). However, BMI values did not differ significantly be- tween the groups with good and poor neurologic outcomes (Table 1). In multivariate logistic analysis, older age, cerebrovascular accident, non-shockable rhythm, non-cardiac etiology, lower GCS, higher SOFA score, and longer downtime were associated with poor neurologic out- comes at discharge (Table 3). BMI as a continuous variable was not as- sociated with poor neurologic outcome at discharge after adjusting for covariates. However, overweight BMI classification (OR, 0.482; 95% CI, 0.258-0.903; p = 0.023) was significantly associated with lower poor neurologic outcome compared to normal BMI classification (Table 3).

BMI Classification and TTM Practice

Table 4 shows demographic data according to BMI classification. Normal and overweight BMI were present in 196 (41.9%) and 189 (40.4%) patients, respectively. Women were predominant in the under- weight group, while men were dominant in the other groups (Table 4). First monitored rhythm, etiology of arrest, witness, bystander CPR, downtime, and proportion of cooling device use did not differ signifi- cantly among groups (Table 4). Induction duration and cooling rates dif- fered among groups, while the initial temperature did not (Table 4). The rewarming duration and actual rewarming rate did not differ. Six- month mortality was different among groups, while neurologic out- come at discharge was not (Table 4).

Multivariate linear regression analysis revealed that BMI (B,

-0.073; 95% confidence interval [CI], - 0.108 to - 0.039; p b 0.001) was associated with cooling rate. The actual rewarming rate was not as- sociated with BMI (B, 0.003; 95% CI, -0.001-0.008; p = 0.142), while

GCS (B, 0.012; 95% CI, 0.002-0.022; p = 0.023) was associated with the actual rewarming rate.

Discussion

In this study, overweight BMI compared to normal BMI was associat- ed with decreased long-term mortality and poor neurologic outcome at discharge in cardiac arrest survivors treated with TTM. Increased BMI was associated with decreased cooling rate and prolongation of induc- tion duration, while BMI was not associated with rewarming rate.

The prevalence of obesity is also rapidly increasing in the Korean population [16]. BMI reflects risk for diabetes and cardiovascular disease and the prevalence of diabetes varies according to ethnic groups. There- fore, the cut-off of BMI classification should be adjusted according to

ethnic groups. The World Health Organization experts recommended alternative cut-off points for BMI classification in Asian populations [13]. We adopted the BMI classification for Asian population, because all cohort of the present study were Asian.

High BMI is associated with all-cause mortality including cardiovas- cular disease [2,3]. BMI is an independent factor for development of sud- den cardiac death [17,18]. However, overweight or obese BMI classification rather than normal or underweight BMI classification is shown to be associated with lower mortality in chronic kidney disease, coronary artery disease, and heart failure [8,9,19,20]. The concept of an obesity paradox, in which mildly elevated BMI correlates with good out- comes, has been accepted. Jain et al. reported that overweight BMI classification (25.0-29.9 kg m^-2) is associated with higher post- resuscitation survival compared to underweight (b 18.5 kg m^-2) or normal (18.5-24.9 kg m^-2) BMI classifications in in-hospital cardiac ar- rest survivors [5]. Testori et al. also reported that overweight (25.0-29.9 kg m^-2) BMI classification rather than normal (18.5-24.9 kg m^-2) BMI remained a predictive factor for six-month favorable neurologic out- come in both in-hospital and out-of-hospital cardiac arrest [6]. Those previous studies demonstrated inverse U-shaped relationship between BMI and clinical outcomes [5,6]. Corresponding to these previous stud- ies, overweight BMI classification rather than normal BMI was associat- ed with good neurologic outcome at discharge and six-month survival after adjusting for several prognostic factors, while BMI as a continuous variable had no association with outcomes in our cohort. Contrary to this finding, a study of cardiac arrest survivors treated with therapeutic hypothermia reported that the risk for mortality was significantly great- er in those with a BMI of N 30 kg m^-2 than those with a BMI b 30 kg m^-2 [4]. However, this study did not assess the relationship of mortality among underweight, normal, and overweight BMI classifications be- cause it only compared obesity and non-obesity with a cut-off point of 30 kg m^-2 [4]. Nutritional status, fat mass, obesity distribution, and lean body mass have been postulated to have more prognostic signifi- cance than BMI in the evaluation of body composition [21-24]. Howev- er, these data were not available in our cohort. Instead of BMI, it has been suggested that lean body mass, body fat, and wrist circumference can be used to predict survival in coronary artery disease or heart failure [21,22]. Gastelurrutia et al. reported undernourishment in normal BMI and even in overweight or obese BMI in among subjects with heart fail- ure, concluding that nutritional status is a more salient prognosticator than BMI [23]. Future studies are required to identify the association be- tween obesity and clinical outcome in cardiac arrest survivors by using lean body mass, wrist circumference, or their combination with BMI.

Higher BMI correlated with prolonged induction duration in our co- hort. We calculated the cooling rate to adjust for the initial temperature and found that higher BMI also correlated with slow cooling rate. Fur- thermore, BMI had a negative effect on cooling rate after adjusting for

Table 3

Multivariate logistic regression analysis for poor neurologic outcome at discharge.

	Adjusteda OR (95% CI)	p	Adjustedb OR (95% CI)	p
Age, yr	1.029 (1.009-1.049)	0.005	1.032 (1.011-1.052)	0.002
Previous CVA	6.929 (1.813-26.476)	0.005	6.770 (1.716-26.707)	0.006
Shockable rhythm	0.210 (0.106-0.416)	b0.001	0.201 (0.100-0.403)	b0.001
Cardiac etiology	0.249 (0.125-0.495)	b0.001	0.252 (0.126-0.505)	b0.001
Glasgow coma scale	0.590 (0.472-0.737)	b0.001	0.583 (0.464-0.734)	b0.001
SOFA score	1.124 (1.024-1.234)	0.014	1.130 (1.029-1.242)	0.011
Downtime, min	1.053 (1.031-1.075)	b0.001	1.055 (1.033-1.077)	b0.001
Induction duration, h	0.947 (0.847-1.059)	0.341	0.949 (0.850-1.061)	0.360
Rewarming duration, h	1.075 (0.997-1.159)	0.060	1.084 (1.003-1.171)	0.043
Body mass index, kg m-2	1.010 (0.926-1.102)	0.819	NA
Normal	NA		Reference
Underweight	NA		0.440 (0.151-1.281)	0.132
Overweight	NA		0.482 (0.258-0.903)	0.023
Obese	NA		0.885 (0.297-2.637)	0.827

OR, odds ratio; CI, confidence interval; CVA, cerebrovascular accident; SOFA, sequential organ failure assessment.

^a Model 1, logistic regression analysis including body mass index as continuous variable.

^b Model 2, logistic regression analysis including body mass index as nominal variable.

Table 4

Baseline characteristics stratified by body mass index classifications.

	Underweight (n = 45)	Normal (n = 196)	Overweight (n = 189)	Obese (n = 38)	p
Male sex	16 (35.6)	134 (68.4)	138 (73.0)	27 (71.1)	b0.001
Age, yr	64.0 (46.5-78.5)	60.0 (47.0-71.0)	59.0 (47.0-69.5)	59.5 (51.0-66.0)	0.479
Comorbidities >= 3	9 (20.0)	34 (17.3)	28 (14.8)	7 (18.4)	0.807
Shockable rhythm	12 (26.7)	56 (28.6)	69 (36.5)	10 (26.3)	0.268
Cardiac etiology	24 (53.3)	100 (51.0)	114 (60.3)	21 (55.3)	0.328
Witness	38 (84.4)	143 (73.0)	148 (78.3)	29 (76.3)	0.349
Bystander CPR	23 (51.1)	96 (49.0)	98 (51.9)	17 (44.7)	0.853
Downtime, min	27.0 (13.0-35.0)	30.0 (17.0-40.0)	30.0 (19.5-40.0)	24.0 (13.0-39.3)	0.320
Glasgow coma scale	3 (3-4)	3 (3-4)	3 (3-4)	3 (3-3)	0.062
SOFA score	10 (8-12)	9 (7-12)	10 (7-12)	10 (8-12)	0.547
Cooling device					0.797
Blanket	10 (22.2)	52 (26.5)	47 (24.9)	8 (21.1)
Endovascular	15 (33.3)	50 (25.5)	46 (24.3)	13 (34.2)
Hydrogel pad	20 (44.4)	94 (48.0)	96 (50.8)	17 (44.7)
Initial temperature, ?C	36.0 (34.6-36.3)	36.0 (35.2-36.9)	36.0 (35.2-36.8)	36.1 (35.6-36.8)	0.178
Induction duration, h	1.5 (1.0-3.0), 41a	2.0 (1.0-3.5), 195a	2.8 (1.8-4.0), 188a	4.0 (2.5-5.6)	b0.001
Cooling rate, ?C/h	1.5 (0.7-2.3), 41a	1.4 (0.9-2.0), 189a	1.1 (0.7-1.6), 181a	0.8 (0.5-1.3)	b0.001
Rewarming duration, h	12.0 (11.0-13.0), 40a	12.0 (10.0-14.0), 186a	12.0 (11.5-13.3), 177a	12.0 (9.8-13.0), 34a	0.420
Target rewarming rate					0.961
0.25 ?C/h	33 (82.5)	152 (81.7)	147 (83.1)	29 (85.3)
0.5 ?C/h Actual rewarming rate, ?C/h	7 (17.5) 0.25 (0.23-0.27), 40a	34 (18.3) 0.25 (0.23-0.30), 186a	30 (16.9) 0.25 (0.23-0.28), 177a	5 (14.7) 0.25 (0.24-0.32), 34a	0.479
Poor neurologic outcome	29 (64.4)	142 (72.4)	113 (59.8)	27 (71.1)	0.061
Six month mortality	32 (71.1)	124 (63.3)	93 (49.2)	22 (57.9)	0.010

CPR, cardiopulmonary resuscitation; SOFA, sequential organ failure assessment;

^a Number of cases for analysis.

confounding factors. A study of cardiac arrest survivors treated with TTM also reported that high BMI was associated with delay in achieve- ment of target temperature [7]. Slower cooling rates in higher BMI is ex- plained by the fact that fatty tissue acts as an insulator and traps heat. However, BMI has no effect on the time to target temperature during temperature management with endovascular cooling devices in stroke patients [25]. Jarrah et al. also reported that cooling rate was not associ- ated with BMI during the induction phase using hydrogel pads in cardi- ac arrest survivors [26]. Thus, the effect of BMI on cooling rates remains controversial. However, the results from our cohort may be more reli- able, since our cohort had a larger sample size than previous studies and we included an endovascular cooling device as well as two kinds of surface cooling devices.

Several studies have examined the association between induction duration and outcome in cardiac arrest survivors treated with TTM [11,27-29]. Wolff et al. showed that a 1-hour increase in induction time was associated with a 31% decrease in the probability of good neu- rologic outcome [27]. In contrast to this study, Perman et al. reported that prolonged induction duration of over 300 min had a higher likeli- hood of yielding good neurologic outcomes compared to an induction duration of below 120 min [28]. The survivors and good neurologic out- come groups also had a prolonged induction duration compared to the non-survivors and poor neurologic outcome group in our cohort. How- ever, induction duration was removed from the final multivariate logis- tic regression analysis model for neurologic outcome and six-month mortality in the present study. Although the induction time is not a ro- bust indicator of outcome, induction time may be a valuable indicator in cardiac arrest survivors treated with TTM because high BMI is associat- ed with a longer induction time.

BMI can affect rewarming rates in the same way that it affects induc- tion rates. Although the optimal rewarming rate is still not known, the current consensus is 0.25 to 0.5 ?C h^-1 [30]. In a study that examined the association between active rewarming and outcomes in cardiac ar- rest survivors treated with TTM, the rewarming rate was not associated with outcome [31]. The rewarming duration was not consistently asso- ciated with neurologic outcome at discharge or six-month mortality in our cohort. Furthermore, the actual rewarming rate was dependent on the GCS, not BMI. It is presumed that a slow controlled rate of

rewarming over about 12 h offsets the effect of BMI, whereas induction should be performed as rapidly as possible.

This study had several limitations. First, this study was a single cen- ter retrospective design. Additionally, there were missing BMI data (44/ 636). A possible selection bias hence limits the generalization of our re- sults. The relatively small sample size of subjects with underweight or obese BMI classification might have been insufficient to identify the as- sociation between these BMI classifications and the clinical outcomes, in comparison with the association with normal BMI. Moreover, the study focused on establishing an association rather than determining causa- tion. Therefore, further multi-center studies are required to confirm our findings. Second, despite our efforts, there may have been potential confounding factors that were not considered in our multivariate anal- yses. Third, the current study does not explain the mechanism of the obesity paradox. Cytokines have been explored to identify the mecha- nism of the obesity paradox in heart failure patients [32]. Future inves- tigation is required to explore the role of cytokines in the obesity related outcomes in cardiac arrest survivors. Fourth, withdrawal of life- sustaining therapies is not generally permitted in Korea. This explains the high proportion of poor neurologic outcome at discharge in the present study and this is why we defined the primary outcome as six- month mortality rather than the in-hospital mortality. Finally, the pro- tective effect of obesity paradox reportedly disappears after 5 years of follow-up in patients with coronary artery disease [33]. The outcome in the present study was determined at the time of hospital discharge and six-month after collapse. Thus, future research is needed to assess long-term outcomes.

In conclusion, overweight BMI classification was found to be associ-

ated with lower six-month mortality and poor neurologic outcome at discharge compared to normal BMI classification in cardiac arrest survi- vors treated with TTM. Higher BMI correlated with slower induction rate. Further studies are necessary to assess the mechanism of the obe- sity paradox.

Conflict of Interest Statement

No authors have any conflicts related to this work.

Funding Sources/Disclosures

This work supported by a grant (CRI16024-1) Chonnam National University Hospital Biomedical Research Institute.

References

1. Kenchaiah S, Evans JC, Levy D, et al. Obesity and the risk of heart failure. N Engl J Med

   2002;347:305-13.

   Adams KF, Schatzkin A, Harris TB, et al. Overweight, obesity, and mortality in a large prospective cohort of persons 50 to 71 years old. N Engl J Med 2006;355:763-78.

2. Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath Jr CW. Body-mass index and mor- tality in a prospective cohort of U.S. adults. N Engl J Med 1999;341:1097-105.
3. Breathett K, Mehta N, Yildiz V, Abel E, Husa R. The impact of body mass index on pa- tient survival after therapeutic hypothermia after resuscitation. Am J Emerg Med 2016;34:722-5.
4. Jain R, Nallamothu BK, Chan PS. American heart association national registry of car- diopulmonary resuscitation i. Body mass index and survival after in-hospital cardiac arrest. Circ Cardiovasc Qual Outcomes 2010;3:490-7.
5. Testori C, Sterz F, Losert H, et al. Cardiac arrest survivors with moderate elevated body mass index may have a better neurological outcome: a cohort study. Resusci- tation 2011;82:869-73.
6. Leary M, Cinousis MJ, Mikkelsen ME, Gaieski DF, Abella BS, Fuchs BD. The association of body mass index with time to target temperature and outcomes following post- arrest targeted temperature management. Resuscitation 2014;85:244-7.
7. Clark AL, Fonarow GC, Horwich TB. Obesity and the obesity paradox in heart failure. Prog Cardiovasc Dis 2014;56:409-14.
8. Gruberg L, Weissman NJ, Waksman R, et al. The impact of obesity on the short-term and long-term outcomes after percutaneous coronary intervention: the obesity par- adox? J Am Coll Cardiol 2002;39:578-84.
9. Callaway CW, Donnino MW, Fink EL, 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.
10. Mooney MR, Unger BT, Boland LL, et al. Therapeutic hypothermia after out-of- hospital cardiac arrest: evaluation of a regional system to increase access to cooling. Circulation 2011;124:206-14.
11. Sendelbach S, Hearst MO, Johnson PJ, Unger BT, Mooney MR. Effects of variation in tem- perature management on Cerebral performance category scores in patients who received therapeutic hypothermia post cardiac arrest. Resuscitation 2012;83:829-34.
12. Consultation WHOE. Appropriate body-mass index for Asian populations and its im- plications for policy and intervention strategies. Lancet 2004;363:157-63.
13. Vincent JL, de Mendonca A, Cantraine F, et al. Use of the SOFA score to assess the in- cidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on “sepsis-related problems” of the European So- ciety of Intensive Care Medicine. Crit Care Med 1998;26:1793-800.
14. Booth CM, Boone RH, Tomlinson G, Detsky AS. Is this patient dead, vegetative, or se- verely neurologically impaired? Assessing outcome for comatose survivors of cardi- ac arrest. JAMA 2004;291:870-9.
15. Kim CS, Ko SH, Kwon HS, et al. Prevalence, awareness, and management of obesity in Korea: data from the Korea national health and nutrition examination survey (1998-2011). Diabetes Metab J 2014;38:35-43.
16. Empana JP, Ducimetiere P, Charles MA, Jouven X. Sagittal abdominal diameter and risk of sudden death in asymptomatic middle-aged men: the Paris Prospective Study I. Circulation 2004;110:2781-5.
17. Jouven X, Desnos M, Guerot C, Ducimetiere P. Predicting sudden death in the popu- lation: the Paris Prospective Study I. Circulation 1999;99:1978-83.
18. Navaneethan SD, Schold JD, Arrigain S, Kirwan JP, Nally Jr JV. Body mass index and causes of death in chronic kidney disease. Kidney Int 2016;89: 675-82.
19. Shah R, Gayat E, Januzzi Jr JL, et al. Body mass index and mortality in acutely decom- pensated heart failure across the world: a global obesity paradox. J Am Coll Cardiol 2014;63:778-85.
20. Clark AL, Fonarow GC, Horwich TB. Waist circumference, body mass index, and sur- vival in systolic heart failure: the obesity paradox revisited. J Card Fail 2011;17: 374-80.
21. Lavie CJ, De Schutter A, Patel DA, Romero-Corral A, Artham SM, Milani RV. Body composition and survival in stable coronary heart disease: impact of lean mass index and body fat in the “obesity paradox”. J Am Coll Cardiol 2012;60: 1374-80.
22. Gastelurrutia P, Lupon J, de Antonio M, et al. Body mass index, body fat, and nutri- tional status of patients with heart failure: the PLICA study. Clin Nutr 2015;34: 1233-8.
23. Srikanthan P, Horwich TB, Tseng CH. Relation of muscle mass and fat mass to cardio- vascular disease mortality. Am J Cardiol 2016;117:1355-60.
24. Lyden P, Ernstrom K, Cruz-Flores S, et al. Determinants of effective cooling during endovascular hypothermia. Neurocrit Care 2012;16:413-20.
25. Jarrah S, Dziodzio J, Lord C, et al. Surface cooling after cardiac arrest: effectiveness, skin safety, and adverse events in routine clinical practice. Neurocrit Care 2011;14: 382-8.
26. Wolff B, Machill K, Schumacher D, Schulzki I, Werner D. Early achievement of mild therapeutic hypothermia and the neurologic outcome after cardiac arrest. Int J Cardiol 2009;133:223-8.
27. Perman SM, Ellenberg JH, Grossestreuer AV, et al. Shorter time to target temperature is associated with poor neurologic outcome in post-arrest patients treated with targeted temperature management. Resuscitation 2015;88:114-9.
28. Lee BK, Park KN, Kang GH, et al. Outcome and current status of therapeutic hypo- thermia after out-of-hospital cardiac arrest in Korea using data from Korea Hypo- thermia Network registry. Clin Exp Emerg Med 2014;1:19-27.
29. Deakin CD, Nolan JP, Soar J, et al. European resuscitation council guidelines for resus- citation 2010 section 4. Adult advanced life support. Resuscitation 2010;81: 1305-52.
30. Bouwes A, Robillard LB, Binnekade JM, et al. The influence of rewarming after ther- apeutic hypothermia on outcome after cardiac arrest. Resuscitation 2012;83: 996-1000.
31. Takiguchi M, Yoshihisa A, Miura S, et al. Impact of body mass index on mortality in heart failure patients. Eur J Clin Investig 2014;44:1197-205.
32. Wang ZJ, Zhou YJ, Galper BZ, Gao F, Yeh RW, Mauri L. Association of body mass index with mortality and cardiovascular events for patients with coronary artery disease: a systematic review and meta-analysis. Heart 2015;101:1631-8.