International Journal of Cardiology 222 (2016) 629–635

Contents lists available at ScienceDirect

International Journal of Cardiology

journal homepage: www.elsevier.com/locate/ijcard

Glucocorticoid use during cardiopulmonary resuscitation may be

beneficial for cardiac arrest☆
Min-Shan Tsai, MD, PhD a, Po-Ya Chuang, MHA b, Ping-Hsun Yu, MD c, Chien-Hua Huang, MD, PhD a,
Chao-Hsiun Tang, PhD b, Wei-Tien Chang, MD, PhD a, Wen-Jone Chen, MD, PhD a,d,⁎
a
Department of Emergency Medicine, National Taiwan University Medical College and Hospital, Taipei, Taiwan
b
School of Health Care Administration, Taipei Medical University, Taipei, Taiwan
c
Department of Emergency Medicine, Taipei Hospital, Ministry of Health and Welfare, Taipei, Taiwan
d
Department of Emergency Medicine, Lotung Poh-Ai Hospital, Yilan County, Taiwan

a r t i c l e i n f o a b s t r a c t

Article history: Background: Various studies have indicated that glucocorticoid supplementation during cardiopulmonary resus-
Received 26 May 2016 citation (CPR), in conjunction with vasopressors, may improve outcomes in instances of cardiac arrest. However,
Received in revised form 1 August 2016 further population-based analysis is warranted with respect to resuscitative and long-term survival benefits con-
Accepted 2 August 2016 ferred by administering glucocorticoids in this setting.
Available online 04 August 2016
Methods: A total of 145,644 adult patients who experienced non-traumatic, cardiac arrest occurred at emergency
room during years 2004–2011 were selected for study from the Taiwan National Health Insurance Research da-
Keywords:
Glucocorticoid
tabase. These patients were grouped as steroid and non-steroid recipients during CPR, and group members were
Cardiac arrest matched in terms of patient characteristics, including presenting complaint, prior steroid use, resuscitative drugs
Cardiopulmonary resuscitation and shocks delivered, treatment setting (medical center or not), socioeconomic status, and year that cardiac ar-
Propensity score rest occurred, through propensity scoring. Logistic regression analysis was performed to determine the impact of
Survival steroid usage on survival to admission, survival to discharge, and 1-year survival.
Taiwan National Health Insurance Results: Compared with matched non-steroid group members (n = 8628), patients given steroid (n = 2876)
Research database displayed significantly higher rates of survival to admission (38.32% vs 18.67%; adjusted OR = 2.97, 95% CI
2.69–3.29; p b 0.0001), survival to discharge (14.50% vs 5.61%; adjusted OR = 1.71, 95% CI 1.42–2.05;
p b 0.0001), and 1-year overall survival (10.81% vs 4.74%; adjusted OR = 1.48, 95% CI 1.22–1.79; p b 0.0001).
Steroid use proved more beneficial in patients with COPD or asthma and in the absence of shockable rhythm dur-
ing CPR.
Conclusion: Glucocorticoid use during CPR is associated with improved survival-to-admission, survival-to-
discharge, and 1-year survival rates.
© 2016 Elsevier Ireland Ltd. All rights reserved.

1. Introduction
Cardiac arrest is a major health problem, with an estimated annual
incidence of 295,000 in the US [1]. Despite advances in cardiopulmonary
resuscitation (CPR) and post-cardiac arrest care, less than 20% of cardiac
arrest victims survived to hospital discharge and only 5–8% of them
were discharged with good neurological outcome [2,3]. Vasopressors,
☆ These authors take responsibility for all aspects of the reliability and freedom from
bias of the data presented and their discussed interpretation.
⁎ Corresponding author at: Department of Emergency Medicine, National Taiwan
University Hospital, No. 7, Chung-Shan S. Road, Taipei 100, Taiwan.
E-mail addresses: mshanmshan@gmail.com (M.-S. Tsai), poyachua@usc.edu
(P.-Y. Chuang), ambigus39@gmail.com (P.-H. Yu), chhuang730@ntu.edu.tw
(C.-H. Huang), chtang@tmu.edu.tw (C.-H. Tang), wtchang@ntu.edu.tw (W.-T. Chang),
wjchen1955@ntu.edu.tw (W.-J. Chen).
such as epinephrine and vasopressin, have long been the mainstay of
CPR drug therapy, but emerging evidence now indicates that glucocor-
ticoids given during CPR improve outcomes of cardiac arrest.
The importance of the adrenal gland in this setting was established
through animal [4] and human studies [5–8]. The interval from collapse
to the start of CPR seems to correlate negatively with circulating cortisol
level [6] and bears a positive association with relative adrenal dysfunc-
tion [9], suggesting that cardiac arrest actually impairs release of cortisol
from the adrenal glands. In the no-flow and low-flow phases of cardiac
arrest and CPR, the inadequately perfused adrenal cortex is rendered
functionally insufficient, culminating in impaired cortisol release and
poor sensitivity to adrenocorticotropic hormone (ACTH) stimulation
[5,6,10]. Both the use of epinephrine and vasopressin during CPR
boost plasma cortisol concentration by improving perfusion to adrenal
cortex and medulla [7,11], which in turn helps maintain vascular tone
and enhances effects of administered vasopressors [12,13].

http://dx.doi.org/10.1016/j.ijcard.2016.08.017
0167-5273/© 2016 Elsevier Ireland Ltd. All rights reserved.
630 M.-S. Tsai et al. / International Journal of Cardiology 222 (2016) 629–635
In a rat animal model, intravenous administration of hydrocortisone
during resuscitation reportedly served to increase the rate of return of
spontaneous circulation (ROSC) and shorten CPR duration [14]. A non-
randomized trial that we previously conducted also has indicated that
use of hydrocortisone during CPR improves the ROSC rate in victims
with out-of-hospital cardiac arrest (OHCA), without worsening side ef-
fects [15]; and in two prospective randomized trials detailed by
Mentzelopoulos et al., a combination regimen of vasopressin, epineph-
rine and methylprednisolone during CPR more often resulted in ROSC
among patients suffering in-hospital cardiac arrest (IHCA), with less
in-hospital mortality. Patients given combination therapy maintained
better mean arterial pressures during resuscitation than those receiving
only epinephrine, implying that this therapeutic strategy improved cor-
onary perfusion pressure, thus facilitating ROSC [12,13]. On the other
hand, pre-hospital administration of dexamethasone in patients with
pulseless idioventricular rhythm failed to show any benefit in terms of
ROSC or hospitalization [16]. Hence, the impact of glucocorticoid on
ROSC during CPR is still uncertain.
Because above mentioned clinical studies were somewhat limited by
small patient numbers and design features, this population-based retro-
spective cohort study was undertaken to examine the possible associa-
tion between glucocorticoid use and outcomes in instances of cardiac
arrest. Data from the Taiwan National Health Insurance Research
Database (NHIRD) was used to determine the relationship during and
beyond the period of resuscitation, in ensuing hospital admissions.
2. Methods
2.1. Data sources
The health insurance administrative data retrieved from The Taiwan National Health
Insurance Research Database (NHIRD) was accessed for this study in the interval from Jan-
uary 1, 2003, to December 31, 2012. This nationwide population-based claims database,
released by the National Health Research Institute (NHRI) for research purposes, provides
comprehensive information on more than 22 million enrollees of Taiwan National Health
Insurance (NHI) program, representing approximately 98.4% of the Taiwan population.
Healthcare utilization and demographics are recorded, including the birthdate, gender, in-
surance status, area of residence, dates of services provided, primary and minor diagnostic
Fig. 1. Process of selecting the study subjects. CPR = cardiopulmonary resuscitation; ED = emergency department.
codes, primary and minor procedural codes, and itemized expenditures for each medical
service rendered [17]. This study was approved by the NHRI and institutional review
board of the National Taiwan University Hospital, and all data were de-identified. Confi-
dentiality assurances were addressed in accordance with data regulations of the NHI
Bureau.
2.2. Selection and grouping of study subjects
Initially, a total of 174,068 patients brought to emergency rooms for CPR (procedure
code: 47029C) between January 1, 2004, and December 31, 2011, were identified as can-
didates. Of these, 172,016 given subsequent medical care were recruited for study.
Patients b18 years old at dates of presentation (n = 4041) and those involved in trau-
ma (n = 12,698) were excluded. To identify cardiac arrest victims during acute hospital
care, patients under emergency room observation for N6 h (n = 7197), and patients not
triaged as level one (n = 2436) were also excluded, leaving 145,644 patients available
for study. These subjects were then grouped as steroid (n = 2912) and non-steroid
(n = 142,732) recipients during CPR (Fig. 1). To eliminate potential effects of past steroid
use, any patient with a history of steroid use prior to cardiac arrest was further excluded,
for a total of 90,494 patients in the secondary sample (steroid, n = 1394; non-steroid, n =
89,100).
2.3. Propensity scoring
Propensity scoring was used to reduce selection bias between steroid and non-steroid
treatment groups. The propensity score (PS), defined as the probability of receiving steroid
or not, was first estimated by modeling a logistic regression drawn from pertinent charac-
teristics, including age, gender, presenting complaint, comorbidities, previous steroid use,
drugs and electric shocks delivered during CPR, treatment setting (tertiary medical center
or not), socioeconomic status, geographic distribution and year that cardiac arrest
occurred [18]. Results of logistic regression for PS estimation are reported in Table A.1
and A.2.
Steroid and non-steroid group members were then matched by PS at a 1:3 ratio
(cases:controls) without replacement. The final PS-matched sample was 2876 patients
in the steroid group (36 patients could not be matched with corresponding controls)
and 8628 patients in the non-steroid group. Without prior steroid use, the final sample
in steroid and non-steroid groups was 1393 and 4179, respectively.
2.4. Definition of variables
Primary study outcomes were survival to admission, survival to hospital discharge,
and 1-year survival. The main focus was steroid use during CPR, which was collected
from claims data generated by emergency care. Allowable steroidal supplements were hy-
drocortisone, methylprednisolone, prednisolone, triamcinolone, dexamethasone, and
betamethasone.
 M.-S. Tsai et al. / International Journal of Cardiology 222 (2016) 629–635 631

Controlled variables were age, gender, presenting complaint, shockable rhythm, epi-
nephrine dosage, and total shocks delivered at time of CPR. To account for cardiac cathe-
terization and post-cardiac arrest steroid use during hospital admission as potential
confounders, these two variables were also controlled for analysis with respect to
survival-to-discharge and 1-year survival rates.
2.5. Statistical analyses
Demographic and clinical characteristics of study subjects at dates of emergency room
visit/CPR were delineated, using either chi-square test (for categorical variables) or
Student's t-test (for continuous variables) for group (steroid vs non-steroid) comparisons.
Logistic regression analysis was used to determine the impact of steroid usage on hospital
admission, hospital discharge, and 1-year overall survival. One-year survival curves were
plotted via Kaplan–Meier method, applying the log-rank test to evaluate steroid effect.
Subgroup analyses of survival to admission were also done based on clinical characteris-
tics; and relationships between steroid use during CPR, clinical characteristics and hospital
admission were further examined. All computations relied on standard software (SAS/Stat
v9.3 for Windows; SAS Institute, Cary, NC, USA). The Strengthening the Reporting of Ob-
servational Studies in Epidemiology (STROBE) checklist for cohort study was used to re-
view important components of retrospective review [19].
3. Results
Demographic and clinical characteristics of study subjects are pre-
sented in Table 1, without (N = 145,644) and with (N = 11,504) PS
matching. Without PS matching, steroid (vs non-steroid) recipients
were more likely to have presenting complaint with cardiac events
(18.58% vs 16.28%), respiratory events (24.69% vs 6.78%), and other
events (17.24% vs 10.28%) (overall p b 0.0001). The steroid patients
also had higher percentage of chronic obstructive pulmonary disease
(COPD) (29.36% vs 17.62%; p b 0.0001), asthma (19.61% vs 6.3%;
p b 0.0001), adrenal insufficiency (1.2% vs 0.65%; p = 0.0002), histories
of steroid use (within 1 year) prior to cardiac arrest (52.13% vs 37.58%;
p b 0.0001), and shockable rhythm during CPR (33.41% vs 20.32%;
p b 0.0001). In PS matched subjects, no significant between-group
(steroid vs non-steroid) differences in presenting complaint, clinical
characteristics, CPR events (shockable rhythm, epinephrine dosage),
or socioeconomic status emerged.
Demographic and clinical characteristics of patients who had not
used steroid within 1 year prior to cardiac arrest are shown in Table 2.
Without PS matching, patients in the steroid group presented with
higher incidence of cardiac events (19.51% vs 16.58%), respiratory
events (19.23% vs 5.74%), and other events (17.65% vs 9.65%) (overall
p b 0.0001). The steroid (vs non-steroid) group of such patients also
displayed a higher proportion of COPD (13.2% vs 9.16%; p b 0.0001),
asthma (4.81% vs 2.17%; p b 0.0001), and shockable rhythm during
CPR (35.58% vs 20.7%; p b 0.0001). No significant between-group differ-
ences in presenting complaint, demographics and clinical characteris-
tics were identified in PS-matched subjects.
Overall, the steroid (vs non-steroid) group registered significantly
higher rates of hospital admission (38.32% vs 18.67%; adjusted OR =
2.71, 95% CI 2.69–3.29; p b 0.0001), hospital discharge (14.50% vs
5.61%; adjusted OR = 1.71, 95% CI 1.42–2.05; p b 0.0001), and 1-year
survival (10.81% vs 4.74%; adjusted OR = 1.48, 95% CI 1.22–1.79;
p b 0.0001) (Table 3). One-year survival curves of both groups are
presented in Fig. 3, with significant between-group differences evident
by log-rank test. The relationships between steroid use during CPR,

Table 1
Baseline characteristics between steroid and non-steroid groups.

Before propensity score matching After propensity score matching

Steroid Non-steroid p-Value Steroid Non-steroid p-Value

n = 2912 n = 142,732 n = 2876 n = 8628

n (%) n (%) n (%) n (%)

Age, years (SD) 67.64(16.61) 68.19(17.35) 0.07 67.66(16.60) 67.51(17.41) 0.69

Gender (male) 1828 (62.77%) 89,824 (62.93%) 0.86 1804 (62.73%) 5484 (63.56%) 0.42
Presenting complaint b0.0001 0.90
Cardiac event 541 (18.58%) 23,241 (16.28%) 541 (18.81%) 1625 (18.83%)
Respiratory event 719 (24.69%) 9681 (6.78%) 685 (23.82%) 2030 (23.53%)
Other event 502 (17.24%) 14,679 (10.28%) 501 (17.42%) 1555 (18.02%)
Unknown 1150 (39.49%) 95,131 (66.65%) 1149 (39.95%) 3418 (39.62%)
Diabetes mellitus 758 (26.03%) 39,757 (27.85%) 0.03 749 (26.04%) 2241 (25.97%) 0.94
Hypertension 1412 (48.49%) 66,347 (46.48%) 0.03 1394 (48.47%) 4159 (48.2%) 0.80
Coronary artery disease 587 (20.16%) 28,204 (19.76%) 0.59 578 (20.1%) 1747 (20.25%) 0.86
Heart failure 417 (14.32%) 19,732 (13.82%) 0.44 412 (14.33%) 1211 (14.04%) 0.70
Atrial fibrillation 136 (4.67%) 7692 (5.39%) 0.09 133 (4.62%) 414 (4.8%) 0.70
Chronic kidney disease 180 (6.18%) 12,628 (8.85%) b0.0001 178 (6.19%) 554 (6.42%) 0.66
Malignancy 277 (9.51%) 15,649 (10.96%) 0.01 276 (9.6%) 859 (9.96%) 0.58
COPD 855 (29.36%) 25,144 (17.62%) b0.0001 827 (28.76%) 2468 (28.6%) 0.88
Asthma 571 (19.61%) 8999 (6.30%) b0.0001 537 (18.67%) 1576 (18.27%) 0.63
Adrenal insufficiency 35 (1.20%) 923 (0.65%) 0.0002 33 (1.15%) 117 (1.36%) 0.39
Autoimmune disease 39 (1.34%) 1627 (1.14%) 0.32 38 (1.32%) 133 (1.54%) 0.40
Steroid use prior to cardiac arrest (1-year) 1518(52.13%) 53,632(37.58%) b0.0001 1486 (51.67%) 4463 (51.73%) 0.96
Shockable rhythm 973 (33.41%) 29,003 (20.32%) b0.0001 950 (33.03%) 2881 (33.39%) 0.72
Epinephrine dosage, mg(SD) 9 (8.73) 9.27(16.27) 0.10 9.03(8.75) 9.05(7.31) 0.89
Tertiary medical center 529 (18.17%) 30,703 (21.51%) b0.0001 526 (18.29%) 1643 (19.04%) 0.37
Urbanization level of residence b0.0001 0.99
1 (high) 614 (21.09%) 34,763 (24.36%) 612 (35.33%) 1862 (35.06%)
2 1024 (35.16%) 59,800 (41.90%) 1016 (13.21%) 3025 (13.12%)
3 386 (13.26%) 12,413 (8.70%) 380 (30.18%) 1132 (30.24%)
≥4 (low) 888 (30.49%) 35,743 (25.04%) 868 (27.89%) 2609 (28.71%)
Geographic distribution b0.0001 0.79
Taipei 809 (27.78%) 42,549 (29.81%) 802 (27.89%) 2477 (28.71%)
Northern 310 (10.65%) 21,111 (14.79%) 309 (10.74%) 911 (10.56%)
Central 513 (17.62%) 29,227 (20.48%) 511 (17.77%) 1516 (17.57%)
Southern 294 (10.10%) 21,786 (15.26%) 292 (10.15%) 857 (9.93%)
Kao-pin 884 (30.36%) 22,178 (15.54%) 860 (29.9%) 2603 (30.17%)
Eastern 102 (3.50%) 5881 (4.12%) 102 (3.55%) 264 (3.06%)

COPD = chronic obstructive pulmonary disease; SD = standard deviation.

632 M.-S. Tsai et al. / International Journal of Cardiology 222 (2016) 629–635

Table 2
Baseline characteristics of patients without steroid use prior to cardiac arrest.

Before propensity score matching After propensity score matching

Steroid Non- steroid p-Value Steroid Non-steroid p-Value

n = 1394 n = 89,100 n = 1393 n = 4179

n (%) n (%) n (%) n (%)

Age, years (SD) 66.32 (17.64) 67.63(18.29) 0.42 66.32(17.64) 66.43(18.18) 0.84
Gender (male) 860 (61.69%) 56,130 (63%) 0.32 859 (61.67%) 2621 (62.72%) 0.48
Presenting complaint b0.0001 0.95
Cardiac event 272 (19.51%) 14,772 (16.58%) 272 (19.53%) 799 (19.12%)
Respiratory event 268 (19.23%) 5115 (5.74%) 267 (19.17%) 816 (19.53%)
Other event 246 (17.65%) 8597 (9.65%) 246 (17.66%) 760 (18.19%)
Unknown 608 (43.62%) 60,616 (68.03%) 608 (43.65%) 1804 (43.17%)
Diabetes mellitus 345 (24.75%) 22,254 (24.98%) 0.85 345 (24.77%) 1042 (24.93%) 0.90
Hypertension 603 (43.26%) 36,821 (41.33%) 0.15 603 (43.29%) 1861 (44.53%) 0.42
Coronary artery disease 225 (16.14%) 14,387 (16.15%) 0.99 225 (16.15%) 668 (15.98%) 0.88
Heart failure 120 (8.61%) 8677 (9.74%) 0.16 120 (8.61%) 337 (8.06%) 0.52
Atrial fibrillation 44 (3.16%) 3360 (3.77%) 0.23 44 (3.16%) 132 (3.16%) 1.00
Chronic kidney disease 70 (5.02%) 6563 (7.37%) 0.0009 70 (5.03%) 212 (5.07%) 0.94
Malignancy 79 (5.67%) 5522 (6.2%) 0.41 79 (5.67%) 258 (6.17%) 0.50
COPD 184 (13.2%) 8164 (9.16%) b0.0001 183 (13.14%) 542 (12.97%) 0.87
Asthma 67 (4.81%) 1930 (2.17%) b0.0001 66 (4.74%) 182 (4.36%) 0.55
Adrenal insufficiency 2 (0.14%) 115 (0.13%) 0.88 2 (0.14%) 2 (0.05%) 0.25
Autoimmune disease 4 (0.29%) 357 (0.4%) 0.5 4 (0.29%) 4 (0.1%) 0.10
Shockable rhythm 496 (35.58%) 18,447 (20.7%) b0.0001 495 (35.53%) 1401 (33.52%) 0.17
Epinephrine dosage, mg(SD) 9.40 (9.26) 9.51(19.64) 0.7 9.41(9.27) 9.37(7.47) 0.87
Tertiary medical center 267 (19.15%) 19,786 (22.21%) 0.01 267 (19.17%) 785 (18.78%) 0.75
Urbanization level of residence b0.0001 0.91
1 (high) 333 (23.89%) 22,715 (25.5%) 333 (23.91%) 1019 (24.38%)
2 459 (32.93%) 37,477 (42.07%) 459 (32.95%) 1334 (31.92%)
3 204 (14.63%) 7830 (8.79%) 203 (14.57%) 625 (14.96%)
≥4 (low) 398 (28.55%) 21,071 (23.65%) 398 (28.57%) 1201 (28.74%)
Geographic distribution b0.0001 0.97
Taipei 443 (31.78%) 28,427 (31.91%) 443 (31.8%) 1325 (31.71%)
Northern 133 (9.54%) 13,874 (15.57%) 133 (9.55%) 399 (9.55%)
Central 228 (16.36%) 17,031 (19.12%) 228 (16.37%) 699 (16.73%)
Southern 126 (9.04%) 12,952 (14.54%) 126 (9.05%) 348 (8.33%)
Kao-pin 421 (30.2%) 13,003 (14.59%) 420 (30.15%) 1270 (30.39%)
Eastern 43 (3.08%) 3806 (4.27%) 43 (3.09%) 138 (3.3%)

COPD = chronic obstructive pulmonary disease; SD = standard deviation.

individual clinical characteristics, and hospital admission are shown in
Fig. 2. Subgroup analyses revealed that steroid use is conducive to hos-
pital admission, regardless of age, gender, coexisting conditions (DM,
COPD, asthma, adrenal insufficiency, autoimmune disease), or shock-
able rhythm during CPR. Significant interactions suggested that steroid
use was more beneficial in patients with COPD (p = 0.0008), asthma
(p b 0.0001) and without shockable rhythm during CPR (p b 0.0001).
Among patients with no histories of steroid use prior to cardiac ar-
rest, the steroid (vs non-steroid) treatment group similarly displayed
higher rates of hospital admission (37.19% vs 19.00%; adjusted OR =
2.72, 95% CI 2.35–3.14; p b 0.0001), hospital discharge (13.64% vs
5.12%; adjusted OR = 1.95, 95% CI 1.48–2.55; p b 0.0001) and 1-year
survival (11.13% vs 4.24%; adjusted OR = 1.93, 95% CI 1.48–2.53;
p b 0.0001) (Table 3).

Table 3
Outcomes between steroid and non-steroid groups.

Before propensity score matching After propensity score matching

Steroid Non-steroid Odds ratio Adjusted OR Steroid Non-steroid Odds ratio Adjusted OR
(95% CI) (95% CI) (95% CI) (95% CI)
Total patients n = 2912 n = 142,732 n = 2876 n = 8628
† †
Survival to admission 1118 (38.39%) 21,650 (15.17%) 3.49 (3.23–3.76) 2.62 (2.41–2.86) 1102 (38.32%) 1611 (18.67%) 2.71 (2.47–2.97)† 2.97 (2.69–3.29)†
Survival to discharge 431 (14.8%) 5040 (3.53%) 4.75 (4.27–5.28)† 1.66 (1.44–1.92)† 417 (14.50%) 484 (5.61%) 2.85 (2.48–3.28)† 1.71 (1.42–2.05)†
One-year survival 321 (11.02%) 4570 (3.2%) 3.75 (3.32–4.22)† 1.67 (1.44–1.92)† 311 (10.81%) 409 (4.74%) 2.44 (2.09–2.84)† 1.48 (1.22–1.79)†
Excluding patients with n = 1394 n = 89,100 n = 1393 n = 4179
Steroid use prior to
cardiac arrest
Survival to admission 518 (37.16%) 13,689 (15.36%) 3.26 (2.92–3.64)† 2.42 (2.14–2.74)† 518 (37.19%) 794 (19.00%) 2.52 (2.21–2.88)† 2.72 (2.35–3.14)†
Survival to discharge 190 (13.63%) 3222 (3.62%) 4.21 (3.60–4.92)† 1.72 (1.39–2.12)† 190 (13.64%) 214 (5.12%) 2.93 (2.38–3.59)† 1.95 (1.48–2.55)†
One-year survival 155 (11.12%) 3202 (3.59%) 3.36 (2.83–3.98)† 1.71 (1.39–2.09)† 155 (11.13%) 177 (4.24%) 2.83 (2.26–3.54)† 1.93 (1.48–2.53)†

Adjusted for Age, Gender, Presenting complaint, Shockable rhythm, Shock No., Epinephrine dose, Cardiac catheterization and post cardiac arrest steroid use (survival to discharge,
One-year survival).
CI = confidence interval; OR = odds ratio.
†
p b 0.0001.
 M.-S. Tsai et al. / International Journal of Cardiology 222 (2016) 629–635
Fig. 2. Interaction between steroid use during cardiopulmonary resuscitation and clinical characteristics on survival to hospital admission. COPD = chronic obstructive pulmonary disease;
DM = diabetes mellitus, CI = confidence interval; OR = odds ratio.
4. Discussion
This nationwide population-based study included a sizeable (N =
145,644) sampling of adult patients who suffered non-traumatic cardiac
arrest immediately or within 6 h of emergency room arrival between
years 2004–2011, assessing the impact of glucocorticoid use during
CPR. The preeminent finding is that glucocorticoid administration
during CPR maybe associated with increased frequency of survival to
admission. Besides, our findings also suggested associations between
glucocorticoid use and improved rates of survival to discharge and
1-year survival. To the best of our knowledge, this is the first use of
population-based data in this particular context.
The human hypothalamic–pituitary–adrenal (HPA) axis plays an im-
portant role in bodily reactions to stress and is readily activated by the
Fig. 3. One-year survival curves of these groups. The log-rank test showed significant
difference between these groups (p b 0.0001).
633
acute rigors of cardiac arrest, reportedly displaying greater sensitivity
and reactivity than does the sympathoadrenal system [20]. Inadequate
cortisol release, refractory to the high plasma ACTH and ADH levels en-
countered during and after CPR, indicates that adrenal insufficiency
happens during cardiac arrest [5,8]. High levels of circulating cytokines
(TNF-α, IL-1, IL-6) released during CPR may also further suppress
cortisol synthesis/release, despite the severe stress imposed [21–24].
Use of epinephrine and vasopressin during CPR appears to enhance
blood flow to the adrenal gland and encourage cortisol release [7,11],
which in turn helps maintain vasomotor tone by inhibiting catechol-
O-methyl transferase, blocking catecholamine reuptake [25], facilitating
catecholamine-induced vasoconstriction [26], and reducing nitric oxide
production [27]. Various animal and human studies have demonstrated
that glucocorticoid supplementation during CPR may be beneficial by
maintaining haemodynamic stability and enhancing myocardial func-
tion, thereby improving ROSC rates [4,7,15,28,29]. Although Paris et al.
found no significant benefit (i.e., ROSC) in pre-hospital administration
of dexamethasone [16], our expansive nationwide analysis nonetheless
showed that glucocorticoid given during CPR increases the frequency of
survival to admission in OHCA victims. In patients overall and in popu-
lation segments, including patients without steroid use prior to cardiac
arrest and patients stratified by age, gender, coexisting conditions (DM,
COPD, asthma, adrenal insufficiency, and autoimmune disease), and
shockable rhythm, the benefits of glucocorticoid use were demonstrable
by subgroup analyses.
The present findings similarly suggest that glucocorticoid use during
CPR are associated with better survival to discharge and 1-year survival,
not only in patients overall, but also in patients without steroid use prior
to cardiac arrest. Above outcomes are aligned with those of Schultz CH
et al., who observed that circulating cortisol concentration during CPR
is associated with post-cardiac arrest haemodynamics and 24-hr
survival [5]. Tavakoli et al. also documented an association between
serum cortisol level measured immediately after ROSC and 7-day
survival [30]. Other sources have linked early refractory shock [31]
and 1-year mortality [32] with adrenal insufficiency in the post-
634 M.-S. Tsai et al. / International Journal of Cardiology 222 (2016) 629–635
cardiac arrest period. Earlier efforts have also shown that long-term sur-
vival is improved by glucocorticoid supplementation during cardiac ar-
rest [6,29]. In a trial conducted by Mentzelopoulos et al., patients given
vasopressors and methylprednisolone in combination during CPR re-
quired less epinephrine and shorter periods of advanced life support, in-
dicating less ischemic damage during cardiac arrest. By maintaining
haemodynamic stability and sustaining less injury during CPR, better
outcomes were favored in patients receiving combination therapy
[13]. In our study population, epinephrine dosage and total shocks de-
livered were matched between the steroid and non-steroid groups.
Given that epinephrine was suggested to administrate every 3–5 min,
the CPR duration was unlikely to have differed noticeably between
these 2 groups. Thus, the benefit of glucocorticoid use during CPR on
survival to discharge and 1-year survival may not only be attributed
from better haemodynamic stability and less ischemic damage.
The recovery of spontaneous circulation in cardiac arrest leads to the
global ischaemia-reperfusion (I/R) injury, which including excess free
radical production, systemic inflammatory response, and activated apo-
ptosis pathway [21,22,33,34]. I/R injury accounts for post-cardiac arrest
myocardial and cerebral damage, and multiple organ failure [22]. Gluco-
corticoids reportedly mitigate free-radical lipid peroxidation and main-
tain cell membrane integrity [33,35,36], as well as downregulating the
expression of pro-inflammatory cytokines, such as TNF-α, IL-6, IL-8,
and reducing leukocyte adhesion [37–39]. Besides, several studies also
have demonstrated the ability of glucocorticoid to ameliorate the apo-
ptosis [40–42]. In order to clarify whether the glucocorticoid use during
CPR qualified as an independent correlate of survival to discharge and 1-
year survival, we adjusted post-cardiac arrest steroid use in the multiple
logistic regression analysis. The persistent benefit of glucocorticoid use
during CPR on long-term survival, suggesting that glucocorticoid use
during resuscitation may help to reduce systemic I/R injury and alleviate
post-cardiac arrest syndrome. As regards the effect of post-cardiac ar-
rest steroid use on the outcome of cardiac arrest survivors, further
well-designed studies are deserved.
5. Limitations
A number of study limitations are acknowledged. As a retrospective
observational study, our results demonstrated only an association of
glucocorticoid use during CPR and improved outcomes in cardiac arrest.
The findings may support the further randomized and controlled trials
to examine the real causation between glucocorticoid use and out-
comes. In patients overall, we matched “steroid use prior to cardiac ar-
rest” by group, ignoring steroid type and dosage, thus raising the issue
of steroid dependence level as a confounder of glucocorticoid benefit.
However, we also evaluated glucocorticoid effect in non-users of steroid
prior to cardiac arrest, revealing a consistent benefit of glucocorticoid
use on resuscitation and survival. As another issue, care delivered
post-cardiac arrest was not fully adjusted in multiple logistic regression
analysis of survival to discharge or 1-year survival. Nevertheless, an at-
tempt was made to eliminate potential discrepancies in care by
matching treatment setting (tertiary medical center or not), institution-
al geographic distribution, and years that cardiac arrest took place be-
tween the steroid and non-steroid groups. Clearly, the present study
targeted a homogenous Asian population, so caution is advised in ex-
trapolating results to other races. In addition, the NHIRD does not record
individual information, including body weight, smoking, witnessed col-
lapse, bystander CPR, specific initial rhythm, cause of cardiac arrest,
exact timing and dose of steroid use during CPR, the indication of steroid
use, blood pressure during and after CPR, and plasma cortisol and ACTH
concentrations, which may confound present findings. Besides, the
therapeutic hypothermia was not covered by the NHIRD before 2015.
Neurologic outcomes likewise were unavailable. However, a favorable
neurologic outcome has shown a strong correlation with 1-year survival
rate in OHCA survivors [43].
6. Conclusion
Outcomes of this large population-based study suggest an associa-
tion between glucocorticoid administrated during CPR and improved
survival-to-admission rates in instances of non-traumatic adult cardiac
arrest. Survival-to-discharge and 1-year survival rates may also benefit
from this therapeutic approach.
Authors' contributions
Tsai MS, Yu PH, Huang CH and Chang WT, and Chen WJ contributed
to the study concept and design; Tsai MS, Chuang PY and Yu PH contrib-
uted to the acquisition of the data; Tsai MS, Chuang PY, Huang CH and
Tang CH analyzed and interpreted the data; Tsai MS, Chuang PY and
Yu PH draft the manuscript; Huang CH, Tang CH, Chang WT and
Chen WJ provided critical revision of the manuscript for important intel-
lectual content; Tsai MS, Chuang PY, Yu PH and Huang CH performed
the statistical analysis; Tang CH, Chang WT and Chen WJ supervised
the study. All authors have read and approved the final version of the
manuscript.
Conflicts of interest
None.
Acknowledgments
This manuscript was supported by a research grant from the
National Taiwan University Hospital, [grant number 104-S2755]. This
manuscript has been edited by native English-speaking experts of
BioMed Proofreading.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.ijcard.2016.08.017.
References
[1] D. Lloyd-Jones, R. Adams, M. Carnethon, et al., Heart disease and stroke statis-
tics—2009 update: a report from the American Heart Association Statistics Commit-
tee and Stroke Statistics Subcommittee, Circulation 119 (3) (2009) 480–486.
[2] P.C. Ko, M.H. Ma, Z.S. Yen, C.L. Shih, W.J. Chen, F.Y. Lin, Impact of community-wide
deployment of biphasic waveform automated external defibrillators on out-of-
hospital cardiac arrest in Taipei, Resuscitation 63 (2) (2004) 167–174.
[3] J.P. Marenco, P.J. Wang, M.S. Link, M.K. Homoud, N.A. Estes 3rd., Improving survival
from sudden cardiac arrest: the role of the automated external defibrillator, JAMA
285 (9) (2001) 1193–1200.
[4] P.J. Foley, W.A. Tacker, J. Wortsman, S. Frank, P.E. Cryer, Plasma catecholamine and
serum cortisol responses to experimental cardiac arrest in dogs, Am. J. Phys. 253
(3 Pt 1) (1987) E283–E289.
[5] C.H. Schultz, E.P. Rivers, C.S. Feldkamp, et al., A characterization of hypothalamic–
pituitary–adrenal axis function during and after human cardiac arrest, Crit. Care
Med. 21 (9) (1993) 1339–1347.
[6] K.H. Lindner, H.U. Strohmenger, H. Ensinger, W.D. Hetzel, F.W. Ahnefeld, M.
Georgieff, Stress hormone response during and after cardiopulmonary resuscitation,
Anesthesiology 77 (4) (1992) 662–668.
[7] E. Kornberger, A.W. Prengel, A. Krismer, et al., Vasopressin-mediated adrenocortico-
tropin release increases plasma cortisol concentrations during cardiopulmonary re-
suscitation, Crit. Care Med. 28 (10) (2000) 3517–3521.
[8] T. Ito, D. Saitoh, A. Takasu, T. Kiyozumi, T. Sakamoto, Y. Okada, Serum cortisol as a
predictive marker of the outcome in patients resuscitated after cardiopulmonary ar-
rest, Resuscitation 62 (1) (2004) 55–60.
[9] F. Pene, H. Hyvernat, V. Mallet, et al., Prognostic value of relative adrenal insufficien-
cy after out-of-hospital cardiac arrest, Intensive Care Med. 31 (5) (2005) 627–633.
[10] A. Chalkias, T. Xanthos, Post-cardiac arrest syndrome: mechanisms and evaluation
of adrenal insufficiency, World J. Crit. Care Med. 1 (1) (2012) 4–9.
[11] A.C. Krismer, V. Wenzel, W.G. Voelckel, et al., Effects of vasopressin on adrenal gland
regional perfusion during experimental cardiopulmonary resuscitation, Resuscita-
tion 56 (2) (2003) 223–228.
[12] S.D. Mentzelopoulos, S.G. Zakynthinos, M. Tzoufi, et al., Vasopressin, epinephrine,
and corticosteroids for in-hospital cardiac arrest, Arch. Intern. Med. 169 (1)
(2009) 15–24.
 M.-S. Tsai et al. / International Journal of Cardiology 222 (2016) 629–635
[13] S.D. Mentzelopoulos, S. Malachias, C. Chamos, et al., Vasopressin, steroids, and epi-
nephrine and neurologically favorable survival after in-hospital cardiac arrest: a
randomized clinical trial, JAMA 310 (3) (2013) 270–279.
[14] H. Smithline, E. Rivers, T. Appleton, R. Nowak, Corticosteroid supplementation dur-
ing cardiac arrest in rats, Resuscitation 25 (3) (1993) 257–264.
[15] M.S. Tsai, C.H. Huang, W.T. Chang, et al., The effect of hydrocortisone on the outcome
of out-of-hospital cardiac arrest patients: a pilot study, Am. J. Emerg. Med. 25 (3)
(2007) 318–325.
[16] P.M. Paris, R.D. Stewart, F. Deggler, Prehospital use of dexamethasone in pulseless
idioventricular rhythm, Ann. Emerg. Med. 13 (11) (1984) 1008–1010.
[17] National Health Research Institute, Background of National Health Insurance Data-
base, http://w3.nhri.org.tw/nhird/en/index.htm (Last accessed July, 2014).
[18] R. D'Agostino Jr., Tutorial in Biostatistics: Propensity Score Methods for Bias Reduc-
tion in the Comparison of a Treatment to a Non-Randomized Control Group, 1998.
[19] J.P. Vandenbroucke, E. von Elm, D.G. Altman, et al., Strengthening the Reporting of
Observational Studies in Epidemiology (STROBE): explanation and elaboration, Int.
J. Surg. 12 (12) (2014) 1500–1524.
[20] M.J. O'Leary, A.C. Timmins, J.N. Appleby, et al., Dissociation of pituitary–adrenal and
catecholamine activation after induced cardiac arrest and defibrillation, Br. J.
Anaesth. 82 (2) (1999) 271–273.
[21] C. Adrie, M. Adib-Conquy, I. Laurent, et al., Successful cardiopulmonary resuscitation
after cardiac arrest as a "sepsis-like" syndrome, Circulation 106 (5) (2002) 562–568.
[22] C. Adrie, I. Laurent, M. Monchi, A. Cariou, J.F. Dhainaou, C. Spaulding, Postresuscitation
disease after cardiac arrest: a sepsis-like syndrome? Curr. Opin. Crit. Care 10 (3)
(2004) 208–212.
[23] J. Bro-Jeppesen, J. Kjaergaard, M. Wanscher, et al., Systemic inflammatory response
and potential prognostic implications after out-of-hospital cardiac arrest: a
substudy of the target temperature management trial, Crit. Care Med. 43 (6)
(2015) 1223–1232.
[24] E. Beale, J. Zhu, H. Belzberg, Changes in serum cortisol with age in critically ill pa-
tients, Gerontology 48 (2) (2002) 84–92.
[25] B.A.J.T. Geddes, R.J. Dvorsky, N.M. Lefcoe, Interactions of glucocorticoids and bron-
chodilators on isolated guinea pig trachea and human bronchial smooth muscle,
Am. Rev. Respir. Dis. 110 (1974) 420–427.
[26] R.I. Dorin, P.J. Kearns, High output circulatory failure in acute adrenal insufficiency,
Crit. Care Med. 16 (3) (1988) 296–297.
[27] M. Matsumura, H. Kakishita, M. Suzuki, N. Banba, Y. Hattori, Dexamethasone sup-
presses iNOS gene expression by inhibiting NF-kappaB in vascular smooth muscle
cells, Life Sci. 69 (9) (2001) 1067–1077.
[28] K.H. Lindner, T. Haak, A. Keller, U. Bothner, K.G. Lurie, Release of endogenous vaso-
pressors during and after cardiopulmonary resuscitation, Heart 75 (2) (1996)
145–150.
635
[29] B.C. White, T.J. Petinga, P.J. Hoehner, R.F. Wilson, Incidence, etiology, and outcome of
pulseless idioventricular rhythm treated with dexamethasone during advanced
CPR, JACEP 8 (5) (1979) 188–193.
[30] N. Tavakoli, A. Bidari, V.S. Shams, Serum cortisol levels as a predictor of neurologic
survival in successfully resuscitated victims of cardiopulmonary arrest, J. Cardiovasc.
Thorac. Res. 4 (4) (2012) 107–111.
[31] G. Hekimian, T. Baugnon, M. Thuong, et al., Cortisol levels and adrenal reserve after
successful cardiac arrest resuscitation, Shock 22 (2) (2004) 116–119.
[32] J.J. Kim, Y.S. Lim, J.H. Shin, et al., Relative adrenal insufficiency after cardiac arrest:
impact on postresuscitation disease outcome, Am. J. Emerg. Med. 24 (6) (2006)
684–688.
[33] E.D. Hall, Neuroprotective actions of glucocorticoid and nonglucocorticoid steroids
in acute neuronal injury, Cell. Mol. Neurobiol. 13 (4) (1993) 415–432.
[34] M.S. Tsai, C.H. Huang, C.Y. Tsai, et al., Ascorbic acid mitigates the myocardial injury
after cardiac arrest and electrical shock, Intensive Care Med. 37 (12) (2011)
2033–2040.
[35] J.A. Spath Jr., D.L. Lane, A.M. Lefer, Protective action of methylprednisolone on the
myocardium during experimental myocardial ischemia in the cat, Circ. Res. 35 (1)
(1974) 44–51.
[36] J.C. Wynsen, K.C. Preuss, G.J. Gross, H.L. Brooks, D.C. Warltier, Steroid-induced en-
hancement of functional recovery of postischemic, reperfused myocardium in con-
scious dogs, Am. Heart J. 116 (4) (1988) 915–925.
[37] G. Valen, T. Kawakami, P. Tahepold, A. Dumitrescu, C. Lowbeer, J. Vaage, Glucocorti-
coid pretreatment protects cardiac function and induces cardiac heat shock protein
72, Am. J. Physiol. Heart Circ. Physiol. 279 (2) (2000) H836–H843.
[38] M. Buttini, A. Mir, K. Appel, et al., Lipopolysaccharide induces expression of tumour
necrosis factor alpha in rat brain: inhibition by methylprednisolone and by rolipram,
Br. J. Pharmacol. 122 (7) (1997) 1483–1489.
[39] T. van der Poll, S.F. Lowry, Lipopolysaccharide-induced interleukin 8 production by
human whole blood is enhanced by epinephrine and inhibited by hydrocortisone,
Infect. Immun. 65 (6) (1997) 2378–2381.
[40] Z.Q. Lu, J.X. Lu, Y.J. Deng, Glucocorticoids offer protection against myocardial injury
in a murine model of sepsis, Int. J. Clin. Exp. Med. 8 (8) (2015) 12211–12218.
[41] D. Aguilar, J. Strom, Q.M. Chen, Glucocorticoid induced leucine zipper inhibits apo-
ptosis of cardiomyocytes by doxorubicin, Toxicol. Appl. Pharmacol. 276 (1) (2014)
55–62.
[42] H.Y. Oh, S. Namkoong, S.J. Lee, et al., Dexamethasone protects primary cultured he-
patocytes from death receptor-mediated apoptosis by upregulation of cFLIP, Cell
Death Differ. 13 (3) (2006) 512–523.
[43] S. Zimmermann, F.A. Flachskampf, A. Alff, et al., Out-of-hospital cardiac arrest and
percutaneous coronary intervention for ST-elevation myocardial infarction: long-
term survival and neurological outcome, Int. J. Cardiol. 166 (1) (2013) 236–241.