Study (trial registry; funding) | Characteristics | Medication Protocol | Effect | Focus* | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Patients (n) | Length of Follow-up | Selection Criteria (Part) | Population | Interventions | Rescue | Prohibited medication | Allowed bias medication | Primary outcome | Results | Significance | ||
Calverley, 20076 (registered; funding from GSK) | 6112 | 3 years | 40–80 years old, COPD diagnosis FEV1: <60% FER: <0.70 before BD Reversibility: <10% No respiratory disease, use of oxygen | 65 years 75% male 43% smoker FEV1 44% predicated value | Salmeterol/Fluticason Salmeterol Fluticason Placebo | Albuterol | Long-acting BD, steroids | Short-acting and other BD | Mortality | 12.6% vs 13.5% vs 16.0% vs 15.2% | NS | A |
Calverley et al,4 2003 (not registered; funding from GSK) | 1465 | 1 year | FEV1: 25% to 70% before BD FER: <0.70 before BD Reversibility: <10% ≥1 exac/year 3 years No respiratory disease, use of oxygen | 63.5 years 72.5% male 51% smoker FEV1 49% predicted value | Salmeterol/ Fluticason Salmeterol Fluticason Placebo | Albuterol | Long-acting β-agonist, steroids | Anticholinergics and theophillin | FEV1 before BD | 10% vs 2% vs 2% vs −3% | P < .01 | B |
Tashkin et al,7 2008 (registered; funded by BI and Pfizer) | 5993 | 4 years | >40 years FEV1: <70%, FER: <0.70 No respiratory disease, use of oxygen Myocardial infarction during last 6 months, unstable arrhythmia | 64.5 years 75% male 30% smoker FEV1 48% predicted value | Spiriva Placebo | — | Short-acting anticholinergics | All nonanticholinergics | FEV1 decline before and after BD | Before BD: 30 vs 30 mL/yr After BD: 40 vs 42 mL/yr | NS | C |
Niewoehner et al,5 2005 (not registered; funded by BI and Pfizer) | 1829 | 6 months | >40 years COPD diagnosis FEV1: <60% FER: <0.70 No asthma Myocardial infarction during past 6 months, cardiac hospital during past year | 67.8 years 99% male 30% smoker FEV1 36% predicted value 29% oxygen | Spiriva Placebo | — | Short-acting anticholinergics | All nonanticholinergics | %Exacerbation | 32.3% vs 27.9% | P = .037 | D |
%Exacerbation hospitalization | 9.5% vs 7.0% | NS | ||||||||||
Vogelmeier et al,8 2011 (registered; funded by BI and Pfizer) | 7376 | 1 year | >40 years FEV1: <70%, FER: <0.70 ≥1 exacerbation during past year No asthma CVD | 62.9 years 74.7% male 48% smoker FEV1 49% predicted value | Spiriva Salmeterol | Albuterol | Anticholinergics, long-acting β-agonist | Short-acting β-agonist | Time to first exacerbation | 187 vs 145 days (first fourth of patients) | P < .001 | E |
↵* A: Acknowledge statistically nonsignificant results for primary outcome, but the focus is on beneficial effect and secondary outcome in main text discussion and conclusion. They claim the study is underpowered. B: Focus is on secondary outcomes in main text discussion. C: Focus is on secondary outcome in result and discussion section of both abstract and main text. Of many nonsignificant post hoc subgroup analyses, they only state the significant one. D: Acknowledge statistically nonsignificant results for the primary outcome (called “borderline significant”), but focus is on beneficial effect in the abstract and main text results. E: Focus is on an exaggerated effect on one fourth of all patients (a third had an exacerbation), which is not stated in the abstract. Focus is on inaccurate description of population in main text discussion and conclusion.
BD, bronchodilator; BI, Boehringer Ingelheim; COPD, chronic obstructive pulmonary disease; CVD, cardiovascular disease; FER, forced expiratory ratio; FEV1, forced expiratory volume in first second; GSK, GlaxoSmithKline; NS, not significant.