Epidemiology of spirometric test failure.

Excessive variability in spirometric measurements derived from the forced expiratory volume manoeuvre , also called "test failure,"' is a phenomenon familiar to all those who use pulmonary function tests, whether it be in a clinical context2 or in the context of population based studies.3 Because the manoeuvre is effort dependent, test failure is usually attributed to a failure of effort or comprehension on the part of the patient or subject (one editorialist capsulised the issue in the phrase "disability or disinclination")4 or to the incompetence of the technician .256 Adequate effort on the part of the subject tends to be particularly in question in disability assessments or when workmens' compensation is involved.4 To counter this, guidelines have been developed to standardise spirometry, both in the clinical and the epidemiological context.7 In addition to technical recommendations on equipment, its selection and validation, these guidelines recommend that a minimum of three satisfactorily performed trials should be required from a subject undergoing tests (criteria for satisfactory performance are clearly defined), with up to eight trials if this is necessary to meet certain reproducibility criteria.56 For FEV, and FVC, the reproducibility criteria recommended by the American Thoracic Society are that the two best trials should not differ by more than 5% or 100 ml, whichever is the greater6; those recommended by the European Economic Community stipulate that the difference should not exceed 300 ml for FVC.' The purpose of this editorial is to bring to attention the gathering evidence, mostly from epidemiological studies in industry, that failure to perform reproducible spirometry may itself be an indicator of respiratory ill health. For instance, in several cross sectional studies of occupational groups test failure has been shown to relate to symptoms.89 In Chinese textile workers the symptoms were those of byssin-osis,8 in Pennsylvania railroad workers chronic bronchitis ,8 and in coal miners, wheezing and shortness of breath were implicated as well as chronic cough and sputum.9 Similarly, subjects who exhibit test failure have, on average, lower levels of FEV,910 and an increased annual loss of FEV, in two cohort studies in which this has been measured.''O Test failure does not, however, appear to be a significant predictor of death when function level and smoking are taken into account.91O A surprising and unexplained feature in two community based studies was that test failure occurred less frequently in smokers than in non-smokers.'"1 The physiological mechanisms underlying the phenomenon …