Efficacy and Safety of Upadacitinib vs Dupilumab in Adults with Moderate-to-Severe Atopic Dermatitis
JAMA Dermatology, 2021 · paper · Clinical trials
A head-to-head phase 3b trial of upadacitinib versus dupilumab in 673 adults with moderate-to-severe atopic dermatitis. At week 16, 72.4% on upadacitinib and 62.6% on dupilumab reached 75% skin improvement; serious infections were higher with upadacitinib.
11 findings
Substance Table 1, Sex, Male, upadacitinib: '183 (53.2)'. But 183/342 = 53.5%, which is the figure in the abstract and text ('183 men [53.5%]'), and the female row '159 (46.5)' sums to 100 only with 53.5.
- p. 4 Table 1, Sex, Male, upadacitinib column
183 (53.2)
- p. 1 Abstract, Results
342 (183 men [53.5%]
- p. 3 Results, Patient Disposition
342 (183 men [53.5%]
- p. 3 Patient Disposition
183 men [53.5%]
Judge's note: Checked in the page images: Table 1 (p4) prints 183 (53.2), while the abstract (p1) and Patient Disposition (p3) print 53.5%. Same n = 342, and Table 1 is baseline data, so MI does not apply. The supplement N (344/348) point raised by one reviewer is handled under C003 and C010.
Second check: another model agreed
Report a problem with this finding
Substance Supplement 2 still shows the pre-correction arm sizes (eFigure 1 "UPA ... (N=348)" and "DUPI ... (N=344)"; eTables 4, 5 and 6 headers "N = 344 / N = 348"), whereas the corrected article uses n = 342 and n = 331 "after the removal of 19 patients from 1 center".
- p. 3 eFigure 1
UPA 30 mg tablets + PBO pre-filled syringe (N=348)
- p. 10 eTable 4 header
(N = 344) (N = 348)
- p. 12 eTable 6 header
(N = 344) (N = 348)
- p. 8 Correction notice
after the removal of 19 patients from 1 center
Only the authors can say what is right here.
Judge's note: The correction notice (main p8) records the February 14, 2024 removal of 19 patients but does not say Supplement 2 was updated. eTable 5 week-24 EASI75 (64.2% vs 59.5%) does not match the article text (223/342 = 65.3% vs 203/331 = 61.2%) that cites it. 348 + 344 - 19 = 673 = 342 + 331. Only the authors or journal can say whether an updated supplement exists.
Second check: another model agreed
Report a problem with this finding
Substance Text: '188 of 344 [56.1%] vs 117 of 323 [36.4%]'. The upadacitinib denominator 344 exceeds the 342 patients in the arm, and Table 2 gives No. = 334 for this end point.
- p. 4 Results text, top of left column
188 of 344 [56.1%]
- p. 5 Table 2, Worst Pruritus NRS improvement ≥4 points, No. row
334
- p. 4 Results, Efficacy Outcomes
188 of 344 [56.1%] vs 117 of 323
- p. 5 Table 2, NRS improvement ≥4 points, No.
323 334
Judge's note: Checked in the images: p4 text prints '188 of 344', and Table 2 (p5) 'No.' row prints 323 and 334. Footnote c restricts the analysis to patients with baseline NRS ≥4, a subset of the 342. 344 is impossible; 334 is evidently intended.
Second check: another model agreed, code check
Report a problem with this finding
Substance Serious infections, upadacitinib: Table 3 prints "4 (1.1)" and the text prints "(4 [1.1%] vs 2 [0.6%])", but 4/342 = 1.17%, which rounds to 1.2%. The same table prints 4 (1.2) for the drug-related SAE row with the same n.
- p. 7 Table 3, Serious infections, upadacitinib
4 (1.1)
- p. 4 Safety Outcomes text
(4 [1.1%] vs 2 [0.6%])
- p. 4 Safety Outcomes
serious infection (4 [1.1%] vs 2 [0.6%])
- p. 4 Safety Outcomes
4 [1.1%] vs 2 [0.6%]
Judge's note: Checked Table 3 (p7) and the Safety Outcomes text (p4). Table 3 states n = 342, and other cells round correctly (eg, 4 (1.2), 7 (2.0)). 1.1% matches the pre-correction denominator of 348: supp-02 eTable 6 (N = 348) also shows 4 (1.1). This looks like a leftover from the February 2024 correction that removed 19 patients; the correction notice does not fix it.
Second check: another model agreed
Report a problem with this finding
Substance Table 3: dupilumab headache is printed as "21 (6.0)", but 21/331 = 6.3%.
- p. 7 Table 3, TEAEs reported by ≥5%, Headache, dupilumab
21 (6.0)
Only the authors can say what is right here.
Judge's note: The Table 3 image confirms n = 331, and the other dupilumab cells match it (29 (8.8), 22 (6.6), 13 (3.9), 9 (2.7)). Neither 21/331 nor 21/344 gives 6.0, so 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
Reference or label eTable 1 criterion 4 cites "Hanifin and Rajka criteria9", but Supplement 2 has no reference list, and in the article reference 9 is Reich et al (AD Up), while Hanifin and Rajka is reference 11.
- p. 5 eTable 1, criterion 4
and Rajka criteria9
- p. 2 Methods, Patients
Hanifin and Rajka criteria11
Judge's note: supp-02 p5 image shows the superscript 9. Grep finds no reference list in supp-02. In the main reference list, 9 is Reich K et al and 11 is Hanifin JM, Rajka G; the main Methods cite it as "criteria11".
Second check: another model agreed
Report a problem with this finding
Reference or label Results text labels the Worst Pruritus NRS dispersion 'The mean (SD) percentage improvement ...' (p3: 32.0% [1.8%], 60.4% [2.2%], 67.8% [1.9%]) and 'mean (SD) Worst Pruritus NRS ... at week 24 (63.7% [2.8%] vs 54.9% [2.9%]' (p4), but these are SEs.
- p. 3 Results, Efficacy Outcomes
The mean (SD) percentage improvement from baseline Worst Pruritus NRS
- p. 4 Results, week 24 paragraph
mean (SD) Worst Pruritus NRS than dupilumab-treated patients at week 24 (63.7% [2.8%] vs 54.9% [2.9%]
- p. 5 Table 2 footnote b
Least-squares mean (SE) [95% CI]
- p. 8 Correction notice
to correct the indication of statistical variance from SD to SE in Table 2’s footnote b
Judge's note: Both 'mean (SD)' labels are visible in the page images. The same values appear as least-squares mean (SE) in Table 2 (footnote b, corrected to SE on July 24, 2024), in the abstract ('mean [SE], 32.0% [1.8%]') and in the p4 sentence 'mean (SE) Worst Pruritus NRS'. The Table 2 CI half-widths match 1.96×SE. The correction notice on p8 changes only Table 2 footnote b, so these two text labels were not corrected.
Second check: another model agreed
Report a problem with this finding
Reference or label eTable 1 criterion 14: "topical emollient treatments, described in Eligibility Criterion 8". In eTable 1 the emollient requirement is criterion 6; criterion 8 is the pregnancy test.
- p. 6 eTable 1, criterion 14, topical treatments bullet
described in Eligibility Criterion 8
- p. 5 eTable 1, criterion 6
Patient has applied a topical emollient (moisturizer) twice daily
- p. 6 eTable 1, criterion 14, last bullet
Topical treatments (with the exception of topical emollient treatments, described in Eligibility Criterion 8)
- p. 6
described in Eligibility Criterion 8)
Judge's note: Checked supp-02 pp5-6: criterion 6 is the emollient requirement and criterion 8 is the serum/urine pregnancy test.
Second check: another model agreed
Report a problem with this finding
Reference or label The text below eTable 3 cites "(eTable 5)" for the responder counts and "given in eTable 6" for the window distances; these are in eTable 2 and eTable 3. eTable 5 is week-24 efficacy and eTable 6 is TEAEs.
- p. 9 Text below eTable 3
WP-NRS ≥4 (eTable 5)
- p. 9 Text below eTable 3
analysis window are given in eTable 6
- p. 9 eTable 3 note
(eTable 5)
- p. 9 eTable 3 note
are given in eTable 6
Judge's note: supp-02 p8: eTable 2 gives the responder counts (1 DUPI, 3 UPA for EASI75 at week 16; 0 for WP-NRS). p9: eTable 3 gives the distances. p11: eTable 5 is week-24 efficacy. p12: eTable 6 is TEAEs.
Second check: another model agreed
Report a problem with this finding
Reference or label The Hepatitis Screen text on p43 cites "(Figure 1, Scenarios A and B)" and related scenarios four times, but the HBV scenario figure is "Figure 2. Interpretation and Management of HBV Serologic Test Results" (p44). Figure 1 is the Study Schematic (p14).
- p. 44
Figure 2.
Judge's note: Page 43 has four hyperlinked 'Figure 1, Scenario...' references. The scenario table is captioned Figure 2 on p44, and Figure 1 on p14 is the study schematic. The cross-references point to the wrong figure.
Second check: another model agreed
Report a problem with this finding
Wording The live-vaccine washout before first dose is "at least 14 days (or longer if locally required)" on p22, but "within 4 weeks prior to the first dose" in Eligibility Criterion 15 (p17), and the Section 5.4 Vaccines text (p25) recommends the live herpes zoster vaccine "at least 4 weeks before first dose".
- p. 22
at least 14 days (or longer if locally required) before first dose of study drug
Only the authors can say what is right here.
Judge's note: Compared on the page images. p22 (Vaccines paragraph) says live vaccinations must be completed at least 14 days before the first dose. p17 Criterion 15 says no live vaccine within 4 weeks before the first dose. p25 (Section 5.4) says the zoster vaccine should be given at least 4 weeks before. The post-dose window also differs: 12 weeks after the last injection on p22/p25 versus 4 weeks after the last dose in Criterion 15. Nothing in the files reconciles these windows, so only the authors can say which applies.
Second check: another model agreed
Report a problem with this finding
24 minor (typos and the like; most checked by the judge only)
Wording The Abstract and Discussion say "rates of conjunctivitis and injection-site reactions were higher" with dupilumab, but no injection-site reaction data appear in Table 3, the Results text or supplement eTable 6.
- p. 7 Discussion, first paragraph on page
the rates of conjunctivitis and injection-site reactions were numerically higher with dupilumab
- p. 1 Abstract, Results
whereas rates of conjunctivitis and injection-site reactions were higher for patients who received dupilumab
- p. 7 Table 3
Conjunctivitis
Only the authors can say what is right here.
Judge's note: Grep finds "injection-site" only in the Abstract and Discussion. Table 3 (p7) and eTable 6 (supp-02 pp12-13) have conjunctivitis rows but no injection-site reaction row. The claim may be true, but it cannot be checked against any data in the files.
Second check: another model agreed
Report a problem with this finding
Wording Text: 'Eighty-seven upadacitinib-treated patients (25.4%) and 85 dupilumab-treated patients (25.7%) received rescue therapy'. The cited eTable 4 (N = 344/348) gives 'Any Rescue Therapy' through week 24 as '85 (24.7)' and '88 (25.3)'.
- p. 3 Results, Patient Disposition
Eighty-seven upa- dacitinib-treated patients (25.4%) and 85 dupilumab-treated patients (25.7%)
- p. 10 eTable 4, Overall Through Week 24, Any Rescue Therapy
85 (24.7) 88 (25.3)
Only the authors can say what is right here.
Judge's note: The supplement image confirms that eTable 4 uses N = 344 (dupilumab) and N = 348 (upadacitinib), a total of 692 = 673 + 19. This matches the population before the February 2024 removal of 19 patients described in the correction notice. The supplement does not state that it uses the uncorrected population, so the article and the supplement it cites disagree. The text figures agree with Figure 1 (61+26=87; 70+15=85). Only the authors can confirm whether the supplement should have been updated.
Second check: another model agreed
Report a problem with this finding
Wording Week-24 text '(223 of 342 [65.3%] vs 203 of 331 [61.2%] ...) (eTable 5 in Supplement 2)', with EASI90 56.6%/49.5% and EASI100 27.8%/13.6%. The cited eTable 5 (N = 344/348) gives EASI75 64.2/59.5, EASI90 55.6/47.6 and EASI100 27.3/13.1.
- p. 4 Results, week 24
(223 of 342 [65.3%] vs 203 of 331 [61.2%]; P = .29
- p. 11 eTable 5, EASI75 row
59.5
- p. 11 eTable 5, EASI90 row
55.6
- p. 8 Correction notice
to add more data to eTable 5 in Supplement 2
Only the authors can say what is right here.
Judge's note: The eTable 5 image confirms the values and the N = 344/348 headers, ie the population before the February 2024 removal of 19 patients. The article does not say that the supplement uses that population. The reviewer's claim that eTable 5 should contain P values misreads the notice: the December 2021 correction added P values to the Results section. The 223/342 versus 65.3% gap is explained by NRI-C with MI (p3). Same underlying cause as C003, but a different table and different values.
Second check: another model agreed, code check
Report a problem with this finding
Wording 'The most frequently reported AE among patients treated with dupilumab was conjunctivitis (29 [8.8%])'. Table 3 shows 'Dermatitis atopic 29 (8.8)' tied with conjunctivitis in the dupilumab arm.
- p. 4 Safety Outcomes text
The most frequently reported AE among patients treated with dupilumab was conjunctivi- tis (29 [8.8%])
- p. 7 Table 3, Dermatitis atopic, dupilumab
29 (8.8)
Judge's note: The Table 3 image (p7) shows both rows at 29 (8.8) for dupilumab. The text does not say that AD worsening was excluded, so 'the most frequently reported' overstates what is a tie.
Second check: another model agreed
Report a problem with this finding
Wording Main text: "none of these events led to study drug discontinuation" (laboratory-related AEs). eTable 6 footnote g: "study drug was permanently discontinued in 1 UPA patient with anemia, 1 UPA patient with neutropenia, and 1 UPA patient with lymphopenia".
- p. 5 Safety Outcomes text
none of these events led to study drug discontinuation
- p. 13 eTable 6 footnote g
study drug was permanently discontinued in 1 UPA patient with anemia, 1 UPA patient with neutropenia, and 1 UPA patient with lymphopenia
Only the authors can say what is right here.
Judge's note: Page images main p5 and supp-02 p13 checked. eTable 6 covers the full monitoring period. Upadacitinib anemia and neutropenia events ran up to day 143 and day 140, so those discontinuations may have come after week 16. The lymphopenia events, however, were on days 1 and 23, both within 16 weeks. Table 3, which covers week 16, lists lymphopenia (0 vs 2) among the laboratory AEs. The main text does not mention lymphopenia, and it is unclear whether the footnote attributes the discontinuations to these AEs. Only the authors can clarify, so this is a query.
Second check: another model agreed
Report a problem with this finding
Wording Discussion: "significantly higher rates of clinically meaningful improvements in itch were reported as early as week 1", but the only week-1 itch end point is the LS-mean % change in Worst Pruritus NRS (−32.0 vs −8.9).
- p. 6 Discussion, paragraph 3
significantly higher rates of clinically meaningful improvements in itch were reported as early as week 1
- p. 5 Table 2, Week 1 row
−32.0 (1.8) [–35.4 to −28.5]
Only the authors can say what is right here.
Judge's note: Page images p5 and p6 checked. In Table 2 the ≥4-point (clinically meaningful) Worst Pruritus NRS responder end point appears only at week 16. At week 1 the table reports the mean percentage change. The paper presents no week-1 responder rates, so the wording describes a kind of result the paper does not report.
Second check: another model agreed
Report a problem with this finding
Wording Limitations: "with primary and ranked secondary end points at week 16". Table 2 lists ranked secondary end points at weeks 4, 2 and 1 as well as week 16.
- p. 7 Limitations, first paragraph
with primary and ranked secondary end points at week 16
- p. 5 Table 2, secondary end points in order of ranking
Week 2
Judge's note: In the Table 2 image (p5), the ranked secondary end points include % change in Worst Pruritus NRS at week 4, EASI75 at week 2 and % change in Worst Pruritus NRS at week 1. The phrase may mean the end points were assessed within the first 16 weeks, so this is a loose-wording point.
Second check: another model agreed
Report a problem with this finding
Wording The eTable 4 title is "Proportion of Patients Receiving 1 or More Rescue Therapy", but the caption is "eTable 4. Proportion of Patients Initiating Rescue Therapy ... before and after week 16". The table shows "Before Week 16" and "Overall Through Week 24".
- p. 10 eTable 4 title vs caption
eTable 4. Proportion of Patients Initiating Rescue Therapy
Judge's note: Seen in the supp-02 p10 image. Either wording could be legitimate, but the title and caption disagree, and "after week 16" does not match the "overall through week 24" section.
Second check: another model agreed
Report a problem with this finding
Typo Unbalanced parentheses: '(248 [72.4%] vs 207 [62.6%]; adjusted difference, 9.7% (95% CI, 2.6%-16.7%; P = .007) (Table 2)'. The outer parenthesis is never closed.
- p. 3 Efficacy Outcomes
adjusted difference, 9.7% (95% CI, 2.6%-16.7%; P = .007)
Judge's note: Checked in the p3 page image. Two parentheses are opened ('(248' and '(95% CI') and only one is closed before '(Table 2)'. In this house style the inner one would be square brackets.
Second check: none (judge only)
Report a problem with this finding
Typo Table 2 prints the week-16 Worst Pruritus NRS difference as "−18.21"; every other Difference entry has one decimal place.
- p. 5 Table 2, % change Worst Pruritus NRS week 16, Difference
−18.21
Judge's note: Page image p5, Table 2: the value reads −18.21, while the other differences (9.7, 20.4, 21.2, −28.0, 26.1, −23.0, 19.8) have one decimal. The value agrees with −67.8 minus −49.6, so only the formatting is inconsistent.
Second check: none (judge only)
Report a problem with this finding
Typo Safety text reads "creatinine phosphokinase elevations", while Table 3 and the abbreviation list read "Creatine phosphokinase".
- p. 5 Safety Outcomes
creatinine phosphokinase
- p. 7 Table 3
Creatine phosphokinase
Judge's note: The page images confirm 'creatinine phosphokinase' on p5 and 'Creatine phosphokinase elevation' and 'CPK, creatine phosphokinase' on p7. This is a misspelling in the text.
Second check: none (judge only)
Report a problem with this finding
Typo Table 2, first secondary row: "% Change from baseline in worst pruritus NRS" is lower case; the other rows read "Worst Pruritus NRS".
- p. 5 Table 2, first secondary end point row
in worst pruritus NRSb
Judge's note: Page image p5 confirms the inconsistent capitalisation within Table 2. Cosmetic only.
Second check: none (judge only)
Report a problem with this finding
Typo Table 2, EASI100 upadacitinib: "[23.6to 33.2]" has no space before "to"; also "−49.6(2.0)".
- p. 5 Table 2, EASI100 upadacitinib
[23.6to 33.2]
Judge's note: Page image p5 shows '[23.6to 33.2]' and '−49.6(2.0)' with the spaces missing. Not an extraction artefact.
Second check: none (judge only)
Report a problem with this finding
Typo Figure 2 legend: "Severity Index (EASI75) B, Proportion" and "in EASI (EASI90). C Proportion" lack the expected panel-label punctuation.
- p. 6 Figure 2 legend
Severity Index (EASI75) B, Proportion
- p. 6 Figure 2 legend
in EASI (EASI90). C Proportion
Judge's note: Page image p6, Figure 2 legend: a period is missing before 'B,' and a comma is missing after 'C'. Panels A and D are punctuated consistently.
Second check: none (judge only)
Report a problem with this finding
Typo "Several key inflammatory cytokines are involved in the pathogenesis of AD signal via JAK1" has two finite verbs; a word such as 'that' is missing.
- p. 6 Discussion, paragraph on cytokines
Several key inflammatory cytokines are involved in the pathogenesis of AD signal via JAK1
Judge's note: Page image p6 shows the sentence exactly as quoted.
Second check: none (judge only)
Report a problem with this finding
Typo "As the effectiveness of new drugs for psoriasis have improved" pairs the singular subject 'effectiveness' with the plural verb 'have'.
- p. 6 Discussion
As the effectiveness of new drugs for psoriasis have improved
Judge's note: Page image p6 confirms the wording. Grammatical slip.
Second check: none (judge only)
Report a problem with this finding
Typo Figure 2 legend: "(EASI90). C Proportion of patients achieving 100% improvement" lacks the comma used after the other panel letters ("A," "B," "D,").
- p. 6 Figure 2 legend
C Proportion of patients achieving 100% improvement
Judge's note: The page image shows "B, Proportion" and "D, Mean", but "C Proportion" has no comma.
Second check: none (judge only)
Report a problem with this finding
Typo Protocol synopsis: "Approximately 650 adults subjects with moderate to severe AD" should read "adult subjects".
- p. 6
Approximately 650 adults subjects with moderate to severe AD
Judge's note: Seen in the supp-03 p6 page image (Study Population row).
Second check: none (judge only)
Report a problem with this finding
Typo Dr Lio's disclosure reads "stock options from Micreos and; other royalties from patented product from Theraplex", with a stray "and" before the semicolon.
- p. 8 Conflict of Interest, Dr Lio
stock options from Micreos and; other royalties
Judge's note: Seen in the page image (p8, Conflict of Interest column).
Second check: none (judge only)
Report a problem with this finding
Typo The supplement text below eTable 3 has "For patients patients who were imputed", and the footnote below it reads "*Study Day day is defined ... or to the randomization date is if no study drug is taken."
- p. 9 Text below eTable 3
For patients patients who were imputed
- p. 9 Footnote below text
*Study Day day is defined
- p. 9 Footnote below text
date is if no study drug is taken
- p. 9
For patients patients who were imputed as responders for EASI 75 at Week
Judge's note: All three slips are visible in the supp-02 p9 image. The table's own footnote above has the clean wording.
Second check: code check
Report a problem with this finding
Typo Section 3.5 of the protocol repeats a phrase: "The following safety evaluations will be performed during the study performed during the study:"
- p. 12
will be performed during the study performed during the study:
Judge's note: I checked the crop of supp-03 p12 (Section 3.5 Safety Endpoints). The phrase 'performed during the study' does appear twice in a row in the printed text. This is a cosmetic duplication.
Second check: none (judge only)
Report a problem with this finding
Typo Reference 4 in the protocol reads "European Academy of Allergoly and Clinical Immunology"; this is a misspelling of 'Allergology'.
- p. 60
European Academy of Allergoly and Clinical Immunology
Judge's note: Crop of supp-03 p60, reference 4 (Akdis et al.). The printed word is 'Allergoly'. Cosmetic misspelling.
Second check: none (judge only)
Report a problem with this finding
Typo The abbreviation list gives "QTcF Friedericia's correction formula"; this misspells Fridericia, which is spelled correctly elsewhere in the protocol.
- p. 64
Friedericia's correction formula
Judge's note: Crop of supp-03 p64 shows 'Friedericia's'. The text layer has the correct spelling 'Fridericia's' at other places in the protocol (e.g., the QTcF > 500 msec criteria). Cosmetic misspelling.
Second check: none (judge only)
Report a problem with this finding
Typo In the Operations Manual Week 2 Central Lab list (p80), "Optional Biomarker: Whole blood RNA" appears twice in a row.
- p. 80
• Optional Biomarker: Whole blood RNA • Optional Biomarker: Whole blood RNA
Judge's note: Viewed the full page image of p80. The Week 2 Central Lab column lists 'Optional Biomarker: Whole blood RNA' as two consecutive identical bullets. The right column already lists the other biomarker items (proteomic, DNA, skin biopsies), so this is a duplicated entry.
Second check: none (judge only)