Interpretative strategies in assessing spirometric test results
Spirometric test results should be interpreted in conjunction
with predicted values and their lower limits of normal, clinical data,
and when available the response to bronchodilator drugs.
There are many publications on predicted values. The ‘lower limit of normal’ is commonly derived in either of three ways:
- Predicted value minus 1.645 the standard deviation of the predicted value. This is a procedure to estimate the lower 5th percentile. It is based on the assumption that the scatter around the predicted value is normally distributed. In practice this condition is often not fully met.
- From the 5th percentile. Unfortunately in most publications the 5th percentile is not mentioned, so that option 1 is then the best approximation.
- Use 80% of the predicted value als the ‘lower limit of normal’. This is scientifically unfounded (ref. 1), but nevertheless often done.
It matters which set of reference values for spirometry is adopted, and how the ‘lower limit of normal’ is determined, in particular in the case of borderline values. This can be shown by the following case of a Caucasian male, standing height 1.78 m, age 55 yr, observed FEV1 3.06 L. Predicted values and lower limits according to some published equations are as follows:
|Langhammer||3.95||3.24||3.16||too low||too low|
|Roca||3.88||3.15||3.10||too low||too low|
As shown in the table, both the sensitivity and the specificity of spirometric test results is influenced by the selection of reference values as well as by the strategy adopted to compute the ‘lower limit of normal’:
- There are appreciable differences between predicted values and their associated ‘lower limits of normal’. On that account the FEV1 of 3.06 L is within the ‘normal range’ according to most prediction equations when using either the 'lower limit of normal' is specified by the author, or approximating it from the predicted mean minus 1.64·RSD. Application of the 80%of predicted rule, which for obscure reasons is still used, leads to a considerable bias and potentially to overdiagnosis and overtreatment.
note that in younger and in elderly subjects, and in taller
subjects, the 80% rule for establishing the ‘lower limit of
normal’ leads to greater errors of judgement.
|Ref. 1 - Do not use per cent predicted|
|1||Sobol BJ. Assessment of ventilatory abnormality in the asymptomatic subject: an exercise in futility. Thorax 1966; 21: 445-449.|
|2||Oldham PD. Percent of predicted as the limit of normal in pulmonary function testing: a statistically valid approach. Thorax 1979; 34: 569.|
|3||Miller A. Prediction equations and ‘normal values’. In: Miller A, ed. Pulmonary function tests in clinical and occupational lung disease. New York, Grune & Stratton, 1986; 197-213|
|4||Miller MR, Pincock AC. Predicted values: how should we use them? Thorax 1988; 43: 265-267.|
|5||Quanjer PhH. Predicted values: how should we use them (letter). Thorax 1988; 43: 663-664.|
|6||ATS Statement. Lung function testing: selection of reference values and interpretative strategies. Am Rev Respir Dis 1991; 144: 1202-1218.|
|7||Quanjer PhH, Tammeling GJ, Cotes JE, Pedersen OF, Peslin R, Yernault JC. Lung volumes and forced ventilatory flows. Eur Respir J 1993; 6 suppl. 16: 5-40. Erratum Eur Respir J 1995; 8: 1629.|