Donanemab in Early Symptomatic Alzheimer Disease: The TRAILBLAZER-ALZ 2 Randomized Clinical Trial
2023 · paper · Clinical trials
An 18-month randomized trial of donanemab versus placebo in 1,736 people with early symptomatic Alzheimer disease. Donanemab slowed clinical decline at 76 weeks, but brain-swelling imaging abnormalities occurred in 24% of treated participants, and three treatment-related deaths occurred.
3 findings
Substance In the combined-population results, the text pairs the MMRM CDR-SB difference "−0.70 (95% CI, −0.95 to −0.45)" with "28.9% [95% CI, 18.26%-39.53%]". That slowing CI belongs to the NCS2 row of Table 2; the MMRM row and the Figure 2D legend give 18.41 to 39.44.
- p. 6 Results, Secondary Outcomes, combined population paragraph
−0.70 (95% CI, −0.95 to −0.45) (28.9% [95% CI, 18.26%-39.53%] slowing
- p. 10 Table 2, CDR-SB Combined, MMRM row
(18.41 to 39.44)
- p. 9 Figure 2 legend, panel D
18.41%-39.44%) of clinical progression
- p. 6 Results, Secondary Outcomes, combined population
(28.9% [95% CI, 18.26%-39.53%] slowing of clinical pro-gression) for CDR-SB
Judge's note: Table 2 on p10 (upright image) shows NCS2 −0.67 (−0.92 to −0.43) with 28.9 (18.26 to 39.53), and MMRM −0.70 (−0.95 to −0.45) with 28.9 (18.41 to 39.44). The Figure 2 legend on p9 gives D 28.9% (18.41%-39.44%). The low/medium paragraph uses the MMRM row consistently (−0.67, −0.95 to −0.40; 36.0%, 20.76-51.15), so the combined sentence mixes rows from the two models.
Second check: another model agreed
Report a problem with this finding
Substance Figure 3F at-risk table, donanemab row, prints "696" at both 180 d and 240 d, although the low/medium tau subset (Figure 3E) falls from 492 to 470 over the same interval.
- p. 12 Figure 3F, No. of participants at risk, Donanemab row
696
Only the authors can say what is right here.
Judge's note: Read the at-risk tables in Figure 3E and 3F on p12. The low/medium population is a subset of the combined population, so the combined donanemab count must fall by at least 22 between 180 d and 240 d. Every other interval obeys this: donanemab 64≥38, 41≥22, 121≥58; placebo 180→240 d is 29≥15. One of the two 696 values is evidently wrong, and only the authors can say which.
Second check: another model agreed
Report a problem with this finding
Reference or label The Plasma P-tau217 section reads "The difference in LSM change in tau SUVR (log10-based) vs placebo was −0.25"; the quantity is log10 plasma P-tau217, not tau SUVR.
- p. 6 Results, Exploratory Outcomes, Plasma P-tau217
The difference in LSM change in tau SUVR (log10-based) vs placebo was −0.25
Judge's note: On p6 the paragraph sits under the Plasma P-tau217 heading and cites Figure 3C and D, whose y-axes on p12 are log10 plasma P-tau217 and whose curves end near −0.22 and −0.19. Tau SUVR is the PET outcome in the preceding section, where the differences are −0.0002 and −0.0041. The label is wrong; the numbers are not.
Second check: another model agreed
Report a problem with this finding
10 minor (typos and the like; most checked by the judge only)
Reference or label "Serious amyloid-related imaging abnormalities of edema/effusion (see Table 3) occurred in 13 participants (1.5%)": Table 3 has no row for serious ARIA-E.
- p. 9 Results, Adverse Events, text below Figure 2
Serious amyloid-related imaging abnormalities of edema/effusion (see Table 3) occurred in 13 participants (1.5%)
Judge's note: Checked Table 3 on p13. It lists ARIA-E overall, asymptomatic, symptomatic and by APOE ε4 status, plus serious AEs overall, but no serious ARIA-E count, so the 13 (1.5%) cannot be found where the text points. The pointer may be meant for ARIA-E in general.
Second check: another model agreed
Report a problem with this finding
Wording "Most ARIA-E events were mild to moderate (see eTable 2 ...) (n = 188 [93.1%] in the donanemab group": 188 of the 205 participants with ARIA-E is 91.7%, and the denominator behind 93.1% is not stated.
- p. 8 Adverse Events
n = 188 [93.1%] in the donanemab group
- p. 8 Adverse Events, ARIA-E severity
n = 188 [93.1%] in
Only the authors can say what is right here.
Judge's note: Read on p8, next to "205 participants (24.0%)". Table 3 on p13 gives donanemab ARIA-E as 205 under both the TEAE and the MRI definitions, so neither gives 93.1% (188/202 would). The placebo figure (17 [100%]) uses the TEAE count of 17 rather than the MRI count of 18, which suggests the denominators come from eTable 2. That supplement is not in the files, so the denominator should be queried with the authors.
Second check: another model agreed
Report a problem with this finding
Wording The Discussion says high tau participants "demonstrated no differences (P< .05)"; as printed, the parenthetical states the condition for a significant difference.
- p. 14 Discussion, paragraph 5
tau participants demonstrated no differences (P< .05) on
- p. 14 Discussion
demonstrated no differences (P< .05)
- p. 14 Discussion paragraph on high tau participants
demonstrated no differences (P< .05) on
- p. 14 Discussion, para on high tau
no differences (P< .05)
Judge's note: Seen on p14. The intended meaning is presumably 'no significant differences at the P < .05 level'; Results p8 gives the high-tau iADRS P = .42. The wording is ambiguous rather than a data error.
Second check: another model agreed
Report a problem with this finding
Typo The Tau PET section cites "(eFigure 5 in Supplement)" without a supplement number, while every other pointer says "Supplement 3".
- p. 6 Results, Tau PET
(eFigure 5 in Supplement)
Judge's note: Seen on the p6 image. The target is evidently Supplement 3, so the pointer is incomplete but still usable.
Second check: none (judge only)
Report a problem with this finding
Typo In "(CDR-G hazard ratio, 0.63 [95% CI, 0.51-0.77; P < .001)" the square bracket is never closed; it should read "0.51-0.77];".
- p. 6 Time-Based Analyses
0.63 [95% CI, 0.51-0.77; P < .001)
Judge's note: Seen on p6 (crop C007-a). The parallel low/medium clause is correctly written as "0.47-0.80]; P < .001)".
Second check: none (judge only)
Report a problem with this finding
Typo In the combined-population paragraph, "for ADAS-Cog13 (Figure 2, Table 2; eFigures 2 and 4" prints 13 on the baseline, whereas the preceding paragraph and Table 2 set it as a subscript (ADAS-Cog₁₃).
- p. 6 Secondary Outcomes
for ADAS-Cog13 (Figure 2, Table 2; eFigures 2 and 4
Judge's note: Compared the C007-a crop with the low/medium paragraph on p6. The image confirms the inconsistency, so it is not an extraction artefact. Cosmetic only.
Second check: none (judge only)
Report a problem with this finding
Typo Table 2, iADRS low/medium tau, placebo 76-week observed Mean (SD): "98.88 (17.95)" in the NCS2 row but "98.88 (17.96)" in the MMRM row.
- p. 10 Table 2, iADRS Low/medium tau, placebo 76 Weeks, NCS2 vs MMRM rows
98.88 (17.95)
- p. 10 Table 2, iADRS Low/medium tau, placebo MMRM row
98.88 (17.96)
Only the authors can say what is right here.
Judge's note: Seen on the p10 upright image. Every other paired observed Mean (SD) cell is identical between the NCS2 and MMRM rows (same n = 444), so this 0.01 discrepancy looks like a transcription slip. Only the authors can say which value is right.
Second check: none (judge only)
Report a problem with this finding
Typo The conflict-of-interest disclosure says Dr Collins was an "employee of and stockholder in from Eli Lilly and Company"; the word "from" is superfluous.
- p. 15 Conflict of Interest Disclosures, Dr Collins
stockholder in from Eli Lilly and Company
Judge's note: Seen in the p15 crop. The stray preposition is in the typeset text, so it is not an extraction artefact.
Second check: none (judge only)
Report a problem with this finding
Typo Company names are misspelled in the disclosures: Dr Apostolova's entry reads "GE Healthcare, Eisa, Roche" and Dr Hansson's reads "Fujirebio, Merk,". They should read Eisai and Merck.
- p. 15 Conflict of Interest Disclosures, Dr Apostolova
GE Healthcare, Eisa, Roche
- p. 15 Conflict of Interest Disclosures, Dr Hansson
Fujirebio, Merk
- p. 15 Disclosures, Dr Apostolova
Siemens, Corium, GE Healthcare, Eisa, Roche
- p. 15 Disclosures, Dr Hansson
Fujirebio, Merk,
Judge's note: I checked the page 15 crops. The page image shows "Eisa" in Apostolova's personal-fees list and "Merk" in Hansson's list, so these are not extraction artefacts. The same section spells "Eisai" correctly in the Salloway, Hansson and Ritchie entries, and "Merck" correctly in Ritchie's entry. The intended spellings are therefore clear from the paper itself.
Second check: none (judge only)
Report a problem with this finding
Typo Dr Apostolova's disclosure twice names the organisation without the apostrophe ("grants from NIA, Alzheimer Association," and "Alnylam, Alzheimer Association, and"). Elsewhere the paper writes "Alzheimer's Association".
- p. 15 Disclosures, Dr Apostolova
Alzheimer Association, AVID Radiopharmaceuticals
Judge's note: The page 15 image shows "Alzheimer Association" twice in Apostolova's entry. The paper itself uses "Alzheimer's Association" on page 2 (line 82 of the text extract, 'Alzheimer's Association Research Framework') and in the references ('National Institute on Aging-Alzheimer's Association'). So the form is inconsistent within the paper. This is a cosmetic punctuation point.
Second check: none (judge only)