Anaesthetic Depth and Complications after Major Surgery: An International, Randomised Controlled Trial
2019 · paper · Clinical trials
A randomized trial in 6644 older, high-risk surgical patients comparing light and deep general anaesthesia guided by the bispectral index. One-year mortality was not significantly lower with light anaesthesia (6.5% vs 7.2%).
10 findings
Substance Results text gives median BIS in the BIS 35 group as '38·8 (36·3–42·2)', whereas the Abstract has '38·8 (36·3 to 42·4)' and Table 2 has '38·8 (36·3–42·4)'.
- p. 5 Results text, paragraph 1
38·8 (36·3–42·2) in the BIS 35 group
- p. 5 Table 2, BIS row
38·8 (36·3–42·4)
- p. 1 Abstract, Findings
38·8 (36·3 to 42·4) in the BIS 35 group
- p. 5 Results text
38·8 (36·3–42·2)
Judge's note: Read on the page images of p 1 (Abstract) and p 5 (text and Table 2). The figures describe the same ITT group and statistic, so the upper quartile 42·2 in the text conflicts with 42·4 in the other two places.
Second check: another model agreed
Report a problem with this finding
Substance Table 1, Peptic ulcer disease, BIS 50: '385 (10%)', but 385/3316 = 11·6%.
- p. 4 Table 1, Peptic ulcer disease, BIS 50
385 (10%)
Only the authors can say what is right here.
Judge's note: Page 4 image reads 385 (10%); column n=3316. No other denominator is stated, and BIS 35 393/3328 = 11·8% is printed as 12%. Either the count or the percentage is wrong; only the authors can say which.
Second check: another model agreed
Report a problem with this finding
Substance Table 2, PACU 'Number who attended', BIS 50: '3314 (91%)'; 3314/3316 = 99·9%. The BIS 35 cell '3030 (90%)' is also 91·0% of 3328.
- p. 5 Table 2, Postanaesthesia care unit, Number who attended
3314 (91%)
Only the authors can say what is right here.
Judge's note: Image confirms 3314 (91%). 91% of 3316 is about 3018, and the analgesia and antiemetic rows (1954, 65%; 626, 21%) both fit a denominator of about 3010, so the count looks like a typo for about 3014. Only the authors can confirm.
Second check: another model agreed
Report a problem with this finding
Reference or label In Procedures, the "Brice questionnaire for awareness ... day 30.21" and the "15-item quality of recovery score ... 30 postoperatively.22" cite refs 21 and 22, but ref 21 is Stark et al (QoR-15) and ref 22 is Brice et al (awareness). The two citations are swapped.
- p. 3 Methods, Procedures
once on day 1, 2, or 3, and again on day 30.21
- p. 3 Methods, Procedures
and 30 postoperatively.22
- p. 8 References 21 and 22
evaluation of a postoperative quality of recovery score: the QoR-15
- p. 3 Procedures
Brice questionnaire for awareness was administered once on day 1, 2, or 3, and again on day 30.21
Judge's note: Checked the p3 text and the p8 reference list (refs 21 and 22) in the images. Separately, ref 21 prints 'Anesthesiology 2013; 18: 1332–40'. Rests on outside knowledge: per Crossref the volume is 118. That would be a further S2 reference error for the authors to check.
Second check: another model agreed, code check
Report a problem with this finding
Reference or label Table 1: 'Preoperative creatinine, mmol/L–¹' with medians 85 (71–104) and 84 (71–103); the values fit µmol/L.
- p. 4 Table 1, Preoperative creatinine
Preoperative creatinine, mmol/L
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: normal serum creatinine is about 60–110 µmol/L, so values near 85 cannot be mmol/L. The unit label appears to be wrong; the data are presumably correct.
Second check: another model agreed
Report a problem with this finding
Reference or label Ref 13 reads "Anesth Analg 2013; 118: 977–80." but the next reference, ref 14, in the same volume on the following pages reads "Anesth Analg 2014; 118: 981–86."; the year of ref 13 is inconsistent (2014 expected).
- p. 7 Reference 13
Anesth Analg 2013; 118: 977–80.
- p. 7 Reference 14
Anesth Analg 2014; 118: 981–86.
- p. 7 Reference 13
Anesth Analg 2013; 118: 977–80
Judge's note: Checked both entries on the p7 image. Ref 13 (Brown CH 4th et al.) is dated 2013, vol 118, pp 977–80. Ref 14 (Short T et al.) is dated 2014, vol 118, pp 981–86. Consecutive page ranges in the same volume of the same journal cannot belong to different publication years. The list's own evidence therefore makes 2013 inconsistent, and 2014 matches ref 14. Which year the original article actually carries is outside knowledge, but the internal inconsistency is plain.
Second check: another model agreed
Report a problem with this finding
Reference or label Ref 21 (Stark PA et al., QoR-15) reads "Anesthesiology 2013; 18: 1332–40.". The volume "18" cannot be right for 2013 (ref 27 in the same list is Anesthesiology 2013; 119), so a digit appears to be missing (118).
- p. 8 Reference 21
Anesthesiology 2013; 18: 1332–40.
- p. 8 Reference 27
Anesthesiology 2013; 119: 507–15.
- p. 8 Ref 21
Anesthesiology 2013; 18: 1332–40
- p. 7 Ref 13
Anesth Analg 2013; 118: 977–80
Judge's note: The p8 image confirms that ref 21 prints volume 18 in bold, so this is not an extraction artefact. Within the same list, Anesthesiology 2010 is vol 112 (ref 5), 2013 is vol 119 (ref 27) and 2015 is vol 122 (ref 25), so a 2013 volume must be about 118–119. The part of reviewer note 5 about ref 13's year duplicates C032 and is judged there.
Second check: another model agreed
Report a problem with this finding
Wording The abstract and Results say "6644 were enrolled, randomly assigned to treatment or control, and formed the intention-to-treat population", but Figure 1 shows "6668 enrolled and randomly assigned". 6644 is the number who had anaesthesia and surgery.
- p. 1 Abstract, Findings
6644 were enrolled, randomly assigned to treatment or control
- p. 4 Results, paragraph 1
6644 were enrolled, randomly assigned to treatment or
- p. 3 Figure 1
6668 enrolled and randomly assigned
Judge's note: In Figure 1 (p3), 3326 + 3342 = 6668 were randomised. After 10 and 14 exclusions, 6644 remain, which matches the ITT definition on p3 (randomised and had induction). The ITT count of 6644 is right, but the text calls it the number enrolled and randomised, which conflicts with the flow diagram.
Second check: another model agreed
Report a problem with this finding
Wording Methods list "a composite of these four cardiovascular outcomes" (MI, cardiac arrest, PE, stroke) as a secondary outcome, but Table 3 reports only a "Composite of mortality, myocardial infarction, cardiac arrest, pulmonary embolism, and stroke", under Exploratory outcomes.
- p. 3 Methods, Outcomes
a composite of these four cardiovascular outcomes
- p. 6 Table 3, Exploratory outcomes
Composite of mortality, myocardial infarction, cardiac arrest, pulmonary embolism, and stroke
- p. 3 Outcomes
a composite of these four cardiovascular outcomes, sepsis, surgical site infection, a composite of these two septic outcomes
- p. 6 Table 3
Composite of mortality, myocardial
Only the authors can say what is right here.
Judge's note: Checked the p3 Outcomes paragraph and the full Table 3 on p6. The four-component composite is not reported anywhere in the article, and the paper does not explain the change.
Second check: another model agreed
Report a problem with this finding
Wording Figure 1 shows "11 358 excluded*", but the listed reasons (2619 + 4978 + 2459 + 181) sum to 10 237, although the footnote says "Some patients had multiple reasons for exclusion".
- p. 3 Figure 1
11 358 excluded*
- p. 3 Figure 1 footnote
*Some patients had multiple reasons for exclusion.
Only the authors can say what is right here.
Judge's note: Checked Figure 1 on p3. The total of 18 026 − 6668 = 11 358 is consistent. If some patients had several reasons, the reasons should sum to at least 11 358, but they fall 1121 short, so a category or count appears to be missing.
Second check: another model agreed
Report a problem with this finding
17 minor (typos and the like; most checked by the judge only)
Substance Results: '4221 (63%) were male and 2423 (37%) were female'; of 6644, these counts are 63·5% and 36·5%.
- p. 4 Results, paragraph 2
4221 (63%) were male and 2423 (37%) were female
Judge's note: Counts match Table 1 (2111+2110, 1205+1218); denominator is the ITT population of 6644. 4221/6644 is 63·53%, which rounds to 64%, and 2423/6644 is 36·47%, which rounds to 36%. The error is small.
Second check: another model agreed
Report a problem with this finding
Substance Results: '3107 (46%) had surgery for cancer'; 3107/6644 = 46·8% (47%).
- p. 4 Results, paragraph 2
3107 (46%) had surgery for cancer
- p. 4 Table 1, Operation for cancer
1531 (46%)
Judge's note: 1531+1576 = 3107, matching Table 1. The Table 1 cell 1531 (46%) is correct (46·2%), so only the text percentage is off. Abdominal surgery 3053/6644 = 46·0% is correct.
Second check: another model agreed
Report a problem with this finding
Wording The Methods list the cardiovascular and septic composites among the "Secondary outcomes", but Table 3 places them under "Exploratory outcomes".
- p. 3 Methods, Outcomes
Secondary outcomes were the incidences of myocardial
- p. 6 Table 3
Exploratory outcomes
Judge's note: The p3 Methods and the Table 3 headings on p6 were checked. The septic composite (sepsis + SSI) matches the Methods definition, yet it is labelled exploratory. The change of tier is not explained.
Second check: another model agreed
Report a problem with this finding
Wording The Methods name the secondary outcome "total intensive care unit stay", but Table 3 reports "Unplanned ICU admission" (n, %, OR), and no ICU length of stay is reported.
- p. 3 Methods, Outcomes
total intensive care unit stay
- p. 6 Table 3
Unplanned ICU admission
- p. 3 Outcomes
a composite of these two septic outcomes, total intensive care unit stay
- p. 6 Table 3
Composite of sepsis and surgical site infection
Only the authors can say what is right here.
Judge's note: Table 3 on p6 was checked. Its other outcomes include hospital stay but not ICU stay. The two measures are different quantities, and the paper does not explain the switch. The tier and composite points in reviewer note 2 are covered by C011 and C012.
Second check: another model agreed
Report a problem with this finding
Wording Discussion: "1-year mortality was 2% lower than anticipated". Methods give an expected 1-year survival of 90% (10% mortality), and observed mortality was 6·5% and 7·2% (450/6644 ≈ 6·8%), so the baseline behind the "2%" is not stated.
- p. 7 Discussion, limitations paragraph
1-year mortality was 2% lower than anticipated
- p. 3 Statistical analysis
The expected probability of 1-year survival was 90%.
- p. 3 Statistical analysis
The expected probability of 1-year survival was 90%
- p. 7 Discussion, limitations
1-year mortality was 2% lower
Only the authors can say what is right here.
Judge's note: p3 Statistical analysis: "The expected probability of 1-year survival was 90%", with a 20% reduction in mortality powered for the BIS 50 group. p5 Results: 6·5% (212) and 7·2% (238). Against the 10% expected mortality, the observed overall rate is about 3·2 points lower, and the BIS 35 arm is 2·8 points lower. A gap of about 2 points appears only if the anticipated figure is the average of 10% and the 8% assumed under the alternative (9%). The paper does not say which baseline it means, so this is a query to the authors, not a clear error.
Second check: another model agreed
Report a problem with this finding
Wording Methods say 'The Holm-Bonferroni adjustment for multiplicity was applied to the secondary outcomes', yet the neuropathic pain 1-year score (p 0·038, Table 3) is called 'a significant difference' in the Results and the Discussion; the paper does not say whether the Table 3 p values are adjusted.
- p. 4 Statistical analysis
The Holm-Bonferroni adjustment for multiplicity was applied to the secondary outcomes.
- p. 6 Table 3, Neuropathic pain 1 year score
0·038
- p. 5 Results
There was a significant difference between the groups in the severity, but not incidence, of neuropathic pain at 1 year.
- p. 7 Discussion, paragraph 3
There was a significant difference between groups in the severity, but not the incidence, of neuropathic pain at 1 year.
Only the authors can say what is right here.
Judge's note: Checked Methods (p 4), Table 3 (p 6), Results (p 5) and Discussion (p 7). Table 3 lists more than ten secondary comparisons, and none of the Table 3 footnotes says the p values are Holm-adjusted. Query: are the printed p values adjusted, and on what basis is 0·038 called significant?
Second check: another model agreed
Report a problem with this finding
Wording Results: '4272 (66%) patients being within five units of the target', yet 2489 patients (1198+1291 in Figure 1) were excluded from the per-protocol set for mean BIS more than five units from target, leaving at most 4155.
- p. 5 Results, paragraph 1
4272 (66%) patients being within five units of the target
- p. 5 Results, per-protocol paragraph
2489 patients were excluded for mean
Only the authors can say what is right here.
Judge's note: Figure 1 shows 1198 and 1291 'mean BIS out of range', which together are the 2489 in the text. The per-protocol exclusion counts add up exactly (1234 and 1332 per arm), so no overlap explains the gap. 4272/6445 = 66%. The definition of 'within five units' may differ, but the paper does not say so.
Second check: another model agreed
Report a problem with this finding
Wording Discussion says "39% of patients received less than 0·7 MAC of volatile anaesthetic throughout surgery", whereas the Results report that "2602 (39%) patients received a mean MAC of less than 0·7 during maintenance of anaesthesia".
- p. 7 Discussion, awareness paragraph
39% of patients received less than 0·7 MAC of volatile anaesthetic throughout surgery
- p. 5 Results
2602 (39%) patients received a mean MAC of less than 0·7 during maintenance of anaesthesia
Judge's note: p5 Results and the Methods (p3: maintenance epoch from 10 min after induction to discontinuation; per-patient mean values) describe a per-patient mean MAC below 0·7 over the maintenance period. The Discussion (p7) rewords this as receiving less than 0·7 MAC "throughout surgery", which reads as a ceiling never exceeded during the whole operation. A mean below 0·7 does not establish that. This is a wording overstatement; the number itself agrees with the Results.
Second check: another model agreed
Report a problem with this finding
Wording The per-protocol primary outcome is reported as 'OR 0·86 (95% CI 0·67–1·09', while the Methods describe the primary outcome as analysed by log-rank test and summarised as an HR; the per-protocol model is not stated.
- p. 6 Results, per-protocol
OR 0·86 (95% CI 0·67–1·09
Judge's note: The Methods (pp 3–4) give HRs from a Cox model for the primary outcome, and the Table 3 footnote also gives an HR for the primary outcome. The per-protocol set excludes patients lost to follow-up, so a binary analysis would be possible, but the paper does not say which model was used.
Second check: another model agreed
Report a problem with this finding
Typo The abstract Funding line names the "Research Grant Council of Hong Kong", while the Acknowledgments name the "Research Grants Council of Hong Kong".
- p. 1 Abstract, Funding
Research Grant Council of Hong Kong
- p. 7 Acknowledgments
the Research Grants Council of Hong Kong
- p. 7 Acknowledgments
Research Grants Council of Hong Kong
Judge's note: Both forms are visible in the images (p1 Funding, p7 Acknowledgments), so the paper names the same funder two ways.
Second check: none (judge only)
Report a problem with this finding
Typo The affiliations read "Hong Kong Special Administrative Area, China" (twice) instead of "Special Administrative Region".
- p. 1 Affiliations sidebar
Hong Kong Special Administrative Area, China
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the official designation is 'Special Administrative Region'. Both instances on p1 were seen in the images (Chinese University of Hong Kong; Pamela Youde Nethersole Eastern Hospital).
Second check: none (judge only)
Report a problem with this finding
Typo The byline gives "Davina J McAllister", but the affiliations give "D S McAllister PGDipHSc" and the Contributors section gives "DSM".
- p. 1 Affiliations
D S McAllister PGDipHSc
- p. 1 Author list
Davina J McAllister
Only the authors can say what is right here.
Judge's note: Both forms were seen on p1. The Contributors text (line 953) uses DSM. Only the authors can say which middle initial is correct.
Second check: none (judge only)
Report a problem with this finding
Typo The ketamine infusion rate is printed as "25 mg/h⁻¹", which mixes a solidus with a negative exponent; it should be mg/h or mg h⁻¹.
- p. 2 Methods, Study design and patients
25 mg/h
Judge's note: The p2 image shows the superscript ⁻¹ after mg/h.
Second check: none (judge only)
Report a problem with this finding
Typo Unit notation combines a slash with a superscript −1: 'g/L–¹' (haemoglobin, albumin), 'mmol/L–¹' (creatinine), '25 mg/h–¹' (exclusion criteria).
- p. 4 Table 1, haemoglobin and albumin rows
Preoperative haemoglobin, g/L
- p. 4 Table 1, albumin row
Albumin, g/L
- p. 2 Exclusion criteria
25 mg/h–¹
- p. 4 Table 1
Preoperative haemoglobin, g/L–¹
Judge's note: Superscript −1 after the slash is visible on the page 4 table and in the page 2 text. The notation is redundant, but the intended meaning is clear.
Second check: none (judge only)
Report a problem with this finding
Typo The text reads "Medical Dictionary of Regulatory Activities (MedDRA)"; the name is "Medical Dictionary for Regulatory Activities".
- p. 3 Outcomes
Medical Dictionary of Regulatory Activities (MedDRA)
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: MedDRA's official name uses 'for'. The p3 image shows 'of'.
Second check: none (judge only)
Report a problem with this finding
Typo Table 1, ASA 3, BIS 35: '3144 (95%)'; 3144/3328 = 94·5%, which rounds to 94%. The total of 3327 is plausibly accounted for by the footnoted ASA 2 patient.
- p. 4 Table 1, ASA physical status, BIS 35
3144 (95%)
- p. 4 Table 1 footnote
*Includes the protocol violation of one ASA physical status 2 patient.
Judge's note: Image confirms 3144 (95%) and 183 (5%). The one-patient shortfall (3144+183 = 3327 vs 3328) fits the footnote about one ASA physical status 2 patient (an ASA 2 patient is also mentioned in the per-protocol exclusions on p 6), so the count is explained. Only a small rounding discrepancy in the percentage remains.
Second check: none (judge only)
Report a problem with this finding
Typo Table 2 row label 'Major regional local anaesthesia' differs from the text's 'major regional anaesthesia'.
- p. 5 Table 2
Major regional local
Judge's note: Table 2 on p 5 reads 'Major regional local anaesthesia'. The Methods (p 2) and Results (p 5) say 'major regional anaesthesia'. This is a labelling inconsistency only.
Second check: none (judge only)