Obstructive and restrictive spirometric patterns: fixed cutoffs for FEV1/FEV6 and FEV6
 J. Vandevoorde ^{1} ,
 S. Verbanck ^{2} ,
 D. Schuermans ^{2} ,
 J. Kartounian ^{1} and
 W. Vincken ^{2}
 ^{1}Dept of General Practice and, ^{2}Respiratory Division, Academic Hospital, University of Brussels, Brussels, Belgium.
 J. Vandevoorde, University of Brussels (Vrije Universiteit Brussel), Dept of General Practice, Laarbeeklaan 103, B1090 Brussels, Belgium. Fax: 32 52373311. Email: Jan.Vandevoorde{at}vub.ac.be
Abstract
The purpose of this study was to determine fixed cutoff points for forced expiratory volume in one second (FEV_{1})/FEV_{6} and FEV_{6} as an alternative for FEV_{1}/forced vital capacity (FVC) and FVC in the detection of obstructive and restrictive spirometric patterns, respectively.
For the study, a total of 11,676 spirometric examinations, which took place on Caucasian subjects aged between 20–80 yrs, were analysed. Receiver–operator characteristic curves were used to determine the FEV_{1}/FEV_{6} ratio and FEV_{6} value that corresponded to the optimal combination of sensitivity and specificity, compared with the commonly used fixed cutoff term for FEV_{1}/FVC and FVC.
The data from the current study indicate that FEV_{1}/ FEV_{6} <73% and FEV_{6} <82% predicted can be used as a valid alternative for the FEV_{1}/FVC <70% and FVC <80% pred cutoff points for the detection of obstruction and restriction, respectively. The statistical analysis demonstrated very good, overall, agreement between the two categorisation schemes. For the spirometric diagnosis of airway obstruction (prevalence of 45.9%), FEV_{1}/FEV_{6} sensitivity and specificity were 94.4 and 93.3%, respectively; the positive and negative predictive values were 92.2 and 95.2%, respectively. For the spirometric detection of a restrictive pattern (prevalence of 14.9%), FEV_{6} sensitivity and specificity were 95.9 and 98.6%, respectively; the positive and negative predictive values were 92.2 and 99.3%, respectively.
This study demonstrates that forced expiratory volume in one second/forced expiratory volume in six seconds <73% and forced expiratory volume in six seconds <82% predicted, can be used as valid alternatives to forced expiratory volume in one second/forced vital capacity <70% and forced vital capacity <80% predicted, as fixed cutoff terms for the detection of an obstructive or restrictive spirometric pattern in adults.
 Chronic obstructive pulmonary disease
 forced expiratory volume in six seconds
 pulmonary function testing
 spirometry
Spirometry is the most frequently performed pulmonary function test and is an essential tool for the diagnosis and followup of respiratory diseases. Handheld office spirometers are now widely available for use in primary care and improvements in spirometry software have resulted in access to improved spirometric tests. Several studies emphasised the importance of spirometry in primary care, as a screening tool for the early detection of chronic obstructive pulmonary disease (COPD) 1–4. This has resulted in the need for easytoperform spirometry tests. Increasing evidence showed that the forced expiratory volume in six seconds (FEV_{6}) 5, can be used as a convenient alternative for forced vital capacity (FVC) 6–9. The use of six seconds expiratory manoeuvres makes office spirometry easier and faster, providing a more explicit endoftest definition and reduces the risk of syncope 10.
An important issue in spirometry is the definition of abnormality. The American Thoracic Society and the European Respiratory Society (ERS) guidelines recommend the use of reference equations, derived from a representative sample of healthy subjects, to determine lower limits of normal (LLN) 11, 12, taking into account that the spirometric indices are influenced by age, height, sex and ethnicity. It has already been demonstrated that FEV_{1}/FEV_{6} is a valid alternative for FEV_{1}/FVC when using LLN based on the third National Health and Nutrition Examination Survey (NHANES III) reference equations 6, 7, 9. At present, spirometers, and in particular handheld spirometers, do not always provide reference equationbased LLN. Moreover, reference equations for FEV_{6} and FEV_{1}/FEV_{6} are only available for the USA population (NHANES III survey) 13 and only recently for European subjects in the 65–85yrs category 14. Adding to the complexity of the problem, reference equations obtained from different studies lead to significant differences in predicted FVC and FEV_{1} 15.
Presently, it is common practice to determine airway obstruction by use of a fixed cutoff point, i.e. when FEV_{1}/FVC is <70%, according to the Guidelines from the Global Initiative for Chronic Obstructive Lung Disease (GOLD) 16.The aim of this study is to determine an alternative for the fixed cutoff points of FEV_{1}/FVC <70% and FVC <80% predicted suitable for the use of FEV_{1}/FEV_{6} and FEV_{6}, respectively.
METHODS
The data of consecutive adult patients, referred to the lung function laboratory of the Academic Hospital of the University of Brussels (AZVUB, Brussels, Belgium), between February 1992 and December 2000, were analysed. Spirometry measurements were performed with a massflow sensor (SensorMedics model 2200, Viasys Health Care, Yorba Linda, CA, USA), by highly trained and experienced pulmonary function technicians, according to the guidelines of the ERS 12.
For the diagnosis of airway obstruction, FEV_{1}/FVC <70% was used as a fixed cutoff point, according to the GOLD guidelines 16. From a receiver–operator characteristic (ROC) curve, the FEV_{1}/FEV_{6} ratio, that corresponded to the optimal combination of sensitivity and specificity (i.e. the greatest sum of both), was determined.
A subject was said to have a restrictive spirometric pattern if there was a reduced FVC in the presence of a normal FEV_{1}/FVC. A fixed cutoff of 80% of the predicted value for FVC was used as the “gold standard”. The FEV_{6} value that corresponded to the optimal combination of sensitivity and specificity was determined from a ROC curve.
To calculate sensitivity and specificity for FEV_{1}/FEV_{6} and FEV_{6} as a predictor for obstruction or a restrictive spirometric pattern, 2×2 tables were used. For both indices the positive predictive value (PPV) and the negative predictive value (NPV) were also calculated. The PPV represents the proportion of patients with abnormal test results who have the disease and the NPV represents the proportion of patients with normal test results who do not have the disease. Furthermore, in each analysis the discordant cases, i.e. false positives and false negatives, were scrutinised.
Finally, agreement between the two categorisation schemes, based either on FVC or on FEV_{6}, was assessed using kappa statistics: the number of obstructive patients were determined using FEV_{1}/FVC <70% as a fixed cutoff point. In the nonobstructive patients, a restrictive spirometric pattern was considered if FVC <80% of the predicted value. Similarly, the cutoffs, obtained by a ROC curve for FEV_{1}/FEV_{6} and FEV_{6}, were used as a fixed cutoff to determine the number of patients with a normal, obstructive or restrictive spirometric pattern. The resulting classifications, based on either FVC or FEV_{6}related indices, were combined in a 3×3 table, and a kappa value was calculated. Kappa represents the agreement between the two categorisation schemes in excess of the amount of agreement that would be expected by chance.
RESULTS
The same study population was used as in a previous study, by the current authors 9, comparing FEV_{6} and FVC using LLN determined with the NHANES III reference equations 13. Spirometric data from 11,676 Caucasian subjects were studied, of whom 7,010 (60%) were male and 4,666 (40%) were female. Subject characteristics are shown in table 1⇓. In this table, FEV_{1}/FVC <70% and FVC <80% pred were used for the diagnosis of obstruction and a restrictive pattern, respectively.
The obstructive group was further classified into subgroups according to the severity of airway obstruction in accordance with the GOLD guidelines 16: FEV_{1}/FVC <70%, in combination with FEV_{1} ≥80% pred (Stage I), or 50%≤FEV_{1}<80% pred (Stage II), or 30%≤FEV_{1}<50% pred (Stage III), or FEV_{1} ≤30% pred (Stage IV).
Spirometric diagnosis of obstruction
Considering FEV_{1}/FVC <70% as being the ‘gold standard’ for obstruction, a ROC curve was used to determine the best corresponding cutoff for FEV_{1}/FEV_{6} (fig. 1⇓). The area under the ROC curve was 98.8% (95% confidence interval (CI): 98.6–98.9%), and the FEV_{1}/FEV_{6} cutoff, corresponding to the greatest sum of sensitivity and specificity, was 73%. When using a FEV_{1}/FEV_{6} cutoff of 76%, sensitivity reached 100%, but specificity dropped to 71.7%. Choosing a fixed cutoff of FEV_{1}/FEV_{6} <70% resulted in a specificity of 100%, with a sensitivity of 84.4%. Table 2⇓ shows the sensitivity and specificity results using FEV_{1}/FEV_{6} <73% as a fixed cutoff. For the total population, FEV_{1}/FEV_{6} sensitivity and specificity were 94.4 and 93.3%, respectively. The PPV and NPV were 92.2 and 95.2%, respectively. The prevalence of obstruction was 45.9%. Similar results were obtained when considering male and female subjects separately (data not shown).
Analysis of the 726 discordant cases (false positives and false negatives combined) showed that 98.8% of the discordant values of FEV_{1}/FEV_{6} are within a ±5% interval of the chosen fixed cutoff of 73%. Only one subject was found to have a FEV_{1}/FEV_{6} which differed by >10% from the cutoff.
In the 426 false positive cases, the mean difference of FEV_{1}/FVC and FEV_{1}/FEV_{6}, with their respective LLN, was 0.9 (sd = 1.0) and −2.1% (sd = 1.0 %). In the 300 false negative cases the mean difference of FEV_{1}/FVC and FEV_{1}/FEV_{6} with their respective LLN was 3.6 (sd = 2.6) and 1.5% (sd = 2.0%).
Spirometric detection of restriction
In all subjects with normal FEV_{1}/FVC (n = 6,319), FVC <80% pred was considered as a ‘gold standard’ for the detection of a restrictive spirometric pattern. A ROC analysis showed that a fixed cutoff of FEV_{6} <82% pred resulted in the best combination (i.e. greatest sum) of sensitivity and specificity (fig. 2⇓). The area under the ROC curve was 99.5% (95% CI = 99.5–99.6%). When using FEV_{6} <84% pred, sensitivity reached 100% with a specificity of 95.5%. Choosing FEV_{6} <80% resulted in a specificity of 100%, but sensitivity dropped to 88%. Table 3⇓ shows the current findings using FEV_{6} <82% pred as a fixed cutoff. For the total population, sensitivity and specificity was 95.9 and 98.6%, respectively. The PPV was 92.2% and the NPV 99.3%. The prevalence of a restrictive pattern was 14.9%. Similar results were obtained for both male and female populations (data not shown).
Analysis of the 116 discordant cases (false positives and false negatives combined) showed that 94.0% of the discordant values of FEV_{6} are within a ±5% interval of the chosen fixed cutoff of 82% pred. None of these results differed by >10% with the cutoff (data not shown).
In the 77 false positive cases, the mean difference of FVC and FEV_{6}, with their respective cutoffs, was 1.1 (sd = 1.4) and −2.5% (sd = 1.6%). In the 39 false negative cases, the mean difference of FVC and FEV_{6} with their respective cutoff was −1.8 (sd = 0.8) and 0.5% (sd = 0.7%).
Overall agreement using kappa statistics
Overall agreement between the two categorisation schemes was assessed using kappa statistics (table 4⇓). In this study, a kappa value of 0.87 (95% CI = 0.86–0.88) was obtained, indicating a very good agreement between FEV_{6} and FVCderived indices.
DISCUSSION
It has already been demonstrated that FEV_{6} is a reliable alternative for FVC to identify obstructive and restrictive spirometric patterns, using the NHANES III reference equations to calculate LLN for each spirometric index 7–9. The main purpose of the present study was to determine a fixed cutoff for the FEV_{1}/FEV_{6} ratio and for the FEV_{6} which are equivalent to the commonly used fixed cutoffs for the FEV_{1}/FVC ratio and for the FVC.
Indeed, with a kappa value of 0.87, a very good overall performance was obtained for FEV_{1}/FEV_{6} <73% and FEV_{6} <82% pred as fixed cutoffs for the detection of obstructive and restrictive spirometric patterns, respectively (table 4⇑). A kappa value of 1 indicates perfect agreement, while a kappa value of 0 indicates that agreement is no better than chance. Landis and Koch 17 have proposed the following as standards for strength of agreement for the kappa coefficient: 0.01–0.20 = slight, 0.21–0.40 = fair, 0.41–0.60 = moderate, 0.61–0.80 = substantial and 0.81–1 = almost perfect agreement.
Spirometric diagnosis of obstruction
A ROC curve analysis showed that FEV_{1}/FEV_{6} <73% was the cutoff with the best combination of sensitivity and specificity, and performed well as a surrogate for FEV_{1}/FVC <70% (table 2⇑). In addition, almost all of the discordant cases were close to the cutoff value.
The findings of the present study apply to a population with an overall prevalence of airway obstruction of 45.9%, which corresponds to reported prevalences of COPD of 30–50% in highrisk populations, i.e. smokers aged >45 yrs and subjects with respiratory symptoms. This makes FEV_{1}/FEV_{6} suitable for screening purposes in primary care.
It should be emphasised that fixed cutoff values should be used with caution, as spirometric indices are highly influenced by age, height, sex and race. From the NHANES III survey, reference equations have become available to calculate predictive values and LLN for FEV_{1}, FVC, FEV_{6}, FEV_{1}/FVC and FEV_{1}/FEV_{6}, with age and height as predictor variables 13. In the study by Hankinson et al. 13, the analysis was performed separately for each sex and each ethnic group, and the LLN was estimated as predicted minus 1.645 times the standard error of the estimate, which corresponds to the fifth percentile.
While fixed cutoff values are more widely used, in an attempt to simplify the diagnosis, there is potential for misclassification. For example, in elderly subjects, where the agerelated decline in FEV_{1}/FVC and FEV_{1}/FEV_{6} may cause a significant overdiagnosis of airway obstruction, pointed out by Hardie et al. 18. This is illustrated here by table 5⇓ where LLN values, derived from the NHANES III reference equations, are depicted for discrete ages between 20–80 yrs. Clearly, fixed cutoffs of FEV_{1}/FVC <70% and FEV_{1}/FEV_{6} <73% are most suitable for use with middleaged subjects. Table 5⇓ also shows that cutoffs could be 7% (for FEV_{1}/FVC) or 5% (for FEV_{1}/FEV_{6}) higher in 20yrold females and 8% (for FEV_{1}/FVC) or 6% (for FEV_{1}/FEV_{6}) lower in 80yrold males. Finally, it must be considered that, regardless of the method used to define abnormality, measured values that lie close to the threshold should be interpreted with caution, due to several sources of variability: 1) diurnal and daytoday variations of spirometric indices 11; 2) betweenmanoeuvre repeatability criteria that allow for a difference up to a maximum of 0.150 L between the two largest values of both FEV_{1} and FVC 5; 3) patients with obstruction having coefficients of variation for FEV_{1} and FVC that are approximately twice those of normal subjects 19.
Spirometric detection of restriction
The present study shows that the commonly used cutoff value, FVC <80% predicted, could be replaced by FEV_{6} <82% pred, especially for excluding a restrictive ventilatory defect. However, a restrictive abnormality is characterised by a reduced total lung capacity, whereas a reduced FVC in the presence of a normal FEV_{1}/FVC can only be used to suggest, but not to diagnose, the presence of a restrictive abnormality 11. A study by Swanney et al. 8 showed that spirometrybased algorithms could not reliably predict a reduced total lung capacity, but are very useful at excluding a restrictive defect. They also demonstrated that FEV_{6} was equivalent to FVC when using LLN calculated with the NHANES III reference equations. In the current study high, negative predictive values were found when comparing FEV_{6} and FVC as a predictor of a restrictive pattern (table 3⇑). This makes the use of FEV_{6} suitable for the exclusion of restriction.
However, fixed percentages of the predicted value should be used with caution. Tables 6⇓ and 7⇓ show the age and height dependency of the LLN for FVC and FEV_{6}, using the NHANES III reference equations, for males and females. Using fixed cutoffs had a tendency towards overdiagnosis of a restrictive pattern in elderly people.
CONCLUSION
This study demonstrates that forced expiratory volume in one second/forced expiratory volume in six seconds <73% and forced expiratory volume in six seconds <82% pred can be used as a valid alternative for forced expiratory volume in one second/forced vital capacity <70% and forced vital capacity <80% pred as fixed cutoff points for the detection of an obstructive or restrictive spirometric pattern in adults. The authors emphasise that the fixed cutoff terms should be used with caution, particularly outside the middleaged population. Ideally, abnormalities in spirometric values should be defined using lower limits of normal, derived from a representative sample of healthy subjects.
Footnotes

For editorial comments see page 245.
 Received March 25, 2005.
 Accepted September 8, 2005.
 © ERS Journals Ltd