Affiliations: All India Institute of Medical Sciences, New Delhi, India,
Veterans Affairs Medical Center, Minneapolis, MN
Correspondence to: Anant Mohan, MD, Department of Medicine, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110029, India; e-mail: email@example.com
We read with great interest the article by Niewoehner et al,1who attempted to develop risk indexes for predicting hospital admissions and exacerbations of COPD. Several such studies have been conducted earlier, with different combinations of clinical and laboratory variables emerging as prognostic markers.2 The present sample size is large, but a follow-up of 6 months seems somewhat inadequate to develop predictors. It would have been useful to cover a full calendar year to note the monthly and seasonal distribution of exacerbations.
The authors used a definition that, according to them, excludes milder illness. In fact, these relatively minor exacerbations comprise a significant proportion of hospital visits, and although not requiring hospitalization, pose a significant burden on health-care resources. The classification of acute exacerbation of COPD by Anthonisen et al3 is useful to determine severity. Also, the fact that the authors’ definition of exacerbation included hospital admission as one of the criteria makes the end points no longer exclusive. This also applies to productive cough as a predictor when it is one of the diagnostic factors, raising questions of colinearity.
Exacerbations developed in 551 patients during 6 months. It would be useful to know the total number of exacerbations episodes during this period in order to determine predictors of multiple exacerbations and hospitalizations. It would be expected that any hospitalized patient would receive steroids/antibiotics or both. That the points for prior hospitalization are much less than those for steroid and antibiotic use in the prior year is hard to reconcile. Cardiovascular comorbidity increases the risk for hospital admissions but not for exacerbations. This is puzzling since all hospital admissions in this study were assumed to be due to exacerbations of COPD only. It would be interesting to see how the score fares in validation samples. The authors could have done a calibration of the scoring system to identify risk for the individual patient, which is what they set out to do in the first place.
The authors have no conflicts of interest to disclose.
Dr. Niewoehner has received grants, honoraria, or advisory fees from Boehringer Ingelheim, Pfizer, AstraZeneca, Adams Respiratory Therapeutics, and Sanofi Pasteur within the past 3 years.
The limitation of the 6-month follow-up period was clearly acknowledged in our article.1 Despite that limitation, our models demonstrate strong predictive attributes that can be viewed as being both statistically significant and clinically meaningful.
Our statement about excluding minor illness refers to events that might meet a symptom-based definition of exacerbation but are not so severe that patients seek medical attention. We made no claim that the end points in our two models (exacerbations and exacerbations resulting in hospitalization) were exclusive. We thought it appropriate to model hospitalizations separately because of their greater importance in terms of morbidity and cost. If a productive cough was present chronically, the protocol required that there be an increase in its severity to be included as part of the symptom complex describing an exacerbation. Hence, we were not using an existing condition to define a primary outcome.
Predictive models can be developed for multiple events, but the methods are complex and rarely used. They add little predictive value when events are relatively uncommon, as is the case in our trial (mean of 0.44 exacerbations per randomized subject). Antibiotic and systemic corticosteroid use for COPD was more common than hospitalization for COPD in the prior year because most exacerbations were treated as outpatients. It is unclear whether the increased risk that cardiovascular comorbidity confers on COPD hospitalization is due simply to misdiagnosis, or whether the presence of such a condition predisposes to COPD exacerbation by some unknown mechanism. As described in the article, we did perform internal validation of the models by bootstrapping techniques, and we also stated the desirability for external validation.1 Reliability plots are located in the corners of both figures in our article, and these demonstrate that calibration was good for both models.1 We also provided information that allows calculation of absolute risk for individual patients in both models.
Become a CHEST member and receive a FREE subscription as a benefit of membership.
Individuals can purchase this article on ScienceDirect.
Individuals can purchase a subscription to the journal.
Individuals can purchase a subscription to the journal or buy individual articles.
Learn more about membership or Purchase a Full Subscription.
Institutional access is now available through ScienceDirect and can be purchased at myelsevier.com.
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited:
Customize your page view by dragging & repositioning the boxes below.
Enter your username and email address. We'll send you a reminder to the email address on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.