Critically Appraised Topic: Does the timing of ECG acquisition post-ROSC for out of hospital cardiac arrests matter in identifying significant coronary lesions?

Study #1- Baldi: This was a retrospective, multicenter cohort study of adults who achieved return of spontaneous circulation from out of hospital cardiac arrests. Only patients who underwent coronary angiography during hospitalization and who had multiple post-ROSC ECGs performed before the angiography were enrolled. Two cardiologists at each participating center were blinded to the angiography findings and the time from ECG to ROSC. Each ECG was classified as either diagnostic of STEMI or not diagnostic according to the 2017 European Society of Cardiology guidelines. The angiography was analyzed by analysis of the procedure report and if needed, an analysis of the angiography itself. Interpretations were divided into 4 main groups: ECG findings falsely negative & positive, and truly negative & positive for STEMI. Positivity for STEMI was defined by CAD worthy of angioplasty. ROSC to ECG timings were grouped into <7 min, 8-33 min, & 33+ min. The final group was of 370 patients, in whom 198 (54%) had a STEMI and 172 (46%) did not. In analysis of the 3 tertiles of time from ROSC to ECG, the percentage of false-positive ECG findings in the first tertile (18.5%) was significantly higher than that in the second tertile (7.2%; odds ratio [OR], 0.34; 95%CI, 0.13-0.87; P = .02) and third tertile (5.8%; OR, 0.27; 95%CI, 0.15-0.47; P < .001). These differences remained significant when adjusting for sex, age, #of segments w ST elevation, QRS, HR, #of Epi, shockable initial rhythm, and for 3+ shocks delivered. The PPV of the ECG predicting STEMI increased from 75.3% to 85.5% and 83.7% in the 1st, 2nd, and 3rd tertiles. Specificity increased from 41% to 75% and 81.6%. 

Study #2 - Sharma: This was a retrospective, observational study performed at Hennepin Healthcare System to look at the accuracy of initial v. follow up ECGs performed after ROSC. Patient included were only those who presented with out of hospital cardiac arrest in VT/VF, who attained ROSC, and underwent coronary angiography during hospitalization. All other cardiac arrests were excluded. Initial ECG and a 2nd ECG and undisclosed time intervals were analyzed by two “experienced readers.” The sample size included 152 patients, 64 of which (42%) had intervenable coronary lesions vs 88 (58%) did not. Of the 152 patients enrolled, only 102 had follow up ECGs with a mean time of 71 minutes. The data for ECGs being able to correctly identify intervenable coronary lesions of those 102 patients are as follows: ECG 1 vs 2 - sensitivity 63% vs 50%, specificity 81% vs 91%, PPV 61% vs 73% and NPV 83% vs 80%. ECG 1 accurately predicted coronary lesions with OR 7.3 (95% CI: 2.9 - 18.6, p<0.0001), whereas ECG 2 had an OR 10.7 (95% CI 3.6 - 31.6, p <0.0001). 


Both studies were released within 6 months of each other and are rather the first of their kind. Baldi has more comprehensive data sets within the study and has more organized data collection parameters. Sharma, although not discussed in detail above, provided a look at other metabolic factors that may affect ECG tracing and concluded that regardless of the metabolic derangement seen in cardiac arrest, it did not affect the ECG’s ability to predict a significant coronary lesion. The studies have some similarities and some differences. ECG interpretation was done by cardiologists specifically in the Baldi study, and by “experts” in the Sharma study, which questions who exactly interpreted it. Albeit they did both include blinding to clinical context and a 3rd interpreter in the case of split opinion on the ECG. The ECG interpretation guidelines Baldi used were specifically the 2017 European Society of Cardiology, a reputable evidence-based source, whereas Sharma had a very subjectively described criteria in the study of unclear origin. A stark difference in the European vs American studies include the definition of “worthy of coronary intervention.” That still remains a heavily debated topic in the field of cardiology but would be a major area of inconsistency within both studies. In regard to the interpretation of the angiography, Baldi’s study had both angiography reports created by cardiologists who knew the clinical picture and ECG findings, in addition to blinded interpreters of the angiography which is an inconsistency as well. These studies are far from perfect, but nonetheless, they provide valuable information and areas of potential research in the future.  


Conclusions: These have been the first to investigate whether the timing of the ECG after ROSC from OHCA was associated with ECGs meeting STEMI criteria. Furthermore, Baldi went into whether or not each catheterization had an intervention on as well. It is important to consider the impact and the applicability of these findings. ED providers with interventionalists available in-house have much less of a decision-making burden than centers where transfer is required. Based on my literature review, I see minimal harm on obtaining an initial ECG. In a patient that had just underwent cardiac arrest, the pros of early and serial ECGs far outweigh the cons, regardless if the patient has further intervention. Sharma points out a strong point in that even if one would like to wait for a second ECG, ignoring the first one may have detriment to the patient, and I agree. The overall takeaway of this literature review would be serial ECGs at 0 & 8 minutes may be useful in clarifying what may initially appear as a STEMI