It is common practice for the results of spirometry to be interpreted in relation to reference values and in terms of whether or not they are considered to be within the 'normal' range [8,9]. Most equipment manufacturers follow these guidelines.
There may be potential causes for variation in clinical measurement, which include:
• Technical variation of the instrument.
• Performance of the test.
• Interpretation of the procedure by the operator.
• Position of patient during the procedure.
These variations must be evaluated and standardized as much as possible.
Clinically, the most important factors responsible for individual variations are: 1, gender; 2, height; 3, age; and 4, ethnic origin, together with the presence or absence of respiratory disease. Compared with a Caucasian population, black races tend to have predicted normal values approximately 13% less. Asians are intermediate.
The distribution of FEV1 and FVC in population studies are near to Gaussian in the middle range but less so at the extremes. Reference values are most commonly calculated by a linear regression equation, but care should be taken in interpreting data outside the age range from which the population of normal individuals is sampled.
In clinical practice at hospitals and in primary care, values of FEV1 and FVC are traditionally expressed as a percentage of the mean normal value for that individual. A value below 80% predicted is said to be abnormal. This has the major advantage that it immediately defines the level of severity of COPD present within a given patient. Many electronic spirometers provide results in this format. However, statistically, a more accurate representation of normal and abnormal values is by using the 95% confidence limits of the regression equation. An abnormal result is one which falls below the 5th percentile range. The figure can be calculated from: lower limit of normal = predicted value - 1.645 x SEE, where SEE is the standard error of the estimate (the average standard deviation, SD, of the data around the regression line). A recent review by Quadrelli et al.  has compared normal values using the two methods above for different prediction equations and found that, particularly in shorter and more elderly people, the lower normal range figure is often in the 60-80% predicted range. Thus, the percentile calculation provides a more accurate assessment of normal limits, where percentage predicted can provide a measure of the degree of deviation from the predicted value. Data for patients whose values lie close to lower limits should be interpreted with caution. It is also not acceptable to use a fixed FEV1/FVC ratio as a lower limit of normal .
Was this article helpful?