If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Center for Resuscitation Science, Department of Emergency Medicine, Beth Israel Deaconess Medical Center, Boston, MA, United StatesDivision of Pulmonary, Critical Care, and Sleep Medicine, Beth Israel Deaconess Medical Center, Boston, MA, United States
To conduct a prognostic factor systematic review on point-of-care echocardiography during cardiac arrest to predict clinical outcomes in adults with non-traumatic cardiac arrest in any setting.
Methods
We conducted this review per PRISMA guidelines and registered with PROSPERO (ID pending). We searched Medline, EMBASE, Web of Science, CINAHL, and the Cochrane Library on September 6, 2019. Two investigators screened titles and abstracts, extracted data, and assessed risks of bias using the Quality in Prognosis Studies (QUIPS) template. We estimated prognostic test performance (sensitivity and specificity) and measures of association (odds ratio). Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology evaluated the certainty of evidence.
Results
In total, 15 studies were included. We found wide variation across studies in the definition of ‘cardiac motion’ and timing of sonographic assessment. Most studies were hindered by high risks of bias from prognostic factor measurement, outcome measurement, and lack of adjustment for other prognostic factors. Ultimately, heterogeneity and risk of bias precluded meta-analyses. We tabulated ranges of prognostic test performance and measures of association for 5 different combinations of definitions of ‘cardiac motion’ and sonographic timing, as well as other miscellaneous sonographic findings. Overall certainty of this evidence is very low.
Conclusions
The evidence for using point-of-care echocardiography as a prognostic tool for clinical outcomes during cardiac arrest is of very low certainty and is hampered by multiple risks of bias. No sonographic finding had sufficient and/or consistent sensitivity for any clinical outcome to be used as sole criterion to terminate resuscitation.
EuReCa ONE-27 Nations, ONE Europe, ONE Registry: a prospective one month analysis of out-of-hospital cardiac arrest outcomes in 27 countries in Europe.
A bedside test to prognosticate clinical outcomes during cardiac arrest resuscitation is a desirable clinical tool.
In addition to screening for evidence of specific etiologies of cardiac arrest, point-of-care echocardiography is increasingly used as a decision aid for termination of resuscitation: the absence of cardiac motion is associated with the absence of return of spontaneous circulation (ROSC).
However, the potential for misinterpretation is under-recognized. For example, prognostic tests that influence clinical care or are utilized within clinical decisions to terminate resuscitation are highly susceptible to bias from ‘self-fulfilling prophecy’, in which clinicians involved with the decision to terminate resuscitation are not blinded to the results of the test in question. Additionally, the timing of point-of-care echocardiography during the course of resuscitation likely influences its ability to successfully predict clinical outcome.
Given the widespread incorporation of point-of-care echocardiography into current clinical practice, a comprehensive and rigorous summary of its intra-arrest prognostic capabilities would provide valuable information to both the resuscitation science community and treating clinicians. Our aim was to perform a prognostic factor systematic review on point-of-care echocardiography during cardiac arrest to inform the 2020 update to international resuscitation guidelines.
Methods
Protocol and registration
The protocol for the current study was prospectively submitted to the International Prospective Register of Systematic Reviews (PROSPERO) on October 2, 2019 (ID pending) and is provided in Supplementary Appendix. This systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.
The PRISMA checklist is provided in Supplementary Appendix. This review was conducted by the Advanced Life Support Task Force of the International Liaison Committee on Resuscitation (ILCOR).
Eligibility criteria, outcomes, and definitions
The study question was framed using the PICOST (Population, Intervention, Comparator, Outcome, Study Design, Timeframe) format: In adults in any setting in non-traumatic cardiac arrest (P), does a particular finding on point-of-care echocardiography during CPR (I), compared to the absence of that finding or a different finding on point-of-care echocardiography during CPR (C) prognosticate clinical outcomes (O). Human randomized and non-randomised studies (both prospective and retrospective), prognosis studies based on RCT data, and case-control studies were eligible for inclusion. Animal studies, ecological studies, case series, case reports, narrative reviews, editorials, comments, letters to the editor, and unpublished studies (e.g., conference abstracts, trial protocols) were excluded (S). There were no limitations on publication period or manuscript language, provided there was an English abstract (T).
The ILCOR Advanced Life Support Task Force prioritized the clinical outcomes ROSC (important), survival to hospital admission (important), survival to hospital discharge (critical), favorable neurologic outcome at hospital discharge (critical), survival beyond hospital discharge (critical), and favorable neurologic outcome beyond hospital discharge (critical).
Point-of-care echocardiography encompassed all means of sonographically viewing of the heart during CPR across the spectrum of sonographic modalities: transthoracic sonography with a multi-purpose bedside ultrasound, formal transthoracic echocardiography, and transesophageal echocardiography. We expected a priori that most prognostic factors would center primarily around ‘cardiac activity’, indicating spontaneous myocardial/valvular contraction or movement. We anticipated this could be variably defined across studies along a spectrum of observed degree of ‘cardiac activity’.
Literature search
After collaboratively developing the search strategy (Supplementary Appendix) to capture each component of the PICO question, an information specialist searched the following electronic bibliographic databases on September 6, 2019: Medline, EMBASE, Web of Science, CINAHL, and the Cochrane Library. We also reviewed the references of both included studies and identified systematic reviews pertinent to this topic.
Study selection
Two investigators, using pre-defined screening criteria, independently screened all titles and abstracts retrieved by the systematic search. After resolving disagreements regarding inclusion and exclusion of articles by discussion or adjudication with a third investigator, they independently reviewed the articles retained for full-text assessment. Disagreements regarding eligibility were resolved by discussion. We calculated Kappa statistics for inter-rater agreement during screening and final inclusion.
Data collection
Two investigators used a pre-defined and piloted standardized data tool to independently extract data pertinent to the PICOST question. These data elements were driven by the Checklist for Critical Appraisal and Data Extraction for Systematic Review of Prediction Modeling Studies (CHARMS-PF) checklist (Supplementary Appendix)
for critical appraisal and data extraction for systematic review of prognostic factors. Discrepancies in the extracted data were identified and resolved via discussion.
Bias assessment
Two investigators independently reviewed the risk of bias of individual studies and disagreements were resolved via discussion. We used the Quality in Prognosis Studies (QUIPS) template to assess risk of bias across six domains: study participation, study attrition, prognostic factor measurement, outcome measurement, study confounding, and statistical analysis/reporting.
The signaling questions and criteria used to rate risk of bias are in Supplementary Appendix. QUIPS contains similar elements to Quality Assessment of Diagnostic Accuracy Studies Version 2 (QUADAS-2), which is used for diagnostic test accuracy systematic reviews.
We also considered industry sponsorship as a potential source of bias.
In addition to this standardized risk of bias assessment, we especially considered two sources of bias related to prognostication during resuscitation of cardiac arrest. First, ‘self-fulfilling prophecy’, when clinicians involved with the decision to terminate resuscitation are not blinded to the results of point-of-care echocardiography, was a key consideration when reviewing studies for risk of bias. We considered this a critical risk of bias that precluded pooling studies. Operationally, we determined a priori that this would include studies with point-of-care echocardiography performed immediately prior to termination of resuscitation, studies in which clinicians were not blinded to sonographic findings, or studies with other evidence of self-fulfilling prophecy. Second, the timing of point-of-care echocardiography during resuscitation was another key confounder since restoring cardiac motion is a primary goal of resuscitative therapies. For example, resuscitative interventions could lead to the restoration of cardiac motion, or cardiac motion could cease over the course of an unsuccessful resuscitation. The timing of a prognostic test assessing cardiac motion could artificially improve or lower its prognostic estimates.
Data analysis and synthesis
Although the Cochrane Prognosis Working Group recommends estimating odds ratios or risk ratios in a prognostic factor systematic review,
the binary nature of the clinical outcomes lends itself well to consideration in a standard 2 × 2 tabular format. While this is not a systematic review of diagnostic test accuracy, elements of test performance have clinical applications in the prognostication of clinical outcomes (i.e. ROSC) with a bedside tool (i.e. point-of-care echocardiography). In addition, we believe that consideration of the true- and false-positive rates of point-of-care echocardiography is more useful to clinicians than traditional measures of association (e.g. odds ratio). Nonetheless, we provide both estimates of test performance and traditional measures of association for interpretation. Test positive denotes presence of the sonographic finding in question (e.g. cardiac motion). Disease positive denotes presence of the clinical outcome in question (e.g. ROSC).
Studies were assessed for clinical, methodological, and statistical heterogeneity. A p-value of <0.10 or I-squared statistic of >50% indicated substantial statistical heterogeneity.
Sufficiently homogenous studies without critical risk of bias were eligible for pooling with a random effects meta-analysis, as per the Cochrane Collaboration Prognosis Working Group, since unexplained heterogeneity is likely to remain in prognostic factor systematic reviews.
We planned the following a priori subgroups: witnessed vs. unwitnessed collapse, shockable vs. nonshockable initial cardiac rhythm, and in-hospital vs. out-of-hospital cardiac arrest.
Using guidance documents from the Cochrane Prognosis Methods Group,
we assessed the certainty of the overall evidence using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology ranging from very low certainty of evidence to high certainty of evidence.
We used GRADEpro software (McMaster University, 2014) to tabulate detailed assessment of overall risk of bias, inconsistency, imprecision, and indirectness.
Results
Study selection
The search identified 2606 unique titles and abstracts, of which 2575 were excluded after initial review (Kappa 0.75) (Fig. 1). After reviewing 31 full-text articles for eligibility, an additional 16 were excluded leaving 15 manuscripts for inclusion (Kappa 1.0). All included studies were observational in nature.
Fig. 1Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) diagram illustrating the selection of articles.
Altogether, 15 observational studies enrolled 2091 subjects between 1999–2017 (four studies did not specify years of enrollment) and were published between 1997–2019, of which 10 were conducted in the Emergency Department, two in the prehospital setting, two in the inpatient setting, and one with mixed settings of enrollment.
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
Emergency Ultrasound Working Group of the Johann Wolfgang Goethe-University Hospital, Frankfurt am Main. Focused echocardiographic evaluation in life support and peri-resuscitation of emergency patients: a prospective trial.
There were seven studies from North America, five from Asia, two from Europe, and one from South America. Most (14/15) studies utilized the subxiphoid view, nine utilized the apical 4-chamber view, nine utilized a parasternal view, and one utilized transesophageal views as needed. There was wide variability between studies in the reported training and credentials of sonographers (Table 1).
Table 1Characteristics of included studies. US: ultrasound. ROSC: return of spontaneous circulation. TOR: termination of resuscitation. OHCA: out of hospital cardiac arrest. ED: emergency department. ICU: intensive care unit. IHCA: in hospital cardiac arrest. Min: minutes. PEA: pulseless electrical activity. DNR: do not resuscitate. FEEL: focused echocardiographic evaluation in life support.
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
Emergency Ultrasound Working Group of the Johann Wolfgang Goethe-University Hospital, Frankfurt am Main. Focused echocardiographic evaluation in life support and peri-resuscitation of emergency patients: a prospective trial.
Emergency physicians (senior resident or above) that passed training course (lecture, hands-on, simulation, live patients, multiple choice test, approved live scans)
Upon reviewing the included articles, we discovered wide variability in the definitions of ‘cardiac motion’ pertaining to anatomy (i.e. left ventricular contractions with associated valvular opening, myocardial contractions, any ventricular movement, any myocardial movement, any movement [including isolated valvular fluttering], or unspecified) and timing (initial, every, any, or subsequent point-of-care echocardiogram; or unspecified) (Supplementary Appendix). Ultimately, we classified studies describing cardiac motion as organized contractility vs. non-organized and/or unspecified motion. We classified studies describing echocardiogram timing as the initial echocardiogram, every echocardiogram, any echocardiogram, a subsequent echocardiogram, or unspecified timing. We collated sonographic evidence of treatable pathology (evidence of hypovolemia, pericardial effusion, cardiac tamponade, or right ventricular dilation) into one category. Two studies report multiple sonographic findings within a given category on the same subjects.
To collate complete data, these are tabulated as the composite variable ‘subject-assessments’. Finally, we report other described miscellaneous sonographic findings.
Bias assessment
Studies tended to have high risks of bias related to prognostic factor measurement, outcome measurement, and lack of adjustment for other prognostic factors (Table 2). Notably, prospective studies either enrolled a convenience sample of subjects or did not specify consecutive subject sampling. No study specified if outcome assessors were blinded to sonographic findings. No industry sponsorships were identified.
Table 2Risk of bias assessment using the Quality In Prognostic Factor Studies (QUIPS) rubric.
Treating clinicians not blinded to ultrasound findings, but either a protocolized delay between ultrasound and termination of resuscitation, or granular data indicate lack of self-fulfilling prophecy.
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
Emergency Ultrasound Working Group of the Johann Wolfgang Goethe-University Hospital, Frankfurt am Main. Focused echocardiographic evaluation in life support and peri-resuscitation of emergency patients: a prospective trial.
Treating clinicians not blinded to ultrasound findings, but either a protocolized delay between ultrasound and termination of resuscitation, or granular data indicate lack of self-fulfilling prophecy.
Treating clinicians not blinded to ultrasound findings, but either a protocolized delay between ultrasound and termination of resuscitation, or granular data indicate lack of self-fulfilling prophecy.
Treating clinicians not blinded to ultrasound findings, but either a protocolized delay between ultrasound and termination of resuscitation, or granular data indicate lack of self-fulfilling prophecy.
No measures of association for clinical outcome beyond survival to hospital admission.
a Enrolled convenience sample of subjects.
b Treating clinicians not blinded to ultrasound findings, but either a protocolized delay between ultrasound and termination of resuscitation, or granular data indicate lack of self-fulfilling prophecy.
c Unspecified if outcome assessors blinded to ultrasound findings.
d Clinical outcomes beyond return of spontaneous circulation not adjusted for other prognostic factors.
e No measures of association for clinical outcome beyond survival to hospital admission.
f Granular data indicate presence of self-fulfilling prophecy.
g Treating clinicians not blinded to ultrasound findings.
h Unspecified timing of ultrasound in relation to termination of resuscitation.
i Imprecise or unclear definition of ultrasound finding.
j Unclear credentials of the sonographer.
k Treating clinicians blinded to sonographic finding of ‘cardiac motion’, but were advised of other findings that might prompt specific interventions.
Two studies contained evidence of self-fulfilling prophecy. In both Atkinson et al. and Gaspari et al., subjects with cardiac motion received longer durations of CPR than those without (Atkinson et al.: 27 min vs. 12 min; Gaspari et al.: 18 min vs. 12 min).
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
Furthermore, subjects in Atkinson et al. with cardiac motion were more likely to be treated with endotracheal intubation (95% vs. 47%) and epinephrine (100% vs. 82%) than those without.
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
In Flato, et al., subjects without cardiac motion (median 12 cycles CPR) had longer durations of CPR than subjects with cardiac motion (median 6 cycles CPR), even though the treating team was not blinded to sonographic findings.
In both Kim et al. and Lien et al., all subjects received mandatory prespecified 30 min of CPR beyond sonographic assessment prior to termination of resuscitation.
The timing of sonographic assessment varied greatly between studies (Table 1 and Supplementary Appendix). Since no more than 4 studies addressed any one combination of sonographic finding and timing, we did not assess for publication bias.
Meta-analyses
Ultimately, no combination of studies assessing a particular sonographic finding with particular timing had sufficiently low risk of bias to perform meta-analyses (Table 2 and Supplementary Appendix).
Certainty of evidence
The overall certainty of evidence was rated as very low for all outcomes primarily due to risk of bias, inconsistency, or imprecision (Table 3 and Supplementary Appendix). The individual studies were at substantial risk of bias due to prognostic factor measurement, outcome measurement, adjustment for prognostic factors, or residual confounding. Because of this and a high degree of clinical heterogeneity, individual studies are difficult to interpret.
Table 3Estimated prognostic test performance and prognostic association for sonographic findings on point-of-care echocardiography during cardiac arrest to predict clinical outcomes. Certainty of evidence was assessed with the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach. CI: confidence interval. IHCA: in-hospital cardiac arrest. OHCA: out-of-hospital cardiac arrest.
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
Gaspari, et al. and Zengin, et al. report multiple sonographic findings within a given category on the same subjects; n reflects composite variable ‘subject-assessments’.
Gaspari, et al. and Zengin, et al. report multiple sonographic findings within a given category on the same subjects; n reflects composite variable ‘subject-assessments’.
Emergency Ultrasound Working Group of the Johann Wolfgang Goethe-University Hospital, Frankfurt am Main. Focused echocardiographic evaluation in life support and peri-resuscitation of emergency patients: a prospective trial.
Gaspari, et al. and Zengin, et al. report multiple sonographic findings within a given category on the same subjects; n reflects composite variable ‘subject-assessments’.
Gaspari, et al. and Zengin, et al. report multiple sonographic findings within a given category on the same subjects; n reflects composite variable ‘subject-assessments’.
Gaspari, et al. and Zengin, et al. report multiple sonographic findings within a given category on the same subjects; n reflects composite variable ‘subject-assessments’.
IHCA & OHCA Observational
0.00 to 1.00
0.84 to 0.94
0.38 to 125.00
Very low
a Studies did not report these data for all enrolled subjects; n is lower than the total of all subjects enrolled.
b Gaspari, et al. and Zengin, et al. report multiple sonographic findings within a given category on the same subjects; n reflects composite variable ‘subject-assessments’.
Presence of organized cardiac motion (unspecified echocardiogram timing)
One observational study of 49 IHCA subjects reported sensitivity (1.00; 95% CI 0.40–1.00), specificity (0.49; 95% CI 0.34–0.64), and OR (8.62; 95% CI 0.44–169.38) for survival to 180 days.
Two observational studies of 229 IHCA and OHCA subjects reported ranges of sensitivity (0.67 to 1.00), specificity (0.51 to 0.89), and odds ratio (13.60 to 16.63) for survival to hospital discharge.
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
Two observational studies of 349 OHCA subjects reported ranges of sensitivity (0.39 to 1.00), specificity (0.91 to 0.91) (identical point estimates), and odds ratio (6.73 to 414.56) for survival to hospital admission.
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
Two observational studies of 229 IHCA and OHCA subjects reported ranges of sensitivity (0.34 to 0.79), specificity (0.68 to 0.96), and odds ratio (8.07 to 13.21) for ROSC.
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
Presence of non-organized and/or unspecified cardiac motion on initial echocardiogram
One observational study of 42 OHCA subjects reported sensitivity (1.0; 95% 0.03–1.00), specificity (0.78; 95% CI 0.62–0.89), and odds ratio (10.26; 95% CI 0.39–273.09) for good neurologic outcome at hospital discharge.
Three observational studies of 1171 IHCA and OHCA subject-assessments reported ranges of sensitivity (0.06 to 0.91), specificity (0.49 to 0.94), and odds ratio (0.38 to 17.00) for survival to hospital discharge.
Four observational studies of 1295 IHCA and OHCA subject-assessments reported ranges of sensitivity (0.11 to 0.92), specificity (0.55 to 0.85), and odds ratio (0.75 to 27.56) for survival to hospital admission.
Three observational studies of 861 IHCA and OHCA subjects reported ranges of sensitivity (0.25 to 0.64), specificity (0.78 to 1.00), and odds ratio (6.33 to 16.11) for ROSC.
Presence of non-organized and/or unspecified cardiac motion on every echocardiogram
Two observational studies of 134 OHCA subjects reported ranges of sensitivity (0.50 to 0.80), specificity (0.92 to 1.00), and odds ratio (45.33 to 148.20) for survival to hospital admission.
Presence of non-organized and/or unspecified cardiac motion (unspecified echocardiogram timing)
One observational study of 49 IHCA subjects reported sensitivity (1.00; 95% CI 0.40–1.00), specificity (0.49; 95% CI 0.34–0.64), and odds ratio (8.62; 95% CI 0.44–169.38) for good neurologic outcome at 180 days.
One observational study of 70 OHCA subjects reported sensitivity (1.0; 95% 0.03–1.00), specificity (0.86; 95% CI 0.75–0.93), and odds ratio (17.00; 95% CI 0.65–446.02) for good neurologic outcome at hospital discharge.
One observational study of 177 OHCA subjects reported sensitivity (0.48; 95% CI 0.28–0.69), specificity (0.77; 95% CI 0.69–0.83), and odds ratio (3.09; 95% CI 1.29–7.37) for survival to hospital discharge.
Three observational studies of 291 OHCA subjects reported ranges of sensitivity (0.72 to 0.86), specificity (0.60 to 0.84), and odds ratio (9.14 to 14.00) for survival to hospital admission.
Emergency Ultrasound Working Group of the Johann Wolfgang Goethe-University Hospital, Frankfurt am Main. Focused echocardiographic evaluation in life support and peri-resuscitation of emergency patients: a prospective trial.
Four observational studies of 317 OHCA subjects reported ranges of sensitivity (0.62 to 1.00), specificity (0.33 to 0.98), and odds ratio (23.18 to 289.00) for ROSC.
Return of organized cardiac motion on subsequent echocardiogram
One observational study of 20 IHCA subjects reported sensitivity (0.50; 95% CI 0.01–0.99), specificity (0.79; 95% CI 0.54–0.94), and odds ratio (3.75; 95% CI 0.19–74.06) for survival to hospital discharge.
One observational study of 20 IHCA subjects reported sensitivity (0.67; 95% CI 0.22–0.96), specificity (1.00; 95% CI 0.77–1.00), and odds ratio (52.50; 95% CI 2.10–1300.33) for ROSC.
Presence of coalescent echo contrast (i.e. visible clotted intra-cardiac blood) after 20–30 min CPR
One observational study of 20 IHCA subjects reported sensitivity (0.00; 95% CI 0.00–0.84), specificity (0.45; 95% CI 0.23–0.68), and odds ratio (0.13; 95% CI 0.01–3.11) for survival to hospital discharge.
One observational study of 20 IHCA subjects reported sensitivity (0.00; 95% CI 0.00–0.46), specificity (0.21; 95% CI 0.05–0.51), and odds ratio (0.02; 95% CI 0.00–0.53) for ROSC.
Three observational studies totaling 1130 IHCA and OHCA subject-assessments reported ranges of sensitivity (0.00 to 0.15), specificity (0.89–0.98), and odds ratio (1.32 to 4.25) for survival to hospital discharge.
One observational study with 531 IHCA and OHCA subject-assessments reported ranges of sensitivity (0.03 to 0.04), specificity (0.95 to 0.99), and odds ratio (0.61 to 4.70) for survival to hospital admission.
Four observational studies totaling 317 IHCA and OHCA subject-assessments reported ranges of sensitivity (0.00 to 1.00), specificity (0.84 to 0.94), and odds ratio (0.38 to 125.00) for ROSC.
As CPR duration increased from <4 min to 14–16 min, sensitivity for ROSC increased from 0% (95% CI 0–46%) to 100% (95% CI 3–100%) and specificity for ROSC increased from 88% (95% CI 47–100%) to 100% (95% CI 16–100%). Both estimates peaked at 10–12 min, which corresponded to the greatest numbers of subjects in any given quantile: sensitivity 100% (95% CI 79–100%) and specificity 100% (89–100%). Only two studies provided estimates of inter-rater reliability for the sonographic finding under investigation (Kappa 0.93 and 0.63, respectively).
We conducted a prognostic factor systematic review of point-of-care echocardiography during resuscitation of adults with non-traumatic cardiac arrest in any setting to predict clinical outcomes. Ultimately, clinical heterogeneity and risk of bias precluded meta-analyses, the certainty of evidence was uniformly very low, and individual studies are difficult to interpret.
The most striking finding of this systematic review was the widely inconsistent definitions and terminology around sonographic evidence of cardiac motion, which included wide variation in the classification of anatomy, type of motion, and timing of point-of-care echocardiography. This finding is consistent with a recent prospective survey study conducted by Hu et al.
Among 127 emergency medicine, critical care, and cardiology physician sonographers shown sonographic video clips from a sample of 15 cases of cardiac arrest, there was only moderate agreement (Krippendorff's α 0.47) of what constituted cardiac standstill. Within subject subgroups by specialty, level of training, and self-reported sonographic skill, agreement ranged from 0.43 to 0.55. Cases with myocardial contractions but profound bradycardia, and valvular fluttering from mechanical ventilation or weak myocardial contractions generated the most disagreement. We strongly encourage the Utstein working group, the World Interactive Network Focused on Critical Ultrasound (WINFOCUS), or other ultrasound and diagnostic imaging professional societies to establish uniform definitions and terminology describing sonographic findings of cardiac activity during cardiac arrest.
Additionally, most of the identified studies suffer from high risk of bias related to prognostic factor measurement, outcome measurement, lack of adjustment for other prognostic factors, and confounding from self-fulfilling prophecy and unspecified timing of point-of-care echocardiography. The evidence supporting use of point-of-care echocardiography as a prognostic tool during cardiac arrest is uniformly of very low certainty due to these risks of bias, inconsistency, and imprecision. Clinicians should interpret sonographic findings during cardiac arrest in light of these limitations. We strongly encourage subsequent investigations of point-of-care echocardiography during cardiac arrest to employ robust methodology that mitigates risks of bias unique to prognostic factor studies, to report the precise credentials of sonographers, to report inter-rater reliability, and to report uniform timing of sonographic assessment. Given the heterogenous nature of cardiac arrest, standardizing the timing of sonographic assessment is challenging. Assessment intervals could be normalized to assorted clinical milestones such as activation of the prehospital emergency response system, arrival of prehospital personnel, or arrival to the Emergency Department.
The primary goal of prognostication during cardiac arrest is to predict clinical outcomes with both classification accuracy and certainty. Operationally, this results in continuing resuscitation efforts in patients with a possibility of survival and terminating resuscitation in futile cases. In this systematic review, we found wide variability in both the point estimates and certainty around these point estimates to prognosticate clinical outcomes. A few sonographic findings (any cardiac activity on initial assessment, return of organized cardiac activity on subsequent assessment, and evidence of treatable pathology) tended to have higher ranges of specificity for the short-term clinical outcomes of ROSC and survival to hospital admission, but the certainty of this evidence is very low. No sonographic finding had sufficient and/or consistent sensitivity for any clinical outcome to be used a sole criterion to terminate resuscitative efforts. It is generally considered more acceptable to continue resuscitation efforts that prove futile than to erroneously terminate resuscitation in a patient who would have otherwise survived. In either case, the prognostic implications of sonographic findings during cardiac arrest are at high risk of over-interpretation or providing false reassurance.
Two forthcoming studies may add to the findings of this systematic review. Javaudin, et al. propose a prospective, multicenter observational study of early point-of-care focused echocardiography as a predictive factor for absence of ROSC after out-of-hospital cardiac arrest.
Early point-of-care focused echocardiographic asystole as a predictive factor for absence of return of spontaneous circulatory in out-of-hospital cardiac arrests: a study protocol for a prospective, multicentre observational study.
Additionally, investigators from Nantes University Hospital (Nantes, France) propose a prehospital, prospective cohort study of sonographic asystole within the first minutes of chest compressions as a predictor for absence of ROSC (ClinicalTrials.gov NCT03494153).
Despite its non-invasive nature, point-of-care echocardiography is not necessarily a benign modality. Several investigations report the introduction of additional interruptions in chest compressions with a transthoracic approach.
One proposed strategy to limit this adverse effect is to record brief sonographic video clips during pulse/rhythm checks which may then be viewed and interpreted during resumption of chest compressions. Additionally, sonographers may serially assess for specific findings on subsequent pulse/rhythm checks instead of extending the duration of an individual pulse/rhythm check for a comprehensive assessment.
Implementation of the Cardiac Arrest Sonographic Assessment (CASA) protocol for patients with cardiac arrest is associated with shorter CPR pulse checks.
Point-of-care echocardiography may still have utility to diagnose treatable etiologies of cardiac arrest, to guide the optimal anatomic location for chest compressions, to suggest prudent therapies, and to intermittently assess response to resuscitative treatments. These applications are not within the scope of this particular systematic review. However, echocardiographic findings associated with treatable etiologies may not necessarily indicate the same pathology during cardiac arrest. For example, isolated right ventricular dilation is an uncertain diagnostic indicator of massive pulmonary embolism. Right ventricular dilation begins a few minutes after onset of cardiac arrest as blood shifts from the systemic circulation to the right heart along its pressure gradient.
Additionally, right ventricular dilation has been uniformly observed in a porcine model of cardiac arrest across various etiologies of hypovolemia, hyperkalemia, and primary arrhythmia.
SHoC Investigators. Is point-of-care ultrasound a reliable predictor of outcome during atraumatic, non-shockable cardiac arrest? A systematic review and meta-analysis from the SHoC investigators.
Accuracy of point-of-care focused echocardiography in predicting outcome of resuscitation in cardiac arrest patients: a systematic review and meta-analysis.
The predictive value of bedside ultrasound to restore spontaneous circulation in patients with pulseless electrical activity: a systematic review and meta-analysis.
Notably, we did not restrict the target population to subjects with a nonshockable initial cardiac rhythm but did restrict it to subjects with non-traumatic cardiac arrest. Additionally, we utilized methodology standards for a systematic review of prognostic factor studies, not diagnostic test accuracy studies; the data extraction, bias assessment, and certainty of evidence assessment tools all differ. Finally, we were more stringent than other systematic reviews in our assessments for heterogeneity and risk of bias, which ultimately precluded meta-analyses. Other systematic reviews have estimated pooled test performance and measures of association for cardiac activity and clinical outcomes. Given the inherent limitations and biases we identified in this systematic review, those pooled estimates should be interpreted with caution.
Conclusions
The evidence for using point-of-care echocardiography as a prognostic tool for clinical outcomes during cardiac arrest is of very low certainty with significant risks of bias in prognostic factor and outcome measurements, lack of adjustment for other prognostic factors, and confounding. The establishment of uniform definitions and terminology describing sonographic findings of cardiac activity during cardiac arrest would greatly facilitate the interpretation of future studies.
Conflict of interest
None of the authors declared conflicts of interest for this systematic review. Dr. Paiva was a co-author on one of the studies included in this systematic review but did not participate in the study selection or risk of bias assessment processes. He did review the findings of this systematic review and participate in Task Force discussions on the interpretation of these data as a Task Force member.
Acknowledgments
The authors extend appreciation to Dr. David Lee Osterbur, information specialist at the Harvard Countway Library of Medicine, Boston, MA, United States, for preparing and conducting the systematic searches. Additionally, Dr. Eddy Lang provided methodologic expertise.
Appendix A. International Liaison Committee on Resuscitation Advanced Life Support Task Force Collaborators
Members of the International Liaison Committee on Resuscitation Advanced Life Support Task Force who met the criteria as a collaborator include:
Lars W. Andersen
Bernd W. Böttiger
Clifton W. Callaway
Charles D. Deakin
Michael Donnino
Cindy H. Hsu
Peter T. Morley
Laurie J. Morrison
Robert W. Neumar
Jerry P. Nolan
Edison F. Paiva
Michael J. Parr
Claudio Sandroni
Barney Scholefield
Jasmeet Soar
Tzong-Luen Wang
Appendix B. Supplementary data
The following are the supplementary data to this article:
EuReCa ONE-27 Nations, ONE Europe, ONE Registry: a prospective one month analysis of out-of-hospital cardiac arrest outcomes in 27 countries in Europe.
Does point-of-care ultrasound use impact resuscitation length, rates of intervention, and clinical outcomes during cardiac arrest? A study from the sonography in hypotension and cardiac arrest in the emergency department (SHoC-ED) investigators.
Emergency Ultrasound Working Group of the Johann Wolfgang Goethe-University Hospital, Frankfurt am Main. Focused echocardiographic evaluation in life support and peri-resuscitation of emergency patients: a prospective trial.
Early point-of-care focused echocardiographic asystole as a predictive factor for absence of return of spontaneous circulatory in out-of-hospital cardiac arrests: a study protocol for a prospective, multicentre observational study.
Implementation of the Cardiac Arrest Sonographic Assessment (CASA) protocol for patients with cardiac arrest is associated with shorter CPR pulse checks.
SHoC Investigators. Is point-of-care ultrasound a reliable predictor of outcome during atraumatic, non-shockable cardiac arrest? A systematic review and meta-analysis from the SHoC investigators.
Accuracy of point-of-care focused echocardiography in predicting outcome of resuscitation in cardiac arrest patients: a systematic review and meta-analysis.
The predictive value of bedside ultrasound to restore spontaneous circulation in patients with pulseless electrical activity: a systematic review and meta-analysis.
30183-0&id=gr1.jpg){kind=link}