Reexamination of the UN10 Rule to Discontinue Resuscitation During In-Hospital Cardiac Arrest

IMPORTANCE Severalclinicaldecisionrules(CDRs)havebeendevelopedtohelppractitionersknow when to safely terminate resuscitative efforts after in-hospital cardiac arrest (IHCA). The UN10 rule, a CDR that uses 3 intra-arrest variables, has been shown to predict a poor chance of survival to discharge. However, its large-scale applicability in clinical settings remains unknown. OBJECTIVE To assess the performance of a parsimonious CDR in a national cohort of individuals with IHCA. DESIGN, SETTING, AND PARTICIPANTS This retrospective cohort study used a nationwide cohort fromtheAmericanHeartAssociationGetWiththeGuidelines–ResuscitationIHCAregistrytoderive a sample of 96509 patients from 716 US hospitals who experienced IHCA from January 1, 2000, to January 26, 2016. Data analysis began in January 2018 and concluded in June 2018. EXPOSURES The UN10 rule uses 3 variables: (1) unwitnessed arrest, (2) nonshockable rhythm, and (3)noreturnofspontaneouscirculationwithin10minutesofresuscitativeefforts.TheCDRindicates futilityifall3criteriaaremet.ThisCDRwasanalyzedaccordingtotheConsolidatedHealthEconomic Evaluation Reporting Standards (CHEERS) reporting guideline.


CONCLUSIONS AND RELEVANCE
Patients who met the UN10 rule were associated with unfavorable neurologic status and low rates of survival after IHCA.Yet their survival rates are higher Findings In this cohort study of 96 509 patients with in-hospital cardiac arrest, the percentage of patients who satisfied the UN10 rule for futility and survived (6.3%) was substantially higher than the initial derivation cohort (0%) and single-center validation cohort (1.1%).
Meaning A revalidation study of the UN10 rule in a large cohort demonstrated moderate predictive ability to identify patients with poor survival; however, survival rates of patients who met all UN10 criteria were much higher than previous studies.

Introduction
2][3][4][5][6] Van Walraven et al 7,8 developed a parsimonious model incorporating 3 readily available intra-arrest variables, to identify patients with in-hospital cardiac arrest (IHCA) who have no chance of survival to discharge.This model, which we call the UN10 rule based on the 3 variables (U, unwitnessed arrest; N, nonshockable rhythm; and 10, return of spontaneous circulation [ROSC] not obtained within 10 minutes), was prospectively validated in 2181 patients at a single hospital nearly 20 years ago.While it is unclear how widely used this model currently is in clinical settings, the application of a simple CDR relying on just 3 intra-arrest variables in code settings could greatly enhance termination decisions.How it performs in a broader sample of hospitalized patients and in the context of a diverse population and contemporary resuscitation care practices remains unknown.

Data Source
We used the American Heart Association Get With the Guidelines-Resuscitation (GWTG-R) registry, a large, multicenter, prospective, observational registry of IHCA in the United States.The registry has been described in detail previously. 9Briefly, trained personnel at participating hospitals record observational data during resuscitation of IHCAs, defined as apnea, absence of central palpable pulse, and unresponsiveness.Cases are identified by available arrest flow sheets, paging system logs, medication administration records, emergency resuscitation equipment, and hospital billing sheets.
Information is standardized using Utstein definitions as developed by international experts. 9The American Heart Association provides oversight for the entire process of data collection, analysis, and reporting.A deidentified database was used for statistical analyses.The institutional review board of the University of Michigan reviewed the study protocol and determined the study was exempt.
Patient consent was waived owing to the use of a deidentified database.This study follows the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guideline. 10

Definitions
Duration of resuscitation was documented in integer minutes and was defined as the time from the onset of resuscitation to ROSC or termination of efforts when the patient was declared deceased.
Return of spontaneous circulation was defined as the restoration of a pulse for at least 20 minutes during the cardiac arrest.Shockable rhythms were defined as arrests due to pulseless ventricular tachycardia or ventricular fibrillation.Nonshockable rhythms were defined as pulseless electrical activity or asystole.We used cerebral performance category (CPC) scores to assess neurologic status of survivors at the time of discharge (1, little to no major neurologic disability; 2, moderate disability; 3, severe disability; 4, coma or vegetative state; and 5, brain death). 11In keeping with prior literature, favorable neurologic survival was defined as survival without severe neurologic disability (ie, CPC score, 1 or 2). 12

Main Outcomes
The primary outcome of this validation study was survival to discharge.A patient was predicted to have no chance of survival to discharge if all 3 of the following conditions were met: (1) unwitnessed arrest (ie, not in person or by monitor), (2) a nonshockable initial rhythm (ie, pulseless electrical activity or asystole), and (3) no ROSC within 10 minutes of starting chest compressions.As previously, these 3 variables, initially derived and validated by van Walraven et al 7,8 will be defined as the UN10 rule.Additionally, we assessed whether these 3 variables predicted survival with a favorable neurologic status.Because some percentage of survivors had missing information on CPC scores at discharge and were assumed to be missing at random, we performed multiple imputation and pooled the results with 20 data sets.Results with and without imputation were not meaningfully different.

Study Population
We identified 197 650 patients 18 years or older with complete clinical and demographic data who experienced an index cardiac arrest at 1 of 725 study hospitals between January 1, 2000, and January 26, 2016 (Figure).After several exclusions, including 4340 individuals with cardiac arrests whose duration prior to achieving ROSC was less than 2 minutes (to ensure a veritable resuscitative effort) as well as 5355 patients who did not achieve ROSC and received less than 10 minutes of attempted resuscitation (to whom the UN10 rule would not apply), our final study population consisted of 96 509 patients with an index IHCA from 716 hospitals.A total of 2827 survivors (15.1%) had missing information on CPC scores at discharge.

Statistical Analysis
We assessed the UN10 rule's performance using a 2 × 2 contingency table to compare observed and predicted survival to discharge; sensitivity, specificity, and predictive values were calculated using 95% CIs.Receiver operator curves are not presented in the original reports of the UN10 rule owing to  b Mean (range).
e Defined as achieving ROSC for more than 1 hour.
f Defined as achieving ROSC for more than 20 minutes.

Discussion
The UN10 rule is a parsimonious CDR that demonstrated nearly perfect predictive ability to determine whether an ongoing resuscitation could be considered futile in initial studies.However, in a large contemporary cohort, the UN10 rule did not discriminate sufficiently to justify futility and discontinuation of resuscitative efforts for patients with IHCA.Given that 4.8% of patients meeting the UN10 rule had favorable neurologic survival and 6.3% survived to discharge, many patients and families may not consider resuscitative efforts futile at these levels.
Models using only intra-arrest variables to predict survival and guide resuscitative efforts remain limited.4][15] The UN10 rule appears to be unique among current CDRs in that it relies solely on intra-arrest variables, which are often readily available during code situations, and it has now been validated in a large, national cohort.
Only recently have registries collected relevant data regarding quality of life measures at the time of discharge for patients following IHCA.Many of the prior models did not incorporate neurologic status into their calculations; however, this has been shown to be very important to survivors and has been included in more recent CDRs. 1,4In the current study, incorporation of neurologically intact survival was imperative because, although survival to discharge of patients meeting the UN10 rule for futility was 6.3%, if neurologically intact survival on discharge was lower than reported in this study (4.8%), then the CDR could have been more confidently reported as a tool for practitioners to terminate resuscitative efforts.

Limitations
This study has several limitations that warrant further discussion.First, we used the GWTG-R database to validate the UN10 rule, and outcomes may differ at nonparticipating facilities.Second, the GWTG-R database is not a comprehensive data set and therefore does not include data such as the quality of chest compressions or duration of interruptions during cardiopulmonary resuscitation, which could alter results.Third, this study only assessed IHCAs with follow-up until discharge; therefore, it cannot be applied to out-of-hospital cardiac arrests and has no measure of outcomes following discharge.Fourth, treatment algorithms have rapidly changed (eg, increasing use of end-tidal carbon dioxide), survival rates for IHCA have increased, and use of palliative care practices in the creation of do not attempt resuscitation orders for patients have evolved since the original van Walraven et al 7,8 studies in 1999 and 2001, which could explain some of the differences between our study and the original cohorts. 16Also, because the GWTG-R registry is a large national sample, it is unknown if hospitals included were using the UN10 rule from the original van Walraven et al 7,8 studies for decision making following IHCA.Therefore, it is possible that this study could have artificially inflated the positive predictive value because health care professionals could have been using the rule to terminate resuscitation if a patient's indexed cardiac arrest met the CDR criteria.

Question
How does a previously developed clinical decision rule (termed the UN10 rule), designed to predict futility during in-hospital cardiac arrest, perform in a large national sample?

Figure .
Figure.Population of Patients Used to Revalidate the UN10 Rule

Table 2 .
Outcomes After In-Hospital Cardiac Arrest Stratified by UN10 Clinical Decision Rule a a The UN10 rule is based on 3 variables: (1) unwitnessed arrest (not in person or by monitor), (2) a nonshockable initial rhythm (ie, pulseless electrical activity or asystole), and (3) no return of spontaneous circulation within 10 minutes of starting chest compressions.

Table 3 .
Sensitivity Analyses Using UN10 Clinical Decision Rule a a The UN10 rule is based on 3 variables:(1)unwitnessed arrest (not in person or by monitor), (2) a nonshockable initial rhythm (ie, pulseless electrical activity or asystole), and (3) no return of spontaneous circulation within 10 minutes of starting chest compressions.