Semaglutide and Cardiovascular Outcomes in Obesity without Diabetes
2023 · paper · Clinical trials
A randomized, placebo-controlled trial of weekly semaglutide in 17,604 adults with cardiovascular disease and overweight or obesity but no diabetes. Cardiovascular death, heart attack, or stroke occurred in 6.5% versus 8.0%; more semaglutide patients stopped treatment for adverse events.
1 finding
Reference or label Table 1 and Table 3 footnotes: "To convert the values for glycated hemoglobin to millimoles per mole, multiply by 10.929 and subtract 2.15". Applied as written, this does not reproduce the paper's own "6.5% (48 mmol per mole)"; 2.15 has to be subtracted before multiplying.
- p. 5 Table 1 footnote *
multiply by 10.929 and subtract 2.15
- p. 8 Table 3 footnote *
multiply by 10.929 and subtract 2.15
- p. 2 Methods, Trial Population
level of 6.5% (48 mmol per mole)
Judge's note: Both footnotes print the same wording (page 5 image and the Table 3 crop on page 8). The Methods (p.2) and the Table 2 footnote †† give 6.5% = 48 mmol per mole. As written, 6.5*10.929-2.15 gives 68.9. In the reverse order, (6.5-2.15)*10.929 gives 47.5, which rounds to 48. The formula's order of operations is reversed.
Second check: another model agreed
Report a problem with this finding
3 minor (typos and the like; most checked by the judge only)
Wording Discussion (p.10): "the 37.8-percentage-point decrease in the high-sensitivity C-reactive protein level with semaglutide". Table 3 reports −37.82% as a relative between-group difference of log-transformed values, not percentage points and not the within-semaglutide change (−39.12%).
- p. 10 Discussion, paragraph on biomarkers
the 37.8-percentage-point de-crease in the high-sensitivity C-reactive protein level with semaglutide
- p. 8 Table 3, High-sensitivity CRP row and footnote *
–37.82 (–39.70 to –35.90)
- p. 10 Discussion, paragraph 3
the 37.8-percentage-point de-crease in the high-sensitivity C-reactive protein level
Judge's note: Table 3 (p.8) gives CRP −39.12 (semaglutide), −2.08 (placebo) and −37.82 for the difference. The footnote says CRP was log-transformed and the results are 'relative changes (i.e., percentage changes)'. The ratio (1-0.3912)/(1-0.0208) = 0.6217 reproduces −37.8%. A simple subtraction of the arm values gives −37.04 percentage points. The 'percentage-point' wording therefore misdescribes the quantity. The number itself is correct.
Second check: another model agreed
Report a problem with this finding
Typo Table 2 footnote ‡‡: "urinary albumin-to-creatine ratio, >300 mg per gram" should read "albumin-to-creatinine ratio".
- p. 7 Table 2 footnote ‡‡
urinary albumin-to-creatine ratio, >300 mg per gram
Judge's note: The crop of the page 7 footnote shows 'creatine'. Macroalbuminuria is defined by the albumin-to-creatinine ratio. This is a spelling slip, not an extraction artefact.
Second check: none (judge only)
Report a problem with this finding
Typo Reference 24: "Andersen PK, Borgun O, Gill RD, Keiding N." The co-author's name is Borgan.
- p. 11 Reference 24
Andersen PK, Borgun O, Gill RD
- p. 11 Reference 24
Andersen PK, Borgun O, Gill RD,
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the co-author of 'Statistical Models Based on Counting Processes' (Springer, 1993) is Ørnulf Borgan. The crop of page 11 shows 'Borgun O', so this is not an extraction artefact.
Second check: none (judge only)