Todd A. Lee, PharmD, PhD; A. Simon Pickard, PhD; David H. Au, MD, MS; Brian Bartle, MPH; Kevin B. Weiss, MD, MPH, MS
Note: At the time of this work, Dr. Weiss was at the Center for Management of Complex Chronic Care, Hines Veterans Affairs Hospital, and the Institute for Healthcare Studies, Northwestern University Feinberg School of Medicine. He has since moved to the American Board of Medical Specialties.
Disclaimer: The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the U.S. Department of Veterans Affairs.
Grant Support: By the U.S. Department of Veterans Affairs Health Services Research and Development (IIR 03-307).
Potential Financial Conflicts of Interest:Honoraria: T.A. Lee (AstraZeneca, Novartis), D.H. Au (GlaxoSmithKline). Consultancies: K.B. Weiss (Merck & Co.). Stock ownership or options (other than mutual funds): D.H. Au (Pfizer). Grants received: T.A. Lee (Altana, Aventis, AstraZeneca, Boehringer Ingelheim, Chiesi, GlaxoSmithKline, Merck & Co., Novartis, Pfizer, Schering-Plough, Sepracor, University of Kentucky). Other: D.H. Au (Assessing the Impact of Recent Updates for Advair and Serevent Special Issues Board).
Reproducible Research Statement:Study protocol and data set: Not available. Statistical code: Available from Dr. Lee (firstname.lastname@example.org).
Requests for Single Reprints: Todd A. Lee, PharmD, PhD, Hines Veterans Affairs Hospital, 5000 South 5th Avenue (151-H), Hines, IL 60141; e-mail, email@example.com.
Current Author Addresses: Dr. Lee and Mr. Bartle: Hines Veterans Affairs Hospital, 5000 South 5th Avenue (151-H), Hines, IL 60141.
Dr. Pickard: 833 South Wood Street, RM164 MC886, Chicago, IL 60612.
Dr. Au: Veterans Affairs Puget Sound Health Care System, Health Services Research and Development (152), 1660 South Columbian Way, Seattle, WA 98108.
Dr. Weiss: American Board of Medical Specialties, 1007 Church Street, Suite 404, Evanston, IL 60201-5913.
Author Contributions: Conception and design: T.A. Lee, D.H. Au, K.B. Weiss.
Analysis and interpretation of the data: T.A. Lee, A.S. Pickard, D.H. Au, B. Bartle, K.B. Weiss.
Drafting of the article: T.A. Lee, A.S. Pickard, D.H. Au, K.B. Weiss.
Critical revision of the article for important intellectual content: T.A. Lee, A.S. Pickard, D.H. Au, B. Bartle, K.B. Weiss.
Final approval of the article: T.A. Lee, A.S. Pickard, D.H. Au, B. Bartle, K.B. Weiss.
Statistical expertise: T.A. Lee, A.S. Pickard.
Obtaining of funding: T.A. Lee.
Administrative, technical, or logistic support: B. Bartle.
Collection and assembly of data: A.S. Pickard, B. Bartle.
Concerns exist regarding increased risk for mortality associated with some chronic obstructive pulmonary disease (COPD) medications.
To examine the association between various respiratory medications and risk for death in veterans with newly diagnosed COPD.
Nested case–control study in a cohort identified between 1 October 1999 and 30 September 2003 and followed through 30 September 2004 by using National Veterans Affairs inpatient, outpatient, pharmacy, and mortality databases; Centers for Medicare & Medicaid Services databases; and National Death Index Plus data. Cause of death was ascertained for a random sample of 40% of those who died during follow-up. Case patients were categorized on the basis of all-cause, respiratory, or cardiovascular death. Mortality risk associated with medications was assessed by using conditional logistic regression adjusted for comorbid conditions, health care use, and markers of COPD severity.
U.S. Veterans Health Administration health care system.
32 130 case patients and 320 501 control participants in the all-cause mortality analysis. Of 11 897 patients with cause-of-death data, 2405 case patients had respiratory deaths and 3159 case patients had cardiovascular deaths.
All-cause mortality; respiratory and cardiovascular deaths; and exposure to COPD medications, inhaled corticosteroids, ipratropium, long-acting β-agonists, and theophylline in the 6 months preceding death.
Adjusted odds ratios (ORs) for all-cause mortality were 0.80 (95% CI, 0.78 to 0.83) for inhaled corticosteroids, 1.11 (CI, 1.08 to 1.15) for ipratropium, 0.92 (CI, 0.88 to 0.96) for long-acting β-agonists, and 1.05 (CI, 0.99 to 1.10) for theophylline. Ipratropium was associated with increased cardiovascular deaths (OR, 1.34 [CI, 1.22 to 1.47]), whereas inhaled corticosteroids were associated with reduced risk for cardiovascular death (OR, 0.80 [CI, 0.72 to 0.88]). Results were consistent across sensitivity analyses.
Current smoking status and lung function were not measured. Misclassification of cause-specific mortality is unknown.
The possible association between ipratropium and elevated risk for all-cause and cardiovascular death needs further study.
Many think we need more information about the safety of respiratory medications for chronic obstructive pulmonary disease (COPD).
This large case–control study examined associations between medications and risk for death in veterans with newly diagnosed COPD. Inhaled corticosteroids were associated with decreased risk for death. Theophylline and ipratropium were associated with increased risk for respiratory and cardiovascular death, respectively.
Potential confounders, such as smoking status and disease severity, were not known. Associations may not reflect causal relationships.
Additional research about the safety of ipratropium, one of the most commonly prescribed medications for COPD, is needed.
Study flow diagram.
COPD = chronic obstructive pulmonary disease.
Table 1. Participant Characteristics
Table 2. Medication Use Related to Chronic Obstructive Pulmonary Disease
Table 3. Adjusted Odds of Mortality
Risk for mortality associated with respiratory medications in the sensitivity analyses for each study end point.
Bars indicate 95% CIs. CHF = chronic heart failure; ICS = inhaled corticosteroid; IPRA = ipratropium; LABA = long-acting β-agonist; SABA = short-acting β-agonist; THEO = theophylline.
Asakura Medical Association Hospital
September 19, 2008
A potential bias
It is interesting to note that patients taking inhaled corticosteroid (ICS) have reduced mortality in comparison with those taking other types of medication in patients with chronic obstructive pulmonary disease. However, before we start discussion about the direct association of the pharmacological effect of ICS with the reduced mortality, potentially confounding factors to the observation must be considered. Particularly, potential bias in patients deserves attention, as the authors stated. Although the authors proposed a bias in the severity of the disease and smoking status as potential confounders, a potential bias in the sociodemographic distribution among the patients seems to be more relevant to the reduced mortality in patients taking ICS. Although observations in patients with asthma, several studies have suggested that patients with disadvantaged sociodemographic background are less likely to use ICS, either through lower adherence to the treatment (1), or maybe through lower rates of prescription (2). On the other hand, socioeconomic status has been documented to have inverse association with mortality from cardiovascular diseases (3) and other diseases (4). Thus, it is prudent to take these potential confounders into account to adequately appreciate the observed differences in the mortality.
1. Apter AJ, Boston RC, George M, et al. Modifiable barriers to adherence to inhaled steroids among adults with asthma: it's not just black and white. J Allergy Clin Immunol 2003;111(6):1219-26.
2. Kozyrskyj AL, Mustard CA, Simons FE. Socioeconomic status, drug insurance benefits, and new prescriptions for inhaled corticosteroids in schoolchildren with asthma. Arch Pediatr Adolesc Med 2001;155(11):1219 -24.
3. Mackenbach JP, Bos V, Andersen O, et al. Widening socioeconomic inequalities in mortality in six Western European countries. Int J Epidemiol 2003;32(5):830-7.
4. Kunst AE, Groenhof F, Mackenbach JP, Health EW. Occupational class and cause specific mortality in middle aged men in 11 European countries: comparison of population based studies. EU Working Group on Socioeconomic Inequalities in Health. Bmj 1998;316(7145):1636-42.
Nicholas J Gross
Stritch-Loyola School of Medicine
September 24, 2008
Risk of Death Associated with COPD Medications
Despite its very large data base, the study has problems that are inherent in observational data analyses where potential confounding factors cannot be reliably controlled for. Of particular concern in the present study are the absence of data about disease severity and smoking status, as the authors acknowledge. Attempts to adjust for these confounders by reference to a small pilot study of questionable relevance (1) or by internal "measures of disease severity" that are not described in any detail are unreassuring. When the effect size of a treatment of interest, ipratropium versus no ipratropium, is quite small, as here, the inability to adjust with confidence for 2 very weighty confounders throws considerable doubt on the conclusion. Furthermore, after adjustment for covariates the authors calculate the odds ratio for mortality with ipratropium use is reduced to 1.11, which is below the 1.15 threshold of 80% power on which their sample size was calculated.
The inherent methodological problems are probably responsible for a number of other apparent anomalies. I will mention just two. In Table I, the comorbidities hypertension and osteoarthritis appear to be substantially less common in respiratory deaths than case controls; conversely, the use of loop diuretics and digoxin are substantially more common in respiratory deaths. If the odds ratio for each were calculated one might conclude that hypertension and osteoarthritis were protective against, while loop diuretics and digoxin were risk factors for, respiratory death in COPD. Would anyone accept those conclusions?
Secondly in Figure 2, one notes the highest odds ratios for mortality are associated with theophylline usage. But, paradoxically, the risk is seen only in association with respiratory deaths, and not at all with cardiovascular deaths. Theophylline does not have major respiratory side effects. It does, of-course, have serious, sometimes fatal, cardiovascular side effects. How does one make sense of this apparent contradiction?
One is also struck that numerous prospective studies (2) except one very small one (3), have exonerated ipratropium of any associated increase in mortality in COPD, the reverse in fact. The present results could be well explained by disease severity, -a major confounding factor that was not or could not be precisely controlled for. We are left with the uncertainty whether sicker patients received more treatment, or whether the treatment itself contributed to patients' mortality. I suspect the former.
1. Joo MJ, Lee TA, Bartle B, Van de Graaff W, Weiss KB. Patterns of healthcare utilization by COPD severity: a pilot study. Lung 2008;186:307-12
2. Salpeter SR, Buckley NS. Systematic review of clinical outcomes in chronic obstructive pulmonary disease: beta-agonist use compared with anticholinergics and inhaled corticosteroids. Clin Rev Allergy Immunol 2006;31:219-30.
3. Ringbaek T, Viskum K. Is there any association between inhaled ipratropium and mortality in patients with COPD and asthma? Respir Med 2003;97:264-72.
NJG has received payments for consulting with Boehringer-Ingelheim, GlaxoSmithKline, Altana, Almirall, Forest, Dey, and Astra-Zeneca. NJG has received speaking fees from Astra-Zeneca, Boehringer-Ingelheim, Dey, and Pfizer. I have no investments in any Pharmaceutical Company.
Biases in the study of COPD medications on mortality
Lee et al report, using an observational study design, that the use of ipratropium is associated with a 34 % increase in the risk of cardiovascular death and the use of inhaled corticosteroids (ICS) is associated with a 20% reduction in all-cause mortality in COPD patients.(1) Some methodological considerations should be addressed in judging the validity of these findings.
The excess risk of cardiovascular death seen with ipratropium (odds ratio 1.34; 95% CI 1.22-1.47) is possibly due to the marked imbalance in the occurrence of COPD exacerbations during the prior 6 months between cases (64%) and controls (22%).(2) While the data analysis adjusted the odds ratio from 1.75 (crude) to 1.34, the statistical approach did not consider the timing of the exacerbation and was thus unable to distinguish whether the cardiac outcomes occur through the increased likelihood of an exacerbation or whether they require the occurrence of an exacerbation. It is possible that a more refined analysis that can take the timing of exacerbations into account could further reduce the adjusted risk.
The protective effect of inhaled corticosteroids (ICS) on all-cause mortality (odds ratio 0.80; 95% CI 0.78-0.83) contradicts the TORCH randomized controlled trial that found a 6% increase in the death rate (rate ratio 1.06; 95% CI 0.89-1.27) with the ICS fluticasone compared with placebo.(3) Unlike observational studies, the randomized controlled trial design avoids the problem of confounding factors, measured and unmeasured, resulting from imbalances between ICS users and non-users.
The study may also have been subject to immeasurable time bias.(4) In a case-control study of mortality involving a chronic illness such as COPD, hospitalizations just prior to death will be common, so that outpatient prescriptions identified from computerized databases cannot be measured during this hospitalized time period. Thus, hospitalizations before death will lead to an apparently lower use of drugs, while being likely associated with an increased risk of death. This phenomenon will thus create the illusion that the drug is effective at preventing mortality. An illustration of this phenomenon using a cohort of COPD patients showed that fatal cases spent more time in hospital prior to death than controls, so that the odds ratio of death associated with ICS use was 0.60, while a data analysis that accounted for the varying out of hospital measurable times changed this odds ratio to 0.98.(4)
While observational studies are essential to evaluate drug safety, their methods can introduce bias.
(1) Lee TA, Pickard AS, Au DH, Bartle B, Weiss KB. Risk for death associated with medications for recently diagnosed chronic obstructive pulmonary disease. Ann Intern Med. 2008;149:380-390.
(2) Soler-Cataluna JJ, Martinez-Garcia MA, Roman SP, Salcedo E, Navarro M, Ochando R. Severe acute exacerbations and mortality in patients with chronic obstructive pulmonary disease. Thorax. 2005;60:925-31.
(3) Calverley PM, Anderson JA, Celli B, Ferguson GT, Jenkins C, Jones PW et al. Salmeterol and fluticasone propionate and survival in chronic obstructive pulmonary disease. N Engl J Med. 2007;356:775-89.
(4) Suissa S. Immeasurable time bias in observational studies of drug effects on mortality. Am J Epidemiol. 2008;168:329-35.
SS and PE have received grants or served on advisory boards for Astra-Zeneca, Boehringer-Ingelheim, GlaxoSmithKline and Sepracor.
Todd A. Lee
Hines VA Hospital, Hines, IL
November 5, 2008
We thank the correspondents for their thoughtful comments. Drs. Suissa and Ernst raise three concerns about our paper. First, to alleviate concerns that the observed risk for ipratropium and cardiovascular mortality was a result of an increased risk of exacerbations in the cases, we conducted an analysis restricted to those without an exacerbation in the 180 days preceding their event date and found a level of risk (OR 1.45, 95% CI 1.14-1.85) similar to the results reported in the article. Second, unmeasured confounding and bias is one possible explanation for discordant results between the TORCH study and our study. Among the other explanations, the difference in outcomes may be related to the patient populations. The population included in our analysis was older, had higher overall and cause-specific mortality rates, was nearly all male, had more co-existing conditions and had differential respiratory medication use prior to entering the study. Third, we were able to measure medications used within the VA healthcare system during inpatient stays and thus the analysis was not subject to concerns about immeasurable time bias.
Dr. Gross raises concerns about observational studies and our ability to adjust for disease severity and smoking status. The inability to measure smoking status was acknowledged in the paper, and the proportion of current smokers in the ipratropium group needed to be nearly 2.5 times higher than in the comparator group to negate the observed association with ipratropium and cardiovascular mortality. COPD exacerbations and hospitalizations were used to characterize disease severity. The observed association with respiratory mortality and theophylline may have been related to residual confounding by severity, given its place in treatment guidelines. However, this is unlikely to be the case with ipratropium as it is used as a first-line medication in the treatment of COPD. The specific hypotheses tested in this study were based on prior research, including a large prospective study of ipratropium that showed an increased risk of cardiovascular hospitalizations and deaths(1). In addition our findings are consistent with a recent meta-analysis that reported similar risks with anticholinergics and cardiovascular outcomes(2).
Finally, the doctor's notes that socioeconomic status may confound the association between medication use and mortality. An advantage of conducting this analysis among US veterans who received care from Veterans Health Administration is the strong emphasis that VA places on equal access to health care that likely minimizes the effects of socioeconomic status compared to other populations.(3,4)
(1) Anthonisen NR, Connett JE, Enright PL, Manfreda J. Hospitalizations and mortality in the Lung Health Study. Am J Respir Crit Care Med. 2002;166:333-39.
(2) Singh S, Loke YK, Furberg CD. Inhaled anticholinergics and risk of major adverse cardiovascular events in patients with chronic obstructive pulmonary disease: a systematic review and meta-analysis. JAMA. 2008;300:1439-50.
(3) Jha AK, Shlipak MG, Hosmer W, Frances CD, Browner WS. Racial differences in mortality among men hospitalized in the Veterans Affairs health care system. JAMA. 2001;285:297-303.
(4) Selim AJ, Fincke G, Berlowitz DR, Cong Z, Miller DR, Ren XS et al. No racial differences in mortality found among Veterans Health Administration out-patients. J Clin Epidemiol. 2004;57:539-42.
Honoraria: T.A. Lee (Astra-Zeneca, Novartis), D.H. Au (GlaxoSmithKline). Stock ownership or options (other than mutual funds): D.H. Au (Pfizer). Grants received: T.A. Lee (Altana, Aventis, AstraZeneca, Boehringer Ingelheim, Chiesi, GlaxoSmithKline, Merck & Co., Novartis, Pfizer, Schering-Plough, Sepracor, University of Kentucky). Other: D.H. Au (Assessing the Impact of Recent Updates for Advair and Serevent Special Issues Board).
Westchester Medical Center, Valhalla, NY 10595
December 3, 2008
Risk for Death Associated with Medications for Recently Diagnosed COPD
To the Editor:
In their pharmacoepidemiology study, Lee and colleagues (1) report the possible association between ipratropium and elevated risk for all-cause and cardiovascular mortality in patients with newly diagnosed chronic obstructive pulmonary disease (COPD). During an Evidence Based Medicine presentation for a group of internal medicine residents we studied this paper with keen interest in research methodologies. The paper refers to a cohort of 145 020 veterans with a new diagnosis of COPD, which excludes those not receiving any respiratory treatment. But all-cause mortality analysis in their study included 320 501 control participants, some of whom obviously would not belong to their cohort. This makes one wonder the source of controls, who are supposed to be patients with COPD who were alive at the end of the study period. If the control group had included patients without COPD, the sensitivity analysis for unmeasured confounders (severity of COPD in this case) would be inaccurate. Also, in table 3, patients not on any respiratory medications or those only on short acting beta-agonists were considered as the reference standard. But as per the study protocol, patients who were not on any respiratory medications were excluded from the study.
We understand that much of medical research is observational and the reporting of observational studies is often of insufficient quality. Thus, we study such research in light of The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies (2). However, this type of fundamental law in the study design would not have been appreciated without an in depth review of the paper.
1. Lee TA, Pickard S, Au DH, Bartle B, Weiss KB. Risk for death associated with medications for recently diagnosed chronic obstructive pulmonary disease. Ann Intern Med. 2008; 149: 380-390.
2. Von Elm E, Altman DG, Egger M, Pocock SJ, GÃ¸tzsche PC, Vandenbroucke JP; STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for Reporting Observational Studies. Ann Intern Med. 2007; 147: 573-577.
Lee TA, Pickard AS, Au DH, et al. Risk for Death Associated with Medications for Recently Diagnosed Chronic Obstructive Pulmonary Disease. Ann Intern Med. 2008;149:380–390. doi: https://doi.org/10.7326/0003-4819-149-6-200809160-00004
Download citation file:
Published: Ann Intern Med. 2008;149(6):380-390.
Cardiology, Chronic Obstructive Airway Disease, Pulmonary/Critical Care.
Results provided by:
Copyright © 2020 American College of Physicians. All Rights Reserved.
Print ISSN: 0003-4819 | Online ISSN: 1539-3704
Conditions of Use