Efficacy and Safety of Inhaled Corticosteroids in Patients With COPD: A Systematic Review and Meta-Analysis of Health Outcomes

I have some comments regarding the recently published article on COPD (Gartlehner, G., R. A. Hansen, S. S. Carson, and K. N. Lohr. 2006. Efficacy and safety of inhaled corticosteroids in patients with COPD: A systematic review and meta-analysis of health outcomes. Ann Fam Med 4:253- 262).[1] This is a competently performed and reported efficacy meta- analysis, with a welcome enhancement that includes safety data from observational studies, as discussed by Dr. Yawn. However, there are many reasons why clinicians should be cautious, since the clinical relevance of the results and the conclusions are limited by the poor quality of the underlying studies. In particular, I would like to draw readers’ attention to the bottom of the last page of the discussion, where I paraphrase four statements:

• No effectiveness studies exist • Generalizabilty is limited • Non-smokers were excluded • Patients with reactive airways were excluded

But exactly how limited is the clinical relevance? I offer some information to help clinicians understand the causes and magnitude of the limitations.

Historically, COPD studies have been systematically designed to enroll non-representative patient populations. This is because North American lung specialist researchers have chosen to adopt a “splitting” approach to asthma and COPD.[2] Only smoking-associated COPD patients are enrolled in COPD studies, and only “pure” asthma patients (who do not smoke and who do not have concomitant COPD) are enrolled in asthma studies. These choices have created an orphan population (estimated to be as much as 50% of lung disease patients) that is not studied at all. Add to this the other stringent exclusion criteria of most asthma and COPD studies, and you will find that, at best, one in ten or twenty (in one example of which I am aware, 1 in 240) of the remaining asthma/COPD patients are actually enrolled.

Specialist lung researchers also promote the view that COPD is “…the only chronic disease for which the finger of blame can be pointed at a single risk factor – tobacco smoking.”[3] Population-based epidemiological and practice-based clinical evidence provide a differing perspective: Studies dating back to the 1970s have identified two distinct types of COPD: (1) smoking-associated COPD and (2) asthma with chronic airways obstruction (AS-CAO).[4] It has been estimated that about 15-20% of patients with COPD are non-smokers, and 30% have AS-CAO.[5] Specialist lung researchers also promote the (“splitting”) view that asthma and COPD are different diseases, yet evidence derived from a 20 year prospective, population-based cohort of over 3000 adults finds that asthma (RR=12.5), not smoking (RR=2.9), is the strongest risk factor for COPD, and that 1 in 5 patients with active asthma develops COPD over a 20-year time period.[6]

A strong argument can be made that the current “splitting” paradigm of asthma and COPD as distinct diseases is a result of historical factors in the USA: (1) the majority of lung research is conducted by lung disease specialists on non-representative populations (referral bias, and self- imposed (“dogmatic”) bias as described above), (2) epidemiological evidence is cited selectively or ignored altogether in favor of “expert opinion,” and (3) primary care lung research is underrepresented in the literature. The practical effect of these factors is to seriously limit the generalizability of the research results, and to raise barriers to accepting novel hypotheses that contradict prevailing dogma.[7]

References

1. Gartlehner G, Hansen RA, Carson SS, Lohr KN (2006) Efficacy and safety of inhaled corticosteroids in patients with COPD: A systematic review and meta-analysis of health outcomes. Ann Fam Med 4: 253-262.

2. Vermeire PA, Pride NB (1991) A "splitting" look at chronic nonspecific lung disease (CNSLD): common features but diverse pathogenesis. Eur Respir J 4: 490-496.

3. Hurd SS, Lenfant C (2005) COPD: good lung health is the key. Lancet 366: 1832-1834. 4. Burrows B (1991) Epidemiologic evidence for different types of chronic airflow obstruction. Am J Resp Dis 143: 1452-1455.

5. Sherrill D, Guerra S, Bobadilla A, Barbee R (2003) The role of concomitant respiratory diseases on the rate of decline in FEV1 among adult asthmatics. Eur Respir J 21: 95-100.

6. Silva GE, Sherrill DL, Guerra S, Barbee RA (2004) Asthma as a risk factor for COPD in a longitudinal study. Chest 126: 59-65.

7. Hahn DL (2006) A theory explaining time trends in asthma prevalence. Eur Respir J 27: 434-435.

Competing interests:   None declared

TO THE EDITORS:

We want to thank Professor Partridge and Dr. Yawn for their interest in our article and their insightful comments. As to Professor Partridge's comment about the meta-analysis, the reason that our meta- analysis did not find a statistically significant difference in mortality is primarily because of a lack of power. The point estimate favors inhaled corticosteroid treatment over placebo (RR 0.81; 95% 0.60-1.08), almost reaching statistical significance, as we briefly touched on in the Discussion section. Death was a rare event in most studies; moreover, even reporting on mortality as an outcome was limited in many studies. Unfortunately, we did not have access to individual patient data. We contacted the study authors for additional unpublished data on mortality, but the response was limited at best and did not produce data that we could incorporate into our own analyses.

In October 2005, Sin et al.. published a pooled analysis of individual patient data that presented a statistically significant reduction of all-cause mortality over 2 to-3 years in patients with chronic obstructive pulmonary disease.(1) We believe that these results are consistent with our findings. Nevertheless, the need is great for future studies that can clarify whether this benefit in survival goes beyond just 2 or 3 years.

We appreciate Dr. Yawn's endorsement of our methods. As it happens, the RTI-UNC Evidence-based Practice Center supported the US Preventive Services Task Force for more than 5 years (1997-2002), and we generally adopted the principle then, which continues to guide our present work, of examining an appropriate variety of observational studies for issues relating to harms, to subgroup analyses, and the like. Like Dr. Yawn, we encourage all those doing systematic reviews to take this wider perspective whenever possible.

Best regards,

Gerald Gartlehner et al.

1. Sin DD, Wu L, Anderson JA, Anthonisen NR, Buist AS, Burge PS, et al. Inhaled corticosteroids and mortality in chronic obstructive pulmonary disease. Thorax 2005;60(12):992-7.

Competing interests:   None declared

Gartlehner G, Hansen R, Carson S and Lohr K (1) present a well done and nicely presented review of inhaled cortico-steroids in COPD management. They reach the same conclusions that have previously been presented by several others including the evidence-based review by the Global Iniative for Obstructive Lung Disease used for the GOLD COPD guideines (www.copdgold.org).

But the authors point out an important difference in their methodology, recognition that an evidence based review must select the appropriate type of evidence for each element of the review. I believe this point deserves greater emphasis.

Other groups doing systematic reviews as the basis for recommendations or guidelines have reached similar conclusions and often go beyond review of RCTs. Specifically as pointed out by the authors, evidence on adverse effects of medications or other interventions is seldom adequately addressed in RCTs. For most efficacious intervetnions, the adverse effects are assumed to be too uncommon to be fully explored in a trial of only 10 to 16 weeks and in a population of only a few hundred people. Therefore, a systematic review often needs to include observational studies and sometimes even case reports to attempt to develop some bounds or assement of adverse effects of intervetnions. For this review of inhaled cortico-steroids (ICS)in COPD this required reviewing the potential side effects of ICS in middle aged and older adults. In the update of the NAEPP or NHLBI asthma guidelines to be completed winter 2006, we have used similar methods by adding oervations studies and case series on side effects of ICS to systmatically assess the adverse effects in a younger age group including children. In addition, the asthma guidelines require review of the data on the adverse effects of long acting beta agonists.

In developing the recommendations of the USPSTF we search for any articles that address adverse effects of screening procedures (almost never RCTs) as well as observational studies. For example, the recent USPSTF recommednation on screening for obesity required review of the observational studies' literature for adverse effects of both surgical procedures and medications used for treatment of obesity to develop a risk/benefit balance. www.ahrq.gov/USPSTF

I salute the authors for their review and for highlighting the need to go beyond the rigid interpretation of evidence based review that involves only review of randomized clinical trials to include the appropriate types of work for the question to be addressed. I hope authors of future systematic reviews will use this methodology rather than limiting themselves to RCTs only.

1. Gartlehner G, Hansen R, Carson S and Lohr K. Efficacy adn Safety of inhaled corptosteroids in patients with COPD: A systematic reveiw and meta-analyssi of health outcomes. Ann Fam Med 2006;4:253-262.

Competing interests:   None declared