Clinical Investigations

Vital Signs After Cardiac Arrest Following Withdrawal of Life-Sustaining Therapy: A Multicenter Prospective Observational Study* Sonny Dhanani, MD1,2,3; Laura Hornby, MSc4,5; Roxanne Ward, BScN, MSc1,2; Andrew Baker, MD6,7; Peter Dodek, MD8,9; Jane Chamber-Evans, BScN, MSc4,10,11; Rob Fowler, MDCM7,12; Jan O. Friedrich, MD6,7; Robert M. Gow, MBBS2,3,13; Demetrios J. Kutsogiannis, MD14,15; Lauralyn Mcintyre, MD16,17,18,19; Franco Momoli, PhD18,19,20; Karine Morin, LLM21; Tim Ramsay, PhD18,19; Damon Scales, MD7,12; Hilary Writer, MD1,2,3; Serafettin Yildirim, BMgmt22; Bryan Young, MD23,24; Sam Shemie, MD4,25,26; on behalf of the Canadian Critical Care Trials Group and in collaboration with the Bertram Loeb Chair and Research Consortium in Organ and Tissue Donation

*See also p. 2447. 1 Division of Pediatric Critical Care, Children’s Hospital of Eastern Ontario, Ottawa, ON, Canada. 2 Children’s Hospital of Eastern Ontario Research Institute, Ottawa, ON, Canada. 3 Faculty of Medicine, University of Ottawa, Ottawa, ON, Canada. 4 Montreal Children’s Hospital, McGill University Health Centre, Montreal, QC, Canada. 5 McGill University Health Centre Research Institute, Montreal, QC, Canada. 6 Department of Critical Care, The Keenan Research Centre in the Li Ka Shing Knowledge Institute, St Michael’s Hospital, Toronto, ON, Canada. 7 Faculty of Medicine, Interdepartmental Division of Critical Care, University of Toronto, ON, Canada. 8 Critical Care Medicine and Center for Health Evaluation and Outcome Sciences, St. Paul’s Hospital, Vancouver, BC, Canada. 9 Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada. 10 Ingram School of Nursing McGill University, McGill University Health Centre, Montreal, QC, Canada. 11 Faculty of Medicine Biomedical Ethics Unit, McGill University, Montreal, QC, Canada. 12 Department of Critical Care Medicine and Department of Medicine, Sunnybrook Health Sciences Centre, Toronto, ON, Canada. 13 Pediatric Cardiology, Children’s Hospital of Eastern Ontario, Ottawa, ON, Canada. 14 Neurosciences Intensive Care Unit, University of Alberta Hospital, Edmonton, AB, Canada. 15 Faculty of Medicine and Dentistry, University of Alberta, Edmonton, AB, Canada. 16 Critical Care, The Ottawa Hospital, Ottawa, ON, Canada. 17 Department of Medicine, University of Ottawa, Ottawa, ON, Canada. 18 The Ottawa Health Research Institute, Ottawa, ON, Canada. 19 The Clinical Epidemiology Program Methods Centre, University of Ottawa, Ottawa, ON, Canada. Copyright © 2014 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0000000000000417

2358

www.ccmjournal.org

Department of Epidemiology and Community Medicine, University of Ottawa, Ottawa, ON, Canada.

20

National GE3LS Program, Genome Canada, Ottawa, ON, Canada.

21

Information Systems, Children’s Hospital of Eastern Ontario, Ottawa, ON, Canada.

22

Department of Clinical Neurological Sciences, London Health Sciences Centre, London, ON, Canada.

23

Faculty of Medicine, University of Western Ontario, London, ON, Canada.

24

Pediatric Critical Care, Montreal Children’s Hospital, Montreal, QC, Canada.

25

Faculty of Pediatrics, McGill University, Montreal, QC, Canada.

26

Dr. Dhanani, Ms. Hornby, and Ms. Ward contributed to the conception or design of the work; the acquisition, analysis, or interpretation of data for the work; drafting the work and revising it critically for important intellectual content; and final approval of the version to be published. Dr. Baker contributed to the conception or design of the work; the acquisition, analysis, or interpretation of data for the work; revising it critically for important intellectual content; and final approval of the version to be published. Dr. Dodek contributed to the acquisition, analysis, or interpretation of data for the work; revising it critically for important intellectual content; and final approval of the version to be published. Ms. Chamber-Evans contributed to the conception or design of the work; revising it critically for important intellectual content; and final approval of the version to be published. Dr. Fowler, Dr. Friedrich, and Dr. Gow contributed to the acquisition, analysis, or interpretation of data for the work; revising it critically for important intellectual content; and final approval of the version to be published. Drs. Kutsogiannis and Mcintyre contributed to the conception or design of the work; the acquisition, analysis, or interpretation of data for the work; revising it critically for important intellectual content; and final approval of the version to be published. Dr. Momoli and Ms. Morin contributed to the conception or design of the work; revising it critically for important intellectual content; and final approval of the version to be published. Dr. Ramsay contributed to the conception or design of the work; the acquisition, analysis, or interpretation of data for the work; drafting the work and revising it critically for important intellectual content; and final approval of the version to be published. Drs. Scales and Writer contributed to the acquisition, analysis, or interpretation of data for the work; revising it critically for important intellectual content; and final approval of the version to be published. Mr. Yildirim and Dr. Young contributed to the conception or design of the work; the acquisition, analysis, or interpretation of data for the November 2014 • Volume 42 • Number 11

Clinical Investigations work; revising it critically for important intellectual content; and final approval of the version to be published. Dr. Shemie contributed to the conception or design of the work; the acquisition, analysis, or interpretation of data for the work; drafting the work and revising it critically for important intellectual content; and final approval of the version to be published. Design and implementation of the study and manuscript development were coordinated at the Children’s Hospital of Eastern Ontario. Supported, in part, by grants from the Children’s Hospital of Eastern Ontario Research Institute, Physician’s Services Incorporated, and Canadian Institutes for Health Research. Dr. Dhanani consulted for the Trillium Gift of Life Network (Chief Medical Officer, Organ Procurement Organization, Ontario, Canada). His institution received grant support from Physicians Service Incorporated and the Children’s Hospital of Eastern Ontario Research Institute (CHEO) (Operational peer-reviewed competitive research grants distributed by the CHEO and the University of Ottawa). Dr. Hornby received grant support from the CHEO Research Institute (Operational peer-reviewed competitive research grants distributed by the CHEO Research Institute and the University of Ottawa), Physician Services Incorporated (Peer reviewed competitive research grant), and the Canadian Institutes of Health Research (Planning grant - Peer reviewed competitive research grant). She is employed by the University of Ottawa (Paid consultant for the Bertram Loeb Research Consortium in Organ and Tissue Donation) and Canadian Blood Services (Paid consultant for the Organ and Tissue Donation and Transplantation Division), received grant support from the Canadian Institutes of Health Research (Funding granted as part of the Canadian National Transplant Research Program, and received support for travel from the CHEO Research Institute. Dr. Ward is employed by the CHEO. Her institution received grant support from the CHEO (Operational peer-reviewed competitive research grants distributed by Children’s Hospital of Eastern Ontario Research Institute and University of Ottawa). Her institution received support for travel, participation in review activities, and manuscript writing and review from the CHEO. Dr. Baker received grant support (Operational peer-reviewed competitive research grants distributed by the CHEO Research Institute and University of Ottawa). His institution received grant support from the Canadian Institute for Health Research (CIHR) (CO-PI on a variety of clinical grants). Dr. Chamber-Evans received grant support (Operational peer-reviewed competitive research grants distributed by the CHEO Research Institute and University of Ottawa) and received support for travel from a CIHR Planning grant (train travel support for planning meetings). Drs. Fowler, Friedrich, and Gow received grant support (Operational peer-reviewed competitive research grants distributed by the CHEO Research Institute and University of Ottawa). Dr. Kutsogiannis received support for travel from the CIHR (Support for travel to one study related meeting in Ottawa, Canada) and is employed by Alberta Health Services (Medical Director, Human Organ Procurement and Exchange Program). Dr. Mcintyre received grant support (Operational peer-reviewed competitive research grants distributed by the CHEO Research Institute and University of Ottawa) and is employed by Genome Canada. Dr. Momoli’s institution received grant support from PSI, CHEO Internal, and CIHR (Operational peer-reviewed competitive research grants distributed by the CHEO Research Institute and University of Ottawa). Dr. Morin received grant support (Operational peer-reviewed competitive research grants distributed by the CHEO Research Institute and University of Ottawa) and is employed by Genome Canada. Dr. Scales received grant support from the Physicians' Services Incorporated Foundation (Fellowship in Translational Health Research). Dr. Yildirim’s institution received grant support from the CHEO (Operational peer-reviewed competitive research grants distributed by the CHEO Research Institute and University of Ottawa). Dr. Young received grant support for stimulator for EEG response, consulted for GE Healthcare (EEG studies), has patent for EEG stimulation, and received support for article research from the CIHR. His institution received grant support from London Health Sciences Centre (Operational peer-reviewed competitive research grants). Dr. Shemie received grant support (Operational peer-reviewed competitive research grants distributed by CHEO Research Institute and the University of Ottawa) and is employed by Montreal Children’s Hospital and Canadian Blood Services. The remaining authors have disclosed that they do not have any potential conflicts of interest. Dr. Dhanani is a Chief Medical Officer of Trillium Gift of Life Network, Ontario. Dr. Shemie is Medical Lead of Canadian Blood Services. For information regarding this article, E-mail: [email protected]

Critical Care Medicine

Objective: Controversies regarding the process and timing of the determination of death for controlled organ donation after circulatory death persist. This study assessed the feasibility of conducting a prospective, observational study of continuous monitoring of vital signs for 30 minutes after the clinical determination of death in five Canadian ICUs. Waveform data were analyzed. Design: Prospective observational cohort study. Setting: One pediatric and four adult Canadian ICUs. Patients: One month of age or older, admitted to the ICU, and for whom a consensual decision to withdraw life-sustaining therapies had been made, with an anticipation of imminent death. Interventions: None. Measurements and Main Results: Invasive arterial blood pressure, electrocardiogram, and oxygen saturation plethysmography activity were recorded and reviewed for 30 minutes after declaration of death. Feasibility was assessed (recruitment, consent rate, protocol compliance, and staff satisfaction). Of 188 subjects screened over 16 months, 41 subjects were enrolled (87% consent rate). Data collection was complete for 30 subjects (73% protocol compliance). In four subjects, arterial blood pressure resumed following cessation of activity. The longest period of cessation of arterial blood pressure before resumption was 89 seconds. The duration of resumed activity ranged from 1 to 172 seconds. No cases of sustained resumption of arterial blood pressure activity were recorded, and no instances of clinical autoresuscitation were reported. In nearly all patients (27 of 30), electrocardiogram activity continued after the disappearance of arterial blood pressure. Conclusions: This is the first observational study to prospectively collect waveform data for 30 minutes after the declaration of death. A future larger study may support initial data suggesting that circulatory function does not resume after more than 89 seconds of absence. Furthermore, persistence of cardiac electrical activity with the documented absence of circulation may not be relevant to declaration of death. (Crit Care Med 2014; 42:2358–2369) Key Words: autoresuscitation; determination of death; donation after circulatory death; organ donation; vital signs; withdrawal of life-sustaining therapies

C

ontrolled organ donation after circulatory death (DCD) has the potential to significantly increase deceased donation 1–3). Although it is generally supported by healthcare professionals and members of the public. However, there is persisting controversy regarding the criteria used to determine death, the time period required to confirm irreversible cessation of cardiac and neurologic function after circulatory arrest, and the potential for autoresuscitation (defined as the spontaneous resumption of circulation) (4, 5). Although clinical autoresuscitation has only been reported after cardiopulmonary resuscitation, there is still concern about the possibility in the uncontrolled and controlled DCD settings. The ethical foundation of deceased donation is the “dead donor” rule, which states that “vital organs should only be taken from dead subjects and, correlatively, living subjects must not be killed by organ retrieval” (6). Adherence to this rule necessitates physiological research to www.ccmjournal.org

2359

Dhanani et al

better inform the medical criteria for the determination of death in this context. Improved understanding of the natural history of the dying process after circulatory arrest is crucial to clarify issues of practice and timing of the declaration of death in the context of organ donation (7–10). The re-emergence of DCD programs is currently hindered by the aforementioned ethical and legal concerns. Addressing the physiological and ethical concerns about DCD will help establish accepted medical practices and public health policy for death after circulatory arrest. A prospective observational study of the natural history of cessation of physiological function after withdrawal of life-sustaining therapies will provide clarity and certainty in the determination of death within the DCD context. The primary objective of this study was to assess feasibility (recruitment, consent rate, and protocol compliance) of collecting vital sign data on the physiological changes that occur during the dying process—before, during, and after the clinical declaration of death after circulatory arrest—among patients in the ICU who had withdrawal of life-sustaining therapies. Secondary objectives were to describe the timing of the disappearance of vital signs after withdrawal of life-sustaining therapies, to describe the frequency of clinical autoresuscitation, and to evaluate key logistic and ethical factors for the design of future larger studies.

withdrawal of life-sustaining therapies: invasive arterial blood pressure monitoring, continuous 3-lead electrocardiogram, and oxygen saturation plethysmography. Subjects who had given consent for organ donation, who had already been declared dead by neurologic criteria, or who had a cardiac pacemaker were excluded. Local practices for the declaration of death were not altered. The subject’s family and staff were able to remain at the bedside throughout the period of data collection.

Data Collection For each patient enrolled as a study subject, the following data were collected: baseline demographics, admission diagnosis, preexisting conditions, interventions administered before and during withdrawal of life-sustaining therapies, time of initiation of withdrawal of life-sustaining therapies, time of declaration of death, and method(s) used to determine death after circulatory arrest. The time of initiation of withdrawal of life-sustaining therapies was defined as the time when the first action (extubation or cessation of pressors) to begin withdrawal of life-sustaining therapies occurred. Clinical declaration of death was not dictated and was performed as per local practice. The time of declaration of death was defined as the time of death that was recorded in the subject’s chart. In addition, any clinically observed instances of autoresuscitation were collected. Any violations of the study protocol were recorded. METHODS Vital sign (arterial blood pressure, electrocardiogram, and oxygen saturation plethysmography) waveform data for each Study Subjects The study was conducted at one pediatric and four adult ICUs subject were transmitted to a central monitor. Data collection started at the initiation of withdrawal of life-sustaining therain Canada. Institutional research ethics board’s approval and pies and continued until 30 minutes after the clinical determiwritten informed consent from the substitute decision maknation of death (Fig. 1). During this time period, the healthcare ers were obtained. We included subjects who were 1 month of age or older, admitted to the ICU, and for whom a consensual team left the bedside monitor on or off, as per local practice. decision to withdraw life-sustaining therapies had been made, Hard copies of the waveform activity data were printed for a 60-minute time period—from 30 minutes before to 30 minwith an anticipation of imminent death. Participation required agreement from the critical care physician and bedside health- utes after the declaration of death. Subsequently, waveform data were reviewed by a minimum of three adjudicators to care team. Subjects were required to have a minimum of the determine when cessation of waveform activity occurred and following continuous vital sign monitors in place at the time of whether there was any return of activity. Discrepancies between adjudicators were resolved by consensus. Electroencephalogram monitoring was in place as standard of care in four patients at one study site. Electroencephalograms were recorded with a fourchannel bipolar electroencephalogram monitor using four skin surface electrodes and a subhairline montage. Electroencephalogram waveform recordings were included for these subjects from withFigure 1. Timeline for monitoring of vital signs before, during, and after withdrawal of life-sustaining therapies. (1) Substitute decision maker (SDM) consents to withdrawal of life-sustaining therapies, (2) SDM consents to drawal of life-sustaining therasubject’s participation in study, (3) act of withdrawal of life-sustaining therapies, (4) variable time to absence of pies until 30 minutes after vital signs (physiological death), (5) variable time to declaration of death, and (6) 60 min of vital sign monitoring clinical declaration of death. collected (30 min prior to 30 min after the declaration of death). 2360

www.ccmjournal.org

November 2014 • Volume 42 • Number 11

Clinical Investigations

Three adjudicators determined the time of isoelectric electroencephalogram and recorded any instances of resumption. Isoelectric electroencephalogram was defined as no activity greater than 2 microvolts (11). A voluntary, self-administered staff satisfaction questionnaire was given to staff at the end of each subject’s participation (Appendix 1). The questionnaire included content items developed to capture staff satisfaction relevant to study procedures. An iterative process with content experts was used to review and revise questions as per Burns et al (12). Physicians completed a clinical declaration of death checklist for each study subject (Appendix 2).

be sufficient by expert consensus to ascertain the pilot study feasibility as per enrollment rate and compliance to protocol. Given that the scientific literature contains only case reports of autoresuscitation, no calculations could be done to estimate the number needed to observe one event. Likewise, it was unknown what recruitment rate would be obtainable because of the uniqueness of study procedure. Thus, sample size was based on an arbitrary enrollment time of 12 months for recruitment with knowledge of on-site ICU mortality rates. We anticipated recruitment rates of one patient per 36 days at the adult sites and one patient per 72 days at the pediatric site.

Sample Size A data-based sample size of 45 (10 subjects per adult site and five subjects for the pediatric site) was determined to

Statistical Analysis Statistical analysis was done using IBM SPSS Statistics for Windows, Version 20.0 (IBM Corp., Armonk, NY). Recruitment,

Figure 2. Screening and enrollment. *More than one reason given for some subjects. #Missed subjects include those who died when research coordinators were not available or too quickly to enroll or physician forgot to tell the family about the study. NDD = neurological determination of death, DCD = donation after cardiac death, SDM = substitute decision maker.

Critical Care Medicine

www.ccmjournal.org

2361

Dhanani et al

Demographics and Subject Characteristics Table 1.

Gender, n (%)  Male

27 (66)

Age in years, median (minimum, maximum)  Pediatric site, mo (n = 4)  Adult sites, yr (n = 37)

13.5 (1, 25) 64 (30, 90)

Illness severity at admission, median (interquartile range)  Pediatric site, Pediatric Risk of Mortality III score (n = 4)

22.5 (16.7–26)

 Adult sites, Acute Physiology and Chronic Health Evaluation II score (n = 37)

26.5 (20–30.7)

Primary reason for admission, n (%)  Neurologic condition

14 (34)

 Cardiac condition

10 (24)

 Respiratory condition

4 (10)

 Sepsis/infection

7 (17)

 Trauma

3 (7)

 Other

3 (7)

consent, and protocol compliance rates were calculated with counts and proportions. Demographic and physiological characteristics of persons participating in the study were assessed using descriptive statistics. For categorical variables, frequencies and percentages were tabulated. For continuous variables, medians, interquartile ranges, and maximum/minimum values were tabulated. In order to assess interobserver adjudicator reliability of reviewed waveform data, intraclass correlation coefficient was used for two randomly chosen adjudicators for each patient.

RESULTS Demographics From December 1, 2010, to May 31, 2012, participating ICUs screened 188 consecutive patients for whom withdrawal of lifesustaining therapies had been recommended by the healthcare team and/or requested by their substitute decision maker/legal guardian. Of these, 106 subjects did not meet study eligibility and 35 were missed as described in Figure 2. Most dying patients were excluded because they were an eligible organ donor or they did not have an arterial blood pressure monitor in situ. Of the remaining 47 eligible subjects whose families were approached for consent, six refused study participation (Fig. 2). Baseline characteristics of the 41 subjects enrolled in the study are presented in Table 1. Feasibility Each site enrolled patients for a preset time period of 12 months. We used a staggered start-up for sites and enrolled 91% of our 2362

www.ccmjournal.org

estimated sample size (41 of 45) over 16 months, with an 87% consent rate (41 of 47) among approached substitute decision makers/legal guardians. Three sites recruited the anticipated number of subjects; the pediatric site recruited four of five subjects and one adult site recruited seven of 10 subjects. Two sites recruited subjects more quickly than expected; one site required only 6 months while the other required 11 months. Protocol compliance was 73%. Twelve protocol violations after enrollment were recorded among 11 subjects (Fig. 2). Having an arterial catheter was an inclusion criterion into the observational feasibility study. All patients had a functioning arterial blood pressure monitor at the time of enrollment; however, nine patients were subsequently not reviewed because of loss of arterial catheter after observation period started and/or not able to be reviewed for technical reasons. Complete monitored waveform data of vital signs by arterial blood pressure and electrocardiogram monitoring were available for analysis in 30 subjects. Oxygen saturation plethysmography waveform signals were reviewed but were deemed by adjudicators to be too unreliable to be included for analysis. Intraclass coefficient for interobserver reliability was 0.921 for electrocardiogram and 0.938 for arterial blood pressure waveform data when assessed for two randomly chosen adjudicators for each patient. Staff satisfaction questionnaires were returned for 37 subjects, ranging from 1 to 3 per enrolled subject. Of the total 75 bedside staff who completed the staff satisfaction questionnaire, 39 (52%) were nurses, 15 (20%) were respiratory therapists, 18 (24%) were physicians, and three (4%) were considered “other.” Total number of questionnaires distributed was not documented. All of the questionnaires had more than 80% of individual questions completed. The majority (69%) agreed/strongly agreed that this was an important study to do; 91% agreed/strongly agreed that they were comfortable treating the subject during the study collection period; and most (80%) agreed/strongly agreed that they were comfortable leaving the monitoring equipment in place for 30 minutes after the declaration of death. Only 3% respondents agreed/strongly agreed that the study had a negative influence on patient care, and 7% respondents agreed/strongly agreed that the study interfered with how they performed their duties. Description of Withdrawal of Life-Sustaining Therapies One subject died before withdrawal of life-sustaining therapies. Therapies present at the time of withdrawal of life-sustaining therapies and at the time of declaration of death were recorded in the remaining 40 subjects. At the start of withdrawal of life-sustaining therapies, 34 subjects (85%) were mechanically ventilated and six (15%) were receiving noninvasive mechanical ventilation. Most subjects (80%) were receiving sedation/analgesia, including midazolam (46%), hydromorphone (27%), fentanyl (22%), and morphine (20%). Fourteen subjects (34%) were receiving vasopressors, most commonly norepinephrine (27%), vasopressin (22%), and phenylephrine (15%). All subjects had vasopressors stopped and were removed from ventilatory support at withdrawal of life-sustaining therapies. Of those mechanically ventilated, 21 November 2014 • Volume 42 • Number 11

Clinical Investigations

or interventions at time of withdrawal of life-sustaining therapies between the four subjects who had resumption of arterial blood pressure and the remainder who did not.

Figure 3. Description of resumption of arterial blood pressure (ABP) activity, occurring after an absence of at least 60 seconds, measured in 4 subjects. In two subjects, the activity returned twice during the monitoring period. All numbers represent time in seconds. Gray areas represent time with no activity and black areas represent times with some measurable activity.

subjects (62%) were extubated at or shortly after withdrawal of life-sustaining therapies and the others remained intubated. Thirty-two subjects (80%) had sedation/analgesia continued or added after withdrawal of life-sustaining therapies; midazolam was the most common agent (55%) used after withdrawal of life-sustaining therapies. All subjects died and clinically observed autoresuscitation was not reported within the study sample. The median time from withdrawal of life-sustaining therapies to clinical declaration of death was 60 minutes (range, 11–2,960 min). A declaration of death checklist was completed by healthcare providers for 39 of 41 subjects. The most common diagnostic tests used to clinically declare death were absent heart sounds (95%), absent breath sounds (90%), and absent palpable pulses (87%). Description of Arterial Blood Pressure Activity (Adult and Pediatric) Arterial blood pressure and electrocardiogram waveform data were reviewed in 33 of 41 subjects. Three subjects had waveform data containing multiple areas of missing tracings, indicating a loss of the waveform signal, and therefore were excluded, reducing the number of subjects with evaluable waveform tracings to 30. In 26 of these subjects, arterial blood pressure activity stopped and did not restart. Four subjects (13%) had a total of six episodes in which a return of arterial blood pressure waveform activity after a minimum of a 60-second absence was identified (Fig. 3). The longest period of cessation before resumption was 89 seconds. This was not identified as autoresuscitation, as it was not sustained resumption of circulation and not reported clinically. The duration of resumed arterial blood pressure activity ranged from 1 to 172 seconds. Measurable values for systolic and diastolic pressures during resumption were recorded for one adult subject and ranged from 12 to 27 mm Hg for systolic and from 11 to 15 mm Hg for diastolic. There were no obvious differences in demographics Critical Care Medicine

Description of Electrocardiogram Activity (Adult and Pediatric) In three of 30 subjects (10%), arterial blood pressure and electrocardiogram activity stopped at the same time. In 22 subjects, electrocardiogram activity persisted continuously for a median of 08:16 (mm:ss; range, 00:05–38:00) after absence of arterial blood pressure activity. In the other five subjects, electrocardiogram activity started and stopped intermittently for a median of 11:11 (mm:ss range, 00:37–36:29) after the absence of arterial blood pressure activity. Three subjects (10%) (two of whom were children) had electrocardiogram activity that continued up to the end of the monitoring period. Terminal electrocardiogram rhythms prior to and after the cessation of arterial blood pressure activity varied with no common pattern identified (Fig. 4). Description of Electroencephalogram Activity (Adult) Electroencephalogram waveform activity was available for review in four adult subjects. At the time of withdrawal of life-sustaining therapies, all four subjects had delta waveform electroencephalogram activity. In three subjects, isoelectric electroencephalogram preceded the cessation of electrocardiogram and arterial blood pressure by 04:00, 10:00, 10:10 mm:ss, respectively. In the remaining subject, delta and theta waveform activity persisted for 26:08 mm:ss after cessation of arterial blood pressure activity. Once ceased, no resumption of electroencephalogram activity was observed in any of the subjects during the full monitoring period, including in the subject in whom there was a documented 40-second resumption of arterial blood pressure activity following 75 seconds of its absence. Description of Pediatric Subjects Only For interest, pediatric subjects (n = 4) were also analyzed separately. Arterial blood pressure and electrocardiogram activity stopped at the same time in one subject. Three pediatric subjects had cessation of arterial blood pressure activity that preceded the cessation of electrocardiogram activity by 11:11, 27:42, 36:29 mm:ss, respectively. Of these, two subjects had electrocardiogram activity until the end of the monitoring period. There was no resumption of arterial blood pressure activity documented in any pediatric subjects. Electroencephalogram activity was not recorded in the pediatric study group.

DISCUSSION This exploratory observational study demonstrated that it is feasible to enroll patients and collect vital sign waveform data from adult and pediatric critical care subjects during the dying process after untreated circulatory arrest in the ICU. We enrolled 91% of our estimated sample size (41 of 45) over 16 months, with an 87% consent rate (41 of 47) among approached substitute decision makers/legal guardians. High consent rates and supportive staff satisfaction survey results suggest that bedside staff and subject families accept research www.ccmjournal.org

2363

Dhanani et al

Figure 4. Example of selected waveform data of a single subject after withdrawal of life-sustaining therapies with selected electrocardiogram, arterial blood pressure, oxygen saturation plethysmography, 4-lead electroencephalogram.

during the dying process. Although a feasibility study, this is the largest observational study in which vital sign waveform data have been described prospectively for an extended time period after death. All subjects died and there were no cases of sustained resumption of arterial blood pressure activity. In nearly all patients, electrocardiogram activity continued after the disappearance of arterial blood pressure. The higher than expected recruitment and consent rates for this study were potentially due to several strategies used in advance to obviate barriers to successful implementation. First, interprofessional consultations ensured robust ethics 2364

www.ccmjournal.org

submissions and consent design. Second, recurrent communication with site investigators occurred. Third, site visits, training, and follow-up review with bedside staff were done. The results of the staff satisfaction questionnaire indicate that the study was well accepted and did not interfere with patient care, suggesting support for future studies in critical care patients at the end of life and in the organ donation process. One previous study prospectively monitored 12 subjects after death, of which three had arterial blood pressure monitoring in situ (5). They reported electrocardiogram activity of short duration up to 10 minutes after circulatory arrest. November 2014 • Volume 42 • Number 11

Clinical Investigations

No spontaneous return of circulation was measured. Another study involved a retrospective chart review of vital sign data in 15 DCD donors. Full data were available for 12 of these patients and were limited to 2 minutes after arrest of circulation (13). They reported that circulation did not resume after more than 1 minute of absent circulation. A more recent, retrospective study of 73 subjects also collected vital sign data during the DCD process, and the time period after the arrest of circulation was 5 minutes (14). In that study, autoresuscitation did not occur during the standard 5-minute waiting observation period. Autoresuscitation is more likely in uncontrolled DCD where patients may have had resuscitation prior to withdrawal of lifesustaining therapies. However, there is limited international practice of uncontrolled DCD. Most jurisdictions continue to worry about autoresuscitation even when discussing controlled DCD. A survey of Canadian ICU practitioners demonstrated a witnessing of autoresuscitation in 37% of respondents (9). In addition, controlled DCD practice remains variable, with current recommended wait times from arrest of circulation to procurement ranging from 2 to 10 minutes (8). Case studies and anecdotal reports of autoresuscitation (10) and events such as the “lazarus phenomenon” have delayed the uptake of DCD protocols. Critics of DCD have correctly pointed out that there is very limited data regarding the pattern of cessation of circulatory function during the process of controlled DCD and that more data are needed to ensure that donors are declared dead according to standardized criteria. This is the first study to prospectively document arterial blood pressure and electrocardiogram activity after declaration of death for greater than 5 minutes. In four subjects, arterial blood pressure waveform activity, defined as present regardless of frequency and magnitude, resumed after at least 60 seconds of cessation. There was no identification of arterial blood pressure activity after cessation longer than 89 seconds, a time period which is below internationally accepted DCD practice recommendations (8). This arterial blood pressure activity persisted for variable durations from 1 to 172 seconds. We considered autoresuscitation to be defined as the unassisted return of spontaneous circulation after a declaration of death. In our study, we did document short-lived return of waveform activity after 60 seconds of absence. This activity was not sustained and not reported clinically. Thus, to report these events as autoresuscitation was not justified. Although documented as activity for the purposes of this descriptive study, the clinical relevance of minimal pulse pressure is uncertain and beyond the scope of this study. The cases of resumption of waveform activity were associated with extremely low pulse pressure. There is little evidence of what blood pressure is meaningful and clinically relevant especially as it relates to perfusion of the brain and coronary arteries. Limited data suggest that cerebral electrical activity in adults ceases below 50 mm Hg systolic blood pressure (15), but the minimal duration or pulse pressure needed to sustain or resume perfusion and function of end organs, especially of the brain, is not known. Ïn this study, the documented resumption Critical Care Medicine

of activity was likely pulseless and would not be effective or sustained circulation and would not have any clinical relevance. In a majority of subjects (27 of 30), electrocardiogram activity continued (median, 05:43 mm:ss; maximum of 38 min) after the absence of arterial blood pressure activity. The absence of associated arterial blood pressure activity suggests that the inclusion of electrocardiogram criteria in the declaration of death may have limited clinical relevance. Despite our small sample size, we feel that absence of electrocardiographic activity may not be necessary as part of the circulatory criteria for determining death. This would lead to focus on cessation of function and circulation rather than simple activity. Standardized definitions are needed for the cessation and potential resumption of clinically meaningful arterial blood pressure and electrocardiogram activity. Most importantly, further study is needed to evaluate the clinical functional significance of any resumption of monitored circulation to meaningful neurologic activity. The question of how long one needs to be acirculatory before autoresuscitation will not occur is important. An examination of how long a patient needs to be acirculatory before neurologic function ceases permanently will also need to be directly addressed. Electroencephalogram activity was measured in a small subgroup of subjects (n = 4). At the time of withdrawal of life-sustaining therapies, all four subjects had delta activity, reflecting coma and thus unconsciousness and the lack of awareness. Isoelectric electroencephalogram preceded the cessation of electrocardiogram and arterial blood pressure in three subjects. In the fourth subject, waveform activity persisted following electrocardiogram and arterial blood pressure cessation. This electroencephalogram activity is difficult to comprehend on a physiological basis given its presence after prolonged loss of circulation suggesting that these waveform bursts could be artifact in nature. Especially given that we only recorded neurologic data in four patients, one cannot make any recommendations about the length of time required for permanent cessation of neurologic function. However, our limited findings would indicate that neurologic function permanently ceases prior to the 5-minute wait period required as part of Canadian DCD practice. It is evident that the key to future studies on the determination of cessation of circulatory function will need to include observation of neurologic function. Enrollment in this study was successful notwithstanding previous research that has demonstrated that studies carried out in ICUs may be associated with low consent rates (16, 17). Staff questionnaires showed support for the research program and comfort with the procedure, indicating that the study did not interfere with end-of-life care. Our study protocol and research process were developed by a research team constructed to include experts in critical care practice and research, ethics, and the law. We feel that this had an important positive impact on why the study protocol met with very little resistance from research ethics boards and medical staff at the study sites. Despite achieving its objectives, this was a descriptive feasibility study not designed to evaluate the prevalence of clinical events in the process of dying. Indeed, although no cases www.ccmjournal.org

2365

Dhanani et al

of clinical autoresuscitation were reported, our relatively small sample size precludes drawing any inferences that such event could never occur. The findings were derived from patients across age groups and from multiple hospitals but may not be generalizable to those who are candidates for DCD. None of the study subjects became organ donors (exclusion criteria), although subjects may have been eligible for DCD. Despite our high recruitment and consent rates, 35 potential subjects were missed, primarily because they died during the evening or on weekends when research personnel were not available. Capture of multiple waveform data over 60 minutes from multiple centers and different monitor platforms was challenging due to technical or logistic problems with the recording or transferring of vital sign waveform data. In addition, review of the data by adjudicators was difficult despite high intraclass correlation. There was risk for subjective misinterpretation among artifact, activity, and clinically meaningful activity, especially without a priori definitions. As an example, oxygen saturation plethysmography was recorded but found to be difficult to interpret and subsequently not reviewed by adjudicators. Discrepancies between adjudicators were most commonly attributed to recording and scale differences rather than differences in judgment. We could not reliably rule out waveform misinterpretation, especially for electroencephalogram data, due to both physiological and nonphysiological artifacts that are common in the ICU. A more specific electronic data capture platform, a waveform reviewing technology, clear a priori definitions, and means of accounting for signal interference are essential for future larger studies.

CONCLUSIONS In this prospective multicenter observational study, we showed that it is feasible and acceptable to study the natural history of death by describing vital signs during the dying process and 30 minutes after declaration of death. This is the first study to systematically collect prospective vital sign data for greater than 5 minutes after circulatory arrest. A larger study is needed to support the results that show no return of circulation after 89 seconds of cessation of arterial blood pressure activity, which is consistent with current practices for determining death prior to controlled DCD. Furthermore, persistence of cardiac electrical activity with the documented absence of circulation may not be relevant to declaration of death. Any documented resumption of activity was likely not effective or sustained circulation and would not have any clinical relevance.

ACKNOWLEDGMENT We thank the families of the patients who agreed for them to participate in this study and the multiple research coordinators, assistants, and technical specialist involved: Irene Watpool,

2366

www.ccmjournal.org

BScN, Amanda Van Beinum, BScN, Geoff Green, PhD, Julia Lee, BScN, Yoon Lee, BScN, Kerry Poretta, RT, Marlene Santos, BSc, MD, Orla Smith, BScN, MSc, Jane Topolovec-Vranic, PhD, Ziva Brigler, Loretta Norton, MSc, PhD (c), Kelli Pearce, BScN, Kirby Scott, MSc, Pat Thompson, MScN, Mark Wan, and Brandon Beaudry. We are grateful to Paulina Mirsky, Program Coordinator, Loeb Chair in Organ Donation, University of Ottawa, for her administrative support of the study and review of the manuscript and to Salwa Akiki for statistical analysis.

REFERENCES

1. Hernadez-Alejandro R, Wall W, Jevnikar A, et al: Organ donation after cardiac death: Donor and recipient outcomes after the first three years of the Ontario experience. Can J Anaesth 2011; 58:599–605 2. Moser M, Sharpe M, Weernink C, et al: Five-year experience with donation after cardiac death kidney transplantation in a Canadian transplant program: Factors affecting outcomes. Can Urol Assoc J 2012; 6:448–452 3. Halpern SD, Barnes B, Hasz RD, et al: Estimated supply of organ donors after circulatory determination of death: A population-based cohort study. JAMA 2010; 304:2592–2594 4. Bastami S, Matthes O, Krones T, et al:Systematic review of attitudes toward donation after cardiac death among healthcare providers and the general public. Crit Care Med 2013; 41:897–905 5. Wijdicks EF, Diringer MN: Electrocardiographic activity after terminal cardiac arrest in neurocatastrophes. Neurology 2004; 62:673–674 6. Robertson JA: The dead donor rule. Hastings Cent Rep 1999; 29:6–14 7. Nakagawa TA, Rigby MR, Bratton S, et al: A call for full public disclosure for donation after circulatory determination of death in children. Pediatr Crit Care Med 2011; 12:375–377 8. Dhanani S, Hornby L, Ward R, et al: Variability in the determination of death after cardiac arrest: A review of guidelines and statements. J Intensive Care Med 2012; 27:238–252 9. Dhanani S, Ward R, Hornby L, et al; Canadian Critical Care Trials Group; Bertram Loeb Research Consortium in Organ and Tissue Donation: Survey of determination of death after cardiac arrest by intensive care physicians. Crit Care Med 2012; 40:1449–1455 10. Hornby K, Hornby L, Shemie SD: A systematic review of autoresuscitation after cardiac arrest. Crit Care Med 2010; 38:1246–1253 11. American Clinical Neurophysiology Society: Guideline 3: Minimum technical standards for electroencephalogram recording in suspected cerebral death. J Clin Neurophysiol 2006; 23:97–104 12. Burns KE, Duffett M, Kho ME, et al: A guide for the design and conduct of self-administered surveys of clinicians. CMAJ 2008; 179: 245–252 13. DeVita MA, Snyder JV, Arnold RM, et al: Observations of withdrawal of life-sustaining treatment from patients who became non-heart-beating organ donors. Crit Care Med 2000; 28:1709–1712 14. Sheth KN, Nutter T, Stein DM, et al: Autoresuscitation after asystole in patients being considered for organ donation. Crit Care Med 2012; 40:158–161 15. Kovách AG, Sándor P: Cerebral blood flow and brain function during hypotension and shock. Annu Rev Physiol 1976; 38:571–596 16. Menon K, Ward RE, Gaboury I, et al: Factors affecting consent in pediatric critical care research. Intensive Care Med 2012; 38:153–159 17. Smith OM, McDonald E, Zytaruk N, et al; PROTECT Research Coordinators; PROTECT Investigators; Canadian Critical Care Trials Group; Australian, New Zealand Intensive Care Society Clinical Trials Group: Rates and determinants of informed consent: A case study of an international thromboprophylaxis trial. J Crit Care 2013; 28:28–39

November 2014 • Volume 42 • Number 11

Clinical Investigations

APPENDIX 1. Staff Satisfaction Questionnaire Pilot Study for the Determination of Death After Cardiac Arrest The Physiological Changes that Occur After Death from Cardiac Arrest: An Observational Pilot Study Staff Satisfaction Questionnaire Physiologically, death has been described as the complete and irreversible cessation of brain and/or cardiorespiratory function. Brain death has been well described and accepted in medical practice. It has been clearly defined based on 1) clinical criteria of absent brainstem function and, when applicable, 2) ancillary tests demonstrating absent cerebral blood flow. However, the determination of death after cardiac arrest is not well described. You have recently witnessed/observed/cared for a patient who participated in a study that was designed to determine the feasibility of collecting epidemiological data on the physiological changes that occur during the dying process and in the postmortem period after cardiac arrest, following withdrawal of lifesustaining therapies. As part of this study, we are seeking feedback from those members of the treating team who were most intimately involved in the process. We are interested in your opinion of the study procedures and whether they affected patient care in the ICU. We ask that you answer the following questions regarding your experience. This questionnaire contains 11 questions in total, addressing issues broadly related to your impressions of the study. The questionnaire will take approximately 5 minutes to complete. Your answers will remain confidential. This information is being collected in the context of a research study. Your participation is entirely voluntary. You may refuse to complete the questionnaire. The research ethics board has reviewed this survey and has provided ethics approval. Your completion of the questionnaire will tell us you have understood the above and have agreed to participate as described. Should you have any questions or comments, please do not hesitate to contact the principal investigator, Dr. Sonny Dhanani, at 613-737-7600, extension 3654, or at the address [email protected] Below are a number of statements that may or may not describe you, your feelings, or your behavior. Please read each statement carefully and select (check mark) to indicate how well you agree or disagree. If the statement does not apply or if you don’t know or prefer not to answer, please check “N/A.” Please indicate how well you agree or disagree with the following statements:

Strongly ­Disagree Disagree Neutral Agree

Strongly Agree

N/A

 1. I felt that the purpose and the objectives of the study were clearly explained to me.  2. I received enough training/information regarding the study.  3. I believe that it is important to do this study.  4. I felt that the study interfered with how I performed my duties.  5. I felt that the study had a negative influence on patient care.  6. I felt that carrying out the study procedures was too time consuming.  7. I felt comfortable in leaving the monitors in place for 30 minutes after the declaration of death.  8. I felt comfortable caring for a patient who was enrolled in this study.  9. I felt comfortable discussing the study procedures with patient’s loved ones.  10. In my opinion, the family had a good understanding of the study.  11. I felt that the study was well accepted by the patient’s loved ones.

1. Please include any additional comments regarding your participation in the Pilot Study for the Determination of Death After Cardiac Arrest.

_____________________________________________________________________ _____________________________________________________________________ (Continued) Critical Care Medicine

www.ccmjournal.org

2367

Dhanani et al

Appendix 1.

(Continued) Staff Satisfaction Questionnaire

_____________________________________________________________________ _____________________________________________________________________ _____________________________________________________________________ _____________________________________________________________________ _____________________________________________________________________ Please tell us about yourself: 1. What is your profession? □  Registered nurse □  Respiratory therapist □ Physician □ Other_______________________________________________________________ ________________________________________________________________ 2. How many years of ICU experience do you have?      (Please indicate the number in the box)

□

On behalf of the University of Ottawa’s Bertram Loeb Research Consortium, the McGill University Health Centre, and the Clinical Research Unit at Children’s Hospital of Eastern Ontario Research Institute, I would like to thank you for your time and effort. Sincerely Sonny Dhanani Critical Care Children’s Hospital of Eastern Ontario

APPENDIX 2. Checklist for the Determination of Death DDePICt Pilot Study for the Determination of Death after Cardiac Arrest Checklist for the Determination of Death

PLEASE COMPLETE THIS FORM 30 MINUTES AFTER THE DECLARATION OF DEATH OR ONCE DATA COLLECTION IS COMPLETE

Summary and Purpose Physiologically, death has been described as the complete and irreversible cessation of brain and/or cardiorespiratory function. Brain death has been well described and accepted in medical practice. It has been clearly defined based on 1) clinical criteria of absent brainstem function and, when applicable, 2) ancillary tests demonstrating absent cerebral blood flow. However, the determination of death after cardiac arrest is not well described. The objective of this checklist is to record how the determination of death after cardiac arrest was performed on DDePICt pilot study subjects. The information from this checklist will help inform the design of a multicentre, prospective observational study of the determinants of death after cardiac arrest in the ICU. Your answers on this checklist will remain confidential. Please note that the checklist is being collected in the context of a research study and not for quality assurance/improvement. Your participation in completing this checklist is entirely voluntary. Should you have any questions or comments, please do not hesitate to contact the principal investigator, Dr. Sonny Dhanani, at 613-737-7600, extension 3654, or at the address [email protected]. By completing this checklist, you are understood to have consented to participating in this study.

1. Declaration of Death A. Specifically, which of the following diagnostic tests were used to determine death after cardiac arrest? (Please check your responses below): 2368

www.ccmjournal.org

November 2014 • Volume 42 • Number 11

Clinical Investigations

APPENDIX 2. (Continued) Checklist for the Determination of Death Absent heart sounds by auscultation

□ YES

□ NO

Absent palpable pulse

□ YES

□ NO

Absent pulse by audible Doppler

□ YES

□ NO

Absent blood pressure by noninvasive monitoring

□ YES

□ NO

Flat arterial line tracing

□ YES

□ NO

Pulseless electrical activity (nonperfusing rhythm)

□ YES

□ NO

Flat electrocardiogram tracing (standard 3-lead electrocardiogram)

□ YES

□ NO

Absent breath sounds by auscultation

□ YES

□ NO

Absent pulse oximetry

□ YES

□ NO

Unresponsiveness to painful stimulus

□ YES

□ NO

Fixed and dilated pupils

□ YES

□ NO

Other

□ YES

□ NO

(no oxygen saturation and/or no plethysmography tracing)

If other, please specify:______________________________________________ B. What other measures, if any, did you use to confirm the determination of death after cardiac arrest (Please check your responses in the table below) Repeat your diagnostic tests

□ YES

□ NO

Get confirmation by a second physician

□ YES

□ NO

Other

□ YES

□ NO

If other, please specify: _______________________________________________________ If confirmation by a second physician was obtained, what was the duration of time between declaration of death and second confirmation?  □ N/A ________ minutes C. If the diagnostic tests were repeated, how many minutes after completing your tests the first time did you repeat the evaluation?

2. Return of Circulation (autoresuscitation) Please indicate the number of minutes in the box or “NA” if not applicable. Autoresuscitation is defined as the unassisted return of spontaneous circulation after a declaration of death. Did autoresuscitation occur:  □ YES □ NO If yes, time of occurrence: ____________________ (hh:mm) Reported by:  □ Physician  □ Nurse  □  Other: Specify_______________ Description of event: ___________________________________________________________________________________________________ ___________________________________________________________________________________________________ ___________________________________________________

3. Who Declared Death Please indicate who performed the declaration of death: □  Staff physician □ Fellow □ Resident □  Other: _________________________________ THANK YOU!! Critical Care Medicine

www.ccmjournal.org

2369