Octavian C. Ioachimescu, MD; Saiprakash B. Venkateshiah, MD; Mani S. Kavuru, MD; Kevin McCarthy, RCPT; and James K.

Transcription

1 Estimating FVC From FEV 2 and FEV 3 * Assessment of a Surrogate Spirometric Parameter Octavian C. Ioachimescu, MD; Saiprakash B. Venkateshiah, MD; Mani S. Kavuru, MD; Kevin McCarthy, RCPT; and James K. Stoller, MD, MS Background: In the context that accurate measurement of FVC is important in diagnosing airflow obstruction and in assessing restriction, strategies to achieve reliable and accurate FVC measurement have drawn much attention. Objectives: Because the rate of achieving end-of-test criteria during spirometry has been shown to be low, with resultant underrecording of the FVC, the current study proposes a regression equation for estimating FVC from measured values of the forced expiratory volume in 2 s (FEV 2 ) and forced expiratory volume in 3 s (FEV 3 ). Methods: The predictive equation for the estimated FVC from volume measurements within the first 3sofexhalation (estimated FVC 3 ) was generated based on 330 consecutive acceptable spirograms performed in the Cleveland Clinic Foundation Pulmonary Function Laboratory. The equation was applied to an independent validation set comprised of spirometry measurements on 370 different consecutive patients. Results: In the validation spirometry sample, in which the prevalence of obstruction was 34% (based on values of the measured FEV 1 /FVC compared to National Health and Nutrition Examination Survey III values), the sensitivity, specificity, and positive and negative predictive values of FEV 1 /estimated FVC 3 for obstruction were 93.8%, 89.1%, 81.2%, and 96.9%, respectively. The misclassification rate was 9.2%. In the same cohort, the mean difference ( SD) between estimated FVC 3 and measured FVC was ml. Conclusions: Given that FVC is frequently underrecorded, with resultant overestimation of FEV 1 /FVC and underdiagnosis of airflow obstruction, we believe that estimating FVC from FEV 2 and FEV 3 can offer practical diagnostic advantages. (CHEST 2005; 128: ) Key words: airflow obstruction; forced expiratory volume in 3 s; FVC; spirometry Abbreviations: CI confidence interval; estimated FVC 3 estimated FVC from volume measurements within the first3sofexhalation; FET forced expiratory time; FEV 2 forced expiratory volume in 2 s; FEV 3 forced expiratory volume in 3 s; FEV 6 forced expiratory volume in 6 s; LLN lower limit of normal; NHANES National Health and Nutrition Examination Survey; NPV negative predictive value; PPV positive predictive value; RMSE residual mean square error In the context that spirometry is essential in assessing respiratory disorders, FVC is an important measurement to diagnose the presence of obstructive lung disease and to evaluate restrictive disease. At the same time, the accurate measurement of FVC poses challenges to ensure that end-of-test criteria *From the Department of Pulmonary, Allergy, and Critical Care Medicine, The Cleveland Clinic Foundation, Cleveland, OH. Drs. Ioachimescu and Venkateshiah contributed equally to this article. Manuscript received November 25, 2004; revision accepted February 10, Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( org/misc/reprints.shtml). Correspondence to: Octavian C. Ioachimescu, MD, Department of Pulmonary, Allergy, and Critical Care Medicine, The Cleveland Clinic Foundation, 9500 Euclid Ave A90, Cleveland, OH 44195; oioac@yahoo.com are met in order to avoid underestimation of FVC, which occurs commonly, as shown by Eaton et al, 1 who demonstrated that among 2,928 expiratory efforts performed in primary care practices, only 28% satisfied an expiratory time 6 s and that 47% were 4 s in duration. Because underestimating FVC can cause the FEV 1 /FVC ratio to be overestimated with resultant underrecognition of airflow obstruction, much attention has been given to assessing spirometric measures that can serve as accurate surrogates for FVC, eg, the forced expiratory volume in 6 s (FEV 6 ). 2 Extending available analyses of FEV 6, 2,3 the current research addresses a strategy for estimating FVC using expiratory efforts of shorter duration, ie, after 3 s. Specifically, the current research addresses the following two questions using two separate cohorts, the developmental and validation 1274 Clinical Investigations

2 groups: how well does the forced expiratory volume in 3 s (FEV 3 ) predict the measured FVC, and what is the diagnostic performance of the FEV 1 /FVC ratio using FVC estimates from the FEV 3 in detecting airflow obstruction? Some of the results of this research have been previously reported in abstract form. 4 Materials and Methods This study examines the diagnostic performance of estimated FVC from volume measurements within the first 3 s of exhalation (estimated FVC 3 ). To develop the prediction equations for estimated FVC 3, data from serial spirometry tests performed on 400 patients tested in the Cleveland Clinic Foundation Pulmonary Function Laboratory between June 6, 2000, and June 19, 2000, were analyzed. The prediction equations and the diagnostic performance of FEV 1 /estimated FVC 3 in detecting airflow obstruction were then assessed in a separate, independent sample of 377 spirograms performed in the same laboratory between June 3, 2002, and June 14, Criteria for acceptability and reproducibility of spirometry measurements were those specified by the American Thoracic Society. 5 As previously described, 6 spirometry (Jaeger Master Lab Pro; Erich Jaeger GmbH; Wurzburg, Germany) was performed using a modified expiratory technique in which the subject was asked to perform a maximally forceful exhalation for 3 s, and thereafter to continue to exhale slowly (ie, in a relaxed fashion) as long as possible. Height was measured to the nearest centimeter without shoes, and weight was recorded to the nearest kilogram. Each spirometer was calibrated daily with a 3-L syringe. Forced expiratory volume in 2 s (FEV 2 ), FEV 3, and all other standard spirometric indexes were measured using software (Erich Jaeger GmbH). To determine the regression equation for estimated FVC 3, several independent variables were examined. Candidate variables were FEV 1, FEV 1 percentage of predicted, FEV 2, FEV 3, and the difference between the volume exhaled in 3 s and 2 s (FEV 3 FEV 2 ). The difference between estimated FVC 3 obtained from the regression equation and measured FVC was calculated. Also, the FEV 1 /estimated FVC 3 ratio was compared to the lower limit of the FEV 1 /FVC ratio from the National Health and Nutrition Examination Survey III (NHANES III) reference set. 3 In keeping with current guidelines, 7 FEV 1 /FVC was used to diagnose airflow obstruction, which was deemed to be present if FEV 1 /FVC ratio fell below the lower limit of normal (LLN) in the NHANES III reference set. 3 Based on Global Initiative for Chronic Obstructive Lung Disease staging criteria 7 (in which the diagnosis of airflow obstruction is based on values of FEV 1 /FVC 0.70), the severity of airflow obstruction was graded by FEV 1 into three categories: mild ( 80% predicted), moderate (30 to 80% predicted), and severe ( 30% predicted). Statistical analysis was performed using statistical software (SigmaStat 2.03; SPSS; Chicago, IL; and JMP 5.01; SAS Institute; Cary, NC). Main statistical tests performed were 2 analysis and Kruskal-Wallis analysis of variance. To assess agreement between FVC and estimated FVC 3, a discriminative analysis was performed using the Bland and Altman method. 8 Values of p 0.05 were considered statistically significant. Results For developing the regression equation, 400 spirometry studies were collected, of which 70 records (17.5%) were excluded (Table 1). Reasons for excluding spirograms included the following: exhalation time 6s(n 55), no mention of race (n 9), females 18 years old (n 3), and males 20 years old (n 3). Of the 330 subjects whose spirograms were used, 284 were white and 46 were black. One hundred seventy-six were male, and 154 were female. One hundred forty patients (42.4%) had obstruction (mild [n 14], moderate [n 90], and severe [n 36]). Eighty-two patients (24.8%) had restriction, none (0%) had both obstruction and restriction, and 108 patients (32.7%) were normal. Mean expiratory time in these 330 spirograms was 10.3 s (95% confidence interval [CI], 9.9 to 10.6 s; SD, 3.2 s). Using these 330 spirometry measurements in the developmental data set, after considering different models, the best-fit model was a linear regression equation for predicting FVC: estimated FVC FEV FEV 3 FEV 2 liters where R , RMSE (residual mean square error) 0.25, p (Fig 1). The mean difference between estimated FVC 3 and FVC was 0.6 ml (95% CI, 16.1 to 16.0 ml; SD, 148 ml). Values of estimated FVC 3 were within 200 ml of the measured FVC in 201 of 330 tests (60%). Table 2 presents demographic and spirometric characteristics of the 330 subjects whose spirometry measurements comprise the developmental data set, stratified by those whose FVC and estimated FVC 3 were within 200 ml, those whose estimated FVC 3 exceeded the measured FVC by 200 ml, and those whose measured FVC exceeded the estimated FVC 3 by 200 ml. This last group had longer forced expiratory time (FETs) and lower FEV 1 /FVC compared to the other two groups. Table 3 presents the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of FEV 1 /estimated FVC 3 in detecting Table 1 Reasons for Excluding Spirograms From the Developmental Data Set Reason for Exclusion No. Short exhalation time (FET 6 s) and no expiratory 32 plateau Abrupt termination of expiration (glottic closure) 7 Cough in first second or hesitant start 6 No race specified (unable to calculate predicted value) 9 Pediatric patient 6 Short exhalation time (FET 6 s) but valid plateau 10 (analyzed separately) Total 70 CHEST / 128 / 3/ SEPTEMBER,

3 Figure 1. Comparison of actual FVC with estimated FVC 3 in the developmental data set (n 330). airflow obstruction, as defined by measured values of FEV 1 /FVC according to NHANES III criteria. The PPV of estimated FVC 3 for identifying obstruction was 82.7%, and the NPV for excluding obstruction was 96.4%. The likelihood ratio for detecting obstruction was 6.5. The misclassification rate was 10.3%. To assess estimated FVC 3 under conditions of short expiratory times, we evaluated the 10 spirograms that had a short exhalation (FET 6 s) but a valid expiratory plateau. In 60% (6 of 10 spirograms), estimated FVC 3 value was within 200 ml of the actual FVC. None of these 10 patients had obstruction. In the context of excluding from both development and validation groups the tests with short expiratory times ( 6 s), we do not know if the above prediction equation applies to spirograms with short exhalation, even when a visually apparent expiratory plateau is reached earlier than 6 s. To validate the diagnostic utility of estimated FVC 3, the values derived from the regression equation in the developmental set were compared to actual FVC values in a separate validation cohort of 377 subjects, who performed technically acceptable spirograms obtained in the Cleveland Clinic Foundation Pulmonary Function Laboratory between June 3, 2002, and June 14, As shown in Figure 2, the mean difference between estimated FVC 3 and actual FVC was 24.7 (95% CI, 0.7 to 48.8 ml; SD, 237 ml). The estimated FVC 3 explained 94.7% of the variance in measured FVC (p 0.001). Mean expiratory time in these 377 spirograms was 10.6 s (95% CI, 10.3 to 11.0 s; range, 6.0 to 18.6 s; median, 10.1 s). In our pulmonary function laboratory, which uses the modified exhalation technique almost exclusively, the mean expiratory time calculated on 4,413 patients tested during the same period 2 years before the study (2000) was 11.8 s. For patients with obstruction (n 1,826, 41% of the spirograms), the mean expiratory time was 13.5 s. The expiratory time distribution by FEV 1 /FVC ratio was as follows: 11.8 s (70 to 80%), 13.5 s (30 to 69%), and 15 s ( 29%). In evaluating the diagnostic performance of the estimated FVC 3 /FVC for detecting obstruction, PPV and NPV were 81% and 96%, respectively (Table 4), quite similar to the values in the developmental cohort (Table 3). Bland and Altman analysis of data in the validation group (Fig 3), plotting the difference between estimated FVC 3 and FVC against their average, showed a 95% CI of the mean of (0.7 to 48.8 ml); the SD of the difference was 237 ml. Finally, the likelihood ratio for diagnosing obstruction using the FEV 1 /estimated FVC 3 in the validation cohort was 8.6, and the misclassification rate was 9.2% (Table 4) Clinical Investigations

4 Table 2 Demographic and Spirometric Characteristics of the Developmental Cohort Stratified by Degree of Agreement Between the Actual FVC and the Estimated FVC 3 * Characteristics FVC and Estimated FVC 3 Within 0.2 L Estimated FVC 3 Greater Than FVC by 0.2 L Estimated FVC 3 Less Than FVC by 0.2 L All p 0.05 (Groups) Patients, No Female/male gender 96/105 48/25 10/46 154/176 1 vs 2, 2 vs 3, 1 vs 3 Black/white race 28/173 12/61 4/52 46/284 NS Age, yr NS Height, cm vs2,2vs3,1vs3 Weight, kg NS BMI vs3 FVC, L vs 2, 2 vs 3 FEV 1,L vs2,1vs3 FEV 1 /FVC, % vs 3, 2 vs 3 FET, s vs3,2vs3 FEV 2,L vs 2, 1 vs 3 FEV 3,L vs 2, 1 vs 3 FVC, % vs 2, 2 vs 3 FEV 1,% vs 2, 1 vs 3 FEV 3 FEV 2,L vs 2 Estimated FVC 3,L vs 2 FEV 1 /estimated FVC 3, (%) vs 2, 1 vs 3 FVC estimated FVC 3,L vs 2, 2 vs 3, 1 vs 3 *Data are presented as mean SD or No. NS not significant. To further assess the estimated FVC 3, we evaluated the correlations with actual FVC and detection of obstruction in the subset of validation set spirograms that achieved (34%) vs those that did not achieve an expiratory plateau (66%). The strength of the correlations, the variances, and the statistical significance of the predictive equations were similar in both subsets (ie, in those achieving an expiratory plateau, R , RMSE 0.217, p ; and in those not achieving an expiratory plateau, R , RMSE 0.238, p ). The sensitivity, specificity, PPV, and NPV were similar in the group of spirograms with expiratory plateau and in the group without a valid expiratory plateau, ie, 92%, 84%, 60%, 97%, and 94%, 91%, 88%, 95%, respectively. Also, dichotomizing the spirograms in the validation set by those in which estimated FVC 3 underestimated measured FVC vs those in which the estimated FVC 3 overestimated the measured FVC group, we observed that the predictive equation produced the following sensitivities, specificities, PPVs, and NPVs: 100%, 92%, 88%, 100% and 68%, 100%, 100%, 81%, respectively. The statistic degree of agreement in detecting obstruction varied between 0.79 and 0.89 among different subgroups (plateau vs nonplateau, underestimated vs overestimated FVC). Discussion The main findings of this study are as follows: (1) FEV 3 can be used to estimate FVC with reasonable accuracy. Although this surrogate parameter confers benefits for interpreting spirometry results when spirometric end-of-test criteria are not satisfied, Table 3 Comparison of FEV 1 /Estimated FVC 3 With FEV 1 /FVC in Patients With or Without Airflow Obstruction in the Developmental Cohort* FEV 1 /FVC Variables Obstruction (Value Less Than LLN) No Obstruction (Value Greater Than LLN) Total FEV 1 /estimated FVC 3 indicates obstruction FEV 1 /estimated FVC 3 indicates no obstruction Total *Sensitivity 95.7% (95% CI, 91.4 to 98.2%); specificity 85.3% (95% CI, 82.1 to 87.1%); PPV 82.7% (95% CI, 79.0 to 84.8%); NPV 96.4% (95% CI, 92.9 to 98.5%); misclassification rate 34/330 (28 6/330) 10.3%. LLN in NHANES III data set. CHEST / 128 / 3/ SEPTEMBER,

5 Figure 2. Comparison of actual FVC with estimated FVC 3 in the validation data set (n 377). estimates of the true FVC are imperfect and use of the FEV 1 /estimated FVC 3 does not eliminate the risk of misclassifying the presence of obstruction. (2) Compared with the actual ratio of FEV 1 /FVC, the ratio of FEV 1 /estimated FVC 3 demonstrated reasonably good, albeit imperfect, diagnostic performance in detecting airflow obstruction and excellent performance in excluding obstruction. For example, in the validation data set, the PPV of FEV 1 /estimated FVC 3 for detecting obstruction was 81.2% and the NPV was 96.9%. (3) Although use of a surrogate parameter has appeal, its imperfect diagnostic performance should not detract from efforts to optimize performance of spirometry, eg, through technician training and active quality control of testing. Based on the acceptable diagnostic performance of estimated FVC 3 and the fact that many patients do not achieve even 6 s of exhalation, our findings suggest that the estimated FVC 3 derived from measured values of FEV 2 and FEV 3 can be a useful spirometric parameter. That failure to achieve spirometric end-of-test criteria or even 6 s of exhalation is common has been shown by Eaton et al. 1 Recognizing that equipment features may have contributed to end-of-test failures in this study, among 2,928 expiratory measures only 28% achieved an expiratory time of 6 s. The remaining 72% of the blows (n 2,103) lasted 6s, including 47% (n 1,380) that were 4 s in duration. Visual inspection of the spirograms showed that Table 4 Comparison of FEV 1 /Estimated FVC 3 With FEV 1 /FVC in Patients With or Without Airflow Obstruction in the Validation Cohort* FEV 1 /FVC Variables Obstruction (Value Less Than LLN) No Obstruction (Value Greater Than LLN) Total FEV 1 /estimated FVC 3 indicates obstruction FEV 1 /estimated FVC 3 indicates no obstruction Total *Sensitivity 93.8% (95% CI, 91.4 to98.2%); specificity 89.1% (95% CI, 82.1 to 87.1%); PPV 81.2% (95% CI, 79.0 to 84.8%); NPV 96.9% (95% CI, 92.9 to 98.5%); misclassification rate 35/377 (27 8/377) 9.2%. LLN in NHANES III data set Clinical Investigations

6 Figure 3. Bland and Altman plot of estimated FVC 3 and actual FVC. The middle line represents the mean, while the other lines represent 1, 2, 1, and 2 SDs of the differences (R , p ). The analysis, which is a conventional method for evaluating new measurement techniques, plots the dispersion of individual differences between the measurements against their means. Mean difference in liters, ; SE, (upper 95%, ; lower 95%, ); SD, (upper 95%, 0.255; lower 95%, 0.221); N 377; R % of the curves that failed to achieve 6sof duration had expiratory plateaus, suggesting that most of the test failures were due to early termination of effort. As shown by Townsend et al, 9 early termination of expiration can confound the interpretation of the spirogram. Specifically, short durations of expiration cause undermeasurement of FVC, thereby overestimating values of FEV 1 /FVC. The diagnostic impact of overestimating FEV 1 /FVC by underrecording FVC was such that after 3 s of expiration, 60 of the 80 subjects (75%) were classified with FEV 1 /FVC ratios 80%, and 2 subjects (3%) had FEV 1 /FVC ratios 69%. The distribution of FEV 1 /FVC values was a distortion of the FEV 1 /FVC distribution that was based on complete expirations, in which only 18 of the subjects (23%) had FEV 1 /FVC ratios 80% and 21 subjects (26%) had FEV 1 /FVC ratios 69%. Therefore, the proportions of individuals with normal and abnormal FEV 1 /FVC values varied dramatically with the length of expiration recorded, with fewer abnormal FEV 1 /FVC values produced after a short-recorded expiration. Most of the subjects with low FEV 1 /FVC were misclassified after only 3sof expiration, and many were misclassified even after 6 s of expiration. In another study, Graham et al 10 showed a risk of reaching spurious spirometric conclusions based on underrecorded expirations in finding that FVC in Vermont granite workers appeared to increase by a mean of 0.54 L over a 5-year period. Evaluation of the spirograms showed that these observed increases in FVC were probably caused by comparing underrecorded baseline FVC values (ie, measured after 2.5 to 3 s of expiration) to FVC values based on longer expirations (exceeding 6 s of expiratory effort) recorded on follow-up tests performed 5 years later. This observation underscores the importance of assessing the FET in evaluating and comparing values of FVC. Based on these pitfalls of underrecorded FVC values, the National Lung Health Education Program and others 2,11 have recommended use of FEV 6 as a surrogate for FVC. In the context that even 6 s of exhalation are often not achieved, an approach of estimating FVC based on measures values of FEV 2 and FEV 3 has appeal. At the same time, satisfying end-of-test criteria and ensuring that the value of FVC is reliable would obviate the need for estimating the FVC (as with equation proposed here) or using a surrogate parameter like the FEV 6. In the context that failure to achieve acceptable CHEST / 128 / 3/ SEPTEMBER,

7 spirometry efforts is common, it is imperative that pulmonary function laboratories undertake both careful quality review and technician training. Still, although it is tempting to propose enhanced technician training as a solution to optimize spirometry performance, several studies suggest that training, while useful, does not completely ensure consistent achievement of end-of-test criteria. For example, despite offering a spirometry workshop to technicians at the start of the study by Eaton et al, 1 only 39% of the 923 expiratory efforts achieved expiratory times 6 s. Of the tests performed by technicians who received no formal spirometry workshop training, only 16% of 1,285 expiratory efforts achieved expiratory times of 6s. 1 Others have reported failure to achieve American Thoracic Society criteria for acceptability and reproducibility in 8 to 20% of spirometry test sessions. More extensive training and electronic spirometer feedback as in the Lung Health Study 12 can improve the rate but still failed to completely ensure uniform achievement of full expiratory times. Specifically, 2.1% of test sessions in the Lung Health Study 12 did not meet American Thoracic Society acceptability and reproducibility criteria. Finally, in the National Heart, Lung, and Blood Institute Registry of Individuals With Severe Deficiency of 1 Antitrypsin, 18 in which technicians underwent extensive initial group training, received regular feedback on their technical performance in spirometry, and experienced site visits for biological control testing, the rate of failing to achieve end-oftest criteria was still 69% even though spirometry reproducibility was quite high. 18 Clearly, though extensive training has been shown to improve technical performance in spirometry, it is impractical to propose widespread spirometric training of the sort described in these highly quality-controlled studies, especially in the face of official calls for widespread office-based spirometry (as by the National Lung Health Education Program). 11 In this regard, estimating FVC from values of FEV 2 and FEV 3, which can be achieved more easily than FVC, also has appeal. It bears mention that a tight quality control review process remains a critical component of good spirometry performance, whether using standard parameters or the FVC estimates discussed here. Indeed, only by such ongoing review can attention to all the methodologic requirements of good spirometry (eg, adequate start and end of test, complete inhalation) be optimized. As we imagine extending use of the estimated FVC 3 to more widespread spirometric practice, a favorable feature is that several large manufacturers of spirometers already report both FEV 2 and FEV 3 (eg, Koko Spirometer; Pulmonary Data Services; Louisville, CO; V max Spectra; SensorMedics; Yorba Linda, CA; Erich Jaeger spirometers; Wurzburg, Germany), making calculation of estimated FVC 3 readily possible. Many other manufacturers currently report the FEV 3 and could likely easily add reporting of FEV 2 if warranted by market demands. With these observations in mind, estimated FVC 3 may be preferable to FEV 6, which is not currently available as a readout on many commercial spirometers. Several shortcomings of this study warrant comment. First, the predictive values of FEV 1 /estimated FVC 3 for airflow obstruction expectedly depend on the prevalence of airflow obstruction in the population studied. In the two study cohorts considered here, airflow obstruction was evident in 34 to 42% of the spirograms. Were airflow obstruction more prevalent, the PPV of FEV 1 /estimated FVC 3 would have been higher than the value of 81% observed in the validation set and the NPV would be expected to be lower. Second, because our regression equation was derived from spirograms obtained using the modified technique that is routinely used in our pulmonary function laboratory (in which the patient is instructed to relax and push gently after the first 3sof maximally forced expiration), whether these results will be robust for spirometric data generated with a traditional technique with sustained forced expiration is uncertain. Because our earlier randomized controlled trial of this modified spirometric technique showed that its use allowed more frequent accomplishment of end-of-test criteria and higher values of FVC than were obtained using the traditional technique, 6 validation of the regression equation for spirometric data obtained with this traditional technique will be needed for those laboratories still using that approach. Third, because our results were derived from a population of adults, their application to pediatric populations is unproven. And fourth, the fact that the calculated FVC could be lower than measured FVC, even if an expiratory plateau was achieved, highlights a limitation of the estimation. Our analysis failed to identify any specific demographic features or spirometric parameters that predict underestimation of the measured FVC by estimated FVC 3. In the context that technical errors in performing spirometry are unlikely to underestimate the FVC, we suggest several conditions that should be satisfied in order to optimally use the estimated FVC 3. Specifically, if the value of estimated FVC 3 is lower than the measured FVC, the measured FVC should be used. Similarly, because our validation set contained no spirograms in which the FET was 6s(ie, range of FETs was 6 to 18.6 s), validation of estimated FVC Clinical Investigations

8 in a population where FET is 6 s is needed before use of the predictive equation can be endorsed. In summary, the current report presents a regression equation by which FVC can be estimated from measured values of FEV 2 and FEV 3 and then validates the regression equation in an independent, validation cohort. Of note, our predictive equation applies to spirograms with expiratory time of at least 6 s not achieving an expiratory plateau. In the context that the FVC is frequently underrecorded with resultant overestimation of FEV 1 /FVC ratio and underdiagnosis of airflow obstruction, we believe that estimating FVC from FEV 2 and FEV 3 may offer diagnostic and practical advantages, especially when complete expiration cannot be achieved. We encourage further investigation of this estimation method in order to better assess its generalizability and application to specific patient groups (eg, pediatric populations, and those tested with conventional spirometric techniques). References 1 Eaton T, Withy S, Garrett JE, et al. Spirometry in primary care practice: the importance of quality assurance and the impact of spirometry workshops. Chest 1999; 116: Swanney MP, Jensen RL, Crichton DA, et al. FEV(6) is an acceptable surrogate for FVC in the spirometric diagnosis of airway obstruction and restriction. Am J Respir Crit Care Med 2000; 162: Hankinson JL, Odencrantz JR, Fedan KB. Spirometric reference values from a sample of the general U.S. population. Am J Respir Crit Care Med 1999; 159: Venkateshiah SB, McCarthy K, Kavuru MS, et al. Identifying obstruction using a three-second exhalation by extrapolating the forced vital capacity (FVC). Chest 2003; 124(Suppl):122S 5 American Thoracic Society. Standardization of spirometry, 1994 update. Am J Respir Crit Care Med 1995; 152: Stoller JK, Basheda S, Laskowski D, et al. Trial of standard versus modified expiration to achieve end-of-test spirometry criteria. Am Rev Respir Dis 1993; 148: Global Initiative for Chronic Obstructive Lung Disease (GOLD). Guidelines, global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: NHLBI/WHO Workshop report. Bethesda, MD: National Heart, Lung and Blood Institute, National Institute of Health, Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: Townsend MC, Du Chene AG, Fallat RJ. The effects of underrecorded forced expirations on spirometric lung function indexes. Am Rev Respir Dis 1982; 126: Graham WG, O Grady RV, Dubuc B. Pulmonary function loss in Vermont granite workers: a long-term follow-up and critical reappraisal. Am Rev Respir Dis 1981; 123: Ferguson GT, Enright PL, Buist AS, et al. Office spirometry for lung health assessment in adults: a consensus statement from the National Lung Health Education Program. Chest 2000; 117: Enright PL, Johnson LR, Connett JE, et al. Spirometry in the Lung Health Study: 1. Methods and quality control. Am Rev Respir Dis 1991; 143: Eisen EA, Wegman DH, Louis TA. Effects of selection in a prospective study of forced expiratory volume in Vermont granite workers. Am Rev Respir Dis 1983; 128: Eisen EA, Oliver LC, Christiani DC, et al. Effects of spirometry standards in two occupational cohorts. Am Rev Respir Dis 1985; 132: Kellie SE, Attfield MD, Hankinson JL, et al. Spirometry variability criteria association with respiratory morbidity and mortality in a cohort of coal miners. Am J Epidemiol 1987; 125: Kanner RE, Schenker MB, Munoz A, et al. Spirometry in children: methodology for obtaining optimal results for clinical and epidemiologic studies. Am Rev Respir Dis 1983; 127: Eisen EA, Dockery DW, Speizer FE, et al. The association between health status and the performance of excessively variable spirometry tests in a population-based study in six U.S. cities. Am Rev Respir Dis 1987; 136: Stoller JK, Buist AS, Burrows B, et al. Quality control of spirometry testing in the registry for patients with severe 1 -antitrypsin deficiency: Alpha 1-Antitrypsin Deficiency Registry Study Group. Chest 1997; 111: CHEST / 128 / 3/ SEPTEMBER,