Diagnostic accuracy of a pocket screening spirometer in diagnosing chronic obstructive pulmonary disease in general practice: a cross sectional validation study using tertiary care as a reference

Abstract

Background: COPD-6™ is a lung function testing device for a rapid pre-spirometry testing to screen-out at-risk individuals not having COPD and indicating those at risk. The aim of this study was to validate COPD-6™ lung function testing (index test) in general practice in discriminating patients with COPD out of the population at risk - smokers/ex-smokers with no previous diagnosis of COPD, using measurements at tertiary care as reference standard.

Methods: Consecutive 227 subjects (115 women, 185 smokers/42 ex-smokers, ≥20 pack-years) with no previous diagnosis of COPD, aged 52.5 (SD 6.8) years from 26 general practitioners (GPs) were recruited, lung function tested with COPD-6™, referred to the tertiary institution for repeated COPD-6™ testing followed by spirometry with a bronchodilator (salbutamol), examination, and pulmonologist consultation for the diagnosis and severity of COPD.

Results: COPD was diagnosed in 43 subjects (18.9 %), with an AUC of 0.827 (95 % CI 0.769-0.875, P < 0.001) for the diagnosis of COPD when lung function was measured using COPD-6™ in GP's office with a specificity of 100 % (95 % CI, 97.95-100 %) but a very low sensitivity of 32.56 % (95 % CI, 20.49-47.48 %). Significant agreement for forced expiratory volume in 1 s measured at GP's office and at lung function lab was found (mean difference 0.01 L, p = 0.667) but not for other measured parameters (p < 0.001 for all).

Conclusions: Our study results point out that active case finding in a population at risk for COPD should be instituted (almost 20 % of undiagnosed COPD). Based on our results lung function testing with COPD-6™ can substitute spirometry testing in cases where it is not readily available to the patient/physician taken into account that the traditional FEV1/FEV6 cutoff value of <0.7 is not the only criterion for diagnosis and/or further referral.

Trial registration: ClinicalTrials.gov Identifier NCT01550679 Registered 28 September 2014, retrospectively registered.

Keywords: COPD; Diagnosis; General practice; Screening; Sensitivity and specificity.

Figures

Fig. 1

Diagram of flow of patients through the study

Fig. 2

ROC curve for the diagnosis of COPD using COPD-6™ at the GP’s office. ROC curve plot (AUC 0.827, 95 % CI 0.769–0.875, P < 0.001) was based on FEV1/FEV6 measurements using COPD-6™ at the GP’s office using COPD diagnosis made by pulmonologist at tertiary care hospital as criterion variable; dotted lines represent 95 % confidence intervals

Fig. 3

Bland-Altman plots for between methods and points of care comparisons for lung function parameters. Bland-Altman plots are presented as plots of difference between two measurements [y-axis] plotted against the mean of two measurements [x-axis]; graphs at columns represents parameters measured using COPD-6™ at the GP’s office: first column – FEV1, second column – FEV6, third column – FEV1/FEV6; rows represent mean of two measurements: first row – COPD-6™ at tertiary care, second row – spirometry (S), third row – postbronhodilator (PB) spirometry

Fig. 4

Proposed clinical algorithm for the triage of patient at risk for COPD (smokers/ex-smokers with >20 pack-years having chronic respiratory symptoms or comorbidities associated with COPD) using lung function testing with COPD-6™ at GP’s office