A Controlled Trial of Erenumab for Episodic Migraine
New England Journal of Medicine, 2017 · paper · Clinical trials
A six-month randomized trial of monthly erenumab (70 or 140 mg) versus placebo in 955 people with episodic migraine. Both doses reduced monthly migraine days and acute-medication use more than placebo, with similar adverse event rates.
3 findings
Substance Results text says "405 (42.4%) were currently using or had previously used migraine-preventive medication", but Table 1 rows 'Previous use only' (131, 133, 124) and 'Current use' (10, 9, 8) sum to 415 of 955 (43.5%).
- p. 6 Results, Patients
405 (42.4%) were cur-
- p. 4 Table 1, Migraine-preventive medication use, rows 'Previous use only' and 'Current use'
131 (41.1)
- p. 4 Table 1, row 'No current or previous use'
178 (55.8)
- p. 4 Table 1, Migraine-preventive medication use, Current use
10 (3.1)
Only the authors can say what is right here.
Judge's note: Read Table 1 on the page 4 image and the Results text on page 6. The three preventive-use rows sum to each column N (319, 317, 319), so the table is internally consistent; 955 minus the 540 with no use is also 415. The printed 42.4% matches 405/955, so count and percentage agree with each other but not with the table. Table 1 footnote § says the table uses actual data rather than randomisation stratification, so 405 may be the stratification-based count, but the text does not say so, and the same paragraph's 27 (2.8%) current users and 370 (38.7%) do match Table 1. Not explained in the article; the authors need to say which figure is intended, hence author_query.
Second check: another model agreed
Report a problem with this finding
Reference or label Discussion cites "other phase 2 and 3 trials of erenumab for the treatment of episodic and chronic migraine,22,24,28", but reference 28 is "Connor KM, Shapiro RE, Diener HC, et al. Randomized, controlled trial of telcagepant for the acute treatment of migraine. Neurology 2009;73:970-7", which is neither an erenumab trial nor a prevention trial.
- p. 9 Discussion, final paragraph
2 and 3 trials of erenumab for the treatment of episodic and chronic migraine,22,24,28
- p. 10 Reference 28
Randomized, controlled trial of tel- cagepant for the acute treatment of mi- graine. Neurology 2009
- p. 9 citation of reference 28
erenumab for the treatment of episodic and chronic migraine,
- p. 10 reference list, entry 28
Connor KM, Shapiro RE, Diener HC, et al. Randomized, controlled
Only the authors can say what is right here.
Judge's note: Confirmed the superscript 22,24,28 on the page 9 image and read references 22, 24 and 28 on page 10. References 22 (AMG 334, episodic, phase 2) and 24 (erenumab, chronic, phase 2) fit; 28 is a telcagepant acute-treatment trial. No entry in the 28-item reference list is a phase 3 erenumab trial, so 'phase 2 and 3' is also unsupported by the citations given. The mismatch is objective from the entry's own title, but the intended reference is not in the list, so only the authors can supply the fix.
Second check: another model agreed
Report a problem with this finding
Wording Discussion says erenumab "at a monthly dose of 70 mg or 140 mg resulted in a mean reduction in the frequency of migraine days that was significantly greater (by almost 2 days) than that associated with placebo"; Table 2 gives a difference of −1.4 days for 70 mg and −1.9 days for 140 mg, so 'almost 2 days' fits only the 140-mg dose.
- p. 8 Discussion, first paragraph
greater (by almost 2 days) than that associated
- p. 5 Table 2, Migraine days per month, Difference vs. placebo
−1.4 (−1.9 to −0.9)
- p. 5 Table 2, Migraine days per month, Difference vs. placebo, 140 mg
−1.9 (−2.3 to −1.4)
- p. 8 Discussion, para 1
significantly greater (by almost 2 days)
Judge's note: Read Table 2 on the page 5 image (minus signs confirmed) and the first Discussion paragraph on page 8. The differences are −1.4 (−1.9 to −0.9) and −1.9 (−2.3 to −1.4). The sentence names both doses with no 'up to', so it overstates the 70-mg effect. Wording issue.
Second check: another model agreed
Report a problem with this finding
3 minor (typos and the like; most checked by the judge only)
Wording Results text says "all three trial groups had a mean of 8.3 migraine days per month", but Table 1 gives 8.2±2.5 for placebo (8.3±2.5 for each erenumab group); 8.3 is the overall mean, as the abstract says ("8.3 in the overall population").
- p. 6 Results, Patients
all three trial groups had a mean of 8.3 migraine days per month
- p. 4 Table 1, row 'Migraine days per month', Placebo column
8.2±2.5
- p. 6 Results, Patients
all three trial groups had a mean of 8.3 migraine
- p. 1 Abstract, Results
8.3 in the overall
Judge's note: Checked the Results paragraph on page 6 and Table 1 on page 4 on the page images. Table 1 is for the full analysis set (all randomised) and the sentence gives no other analysis set. The weighted mean of the three groups is 8.27, which rounds to 8.3 overall, so the sentence should refer to the overall population rather than each group. The difference is 0.1 day in one group, so minor.
Second check: another model agreed
Report a problem with this finding
Wording Discussion says "38.7% of the patients had previously not had a beneficial effect from migraine-preventive drugs", but 38.7% (370/955) is the Table 1 'History of preventive treatment failure' count, which includes discontinuation for unacceptable side effects; lack of efficacy alone is 262/955 (27.4%).
- p. 9 Discussion, limitations paragraph
38.7% of the patients had previously not had a beneficial effect from migraine-preventive drugs
- p. 6 Results, Patients
370 (38.7%) had discontinued their use of a previous preventive medication because of insufficient efficacy or unacceptable side effects
- p. 4 Table 1, History of preventive treatment failure / Lack of efficacy
90 (28.2)
Judge's note: Read the limitations paragraph on page 9, the Results sentence on page 6 ("370 (38.7%) had discontinued ... because of insufficient efficacy or unacceptable side effects") and Table 1 on page 4 (failure 127, 127, 116; lack of efficacy 90, 89, 83; footnote says the categories are not mutually exclusive). 'Not had a beneficial effect' may be loose shorthand for any treatment failure including intolerance, and the next sentence uses the full 'lack of efficacy or unacceptable side effects' wording, so this is a minor wording imprecision rather than a wrong number.
Second check: another model agreed
Report a problem with this finding
Typo In Figure 1 the shaded 'Primary End Point' panel labelled "Change from baseline to months 4–6" begins at the month-3 tick, so the month-3 data points sit on its edge although the primary end point covers months 4, 5 and 6 only.
- p. 6 Figure 1 plot area
Change from baseline to months 4–6
Judge's note: Viewed Figure 1 on the page 6 image: the grey shading starts at the month-3 tick and runs to month 6. This can be read legitimately as the interval after the month-3 visit (months 4 to 6 as elapsed periods), so it is a presentation point rather than an error; the caption states months 4, 5 and 6 correctly.
Second check: none (judge only)