The New England Journal of Medicine FEBRUARY 22, 2007

Towards a Revolution in COPD Health (TORCH): A Randomized Double-Blind, Placebo-Controlled Study of Salmeterol and Fluticasone Propionate in Chronic Obstructive Pulmonary Disease

Calverley PMA, Anderson JA, Celli B, Ferguson GT, Jenkins C, Jones PW, Yates JC, Vestbo J

Bottom Line

The TORCH trial evaluated whether the combination of inhaled salmeterol and fluticasone propionate would reduce all-cause mortality in patients with moderate-to-severe COPD compared with placebo, salmeterol alone, or fluticasone alone.

Key Findings

1. The combination therapy (salmeterol/fluticasone) did not achieve a statistically significant reduction in all-cause mortality compared with placebo (Hazard Ratio 0.825; 95% CI 0.681–1.002; P=0.052).
2. Combination therapy significantly reduced the annual rate of moderate-to-severe COPD exacerbations by 25% (HR 0.75; 95% CI 0.69–0.81; P<0.001) compared with placebo.
3. Patients receiving the combination regimen experienced a significantly slower rate of lung function (FEV1) decline compared with placebo (a difference of 16 mL/year; 95% CI 7–25; P<0.001).
4. Health-related quality of life, measured by the St. George's Respiratory Questionnaire, showed consistent improvement with combination therapy compared to placebo.

Study Design

Design
RCT
Double-Blind
Sample
6,112
Patients
Duration
3 yr
Median
Setting
Multicenter, Global
Population Patients with moderate-to-severe chronic obstructive pulmonary disease (FEV1 <60% predicted)
Intervention Salmeterol (50 µg) and fluticasone propionate (500 µg) twice daily
Comparator Placebo, salmeterol alone, or fluticasone propionate alone
Outcome All-cause mortality over 3 years

Study Limitations

The trial narrowly missed its primary endpoint of all-cause mortality statistical significance (P=0.052), which has been attributed by some as potentially affected by the statistical design and interim analysis adjustments.
High rates of participant withdrawal from the study, particularly in the placebo arm, may have introduced bias and complicated the interpretation of long-term mortality data.
The population included patients with pre-bronchodilator FEV1 <60% predicted, which may not have captured the subset of patients at the highest risk of mortality.
The study design used older treatment standards as a comparator, which may not reflect contemporary clinical practices.

Clinical Significance

While the study did not definitively prove that combination therapy reduces all-cause mortality in COPD, it provided robust evidence for significant improvements in key clinical outcomes, including reduced exacerbation rates and slower decline in lung function, influencing long-term COPD management strategies.

Historical Context

The TORCH trial was one of the largest and most ambitious clinical trials in the history of COPD research, representing a landmark effort to determine if pharmacological intervention could impact survival, a goal that had previously only been achieved with long-term oxygen therapy.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

How do the pharmacological mechanisms of salmeterol and fluticasone propionate complement each other in the management of COPD, and why was all-cause mortality chosen as a primary endpoint instead of just symptomatic relief?

Key Response

Salmeterol is a long-acting beta2-agonist (LABA) that induces bronchodilation by increasing intracellular cAMP, while fluticasone is a corticosteroid that reduces airway inflammation. In COPD, the combination aims to address both the fixed and reversible components of airflow limitation. All-cause mortality was selected for TORCH to determine if these drugs could alter the natural history of the disease, moving beyond simple symptom control to disease-modifying outcomes.

Resident
Resident

In the TORCH trial, the p-value for the primary endpoint of all-cause mortality for combination therapy vs. placebo was 0.052. How should this influence your clinical decision to prescribe ICS/LABA in moderate-to-severe COPD patients?

Key Response

While the primary endpoint did not reach the predefined statistical significance (p<0.05), the hazard ratio (0.825) suggested a 17.5% reduction in mortality. Furthermore, significant improvements were seen in secondary endpoints such as exacerbation rates and health status (St. George's Respiratory Questionnaire). Clinically, this supports the use of combination therapy for patients with frequent exacerbations or persistent symptoms despite monotherapy, even if a definitive survival benefit was not statistically proven.

Fellow
Fellow

The TORCH trial identified an increased risk of pneumonia in the fluticasone-containing arms. In a patient with frequent COPD exacerbations but a history of recurrent respiratory infections, how do you balance the evidence for ICS/LABA against the risk of adverse events?

Key Response

The trial showed a significant increase in the reporting of pneumonia as an adverse event in groups receiving fluticasone (19.6% vs 12.3% in placebo). For fellows, this necessitates a phenotype-driven approach: patients with high blood eosinophil counts or frequent exacerbations benefit most from ICS, but those with low eosinophils or high infectious risk might be better managed with LAMA/LABA combinations to avoid steroid-related pneumonia risk.

Attending
Attending

Reflecting on the TORCH results, how do you teach the concept of 'clinical significance' versus 'statistical significance' to trainees when discussing a trial that failed its primary endpoint but transformed standard of care?

Key Response

TORCH is a landmark example of a 'negative' trial that had a 'positive' impact. Despite missing p=0.05 by a narrow margin, the consistency across secondary endpoints—reduced exacerbations and improved lung function—provided enough evidence for regulatory bodies and guideline committees to incorporate ICS/LABA into the standard COPD treatment algorithm, emphasizing that p-values are thresholds for certainty, not binary indicators of clinical utility.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The TORCH trial experienced a high dropout rate, particularly in the placebo arm (44%). Critically evaluate how these differential withdrawal rates might have introduced bias into the Intention-To-Treat (ITT) analysis of mortality.

Key Response

High dropout in the placebo group, often due to symptomatic worsening, can lead to 'healthy survivor' bias. If the sickest patients in the placebo group withdrew and their subsequent mortality was not fully captured, the trial might have underestimated the true treatment effect of the salmeterol/fluticasone combination. This highlights the difficulty of long-term mortality trials in symptomatic diseases where placebo maintenance is ethically and practically challenging.

Journal Editor
Journal Editor

Given the hierarchical testing strategy employed in the TORCH trial, what are the editorial implications of interpreting secondary endpoints (like exacerbation rate) when the primary endpoint (all-cause mortality) did not reach statistical significance?

Key Response

Strictly speaking, in a hierarchical statistical model, if the primary endpoint fails, subsequent p-values for secondary endpoints should be considered exploratory rather than confirmatory to avoid Type I error inflation due to multiplicity. As an editor, one must ensure the authors do not 'over-sell' secondary successes as definitive findings, though the consistency of the secondary data in TORCH provided a compelling narrative that swayed the field.

Guideline Committee
Guideline Committee

How did the TORCH trial data influence the evolution of the GOLD (Global Initiative for Chronic Obstructive Lung Disease) recommendations regarding the use of ICS/LABA compared to LAMA monotherapy?

Key Response

TORCH provided the foundational evidence for using ICS/LABA to reduce exacerbations in patients with moderate-to-severe COPD. However, because it failed its primary mortality endpoint and showed increased pneumonia risk, current GOLD guidelines have become more granular, favoring LAMA/LABA for most patients and reserving ICS/LABA (or triple therapy) for those with high exacerbation risk and elevated blood eosinophils (≥300 cells/μL), reflecting a move toward precision medicine based on the safety/efficacy trade-offs first highlighted by TORCH.

Clinical Landscape

Noteworthy Related Trials

2000

ISOLDE Trial

n = 751 · NEJM

Tested

Fluticasone propionate 500 mcg twice daily

Population

Patients with stable chronic obstructive pulmonary disease

Comparator

Placebo

Endpoint

Rate of decline in FEV1

Key result: Inhaled fluticasone did not significantly affect the rate of decline in FEV1, but it did reduce the frequency of exacerbations and improved quality of life scores.
2008

UPLIFT Trial

n = 5,993 · NEJM

Tested

Tiotropium 18 mcg daily

Population

Patients with COPD (GOLD stage II or greater)

Comparator

Placebo plus standard care

Endpoint

Rate of decline in FEV1 and all-cause mortality

Key result: Tiotropium did not significantly slow the rate of FEV1 decline over 4 years, though it improved lung function and reduced exacerbation rates.
2018

IMPACT Trial

n = 10,355 · NEJM

Tested

Fluticasone furoate/umeclidinium/vilanterol (triple therapy)

Population

Patients with symptomatic COPD and a history of exacerbations

Comparator

Fluticasone furoate/vilanterol or umeclidinium/vilanterol

Endpoint

Annual rate of moderate or severe COPD exacerbations

Key result: Triple therapy resulted in a lower rate of moderate or severe COPD exacerbations compared to dual therapies.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis