Tirzepatide Once Weekly for the Treatment of Obesity
2022 · paper · Clinical trials
A 72-week randomized trial of once-weekly tirzepatide versus placebo in 2539 adults with obesity, without diabetes. Average weight fell 15.0% to 20.9% across doses versus 3.1% with placebo; gastrointestinal side effects were most common.
2 findings
Substance Table 2, difference from placebo in change in waist circumference, 5-mg column: "−10.1 (−1.6 to −8.6)". The lower bound is above the upper bound and the interval does not contain the estimate (probably −11.6).
- p. 7 Table 2, 'Difference from placebo in change in waist circumference — cm', Tirzepatide 5 mg column
−10.1 (−1.6 to −8.6)
Judge's note: The page 7 image shows '−10.1 (−1.6 to −8.6)'. −1.6 > −8.6, so the bounds are out of order, and −10.1 lies outside the interval. The upper half-width is 1.5. The 10-mg and 15-mg intervals in the same row are about ±1.45, which points to a lost leading digit (−11.6).
Second check: another model agreed, code check
Report a problem with this finding
Substance Table 3, systolic blood pressure, placebo least-squares mean: "−1.0 (−2.3 to −0.3)". The interval is not centred on the estimate; an upper bound of +0.3 would be.
- p. 8 Table 3, Systolic blood pressure — mm Hg, Placebo column
−1.0 (−2.3 to −0.3)
Only the authors can say what is right here.
Judge's note: The page 8 image shows '−1.0 (−2.3 to −0.3)'. The estimate does lie inside the interval (one reviewer said otherwise, which is wrong), but the interval is lopsided: 1.3 below and 0.7 above. Systolic blood pressure is analysed in mm Hg with no log transform; footnote ‖ applies only to lipids and insulin. Every other untransformed interval in the row and table is symmetric. The pooled and difference interval widths imply a placebo half-width of about 1.34, which matches −2.3 to +0.3. A lost sign is likely, but only the authors can confirm.
Second check: another model agreed
Report a problem with this finding
8 minor (typos and the like; most checked by the judge only)
Substance Table 3, non-HDL cholesterol, placebo: "−2.3 (−4.9 to −0.2)". This log-scale interval is lopsided in the opposite direction to back-transformation, by more than rounding can explain.
- p. 8 Table 3, Non-HDL cholesterol — mg/dl, Placebo column
−2.3 (−4.9 to −0.2)
Only the authors can say what is right here.
Judge's note: The page 8 image confirms the values. Footnote ‖: lipids were analysed after log transformation. On the log scale the lower bound is 0.0270 from ln(0.977) and the upper bound 0.0213. Back-transformation should make the percentage interval longer above the estimate, not shorter. The other placebo lipid intervals (total cholesterol, HDL) are symmetric on the log scale within about 0.0005. The pooled (log half-width ≈0.0116) and difference (≈0.028) intervals imply a placebo log half-width of about 0.026, which gives roughly −4.8 to +0.2. A lost sign on the upper bound is possible; the authors should check.
Second check: another model agreed
Report a problem with this finding
Wording Table 2 places the ≥25% weight-reduction row under the heading "Additional secondary end point", but its footnote ‖ says "This was an exploratory end point not controlled for type 1 error", and the Results and Discussion call it a "prespecified exploratory end point".
- p. 7 Table 2, section heading above the ≥25% row
Additional secondary end point
- p. 7 Table 2, footnote ‖
This was an exploratory end point not controlled for type 1 error
- p. 4 Results, Change in Body Weight
for the prespecified exploratory end point of a reduction in body weight of 25% or more
- p. 9 Discussion, paragraph 3
met the prespecified exploratory end point of weight reduction of 25% or more
Only the authors can say what is right here.
Judge's note: The crops of page 7 (heading and footnote ‖), page 4 and page 9 confirm both labels. Footnote ¶ uses 'additional secondary end points' for a different category (the 5-mg targets). So the tier given to the ≥25% end point is inconsistent within the table.
Second check: another model agreed
Report a problem with this finding
Typo Introduction reads "A molecule that combines both GIP and GLP receptor agonism"; everywhere else the paper says GLP-1 receptor.
- p. 2 Introduction, paragraph 2
GIP and GLP receptor agonism theoretically may
- p. 2 Introduction, last sentence of paragraph 2
both GIP and GLP receptor agonism
Judge's note: The page 2 crop shows "GIP and GLP receptor agonism". The next sentences say "GIP and GLP-1 receptors" and "GLP-1 receptors". The '-1' is missing.
Second check: none (judge only)
Report a problem with this finding
Typo Participants: "who reported one or more unsuccessful dietary effort to lose weight"; 'one or more' needs the plural 'efforts'.
- p. 2 Participants
one or more unsuccessful dietary effort
Judge's note: The page 2 crop shows the singular 'effort' after 'one or more'. Grammatical slip.
Second check: none (judge only)
Report a problem with this finding
Typo Statistical Analysis: "...in clinical trials (R1)14)," has an unmatched closing parenthesis.
- p. 3 Statistical Analysis
clinical trials (R1)14)
Judge's note: Full page 3 image. The parenthesis opened at "(Interna-" closes at "[E9])", and "(R1)" is balanced on its own. That leaves the ')' after superscript 14 with nothing to close. Later the text uses "ICH E9 [R1] addendum14". Punctuation only.
Second check: none (judge only)
Report a problem with this finding
Typo Efficacy estimand placebo change: "(95% CI, –3.2 to–1.6)" has no space between 'to' and the number.
- p. 4 Efficacy estimand paragraph
−3.2 to−1.6
Judge's note: The page 4 crop shows 'to–1.6' with no space, so this is in the PDF, not an extraction artefact. Cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo Table 3, SF-36 treatment difference: "1.9 (1.0 to 2. 9)" has a stray space inside the number 2.9.
- p. 8 Table 3, SF-36 row
1.9 (1.0 to 2. 9)
Judge's note: The page 8 crop shows '2. 9' with a visible space in the typeset table. Cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo Reference 5 author list: "Aronne LJ, Hall KD, M Jakicic J, et al.". The initial comes before the surname, unlike every other entry.
- p. 12 Reference 5
Hall KD, M Jakicic J
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the author's name is John M. Jakicic ('Jakicic JM'). The page 12 crop shows 'M Jakicic J', which breaks the 'Surname Initials' format used everywhere else in the reference list.
Second check: none (judge only)