The ALS TF thanks all of those who have provided comments. We have taken all comments into consideration, and agree that we cannot tell from the evidence to-date whether DSED or VC is superior to the other. We have therefore amended our treatment recommendation to clarify that either strategy can be considered, while acknowledging that the certainty of evidence is slightly lower for VC than for DSED. We also agree that consideration of pad placement is important, and have discussed this in the evidence-to-decision and justification section. We agree that a future review looking specifically at how pad placement affects outcome may be important. We have also emphasized that if DSED is used, the approach used in the trial should be followed, as they did not report any defibrillator damage with their approach.
Previous Comments by viewers
Conflicts of interest: Grants
We thank the ILCOR task force for their careful review of our trial. We would like to provide some clarity regarding the trial methodology and respectfully challenge some of the rationale behind the downgrading of the quality of the evidence The task force rated down for Risk of Bias (RoB) and Imprecision. The explanation provided by the task force for rating down for RoB was that paramedics weren't blinded to the defibrillation strategy. In our opinion, this does not represent an important source of bias in the trial for the following reasons: (1) Paramedics did not choose to enroll individual patients; rather patients were enrolled if they met eligibility criteria and were allocated to a defibrillation strategy according to the cluster randomization; (2) all outcomes were objective and unlikely to be influenced by any subjective assessments; (3) our results showed that quality of CPR was excellent across groups, with no important differences in scene time, drug administration, or time to drug administration; if conscious or unconscious bias was a concern, one would expect these to be different (e.g., paramedics might spend less time on scene if standard care was thought to be inferior, or frequently crossover from standard to other interventions). The statement also notes that paramedics determined some secondary outcomes (VFT and ROSC). This is inaccurate: VFT was never determined by paramedics but was determined by blinded abstractors after review of each case; paramedics determined ROSC but this was according to standard practice and each case was independently reviewed and verified by the medical director or local site investigator. We do not believe these approaches to outcome assessment would have introduced important bias – and we would argue the task force should not have downgraded for Risk of Bias due to lack of blinding. The task force also downgraded for imprecision. We would argue the task force should not have downgraded for imprecision for the DSED vs standard comparison, but we are comfortable with the decision to downgrade for imprecision for the VC vs standard comparison. GRADE suggests an approach that relies on thresholds and CIs of the absolute effect (CI approach) as a primary criterion for imprecision rating, NOT the optimal information size approach. We therefore disagree with the stated rationale which currently says, "Optimal information size not met. Using the hypothesized survival rate with standard defibrillation in the original trial (12%) and the hypothesized absolute increase in survival of 8% for both DSED and VC, a sample size of 310 patients per study group was calculated by the authors. The actual sample size of 125 (DSED), 136 (SD) and 144 (VC) was well below this number, introducing the possibility of imprecision in the results". Note that the absolute increase in survival for DSED was actually 17% not 8%, suggesting the hypothesized effect size for DSED used in the sample size was likely an underestimate. In our opinion, the thresholds and CIs of the effect for DSED (aRR 2.21, 95% CI 1.33 to 3.67) were not close to 1 and do not suggest imprecision. We appreciate the opportunity to comment and welcome any feedback from the ILCOR task force.
Response from ALS Chair Kate Berg: Thank you Dr. Cheskes for the comments regarding your trial. Risk of bias and certainty of evidence were discussed extensively with the task force, and decisions were not easy to come to. The task force opinion is that there is a potential risk of bias in unblinded studies (unavoidable in this case) because whether the study intervention or a different intervention is used may depend on the clinician’s opinion about the intervention or their opinion about the specific patient. There was a significant amount of crossover between the 3 study treatments, which was not evenly distributed. Exactly why this occurred or how it could have influenced the results is not entirely clear, but the task force felt that this may indicate a potential source of risk of bias. Thank you for clarifying how termination of VF/VT was determined. We can make that adjustment in the relevant GRADE tables. Regarding the assessment of imprecision, we think the optimal information size was appropriate to invoke. GRADE recommends the optimal information size (OIS) as a “second, necessary criterion to consider.” The concept of optimal information size is meant to prompt downgrading for imprecision when you do not think the sample size was adequate to be a true population estimate, based on a conventional sample size calculation. Since the goal is to detect when an effect size reported in a smaller trial may not reflect the true effect size, systematic reviewers are advised to calculate a sample size based on relative risk difference of 20-30%, not on the effect size observed in the trial in question (see GRADE handbook section on imprecision 220.127.116.11, #2 under “dichotomous outcomes,” and then see Example 3). We thought the original sample size calculation conducted by your group was reasonably close to the GRADE recommendation for that calculation so relied on that for the OIS. The fact that survival increased more than your group initially hypothesized is notable and may reflect the true effect, but there is risk of overestimation of treatment effect when trials are stopped early (Liu et al, Trials 23, 747, Sept 2022; Guyatt et al, BMJ ;344:e3863, 2012), so we think determining OIS based on a more conservative sample size calculation is prudent. Overall, the task force opinion is that a single trial of approximately 130 patients per study group almost by definition cannot provide higher certainty evidence for what is a considerable practice change, and we think GRADE methodology supports that determination.
2022.12.19 21:39:47 (modified: 2022.12.21 22:04:26)
Conflicts of interest: Financial relationships
I appreciate the ILCOR task force’s prompt analysis of the DOSE VF Trial. There are a few details that I would like to submit for consideration. First, the primary endpoint of the trial showed that both vector change (VC) and DSD had a benefit over standard care. The trial stopped early, was consequently underpowered and did not compare VC and DSD, so cannot provide accurate relative assessment of the two techniques. That the CoSTR recommends DSD over VC is not consistent with these limitations and is not consistent with the authors’ conclusion in the paper. Due to the limited information available through this trial, it may be premature to draw a conclusion on the superiority or inferiority of VC and DSD Second, it is worth noting that it is likely that none of the defibrillators available today were designed for DSD, and that DSD may be an off-label use of these devices. Defibrillator damage can occur with some DSD techniques, as has been reported in the literature, but was reported not to have occurred in DOSE VF. It is appropriate you have included the statement: “If a double sequential defibrillation strategy is utilized, we suggest an approach similar to that in the available trial, with a single operator activating the defibrillators in sequence. (good practice statement)”. A second aspect of the DSD practiced in the trial is also worth reinforcing: the specific choice of anterior-lateral and anterior-posterior pad placement for the two shock vectors. Other DSD vectors and other shock timing practices may have higher risk of incapacitating the defibrillator. Lastly, I am concerned that the comment period for this complex topic is ending too early for there to have been robust comment from the public. No letters to the editor in response to the report of the single cluster-randomized study published in NEJM have even yet been published. It may be beneficial to extend the comment period and increase visibility to this draft through appropriate channels. I would like to sincerely thank the ILCOR task force for your consideration of these comments.
2023.01.19 10:01:01 (modified: 2023.01.19 10:02:58)
From ALS Chair Kate Berg: Response: Dr. Chapman, thank you very much for your comments. First, while we agree that the available data did not allow for statistical comparison of DSED and VC, it is also the case that DSED compared with SD showed statistically significant benefits for all outcomes, while this was not the case for VC compared with SD. This is why the TF came to the conclusions it did, after extensive discussion. We include some of this discussion in the evidence to decision section and have included it as a knowledge gap. We agree about the importance of the sample size being smaller than planned due to the trial stopping early, and that the effect size might be an overestimate, which is why the evidence was downgraded for imprecision. The lack of comparison between DSD and VC, however, was not as a result of stopping the study early. The study was never designed to compare DSD to VC. Second, we agree that none of the currently available defibrillators are designed for DSED. We have discussed the importance of pad placement for SD, possible concerns about defibrillator damage and the reasoning for recommending the approach (single operator activating the shock with one hand, sequentially) used in the trial as well, as we agree that these are important considerations. Third, we are happy to ask ILCOR leadership to extend the public comment period.
2023.01.25 08:35:56 (modified: 2023.01.31 04:06:09)
On a practical note, the resources both in cost to increase defibrillator availability and competency training to perform DSED would be higher than with VC. For this reason and others, a DSED to VC comparison study is needed. Until a single DSED defibrillator is manufactured a meaningful DSED to VC comparison will prove challenging.
2023.01.30 14:46:23 (modified: 2023.01.31 04:00:37)
I agree that before considering VC or DSD, proper position of the defibrillation electrodes should be optimal. Many physicians do not position the left electrode in the optimal, mid axillary position. Most place it too anteriorly. (Heames RM, Sado D, Deakin CD. Do doctors position defibrillation paddles correctly? Observational study. BMJ. 2001 Jun 9;322(7299):1393-4. doi: 10.1136/bmj.322.7299.1393. PMID: 11397743; PMCID: PMC32253.). In a study of manufacturers illustration of paddle position, it was found that none of them showed the ideal one. (Foster AG, Deakin CD. Accuracy of instructional diagrams for automated external defibrillator pad positioning. Resuscitation. 2019 Jun;139:282-288. doi: 10.1016/j.resuscitation.2019.04.034. Epub 2019 May 5. PMID: 31063839.) Training manikins also shows a sub-optimal position (https://laerdal.com/ca/informa...). AHA videos, such as this one at 7:08 min shows an incorrect position https://www.youtube.com/watch?.... I suggest to change the wording of anterolateral position (even worse: antero-apical) to antero-axillary. That one of the most important therapy for VF cardiac arrest is not done or teached in an optimal fashion should be a cause of great concern for the ILCOR comitee. There should be great emphasis to better teach proper position of the defibrillating pads in future training materials.
2023.01.31 08:23:37 (modified: 2023.02.01 21:41:19)
The study was done with a rectilinear biphasic defibrillator at 200J. Most likely the conclusion would be the same with a biphasic truncated exponential defibrillator at 360J, but of course this can't be known for sure unless this would be tested.
Mohammad Abdul MANNANCord pulsation (6 sec x 10) for assessing heart rate is better for both LMICs and others to avoid pulseless electrical activity (PEA) and other errors of pulse oximetry.
Jamie TegartWe currently perform the recommended 3:1 compression:ventilation as per NRP guidelines. I would be curious what exactly is meant by the cardiac compressions with sustained inflation (SI). Two studies have commented on the positive results of using SI and I think more clinical studies could be warranted.
Jamie TegartOur site currently only teaches the two thumb techinique with emphasis on starting at the head of the bed. This is to help reduce the need to pause while setting up for UV insertion, as well as reduce fatigue and uneven pressure from compressing at the side of the bed. I would be interested about some of the methods mentioned which potentially provide improved compression rate and depth along with decreased fatigue.
Jamie TegartI think that a HR of under 60 is still an acceptable rate to use, for now, as that is what is currently practiced at most sites. While there is enough evidence to agree that a rising chest & audible air entry are acceptable indicators of effective ventilation, along with a rising HR as the primary indicator. I would be curious to review length of time providing effective ventilation before starting chest compressions. Our current practice is 30 seconds. Since we know that most infants are in a respiratory arrest and not a cardiac maybe more emphasis on the length of time effective ventilation is being done for would be beneficial
Dominic LaroseThe study was done with a rectilinear biphasic defibrillator at 200J. Most likely the conclusion would be the same with a biphasic truncated exponential defibrillator at 360J, but of course this can't be known for sure unless this would be tested.
Elene VanderpasAs previous evidence has shown that the 2-thumb technique is superior to the 2-finger technique, why does the American Heart Association's instructional video for parents/caregivers, "Infant CPR Anytime", teaches the 2-finger technique?
Jamie TegartRegardless of gestation or response to initial steps all our initial HR (heart rates) are achieved by auscultation. We do place ECG & SpO2 devices on infants who remain non vigorous after initial stimulation. ECG is definitely quicker and more accurate than pulse oxymetry for results, especially in those with a lower HR.
Silvia Heloisa Moscatel LoffredoMatter of great importance. Newborns, especially when born prematurely and submitted to advanced resuscitation, reaching the stage of chest compressions, require actions that minimize the deleterious effects of hypoxemia, as well as diffuse multisystemic injuries caused by hyperoxia.Thus, healthcare institutions must recognize the need to monitor oxygen saturation, the use of pulse oximetry, allowing further studies to clarify how supplementation of supplementary oxygen should be indicated.
Silvia Heloisa Moscatel LoffredoConsidering the expressive percentage of neonatal deaths due to asphyxia and that the systematic review showed that heart rate assessment can be obtained more quickly and accurately using the Electrocardiogram (ECG) during newborn resuscitation in the delivery room when compared to other evaluation methods, I believe that these studies will be able to instrumentalize the limited resources institution's managers, showing them that obtaining the ECG equipment to qualify asphyxiated newborn assistance may impact the reduction of expenses with treatment of complications of the asphyxia process.