Article Text

PDF

Smoking, age, and laryngeal height predicted obstructive airway disease in adults
  1. Brian J Smith, MBBS
  1. North West Adelaide Health Services Woodville, South Australia, Australia

    Statistics from Altmetric.com

    
 
 QUESTION: How accurate are the clinical examination and patient history in the diagnosis of obstructive airway disease (OAD)?

    Design

    Blinded comparison of items from the clinical examination and patient history (self reported OAD, smoking history, wheezing on auscultation, maximum and minimum laryngeal height, and laryngeal descent) with spirometric results (FEV1 and FEV1–FVC ratio 5th centile).

    Setting

    20 investigator groups in primary, secondary, and tertiary care centres in 14 countries.

    Patients

    332 patients were recruited, and 309 (mean age 56 y, 57% men, 89% white) were included in the analysis. Patients were recruited so that approximately one third had confirmed OAD, one third had suspected OAD, and one third had no evidence of OAD. Exclusion criteria were reversible airway obstruction, terminal illness, age <18 years, severe respiratory distress, or inability to obtain spirometry.

    Description of tests and diagnostic standard

    Clinical examination items were derived from previous research and consensus: self reported history of chronic OAD, smoking history, laryngeal height and descent, and wheezing on auscultation. Spirometry was done within 30 minutes of clinical examination with the better attempt of 2 efforts recorded. OAD was confirmed if FEV1 and the FEV1–FVC ratio were <5th centile. Spirometrists and clinicians were blinded.

    Main outcome measures

    Positive and negative likelihood ratios (+LRs, –LRs) using multivariate analysis.

    Main results

    The reduced multivariate analysis indicated 4 items that, when all were present, were highly predictive of OAD in all patients: self reported chronic OAD, smoking >40 pack years, age ≥45 y, and maximum laryngeal height ≤4 cm. The presence of 3 items predicted OAD in patients with no history of OAD: smoking >40 pack years, age ≥45 y, and maximum laryngeal height ≤4 cm (LRs in the table). Wheezing and laryngeal descent were not predictive of OAD.

    Multivariate likelihood ratios (LRs) calculated from clinical examination and history for diagnosis of obstructive airway disease (OAD)*

    Conclusion

    Obstructive airway disease was highly likely in the presence of smoking >40 pack years, age ≥45 years, maximum laryngeal height ≤4 cm, and self reported history of OAD.

    Commentary

    The Clinical Assessment of the Reliability of the Examination (CARE) investigators have used the internet to recruit consecutive patients with a range of OAD likelihoods. The spirometric criteria for OAD differ from the commonly used Thoracic Societies criteria, but are sound. The process and criteria for excluding patients with asthma from the study are uncertain, but they may be important because airways disease includes a spectrum of reversibility. A similar process of excluding patients with asthma in clinical practice is necessary for the reported nomogram to retain external validity. Intraobserver and interobserver variations were not measured, but they may be important because a similar study showed only marginal agreement among observers.1 The study, however, provides important utility data for each of the clinical features tested, 2 of which can now be confidently removed from clinical practice when determining the likelihood of OAD, namely, wheezing on auscultation and laryngeal descent.

    The ability to rule out OAD is also important for the use of pulmonary function test resources. Comparison between the nomogram and clinicians' judgment would now be interesting. For example, a patient who is a non-smoker and <45 years old and who denies a history of OAD may either have OAD ruled out using clinical judgment or be sent for pulmonary function testing. Patients with an intermediate probability of OAD will benefit most from the nomogram.

    Internet use for recruitment of >300 patients in 1 month, with a very low data error rate, is impressive. Future CARE reports are awaited with interest because they may affect not only the interpretation of diagnostic tests once clinical pretest likelihoods have been determined but also the economical use of diagnostic tests.

    References

    View Abstract

    Footnotes

    • Sources of funding: UK National Health Service Research and Development Programme and Medical Research Council of Canada.

    • For correspondence: Dr F A McAlister, Division of General Internal Medicine, 2E3.24 Walter Mackenzie Centre, University of Alberta Hospital, 8440 112 Street, Edmonton, Alberta T6G 2R7, Canada. Fax +1 780 407 2680.

    Request permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.