Targeted Temperature Management at 33°C versus 36°C after Cardiac Arrest
2013 · paper · Clinical trials
A trial randomising 950 unconscious survivors of out-of-hospital cardiac arrest to body-temperature targets of 33°C or 36°C. It found no difference in death (50% vs 48%) or in poor neurologic function at 180 days.
0 findings
None that passed a second check.
6 minor (typos and the like; most checked by the judge only)
Wording Methods name a paired test for what appear to be comparisons between independent randomized groups: "The Wilcoxon signed-rank test was used to compare distributions of continuous outcome measures."
- p. 3 Statistical Analysis
The Wilcoxon signed-rank test was used to compare distributions of continuous outcome measures.
- p. 3 Statistical Analysis
The Wilcoxon signed-rank test was used to
Only the authors can say what is right here.
Judge's note: Checked on the page-3 image. The trial compares two independently randomized groups, and the article describes no paired or matched continuous comparison, so the signed-rank test does not fit as named; the rank-sum (Mann–Whitney) test would. The article does not report the continuous-outcome results this test was applied to; they may be in the Supplementary Appendix, which is not among the files. Only the authors can say whether this is a misnomer.
Second check: another model agreed
Report a problem with this finding
Wording The Table 1 footnote says "P>0.05 for all comparisons", but the ischemic heart disease row prints 145 (31) vs 115 (25), and a standard two-group test of these counts gives P of about 0.04–0.05.
- p. 5 Table 1 footnote *
P>0.05 for all comparisons.
- p. 4 Table 1, Medical history, Ischemic heart disease row
145 (31)
- p. 5 Results, Patients
similar prerandomization characteristics (Table 1)
Only the authors can say what is right here.
Judge's note: The Table 1 image on page 4 shows ischemic heart disease 145 (31) of N=473 vs 115 (25) of N=466. The footnote on page 5 shows "P>0.05 for all comparisons." My own pooled two-proportion z is about 2.05 (two-sided P ≈ 0.04); Fisher's exact and continuity-corrected tests also give P below 0.05. The table does not name its test, and a stratified or other test is possible, so this goes to the authors as a query about which test produced the footnote.
Second check: another model agreed
Report a problem with this finding
Wording In Table 3, CPC at follow-up, category 4, the 36°C group is printed as "2 (0.5)", but 2/464 = 0.43%, which rounds to 0.4 (Table 1 prints a similar case, 2/473, as "<0.5").
- p. 8 Table 3, CPC at follow-up, category 4, 36°C group
2 (0.5)
- p. 4 Table 1, unknown first rhythm, 33°C group (comparison convention)
2 (<0.5)
Judge's note: The page-8 image shows a column total of 464 and "2 (0.5)". The other percentages in the column agree with 464 (for example 183/464 = 39%). 2/464 = 0.431%, which is not a rounding tie. Table 1 on page 4 prints 2/473 as "2 (<0.5)".
Second check: another model agreed
Report a problem with this finding
Wording "With the use of the best reported CPC … the relative risk … was 1.04 (95% CI, 0.89 to 1.17; P = 0.67)". The Table 3 best-CPC counts give a crude RR of about 1.07, and P = 0.67 is the same value printed for the modified Rankin trend.
- p. 7 Results, Follow-up and Outcomes
was 1.04 (95% CI, 0.89 to 1.17; P = 0.67)
- p. 8 Table 3, best CPC during trial
201 (43)
Only the authors can say what is right here.
Judge's note: Table 3 (page 8 image): best CPC categories 3 and 4 are 37+201 of 472 (33°C) vs 37+183 of 466 (36°C), with category 5 NA. That gives a crude RR of 1.068. For Table 2, crude ratios agree with the printed site-adjusted ones (1.026 vs 1.02, 1.014 vs 1.01, 1.012 vs 1.01), so the larger gap here is unusual. The printed CI is also centred on about 1.02 on the log scale, and P = 0.67 repeats the mRS trend P. Site adjustment (Cochran–Mantel–Haenszel) could explain part of this, so it is a query to the authors about which analysis produced these numbers.
Second check: another model agreed
Report a problem with this finding
Typo The investigators footnote on page 1 has two verbs run together: "…(TTM) trial is provided listed in the Supplementary Appendix".
- p. 1 Front-page side note (asterisk footnote on TTM Trial Investigators)
provided listed in the Supplementary
- p. 1 footnote * on first page
is provided listed in the Supplementary Appendix
Judge's note: The page-1 image shows "is provided listed in the Supplementary Appendix" printed in the typeset footnote, so this is not an extraction artefact. Only one of "provided" or "listed" belongs.
Second check: none (judge only)
Report a problem with this finding
Typo The affiliations appendix prints "Departments of Neurology (T.C)" without the final period; elsewhere the initials are "T.C."
- p. 9 Appendix, affiliations
Departments of Neurology (T.C)
Judge's note: The page-9 image shows "(T.C)". The same line has "(N.N., T.C., D.E., …)" with the period.
Second check: none (judge only)