Bispectral Index Monitoring to Prevent Awareness during Anaesthesia: The B-Aware Randomised Controlled Trial
The Lancet, 2004 · paper · Clinical trials
A randomized trial of 2463 high-risk surgical patients testing whether brain-wave-based bispectral index monitoring reduces awareness during general anaesthesia. Confirmed awareness occurred in 2 patients monitored versus 11 with routine care, an 82% reduction in risk.
Checked: PDF from medic.usm.my, fetched 4 Oct 2026. Page numbers are those of that PDF.
6 findings
Substance Trial profile prints '1263 assigned routine care'. With '1248 assigned BIS monitoring' that totals 2511, not '2503 randomised'. Results text exclusions (13 cancelled, 2 consent withdrawn, 2 under-age) plus 1238 analysed give 1255.
- p. 2 Trial profile figure
1263 assigned routine care
- p. 2 Trial profile figure
2503 randomised
- p. 2 Trial profile
1263 assigned routine
- p. 4 Results
withdrawal of consent (six, two)
Judge's note: Checked on the page-2 image: the box reads 1263. The BIS arm reconciles (1225+13+6+4=1248). The text on page 4 gives 40 withdrawn and 2463 analysed, which matches 1248+1255=2503. Nothing in the paper explains 1263.
Second check: another model agreed
Report a problem with this finding
Substance Results text says "Of the 1227 patients in the BIS group, 650 received anaesthesia with the BIS below 60 at all times", but the BIS group has 1225 patients everywhere else (Table 4 and 5 headers, results page 4).
- p. 5 Results, BIS data paragraph
Of the 1227 patients in the BIS group, 650 received
- p. 5 Results, BIS readings paragraph
Of the 1227 patients in the BIS group
Only the authors can say what is right here.
Judge's note: Page 5 image confirms 1227. Page 4 gives 1225 randomised to BIS; 14 of them did not receive BIS (1225-14=1211), and 6 routine-care patients received it by mistake (1211+6=1217). No set of 1227 can be derived from these figures. The intended number could be 1225 or the number actually monitored, so only the authors can say which.
Second check: another model agreed
Report a problem with this finding
Reference or label Trial-profile routine-care exclusion box lists only '13 surgery cancelled 2 under-age'. The Results text gives 'withdrawal of consent (six, two)', so 2 routine-care consent withdrawals are missing from the figure.
- p. 2 Trial profile, routine-care exclusion box
13 surgery cancelled 2 under-age
- p. 4 Results para 1
withdrawal of consent (six, two)
Judge's note: Page-2 image: the routine box shows only 13 cancelled and 2 under-age. Page-4 text: 13+13+6+2+4+0+0+2=40 withdrawn, which matches 2503-2463. The figure shows only 38. This is a separate omission from the wrong 1263 arm count in C003.
Second check: another model agreed
Report a problem with this finding
Reference or label Table 3 lists 16 RC rows and 21 rows in the BIS block, one of them unlabelled ('F 55 Open heart surgery — Thought she was awake during surgery, but could not remember anything'): 37 rows against the text's '36 reports … (20 patients in the BIS group vs 16 in the routine care group)'.
- p. 4 Table 3
Thought she was awake during surgery, but could not remember anything
- p. 5 Results, possible awareness
36 reports of possible awareness that were not confirmed by the study endpoint committee (20 patients in the BIS group vs 16 in the routine care group
- p. 6 Discussion, para 5
including 36 reports that were not deemed to represent
- p. 5 Results, para on dreaming/possible awareness
there were 36 reports of possible awareness that were not
Only the authors can say what is right here.
Judge's note: Rows counted on the page-4 image: 16 RC; 14 labelled BIS rows, then the unlabelled F/55 row, then 6 labelled BIS rows. Labelled rows alone give 36, which matches the text. The extra row has its own sex, age, surgery and narrative. The row above it ('Heard conversation, possibly in intensive care unit,') ends in a stray comma, which hints at a layout fault. Only the authors can say whether the row is extra or its label is missing.
Second check: another model agreed
Report a problem with this finding
Reference or label Table 3 row "F 55 Open heart surgery … Thought she was awake during surgery, but could not remember anything" has a blank Group cell. Counting rows gives 16 RC, 20 labelled BIS plus this unlabelled row (37 in all), while the text says "36 reports … (20 patients in the BIS group vs 16 in the routine care group)".
- p. 4 Table 3, row after 'Colonic resection'
F 55 Open heart surgery
- p. 5 Results
20 patients in the BIS group vs 16
Only the authors can say what is right here.
Judge's note: Checked the page 4 image and crop C016-a. The Group cell is empty only for the F/55 open-heart row. I counted the rows on the image: 16 RC and 20 labelled BIS, plus the unlabelled row, so 37 rows. Page 5 text gives 36 (20 vs 16). The row has its own sex, age and surgery, so it cannot be a continuation of the M/68 colonic resection row above it. Only the authors can say which group it belongs to and how it squares with the stated total of 36.
Second check: another model agreed
Report a problem with this finding
Reference or label Table 4 gives the fentanyl unit as mg: "Fentanyl (mg)† 500 (100–1000; n=863) 600 (100–1175; n=840)". The values fit µg.
- p. 5 Table 4, Opioid dose
Fentanyl (mg)†
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: clinical fentanyl doses are given in micrograms, and a median of 500 mg would be impossible. The image (crop C019-a) clearly prints "(mg)", so this is not a lost µ in extraction; the same page set shows µmol/L correctly on p3.
Second check: another model agreed
Report a problem with this finding
17 minor (typos and the like; most checked by the judge only)
Wording 'the absolute reduction in the risk of awareness was 0·74%. The number needed to treat (NNT) was 138'. 1/0.0074 is about 135. NNT 138 corresponds to the randomised denominators (11/1238 - 2/1225 = 0.725%), while 0·74% corresponds to the interviewed denominators.
- p. 4 Results, primary outcome
the absolute reduction in the risk of awareness was 0·74%. The number needed to treat (NNT) was 138
- p. 1 Abstract Interpretation
needed to treat of 138
- p. 7 Discussion, final para (continued)
138. At an acquisition cost of BIS monitoring of US$16 per
- p. 4 Results, primary outcome
in the risk of awareness was 0·74%. The number needed
Only the authors can say what is right here.
Judge's note: Page-4 image. 0·74% equals 0·91% minus 0·17%, which use the 1195/1203 patients with interview data. 138 equals 1/(11/1238 - 2/1225). The paper does not say that the ARR and the NNT use different denominators. The NNT of 138 also feeds the US$2200 cost estimate in the Abstract and Discussion. Query which analysis set each figure comes from.
Second check: another model agreed
Report a problem with this finding
Wording Table 1 row 'Excluding ICU patients (n=1123)': 1123 is the PACU count (547+576), which also leaves out the 62 HDU patients. Excluding only the 1272 ICU patients would leave about 1191.
- p. 2 Table 1
Excluding ICU patients (n=1123)
Only the authors can say what is right here.
Judge's note: Table 5 (page 5): PACU 547/576, HDU 37/25, ICU 639/633. The results text calls 1123 'patients admitted to the post-anaesthesia care unit'. The row label describes a larger population than the one counted, so it would read better as 'PACU patients'.
Second check: another model agreed
Report a problem with this finding
Wording Methods: 'Recovery times were measured from completion of wound dressing'. Discussion: the start of recovery was defined as the moment 'when wound closure was complete'.
- p. 3 Methods, secondary outcomes
Recovery times were measured from completion of wound dressing
- p. 6 Discussion
no longer needed: when wound closure was complete
Only the authors can say what is right here.
Judge's note: Page 3 (Methods) and page 6 (Discussion) images. The same Discussion paragraph treats wound dressing and wound closure as distinct time points, so the two definitions of time zero conflict. Only the authors can say which was used.
Second check: another model agreed
Report a problem with this finding
Wording Results say that excluding "patients recalling intraoperative events but not the pain of surgery" leaves "two patients in the BIS group and nine in the routine care group", but Table 2 mentions pain for only about 5 of the 11 routine-care confirmed cases.
- p. 5 Results
leaving two patients in the BIS group and nine in the routine care group
- p. 3 Table 2
Heard the surgeon talking during surgery
- p. 6 Discussion, para 4
were several reports of awareness without recollection of
- p. 5 Results, para 1
the BIS group and nine in the routine care group, there
Only the authors can say what is right here.
Judge's note: Checked the Table 2 image on page 3. Pain is mentioned for the RC Whipple, hepatoma, F67 coronary, aneurysm and embolectomy cases. The single lung transplant was 'uncomfortable' and the oesophagectomy 'felt something'. The anterior resection, atrial tumour, valvular/coronary and heart-lung transplant cases mention only hearing. Both BIS cases mention pain, which fits the 2. The Discussion (p6) says 'several reports of awareness without recollection of the pain'. The OR of 0·22 does fit 2 vs 9. The classification may rest on fuller records than the table summaries, so this needs clarification from the authors.
Second check: another model agreed
Report a problem with this finding
Wording Table 5 "At 30 days (n with data) 1126 1150" is larger than the number interviewed at 30 days in the text, "2243 at 30 days (1114, 1129)".
- p. 5 Table 5, Patients' satisfaction at 30 days
At 30 days (n with data) 1126 1150
- p. 4 Results para 1
2243 at 30 days (1114, 1129)
- p. 5 Table 5, At 30 days (n with data)
1126
- p. 5 Table 5
At 30 days (n with data)
Only the authors can say what is right here.
Judge's note: Checked the page 4 and 5 images. The 24 h counts (1161, 1169) match the text exactly, but the 30-day satisfaction counts are higher than the number interviewed. The 30-day percentages in Table 5 do fit 1126/1150 (751/1126 is 67%, 781/1150 is 68%). I searched the text and found no other source for satisfaction data at 30 days.
Second check: another model agreed
Report a problem with this finding
Wording Several Table 5 complication percentages do not match the column denominators (1225/1238): sepsis "222 (18·3%)" and "221 (18·0%)", any of the above "283 (23·2%)" and "288 (23·4%)", acute renal failure "55 (4·5%)", and death "50 (4·1%)".
- p. 5 Table 5, Postoperative complications
222 (18·3%)
- p. 5 Table 5, Postoperative complications
283 (23·2%)
- p. 5 Table 5 Acute renal failure
55 (4·5%)
Only the authors can say what is right here.
Judge's note: Checked the page 5 image. MI, stroke, BIS renal failure and BIS death fit n=1225/1238. The routine-care figures fit a denominator of about 1229, which suggests some data were missing. In the BIS column, the denominators implied by sepsis (about 1213) and by any complication (about 1220) differ. There is no footnote on missing data, so the authors need to state the denominators.
Second check: another model agreed
Report a problem with this finding
Wording Table 5 destination after surgery does not add up to the group n: BIS 547+37+639=1223 (n=1225), routine care 576+25+633=1234 (n=1238). Missing data are not footnoted.
- p. 5 Table 5
576 (47%)
Judge's note: The page 5 image confirms the counts. The ICU numbers match the text (1272 = 639+633), and the difference is probably missing destination data, but the table does not say so.
Second check: another model agreed
Report a problem with this finding
Wording The Discussion says "higher doses of midazolam and propofol … were administered to this group", but in Table 4 the propofol induction dose is 80 mg in both groups (p=0·752) and the infusion rate is 5 mg/kg per h in both (p=0·384). Only the TIVA target plasma concentration differed (2 vs 2·4 mg/L).
- p. 6 Discussion, para 3
since higher doses of midazolam and propofol, drugs with
- p. 5 Table 4, Propofol rows
80 (50–150)
- p. 6 Discussion
higher doses of midazolam and propofol
- p. 5 Table 4
80 (20–150)
Judge's note: Checked Table 4 on the page 5 image and the Discussion on page 6. The Results text correctly says 'target concentration of propofol'. Calling this a higher propofol dose is loose wording; it is defensible if a target concentration is read as a dose.
Second check: another model agreed
Report a problem with this finding
Typo Table 1, Extent of surgery, Minor, BIS group: '104 (9%)'. 104/1225 is 8.5%, which rounds to 8% (routine-care 104/1238 is printed as 8%).
- p. 2 Table 1, Extent of surgery, Minor
104 (9%)
- p. 2 Table 1, Extent of surgery
Minor 104 (9%)
Judge's note: Page-2 image shows 104 (9%). The column sums to 1225 (104+216+905), and the other rows round correctly (216 to 18%, 905 to 74%). It is a rounding misprint.
Second check: none (judge only)
Report a problem with this finding
Typo Table 1, Acute trauma with hypovolaemia, routine care: '34 (2·8%)'. 34/1238 is 2.75%, which rounds to 2.7%.
- p. 2 Table 1, Acute trauma with hypovolaemia, routine care
34 (2·8%)
Judge's note: Page-2 image shows 34 (2·8%). The other routine-care percentages fit n=1238 (92 gives 7.4%, 37 gives 3.0%, 98 gives 7.9%).
Second check: none (judge only)
Report a problem with this finding
Typo Misspelling 'were displayed for the anaesthestist' (should be 'anaesthetist').
- p. 2 Procedures
were displayed for the anaesthestist
- p. 2 Procedures, para 3
displayed for the anaesthestist
- p. 2
displayed for the anaesthestist; alarm limits were not
Judge's note: Visible in the page-2 image, Procedures, right column. The paper spells 'anaesthetist' correctly elsewhere, two lines further down.
Second check: none (judge only)
Report a problem with this finding
Typo Table 2, single lung transplant row: 'with a single-lumen tracheal tracheal tube' has a duplicated word.
- p. 3 Table 2, single lung transplant row
single-lumen tracheal tracheal tube
- p. 3
single-lumen tracheal tracheal tube near the end
- p. 3
ntubation with a single-lumen tracheal tracheal tube near the end o
Judge's note: The duplication is visible in the page-3 image of Table 2.
Second check: code check
Report a problem with this finding
Typo Table 5, satisfaction at 24 h, BIS group: "Somewhat or less than satisfied 52 (5%)". 52/1161 is 4.5%, which rounds to 4%, and the column then sums to 101%.
- p. 5 Table 5, Patients' satisfaction at 24 h
52 (5%)
Judge's note: Page 5 image confirms 52 (5%) with n=1161. The other cells in the column round correctly (773/1161 is 67%, 336/1161 is 29%).
Second check: none (judge only)
Report a problem with this finding
Typo Table 4, marked hypotension, all cases, BIS group: "717 (58%)". 717/1225 is 58.5%, which rounds to 59%.
- p. 5 Table 4, Marked hypotension, All cases
717 (58%)
Judge's note: Page 5 image confirms 717 (58%). The rest of the table rounds normally rather than truncating (179/1225 is 14.6%, printed 15%; 1172/1238 is 94.7%, printed 95%), so truncation does not explain it.
Second check: none (judge only)
Report a problem with this finding
Typo US spelling in the authors' own prose, Discussion p6: "unpredictably high anesthetic requirement"; the same paragraph and the rest of the article use 'anaesthetic'.
- p. 6 Discussion, para 2
unpredictably high anesthetic requirement
Judge's note: I checked the crop of p6. The sentence reads 'anesthetic' while the sentences around it use 'anaesthetic'. A text search finds 'anesthe' outside the reference titles only on line 1119. This is a cosmetic spelling inconsistency.
Second check: none (judge only)
Report a problem with this finding
Typo Trial group list, p7: "V Jiranantarat (Suriraj Hospital, Thailand)"; the usual spelling of the hospital is Siriraj.
- p. 7 B-Aware trial group list
Suriraj Hospital, Thailand
- p. 7 B-Aware trial group list
V Jiranantarat (Suriraj Hospital, Thailand)
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the Bangkok hospital's name is spelled 'Siriraj'. The p7 image does print 'Suriraj Hospital, Thailand'. Nothing in the files settles the spelling, so this is put to the authors.
Second check: none (judge only)
Report a problem with this finding
Typo Reference 27 prints "Aldrete J, Kroulike D. A postanesthetic recovery score. Anesth Analg 1970; 49: 924–34"; the co-author's name is usually given as Kroulik.
- p. 7 Reference 27
Aldrete J, Kroulike D.
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the second author of Anesth Analg 1970;49:924–34 is D Kroulik. The p7 image shows 'Kroulike'. The rest of the citation is not in question.
Second check: none (judge only)