Abrocitinib versus Placebo or Dupilumab for Atopic Dermatitis
2021 · paper · Clinical trials
A phase 3 trial of 838 patients with atopic dermatitis comparing oral abrocitinib (two doses), dupilumab and placebo. Both abrocitinib doses beat placebo at week 12; only the 200-mg dose beat dupilumab, and only for itch at week 2.
0 findings
None that passed a second check.
6 minor (typos and the like; most checked by the judge only)
Wording Week-16 abrocitinib-vs-dupilumab comparisons are outside the multiplicity-controlled sequence, yet are described inferentially ("both doses of abrocitinib were not significantly different from dupilumab with respect to an EASI-75 response at week 16"; Discussion: 100-mg dose "not significantly different from dupilumab with respect to the three key secondary end points"). The 200-mg IGA comparison, which has the same status, is called one from which "no conclusion could be drawn".
- p. 4 Statistical Analysis, sequence B
sequence B was for the four comparisons of an IGA response (200-mg abrocitinib vs. placebo, and 100-mg abrocitinib vs. placebo) and an EASI-75 response
- p. 7 Table 2 footnote †
Comparisons between the dupilumab group and other trial groups were not multiplicity-controlled, except with respect to itch response at week 2.
- p. 8 Results, Key Secondary End Points
Therefore, both doses of abrocitinib were not significantly different from dupilumab with respect to an EASI-75 response at week 16.
- p. 11 Discussion, paragraph 1
dose of abrocitinib was not significantly differ-ent from dupilumab with respect to the three key secondary end points of the trial.
Only the authors can say what is right here.
Judge's note: Methods, p. 4: sequence B covers only the four week-16 comparisons against placebo. Table 2 footnote † (p. 7): "Comparisons between the dupilumab group and other trial groups were not multiplicity-controlled, except with respect to itch response at week 2." Footnote ‖: the dupilumab hierarchy failed at the 100-mg itch comparison (P=0.20), and no P values are given for any week-16 dupilumab comparison. Even so, Results p. 8, the Discussion on p. 11 and the Abstract use 'not significantly different' and 'superior … but not' language for the week-16 EASI-75 and IGA comparisons with dupilumab. For 200 mg vs dupilumab on IGA (9.4; 95% CI 0.4 to 18.5), the paper instead says that the lack of a CI-adjustment plan precludes conclusions. The wording is uneven across comparisons that the paper itself describes as having the same uncontrolled status, so the authors should say which inferential framework applies to the week-16 dupilumab comparisons. The minor spelling point 'Hierarchal' (footnote ‖) vs 'hierarchical' (p. 4) is also visible but is a variant spelling only.
Second check: another model agreed
Report a problem with this finding
Wording Table 3 row "Adverse event leading to study discontinuation" is described in the Results as "adverse events that led to discontinuation of the trial regimen".
- p. 10 Table 3, row 4
Adverse event leading to study discontinuation
- p. 8 Results, Safety paragraph 1
adverse events that led to discontinua-tion of the trial regimen were similar across the trial groups
Only the authors can say what is right here.
Judge's note: The page-10 crop shows the Table 3 label "study discontinuation". The page-8 Safety text calls the same counts discontinuation "of the trial regimen". Elsewhere the paper treats withdrawal from the trial and discontinuation of the regimen as different events (for example, the patient "withdrawn from the trial"; nonresponse imputation after withdrawal). Nothing in the files settles which event the row counts, so the authors should clarify.
Second check: another model agreed
Report a problem with this finding
Typo Drug name misspelled in sequence A description: "100-mg abrocti-nib vs. dupilumab" (should be "abrocitinib").
- p. 4 Statistical Analysis, paragraph 2 (sequence A)
100-mg abroctinib vs. dupilumab
Judge's note: The page-4 crop shows "abrocti-" at the end of the line and "nib" on the next line, so the printed word is "abroctinib" with the second "i" missing. Everywhere else the article prints "abrocitinib". Cosmetic misspelling.
Second check: none (judge only)
Report a problem with this finding
Typo Misplaced hyphen: "modified intention-to treat-popu-lation" should read "modified intention-to-treat population".
- p. 4 Statistical Analysis, paragraph 3
modified intention-to treat-popu-lation
Judge's note: The page-4 crop shows "intention-to treat-popu-" with "lation" on the next line. The hyphen in "popu-lation" is a line break, but the hyphen between "treat" and "popu" sits where the hyphen between "to" and "treat" belongs. Cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo Table 2 last row is labelled "Difference from dupilumab (95% CI)" without the "— percentage points" unit that every other difference row carries.
- p. 7 Table 2, last row
Difference from dupilumab (95% CI)
Judge's note: On the page-7 image, every other difference row in Table 2 reads "(95% CI) — percentage points". Only the final EASI-75 week-16 dupilumab row leaves out the unit. The unit can be inferred, so this is a cosmetic inconsistency.
Second check: none (judge only)
Report a problem with this finding
Typo Stray word in Figure 2 legend: "Panel A shows the percentages of patients in who had an Investigator's Global Assessment (IGA) response" (delete "in").
- p. 9 Figure 2 legend
Panel A shows the percentages of patients in who had
Judge's note: The page-9 crop shows the text as quoted. Panel B's description reads "percentages of patients who had". Cosmetic.
Second check: none (judge only)