The British Thoracic Society recommends spirometry, not peak expiratory flow, for diagnosing patients as having chronic obstructive pulmonary disease.1 Recording data from spirometry in patients' notes has been proposed as a marker of quality of care.2 But general practitioners are more familiar with peak expiratory flow rate, and have questioned using more complex spirometry tests to identify chronic obstructive pulmonary disease.3 We analysed data from the third national health and nutrition survey (NHANES III) to investigate how useful peak expiratory flow rate is for detecting people with chronic obstructive pulmonary disease in the community.
Participants, methods, and results
We included only white people aged 50-90 years because chronic obstructive pulmonary disease is uncommon in younger people, and we had insufficient statistical power to study other ethnic groups. We excluded people with self reported asthma.
For the remaining 3874 participants, we extracted information on lung function, history of smoking, and respiratory symptoms. Using equations developed from NHANES III, we calculated percentage predicted lung function.4 We defined peak expiratory flow rates less than 80% of that predicted as abnormal, and people as having chronic obstructive pulmonary disease if we found evidence of airflow obstruction (forced expiratory volume in one second less than 80% of that predicted and a forced expiratory volume to vital capacity ratio of less than 70%), ever smoking at least 100 cigarettes, and had one or more respiratory symptom. These patients were subdivided on the basis of their predicted forced expiratory volume in one second into those with mild (60%-80%), moderate (40%-59%), or severe (< 40%) disease.
We calculated the specificity and sensitivity of an abnormal peak expiratory flow rate. Classic statistical methods were not applicable because of the complex survey design used in NHANES III (www.cdc.gov/nchs/data/nhanes/nhanes3/cdrom/nchs/manuals/nh3guide.pdf); we used specialised survey commands within STATA, weighting our data appropriately.
We identified 265 people with chronic obstructive pulmonary disease—a prevalence of 7.8%, after allowing for NHANES III's sampling method. Of these, 143 (54%) were men, and the overall mean age was 65 years. Among people with a diagnosis of chronic obstructive pulmonary disease, 235 had a peak expiratory flow rate of less than 80% of what we predicted. The sensitivity (adjusted for sampling methods) of an abnormal peak expiratory flow rate in detecting all people with chronic obstructive pulmonary disease was 91%, and for people with moderate or severe chronic obstructive pulmonary disease was 100% (table). The specificity of an abnormal peak expiratory flow rate was lower at 82%, although 62% of the false positive cases were smokers and 47% had airflow obstruction on spirometry.
Table 1.
Diagnoses of chronic obstructed pulmonary disease and peak expiratory flow rate
| Peak expiratory flow rate | None | Mild (60%-80%) | Moderate (40%-59%) | Severe (<40%) | Total |
|---|---|---|---|---|---|
| <80% predicited | |||||
| No of participants | 679 | 126 | 76 | 33 | 235 |
| Crude % | 17.5 | 3.3 | 2.0 | 0.9 | 6.1 |
| Adjusted* % | 16.6 | 4.0 | 2.2 | 0.8 | 7.1 |
| ≥80% predicted or greater | |||||
| No of participants | 2930 | 30 | 0 | 0 | 30 |
| Crude % | 75.6 | 0.8 | — | — | 0.8 |
| Adjusted* % | 75.6 | 0.7 | — | — | 0.7% |
Adjusted to allow for sampling in NHANES III. Crude values for peak expiratory flow rate <80%: sensitivity 89%, specificity 76%, positive predictive value 26%. Values for peak expiratory flow rate <80% predicted allowing for sampling procedure: sensitivity 91%, specificity 82%, positive predictive value 30%
Comment
A peak expiratory flow rate of less than 80% will detect more than 90% of people with chronic obstructive pulmonary disease in the community, including all of those with moderate or severe disease—that is, patients most likely to benefit from treatment with bronchodilators. The specificity of the test was more limited and for this reason the false positive rate is appreciable. The high prevalence of smokers and subjects with airflow obstruction in the false positive group, however, suggests that a peak expiratory flow rate of less than 80% predicted may be picking up milder cases of chronic obstructive pulmonary disease that did not meet our stringent diagnostic criteria.
Our findings suggest that peak expiratory flow rate is good at detecting patients with chronic obstructive pulmonary disease in the community. Spirometry measurements provide additional information, but are more complex and time consuming, and their benefit in primary care has not been quantified. The new general medical services contract should therefore concentrate more on interventions of proved benefit to patients with chronic obstructive pulmonary disease, such as smoking cessation,5 and less on the need for spirometry.
We thank David Coultas for help designing this study and advice on using the NHANES dataset and Tricia McKeever, Liam Smeeth, and John Britton for their comments on the manuscript.
Contributors: HJ did a literature review, did all statistical analysis, and drafted the paper. RH had the idea for the study, helped with statistical analyses, and was responsible for drafing the paper. RH is guarantor.
Funding: No additional funding. RH is a Wellcome Trust advanced fellow.
Competing interests: RH has been reimbursed by Bayer to attend two conferences and has also received a consultancy fee from Bayer which enabled him to attend a conference in Tashkent for planning research projects in Karakarlpakstan.
Ethical approval: None sought.
References
- 1.British Thoracic Society. Guidelines for the management of chronic obstructive pulmonary disease. Thorax 1997;52(suppl): S1-28. [Google Scholar]
- 2.British Medical Association. Investing in general practice: the new general medical services contract. London: BMA, 2003. www.bma.org.uk/ap.nsf/Content/NewGMSContract (accessed 10 Jul 2003).
- 3.Nolan D, White P. FEV1 and PEF in chronic obstructive pulmonary disease management. Thorax 1999;54: 468. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hankinson JL, Odencrantz JR, Fedan KB. Spirometric reference values from a sample of the general US population. Am J Crit Care Med 1999;159: 179-87. [DOI] [PubMed] [Google Scholar]
- 5.Anthonisen NR, Connett JE, Kiley JP, Altose DM, Bailey WC, Buist AS, et al. Effects of smoking intervention and use of an inhaled anticholinergic bronchodilator on the rate of decline of FEV1. The Lung Health Study. JAMA 1994;272: 1497-505. [PubMed] [Google Scholar]