FEV1%predicted versus Lower Limit of Normal for FEV1
Background. As delineated earlier authors of prediction equations have in the last decades gone to great lengths to delineate the Lower Limit of Normal (LLN) for their prediction equation. By convention this LLN for FEV1 and FVC, or for FEV1/FVC, is designed to be the 5th percentile for the index.
It is as if scientists, who provide the prediction equations, and clinical workers live in different worlds. It is common clinical. In clinical work it is common practice to convert a measured value to percent predicted, a proc edure none of the authors ever recommended. This would not matter much if the scatter were proportional to the level of the predicted value. But it is not. Invariably in adults there is no discernable trend that the scatter about the predicted value is larger in tall than in shorter subjects, or larger in young than in old subjects. Yet the predicted values very considerably between these subjects. This implies that any percentile is equally far removed from the predicted value in tall, short, young and old subjects. Therefore, as the predicted value declines with age, and is smaller in shorter subjects, the 5th percentile will be a much lower percentage in those with low predicted values than in those with high predicted values. Transforming a measured value into percent predicted therefore inevitably introduces a height and age based bias. Its use should hence be abandoned in favor of the use of the published lower limit of normal.
GOLD guidelines: over-diagnosis of COPD and its severity. Let us consider how the above affects the GOLD staging of severity of airway obstruction. For your convenience the table on the right is reproduced from the GOLD 2006 update; we certainly do not recommend that you apply these guidelines. According to the GOLD guidelines an FEV1/FVC ratio of < 0.70 implies airway obstruction. We have already explained that this is an improper criterion as it leads to false-positive findings particularly in elderly and tall subjects. If the FEV1/FVC ratio < 0.70 and the FEV1 is > 80 percent predicted, this is regarded as mild airway obstruction (GOLD stage 1), if it is less than 80 percent predicted this signifies moderate airway obstruction (GOLD stage 2). It so happens that particularly above age 50 yr the LLN for FEV1 is very frequently well below 80% predicted. Therefore many people in whom FEV1/FVC is within the normal range but below 0.7, and whose FEV1 is also within the normal range, will not only be erroneously judged to have mild airway obstruction, but a very significant proportion will even be erroneously judged to have moderate airway obstruction. Thus there will not only be considerable over-diagnosis of the prevalence of airway obstruction, but also of the classification of its severity. Predictably such erroneous judgments are going to be quite frequent in subjects older than 50 yr.
The above illustration is based on predicted values for FEV1/FVC and FEV1 published by Falaschetti et al. The prediction equations are based on data collected in the 1995 and 1996 Health Surveys for England and pertain to a few thousand males, non-smoking males with no history of asthma or respiratory symptoms. From about age 47 yr the LLN for FEV1/FVC drops below the 0.7 line, so according to the GOLD guidelines there is airway obstruction. From age 47 to age 49 yr the LLN for FEV1 is above 80% predicted, so that people above that LLN would be regarded to have mild airway obstruction. However, if a subject tracks along that LLN for FEV1, as from age 50 yr the FEV1 will be less than 80% predicted. Hence a perfectly normal evolution of FEV1 with age automatically means that the subject will now be regarded to have moderate airway obstruction.
GOLD guidelines in follow-up studies. This goes to show that in follow-up studies using GOLD guidelines (i.e. using percent predicted), one should expect that people above age 50 yr migrate from ‘normal’ to ‘mild airway obstruction’, and from mild to moderate airway obstruction, because the grading system is flawed. Due to this bias in the system there is no way of telling from spirometric data whether this progression is spurious or is a result of lung disease. It means in practice that GOLD stages 1 and 2 in middle-aged and elderly people should be taken with a grain of salt, and that clinical criteria should come into play if an intervention is considered.
GOLD guidelines in clinical trials. Using GOLD guidelines for classification in clinical trials that take place over several years can lead to the false interpretation of spirometric data. For example, presently there are several large-scale multi-year worldwide clinical trials involving inhaled insulin that use FEV1 as a primary safety outcome measure. Individuals within clinical groupings may be classified as developing airway obstruction with increasing severity simply because the study progresses. These increased incidence rates for adverse advents could result in medical treatments being discontinued, or new therapies not being approved.
Conclusion: avoid the above problems by using the lower limit of normal and not percent predicted. Consequently, do not apply the GOLD guidelines to your patient’s spirometric data.
1.Falaschetti E, Laiho J, Primatesta P, Purdon S: Prediction equations for normal and low lung function from the Health Survey for England. Eur Respir J 2004; 23: 456-463.