Secondary Logo

Journal Logo

Predicted lower limit of normal reduces misclassification risk of airflow limitation in asymptomatic elderly never-smokers

WANG, Yang; XIAO, Wei; MA, De-dong; JIANG, Yuan-yuan

doi: 10.3760/cma.j.issn.0366-6999.20130801
Original article
Free

Background It remains controversial what better defines abnormal lung function at the extremes of age. This study aimed to establish new spirometric reference equations for residents in Jinan and compare the most accepted two spirometric definitions of chronic obstructive pulmonary disease (COPD)—a fixed post-bronchodilator ratio of forced expiratory volume in one second to forced vital capacity (FEV1/FVC) below 70% vs. FEV1/FVC <lower limit of normal (LLN)—in an elderly population (age ≥60 years).

Methods Among 10 002 cases, only never-smokers who had no respiratory symptoms or diagnoses of cardiopulmonary diseases and whose lung function measurements strictly followed the American Thoracic Society guidelines, were included in the statistical analysis. For each gender, prediction equations of FEV1, FVC, FEV1 /FVC and corresponding LLN values were developed using multiple regression models.

Results In this cross-sectional study, spirometric reference values were statistically derived from a sample of 1328 healthy never-smokers (672 males, 656 females), age 56–84 years. As expected, age and height were the main predictors for all lung function parameters. Using the two spirometric definitions, the number of pulmonary function test results interpreted as obstructed varied markedly in the elderly population for both genders (males: χ2=19.59, P < 0.01; females: χ2=5.67, P=0.017).

Conclusions The Global Initiative for Chronic Obstructive Lung Disease (GOLD) criterion probably leads to over-diagnosis in asymptomatic elderly never-smokers. Our study suggests that the predicted LLN values will reduce the misclassification risk instead. However, cohort studies are urgently needed for clinical validation.

Department of Respiratory Medicine, Qilu Hospital, Shandong University, Jinan, Shandong 250012, China (Wang Y, Xiao W, Ma DD and Jiang YY)

Correspondence to: Prof. XIAO Wei, Department of Respiratory Medicine, Qilu Hospital, Shandong University, Jinan,

Shandong 250012, China (Tel & Fax: 86–531–82169506. Email: xiaowei4226@163.com)

(Received March 20, 2013)

Edited by WANG Mou-yue and LIU Huan

As the best standardized, most reproducible, and most objective measurement of airflow limitation available, spirometry remains the gold standard for diagnosing chronic obstructive pulmonary disease (COPD) and monitoring its progression.1 The Global Initiative for Chronic Obstructive Lung Disease (GOLD) has defined incompletely irreversible airflow limitation as a post-bronchodilator ratio of forced expiratory volume in one second to forced vital capacity (FEV1/FVC) below 70%, which is widely used in clinical practice. Regardless of age and gender, this criterion is accepted for the purpose of simplicity and ease of remembering. However, because lung volumes demonstrate marked negative age dependency, the fixed ratio may lead to over-diagnosis in the elderly population (age ≥60 years)2,3 and under-diagnosis in adults younger than 45 years, especially with mild disease.FEV1/FVC <70% as a fixed cut-off: a longitudinal evaluation of clinical and functional outcomes. Thorax 2008; 63: 1040-1045.','400');" onMouseOut="javascript:ImageWrapperControl_ImageMouseOut();">4 Therefore, to minimize the potential misclassification, GOLD (Updated 2010) recommended the use of lower limit of normal (LLN) values for FEV1/FVC as the spirometric cut-off.1

The LLN values are calculated using reference equations, which are statistically derived from representative samples of a “normal” population and generally classify the bottom 5% of them as abnormal. In recent years, numerous reference equations have been published throughout the world. However, most of them were derived from Caucasians and failed to update over time. As is known, there has been increasing evidence that ethnicity and cohort effects do affect lung volumes.5–7 Hence, few reference equations could be applied to mainland Chinese, leading to the potential risk of misinterpretation of pulmonary function test (PFT) results at the extremes of the age range. Specific reference equations for mainland Chinese are urgently needed to meet the demand of both clinical practice and epidemiological studies.

Our group conducted a large-scale, population-based epidemiological study on COPD prevalence of Jinan citizens in 2009. We found that 38 cases of asymptomatic elderly never-smokers were classified as COPD patients using the 70% fixed ratio method. Comparing to the LLN method, however, nearly half of them (15/38) could be possibly over-diagnosed.8 Unfortunately, considering the limited number of cases observed and that the selected reference equations were derived from Caucasian subjects (Hankinson data set6) decades ago, this conclusion could be for reference only. Overall, this study aimed to establish new spirometric reference equations for residents in Jinan and compare the most accepted two spirometric definitions of COPD—post-bronchodilator FEV1/FVC <a fixed ratio of 70% vs. FEV1/FVC <LLN—in the elderly population (age ≥60 years).

Back to Top | Article Outline

METHODS

Study design

This study was conducted in accordance with the amended Declaration of Helsinki. As a retrospective cohort study, only medical data of patients which involved no identifying information were used and the protocol has been approved by the independent ethics committee (Medical Ethics Committee of Qilu Hospital, Shandong University; Approval No. 1165).

A total of 10 002 subjects volunteered or were advised to perform spirometry in Qilu Hospital, one of China's top tertiary medical centers, between January 2010 and December 2011. All the subjects involved were mainland Chinese residing in Jinan, China. Reference values of lung function (LF) parameters were derived from persons with no known previous or present conditions that adversely affect their ventilatory function, the so-called “normal” population. Prediction equations for mean forced expiratory volume in one second (FEV1), forced vital capacity (FVC) and FEV1/FVC were established using multiple regression analysis. For both males and females, age, standing height, weight and body mass index (BMI) were all considered as alternative independent variables in the model.

Back to Top | Article Outline

Methods

Before the spirometry was performed, a detailed application form was filled in by the physicians, with the following items as mandatory fields: general data, medical history, physical examination, lab test results available, and their impression. Both electrocardiogram and chest imaging (X-ray or computed tomography scanning) were required, and then ultrasonic cardiogram was only performed when needed. All the spirometric measurements were done using a computerized spirometer (Masterscreen, Jaeger, North Rhine-Westphalia, Germany). Each copy of PFT results was carefully reviewed for further use.

In this study, only never-smokers who had no respiratory symptoms or diagnosed cardiopulmonary diseases and whose lung function measurement strictly followed the American Thoracic Society (ATS) guidelines were included in the statistical analysis. According to the recommendations by the European Respiratory Society, Hankinson and Johannessen et al,5,6,9 subjects were excluded only if they had one of the criteria presented in Table 1. It is worth noting that never-smokers are those who have smoked no more than 20 packs of cigarette in their lifetime.10

Table 1

Table 1

Despite being a retrospective cohort study, quality control was strongly emphasized. The lung function system was calibrated twice a day. To meet the ATS acceptability and reproducibility criteria11 each subject was instructed by an experienced technician to complete a minimum of three FVC maneuvers. Three acceptable tracings were automatically saved, with the largest FEV1 and FVC reproducible within 200 ml of each other. Two senior technicians, also qualified as clinical physicians, separately selected the best curve and analyzed the data. The spirometric measurements were eliminated if the supervisors were in disagreement. All spirometry data of subjects included in the statistical analysis were entered into the standardized Excel Database.

Subjects who had normal diffusion capacity, no respiratory symptoms or diagnosed cardiopulmonary diseases were defined as asymptomatic COPD patients if there proved to be airflow limitation (FEV1/FVC <LLN or 70%) present. The difference of COPD diagnosis in the elderly population between the LLN method and the 70% fixed ratio method was tested. Predicted spirometric values calculated from the current study and those from previous studies derived from either Caucasian2,6,12 or Chinese13,14 subjects were also compared.

Back to Top | Article Outline

Statistical analysis

Statistical analyses were performed using SPSS software (SPSS for Windows, Version 17.0; SPSS; Chicago, IL, USA) and a P value of 0.05 was deemed statistically significant. Prediction equations for mean FEV1, FVC and FEV1/FVC were established using multiple regression analysis. Assumptions on normality of distribution were tested using the one-sample Shapiro-Wilk test. To improve the goodness of fit, great efforts were put into transformation of LF parameters. In addition, the interaction between independent variables was considered.15,16

The final model was mainly determined by a combination of determinant coefficient (R2, which is the fraction of explained variability), statistical significance of partial regression coefficient, residual standard deviation (RSD), simplicity and ease of remembrance. Additionally, compatibility with published prediction equations was also taken into account.

Prediction equations for the fifth percentile were derived from the regression model as: LLN=predicted mean-1.645× RSD, if the residuals had a distribution close to Gausian.17 Similarly, normalization of residuals was required when the criteria were not met.

To test the difference on COPD diagnosis in the elderly population between the LLN method and the 70% fixed ratio method, chi-square test was performed for the elderly subjects derived from the original subject population. A paired t-test was used to test differences between predicted values derived from different equations for both average Chinese males (height 170 cm, weight 65 kg) and females (height 160 cm, weight 55 kg).

Back to Top | Article Outline

RESULTS

The study was conducted from April 10, 2010 to March 22, 2012. A total of 1328 never-smokers (672 males, 656 females) were recognized as normal and included in statistical analysis (Table 2). All the subjects involved were residents in Jinan. It is worth noting that up to 67.24% (893/1328) of them were elderly, 72.92% (490/672) of males and 61.43% (403/656) of females were ≥60 years old.

Table 2

Table 2

Back to Top | Article Outline

Prediction equations for LF parameters and LLN

In developing the regression models, age and height were found to be the main predictors for all LF parameters. By contrast, weight and BMI were excluded, due to their small contribution to improvement of goodness of fit.

Transformation of LF parameters (logarithm, square or square root) did not significantly improve the R2. Instead, considering the interaction between age and height, LF parameters were divided by age and normalized by taking the natural logarithm. That is, all LF parameters were

converted into the form of Ln Symbol. This led to marked improvement of R2, especially for FEV1/FVC (Table 3). As the residuals demonstrated a log-normal distribution, LLN was calculated with the lognormal method.

Table 3

Table 3

Symbol

Symbol

The prediction equations for LF parameters and LLN were modeled as:

Where H represents height, A age and βn partial regression coefficient. Note that InterceptPRD is used for the prediction equation and InterceptLLN for LLN equation (Table 3). All the regression models and their partial regression coefficients were statistically significant (P < 0.001).

Back to Top | Article Outline

Spirometric definitions of airflow limitation: FEV1/FVC <70% vs. FEV1/FVC <LLN

Of the 1328 subjects, 126 (9.49%, 81 males and 45 females) were classified as asymptomatic COPD patients, using the fixed 70% method. Meanwhile, up to 100 of 126 COPD patients were elderly, 13.88% (68/490) of the males and 7.94% (32/403) of the females. In contrast, of the 1328 subjects, only 57 (4.29%, 30 males and 27 females) were classified as asymptomatic COPD patients, using the LLN method (FEV1/FVC <LLNCS was interpreted as obstructed and LLNCS meant LLN values calculated using equation 2). Forty-three of 57 COPD patients were elderly, 5.15% (27/490) of males and 3.97% (16/403) of females.

Using the two spirometric definitions, the number of PFT results interpreted as obstructed varied markedly in the elderly population for both genders (males: χ2=19.59, P <0.01; females: χ2=5.67, P=0.017). A total of 57 subjects age ≥60 years, whose FEV1/FVC ratio fell between LLNCS and 70%, were defined as “misclassified subjects” or “over-diagnosed subjects”. Their average age was (73.04±6.38) years, and LLNCS averaged (67.06±1.92)%. Concerning the classification of COPD severity, 36 (24 males and 12 females) were classified as GOLD stage I, while the remaining 21 (17 males and 4 females) were classified as GOLD stage II.

Tables 4 and 5 show the number of subjects diagnosed as COPD using different LLNLead Author values as the spirometric cut-offs. LLNLead Author values were derived from different reference equations and marked by the last names of lead authors. For both genders, significant differences were found on COPD diagnosis in the elderly population between the LLN method and the 70% fixed ratio method in most cases (Tables 6 and 7).

Table 4

Table 4

Table 5

Table 5

Table 6

Table 6

Table 7

Table 7

Back to Top | Article Outline

Comparisons between predicted spirometric values derived from different equations

Compared to the newly developed reference values, significant differences were found in most cases (P <0.05), as shown in Tables 8 and 9. The degree of overestimation also varied. As expected, the newly developed values were much closer to those derived from Chinese than from Caucasian subjects.

Table 8

Table 8

Table 9

Table 9

Back to Top | Article Outline

DISCUSSION

Our study established new spirometric reference equations for residents in Jinan (age 56–84 years) and confirmed that the GOLD definition of airflow limitation would probably lead to over-diagnosis in asymptomatic elderly never-smokers. To demonstrate the LF parameters' dependency on age and height, ATS and the European Respiratory Society reached a consensus that statistically derived LLN should be used in lieu of the fixed ratio,5 which was also approved by the GOLD guidelines (Updated 2010). Our study suggests that predicted LLN values reduce the misclassification risk and better describe the change in lung volumes with aging, especially in the elderly population. Take an average Chinese male aged 80 years for example: his PFT results could be interpreted as airflow limitation only when the FEV1/FVC ratio falls below 61.01%.

Due to known ethnic differences in lung function,6,7 ATS and the European Respiratory Society agreed on the use of ethnic-specific reference equations whenever possible.5 If such equations were unavailable, ethnic adjustment factors could be applied. Note that adjustment factors should not be applied to the FEV1/FVC ratio. By far, most published reference equations were derived from Caucasian and few could be applied to mainland Chinese.

The first spirometric reference equations specified for Chinese were established by Wu et al18 in the 1950s and widely used for nearly two decades. Later, Ip et al. published reference equations for Hong Kong and overseas Chinese.13,19,20 Lin and his colleagues conducted a large-scale, multi-center study (six provinces involved) on spirometric prediction equations in 1990.14,21–25 Actually, due to the small sample size and overseas Chinese based data sets, few of them was widely used. Vollemer et al26 summarized the reference values for mainland Chinese aged 35–56 years, but failed to publish corresponding reference equations. In recent years, there has been a remarkable lack of studies of this kind.

China has the largest population in the world. Therefore, new updates of specific reference equations for mainland Chinese are urgently needed. Our study established specified spirometric reference equations for residents in Jinan. Even as a single-center study, our method could be promoted and replicated in the future nationwide and multicenter studies.

One strength of our study is the stringent exclusion criteria we applied. Due the absence of recognized exclusion criteria, we combined the previous recommendations and put forward a stringent definition of “normal population”. As is known, strict application of exclusion criteria leads to additional difficulties in recruiting a reference population, especially elderly subjects. In a sharp contrast to previous studies, up to 67.24% (893/1328) of the reference population were the elderly (age ≥60 years) in the current study. All credit to the majority of subjects included were middle-aged and elderly people who performed spirometry for routine health examination. Both a substantial proportion of elderly subjects and the large sample size, greatly improve the representativeness and reliability of the reference equations.

To establish the best-fit reference equations, great efforts were put into data transformation. We did not replicate previous researchers' work blindly; instead, we took the interaction between independent variables into account, improving the goodness of fit to a great extent. GOLD guidelines recommended the use of a post-bronchodilator spirometry test.1 Nearly 1/3 of post-bronchodilator FEV1/ FVC values could go back to the normal range.27 Similar to most published equations, ours overestimated the number of PFT results interpreted as obstructed, as only pre-bronchodilator values were achieved.Lower limit of normal or FEV1/FVC <0.70 in diagnosing COPD: an evidence - based review. Respir Med 2011; 105: 907-915.','400');" onMouseOut="javascript:ImageWrapperControl_ImageMouseOut();">28 Compared to the LLN method, around half of the asymptomatic elderly never-smokers (57/100) could probably be over-diagnosed using the fixed 70% method. According to Brazzale and his colleagues, approximately 28% of those over 65 years were over-diagnosed with the GOLD definition.29 Apparently, more related research and data are needed.

Unfortunately for a cross-sectional study, it is not possible to determine which spirometric definition of airflow limitation is superior in view of the absence of a gold standard test for COPD. The only valid conclusion is that they may lead to disagreement in the interpretation of PFT results. GOLD guidelines (Updated 2010) recommended the use of LLN, but pointed out at the same time that longitudinal studies to validate the use were urgently needed.1 On one hand, the fixed 70% method led to a missed opportunity of early diagnosis and early intervention in young adults, especially of mild disease.FEV1/FVC <70% as a fixed cut-off: a longitudinal evaluation of clinical and functional outcomes. Thorax 2008; 63: 1040-1045.','400');" onMouseOut="javascript:ImageWrapperControl_ImageMouseOut();">4 On the other hand, over-diagnosis of elderly subjects was supposed to cause unnecessary costs together with individual and social harm. However, as a great many asymptomatic elderly individuals feel it is unnecessary to get tested, the economic cost due to the misclassification could be less than expected and difficult to measure in the general population. Mannino et al carried out a cohort study up to 11 years, suggesting that the over-diagnosed elderly subjects (age >65 years) are more likely to die and have a COPD-related hospitalization than the health control individuals during follow-up.30 This conclusion did not match the expectations and remained to be elucidated.

Overall, our study confirmed that the GOLD criterion of airflow limitation (the fixed 70% method) would probably lead to over-diagnosis in asymptomatic elderly never-smokers. Instead, predicted LLN values reduce the misclassification risk. Longitudinal studies are urgently needed to determine which spirometric definition better predicts the outcomes of the misclassified subjects. Large-scale and multicenter studies are expected to be conducted in order to establish specific reference equations for mainland Chinese. Furthermore, due to the presence of cohort effects, consideration should be given to updating reference equations on a regular basis, e.g. every 10 years.5

Back to Top | Article Outline

Acknowledgments:

The authors would like to be grateful to Dr. WANG Ai-hua and Dr. ZHENG Chun-yan for expert advice on interpretation of PFT results. The authors also thank all the technicians of the lung function laboratory for their technical support for the study, especially Ms. LIU Chun-hong and Ms. ZHANG Haiqing. All the contributors above are staff members of Qilu Hospital, Shandong University, China.

Back to Top | Article Outline

REFERENCES

1. GOLD Science Committee. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease. NHLBI/WHO Global Initiative for Chronic Obstructive Lung Disease (GOLD) workshop summary, 2010. (Accessed July 23, 2011 at http://www.goldcopd.org)
2. Hardie JA, Buist AS, Vollmer WM, Ellingsen I, Bakke PS, Mørkve O. Risk of over-diagnosis of COPD in asymptomatic elderly never-smokers. Eur Respir J 2002; 20: 1117-1122.
3. Roberts SD, Farber MO, Knox KS, Phillips GS, Bhatt NY, Mastronarde JG, et al. FEV1/FVC ratio of 70% misclassifies patients with obstruction at the extremes of age. Chest 2006; 130: 200-206.
4. Cerveri I, Corsico AG, Accordini S, Niniano R, Ansaldo E, Antó JM, et al. Underestimation of airflow obstruction among young adults using FEV1/FVC <70% as a fixed cut-off: a longitudinal evaluation of clinical and functional outcomes. Thorax 2008; 63: 1040-1045.
5. Pellegrino R, Viegi G, Brusasco V, Crapo RO, Burgos F, Casaburi R, et al. Interpretative strategies for lung function test. Eur Respir J 2005; 26: 948-968.
6. 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: 179-187.
7. Korotzer B, Ong S, Hansen JE. Ethical difference in pulmonary function in healthy nonsmoking Asian-Americans and European-Americans. Am J Respir Crit Care Med 2000; 161 (4 Pt 1): 1101-1108.
8. Wang C, Xiao W, Ma D. Epidemiological status and related risk factors of COPD in urban residents in Jinan. J Shandong Univ (Health Sciences) (Chin) 2011; 49: 128-132.
9. Johannessen A, Omenaas ER, Eide GE, Bakke PS, Gulsvik A. Feasible and simple exclusion criteria for pulmonary reference populations. Thorax 2007; 62: 792-798.
10. Vollmer WM, Gíslason T, Burney P. Comparison of spirometry criteria for the diagnosis of COPD: results from BOLD study. Eur Respir J 2009; 34: 588-597.
11. American Thoracic Society. Standardization of spirometry, 1994 update. Am J Respir Crit Care Med 1995; 152: 1107-1136.
12. 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.
13. Ip MS, Lau AC, Yu WC, Tang KS, Choo K, Chan-Yeung MM; Hong Kong Thoracic Society; American College of Chest Physicians (Hong Kong and Macau Chapter). Updated spirometric reference values for adult Chinese in Hong Kong and implications on clinical utilization. Chest 2006; 129: 384-392.
14. Lin SW, Swi KS, Kou YS. Normal values of Lung function in the population of northern China. In: Mu KJ, Liu SW, eds. Nationwide normal values of lung function. Beijing, BJ: PUMC & Beijing Medical University Publication; 1990: 1-23.
15. Cole TJ. Linear and proportional regression models in the prediction of ventilatory function. J R Stat Soc Ser A 1975; 138: 297-338.
16. Dockery DW, Ware JH, Ferris BG Jr, Glicksberg DS, Fay ME, Spiro A 3rd, et al. Distribution of forced expiratory volume in one second and forced vital capacity in healthy, white, adult never-smokers in six U.S. cities. Am Rev Respir Dis 1985; 131: 511-520.
17. Stocks J, Quanjer PH. Reference values for residual volume, functional residual capacity and total lung capacity. ATS Workshop on Lung Volume Measurements. Official Statement of The European Respiratory Society. Eur Respir J 1995; 8: 492-506.
18. Wu MC, Yang SP. Pulmonary function study in healthy Chinese. Report 1: Lung volume and its subdivisions. J Formosan Med Ass 1962; 61: 110-129.
19. Chia SE, Wang YT, Chan OY, Poh SC. Pulmonary function in healthy Chinese, Malay and Indian adults in Singapore. Ann Acad Med Singapore 1993; 22: 878-884.
20. Chin NK, Ng TP, Hui KP, Tan WC. Population based standards for pulmonary function in non-smoking adults in Singapore. Respirology 1997; 2: 143-149.
21. Jia IQ, Mo BL, Guo XJ, Wang IP. Normal values of lung function in the population of South-West China. In: Mu KJ, Liu SW, eds. Nationwide normal values of lung function (Chin). Beijing: PUMC & Beijing Medical University Publication 1990: 38-48.
22. Hou S, Zhan YG, Guo SC, He I. Normal values of lung function in the population of Guangdong. In: Mu KJ, Liu SW, eds. Nationwide normal values of lung function (Chin). Beijing: PUMC & Beijing Medical University Publicationl990: 67-80.
23. Liu GH, Ma XQ. Normal values of lung function in the population of the North-East China. In: Mu KJ, Liu SW, eds. Nationwide normal values of lung function (Chin). Beijing: PUMC & Beijing Medical University Publication 1990: 24-30.
24. Sun B, Li RH. Normal values of lung function in the population of north-west China. In: Mu KJ, Liu SW, eds. Nationwide normal values of lung function (Chin). Beijing: PUMC & Beijing Medical University Publication 1990: 1-23.
25. Zhu X, Li WD. Normal values of lung function in the population of Shanghai. In: Mu KJ, Liu SW, eds. Nationwide normal values of lung function (Chin). Beijing: PUMC & Beijing Medical University Publication 1990: 57-66.
26. Vollmer WM, Tsai R, Wu Y, Li YH, Johnson LR, Williams OD, et al. Patterns of lung function in asymptomatic nonsmoking men and women in the People's Republic of China. Ann Epidemiol 2002; 12: 295-302.
27. Pérez-Padilla R, Hallal PC, Vázquez-García JC, Muiño A, Máquez M, López MV, et al. Impact of bronchodilator use on the prevalence of COPD in population-based samples. COPD 2007; 4: 113-120.
28. Mohamed Hoesein FA, Zanen P, Lammers JW. Lower limit of normal or FEV1/FVC <0.70 in diagnosing COPD: an evidence - based review. Respir Med 2011; 105: 907-915.
29. Brazzale DJ, Upward AL, Pretto JJ. Effects of changing reference values and definition of the normal range on interpretation of spirometry. Respirology 2010; 15: 1098-1103.
30. Mannino DM, Buist AS, Volmer WM. Chronic obstructive pulmonary disease in the older adult: what defines abnormal lung function? Thorax 2007; 62: 237-241.
Keywords:

pulmonary disease, chronic obstructive; spirometry; FEV1/FVC; lower limit of normal

© 2013 Chinese Medical Association