reducing mortality in copd-the torch study

8/19/2019 Reducing Mortality in Copd-The Torch Study

1/27

REDUCING MORTALITY INCOPD – THE TORCH STUDY

Professor PMA CalverleyUniversity Hospital Aintree

LiverpoolUK


2/27

COPD STILL KILLS

• In 2002 the estimated annual mortalitydue to COPD in Europe was 260,000 -2.7% of all deaths

• In the UK 30,000 people die from COPDeach year • Deaths due to COPD are under-

diagnosed• COPD contributes to death from other

causes


3/27

FACTORS ASSOCIATED WITHDEATH

• Age

• Baseline FEV1• Smoking status

• BMI• Exercise capacity• Exacerbation frequency• Health status


4/27

FACTORS YOU CAN MODIFY• Age

• Baseline FEV1• Smoking status

• BMI• Exercise capacity• Exacerbation frequency

• Health status


5/27

TREATMENT THAT REDUCES

MORTALITY

• Smoking cessation – the earlier, thebetter

• LTOT if hypxaemic• LVRS in selected cases

• DRUGS – good or bad?


6/27

Adjusted Survival Function ofCOPD Patients by Therapy

Probability of survival

Soriano et al. Eur Respir J 2002

1.0

0

0.9

0.8

0.6

0.7

0 24 36

Months

Comparison

SALM

6 12 18 30

SALM/FP

FP


7/27

SFC 50/500 µg bd (N=1533)

TORCH: study design

SAL 50 µg bd (N=1521)

Placebo (N= 1524)3-year study duration

FP 500 µg bd (N=1534)

2 weekrun-in

Vestbo et al . Eur Respir J 2004; Calverley et al . NEJM 2007


8/27

Worldwide participation in 42

countries


9/27

TORCH: main objectives• Primary objective

– The effect of SFC 50/500 g vs placebo onall-cause mortality over 3 years in patients withmoderate-to-severe COPD

• Secondary objectives – The effect of SFC 50/500 g on the rate of moderate

and severe exacerbations

– The effect of SFC 50/500 g on health status(SGRQ)

Vestbo et al . Eur Respir J 2004Calverley et al . NEJM 2007SGRQ = St. George’s Respiratory Questionnaire


10/27

Clinical Endpoint Committee(CEC)

• Independently review all fatal events (postrandomisation) occurring during the course ofthe study and assign a primary cause of death

• Request all available documentation(e.g. clinical records, death certificates,and site investigator narratives)

Mc Garvey et al Thorax 2007


11/27

PopulationsPopulation Placebo SAL FP SFC Total

Safety 1544 1542 1552 1546 6184

Intent-to-treat

(ITT) efficacy

1524 1521 1534 1533 6112

Calverley e t al. NEJM 2007


12/27

Demographics

Age 65 (8)

Males 76%Current smokers 43%Pack years 49 (27)% pred baseline FEV 1 (post bronc) 44 (13)% pred reversibility 3.7 (3.7)≥ 1 exacerbations in previous year 57%

ITT

N=6112Mean (sd )

Calverley et al. NEJM 2007


13/27

Deaths

Total927

Randomised911

Intent-to-Treat904

Excluded Sites7

Within 3 years875

Post 3 years29

On treatment464

Post treatment23

On treatment6

Post treatment411

Non Randomised16

On Treatment = up to 14 days after treatment stop


14/27

Overall causes of death as

adjudicated by the Clinical EndpointCommitteeUnknown

7%

Cardiac27%

Cancer 21%

Other 10%

Respiratory35%


P d d


15/27

Premature study drugdiscontinuation

SALM FP048

12162024283236404448

0 12 24 36 48 60 72 84 96 108 120 132 144 156

Probability of withdrawal (%)

Placebo

Time to withdrawal from study medication (weeks)

152415211534

1533

Number at risk

114112401247

1296

100510931112

1164

884986971

1042

SFC

Statistical comparisons: SALM/FP, SAL & FP vs placebo p < 0.001; SALM/FP vs SAL p = 0.048; SALM/FP vs FP p = 0.0Vertical bars are standard errors Calverley et al. NEJM 2007

P i l i ll li


16/27

Primary analysis: all-cause mortalityat 3 years

Vertical bars are standard errors

1524

1533

1464

1487

1399

1426

1293

1339

Number

alive

024

68

1012141618

0 12 24 36 48 60 72 84 96 108 120 132 144 156Time to death (weeks)

Probability of death (%)

SFC 12.6%Placebo 15.2%

HR 0.825, p=0.05217.5% risk reduction

2.6%absolute reduction



17/27

Supportive analysis: All-cause mortalityat 3 years - Cox’s proportional hazards

†Cox's proportional hazards model estimate at mean age, FEV 1, body mass index and proportionalcoefficients for smoking status, gender and region

Placebo

(N = 1524)

SFC

(N = 1533)

Number of deaths 231 193

Percentage of deathsby3 years

15.2 12.6

HR (95% CI) † 0.811 (0.670, 0.982)

p-value † 0.031



18/27

SALM FP

1816

141210

86420

Time to death (weeks)


All-cause mortality at 3 years


15241533

15211534

14641487

14811487

13991426

14171409

12931339

13161288

Placebo SFC

0 12 24 36 48 60 72 84 96 108 120 132 144 156

Number alive



19/27

COPD-related mortality by 3 years


0

1

23

4

5

6

7

8

0 12 24 36 48 60 72 84 96 108 120 132 144 156Time to death (weeks)


1524153315211534

1499151315021515

1476149014751485

1433146014281428

Number alive

Placebo SALM FP SFC


Cause of death


20/27

Cause of deathon treatment (adjudicated by CEC)

0.0

1.02.0

3.0

4.0

5.0

6.0

7.0

Cardio-vascular

Pulmonary Cancer Other Unknown

Deaths (%)

Placebo SFC



21/27

Supportive evidence for a

significant effect


22/27

Rate of moderate and severeexacerbations over three years

*p < 0.001 vs placebo; †p = 0.002 vs SALM; ‡p = 0.024 vs FP

Mean number of exacerbations/year

1.13

0.93 *0.85 *†‡

0.97 *

25% reduction

00.2

0.4

0.6

0.8

11.2

Placebo SALM FP SFC

TreatmentCalverley et al. NEJM 2007

THE OTHER PREDICTORS OF


23/27

THE OTHER PREDICTORS OFMORTALITY

Calverley et al NEJM 2007


24/27

PUTTING THE MORTALITYEFFECT IN CONTEXT


25/27

RISK VS BENEFIT• Health status data are encouraging

• No clear evidence for interaction with highbackground level of co-morbidity• Fewer hospitalisations with salmeterol and

SFC• Pneumonia a new signal – not reflected in

current markers of morbidity/mortality.More research is needed to understandthis


26/27

Effect of statins on all-cause mortality inpatients with coronary heart disease

• Meta-analysis of 17 trials• N = 40974

• Mean/median follow up 0.3–6.1 years

• Relative Risk Reduction = 16%

• Absolute Risk Reduction = 1.8%

Wilt et al . Arch Intern Med 2004

CONCLUSIONS


27/27

CONCLUSIONS

• Many COPD disease patents die from andwith the disease

• Establishing the cause of death is difficult butnot impossible.

• Pharmacotherapy reduces the risk of dying inmoderate/severe COPD

• Statistical and methodological issues have abig effect on clinical trial design• It is hard to test a readily available

symptomatic therapy – contrast the originaloxygen trials and TORCH• Treatment can reduce COPD mortality….so

we should tell people this!

reducing mortality in copd-the torch study

Documents