Ultrasound (US) is the use of high-frequency sound waves that create echoes when interfacing with underlying organs and tissue structures (UKAS 2008). The echoes are received and displayed as a visual image to the operator. US has been researched and implemented within pre-hospital care from the early 1990s within aeromedical transfers, and has been established as a point-of-care diagnostic tool over the past five years in doctor-led systems (Brun et al, 2013). The development of portable technology has enabled US scanning to become part of initial patient consultations. This review seeks to research its value and potential for early management and prediction for survival in out-of-hospital cardiac arrest (OHCA) in the medical patient compared with existing cardiac arrest markers.
OHCA management can be difficult; decision-making is often multi-factorial, and existing practices are unreliable and of varying sensitivity (Kern et al 2008). For example, it is well documented that an asystolic Electro-Cardiogram (ECG) may actually be Ventricular Fibrillation (VF) when seen with US (Kern et al 2008). End Tidal Carbon Dioxide (ETCO2) values do not give a reflection of long-term survival (Heradstveit et al, 2012; Touma and Davies, 2013). This prompted the question of how sensitive the current examination technique was and how it might change practice and decision-making within this cohort of patients.
Search strategy
Eight studies were evaluated to review the use of pre-hospital ultrasound. The OHCA studies reviewed evaluate the use of US both in and out-of-hospital. The evidence was reviewed using the Critical Appraisal Skills Programme framework (CASP 2013). The initial search was conducted from 2005 in order to expand the literature search. This was due to limited rigorous level 2a (Systematic review of prospective cohort studies) and no meta-analysis (or level 1 systematic review with homogenous Randomised Controlled Trials [RCTs]) within the last 5 years. The five databases used were set to a 10-year limit to achieve the strongest (2a) and most contemporaneous literature available on US using electronic databases (PubMed, Medline, Ovid, Cochrane and Embase).
To minimise bias, we adhered to inclusion criteria (Table 1) that aimed to capture the highest quality empirical evidence without influencing the conclusions of the review. As OHCA is a more frequent presentation in adults, we included a lower threshold of 18 years for this review. The search strategy is presented in Table 1.
| Keywords |
|
| Search terms |
|
| Limitations |
|
| Inclusion criteria | The searches were limited to: Systematic reviews, meta analyses, RCTs, Prospective observational studies, retrospective reviews Human Studies Adult studies (>18 years old) Pre-hospital Care and Emergency Department Care |
| Exclusion criteria | Paediatric studies, review articles, editorials and letters, case reports, expert opinions |
| Screening evidence | Titles and abstracts were screened for relevance; inclusion and exclusion criteria were applied to titles and abstracts with non qualifying articles excluded. Reference lists of all included papers were reviewed with ‘backward chaining’ employed to gather pertinent papers for consideration. This found 12 additional papers. |
| Final number | Studies included in order of Empirical Evidence: 1 Systematic Review (Cochrane): |
Results
Blyth et al (2012) undertook the only systematic review into OHCA and whether it predicted Return of Spontaneous Circulation (ROSC). The searched four databases (med-line, CINAHL, Embase and PubMed) with no date limitations. This may have captured more literature, but a contemporary evidence base is fundamental in pre-hospital care with an increasingly evolving EBP guideline base (Kazdin 2008).
Blyth et al (2012) showed that US has a strong prognostic value in predicting ROSC (n=568 patients) although not a definitive marker (negative LR of 0.18). This should be interpreted with care as early initiation of the chain of survival (witnessed arrest, early CPR and early defibrillation) and agonal gasping (39% survival rate versus 9% survival rate in non-gaspers - Berger 2008) can both be predictors of survival despite US guided Ventricular Wall Motion (VWM). Many studies did not report data other than ROSC (not to hospital discharge). The clinical relevance from this systematic review indicated that use of US could not be overtly linked with long-term survival (i.e. to hospital discharge).
Breitkreutz et al (2010) undertook an in-hospital prospective trial over 5.5 years to evaluate the feasibility of Focused Echocardiographic Evaluation in Life support (FEEL) in OHCA. This was related to the occurrence of potentially treatable conditions detected, and US influence on patient management in cardiac arrest. It also contrasted clinical assessment of current practices (ECG, ETCO2, SP02, RR, HR) with FEEL/US examination for the basis of patient management.
Results showed that US performed as part of a CPR algorithm changed management 89% of the time in cardiac arrest decision-making. These results show a change in management in over 50% of the cases in each patient group (OHCA and peri-arrest). This is significant, together with the ability to differentiate true Pulseless Electrical Activity (PEA) (no VWM and true electro mechanical dissociation) from pseudo PEA (some forward flow and minimal VWM). The study also showed that US findings in the peri-arrest patient changed management 66% of the time. This also reflects the US can be used in peri-arrest patients to change management before cardiac arrest ensues.
There were limitations to this study involving a lack of clarity of exclusion criteria; this can introduce confounding bias as heterogeneous comparison between trauma patients and medical cardiac arrest patients carry a very different prognosis depending on sequelae of injury and pathology (i.e. trauma patients in cardiac arrest may have a worse prognosis than medical cardiac arrest patients). Lack of follow up to hospital discharge also limits applicability of results and external validity.
Salen et al (2005) undertook a prospective observational study of US managed non-trauma pulseless adults in cardiac arrest across 4 Emergency Departments (EDs). The study was conducted over a 12-month period with 70 subjects enrolled, (36 in asystole and 34 in PEA). This study was examining US as an adjunctive diagnostic aid and as a predictor of resuscitation outcomes in cardiac arrest patients in PEA and asystole.
The results showed that 36 patients presented in asystole and 34 in PEA. Eight subjects presenting with PEA had ROSC and survived through to the intensive care unit; all subjects in an asystolic rhythm died. Although this study is underpowered due to poor sample sizes the results reveal the terminal outcome of asystole. 68% of subjects (n=23) presenting with PEA had no evident VWM (true EMD) and did not achieve ROSC. This study shows that US differentiates the true difference between these subsets of patients (Pseudo PEA versus true PEA) unlike ECG findings, this is a clear clinical benefit of US.
Aichinger et al (2012) performed a prospective observational review of pre-hospital non-trauma cardiac arrest patients (OHCA) by physician operators (n=24). The study's aim was to evaluate US (described as Emergency Echocardiography – EE) in predicting cardiac arrest outcome by the above operators.
The results showed that cardiac stand still on US were associated with 3.1% survival to hospital admission (n=32). Forty percent of patients with VWM on the initial US survived to hospital admission (p = 0.008). This is statistically significant showing the validity of US in predicting survival to admission, yet not a true reflection of survival. This infers US does not show longevity of survival past hospital admission, although with such small sample numbers (n=1) this is underpowered and not externally valid. Cardiac standstill on initial US resulted in a Positive Predictive Value (PPV) of 96.9% for death at the scene and a Negative Predictive Value (NPV) of 40%. The positive likelihood ratio was 4.2, and the negative likelihood ratio was 0.2. These results agree with Cureton et al (2009) and Hayhirst (2015) showing US is more sensitive in predicting death than survival (NPV of 100% for survival). The clinical bottom line from this study is that US is still a useful adjunct as it can reflect chances of mortality when managing an OHCA.
Due to the homogeneity of population studied (stringent inclusion and exclusion study criteria) the results contain clinical value and relevance to current pre-hospital practice. This study was also performed in the pre-hospital environment that allows easier extrapolation of results and relevant clinical applicability. Limitations included the poor sample sizes (n=32) yet this appeared to be a common theme throughout the empirical literature on US.
Cebicci et al (2014) undertook a retrospective in-hospital review into specialist physician led US in cardiac arrest. The researchers hypothesized that US would be a sensitive predictor of
24-hour survival in cardiac arrest patients (n=410) in comparison to current management with ECG recordings.
The results showed that arrest duration before Cardio-pulmonary Resuscitation (CPR) was 9.0 (range 1-20) minutes and CPR duration was 30.0 (6-45) minutes for OHCA (n=348), while for patients who arrested in the hospital (n=62) arrest duration before CPR and CPR duration were 2 (1-5) minutes and 21 (5-60) minutes respectively. This shows that the mean response time was shorter in-hospital and down time has been shown to be associated with survival (Sherren et al, 2013). This is a confounding study factor to survival affecting results regardless of US scanning, this should have been noted in the exclusion criteria (time capped at 30 minutes) as can alter results. Despite this, 24-hour survival rates of patients who had US cardiac activity were significantly higher in all three groups (asystole, VF/VT, PEA) (p=0.001) in comparison with non-US groups. The clinical relevance of this study shows that US informs decisions in all three groups as to whether to continue resuscitation. These informed decisions identify patients with greater chances of survival compared to ECG rhythms alone.
Limitations included a relatively small sample group (n=410), with methodological bias (The retrospective nature of the study), undefined operator experience and highly variable cardiac arrest downtimes. The study also did not follow up patients beyond 24 hours to hospital discharge; this can be viewed as a more sensitive marker of survival.
Hayhirst et al (2015) undertook a retrospective review of US of in-hospital cardiac arrests (n=56) over two years, reviewing whether US can be used as a predictor of survivability.
Out of 52 cases examined, 43 of these were medical cardiac arrests; no sub-group analysis was undertaken on this group separating it out from trauma cardiac arrests, this introduces confounding bias. Ultrasound detection of VWM had a PPV of 55% for predicting ROSC. The absence of VWM on ultrasound had an NPV of 97% for death. NPV of this cohort is more of an indicator (specificity) than PPV (sensitivity) as the cohort had a small sample size.
The clinical relevance of this study is one that agrees with Aichinger et al (2012); that of US being more specific in predicting death with regards to no VWM than to survival. This is still a useful examination and has predictive value yet more so in death than survival. Limitations of this study included poor study recruitment with no exclusion criteria reflecting high levels of confounding within the methodology.
Discussion
The literature reflects a common theme; that US is a sensitive marker of mortality rather than of survival amongst medical OHCA patients. Blyth et al (2012) indicates that US can be an immediate indicator of ROSC if used in conjunction with other tools and information (such as ETCO2, ECG, OHCA down-time, by-stander CPR, agonal respirations). The culmination of these results shows that US is more specific at identifying true negative results (NPV, 100% - Aichinger et al, 2012) of VWM and therefore a more sensitive marker of mortality than of survivability (Blyth et al, 2012; Cureton et al, 2012; Hayhirst et al, 2015).
Key themes of limitations in the literature include minimal inclusion and exclusion criteria within the methodologies; this introduces confounding bias and a lack of rigor throughout the results. Those studies with stricter criteria and longer study durations to recruit patients had more sensitive results with narrower confidence intervals (Breitkretz et al 2010, Cebicci et al 2014). None of the studies examined paramedic US performance, instead focused on physician led US studies both pre-hospital and in-hospital or radiographer performance and interpretation. This makes extrapolation of results difficult due to disparity of training and exposure to cardiac arrest amongst these cohorts of staff. This reveals a significant gap in empirical research as no paramedic or critical care paramedic scheme has published work on US within adult medical OHCA in the pre-hospital environment.
The results from the initial literature search indicated insufficiently powered research in the OHCA US literature. This was a direct reflection of short study durations over single sites. Multi-centered trials over longer durations would improve rigor, power and external validity of the results. It was not expected that US would show stronger association with mortality than survival. It was also surprising to find an absence of paramedic or critical care paramedic studies; this reflects the infancy of the profession in evidence-based research on this topic. Limitations included the heterogeneity of evidence and lack of long-term survival data. Therefore, US cannot be substantiated as a reliable independent prognostic tool for survival but rather as a tool kit approach and a real time marker of VWM in the cardiac arrest patient.
Conclusion
The overall value of US appears to be limited by the current literature. Stronger powered RCT level research would be beneficial in OHCA. US evidence is too weak to substantiate survival to discharge, or even to intensive care (Schuster et al, 2009; Cureton et al, 2012; Hayhirst et al, 2015). Despite this, there is clinical value and sensitivity in the examination over other pre-hospital examinations such as ECG recordings (Schuster et al 2009) in the differentiation of pseudo and true PEA. To this end US can be considered as part of a diagnostic tool kit for OHCA in relation to rhythm differentiation and indication of mortality.
Some limitations of this review were that it did not address the training requirement for US. This is fundamental to attaining the correct view for interpretation (Booth et al, 2014). It is important to acknowledge that sufficient VWM is not associated with good outcome to discharge. Therefore, US only gives accurate real-time indication of output and blood flow (Cebicci et al, 2014) together with the fore-mentioned markers above. Guidance for practice can be seen in Table 2.
|
|
Note: Primary underlying pathology may require addressing regardless of VWM depending on point
1. The scan can be repeated every rhythm check (5th cycle of CPR) for real-time indication of kinetic VWM in association with ETCO2 and ECG.