Termination of Resuscitation for In-hospital Cardiac Arrest (EIT NEW): Systematic Review

Conflict of Interest Declaration

The ILCOR Continuous Evidence Evaluation process is guided by a rigorous ILCOR Conflict of Interest policy. No Task Force members and other authors were recused from the discussion due to conflict of interest. The following Task Force members and other authors declared an intellectual conflict of interest and this was acknowledged and managed by the Task Force Chairs and Conflict of Interest committees: (none)

CoSTR Citation

Lauridsen KG, Baldi E, Smyth M, Cheng A, Bhanji F, Bigham BL, Bray JE, Breckwoldt J, Duff JP, Gilfoyle E, Hsieh MJ, Iwami T, Lockey AS, Ma M, Monsieurs KG, Okamoto D, Pellegrino JL, Yeung J, Finn J, Greif R. - on behalf of the International Liaison Committee on Resuscitation Education, Implementation and Teams Task Force.

Termination of Resuscitation for In-hospital Cardiac Arrest Consensus on Science with Treatment Recommendations. International Liaison Committee on Resuscitation (ILCOR) Education, Implementation and Teams Task Force, 2020, January 7. Available from: http://ilcor.org

Methodological Preamble

The continuous evidence evaluation process for the production of Consensus on Science with Treatment Recommendations (CoSTR) started with a scoping review that was advanced to a systematic review conducted by ILCOR Education, Implementation and Team (EIT) Task Force under the lead of Lauridsen, with involvement of clinical content experts (Baldi, Greif, Smyth). Evidence was sought and considered by the EIT Task Force.

We searched for studies investigating any clinical decision rule to terminate in-hospital resuscitation. We defined clinical decision rules as cardiac arrest characteristics to be applied during resuscitation to predict survival (return of spontaneous circulation, survival to hospital discharge) and thereby terminate resuscitation if deemed futile. Measures of prediction were negative predictive value, sensitivity, specificity, positive predictive value.

We found three studies investigating the usability of the UN10 rule to predict survival to hospital discharge based on unwitnessed arrest, a non-shockable rhythm, and 10 minutes of cardiopulmonary resuscitation without return of spontaneous circulation. All studies were cohort studies and no studies utilized randomization or prospective implementation of a clinical decision rule.

When it was possible, we extracted all prediction endpoints from the studies. We streamlined reporting of outcomes as true positives being correct prediction of death and true negatives being correct prediction of survival. Accordingly, the aim of any clinical decision rule would be to have a perfect- or near perfect positive predictive value (lower end of 95% confidence interval >99%).

PICOST

The PICOST (Population, Intervention, Comparator, Outcome, Study Designs and Timeframe)

Population: Adults and children with in-hospital cardiac arrest

Intervention: Use of any clinical decision rule

Comparators: No clinical decision rule

Outcomes: No return of spontaneous circulation, death before hospital discharge, survival with unfavorable neurological outcome, death within 30 days.

Study Designs: Randomized controlled trials (RCTs) and non-randomized studies (non-randomized controlled trials, interrupted time series, controlled before-and-after studies, cohort studies) were eligible for inclusion. Unpublished studies (e.g., conference abstracts, trial protocols), animal studies, simulation studies, and studies not in English were excluded.

Timeframe: All years until November 11, 2019.

Prospero Registration: Submitted on January 7^th 2020

Consensus on Science

For the critical outcome of positive predictive value (PPV) in predicting death before hospital discharge for adults with in-hospital cardiac arrest, we identified very low quality evidence from three historical cohort studies {Van Walraven 1999 129; Van Walraven 2001 1602; Petek 2019 e194941} investigating the UN10 rule (downgraded for risk of bias, indirectness, imprecision, and inconsistency). Due to clinical heterogeneity in study cohorts no meta analysis was conducted. Reported positive predictive values were 100% (95% CI: 97.1-100%) {Van Walraven 1999 129}, 98.9% (95% CI: 96.5%-99.7%) {Van Walraven 2001 1602} and 93.7% (95% CI: 93.3%-94.0%) {Petek 2019 e194941}

For the important outcome of specificity in predicting death before hospital discharge, we identified very low quality evidence from three historical cohort studies {Van Walraven 1999 129; Van Walraven 2001 1602; Petek 2019 e194941} investigating the UN10 rule (downgraded for risk of bias, indirectness, imprecision, and inconsistency). Due to clinical heterogeneity in study cohorts no meta analysis was conducted. Reported specificities were 100% (95% CI: 97.1%-100%) {Van Walraven 1999 129}, 99.1% (95% CI: 97.1%-99.8%) {Van Walraven 2001 1602}, and 94.6% (95% CI: 94.3%-94.9%) {Petek 2019 e194941}.

For the critical outcome of sensitivity in predicting death before hospital discharge, we identified very low quality evidence from three historical cohort studies {Van Walraven 1999 129; Van Walraven 2001 1602; Petek 2019 e194941} investigating the UN10 rule (downgraded for risk of bias, indirectness, imprecision, and inconsistency). Due to clinical heterogeneity in study cohorts no meta analysis was conducted. Reported sensitivities were: 12.2% (95% CI: 10.3%-14.4%) {Van Walraven 1999 129}, 14.4% (95% CI: 12.4%-16.0%) {Van Walraven 2001 1602}, and 19.1% (95% CI: 18.8%-19.3%) {Petek 2019 e194941}.

For the important outcome of negative predictive value in predicting death before hospital discharge, we identified very low quality evidence from three observational studies {Van Walraven 1999 129; Van Walraven 2001 1602; Petek 2019 e194941} investigating the UN10 rule (downgraded for risk of bias, indirectness, imprecision, and inconcistency). Due to clinical heterogeneity in study cohorts no meta analysis was conducted. Reported negative predictive values were 10.8% (95% CI: 8.9-12.8%) {Van Walraven 1999 129}, 17.0% (95% CI: 15.3-18.7) {Van Walraven 2001 1602}, and 22.0% (95% CI: 21.9%-22.0%) {Petek 2019 e194941}.

For the important outcomes of positive predictive value, specificity, sensitivity, and negative predictive values in predicting survival to hospital discharge with unfavorable neurological outcome, we identified very low quality evidence from one observational study {Petek 2019 e194941} investigating the UN10 rule (downgraded for risk of bias, indirectness, and imprecision). This study reported a positive predictive value of 95.2% (95% CI: 94.9%-95.6%), a specificity of 95.3% (95% CI: 95.0%-95.6%), a sensitivity of 18.8% (95% CI: 18.5%-19.0%), and a negative predictive value of 19.1% (95% CI: 18.8%-19.3%) {Petek 2019 e194941}.

We identified no studies investigating prediction of no return of spontaneous circulation or death within 30 days.

Treatment Recommendations

We did not identify any clinical decision rule that was able to reliably predict death following in-hospital cardiac arrest.

We recommend against use of the UN10 rule as a sole strategy to terminate in-hospital resuscitation (strong recommendation, very low quality of evidence).

Justification and Evidence to Decision Framework Highlights

In making this recommendation, the EIT task force considered the following:

• Several other scores have been developed aiming at predicting the chance of surviving based on pre-arrest factors only including the GO-FAR score {Ebell 2013 1872} and comorbidity scores {Ebell 1997 171}. While these scores may be suitable to trigger do-not-resuscitate discussions, they are not aimed at deciding when to terminate resuscitation during a resuscitation attempt and were therefore not included in this review.
• We identified the Resuscitation Predictor Scoring Scale {Cooper 2003 6} aiming to identify patients with low likelihood of surviving a cardiac arrest after 15 minutes of resuscitation. This score was not included in the review as the score aimed at identifying patients with low likelihood but not patients with no likelihood of surviving the cardiac arrest.
• Several studies (primarily pre-hospital) have looked at other factors such as end-tidal CO2 and echocardiographic findings to terminate resuscitation. These have been included in reviews by the ILCOR advanced life support task force, and end-tidal CO2 and echocardiographic findings may be considered together with other factors to decide when to terminate in-hospital resuscitation.
• All identified studies were based on historical cohorts and carry a risk of a self-fulfilling prophesy bias as clinicians may have terminated resuscitation on patients who potentially had a chance of surviving in the observed studies. Prospective studies are needed in order to reliably assess the effect of such clinical decision rules.
• Two of the included studies {Van Walraven 1999 129} and {Van Walraven 2001 1602} included patients resuscitated in the 1980s and 1990s, where resuscitation practices differed from present time and where reported survival rates were lower compared to present time {Benjamin 2018 e67}. The third study {Petek 2019 e194941} included patients resuscitated between 2000 and 2016 but a large percentage of the arrests occurred before 2010. As previously stated, survival rates are now higher than previous decades.
• The task force prioritized a perfect positive predictive value (no survivors predicted to be dead) for any clinical prediction rule due to the risk of terminating resuscitation on a patient who could have survived.
• The task force discussed that it is reasonable not to terminate resuscitation as long as the patient has a shockable rhythm. No single clinical factor or no single decision rule has been identified as sufficient to terminate resuscitation. Therefore, the EIT T/F members suggested that a decision to terminate an IHCA resuscitation should continue to be based on a combination of factors that are known to be associated with a low chance of survival, e.g. end-tidal CO2, cardiac stand-still on echocardiography, duration of resuscitation, patient age, and patient comorbidities.

Knowledge Gaps

We identified several knowledge gaps in the published literature.

• There are no clinical decision tools to predict the absence of return of spontaneous circulation during in-hospital resuscitation.
• There are clinical decision tools that combine existing decision tool elements such as resuscitation duration and cardiac arrest rhythm with e.g. end-tidal CO2 and/ or findings on cardiac ultrasound.
• No studies were found on the use of clinical decision tool to terminate resuscitation for pediatric in-hospital cardiac arrest.
• There is a lack of prospective clinical validation studies and randomized trials investigating the use of a clinical decision tool to terminate resuscitation during in-hospital cardiac arrest.
• It is unknown how the use of a clinical decision tool affects resuscitation practices, cost-benefit, or how it affects survival outcomes.

Attachments

Evidence-to-Decision Table: EIT_NEW-TOR-for-IHCA_TF-SR

References

Benjamin EJ, Virani SS, Callaway CW, Chamberlain AM, Chang AR, Cheng S, et al. Heart Disease and Stroke Statistics-2018 Update: A Report From the American Heart Association. Circulation 2018;137:e67–492.

Cooper, S., & Evans, C. (2003). Resuscitation Predictor Scoring Scale for inhospital cardiac arrests. Emergency Medicine Journal: EMJ, 20(1), 6–9.

Ebell, M. H., Jang, W., Shen, Y., & Geocadin, R. G. (2013). Development and validation of the Good Outcome Following Attempted Resuscitation (GO-FAR) score to predict neurologically intact survival after in-hospital cardiopulmonary resuscitation. JAMA Internal Medicine, 173(20), 1872–1878.

Ebell MH, Kruse JA, Smith M, Novak J, Drader-Wilcox J. Failure of three decision rules to predict the outcome of in-hospital cardiopulmonary resuscitation. Med Decis Making 1997;17:171–7.

Petek, B. J., Bennett, D. N., Ngo, C., Chan, P. S., Nallamothu, B. K., Bradley, S. M., Tang, Y., Hayward, R. A., van Walraven, C., & Goldberger, Z. D. (2019). Reexamination of the UN10 Rule to Discontinue Resuscitation During In-Hospital Cardiac Arrest. JAMA Network Open, 2(5), e194941.

van Walraven, C., Forster, A. J., Parish, D. C., Dane, F. C., Chandra, K. M., Durham, M. D., Whaley, C., & Stiell, I. (2001). Validation of a clinical decision aid to discontinue in-hospital cardiac arrest resuscitations. JAMA, 285(12), 1602–1606.

van Walraven, C., Forster, A. J., & Stiell, I. G. (1999). Derivation of a clinical decision rule for the discontinuation of in-hospital cardiac arrest resuscitations. Archives of Internal Medicine, 159(2), 129–134.