Neurology

Early identified risk factors and their predictive performance of brain death in out-of-hospital cardiac arrest survivors

a b s t r a c t

Background: Early prediction of brain death (BD) after the return of spontaneous circulation (ROSC) in patients with cardiac arrest would be useful for the proper distribution of good quality transplantable organs and Medical resources. We aimed to early identify independent risk factors of BD and their Predictive performance in out-of- hospital cardiac arrest (OHCA) survivors.

Methods: This retrospective observational study included adult OHCA survivors from May 2018 to February 2021. Independent risk factors for progression to BD were identified by performing multivariate logistic regression anal- ysis, including clinical, laboratory, biological parameters and prognostic factors. obtained within 6 h after ROSC. Neuron-specific enolase level were categorized into quartile. The primary outcome was BD occurrence.

Results: Overall, 108 patients were included in this analysis, 31 (29%) of whom had BD. In multivariate logistic re- gression analysis, initial serum NSE levels in the fourth quartile compared to the first quartile (odds ratio [OR], 88.5; 95% confidence interval [CI]: 7.0-1113.6) and absence of pupil light reflex (PLR) (OR, 40.3; 95% CI: 3.8-430.3) were independently associated with BD. According to the receiver operating characteristic curve anal- ysis, initial serum NSE levels and PLR showed good-to-excellent and fair-to-good prognostic performance, respectively (area under the curve [AUC], 0.90; 95% CI: 0.83-0.95 vs. 0.81; 95% CI: 0.72-0.88). Additionally, the combination of both the risk factors (AUC, 0.96; 95% CI: 0.90-0.99) showed significantly higher predictive per- formance for BD than when using them individually (P = 0.04 and P < 0.01, respectively).

Conclusion: High levels of initial serum NSE and PLR obtained within 6 h after ROSC may help early predict progression to BD in OHCA survivors. A large prospective multicenter study should be conducted to confirm these results.

(C) 2022

  1. Introduction

hypoxic-ischemic brain injury (HIBI) is the most frequent complica- tion of cardiac arrest that results in Poor neurologic outcomes and low

Abbreviations: AUC, area under the receiver operating characteristic curve; BBB, blood-brain barrier; BD, brain death; CI, confidence interval; CNUH, Chungnam National University Hospital; CPC, cerebral performance category; CT, computed tomography; GWM, gray-to-white matter ratio; HIBI, hypoxic-ischemic brain injury; ICU, intensive care unit; NSE, neuron-specific enolase; OD, odds ratio; OHCA, out-of-hospital cardiac ar- rest; PLR, pupil light reflex; ROC, receiver operating characteristic; ROSC, return of sponta- neous circulation; TTM, Targeted temperature management; WLST, withdrawal of life- sustaining therapy.

* Corresponding author at: Department of Emergency Medicine, College of Medicine, Chungnam National University, 266 Munwha-ro, Jung-gu, Daejeon 35015, Republic of Korea.

E-mail address: [email protected] (J.S. Park).

1 Bong Kyu Lee and Jin Hong Min contributed equally to this work.

survival rates in Out-of-hospital cardiac arrest survivors [1,2]. Brain death (BD) occurs in 10%-15% of OHCA survivors, and increas- ingly, this is a source of high-quality organ supply [3-6]. In view of the recent coronavirus disease pandemic, the importance of proper dis- tribution of medical resources, including mechanical ventilators, to the intensive care unit (ICU) is increasing. If the potential for BD can be predicted in advance in a patient, then there is an opportunity for earlier brain death Diagnostic procedures, and medical resources can be distributed properly, securing an opportunity for Organ donation. How- ever, hasty withdrawal of life-sustaining therapy (WLST) is problematic in patients with neurological recovery potential. Therefore, early and accurate prediction of BD in OHCA survivors is important.

Recently, a few studies were conducted on the prediction of early BD in OHCA survivors; however, the results have limited clinical applica- tion owing to low accuracy and sensitivity [4-6]. This is because those studies used only clinical, laboratory, and biological parameters and

https://doi.org/10.1016/j.ajem.2022.03.054

0735-6757/(C) 2022

did not include known prognostic factors, such as clinical examination, electrophysiology, biomarker, and neuroimaging findings, in the esti- mation.

In the current guidelines for post-cardiac arrest care, it is recom- mended that most prognostic factors be used 48 h after return of spon- taneous circulation (ROSC), except for neuroimaging findings (such as, brain computed tomography [CT]) [7,8]. However, it can be assumed that significant changes in prognostic factors can occur immediately after ROSC if severe HIBI occurs that can cause BD.

Therefore, we hypothesized that combining the prognostic factors used in the current guidelines for predicting neurologic prognosis could lead to a more accurate prediction of progression toward BD [7,8]. This study aimed to identify independent risk factors of BD and their predictive performance of BD before targeted temperature man- agement (TTM) in OHCA survivors using clinical, laboratory, and biolog- ical parameters together with multimodal prognostic factors.

  1. Methods
    1. Study design and population

This study was a retrospective analysis of prospectively collected data from patients who were alive for 24 h after ROSC (e.g., adult coma- tose OHCA survivors) and were treated with TTM at our hospital be- tween May 2018 and May 2021. This study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the Institutional Review Board of Chungnam National University Hospi- tal (No. CNUH 2021-09-070). The extracted data included clinical data only; it did not include any personally identifiable information. There- fore, the requirement for informed consent was waived.

We excluded patients who met any of the following criteria: (1) inel- igibility for TTM (i.e., Brain hemorrhage, active bleeding, known termi- nal illness, or poor pre-arrest neurological status); (2) patients with Traumatic cardiac arrest; (3) missing data about cardiac arrest charac- teristics and laboratory data; and (4) patients treated with extracorpo- real membrane oxygenation. The enrolled patients received standard post-cardiac arrest care according to the recent international guidelines. TTM was performed in all patients who were unable to obey commands after ROSC, except when they were ineligible for TTM and when their guardian did not provide consent. In this study, none of the patients un- derwent WLST during TTM.

    1. TTM protocol

Comatose OHCA survivors were managed according to our previ- ously published TTM protocol [9]. A target temperature of 33 ?C was maintained for 24 h using an Arctic Sun(R) (Energy Transfer Pads(TM); Medivance Corp, Louisville, CO, USA) feedback-controlled Surface cooling device. Upon completion of the TTM maintenance period, the patients were rewarmed to 37 ?C at a rate of 0.25 ?C/h, and the temper- ature was monitored using a bladder temperature probe. All patients re- ceived sedatives and Neuromuscular blocking agents during the TTM. We also continuously monitored all patients receiving TTM for seizures using bispectral index and/or amplitude-integrated electroencephalog- raphy (aEEG). To predict prognosis, clinical examination, electrophysi- ology (electroencephalography [EEG] and/or aEEG), biomarker (serial serum neuron-specific enolase [NSE]), and imaging (brain magnetic resonance image and CT) findings were obtained for all patients who underwent TTM in accordance with international post-cardiac arrest care guidelines [10].

    1. Data collection

We analyzed the data of OHCA survivors treated with TTM extracted from the hospital’s registry. The extracted clinical information included age, sex, and Charlson Comorbidity Index. Moreover, we collected data

on cardiac arrest characteristics, such as witnessed arrest, bystander cardiopulmonary resuscitation, first monitored rhythm, aetiology of cardiac arrest, time from collapse to cardiopulmonary resuscitation (No-flow time), and time from cardiopulmonary resuscitation to ROSC (low-flow time). We defined the predictive value of some prognostic factors early after ROSC as follows: 1) initial serum NSE levels were de- fined as those measured within 6 h after ROSC. To evaluate the associa- tion of serum NSE levels with BD, the cohort was divided into four categories based on the distribution of the serum NSE level data, using quartiles as cut-off values between categories; 2) the absence of pupil light reflex (PLR) was defined as the absence of bilateral PLR during 4-6 h after ROSC. In our institution, PLR was measured every hour after ROSC and no Neuromuscular blockers were used, but sedatives were occasionally used; 3) The gray-to-white matter ratio (GWR) values were obtained using brain CT within 6 h after ROSC. GWR values were calculated by two certified neuroradiologists who measured Hounsfield units of the putamen, caudate nucleus, posterior limb of the internal capsule, and Hounsfield units of the corpus callosum at the basal ganglia level ([putamen + caudate nucleus]/[posterior limb of the internal capsule + corpus callosum]). The average value was used; 4) the presence of seizures was identified either clinically or by aEEG monitoring within 6 h after ROSC. A seizure can be identified as a rapid rise in both the lower and upper margins of the amplitude trace in the aEEG, or simultaneous Seizure activity indicated by repeated spikes or sharP waves in the raw trace in the aEEG [11]. We assessed neurological outcomes at hospital discharge, which were evaluated ac- cording to the Glasgow-Pittsburgh Cerebral Performance Category scale. We dichotomized the results into good (CPC 1-2) and poor (CPC 3-5) outcomes.

    1. Outcomes

The primary outcome of this study was the occurrence of BD at the time of hospital discharge, as defined by Korean law [12,13]. In South Korea, a brain-dead patient can be legally recognized as dead, but diag- nosis of BD can only be established for the purpose of organ donation [14]. To be diagnosed with BD, the patient’s family must consent to the evaluation process for the purpose of possible organ donation. BD was assessed by a BD committee according to Korean law, using the fol- lowing criteria, all of which are required: complete clinical Neurological examination, including documentation of coma, absence of all brainstem reflexes confirmed with an apnea test, and electrical inactiv- ity by EEG [13]. For patients who did not wish to donate organs, in the absence of confounding factors, two emergency medicine specialists used clinical information and/or EEG and/or brain magnetic resonance imaging (MRI) scans to reach a consensus on whether or not the cause of death was BD. All of the following criteria must be included:

1) Coma when sedatives and NMB are not used. 2) No spontaneous breathing and breathing is maintained with a ventilator. 3) Brainstem reflexes (PLR, corneal reflex, doll’s eye reflex) are absent in medical re- cords. 4) If the expert opinion of EEG or MRI is diagnosed as BD (Fig. 1). In our institution, neurological examination, EEG, biomarkers (NSE), and MRI were performed to predict Neurological prognosis between 72 and 96 h after ROSC for all patients who received TTM. The patients’ clinical information included doll’s eye reflex, fixed pupil light reflex, corneal reflex, spontaneous respiration, etc., and EEG and MRI were comprehensively reviewed by specialists.

    1. Statistical analysis

Categorical variables were analyzed using the ?2 test or Fisher’s exact test and are reported as numbers and percentages. As all continu- ous variables have non-normal distributions, we analyzed them using the Mann-Whitney U test and presented them as medians with inter- quartile ranges. Logistic regression analyses were conducted to identify the independent risk factors for BD. All variables with a P-value of less

Image of Fig. 1

Fig. 1. A 58-year-old man whose cause of death was determined to be brain death. Contrast-enhancED magnetic resonance imaging and electroencephalography (EEG) findings performed 72 h after ROSC in an out-of-hospital cardiac arrest survivor who received target temperature management.

(A) T2 weighted imaging reveal diffuse swelling of both cerebral gyri with hyperintense cortex. (B,C) DWI shows increased signal intensity in both hemisphere and ADC imaging shows severe drop. (D) TOF imaging shows absence of blood flow signal in the cerebral vessels. (E) EEG shows electrocerebral inactivity suggesting brain death.

ADC, apparent diffusion coefficient; DWI, diffuse weighted imaging; TOF, time of flight angiography.

than 0.1 in univariate analyses were included in the multivariate logistic regression model. The backward selection method was used to develop the final adjusted model. The goodness of fit of the final model was eval- uated using the Hosmer-Lemeshow test. We reported the logistic re- gression analysis results as odds ratios (ORs) with 95% confidence intervals (95% CIs). receiver operating characteristic curves were constructed to determine the predictive performance of each in- dependent risk factor and a combination of both the risk factors, and the Delong test was used to compare the area under the ROC curve (AUC). Cut-off values with 99% and 100% specificity, and Youden’s index (sensitivity + specificity – 1) were calculated to predict BD after ROSC. Statistical analyses were performed using IBM SPSS Statistics, ver. 25.0 (IBM Corp., Armonk, NY, USA) and MedCalc, version 15.2.2 (MedCalc Software, Mariakerke, Belgium). Statistical significance was set at P < 0.05.

  1. Results
    1. Patient characteristics

Among the 122 non-traumatic adult OHCA comatose survivors treated with TTM during the study period, nine patients received extra- corporeal membrane oxygenation treatment and five patients died of circulatory failure within 24 h after ROSC. A total of 108 patients were included in this study: 77 (71.3%) patients were non-BD and 31 (28.7%) patients were BD (Fig. 2). Of the enrolled patients, the cases of not surviving discharged are as follows, 4 (3.7%) died of circulatory

failure, 6 (5.6%) of multiple organ failure, 6 (5.6%) of WLST, 13 (12.0%)

for organ donation, and 12 (11.1%) of Cerebral injury. Patients’ baseline demographics, clinical characteristics, and Physiological parameters stratified according to outcome are shown in Table 1. The non-BD group had higher rates of witnessed events, bystander cardiopulmonary resuscitation, shockable rhythm, GWR value, and presence of PLR, shorter no- and low-flow times, and lower levels of initial serum NSE than the BD group. However, there were no differences between groups in age, proportion of male patients, Charlson comorbidity index, pres- ence of seizure, initial serum NSE measurement time, and Brain CT scan time.

    1. Multivariable logistic regression analysis for BD

From the multivariate final prediction model, only the absence of PLR (OR, 40.3; 95% CI: 3.8-430.3) and initial serum NSE levels in the fourth quartile (OR, 88.5; 95% CI: 7.0-1113.6) compared to those in the first quartile were independently associated with progression to- ward BD (Table 2).

    1. Prognostic performance, sensitivity, and specificity of independent fac- tors

Table 3 and Fig. 3 show the prognostic performance of independent risk factors when used individually and as a combination for progres- sion toward BD. According to the ROC analysis, the AUCs of initial serum NSE levels (AUC, 0.90; 95% CI: 0.83-0.95) and PLR (OR, 0.81;

Image of Fig. 2

Fig. 2. Flow diagram of the patient Selection process.

CPC, cerebral performance category; ECMO, extracorporeal membrane oxygenation; OHCA, out-of-hospital cardiac arrest; TTM, target temperature management; WLST, withdrawal of life-sustaining therapy.

95% CI: 0.72-0.88) showed good-to-excellent and fair-to-good prognos- tic performance, respectively. Additionally, the combination of initial serum NSE levels and PLR (AUC, 0.96; 95% CI: 0.90-0.99) showed signif- icantly higher performance in the prediction of BD progression than when using them individually (all P < 0.05). When the cut-off level was defined as high specificity, it showed relatively high sensitivity in the following order: combination of both, initial serum NSE levels, and PLR. The combination and initial serum NSE levels showed a sensitivity of 74% and 61%, respectively, at 99% specificity.

  1. Discussion

In this retrospective study, we showed that the fourth quartile of ini- tial serum NSE levels compared to the first quartile and absence of PLR were independently associated with progression toward BD. In addi- tion, the combination of two independent factors (namely, PLR and ini- tial serum NSE levels) showed better predictive performance of progression toward BD than when using each factor, and this combina- tion showed 74% sensitivity and 99% specificity.

To the best of our knowledge, there are only three previous studies

on the prediction and early identification of risk factors for BD after ROSC in OHCA survivors [4-6]. BD progression rates in their study

ranged between 14.8% and 19.6%, which are lower than that of 28.7% in our study. There are two possible reasons for the higher diagnosis rate of BD in our study. First, WLST was not allowed during TTM in our study. Second, looking at the characteristics of all patients in our study, the GWR measured by brain CT obtained quickly after ROSC was low, and less than half of the patients had PLR within 6 h, so it is possible that the cohort was exceptionally brain injured. The previous three studies attempted to identify risk factors related to BD using clin- ical, laboratory, and biological parameters at the prehospital and hospi- tal stages of patients, but the results varied. The earliest published study used only data available at ICU admission and failed to identify risk fac- tors for BD, including in patients who died within the first 24 h [4]. How- ever, two recently published studies identified five and seven risk factors for progression toward BD. Coure et al. reported five risk factors associated with BD, with neurological causes of cardiac arrest being the strongest risk factor for BD, followed by the need for administration of vasoactive drugs at day 1, duration of low-flow of >16 min, female sex, and age [5]. The following seven independent predictors of BD were found in Madelaine et al.’s validation study and development co- hort: female sex, Non-shockable rhythm, non-cardiac causes of OHCA, neurological causes of OHCA, administration of vasoactive drugs at ICU admission, administration of vasoactive drugs 24 h after ICU admission,

Table 1

Baseline demographics, clinical characteristics, and physiological parameters in the total cohort.

Characteristics

Cohort (n = 108)

No brain death (n = 77)

Brain death (n = 31)

P-value

Age, median (IQR)

55 (41-69)

57 (40-69)

58 (47-69)

0.675

Male, n (%)

78 (72.2)

59 (76.6)

19 (61.3)

0.080

CCI (IQR)

1 (0-2)

1 (0-2)

0 (0-2)

0.824

Cardiac arrest characteristics

Witness, n (%)

72 (66.7)

58 (75.3)

14 (45.2)

0.008

Bystander CPR, n (%)

76 (70.4)

59 (76.6)

17 (54.8)

0.018

Shockable rhythm, n (%)

27 (25.0)

27 (35.1)

0 (0)

0.009

Cardiac aetiology, n (%)

38 (35.2)

32 (41.6)

6 (19.4)

0.124

No flow time, min, median (IQR)

2.0 (0.0-12.8)

1.0 (0.0-9.0)

7.0 (2.0-23.5)

0.001

Low flow time, min, median (IQR)

19.0 (10.0-30.0)

15.0 (9.0-24.0)

36.0 (23.5-48.5)

<0.001

Time taken to initial serum NSE sample from ROSC, hour, median (IQR)

4.8 (3.6-5.6)

4.8 (3.4-5.7)

5.0 (4.3-5.2)

0.314

Initial serum NSE, ng/mL, median (IQR)

31.5 (20.7-54.8)

26.0 (18.8-33.8)

81.1 (51.9-270.0)

<0.001

Time taken to Brain CT from ROSC, min, median (IQR)

79.0 (39.0-132.0)

79.0 (45.0-169.0)

99.0 (43.0-132.0)

0.568

GWR, median (IQR)

1.21 (1.11-1.29)

1.25 (1.17-1.35)

1.11 (1.06-1.32)

0.006

Seizure, n (%)

26 (24.1)

22 (28.5)

4 (12.9)

0.058

presence of PLR, n (%)

51 (47.2)

50 (64.9)

1 (3.2)

<0.001

CCI, Charlson comorbidity index; CPR, cardiopulmonary resuscitation; CT, computed tomography; GWR, gray-white matter ratio; IQR, interquartile range; NSE, neuron-specific enolase; PLR, pupil light reflex; ROSC, return of spontaneous circulation.

Table 2

Multivariable logistic regression model with brain death as the dependent variable.

Variable

Adjusted OR

95% CI

P value

PLR

Present

1.0 (reference)

Absent

40.3

3.8-430.3

0.002

Quartile of initial serum NSE

First quartile (< 20.8 mg/dl)

1.0 (reference)

Second quartile (20.8-31.6 mg/dl)

2.2

0.2-24.4

0.545

Third quartile (31.6-56.8 mg/dl)

4.4

0.4-46.5

0.217

Fourth quartile (>= 56.8 mg/dl)

88.5

7.0-1113.6

0.001

Adjusted ORs were the result of the multivariable logistic regression analysis by using the stepwise backward selection method. (Hosmer-Lemeshow test, P = 0.624).

The odds ratio was adjusted with witness arrest, bystander CPR, shockable rhythm, cardiac aetiology, no flow time, low flow time, GWR, and seizure.

CI, confidence interval; GWR, gray-white matter ratio; NSE, neuron-specific enolase; OR,

odds ratio; PLR, pupil light reflex.

and natremia 24 h after admission [6]. However, both the studies have limitations when applying these findings to clinical practice. First, the predictive performance was 0.84 (95% CI: 0.78-0.89) and 0.82 (95% CI: 0.77-0.86), showing fair-to-good performance. Second, in the BD prediction study using the scoring system of Madelaine et al., when the specificity was 98.4%, the sensitivity was reported to be ex- tremely low at 26.2%. In both the studies, the researchers reported that a limitation of the study was that prognostic factors, such as elec- troencephalography findings, biomarkers, neuroimaging, and neurolog- ical examination, were not included [5,6]. However, in our study, multivariate analysis was performed by including prognostic factors along with clinical and biological parameters used in the previous stud- ies, and the results showed that only PLR, a neuroexamination finding, and initial serum NSE levels, a biomarker, were independently associ- ated with prediction of progression toward BD. Although clinical, laboratory, and biological parameters were highly correlated with pro- gression toward BD in previous studies, they were not effective predic- tive factors in the present study because they were less relevant than those known as Prognostic tools. In addition, the combination of these two independent factors showed an excellent predictive performance of 0.96 (95% CI: 0.90-0.99) and showed a high sensitivity of 74% at a specificity of 99% compared to those reported in previous studies.

NSE is the only blood biomarker recommended by the guidelines, and although quantifiable and objective, it has the disadvantage of the presence of extra-neuronal sources, such as hemolysis and neuroendo- crine tumors [8,10,15,16]. NSE is confined to neurons under normal con- ditions and is present only in negligible amounts in peripheral blood. However, when the blood-brain barrier (BBB) is disrupted after HIBI, NSE leaks from the cerebrospinal fluid into the systemic circulation [15]. In a previous study, severe BBB disruption occurred within 24 h after ROSC in the poor outcome group, and in another study [17], eight out of nine patients who experienced severe BBB disruption im- mediately after ROSC had a poor prognosis [18]. Based on these findings, it is suggested that the severity of HIBI is related to the degree of BBB disruption and time. The international guidelines recommend that NSE levels measured 48-72 h after ROSC can be used as a tool to predict

Image of Fig. 3

Fig. 3. Comparison of the area under the receiver operating characteristic curve for predicting brain death, for individual risk factors (pupil light reflex and initial serum neu- ron-specific enolase levels) and a combination of both.

AUROC, area under the receiver operating characteristic curve; CI, confidence interval; NSE, neuron-specific enolase; PLR, pupil light reflex.

neurological prognosis [8]. However, if the level is much higher than the normal immediately after ROSC, then it can be assumed that there was severe HIBI; therefore, it is estimated that the initial serum NSE levels can be used as a BD prediction factor. In the present study, when the specificity was 99%, the BD prediction sensitivity of the initial serum NSE level within 6 h after ROSC was superior to that of the TTM trial in- volving the largest cohort study for prognosis prediction using NSE levels measured at 48 and 72 h (61% vs. 47%, 61% vs. 54%) [19]. More- over, in the multivariable logistic regression analysis, the serum NSE levels meaningful for BD progression were related only to the high levels in the fourth quartile (OR, 88.5; 95% CI: 7.0-1113.6) compared to the levels in the first quartile.

PLR is associated with Cranial nerves II and III (midbrain) and has been used as an indirect marker of brainstem blood flow and may reflect hypoxic-ischemic damage to neurons in the brainstem [20-23]. From a prognostication standpoint, the PLR has traditionally and consistently been a valuable marker. In addition, because PLR varies with time de- pending on the degree of HIBI, the measurement time is important. The absence of bilateral PLR in patients remaining comatose on day 3 after cardiac arrest was considered highly specific for poor neurological outcomes [8]. However, in a prospective multicenter study of 456 pa- tients from day 1 to 3 after ROSC, the proportion of poor neurological outcome among the bilateral PLR absent groups was not high at 57%- 62%, but it gradually increased [22]. In addition, in a retrospective study analyzing 10,151 OHCA survivors, the sensitivity and specificity

Table 3

PLR; initial serum NSE levels; and combination specificity, cut-off values, and sensitivity.

Independent factors

Specificity (95% CIs)

Cut-off values

Sensitivity (95% CIs)

PPV (95% CIs)

NPV (95% CIs)

PLR

100 (95-100)

Presence

0 (0-11)

71 (71-71)

Youden index

65 (53-76)

Absence

97 (83-100)

53 (45-60)

98 (88-100)

Initial serum NSE

100 (95-100)

99 (93-100)

246 mg/dl

68.7 mg/dl

29 (14-48)

61 (42-78)

100

95 (73-99)

78 (74-81)

86 (80-91)

Youden index

94 (86-98)

49.3 mg/dl

81 (63-93)

83 (68-92)

92 (85-96)

Combination

100 (95-100)

29 (14-48)

100

78 (74-81)

99 (93-100)

– 74 (55-88)

96 (76-99)

91 (84-95)

Youden index

88 (79-95)

– 90 (74-98)

76 (63-85)

96 (89-99)

PLR, pupil light reflex; NSE, neuron-specific enolase; CI, confidence interval; PPV, positive predict value; NPV, negative predict value.

of the absence of PLR for poor outcomes were 72.2% and 68.8%, respec- tively, indicating that it did not have enough accuracy to determine the prognosis [23]. However, absence of PLR was associated with poor neu- rological outcome in multivariate analysis (OR, 3.1; 95% CI: 2.7-3.5). In our study, the determination of the presence or absence of PLR was performed between 4 and 6 h after ROSC, an earlier time than that in previous studies; therefore, absence of PLR and severe HIBI can be inferred, which may explain their association with BD progression in multivariate analysis.

The current guidelines recommend the implementation of cooling as early as possible in OHCA survivors [24]. If it is possible to predict BD at an earlier time before TTM, then it can facilitate appropriate distribution of medical resources and reduce the burden on caregivers and families, thereby securing an opportunity for organ donation.

Our study had several limitations. First, this was a single-center study with a small sample size, leading to limitations in the generalizability of the results. However, compared to other similar investigations, no stud- ies have analyzed prognostic factors. Second, this study’s retrospective nature might have distorted the results because of selection bias or miss- ing data. Third, a validation study was not conducted. To increase the ac- curacy of the results, a future validation study is needed. Fourth, the clinical application of the serum NSE levels suggested in this study is lim- ited. The median time to taking off samples for NSE is significantly longer compared to the time until brain CT examination, which may be a limita- tion in making TTM decisions, and analysis of serum NSE levels is not generally available in all institutions. Fifth, in South Korea, BD is decided by the BD Committee only when organ donation is necessary. Therefore, in this study, in patients who did not wish to donate organs, two emer- gency medicine specialists analyzed their clinical information to reach a consensus on whether or not the cause of death was BD. This is likely to cause detection bias in the results. Sixth, PLR was assessed using stan- dard clinical pupil assessment as either present or absent. Standard clin- ical pupil assessment is influenced by pupil size and medication, such as opioids and sedatives. In addition, several recent studies have reported that quantitative PLR through automated quantitative pupillometry shows better results than qualitative standardized pupil assessment [25-27]. A large prospective multicenter study should be conducted to confirm these results. This study is not definitive with regard to the early prediction of BD, but based on this prediction scheme, we hope to extend it to take into account new predictors such as burst suppression pattern of aEEG, electrocerebral inactivity of EEG, and magnetic reso- nance image findings taken before TTM, etc.

  1. Conclusions

In this study, high levels of initial serum NSE and absence of PLR were independently associated with progression toward BD, and the combination of the two risk factors had a higher predictive perfor- mance. If the physician can predict BD earlier after ROSC, then it will be helpful in determining the treatment direction to secure the oppor- tunity for the distribution of medical resources and increasing the feasi- bility for organ donation.

Ethics approval and consent to participate

The study was conducted according to the guidelines of the Declara- tion of Helsinki and approved by the Institutional Review Board (or Ethics Committee) of Chungnam National University Hospital (No. CNUH 2021-09-070). The extracted data included clinical data only; it does not include any personally identifiable information. Therefore, the need for informed consent was waived.

Availability of data and materials

The data presented here are available on request from the correspond- ing author. The data are not publicly available due to Ethical concerns.

Funding

This research did not receive any specific grant from funding agen- cies in the public, commercial, or not-for-profit sectors.

CRediT authorship contribution statement

Bong Kyu Lee: Writing – original draft, Formal analysis. Jin Hong Min: Data curation, Writing – original draft. Jung Soo Park: Writing – original draft, Investigation, Funding acquisition. Changshin Kang: Data curation, Formal analysis, Writing – review & editing. Byung Kook Lee: Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

None.

References

  1. Laver S, Farrow C, Turner D, Nolan JJ. Mode of death after admission to an intensive care unit following cardiac arrest. 2004;30(11):2126-8.
  2. Nolan JP, Neumar RW, Adrie C, et al. Post-cardiac arrest syndrome: epidemiology, pathophysiology, treatment, and prognostication: a scientific statement from the In- ternational liaison Committee on Resuscitation; the American Heart Association Emergency cardiovascular 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. 2008;79(3):350-79.
  3. Sandroni C, D’Arrigo S, Callaway CW, et al. The rate of brain death and organ dona- tion in patients resuscitated from cardiac arrest: a systematic review and meta- analysis. 2016;42(11):1661-71.
  4. Adrie C, Haouache H, Saleh M, et al. An underrecognized source of organ donors: pa- tients with brain death after successfully resuscitated cardiac arrest. 2008;34(1): 132-7.
  5. Cour M, Turc J, Madelaine T, Aoic Argaud LJ. Risk factors for progression toward brain death after out-of-hospital cardiac arrest. 2019;9(1):1-6.
  6. Madelaine T, Cour M, Roy P, et al. Prediction of Brain Death After Out-of-Hospital Cardiac Arrest: Development and Validation of the Brain Death After Cardiac Arrest Score; 2021.
  7. Rossetti AO, Rabinstein AA, Oddo MJTLN. Neurological prognostication of outcome in patients in coma after cardiac arrest. 2016;15(6):597-609.
  8. Kim YM, Jeung KW, Kim WY, et al. 2020 Korean guidelines for cardiopulmonary re- suscitation. Part 5. Post-cardiac arrest care. Clin Exp Emerg Med. 2021;8(S):S41-64.
  9. Kang C, Jeong W, Park JS, et al. Comparison of prognostic performance between neuron-specific enolase and S100 calcium-binding protein B obtained from the ce- rebrospinal fluid of out-of-hospital Cardiac arrest survivors who underwent targeted temperature management. J Clin Med. 2021;10(7):1531.
  10. Nolan JP, Soar J, Cariou A, et al. European resuscitation council and European Society of Intensive Care Medicine Guidelines for post-resuscitation care 2015: section 5 of the European resuscitation council guidelines for resuscitation 2015. Resuscitation. 2015;2015(95):202-22.
  11. Tao JD, Mathur AM. Using amplitude-integrated EEG in neonatal intensive care. J Perinatol. 2010 Oct;30(Suppl):S73-81.
  12. Korean Organ Donation Agency. KODA Donation Flow. Accessed 10 January 2020, at. https://www.koda1458.kr/organ/judgment.do.
  13. Kim DH, Kwon OY, Yang TW, et al. Reflex and spontaneous movements in adult pa- tients during the process of determining brain death in Korea. J Korean Med Sci. 2020;35:e71.
  14. Park J, Kim CJ. Recent decrease in organ donation from brain-dead potential organ donors in Korea and possible causes. J Korean Med Sci. 2020;35:e94.
  15. Geocadin RG, Callaway CW, Fink EL, et al. American Heart Association emergency cardiovascular care committee. Standards for studies of neurological prognostication in comatose survivors of cardiac arrest: a scientific statement from the American Heart Association. Circulation. 2019;140:e517-42.
  16. Moseby-Knappe M, Mattsson-Carlgren N, Stammet P, et al. serum markers of brain injury can predict good neurological outcome after out-of-hospital cardiac arrest. In- tensive Care Med. 2021 Sep;47(9):984-94.
  17. Park JS, You Y, Min JH, et al. Study on the timing of severe blood-brain barrier disrup- tion using cerebrospinal fluid-serum albumin quotient in post cardiac arrest patients treated with targeted temperature management. 2019;135:118-23.
  18. Jeon GR, Ahn HJ, Park JS, et al. The association between neurological prognosis and the degree of blood-brain barrier disruption in cardiac arrest survivors who under- went target temperature management. Neurocrit Care. 2022;202(35):815-24.
  19. Stammet P, Collignon O, Hassager C, et al. TTM-trial investigators. Neuron-specific enolase as a predictor of death or poor neurological outcome after out-of-hospital

cardiac arrest and targeted temperature management at 33?C and 36?C. J Am Coll Cardiol. 2015 May 19;65(19):2104-14.

  1. Zafar SF, JIJ Suarez. Automated pupillometer for monitoring the critically ill patient: a

critical appraisal. 2014;29(4):599-603.

  1. Olson DM, MJC Fishel. The use of automated pupillometry in critical care. 2016;28 (1):101-7.
  2. Oddo M, Sandroni C, Citerio G, et al. Quantitative versus standard pupillary light re- flex for early prognostication in comatose cardiac arrest patients: an international prospective multicenter double-blinded study. 2018;44(12):2102-11.
  3. Javaudin F, Leclere B, Segard J, et al. Prognostic performance of early absence of pu- pillary light reaction after recovery of out of hospital cardiac arrest. Resuscitation. 2018;127:8-13.
  4. Panchal AR, Bartos JA, Cabanas JG, et al. Part 3: adult basic and advanced life support: 2020 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. 2020;142(16_Suppl_2):S366-468.
  5. Chen JW, Gombart ZJ, Rogers S, Gardiner SK, Cecil S, RMJS Bullock. Pupillary reactiv- ity as an early indicator of increased intracranial pressure: the introduction of the Neurological Pupil index. 2011;2:82.
  6. Chen JW, Vakil-Gilani K, Williamson KL, Cecil SJS. Infrared pupillometry, the Neuro- logical Pupil index and unilateral pupillary dilation after traumatic brain injury: im- plications for treatment paradigms. 2014;3(1):1-10.
  7. Rollins MD, Feiner JR, Lee JM, Shah S, Larson MJA. Pupillary effects of high-dose opi- oid quantified with infrared pupillometry. 2014;121(5):1037-44.