The Safety of Addition of Nitrous Oxide to General Anaesthesia in At-Risk Patients Having Major Non-Cardiac Surgery (ENIGMA-II)
2014 · paper · Clinical trials
A randomized trial of general anaesthesia with or without nitrous oxide in 7112 at-risk patients having major non-cardiac surgery. Nitrous oxide did not raise death or cardiovascular complications or infection, but increased severe nausea and vomiting.
Checked: PDF from medic.usm.my, fetched 4 Oct 2026. Page numbers are those of that PDF.
7 findings
Substance Table 1, Heart failure, nitrous oxide arm: '268 (9%)' (n=3495). 268/3495 is 7.7%, which would print as 8%.
- p. 3 Table 1, Pre-existing medical conditions, Heart failure, nitrous oxide column
268 (9%)
- p. 3 Table 1, pre-existing conditions
Heart failure 268 (9%) 276 (8%)
- p. 3 Table 1, Pre-existing medical conditions
Heart failure 268 (9%)
Only the authors can say what is right here.
Judge's note: The page 3 image shows 268 (9%) beside 276 (8%). The column n is 3495 and the paper states no other denominator for this row. Neighbouring rows match n/3495. 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 4 nitrous oxide arm: 'Any adverse event 624 (17%)' while Mild 103 + Moderate 249 + Severe 239 = 591; 624/3483 is 17.9%.
- p. 6 Table 4, Any adverse event, nitrous oxide column
624 (17%)
- p. 6 Table 4, Mild/Moderate/Severe, nitrous oxide column
249 (7%)
- p. 6 Table 4
Any adverse event 624 (17%)
- p. 6 Table 4, Mild
103 (3%)
Only the authors can say what is right here.
Judge's note: Checked on the page 6 image. 591/3483 gives 17.0%, which matches the printed 17%, so 624 looks inconsistent. In the comparator arm the grades sum to 558 against 561, and 561/3509 = 16%. The paper does not explain ungraded events. Only the authors can say which figure is right.
Second check: another model agreed
Report a problem with this finding
Reference or label Table 1 'Creatinine (mmol/L)' with medians 84 (70–105) / 85 (70–106); these values fit µmol/L.
- p. 3 Table 1, Preoperative laboratory tests
Creatinine (mmol/L)
Judge's note: The crop image clearly shows 'mmol/L', so this is not a lost µ in extraction. A serum creatinine of about 84 is a typical value in µmol/L. The unit label is wrong.
Second check: another model agreed
Report a problem with this finding
Reference or label Table 3: 'Admitted to HDU 210 (6·0)*', 'Admitted to ICU 379 (11)*' with footnote '*n=3052. †n=3069.' — the printed percentages fit n=3483/3509, not the footnoted n.
- p. 5 Table 3, Admitted to HDU / Admitted to ICU, and footnote
*n=3052. †n=3069.
- p. 5 Table 3, Admitted to HDU
210 (6·0)*
Only the authors can say what is right here.
Judge's note: Checked on the page 5 image. 210/3052 is 6.9% and 379/3052 is 12.4%, but the printed values match the full denominators. The text says 17% in each group went to ICU or HDU, which also fits the full denominators. The meaning of the footnoted n is unclear, or the footnote is wrong.
Second check: another model agreed
Report a problem with this finding
Reference or label Table 3: 'Unplanned ICU admission 94 (2·7)‡ / 113 (3·2)§' and hospital stay marked ‡/§ with footnote '‡n=421. §n=439.' — the percentages fit 3483/3509, not 421/439.
- p. 5 Table 3, Unplanned ICU admission and Hospital stay rows; footnote
‡n=421.
Only the authors can say what is right here.
Judge's note: Checked on the page 5 image. 94/421 is 22%, but 94/3483 is 2.7%. Hospital stay was analysed in all patients (Cox regression), so a subset n of 421/439 does not fit that row either. The footnotes may belong to the mechanical ventilation row; only the authors can say.
Second check: another model agreed
Report a problem with this finding
Reference or label Table 3 header 'Risk ratio (95% CI)†' has a dagger whose only footnote is '†n=3069', which belongs to the HDU/ICU no-nitrous cells.
- p. 5 Table 3 column header and footnote
Risk ratio (95% CI)†
Judge's note: The crop and page 5 image show the † on the header. The footnote defines † only as n=3069, so the header marker is either undefined or reuses the symbol by mistake.
Second check: another model agreed
Report a problem with this finding
Wording Statistical analysis: 'detect a clinically important reduction in the primary outcome of death and cardiovascular events from 6% to 8%' — 6% to 8% is an increase.
- p. 4 Statistical analysis, first paragraph
clinically important reduction in the primary outcome of death and cardiovascular events from 6% to 8%
- p. 4 Statistical analysis
reduction in the primary outcome of death and cardiovascular events from 6% to 8%
- p. 4 Statistical analysis
from 6% to 8%
- p. 4
detect a clinically important reduction in the primary outcome of death and cardiovascular events from 6% to 8%
Judge's note: The crop of page 4 shows the text as quoted. The direction word contradicts the order of the figures. The fix is either 'increase ... from 6% to 8%' or 'reduction ... from 8% to 6%'.
Second check: another model agreed
Report a problem with this finding
7 minor (typos and the like; most checked by the judge only)
Wording Discussion calls the ENIGMA (ref 16) cardiovascular and surgical-site-infection results 'These significant findings', but the Introduction reports the ref 16 30-day ischaemic cardiac result as a 'non-significant increase (from 0·7% to 1·3%, p=0·26)'.
- p. 7 Discussion, para 2
These significant findings could therefore represent a type I error
- p. 1 Introduction, para 3
significant increase (from 0·7% to 1·3%, p=0·26)
Only the authors can say what is right here.
Judge's note: The page 7 Discussion attributes increased cardiovascular complications and SSI to the ENIGMA trial (ref 16) and calls them 'significant findings'. The Introduction on pages 1–2 describes the ref 16 cardiac result as non-significant. The only significant cardiovascular result it cites is late MI from ref 17. As written, the Discussion wording overstates the ref 16 cardiovascular result. The authors should say which findings they mean.
Second check: another model agreed
Report a problem with this finding
Wording Table 1 'Type of surgery' counts in the nitrous oxide arm sum to 3499, but n=3495; in the other arm they sum to exactly 3516.
- p. 3 Table 1, Type of surgery, nitrous oxide column
1348 (39%)
- p. 3 Table 1, Type of surgery
Vascular 1348 (39%)
Only the authors can say what is right here.
Judge's note: I checked the page 3 image. The ten categories in the comparator arm add up exactly to n, which suggests they are mutually exclusive, yet the nitrous oxide arm exceeds n by 4. The paper gives no rule allowing more than one surgery type per patient. One count is probably off; the discrepancy is small.
Second check: another model agreed
Report a problem with this finding
Wording Table 3, Fever Day 3, nitrous oxide arm: '182 (6%)'. Against the column n=3483 this is 5.2%, and the comparator 178/3509 (5.1%) prints as 5%.
- p. 5 Table 3, Fever, Day 3, nitrous oxide column
182 (6%)
- p. 5 Table 3, Fever day 3
182 (6%) 178 (5%)
- p. 4 Table 2, Morphine
1622 (47%)
Only the authors can say what is right here.
Judge's note: Checked on the page 5 crop. With the full denominators, 182/3483 rounds to 5%. However, other rows in Table 3 (raised troponin) clearly use smaller denominators that are not stated, so a smaller day-3 denominator cannot be ruled out. Applying the same smaller denominator to both arms would keep RR 1.03, so the reviewer's RR argument does not settle this. The authors should state the day-3 denominator. The morphine item (1622 (47%)) is dropped: 1622/3488 = 46.5%, so 7 missing values would explain it.
Second check: another model agreed
Report a problem with this finding
Typo Table 2 footnote: 'Data are median (IQR), n (5), mean (SD), or p value.' — 'n (5)' should be 'n (%)'.
- p. 4 Table 2 footnote
n (5)
- p. 4 Table 2 footnote
Data are median (IQR), n (5), mean (SD), or p value.
- p. 4 Table 2 footnote
Data are median (IQR), n (5), mean (SD)
- p. 4 Table 2
2 (1–3) 1 (1·0–4·0)
Judge's note: The image shows 'n (5)', so this is a typesetting error, not an extraction artefact. The PACU ventilation formatting ('2 (1–3)' vs '1 (1·0–4·0)') is only an inconsistency in decimal places and is not counted.
Second check: none (judge only)
Report a problem with this finding
Typo Results: 'in 283 patients (8%) in patients assigned to nitrous oxide' — redundant 'in patients'.
- p. 5 Results, last paragraph, right column
in 283 patients (8%) in patients assigned to nitrous oxide
Judge's note: The page 5 image shows the wording as quoted. This is a copy-editing slip.
Second check: none (judge only)
Report a problem with this finding
Typo Table 4 footnote 'Data are n%.' — the rows are n (%) and the category rows are counts only.
- p. 6 Table 4 footnote
Data are n%.
Judge's note: The page 6 image shows 'Data are n%.' The notation is malformed and does not describe the count-only category rows.
Second check: none (judge only)
Report a problem with this finding
Typo Discussion: 'Quality of recovery scores was slightly reduced' — plural 'scores' with singular verb; the Results say 'quality of recovery scores were lower'.
- p. 7 Discussion, para 1 (continued)
recovery scores was slightly reduced
Judge's note: The wording is visible on the page 7 image, running across the column break from page 6. The phrase could be parsed with 'Quality' as head noun, but the paper itself uses 'scores were' in the Results. This is a cosmetic point only.
Second check: none (judge only)