An Almost Pragmatic Trial for COPD Screening

Dr. Ulrik and her collaborators1 have reported the positive results of an almost pragmatic trial2 supporting COPD case finding/screening. The study is done in a large number of community practices presumably without extensive research support. The patients are not excluded based on anything other than lack of risk factors for COPD (smoking, occupational exposures and age). The question is a practical one asked by everyone from practicing physicians to those developing national and international guidelines and policies for COPD management—ìDoes case finding identify undiagnosed early COPD?î The answer from the authors is a clear yes with reported NNS (numbers needed to screen) that are some of the lowest for any ìscreeningî procedure. Of course this is not a screening procedure. But the answer is useful–if you identify a high risk pool of patients, the procedures presented can identify new cases of COPD.

However, another perhaps more important question remains, does this early identification make a difference? Many of the cases Ulrik et al identified were mild with FEV1 of >80% of predicted.1 The people had some symptoms or they would not have been included in the group for further workup.1 So ideally, we would be able to provide some therapy for those symptoms. For patients with MRC scores of 0 and 1 we can provide PRN short acting bronchodilators and that may improve the scores from 1 down to 0.3 Our ability to improve or modify the other symptoms assessed such as cough, is less clear. Every patient, physician and pharmaceutical company wishes they had an effective therapy for cough.4,5 The ability to improve sputum production is also not clear. Smoking abstinence—not just cessation—may improve both cough and bothersome sputum production as well as modify the progression of lung function decline.2 A diagnosis of COPD has been shown by some to improve rates of cessation although duration of cessation is less clear.6

The new 2011 GOLD guidelines might be useful for determining which of the patients with early diagnoses could benefit from pharmacotherapy.3 Moving beyond simply staging using FEV1, the 2011 GOLD guidelines also use grades from A to D that provide information on the patient's symptom and risk burden.3 Grade A includes COPD patients with few of the symptoms that are most amenable to therapy—e.g. dyspnea is mild, and risk of exacerbations is low plus they have near normal FEV1. So identifying those in grade A may not result in significant improvement in the burden of COPD over the next few years for those patients other than perhaps increasing smoking cessation rates—not a sure thing. Those in grade B, have more dyspnea and therefore may be more likely to be improved with therapy and more likely to add strength to the evidence for the benefit of early diagnosis. Those in category C and D are clearly likely to be improved. Ulrik and colleagues could use the data they have collected (e.g. MRC score and number of previous lower respiratory infections)1 to provide additional information on the potential benefit in this primary care practice based group of individuals with newly identified COPD based on GOLD COPD grades. The refinements of the GOLD guidelines may prove to be useful for both pragmatic research and clinical practice.3

Ulrik and colleagues’ calculations for NNS are not entirely clear and perhaps the results of 3.9 to 4.6 are over-optimistic.1 This is all tied up in the almost a pragmatic trial issue.2 Ulrik and her colleagues pre-selected or pre-screened the patients to include only those who were over 35 years of age and had ever been smokers or had occupational exposures.1 Who did that assessment? Few EHRs could sort out occupational exposure and unfortunately, search of the EHR may miss up to 50% of ever smokers.7 Therefore the NNS needs to take into account this pre-screening work noting how many were prescreened to identify the group that were determined to be ìhigh riskî. Including information on the numbers needed to pre-screen and then include in ìcase-findingî would provide a more accurate estimate of the work required by a practice to identify a new case of COPD. Future trials should become more fully pragmatic by reporting every step of the process required to identify high risk patients and then go on to report what Ulrik et al have reported.1

Both the investigators and the sponsors have provided us with important additional information on COPD screening. This is one more piece to move us toward an adequate evidence basis for routine COPD case finding/screening. It is particularly exciting to see that as government research funds become less available, two pharmaceutical companies have stepped up to support this practice based study.