A Randomized Trial of Epinephrine in Out-of-Hospital Cardiac Arrest
2018 · paper · Clinical trials
A double-blind trial of epinephrine versus saline placebo in 8,014 adults with out-of-hospital cardiac arrest. Epinephrine raised 30-day survival (3.2% vs 2.4%), but favorable neurologic outcome did not differ significantly, with more survivors severely impaired.
4 findings
Reference or label Discussion: the lower-than-anticipated survival rate "is similar to findings in other studies.7,36"; reference 7 is Larsson et al., "Alpha-adrenoceptor blockade by phentolamine inhibits adrenaline-induced platelet activation in vivo without affecting resting measurements" (Clin Sci 1992), a platelet-activation study.
- p. 9 Discussion, limitations paragraph, right column
is similar to findings in other studies.7,36
- p. 10 Reference list, entry 7
Alpha-adrenoceptor blockade by phentolamine inhibits adrenaline-induced platelet activation in vivo without affecting resting measurements
- p. 2 Introduction, first use of reference 7
let activation, which promotes thrombosis7
- p. 10 Reference list, entry 7
Alpha-adrenoceptor block- ade by phentolamine inhibits adrenaline- induced platelet activation in vivo
Only the authors can say what is right here.
Judge's note: The page 9 image shows the superscript "7,36" after "other studies." (not an extraction artefact). The reference list on page 10 gives entry 7 as the Larsson platelet-activation paper, which the Introduction cites, fittingly, for "platelet activation, which promotes thrombosis7". Its title describes a pharmacology study and it cannot support a statement about survival rates in cardiac-arrest studies. Reference 36 (Gueugniaud et al., vasopressin and epinephrine vs. epinephrine alone, page 11) fits. The reference number 7 appears to be wrong here; only the authors can say which reference was meant.
Second check: another model agreed
Report a problem with this finding
Reference or label Figure 2 x-axis ticks are labelled "0 1 10 100" although the legend says "The data are presented on a log10 scale"; the left-most tick, one evenly spaced decade below 1, is 0.1%, not 0.
- p. 9 Figure 2, x-axis tick labels
0 1 10 100
- p. 9 Figure 2 legend
The data are presented on a log10 scale of the percentages of patients in each group.
Judge's note: Seen on the crop and the page 9 image: the four ticks are equally spaced and labelled 0, 1, 10, 100. Zero cannot appear on a log scale, and equal spacing puts the first tick at 0.1. The bars agree: the first placebo segment (15 of 3994 = 0.38%) ends about 56% of the way from the first tick to the tick at 1 (measured on the crop), where 0.38% falls on a 0.1-to-1 decade (58%); the first epinephrine segment (12 of 4007 = 0.30%) ends at about 48%, as expected. The label "0" should read 0.1.
Second check: another model agreed
Report a problem with this finding
Reference or label Reference 14 is printed "Resuscitation 2014;96:328-40.", but the same list dates Resuscitation volume 85 to 2014 (ref 9) and volume 95 to 2015 (refs 3, 4, 13, 15), so volume 96 cannot be from 2014; the year is probably 2015.
- p. 10 Reference list, entry 14
Resuscitation 2014;96:328-40.
- p. 10 Reference list, entry 9
trials. Resuscitation 2014;85:732-40.
- p. 10 Reference list, entry 15
Resuscitation 2015;95:202-22.
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: that Resuscitation numbers its volumes consecutively in time, and that the Utstein update appeared in 2015;96:328-40. On the page 10 image, ref 14 reads "Resuscitation 2014;96:328-40.", ref 9 "Resuscitation 2014;85:732-40.", and refs 3, 4, 13 and 15 all give "2015;95". The year and volume in ref 14 are therefore inconsistent with the rest of the list. Whether the year or the volume is the wrong element is for the authors to confirm, though the year is the likely one.
Second check: another model agreed
Report a problem with this finding
Wording Text says "The results with respect to survival at 3 months and neurologic outcomes at 3 months were similar in the two groups", but Table 3 gives survival at 3 months as 121/4009 (3.0) vs 86/3991 (2.2), unadjusted odds ratio 1.41 (1.07–1.87), the same pattern as the 30-day result the paper calls significantly higher.
- p. 4 Results, Primary and Secondary Outcomes, right column
The results with respect to survival at 3 months and neurologic outcomes at 3 months were similar in the two groups
- p. 8 Table 3, row Survival at 3 mo
121/4009 (3.0)
- p. 8 Table 3, row Survival at 3 mo, unadjusted odds ratio
(1.07–1.87)
Only the authors can say what is right here.
Judge's note: Read page 4 and Table 3 on page 8 from the images. Survival at 3 months: 3.0% vs 2.2%, odds ratios 1.41 (1.07–1.87) and 1.47 (1.08–2.00), which is almost identical to survival at discharge (1.41, 1.08–1.86) and at 30 days (1.39, 1.06–1.82). The sentence as printed says the two groups were similar; the authors probably meant that the 3-month results were similar to the earlier ones. Nothing elsewhere in the article explains the wording. A wording query, not a data error.
Second check: another model agreed
Report a problem with this finding
9 minor (typos and the like; most checked by the judge only)
Wording Text: "further information indicated that 87 patients (1.1%) were ineligible to participate in the trial"; Figure 1 lists the 87 as excluded after randomization, of whom 47 "Had unknown reason", 4 "Had broken or contaminated syringes" and 2 "Had no intravenous access".
- p. 4 Results, Patients and Interventions
further information indicated that 87 pa-
- p. 5 Figure 1, box "87 Were excluded after randomization"
47 Had unknown reason
- p. 5 Figure 1, box "87 Were excluded after randomization"
4 Had broken or contaminated syringes
Judge's note: Figure 1 (page 5 image): 4 + 6 + 22 + 2 + 4 + 2 + 47 = 87, so the count agrees with the text and 87/8103 = 1.1%. Only 34 of the 87 have a reason that matches an eligibility criterion; 53 have a syringe or access problem or no known reason. "Ineligible" is therefore looser than the figure supports. The figure's own heading is "excluded after randomization". No number is wrong; a wording point.
Second check: another model agreed
Report a problem with this finding
Wording Discussion: "Among the survivors, almost twice the number in the epinephrine group as in the placebo group had severe neurologic impairment"; the numbers are 39 vs. 16 (2.4 times), and "almost twice" fits only the proportions of survivors, 31.0% vs. 17.8% (1.7 times).
- p. 9 Discussion, left column
Among the survivors, almost twice the number in the epinephrine group as in the placebo group had severe neurologic impairment.
- p. 9 Figure 2 legend
(39 of 126 patients [31.0%] vs. 16 of 90 patients [17.8%])
- p. 9 Discussion, paragraph beginning 'Clinical decision making'
almost twice the number in the epinephrine group as in the pla- cebo group had severe neurologic impairment
- p. 7 Discussion para 1
survived with severe neurologic dis-ability in the epinephrine group
Only the authors can say what is right here.
Judge's note: Page 9 image: Figure 2 bar labels give scores 4 and 5 as 12 + 27 = 39 (epinephrine) and 8 + 8 = 16 (placebo), matching the legend's "39 of 126 patients [31.0%] vs. 16 of 90 patients [17.8%]". The count ratio is 2.44 and the proportion ratio 1.74, so the sentence is accurate for the proportion and not for the "number" it names. Loose wording; no figure is wrong. The third reviewer note ("disability" on page 7 against "impairment" elsewhere) is a style point and is not part of this verdict.
Second check: another model agreed
Report a problem with this finding
Typo Methods name the U.K. regulator as "the Medicines and Healthcare Products Regulatory Authority"; the body's name ends in "Agency" (MHRA).
- p. 2 Methods, Trial Design and Oversight
and Healthcare Products Regulatory Authority ap-
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the U.K. medicines regulator is the Medicines and Healthcare products Regulatory Agency. The crop of page 2 shows the printed text "Medicines and Healthcare Products Regulatory Authority ap-proved the protocol" exactly as quoted. The name cannot be checked from the files, so this is a query for the authors rather than a correction.
Second check: none (judge only)
Report a problem with this finding
Typo Appendix affiliations: "Oxford (J.J.M.B)" lacks the final period (printed "J.J.M.B." two lines above), and "London Ambulance Service NHS Trust, (F.M., R.T.F.)" has a stray comma before the parenthesis.
- p. 10 Appendix, affiliations paragraph
Oxford (J.J.M.B)
- p. 10 Appendix, affiliations paragraph
London Ambulance Service NHS Trust, (F.M.,
- p. 10 Appendix affiliations
London Ambulance Service NHS Trust, (F.M., R.T.F.), London
- p. 10 Appendix, affiliations
South East Coast Ambulance Service, Crawley (F.M.)
Judge's note: Both punctuation slips are visible on the page 10 image and the appendix crop; every other affiliation has no comma before the parenthesised initials, and the same author's initials appear as "J.J.M.B." under South Central Ambulance Service. The third reviewer point is not confirmed: F.M. listed under two services is an ordinary dual affiliation, and the six ambulance services in the author affiliations need not equal the "five National Health Service ambulance services" that conducted the trial (page 2), since an author's employer is not necessarily a trial site.
Second check: none (judge only)
Report a problem with this finding
Typo Manufacturer printed as "(Physiocontrol)"; the company's name is written Physio-Control.
- p. 3 Randomization and Treatment, para 2
defibrillator downloads when avail-able (Physiocontrol)
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the defibrillator manufacturer styles its name "Physio-Control". The crop of page 3 shows "defibrillator downloads when avail-able (Physiocontrol)" as quoted. Cosmetic spelling of a company name; not checkable from the files.
Second check: none (judge only)
Report a problem with this finding
Typo "Data regarding the quality of the CPR results were obtained with the use of defibrillator downloads": the word "results" appears to be superfluous (quality of CPR is meant).
- p. 3 Randomization and Treatment, para 2
Data regarding the quality of the CPR results were obtained
Judge's note: The crop of page 3 shows the sentence as quoted. The Discussion (page 9) speaks of "Information about the quality of CPR", which suggests "results" is a stray word. The meaning is recoverable; a wording point only.
Second check: none (judge only)
Report a problem with this finding
Typo The drug is called "adrenaline" in "an assumption of no benefit from adrenaline" (page 4) and in Figure 1 "1192 Received adrenaline before ambulance arrival", where the article otherwise uses "epinephrine".
- p. 4 Results, Bayesian analysis sentence
tion of no benefit from adrenaline, the posterior
- p. 5 Figure 1, exclusion box
1192 Received adrenaline before ambulance arrival
- p. 4 Results, Bayesian analysis paragraph
assump-tion of no benefit from adrenaline
- p. 5 Figure 1 box
Received adrenaline before ambulance arrival
Judge's note: Both instances seen on the page images (page 4 text, Figure 1 on page 5). The Introduction gives "Epinephrine (adrenaline)" and the Methods exclusion criterion reads "the administration of epinephrine before the arrival of the trial-trained paramedic". The terms are synonyms and the meaning is clear; a terminology inconsistency only.
Second check: none (judge only)
Report a problem with this finding
Typo Table 3 row header "Median length of hospital stay (IQR)" gives no unit, whereas the parallel row above reads "Median length of stay in ICU (IQR) — days".
- p. 8 Table 3, row header
Median length of hospital stay (IQR)
- p. 8 Table 3, preceding row header
Median length of stay in ICU (IQR) — days
Judge's note: Table 3 on the page 8 image: the ICU row carries "— days" and the hospital-stay row does not. The unit is evidently days and can be inferred; cosmetic. The second point in the reviewer's note ("Patients who died" shows 0 with no interquartile range) is also as printed, but may simply mean that the quartiles were all 0; not confirmed as a defect.
Second check: none (judge only)
Report a problem with this finding
Typo Funding statement: "by the Comprehensive Research Network Intensive Care Foundation and Out of Hospital Cardiac Arrest Registry, which is funded by the British Heart Foundation and Resuscitation Council; and by Health Care Wales"; the funder names appear to be run together or misnamed.
- p. 9 Funding statement
Comprehensive Research Network Intensive Care Foundation and Out of Hospital Cardiac Arrest Registry
- p. 9 Funding statement
by Health Care Wales
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: that the Welsh body is called Health and Care Research Wales, and that the Comprehensive (Clinical) Research Network and the Intensive Care Foundation are separate organisations. The page 9 image shows the text as quoted, with no comma or "and" between "Comprehensive Research Network" and "Intensive Care Foundation". The correct names cannot be checked from the files; a query for the authors.
Second check: none (judge only)