Trend of Outcome Metrics in Recent Out-of-Hospital-Cardiac-Arrest Research: A Narrative Review of Clinical Trials

Cardiopulmonary resuscitation (CPR) research traditionally focuses on survival. In 2018, the International Liaison Committee on Resuscitation (ILCOR) proposed more patient-centered outcomes. Our narrative review assessed clinical trials after 2018 to identify the trends of outcome metrics in the field OHCA research. We performed a search of the PubMed database from 1 January 2019 to 22 September 2023. Prospective clinical trials involving adult humans were eligible. Studies that did not report any patient-related outcomes or were not available in full-text or English language were excluded. The articles were assessed for demographic information and primary and secondary outcomes. We included 89 studies for analysis. For the primary outcome, 31 (35%) studies assessed neurocognitive functions, and 27 (30%) used survival. For secondary outcomes, neurocognitive function was present in 20 (22%) studies, and survival was present in 10 (11%) studies. Twenty-six (29%) studies used both survival and neurocognitive function. Since the publication of the COSCA guidelines in 2018, there has been an increased focus on neurologic outcomes. Although survival outcomes are used frequently, we observed a trend toward fewer studies with ROSC as a primary outcome. There were no quality-of-life assessments, suggesting a need for more studies with patient-centered outcomes that can inform the guidelines for cardiac-arrest management.


Introduction
The number of trials in the field of out-of-hospital cardiac arrest (OHCA) has grown exponentially throughout the last decade, largely with a focus on increased survival as a key metric for the effectiveness of interventions [1].In 2018, in an effort to clarify meaningful outcomes for ongoing research, the International Liaison Committee on Resuscitation published the Core Outcome Set for Cardiac Arrest (COSCA) in Adults [2].The COSCA initiative process painstakingly reviewed the literature for outcome data, created a priority list that was based on clinicians', patients', and their relatives/partners' preferences, and derived an outcome set based on the consensus of an international advisory panel.
The literature review utilized for the COSCA process [3] affirmed that, within cardiopulmonary resuscitation research, there was a large variation in outcome metrics, such as the types of outcomes, the timing of when to measure outcomes, and the methods.For example, in the 61 included randomized controlled trials, survival was the most reported outcome (85.2%); however, there were 39 individual ways to assess this outcome.Furthermore, many outcomes (41%) were physiologic variables related to body structure or body function, such as heart rate or biomarkers.While the methods of measurement of physiologic data points were also heterogeneous, these outcomes are likely less relevant to patient-centered outcomes.Notably, none of the studies included health-related quality-of-life measurements.
After the outcome data were extracted from the COSCA systematic review, surveys were completed by clinicians, patients, and their relatives [4][5][6].Importantly, patients and partners consistently ranked life-impact outcomes at 1 year, including emotional wellbeing and family impact, as important [5].This is largely consistent with other studies on post-intensive-care syndrome (PICS), demonstrating that outcomes after surviving critical illness, including neurocognitive injury, physical debility, and psychosocial impact, are all patient-centered metrics that have historically been of little focus and poorly understood, yet have wide implications [7,8].A recent study did look at out-of-hospital-cardiac-arrest survivors and the incidence of PICS at 3-and 12-month follow-up [9].That study found that 50% of survivors experienced PICS at 3-months and 47% at 12-month follow-up [9].
Based on the systematic review, survey results, and panel discussion, the COSCA advisory group recommended that researchers include several core outcomes in ongoing cardiopulmonary resuscitation research [2].These outcomes focus on three domains: survival, neuroprognostication, and health-related quality of life.Specifically, the panel recommended measuring (a) survival at hospital discharge, at 30 days, or both; (b) neurologic function measured by mRS at hospital discharge, at 30 days, or both; and (c)-health-related quality of life measured with least one tool at 90 days and at intervals up to 1 year after cardiac arrest.They recommended using the Health Utilities Index (HUI3), the Short-Form 36-Item (SF-36v2) Health Survey, and the EuroQol 5D-5L (EQ-5D-5L) as tools to determine this outcome of quality of life.
Intuitively, the concepts of cardiopulmonary resuscitation outcomes do overlap.The return of spontaneous circulation, for example, is necessary for calculating more distant neurologic outcomes, even at 30 days [10].Similarly, quality-of-life metrics are dependent on neurologic recovery.The return of spontaneous circulation as a primary outcome, however, is not necessarily a valuable, patient-centered outcome.When developing large trials and publishing association guidelines, it is important to focus on patient-centered outcomes that are consistently measured and meaningful.The COSCA outcome set provided that framework.
This narrative review aims to search the published literature since the publication of the COSCA outcomes in 2018 to determine the trend of outcome metrics that have been measured and to compare whether these outcomes align with the COSCA recommendations.

Study Selection
The PubMed database was searched from 1 January 2019 to 22 September 2023, using the search terms "(intervention) AND ("Out-of-Hospital Cardiac Arrest" [Mesh] OR "Heart Arrest" [Mesh])".We included studies starting in January 2019, rather than in the COSCA publication year of 2018, in order to increase the likelihood that researchers would have time to incorporate additional patient-centered outcomes as recommended by the COSCA guidelines into their research methods.Our inclusion criteria were randomized controlled trials, prospective observational trials, or secondary analyses of prospective observational studies in adult human subjects that evaluated any diagnostic or therapeutic interventions in out-of-hospital cardiac arrest and reported any patient-related outcomes.We excluded studies that did not report any patient-related outcomes, such as studies assessing levels of biomarkers, non-original publications (reviews, meta-analyses), and conference proceed-ings.Studies not available in full-text English language were excluded.Two investigators independently screened the titles and abstracts for eligibility, and a third investigator adjudicated any discrepancies.All studies required agreement from at least two investigators to be included in the analysis.This review did not involve any human subjects; thus, it was not submitted to the Institutional Review Board at the Principal Investigator's institution.

Data Collection
The data for the assessments included the demographic information (year of publication, country of study, study design, sample size) of each article and the patient-related primary outcomes and secondary outcomes (survival, neurofunctional outcomes, quality of life).In the first trial for data collection, the interrater agreement between investigators was 96%, so our standardized datasheet was well designed, and the data were validated.
For the primary outcome, 31 (35%) studies used an assessment for neurocognitive functions, while 27 (30%) used survival as their outcome.There were 8 (9%) studies using any neurocognitive assessment at hospital discharge, and most studies (22, 25%) assessed neurocognitive function beyond 30 days.In terms of survival as a primary outcome, four (4%) and seven (8%) studies used survival to hospital admission and hospital discharge, respectively.There were 4 (5%) and 12 (13%) studies that assessed survival at 30 days and beyond 30 days, respectively (Table 1).
For secondary outcomes, neurocognitive function, at any time of assessment, was present in 20 (22%) studies, and survival at any time of assessment was used as the secondary outcome in 10 (11%) studies.Twenty-six (29%) studies used both survival and neurocognitive function.Thirty-one (35%) studies listed other primary outcomes outside of the COSCA guidelines (Figure 1A).Among all the studies, there were higher percentages of studies being published in 2019 that used neurocognitive functions as a primary outcome.The percentages of studies that used survival at any time period as a primary outcome appeared to be unchanged between 2019 and now (Figure 1C).On the other hand, the number of studies that used both neurocognitive function and survival or the number of studies that used only neurocognitive function as a secondary outcome remained the same since 2019.The number of studies that reported only survival as their secondary outcome was decreasing in 2021-2022 (Figure 1D).Among all the studies, there were higher percentages of studies being published in 2019 that used neurocognitive functions as a primary outcome.The percentages of studies that used survival at any time period as a primary outcome appeared to be unchanged between 2019 and now (Figure 1C).On the other hand, the number of studies that used both neurocognitive function and survival or the number of studies that used only neurocognitive function as a secondary outcome remained the same since 2019.The number of studies that reported only survival as their secondary outcome was decreasing in 2021-2022 (Figure 1D).
Figure 2 depicts the types of outcome assessments according to different types of study designs.A majority of the randomized trials used either neurocognitive function or survival as their primary outcome (Figure 2A) or secondary outcome (Figure 2B).Among all the studies, there were higher percentages of studies being published in 2019 that used neurocognitive functions as a primary outcome.The percentages of studies that used survival at any time period as a primary outcome appeared to be unchanged between 2019 and now (Figure 1C).On the other hand, the number of studies that used both neurocognitive function and survival or the number of studies that used only neurocognitive function as a secondary outcome remained the same since 2019.The number of studies that reported only survival as their secondary outcome was decreasing in 2021-2022 (Figure 1D).
Figure 2 depicts the types of outcome assessments according to different types of study designs.A majority of the randomized trials used either neurocognitive function or survival as their primary outcome (Figure 2A) or secondary outcome (Figure 2B).

Discussion
In this narrative review, we have demonstrated the trends and changes in the selection of outcomes in landmark studies in adult cardiac-arrest care.Since the publication of the COSCA initiative in 2018, we have observed an increasing trend of studies adopting outcome measures as recommended by the COSCA initiative [2].Although our observation demonstrates a trend toward the adoption of the recommendations by Haywood et al. [2], up to 30% of our included studies still opted for other outcomes of interest.We hope that this narrative review will highlight the importance of clinical outcomes beyond survival and encourage the incorporation of higher-level outcomes in future studies.
ROSC has long been the outcome of interest, but there are several concerns with the selection of ROSC as an outcome.First, multiple studies have shown that improved ROSC rates may not be associated with a more meaningful improvement in more-distant outcomes such as neurocognitive function or even survival [11].In fact, some studies have shown that rates of improved ROSC may even be associated with worse neurologic outcomes [12].Given the increasing evidence from surveys of the general population and patients indicating a strong preference for functional outcomes rather than ROSC [13][14][15][16], which may not necessarily even translate to improved rates of survival to admission (a brief episode of ROSC may still be considered a "positive" result in a study), it is imperative for higher-impact studies to avoid the use of ROSC as an outcome.
A higher-level outcome is survival to hospital admission.However, it is often argued that admission to hospital does not translate into discharge from the hospital (e.g., patient will be admitted to intensive care but die shortly after), and this is not a patient-oriented outcome; therefore, we urge that caution should be exercised in interpreting these results.Cardiac arrest is often a sudden-onset disease, unanticipated by the patient, family, and friends.It creates immense emotional distress for the family and can lead to lasting psychological harm [16].Survival to hospital admission may allow family and friends time to process this life-altering event and provide much-needed closure.As an added benefit, it may improve the chances of organ donations to help other patients in need [17].Nevertheless, its utility as a primary outcome is questionable.
Survival to hospital discharge, another higher-level outcome, is historically considered a superior choice, although survival to hospital discharge may not translate to good neurologic outcomes in the survivors [18], as more than 50% of discharged patients would have very poor neurologic function, and approximately 24% of cardiac-arrest survivors rely heavily on constant care [19], which, according to surveys, is not a desired outcome by many [20].Survivors with the ability to communicate their wishes may be able to later express this to their clinicians and families; however, in the absence of the ability to clearly state their wishes, this may create both ethical and psychological dilemmas.
This development has led to a recent shift to an even higher order of outcome that is preferred for large-scale, multicenter, and often multinational studies that are designed to inform practice guidelines.These outcomes are measured in standardized forms and include the cerebral performance category (CPC) and the modified Rankin scale (mRS).These measurements allow for a fairly reliable differentiation of the functional neurologic outcomes in survivors of cardiac arrest and for interrater reliability in the mRS or CPC [2].As evidenced by our focused review, since the publication of the COSCA initiative, many large, multicenter, randomized controlled trials have adopted such neurologic function measurements as outcomes.However, there is a variation even when neurologic outcomes are reported.Survival to hospital discharge with a good neurologic outcome, being defined as a cerebral performance category (CPC 1-2), was reported in some studies, but there is a lack of consensus on the timeline for assessing an improvement in neurologic outcome.Although we excluded meta-analyses in this review, trial sequential analyses have been incorporated promisingly in evaluating neurological outcomes among existing studies [21,22].In this analysis, the neurologic outcomes based on CPC or the mRS at the time of discharge or on day 28 after arrest was assessed as a secondary outcome in eight studies, respectively.We only identified two studies that assessed CPC and survival beyond 30 days.Additionally, it must be noted that neurologic function scores do not completely capture the full spectrum of cognition and psychological well-being of the survivors [2].Even among OHCA survivors with a perceived good cognitive outcome (CPC ≤ 2), a high proportion of survivors have reduced-memory-retrieval deficits and cognitive impairment six months after arrest [23].As such, members from the COSCA initiative, while recommending against the use of CPC in cardiac-arrest survivors, unanimously recommended mRS as the choice of neurological assessment.However, a majority of studies identified in our review still used the CPC scale as either their primary or secondary outcomes.
Due to the shortcoming of the neurologic outcome assessments, more sophisticated questionnaires such as the Health Utilities Index, the Short-Form 36-Item Health Survey, and the EuroQol 5D-5L were proposed to provide a more holistic view of the survivor's health.Few studies are able to evaluate functional outcomes or survival along a longitudinal timeline.Up to 55% of survivors have poor functional outcome at 6 months, defined as a score of 4 to 6 on the modified Rankin scale [24].Even among those survivors, the health-related quality-of-life score was ranked at a moderate level of 74-75 based on the EuroQol group's visual analogue scale, with the reference range of scores of 0-"the worst health you can imagine" to 100-"the best health you can imagine" [25].While these tools provide valuable insight into the long-term outcomes of cardiac-arrest care, they may not be optimal outcomes for interventions that are aimed at short-term outcome measurements and should not be selected as the primary outcome in such studies.Nonetheless, it is noteworthy that none of the studies in our analysis used any measure for quality of life as their outcome assessment.
It also points to the importance of an interconnected health system to capture and evaluate patients for longitudinal outcomes.Another advantage of an interconnected system is that it allows the evaluation of associated health care costs and resource utilization assessment [25].Post-cardiac-arrest hospitalizations resulted in a high associated health care cost, with an increased length of stay, medical procedures, and systems of care [26].The cost effectiveness of interventions should be discussed, but few studies were able to evaluate the economic impact of cardiac arrest in secondary analyses or outcome data [25,27,28].
Our review does have many limitations.First of all, it is possible that many of the trials were designed and implemented long before the publication of COSCA; thus, the authors might not have been able to change their studies' outcomes according to the recommendation.Furthermore, we did not assess publications before the publication of the COSCA initiative to ascertain a trend in these outcome metrics before and after COSCA initiative's recommendation.We also searched only on the PubMed database and acknowledge that we could have missed relevant studies listed on other databases.Additionally, we only included studies that reported at least one patient-related outcome; therefore, it is likely that we have artificially increased the rates of patient-related outcomes in our analysis by eliminating a large number of studies that investigated non-patientrelated outcomes such as biomarker levels and quality of chest compression.

Conclusions
Our analysis observed a trend toward an increasing number of studies using neurocognitive assessment as outcomes among the cardiopulmonary resuscitation publications since 2019.There was also a decreasing trend for the use of survival as the only outcome metric among these studies.Further studies in the field of cardiopulmonary resuscitation are necessary to confirm these trends in compliance with the recommendation by the COSCA trial.Other than randomization to 1 of 2 arms, there was no specified study intervention [112] Survival to hospital discharge

Any survival outcome
Both survival and neurocognitive function

Figure 1 .
Figure 1.Trend of outcomes among publications involving patients with out-of-hospital cardiac arrest being included in this review.(A) Percentages of different categories of primary outcome, among the analyzed publications.(B) Percentages of different categories of all secondary outcomes, among the analyzed publications.(C) Percentages of different categories of primary outcome, in each year from 2019 to 2023.(D) Percentages of different categories of secondary outcome, in each year from 2019 to 2023.

Figure 1 .
Figure 1.Trend of outcomes among publications involving patients with out-of-hospital cardiac arrest being included in this review.(A) Percentages of different categories of primary outcome, among the analyzed publications.(B) Percentages of different categories of all secondary outcomes, among the analyzed publications.(C) Percentages of different categories of primary outcome, in each year from 2019 to 2023.(D) Percentages of different categories of secondary outcome, in each year from 2019 to 2023.

Figure 1 .
Figure 1.Trend of outcomes among publications involving patients with out-of-hospital cardiac arrest being included in this review.(A) Percentages of different categories of primary outcome, among the analyzed publications.(B) Percentages of different categories of all secondary outcomes, among the analyzed publications.(C) Percentages of different categories of primary outcome, in each year from 2019 to 2023.(D) Percentages of different categories of secondary outcome, in each year from 2019 to 2023.

Figure 2 .
Figure 2. Trend of primary outcome (A) and secondary outcome (B) measurements according to types of study designs.

Figure 2 .
Figure 2. Trend of primary outcome (A) and secondary outcome (B) measurements according to types of study designs.

Table 1 .
A Characteristics of studies included in the analysis.