Rivaroxaban versus Warfarin in Nonvalvular Atrial Fibrillation
Indian Heart Journal, 2012 · paper · Clinical trials
A double-blind trial of rivaroxaban versus warfarin in 14,264 patients with nonvalvular atrial fibrillation. Rivaroxaban was noninferior for preventing stroke or systemic embolism, with similar overall bleeding but less intracranial and fatal bleeding.
2 findings
Substance Intention-to-treat primary end point: abstract and Results text give "hazard ratio, 0.88; 95% CI, 0.74 to 1.03", Table 2 gives "0.88 (0.75–1.03)" for the same row (269 vs 306 events).
- p. 1 Abstract, Results
hazard ratio, 0.88; 95% CI, 0.74 to 1.03
- p. 5 Primary Outcome text, left column
ard ratio, 0.88; 95% CI, 0.74 to 1.03; P<0.001 for
- p. 5 Table 2, Intention-to-treat population row
0.88 (0.75–1.03)
- p. 1 Abstract, Results
0.88; 95% CI, 0.74 to 1.03
Only the authors can say what is right here.
Judge's note: Checked on the page images of p1 (abstract), p5 text and p5 Table 2. Same population (7081/7090), same events (269 vs 306), same point estimate and upper bound, same P values (<0.001, 0.12); only the lower bound differs. The during-treatment and after-discontinuation rows have their own intervals, so no different analysis explains it. Table 2 footnotes give no explanation. A rough check from the event counts puts the lower bound near 0.747, so the article alone cannot say which is right; the authors must.
Second check: another model agreed · checked by hand
Report a problem with this finding
Substance For the safety, as-treated population (189 vs 243 events, HR 0.79, 95% CI 0.65–0.95), the text on p. 5 gives "P = 0.01 for superiority" but Table 2 prints 0.02 in the Superiority column of the same row.
- p. 5 Primary Outcome text, top of left column
0.65 to 0.95; P = 0.01 for superiority
- p. 5 Table 2, Safety, as-treated population row, Superiority column
0.79 (0.65–0.95)
- p. 5 Table 2, row 'Safety, as-treated population', Superiority column
0.02
- p. 5 Primary Outcome text
P = 0.01 for superiority
Only the authors can say what is right here.
Judge's note: I checked the page 5 crops. The text sentence starts on p. 4 with "In the as-treated safety population, primary events occurred in 189 patients..." and continues on p. 5 with "(hazard ratio, 0.79; 95% CI, 0.65 to 0.95; P = 0.01 for superiority)". In Table 2, the row 'Safety, as-treated population' (7061/189/1.7 vs 7082/243/2.2, 0.79 (0.65–0.95)) shows 0.02 under Superiority; the image confirms the alignment. Population, events, HR and CI are the same, and Methods describe one superiority test in this population, so one P value is printed two ways. Nothing in the text explains the difference. Both values fall below 0.05, so the conclusion of significance is the same either way. Only the authors can say which value is correct. This P value does not appear in the abstract.
Second check: another model agreed
Report a problem with this finding
8 minor (typos and the like; most checked by the judge only)
Wording CHADS2 described as "on a scale ranging from 1 to 6" (Methods) and "The CHADS2 score for the risk of stroke ranges from 1 to 6" (Table 1 footnote); a patient with none of the listed risk factors scores 0, so the scale runs 0 to 6.
- p. 2 Study Participants
on a scale ranging from 1 to 6
- p. 4 Table 1, footnote †
ranges from 1 to 6
- p. 2 Methods, Study Participants
of 2 or more, on a scale ranging from 1 to 6, with
- p. 4 Table 1, footnote †
The CHADS2 score for the risk of stroke ranges from 1 to 6
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the standard CHADS2 definition (1 point each for heart failure, hypertension, age 75 or more, diabetes; 2 for prior stroke/TIA), which gives a minimum of 0. Both sentences are printed as quoted (p2, p4 images). The paper lists the components but not the points. The wording describes the scale, not the observed range (the footnote separately reports three patients with a score of 1). No effect on any result.
Second check: another model agreed
Report a problem with this finding
Wording Methods (p. 3) call sub-threshold overt bleeding "minor episodes", but the Table 3 footnote § (p. 8) reads "Minimal bleeding events were not included in the principal safety end point."; the article never defines "minimal".
- p. 3 Outcomes, end of second paragraph
were classified as minor episodes
- p. 8 Table 3, footnote §
Minimal bleeding events were not included in the principal safety end point.
Judge's note: I checked both crops. Methods, p. 3: "Other overt bleeding episodes that did not meet the criteria for major or clinically relevant nonmajor bleeding were classified as minor episodes." Table 3 footnote §, p. 8: "Minimal bleeding events were not included...". A search of the text finds "minimal" only in that footnote. The two terms seem to name the same category, so the meaning is probably clear. Still, one category carries two labels and the footnote's term is undefined in the article. Either word could be the right one, so this is a minor consistency point.
Second check: another model agreed
Report a problem with this finding
Wording Text: "62.5% had heart failure, and 40.0% had diabetes"; Table 1 gives diabetes 2878 (40.4) and 2817 (39.5), which pool to 39.9% of 14,264.
- p. 4 Patient Characteristics and Treatments, text
and 40.0%
- p. 4 Table 1, Diabetes mellitus row
2878 (40.4)
- p. 4 Table 1, Diabetes mellitus row
2817 (39.5)
- p. 4 Results, Patient Characteristics and Treatments
62.5% had heart failure, and 40.0%
Only the authors can say what is right here.
Judge's note: Table 1 values read from the p4 image. Pooled diabetes is 5695/14264 = 39.93%, which rounds to 39.9, not 40.0. The other pooled figures in the same sentence reproduce from Table 1 with the same denominator (hypertension 90.5, heart failure 62.5, prior stroke/SE/TIA 54.8, women 39.7, VKA 62.4). No footnote or text gives a different denominator for diabetes. A 0.1 percentage-point difference in a descriptive statistic; possibly the mean of the two arm percentages (39.95). Authors to confirm.
Second check: another model agreed
Report a problem with this finding
Typo "a projected event rate of 2.3% per 100 patient-years in the warfarin group": the percent sign and 'per 100 patient-years' duplicate each other; should be '2.3 per 100 patient-years' or '2.3% per year'.
- p. 3 Statistical Analysis, right column
projected event rate of 2.3% per 100
Judge's note: Printed as quoted on the p3 image. The same page says "Event rates per 100 patient-years are presented as proportions of patients per year", and elsewhere the paper uses either '% per year' (text) or 'no./100 patient-yr' (tables), never both together. Intended meaning is clear from Table 2 (warfarin 2.2 to 2.4); cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo Disclosures spell the same company two ways: "Amgen, Amylin, AstraZeneca" (Dr. Mahaffey) and "grant support from Novartis, Merck, and Amilyn/Lilly" (Dr. Califf).
- p. 8 Disclosures, Dr. Mahaffey
Amgen, Amylin, AstraZeneca
- p. 8 Disclosures, Dr. Califf
Amilyn/
- p. 8
grant support from Novartis, Merck, and Amilyn/Lilly
- p. 8 Disclosures, Dr. Califf
Amilyn/ Lilly
Judge's note: Both spellings are visible on the p8 image, in the same disclosure paragraph. The inconsistency is internal to the page. That 'Amylin' is the correct form (Amylin Pharmaceuticals, partnered with Lilly) is outside knowledge, but the fix is evident.
Second check: another model agreed
Report a problem with this finding
Typo Disclosures name "Regado Biotechnologies" (Dr. Mahaffey) and "Regado Biosciences" (Dr. Becker) in the same paragraph.
- p. 8 Disclosures, Dr. Mahaffey
Regado Biotechnologies
- p. 8 Disclosures, Dr. Becker
Regado Biosciences
- p. 8
Pozen, Regado Biotechnologies, Sanofi Aventis,
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: that the company's name is Regado Biosciences and that both mentions refer to it. Both forms are printed as quoted on the p8 image. The inconsistency is on the page; which form is right, or whether each was copied from the author's own disclosure form, cannot be settled from the article.
Second check: another model agreed
Report a problem with this finding
Typo Disclosures (Dr. Mahaffey): "Novartis, Pfizer, Proctor & Gamble, Sanofi Aventis"; the company's name is spelled Procter & Gamble.
- p. 8 Disclosures, Dr. Mahaffey
Proctor & Gamble
- p. 8 Disclosures, Dr. Mahaffey
Ikarja
- p. 8
Novartis, Pfizer, Proctor & Gamble, Sanofi Aventis,
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the correct spelling of the company name, Procter & Gamble. The p8 image shows 'Proctor & Gamble'. Nothing in the article gives the other spelling. The 'Ikarja' half of this cluster is handled under C012.
Second check: another model agreed
Report a problem with this finding
Typo Disclosures (Dr. Mahaffey): "GlaxoSmithKline, Guidant, Ikarja, Johnson & Johnson"; 'Ikarja' is probably a misspelling of Ikaria.
- p. 8
GlaxoSmithKline, Guidant, Ikarja, Johnson & Johnson,
- p. 8 Disclosures, Dr. Mahaffey
Ikarja
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: that the intended company is Ikaria and that no company named Ikarja exists. The p8 image shows 'Ikarja' (with a j). The name occurs once in the article, so there is no internal comparison. Authors to confirm.
Second check: another model agreed