The effect of carbon dioxide on near-death experiences in out-of-hospital cardiac arrest survivors: a prospective observational study

Introduction Near-death experiences (NDEs) are reported by 11-23% of cardiac arrest survivors. Several theories concerning the mechanisms of NDEs exist - including physical, psychological, and transcendental reasons - but so far none of these has satisfactorily explained this phenomenon. In this study, we investigated the effect of partial pressures of O2 and CO2, and serum levels of Na and K on the occurrence of NDEs in out-of-hospital cardiac arrest survivors. Methods A prospective observational study was conducted in the three largest hospitals in Slovenia. Fifty-two consecutive patients (median age 53.1 years, 42 males) after out-of-hospital cardiac arrest were included. The presence of NDEs was assessed with a self-administered Greyson's NDE scale. The initial partial pressure of end-tidal CO2, the arterial blood partial pressures of O2 and CO2 and the levels of Na and K in venous blood were analysed and studied. Univariate analyses and multiple regression models were used. Results NDEs were reported by 11 (21.2%) of the patients. Patients with higher initial partial pressures of end-tidal CO2 had significantly more NDEs (P < 0.01). Patients with higher arterial blood partial pressures of CO2 had significantly more NDEs (P = 0.041). Scores on a NDE scale were positively correlated with partial pressures of CO2 (P = 0.017) and with serum levels of potassium (P = 0.026). The logistic regression model for the presence of NDEs (P = 0.002) explained 46% of the variance and revealed higher partial pressures of CO2 to be an independent predictor of NDEs. The linear regression model for a higher score on the NDE scale (P = 0.001) explained 34% of the variance and revealed higher partial pressures of CO2, higher serum levels of K, and previous NDEs as independent predictors of the NDE score. Conclusions Higher concentrations of CO2 proved significant, and higher serum levels of K might be important in the provoking of NDEs. Since these associations have not been reported before, our study adds novel information to the field of NDEs phenomena.


Introduction
Near-death experiences (NDEs) are an unexplained but quite common experience in many cardiac arrest patients after successful resuscitation [1]. One definition describes NDEs as deep psychological experiences with feelings of transcendence or mystical encounter that typically occur in persons close to death or in situations of intense physical or emotional danger [2]. These elements may include cognitive components such as accelerated thought processes and a 'life review', affective components such as peacefulness and joy, or transcendental components such as apparent encounters with mystical entities or deceased persons [2].
Few prospective studies reported an incidence of NDEs of 11 to 23% in cardiac arrest survivors [3,[25][26][27]. Younger patients seem to experience NDEs more often [18,25,28]. Also, a higher serum partial pressure of oxygen (pO 2 ) has been shown to be associated with the occurrence of NDEs [3]. Other factors that might be important are the cardiac aetiology of cardiac arrest [27], previous near-death or paranormal experiences [27], out-of-hospital cardiac arrest [25], female sex [25], and fear of death [25].
The aim of this study was to investigate the effect of serum pO 2 , serum partial pressure of carbon dioxide (pCO 2 ), and partial pressure of end-tidal carbon dioxide (petCO 2 ) on the occurrence of NDEs in out-of-hospital cardiac arrest survivors. In addition, we also investigated the effect of serum levels of sodium and potassium on the occurrence of NDEs.

Study population
We studied out-of-hospital cardiac arrest survivors who were successfully resuscitated in out-of-hospital settings and consecutively admitted to intensive care units from the beginning of January 2008 to the end of June 2009. The inclusion criteria were: 18 years old or older, presence of the cardiac aetiology of cardiac arrest (as confirmed during the resuscitation and later hospital work up), clinical death (defined as a cessation of breathing and effective cardiac output -electrocardiogram patterns of ventricular fibrillation, pulseless ventricular tachycardia, pulseless electrical activity, and asystolia detected by pre-hospital resuscitation teams), a post-resuscitation cerebral performance categories scale score of 1 [29], and the patients' informed consent.
The National Medical Ethics Committee approved the study -No. 79/10/07.

Settings
We conducted a multicentre study in the intensive care units of three of the largest hospitals in Slovenia: the Clinical Centre of Ljubljana, the Clinical Centre of Maribor, and the General Hospital of Celje. The majority of cardiac arrest survivors in Slovenia are transferred to these three hospitals. At the same time, each of these hospitals is closely connected to several regional outpatient emergency medical centres.
Regional out-patient emergency medical centres are part of primary care out-patient healthcare centres.
Teams of two medically trained paramedics and one emergency physician provide urgent medical care for the population of their catchment areas. Critically ill patients are transferred to the nearest regional hospital.

Data collection
Eligible patients were approached during their hospital stay by a member of the research team, who explained the purpose of the study, assured their complete anonymity, and obtained their informed consent ( Figure 1). No patients refused the interview. Then they filled in a self-administered questionnaire about the NDEs [see Additional file 1] [20], which consists of 16 questions about the cognitive, affective, paranormal, and transcendental component of NDEs. The questions could be answered on a three-point scale (from 0 to 2), with a minimum score of 0 and a maximum of 32. The total number of scores of 7 or above defines the existence of a NDE. The questionnaire was translated from English to Slovene using the guidelines recommended by Guillemin and co-workers [30]. Other data obtained with the interview with the patients were: sex, age, level of education, religious belief, previous NDEs, and fear of death before and after the cardiac arrest (Table 1). Data obtained from patients' files were: time until the beginning of resuscitation, time until return of spontaneous circulation (ROSC), drugs received during resuscitation, the initial petCO 2 (in kPa), pO 2 and pCO 2 (both in kPa) in peripheral arterial blood, and serum levels of sodium and potassium (in mmol/l) in peripheral venous blood. Only the blood sample analysis that was performed on the samples taken in the first five minutes upon the admission of the patients to the hospital was taken into account.

Statistical analysis
To analyse the data, we used the statistical package for the social sciences, version 13.0 (SPSS Inc, Chicago, IL, USA). The limit of statistical significance was set at P < 0.05. Descriptive statistics were computed. For the questionnaire, we calculated the reliability coefficient, Cronbach α. Patients with a NDE score of 7 or above were assigned to the NDE group, others were assigned to the non-NDE group [20]. To identify statistically significant differences between different variables, we used an independent samples t-test, chi-squared test, and a Wilcoxon rank sum test. Linear correlation analysis was performed to reveal possible correlations. To identify a possible model for the explanation of differences, linear and binary logistic regressions were performed. The variables that showed statistically significant differences in univariate analysis were entered into multivariate analysis.

Resuscitation attempted n = 426
Return of spontaneous circulation n = 178

Discharged alive n = 76
Patients with exclusion criteria n = 24 Patients, included in the research n = 52 Figure 1 The flowchart of patients' recruitment. The flowchart starts with the number of out-of-hospital cardiac arrest patients, in whom resuscitation was attempted, followed by the number of patients with return of spontaneous circulation, then the number of patients discharged from the hospital alive, and finally the number of patients that were included in the study.

Descriptive data
The study included 52 patients ( Figure 1). NDEs were reported by 11 (21.2%) of them ( Table 1). The mean (standard deviation) NDE score of all patients was 3.2 ± 5.0 points. The average NDE score of patients in the NDE group was 11.5 ± 4.4, and of the non-NDE group was 0.9 ± 1.6. The Cronbach α of the questionnaire was 0.  (Table 2).

Multivariate analysis
Higher pCO 2 was an independent predictor of NDEs. The logistic regression model explained 46% of the variation (Table 3). A higher NDE score was independently associated with higher pCO 2 , higher serum levels of potassium, and previous NDEs. The linear regression model explained 34% of the variation (Table 4).

Discussion
Our prospective study reports a 21.2% incidence of NDEs in out-of-hospital cardiac arrest survivors. It also suggests that the occurrence of NDEs is connected to higher initial petCO 2 , higher arterial blood pCO 2 , and previous NDEs. Higher serum levels of potassium might also play a role. To our knowledge, this is the first prospective study to report a possible correlation between NDEs and CO 2 . It is still not clear whether NDEs occur before, during, or after the period of cardiac arrest [3]. During cardiac arrest, the petCO 2 falls to very low levels, reflecting the very low cardiac output achieved with cardiopulmonary resuscitation [31]. Higher levels of petCO 2 therefore indicate better cardiac output and higher coronary perfusion pressure [32]. Our findings concerning the association between initial petCO 2 and the occurrence of NDEs therefore support the hypothesis that NDEs occur during the cardiac arrest.
On the other hand, the association between higher pCO 2 upon admission and the occurrence of NDEs might suggest that NDEs occurs after the cardiac arrest. But higher pCO 2 upon admission might simply reflect higher initial petCO 2 . Nevertheless, it is known that CO 2 changes the acid-base equilibrium in the brain, which can provoke unusual experiences in the form of bright light, visions, and out-of-body or even mystical experiences [3,5]. Some earlier studies have shown that inhaled CO 2 , used as a psychotherapeutic agent, could cause NDE-like experiences [33,34]. Therefore, we can conclude that CO 2 might be one of the major factors for provoking NDEs, regardless of when NDEs occur. As far as we know, serum levels of potassium were assessed only in one study [3]. The mean level of potassium in the NDE group was slightly lower in comparison to the control group, but no significant differences were found. As our study managed to associate serum levels of potassium only with the higher NDE score, and not also with the higher incidence of NDEs, no firm conclusions can be drawn at this point. Also, the possible mechanism of the effect of potassium in the NDEs has not yet been established. Alternative theories found the explanation for NDEs in quantum theory, which suggests that consciousness may arise from quantum processes within neuronal microtubules [35]. The recent work of Bernroider and Roy suggests that quantum entanglement in the ion channels (especially in the potassium channel) of brain cells underlies information processing in the brain and, ultimately, also consciousness [36]. Although untenable and purely theoretic, this possible connection between potassium channels in the brain and the mechanism of consciousness (and therefore the possible mechanism of NDEs) deserves further investigation. Available data on the role of oxygen in provoking NDEs is ambiguous. Although one physiological theory [5] suggests that anoxia (or hypoxia) might be the cause for NDEs, Parnia and colleagues [3] found a higher mean pO 2 in peripheral blood; however, due to an insufficient sample quantity, a univariate analysis was not performed. In our study, the NDE group had a lower mean pO 2 than the non-NDE group, but this difference was not statistically significant (Table 2). Nevertheless, this finding is in favour of the theory of anoxia [5] and supported by several studies that reported NDE-like experiences in decreased cerebral perfusion (resulting in local cerebral hypoxia) in rapid acceleration during training of fighter pilots [37], in hyperventilation followed by the valsalva maneuver [38], and in people exposed to high altitudes [6]. The proposed mechanism is the induction of hyperactivity of N-methyl D-aspartate (NMDA) receptors by hypoxia, which induces hallucination and might induce NDEs [10].
Previous prospective studies on NDEs reported an 11 to 23% incidence between cardiac arrest survivors [3,[25][26][27], which is consistent with the incidence found in our study. We have not demonstrated the connection between younger age and a higher incidence of NDEs. In fact, the mean age of the NDE group was lower than the non-NDE group, but this difference was not statistically significant. Previous studies have shown that NDEs more often occur in patients younger than 60 years of age [3,27,28]. The age difference in our study might be overlooked due to an insufficient number of subjects. Figure 2 Differences in pCO 2 among near-death experience and non-near-death experience groups. The graph presents the statistically significant differences in initial partial pressure of end-tidal carbon dioxide (petCO 2 ) and partial pressure of carbon dioxide (pCO 2 ) in arterial blood upon admission to hospital (assessed in the first five minutes upon admission). A, near-death experience group; B, non-near-death experiences group.  It is also true that almost 70% of patients in our sample were younger than 60 years. The mean age of patients in our sample was lower (for almost 10 years) than in the two largest prospective studies of NDEs in cardiac arrest survivors [25,27]. This difference might also be the reason why we were not able to demonstrate any age differences in the occurrence of NDEs.
Our study confirmed the findings of other studies on NDEs that sex [25,27], level of education [25,28], fear of death [25], time until ROSC [25,28], medication during resuscitation [25,28], serum level of sodium [3], and religious belief [25] are not associated with NDE occurance. It also confirmed previously reported findings [25] that patients with previous NDEs are more likely to have repeated NDEs in case of a new cardiac arrest episode.
The questionnaire proved to be a reliable instrument for assessing NDEs also in Slovenian. The Cronbach's α of the questionnaire in the original study was 0.88 [20] and our result (0.875) was almost the same.
Our study suggests that some physiological factors or processes might be important in provoking NDEs. On the other hand, the experiences induced by neurophysiological processes mostly consist of fragmented and random memories and confused experiences unlike the real NDEs that are clear, highly structured and easily recalled [3,25]. It is not thought possible to explain NDEs only in terms of physiological processes. Most likely multiple physiological factors are involved [5]. Clearly, the presence of NDEs pushes the current knowledge of human consciousness and mind-brain relation to the edge of our understanding.
The main strength of our study is its prospective design. With a consecutive recruitment of the patients and the inclusion of three of the largest Slovenian hospitals, the selection bias was reduced as much as possible. The use of a standardised scale for NDEs ensured the consistency of NDEs reports. The number of patients in the sample is the main weakness of our study. Therefore, some important differences might have been overlooked and the results should be interpreted with care. Also, receiver-operator characteristic curves for defining a threshold CO 2 were not produced due to too small a number of patients. The weakness is also the fact that almost 70% of the patients in a sample were younger than 60 years old, which could affect the incidence and the demonstration of age differences in NDEs.
Further multicentre studies should investigate the effect of CO 2 and potassium on the incidence of NDEs in a larger prospective sample of cardiac arrest patients or unconscious patients. The clinical reliability and relevance of our findings should be extensively studied.

Conclusions
As much as one-fifth of out-of-hospital cardiac arrest patients report NDEs during cardiac arrest. Higher initial petCO 2 and higher arterial blood pCO 2 proved to be important in the provoking of NDEs. Higher serum levels of potassium might also be important. As these associations have not been reported before, our study adds new and important information to the field of NDE phenomena. As quality of life of NDE patients might be affected, NDEs warrant further study. Likewise, more rigorous measures to establish good acid-base equilibrium should be adopted in resuscitation guidelines.

Key messages
• The incidence of NDEs in out-of-hospital cardiac arrest survivors is 21.2%.
• NDE occur more often in patients with higher petCO 2 and pCO 2 .
• Higher serum levels of potassium correlate with higher score on Greyson's NDE scale.
• NDEs occur more often in patients with previous NDEs.
Additional file 1: The near-death experience scale.
Abbreviations NDE: near-death experience; ROSC: return of spontaneous circulation; petCO 2 : partial pressure of end-tidal carbon dioxide; pCO 2 : partial pressure of carbon dioxide; pO 2 : partial pressure of oxygen; NMDA receptors: N-methyl D-aspartate receptors.