Physician-diagnosed chronic obstructive pulmonary disease (COPD) compared to spirometry-defined COPD; the substantial difference

Problem

Underdiagnosis of COPD has shown to be common. Medication therapy in COPD is unsatisfying, with low adherence to treatment regimen. Therefore, we aimed to investigate the prevalence of physician-diagnosed COPD in a population with spirometry-defined COPD, and also to estimate factors associated with physician-diagnosed COPD. A secondary aim was to investigate the prevalence of medication use and to compare between those with/without COPD diagnosis.

Approach

The Swedish participants of the international Prospective Urban Rural Epidemiological study, aged 35-70 years, were included (n = 3 941). Data from mandatory health care registers and data on dispensed medication from pharmacies was linked to all participants. Spirometry-defined COPD was defined as fixed ratio FEV1/FVC < 0.7, and physician-diagnosed COPD was defined as having a code for COPD registered in the medical record before baseline. Logistic regression was performed, controlling for potential confounders.

Findings

Totally 541 individuals (13.7 %) had spirometry-defined COPD and among them, 20 (3.7 %) had physician-diagnosed COPD. At least one prescription of long- or short-acting anticholinergics were dispensed to 70 % (n = 14) of the physician-diagnosed individuals and to 5 % (26 out of 521) of those with spirometry-defined COPD. Having one or more than one respiratory symptom was significantly associated with physician-diagnosed COPD (OR 16.0, CI 3.3-77.0 and OR 37.9, CI 8.1-178.0, respectively). Moderate or severe/very severe air flow obstruction was also significantly associated with physician-diagnosed COPD (OR 17.4, CI 3.9-77.7 and OR 44.9, CI 6.6-304.9, respectively). Age was associated with COPD diagnosis (OR with 1.1, CI 1.01-1.15) in model with symptoms and marginally associated (OR 1.1, CI 0.98-1.12) in model with airway obstruction, while sex, BMI and occupation were not associated with COPD diagnosis in any model. There were too few never-smokers among those with physician-diagnosed COPD to be analyzed in regression model.

Consequences

We found that COPD is both underdiagnosed and medicated to a low extent in Swedish health care. A physician-diagnosed COPD more likely occurred among individuals with more respiratory symptoms, with moderate or severe/very severe air flow obstruction and with higher age. A substantial proportion of the respiratory medication was used by undiagnosed individuals, indicating that patients remain undiagnosed and that these should be diagnosed properly as COPD and expected for further medication therapy.