Markers of disease severity in chronic obstructive pulmonary disease

Background and objectives: Diagnosis and assessment of treatment effect in chronic obstructive pulmonary disease (COPD) have relied primarily on the examination of a complex set of symptoms and the use of spirometry. However, these methods require long periods of assessment to determine whether patients show clinically relevant improvements after intervention. We therefore wanted to determine how existing clinical and laboratory measures change with COPD severity and identify disease markers that can serve as better endpoints for diagnosis and assessment of COPD progression and treatment effect. Methods: Using standard COPD keywords and terms, we searched PubMed, ISI Web of Science, and Cochrane Review databases for retrospective and prospective clinical studies published since 1966. We identified 652 studies (n= 146,255) from 1978 to September 2003 based on the availability of spirometric and demographic data, investigation of possible markers, absence of acute exacerbations and comorbidities, and the withdrawal of standard COPD medication. Central tendencies and dispersions of subject baseline measures were collected according to study sample size, smoking status, and mild, moderate and severe COPD stages. A fixed effect meta-analysis was then conducted on each measure at various disease stages. Results: Arterial oxygen tension, Sputum neutrophils and IL-8, and serum TNF-alpha and C-Reactive Protein showed a trend toward separation between COPD stages. Other measures such as pack-years and St George's Respiratory Questionnaire only distinguished between disease and disease-free states. Conclusions: We observed little separation between disease stages for many measures used in CON) diagnosis and clinical trials. This demonstrates the poor sensitivity of such endpoints to define a patient's clinical status and to quantity treatment effect. Therefore, we recommend that longitudinal studies and disease modelling be the primary methods for assessing whether potential markers of disease progression can be used for COPD diagnosis and clinical trials. (c) 2005 Elsevier Ltd. All rights reserved.