| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Department of Medicine (Drs. Ip and Chan-Yeung), The University of Hong Kong, Queen Mary Hospital; Department of Medicine and Therapeutics (Dr. Ko), The Chinese University of Hong Kong, Prince of Wales Hospital; Department of Medicine (Dr. Lau), Pamela Youde Nethersole Eastern Hospital; Department of Medicine (Dr. Yu), Princess Margaret Hospital; Department of Medicine (Dr. Tang), Tuen Mun Hospital; and Department of Medicine (Dr. Choo), Northern District Hospital, Hong Kong SAR, China.
Correspondence to: Mary Sau-man Ip, MD, FCCP, Department of Medicine, The University of Hong Kong, Queen Mary Hospital, Pokfulam, Hong Kong SAR, China; e-mail: msmip{at}hkucc.hku.hk
Abstract
Study objectives: The accuracy of reference values of lung function is important for assessment of severity and functional impairment of respiratory diseases. The aim of the study was to establish updated prediction formulae of spirometric parameters for Hong Kong Chinese and to compare the reference values with those derived from other studies in white and Chinese subjects.
Design: Cross-sectional multicenter study.
Setting: Lung function laboratories of eight regional hospitals in Hong Kong.
Participants: Subjects were recruited by random-digit dialing. One thousand one hundred seventy-six subjects who fulfilled recruitment criteria underwent spirometry.
Measurements: Spirometry was performed according to American Thoracic Society recommendations, and the technique was standardized among the eight participating lung function laboratories.
Results: Evaluable data of 1,089 (494 men and 595 women) healthy nonsmokers aged 18 to 80 years were analyzed. Age and height were found to be the major determinants of FEV1 and FVC, with a linear decline of height-adjusted values with age in both sexes. Spirometric values of this population have increased compared to Chinese populations of similar sex, age, and height two decades ago. Reference values derived from white populations were higher than our values by 5 to 19%, and the degree of overestimation varied with age, sex, and lung function parameter. We also demonstrated that the blanket application of correction factors for Asian populations may not be appropriate. In this study cohort, the distribution-free estimation of age-related centiles was more appropriate for the determination of lower limits of normal.
Conclusions: Our findings underscore the need to use reference values based on updated data derived from local populations or those matched for ethnicity and other sociodemographic characteristics.
Key Words: Chinese adults • lung function • reference values
Lung function testing is an important tool in the diagnosis and evaluation of the severity or functional impairment of many respiratory diseases,12 as well as in surveillance programs in some industries.34 The interpretation of lung function tests usually depends on comparison with reference values derived from a "normal" population. The accuracy of the reference values has important implications for both the individual and health-care practice.
A variety of individual, behavioral, and environmental factors affect lung function development in childhood and adolescence, and the subsequent lung function decline with age.5678 Genetic factors may control body habitus and lung function development. It has been well documented9101112 that different ethnic groups have different lung function values, and reference values for lung function tests should not be applied across different ethnic groups without thor ough evaluation. Exogenous factors such as smoking, nutrition, exercise, air quality, and occupational exposures may affect both lung function development and decline.7131415 Such exogenous factors vary among individuals and are prone to change over time in an individual as well as the community at large.1011 Given the importance of normal values in decision making in health care, such cohort effects should be addressed through regular updating of reference values.
Compared to children of the same age and height two decades ago, it has been demonstrated that several parameters of lung function in children and adolescents in Hong Kong have increased, especially in younger girls.1011 The purpose of this study was to determine normal lung function values for the adult population in Hong Kong and to compare the results with those derived from other populations as well as those from Hong Kong Chinese two decades ago.
Materials and Methods
The study was carried out between January 2001 and March 2003 on subjects residing in the Hong Kong Special Administrative Region, China, where 98% of the population are ethnic Chinese. Ethics committee approval was obtained from The University of Hong Kong (HKU) and each of the eight hospitals where the study was performed. Written informed consent was obtained from each subject before lung function testing.
Study Population
Using random-digit dialing, adults aged 18 to 80 years were identified from the general population in Hong Kong by a research nurse over the telephone. Of 10,402 telephone calls that were answered, 2,066 calls were rejected as the receiver refused to respond further. Of the 8,336 subjects who answered the screening questionnaire based on the American Thoracic Society (ATS) questionnaire for chronic respiratory symptoms,16 3,465 subjects were excluded because they had at least one of the following: a history of smoking in the past year or for one or more pack-years (n = 1,431); a history of lung diseases including asthma, COPD, pulmonary tuberculosis, lung fibrosis, or spinal abnormalities; pleurodesis and chest tube insertion; symptoms of chronic cough, wheeze, or dyspnea; history of thoracic surgery; history of major acute illness in the past 3 months; or a history of respiratory tract infection in the past 4 weeks. Two thousand two hundred eighty-two subjects refused to participate in lung function testing, while spirometry was not performed in 1,413 subjects who were in excess of the quota for their age group (set at 50 subjects for 18 to 20 years old and each subsequent 5-year age group). Eventually, 1,176 subjects underwent spirometry. Eighty-seven subjects were further excluded as their test results were judged as "unusable" due to technically unsatisfactory test performance. Reasons for exclusion were inability to obtain three acceptable spirograms due to cough, premature termination of exhalation, variable effort or leakage, or inability to obtain two reproducible tests, following the test protocol as described in "Methods" section below. Analysis of spirometric data was thus carried out on a total of 1,089 18- to 80-year-old nonsmokers.
Methods
Lung function testing was performed in the lung function laboratories of eight hospitals equipped with an identical computerized spirometer (
max 22; SensorMedics; Yorba Linda, CA), following the procedures recommended by the ATS guidelines.17 For each subject, standing height was measured following standardized technique.18 The lung function system was calibrated with a 3-L syringe at different flow rates at least once daily. All spirometric measurements were performed in a sitting position. FEV1 and FVC were measured in each subject. All technicians, who were experienced in lung function testing, were further trained before the study by a senior technician from HKU to ensure that spirometry was done according to ATS criteria.17 The technicians were instructed to conduct a minimum of three and a maximum of eight maneuvers to ensure that the subjects produced the highest possible peak flow rates. Satisfactory exhalation was considered achieved if there was either no change in exhaled volume (plateau on the volume-time curve) for 1 s after an exhalation time of at least 6 s; or a reasonable duration or plateau in the volume-time curve; or inability of the subject to continue to exhale.17 The lung function machine automatically saved three acceptable tracings with the largest FEV1 and FVC that were reproducible within 200 mL of each other. The largest FEV1 and the largest FVC from these curves were used in the analysis according to the ATS criteria.17 The "best" curve, defined as the one with the largest sum of FEV1 plus FVC, was used to determine the average forced expiratory flow at 25 to 75% of expired volume (FEF25–75%). All the results were corrected to body temperature and pressure, saturated units.
Quality Control
The equipment from the eight laboratories was standardized and calibrated at the onset of the study by a technician from the manufacturer. Spirometry was performed by this technician in each of the eight laboratories, and the FVC ranged from 3.21 to 3.27 L, a variation of 1.9%. In addition, a senior technician at the HKU laboratory provided periodic on-site observation and feedback at all participating laboratories. A respiratory specialist from each participating hospital and the senior technician at HKU scrutinized all raw data. Only tests that were considered technically acceptable and reproducible according to the ATS criteria17 were included for final analysis.
Statistical Analysis
Statistical analyses were performed (SPSS for Windows Version 11.2; SPSS; Chicago, IL). Graphic procedures and testing of residuals for normality of distribution were performed using a one-sample Shapiro-Wilk test. Reference equations were developed using multiple regression equations. The spirometric values FEV1, FVC, FEV1/FVC, and FEF25–75% were regressed against the independent variables of age, standing height, weight, and body mass index. Different methods of transformations such as square, logarithm, and square root of spirometric parameters were also used in the multiple regression equations to improve the predictability of the equations. The SEE and R2 were calculated for each equation. The final model and inclusion of independent variables were based on a combination of R2, SEE, simplicity, and ease of use.
Comparisons of regression curves and prediction values of various lung function parameters were made between those derived from the current study and those derived from other studies based on white populations (Crapo et al,19 Knudson et al,20 Hankinson et al,21 Quanjer et al,22 Falaschetti et al,23) and Chinese populations (Lam et al,24 Da Costa,25 Hou et al26).
Prediction equations for the fifth percentile of each spirometric variable were derived based on the distribution-free estimation method described by Healy and Rabash.27 In brief, the data of each gender were first sorted by age, and the empirical fifth percentile of the residuals was computed for consecutive observations of 200 cases in a rolling continuum. These percentiles were then regressed against a polynomial sequence of median age. The lower limit of normal (LLN) was calculated from the prediction equation thus derived for each lung function variable. To test the goodness of fit of this model, we calculated the percentages of subjects, in four age strata, who would be considered as subnormal by its equations, and compared with the percentages yielded by the method assuming constant difference between the mean and fifth percentile (mean – 1.645 residual SD).
Results
The age and gender distribution of the participants are shown in Figure 1 . Table 1 shows the prediction equations of spirometry chosen for our population as a result of the above analysis. In developing the regression model, age and height were found to be significant independent variables for all pulmonary function parameters. Adding weight and body mass index did not improve the R2. Transformation of spirometric parameters, already mentioned in the "Methods" section, did not improve the R2 for all parameters with the exception of FEV1/FVC, where R2 increased from 0.318 to 0.615 in men and 0.407 to 0.589 in women after transformation by dividing the parameter by the square of height. The normalized residuals corresponding to these models did not differ significantly from the Gaussian distribution in all parameters as determined by the Shapiro-Wilk test. The height-adjusted FEV1 and FVC against age for male and female subjects were mostly within two SD lines, and the regression lines of these spirometric parameters were linear (data not shown).
|
|
|
|
|
|
|
|
|
This is a study of reference values of spirometric lung function in a random population of Chinese adults in Hong Kong. All relevant data were obtained by trained technicians from healthy nonsmokers using standardized equipment, techniques, acceptability, and reproducibility. The predicted spirometric values derived from this study showed varying degrees of difference when compared with those derived from studies on white subjects as well as other studies on Chinese.
Differences in predicted values obtained in various studies may be attributed to technical factors involved in lung function testing. Different lung function devices have been used, and technology has evolved. The more recent studies have employed computerized systems that portend to high precision, but between-instrument variability would still exist and contribute to variations in measurement.17 It is not possible for us to investigate for differences in measurement due to the different instruments in the various studies. Nonetheless, we have attempted to minimize the effect of instrument variation within our study by using the same model of lung function system in the eight centers. The sitting posture has also been shown to result in slightly lower spirometric values than the standing posture.28 However, the postural effects are small and probably much less important in determination of the measurements than the quality with which the tests were conducted.28 We have been rigorous in the provision of quality control to ensure that all test procedures were done and included for analysis in accordance to published guidelines,17 in order to minimize potential technical sources of error or nonuniformity.
Similar to many previous studies2930313233 that have shown that Asians, such as Chinese, Indians, Japanese, and Malaysians, have smaller lung volumes than whites, we found that the FEV1 and FVC values of our adult population were lower than those of whites for all age groups with the same age and height. One study32 reported differences in lung function between Chinese residing in China and those overseas. Such differences were likely related to environmental factors that might affect lung function.567834 To avoid incorrect interpretation of results by using prediction equations derived from a population of different ethnicity, it would be necessary for countries to establish their own reference values for their populations. As it requires considerable resources to carry out a lung function study on a random population sample, the use of correction factors for ethnicity has been suggested.22323536 The ATS recommended a correction factor of – 12% for African-American subjects,35 and a factor of – 15% was found to be appropriate when applied to the data obtained by Knudson et al.3637 In our analysis, we found that the degree of overestimation depended on which prediction equation for white subjects was used. Compared to our predicted values, the overestimation was largest with the prediction of Crapo et al19 and least with that of Knudson et al.20 The degree of overestimation also differed with age and sex. Thus, blanket application of a single correction factor to all prediction equations derived from white populations is inappropriate. The European Coal and Steel Community has suggested that a conversion factor of 1 (no adjustment) should be applied to white subject predicted values for Hong Kong Chinese, while a factor of 0.89 should be applied to white subject predicted values for Japanese.22 Our findings suggest that this would result in a significant overestimation for Chinese in Hong Kong, and thus Southern Chinese. Zheng and Zhong32 suggested that, based on data from several studies conducted in mainland China, application of conversion factors of 0.93 to 0.95 to the European Coal and Steel Community values of various spirometric parameters may be more appropriate for Chinese. Although these conversions resulted in values closer to those obtained in this study, the differences in any one lung function parameter may vary substantially in different age groups, suggesting that one conversion factor would not suffice for the entire range of age in the community.
The reference values for spirometric lung function parameters derived from this study were similar to those of Hou et al26 found in Southern Chinese. The predicted FEV1 or FVC for individuals of the same sex, age, and height differed by < 5% using these prediction equations. This is not unexpected, as Chinese from Hong Kong predominantly originated from Southern China. The previous study by Zheng and Zhong32 also found that normal lung function values differed between Northern and Southern Chinese, with the former having higher values of FEV1 and FVC.
The predicted spirometric values, in particular FEV1, of Da Costa25 for Chinese were lower compared with those derived from the current study, and this may be due to several factors. The study by Da Costa25 was done on a relatively smaller number of Chinese (n = 207) in Singapore and included current and ex-smokers with "light smoking history." Smoking has important bearing on lung function, but this is probably not the major reason for the difference seen between the results of the two studies, especially in the younger age group. The detrimental effect of smoking on lung function is expected to be less in younger adults, as smoking duration is usually shorter. However, our comparison shows that the differences in both FEV1 and FVC in men between the two studies were largest at 20 years of age, compared to 40 years and 60 years of age, while the differences were essentially constant in women in various age groups. Smoking could have contributed in part to the lower values, especially in FEV1, seen at the older age range. People of the same ethnicity residing in different parts of the world may differ in their physical characteristics, diet, and environmental exposures. Furthermore, there has been a lapse of almost 30 years between his study25 and the current study. Ip and colleagues1011 reported a significant increase in height-adjusted values of FVC and FEV1 in both boys and girls in Hong Kong over the entire height range, compared to those obtained in 1985. Their findings and those of the current study support the possibility that environmental factors and improved nutrition may contribute significantly to differences in lung function values over time in the same ethnic group and also between groups of same ethnicity but residing in different localities.
The differences in predicted values derived from our study and those of Lam and coworkers24 are more difficult to interpret. Lam and coworkers24 carried out their study in Hong Kong in the early 1980s, and asymptomatic current smokers were included. The slope of the regression equation derived from the data of Lam et al24 was steeper when compared to other studies. Using their regression equations, the predicted FEV1 was higher in the younger age groups but lower in the older age groups, compared with those based on our regression equations. Subjects in the study by Lam et al24 were volunteers from various institutions, while our subjects were recruited randomly from the community. The inclusion of smokers in the study by Lam et al24 might also have accounted for a more rapid decline in lung function, further contributing to the steep regression slope in relation to age.
We also noted that our FEV1/FVC ratios were higher, except for the elderly age groups, than that in white subjects but similar to that reported by Hou et al.26 The higher ratios in Chinese were probably related to a larger discrepancy in FVC, relative to that in FEV1, between our subjects and whites, and this phenomenon was more prominent in younger age groups.
The delineation of the LLN values is of vital importance in the clinical utilization of reference values. There are different methods for derivation of LLN. Traditional methods assumed a constant difference between the mean and fifth percentiles in the entire age range of the study population. Reports2338 have demonstrated that the distribution-free estimation of age-related centiles for derivation of LLN was more appropriate to data obtained from their study population. In our study cohort, this method also yielded better approximation to the fifth percentiles in general and a more stable age-related profile. We thus propose that this method should be applied. Nonetheless, it is important to note that, due to the relatively small sample size in the elderly, the LLN values for this age group, derived from any prediction formulae, have to be interpreted with caution.
Conclusions
Our findings reinforce the importance of using lung function values derived from nonsmoking populations matched for ethnicity, and such data may need to be examined periodically for secular trends. The use of an adjustment factor applied to prediction equations derived from external populations for race correction may lead to overestimation or underestimation of predicted values. When it is not feasible to obtain reference lung function values from the same community, it is important to adopt such values derived from the same ethnic group with similar sociodemographic characteristics, and preferably as updated as data availability allows.
Acknowledgements
The authors thank Ms. Agnes Lai for coordination of the study and data management; Mr. K. M. Lo for providing expert technical advice on standardization of lung function testing procedures; Ms. Anne Dybuncio (University of British Columbia) and Dr. Daniel Fong (The University of Hong Kong) for statistical analysis; Dr. Michael Schalzer (University of British Columbia) for statistical advice; and nurses and technicians of the participating lung function laboratories for their technical support. We also thank the following for their expert advice and support for the study: Dr. Christopher Lai; Dr. Ling Sai On (Kowloon Hospital); Dr. Johnny Chan (Queen Elizabeth Hospital); Dr. Dominic Choy (Prince of Wales Hospital); Dr. Loretta Yam (Pamela Youde Nethersole Eastern Hospital); Dr. David Hui (Prince of Wales Hospital); Professor W. K. Lam (Queen Mary Hospital); Dr. W. K. Lam (Northern District Hospital); Dr. Thomas Mok (Kowloon Hospital); Ms. Carmen Chan (Princess Margaret Hospital); Ms. C. F. Sin (Tuen Mun Hospital); Dr. C. Y. Tam (Tuen Mun Hospital); Dr. Lam Bing (Queen Mary Hospital); Professor Kenneth Tsang (Queen Mary Hospital); Dr. William Chen (Kowloon Hospital); Dr. K. K. Wong (Northern District Hospital).
Footnotes
Abbreviations: ATS = American Thoracic Society; FEF25–75% = forced expiratory flow at 25 to 75% of expired volume; HKU = The University of Hong Kong; LLN = lower limit of normal
This study is funded by a Research Grant from the Pneumoconiosis Compensation Fund Board, Hong Kong.
Received for publication January 12, 2005. Accepted for publication June 11, 2005.
References
This article has been cited by other articles:
|
|
 |
|
  S. K. F. Leung, W. W. Yew, P. C. Wong, D. Y. T. Fong, A. C. W. Lau, and M. S. M. Ip Lung Function Testing Prediction Equations: Do They Fit All? Chest, May 1, 2008; 133(5): 1288 - 1289. [Full Text] [PDF]
|
|
|
|
|
|
A. C.-W. Lau, M. S.-M. Ip, C. K.-W. Lai, K.-L. Choo, K.-S. Tang, L. Y.-C. Yam, and M. Chan-Yeung Variability of the Prevalence of Undiagnosed Airflow Obstruction in Smokers Using Different Diagnostic Criteria Chest, January 1, 2008; 133(1): 42 - 48. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. W. S. Ko, M. Ip, P. K. S. Chan, M. C. H. Chan, K.-W. To, S. S. S. Ng, S. S. L. Chau, J. W. Tang, and D. S. C. Hui Viral Etiology of Acute Exacerbations of COPD in Hong Kong Chest, September 1, 2007; 132(3): 900 - 908. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
18 Years Old*
