Salmeterol and Fluticasone Propionate and Survival in Chronic Obstructive Pulmonary Disease
New England Journal of Medicine, 2007 · paper · Clinical trials
A three-year randomized trial in over 6,000 COPD patients comparing salmeterol plus fluticasone, each alone, or placebo. The combination's reduction in deaths did not reach the preset significance level, but it reduced exacerbations; pneumonia was more common with fluticasone.
1 finding
Wording Discussion: "sustained improvements in health status and FEV1; the values for both were better at the end of the trial than at baseline". In Fig. 2E the combination-therapy FEV1 curve is below zero (about −20 ml) at week 156.
- p. 12 Discussion, para 4
the values for both were better at the end of the trial than at baseline
- p. 7 Figure 2E (FEV1)
Adjusted Mean Change in FEV1 (ml)
- p. 7 Figure 2E (FEV1)
Combination therapy
- p. 7 Figure 2E
Adjusted Mean Change in FEV1
Only the authors can say what is right here.
Judge's note: In Fig. 2E (p7 upright image) the combination-therapy adjusted mean change in FEV1 peaks at about +72 ml at week 24. It is about +5 ml at week 120 and about −20 ml at week 156, below the zero baseline line. In Fig. 2D the combination health-status change at week 156 is about −1.2 units, which is still an improvement. So the claim holds for health status but not for FEV1 at the end of the trial. The 3-year averaged change (+0.029 L, Table 3) is positive, which may be what the authors meant.
Second check: another model agreed
Report a problem with this finding
4 minor (typos and the like; most checked by the judge only)
Wording Methods: "The steering committee, made up of six academic investigators and two employees of the sponsor". The Appendix lists nine Steering Committee members, three of them at GSK sites (Anderson, Greenford; Knobil and Yates, Research Triangle Park).
- p. 3 Statistical Analysis, last paragraph
The steering committee, made up of six aca-demic investigators and two employees of the sponsor
- p. 14 Appendix, Steering Committee
K. Knobil, J.C. Yates, Research Triangle Park, NC
Only the authors can say what is right here.
Judge's note: Appendix on p14: Calverley, Anderson, Celli, Ferguson, Jenkins, Jones, Knobil, Yates, Vestbo, which is 9 names. The p1 affiliations place Anderson (Greenford) and Yates (Research Triangle Park) at GlaxoSmithKline R&D, and the disclosures say both are GSK employees. Knobil is listed with Yates at Research Triangle Park, but no employer is given. That makes 6 academics and 3 sponsor-site members, against the 6+2 in the Methods. A change in membership could explain it, but nothing in the files says so, so only the authors can resolve it.
Second check: another model agreed
Report a problem with this finding
Wording Results: "Annual admission rates were 17% lower in the combination-therapy and salmeterol groups than in the placebo group". Table 2 gives rate ratios of 0.83 (combination) and 0.82 (0.69–0.96) (salmeterol).
- p. 10 Results, Exacerbations paragraph
rates were 17% lower in the combination-therapy and salmeterol groups than in the placebo group
- p. 9 Table 2, Severe (requiring hospitalization)
0.82 (0.69–0.96)
Judge's note: Table 2 on p9 (upright image), severe exacerbations requiring hospitalization: combination vs placebo 0.83 (0.71–0.98), P=0.03; salmeterol vs placebo 0.82 (0.69–0.96), P=0.02. As printed, 0.82 corresponds to an 18% reduction. 17% would only apply if the unrounded ratio sat at the 0.825 boundary, so this is a minor consistency point.
Second check: another model agreed
Report a problem with this finding
Wording Results p10: "the health status (a reduction of 3.1 units ...) and spirometric measurements (an increase in FEV1 of 0.092 liter) in the combination-therapy group". These are the differences vs placebo in Table 3, not changes within the group, which the same paragraph gives as 3.0 units and 0.029 liter.
- p. 10 Results, Exacerbations, Health Status, and Lung Function, para 2
the health status (a reduction of 3.1 units in the score for the St. George’s Respiratory Question-naire) and spirometric measurements (an increase in FEV1 of 0.092 liter) in the combination-therapy group
- p. 10 same paragraph, earlier
with a mean reduction of 3.0 units averaged over 3 years
- p. 11 Table 3
–3.1 (–4.1 to –2.1)
Judge's note: Table 3 (p11 upright image): the combination group's adjusted mean changes are −3.0 units (SGRQ) and +0.029 L (FEV1). The differences vs placebo are −3.1 (−4.1 to −2.1) and 0.092 (0.075 to 0.108). The p10 text first gives the within-group values (3.0 units; 0.029 liter), then attributes 3.1 and 0.092 to 'the combination-therapy group' and says these were better than placebo, salmeterol and fluticasone. Against salmeterol and fluticasone the differences are smaller (−2.2/−1.2 units; 0.050/0.044 L). Adding 'as compared with placebo' would fix the wording.
Second check: another model agreed
Report a problem with this finding
Typo Figure 2 legend has a duplicated word: "the numbers of of patients alive or dead from non–COPD-related causes (Panel C)".
- p. 6 Figure 2 legend
the numbers of of patients alive or dead from non–COPD-related causes
- p. 6 Figure 2 legend, sentence on the tables below the graphs
the numbers of of patients alive or dead from non–COPD-related causes (Panel C)
- p. 6 Figure 2 legend
the numbers of of patients
Judge's note: The p6 crop of the Figure 2 legend shows "numbers of" at the end of one line and "of patients" at the start of the next. The duplication is on the printed page, so it is not an extraction artefact.
Second check: none (judge only)