Predictive Equations for Lung Function Based on a Large Occupational Population in North China

J Occup Health 2009; 51: 471 477 Journal of Occupational Health Predictive Equations for Lung Function Based on a Large Occupational Population in North China Yonghui WU 1, Zhongyi ZHANG 2, Baoqi GANG 1 and Edgar J. LOVE 3 1 Department of Occupational Health, School of Public Health, Harbin Medical University, P.R. China, 2 Department of Cardiology, School of Medicine, King s College London, UK and 3 Department of Community Health Sciences, Faculty of Medicine, University of Calgary, Canada Abstract: Predictive Equations for Lung Function Based on a Large Occupational Population in North China: Yonghui WU, et al. Department of Occupational Health, School of Public Health, Harbin Medical University, P.R. China Objectives: The currently used predictive equations of lung function in North China were derived from early study and have not been updated for nearly two decades. Methods: Using American Thoracic Society (ATS) standards, sexspecific spirometric predictive equations for forced vital capacity (FVC), forced expiratory volume in one second (FEV 1 ), ratio of FEV 1 to FVC (FEV 1 %) and forced expiratory flow at 25 75% of forced vital capacity (FEF 25 75% ) were derived from 2,897 asymptomatic, lifelong non-smokers (1,208 males, 1,689 females) from a large occupational population in North China. Stepwise multiple regressions were carried out to identify the best predictors of lung function parameters and predictive equations. Independent variables considered for inclusion in predictive equations including age, height, weight and chest circumference were examined. Results: Age and height were found to be necessary variables for all lung function parameters. Weight was a significant variable in only half of our equations. Chest circumferences (expired or inspired) was excluded as they are not practical in use. Data from 255 apparently healthy non-smokers were used to validate the equations by comparing percentage predicted values and proportion of subjects with normal predicted values with those from the study group, and a high accordance was obtained. Other equations published and used in North China do not appear to offer advantages over these equations. Conclusions: These newly developed predictive Received Jan 18, 2009; Accepted Jul 17, 2009 Published online in J-STAGE Sep 25, 2009 Correspondence to: Z. Zhang, Department of Cardiology, The James Black Centre, School of Medicine, King s College London, 125 Coldharbour Lane, London SE5 9NU, UK (e-mail: zhongyi.zhang@kcl.ac.uk) equations should ideally be applied to calculate lung function for adult individuals and populations as reference values in North China. (J Occup Health 2009; 51: 471 477) Key words: Chinese adults, Influence factors, Predictive equations, Spirometry Spirometry can be a valuable tool and is increasingly used as an objective measure of lung function for a variety of purposes. Due to differences in study populations, techniques or procedures, equations used to predict normal lung function vary 1 3). It is imperative to establish appropriate predictive equations as reference for a given population. China is a huge country with a vast territory and variety of ethnic groups; thus, it is important to establish predictive equations for lung function parameters applicable to the populations in different geographical regions. Additionally, lung function is estimated by using predictive equations based on studies performed 19 yr ago in relatively small populations 4). These equations may not be suitable nowadays for predicting lung function. Furthermore, these equations are predominately applied in clinical practice to determine low lung function and focus on respiratory diseases. We have previously indicated that evaluation of respiratory function for an occupational population requires a more strict or sensitive criteria than that for population of patients with respiratory disease because the majority of the employees comprising the former population may experience no or minor impairment of lung function 5). The aim of this study was to develop predictive equations of lung function based on measurements from a large occupational population in North China with a wide spectrum of age. Subjects and Methods Study population The study subjects included current employees from

472 J Occup Health, Vol. 51, 2009 five industries and retirees who attended the health clinics of these industries in Harbin, Heilongjiang Province, the People s Republic of China (PRC). Harbin is a relatively new city with a large migrant population, and the majority of residents moved to this city decades ago from other regions in North China. Our study subjects included some migrants and their offspring. The subjects did different types of work involving mild to moderate physical effort. Written consent was obtained from each of the 5,154 subjects (aged 18 to 84 yr) invited to participate in this study. In total, 5,063 consented to take part, a response rate of 98.2%. Using stepwise exclusion procedures, the subjects who were not appropriate for the study were identified. Data from 255 apparently healthy non-smokers were used to validate the equations derived from the study group. Our validation group (aged 18 65 yr, 157 men, 61.6%) is a separate, independent set and was recruited from 419 local residents/staff using the same exclusion procedures. Questionnaire and medical examination All participants were interviewed by specially trained interviewers. A respiratory questionnaire developed by the American Thoracic Society 6) was given to each subject at the start of the interview. The interviewers collected data on demographic and clinical characteristics, including chronic respiratory symptoms, lifetime occupational exposure and smoking history (type of smoking, number of cigarettes smoked per day and how many years the individual had smoked). A comprehensive clinical examination was carried out by physicians who took into account other illnesses that may produce similar symptoms. Chest X-ray and a 12-lead resting ECG were obtained for all subjects. Subjects who had a history or symptoms of past or present respiratory, cardiac or thoracic disease, a history of occupational exposure to substances known to cause lung injury, a history of smoking or an abnormal ECG were excluded from this study. Lung function test The lung function test was performed using two portable spirometers (HI-198, Chest M.I., Inc., Tokyo, Japan) by trained personnel and was monitored by one of the co-investigators. The HI-198 spirometer has been certified as meeting the ATS criteria for spirometry equipment performance characteristics. The volume accuracy of the spirometers was checked daily using a 3- L calibrated syringe before performance of tests. The devices were tested with three to five charges/discharges within the error range of the true volume (± 3%). Spirometry was performed using standardized procedures conforming to all ATS recommendations for adequacy of spirometry performance 7). The tests were conducted at room temperatures ranging from 17 to 21 C and relative humidity of 31 to 37%. Instructions and a demonstration were given before a practice attempt, and then the study subjects were required to inspire to the maximum extent possible through the instrument and then forcibly expire through the instrument for as long as possible. The tests were carried out with the subjects in a standing position wearing nose clips. Height was measured using a stadiometer with subjects wearing indoor clothing without shoes, standing erect with heels close together and arms hanging naturally at their sides. Weight was measured while the subjects were in light indoor clothing. Both measurements were performed by trained personnel. Age was calculated based on the date of interview and date of birth. Chest circumferences at extreme inspiration and extreme expiration were measured with a tape measure at the level of the nipples. The readings of lung function measurement were shown on a small liquid crystal display and printable. The values were automatically corrected for BTPS conditions (body temperature and ambient pressure, saturated with water vapor at these conditions). The following parameters were determined for each subject: forced vital capacity (FVC), forced expiratory volume in one second (FEV 1 ), ratio of FEV 1 to FVC (FEV 1 %) and forced expiratory flow at 25 75% of forced vital capacity (FEF 25 75% ). All subjects were asked to perform the test at least in triplicate, and the largest FVC and largest FEV 1 were recorded after examining the data from all of the usable curves. FEF 25 75% was taken from the maneuver with the largest sum of FVC and FEV 1. A test was considered technically unreliable if any of the following occurred: the start of expiration was unsatisfactory, the patient held his breath, the mouthpiece was obstructed or the lips were not properly sealed around the mouthpiece. Statistical analysis All data obtained were entered into a personal computer for statistical analysis using the Statistical Package for Social Sciences (version 10.0 for Windows, SPSS Inc. Chicago, IL, USA). Significance was noted at the 0.05, 0.01 or 0.001 levels. The normality of data distributions was checked by the Kolmogorov-Smirnov test. Stepwise multiple linear regression analysis was performed with the respective uses of FVC, FEV 1, FEV 1 % and FEF 25 75% as the dependent variable and as the independent variables age, height, weight, inspiration chest circumference (ICC), expiration chest circumference (ECC) and the difference between the inspiration chest circumference and expiration chest circumferences (ICC-ECC). A variable was added to the model if its associated p value for the F test was less than 0.05 and removed if its p value was greater than 0.1. Results Characteristics of study subjects Of the 5,063 subjects who agreed to participate, 2,897

Yonghui WU, et al.: Predictive Equations of Lung Function for Chinese Adults 473 Table 1. Numbers of subjects included and excluded by gender and age Age (yr) Male (1,208) Female (1,689) Included Excluded Included Excluded 18 19 104 82 101 12 20 24 128 97 132 24 25 29 102 122 106 33 30 34 103 145 120 40 35 39 129 133 276 38 40 44 100 205 271 39 45 49 101 195 199 31 50 54 107 193 115 35 55 59 111 190 121 43 60 64 103 241 109 28 65 84 120 217 139 23 Total 1,208 1,820 1,689 346 Table 2. Anthropometric and spirometric characteristics of the study and validation groups by gender Measurements Male Female Study group (1,208) Validation group (157) Study group (1,689) Validation group (98) Age (yr) 42.4 ± 16.40 40.2 ± 11.04 42.3 ± 14.14 41.3 ± 15.06 Height (cm) 170.0 ± 5.64 171.5 ± 4.66 158.5 ± 5.27 158.6 ± 5.97 Weight (kg) 70.5 ± 10.59 60.2 ± 9.86 61.2 ± 8.96 60.2 ± 8.07 ICC (cm) 91.9 ± 8.27 91.7 ± 8.01 89.3 ± 7.13 88.2 ± 8.22 ECC (cm) 86.4 ± 7.94 85.2 ± 7.74 84.8 ± 20.66 85.0 ± 17.23 ICC ECC (cm) 5.5 ± 2.26 5.2 ± 2.17 5.0 ± 2.06 4.3 ± 2.10 FVC (ml) 4,383.1 ± 645.57 4,379.3 ± 471.25 3,368.8 ± 463.79 3,363.0 ± 455.21 FEV 1 (ml) 3,573.4 ± 595.95 3,571.3 ± 559.12 2,854.9 ± 469.48 2,851.2 ± 471.31 FEV 1 % (%) 81.3 ± 4.70 81.0 ± 4.26 84.0 ± 5.65 83.6 ± 4.14 FEF 25 75% (L/s) 5.2 ± 1.23 5.1 ± 1.71 4.0 ± 1.10 3.8 ± 1.15 ICC=inspiration chest circumference; ECC=expiration chest circumference. Data are presented as Means ± SD. participants (aged 18 84 yr, 1,208 men, 41.7%) fulfilled the study criteria, and their spirometric results were used in the final analysis. The main reasons for excluding subjects from the analysis were lung function test failure (70), unreliable lung function test (61), incomplete records from the questionnaire (134) or medical examination (90), current or ex-smoker (1,016), history of occupational exposure to dust/gases/fumes (290), pulmonary or cardiac diseases (125), pulmonary or cardiac symptoms (363), history of chest surgery (4) and past career as a professional athlete (13). In each subgroup according to gender and age at 5-yr intervals, there were at least 100 subjects available for the final analysis (Table 1). The Kolmogorov-Smirnov test and frequency distribution histogram supported a normal or nearly normal distribution of the data analyzed. Anthropometric and spirometric characteristics of the study population and validation group are presented in Table 2. No significant differences were noted for all the parameters between the two groups according to gender. Bivariate correlation analysis Table 3 summarizes the statistical association (Pearson s product-moment correlation) between lung function measurements and age, height and weight and chest circumference. A strong and significant association was found between each of the lung function measurements and age in both genders. FVC, FEV 1 and FEF 25 75% were also significantly associated with height in both genders. There was no significant association between any of lung function parameters and weight except for a weak correlation of FEF 25 75% with weight in males (p<0.05) and FEV 1 % with weight in females (p<0.01). In regard to chest circumference, a p<0.05

474 J Occup Health, Vol. 51, 2009 Table 3. Association of lung function parameters with age, height, weight and circumference by gender Parameters FVC FEV 1 FEV 1 % FEF 25 75% Male Age 0.75 (251.920)*** 0.76 (281.024)*** 0.48 (4.128)*** 0.59 (0.891)*** Height 0.70 (221.731)*** 0.59 (241.342)*** 0.01 (2.101) 0.38 (0.872)*** Weight 0.07 (197.214) 0.05 (201.030) 0.03 (1.912) 0.02 (0.901)* ICC 0.02 (240.125) 0.07 (256.114)* 0.16 (3.474) 0.04 (0.721) ECC 0.06 (233.108)* 0.001(218.571) 0.12 (3.093) 0.09 (0.612)* ICC-ECC 0.26 (201.479) 0.27 (198.254) 0.17 (1.911) 0.18 (0.757)* Female Age 0.65 (265.012)*** 0.66 (271.831)*** 0.28 (3.121)*** 0.52 (0.877)*** Height 0.69 (219.230)*** 0.56 (223.713)*** 0.01 (1.954) 0.27 (0.803)** Weight 0.01 (121.161) 0.06 (145.912) 0.11 (1.362)** 0.06 (0.885) ICC 0.17 (245.971) 0.19 (269.630) 0.10 (3.153)* 0.167(0.617) ECC 0.09 (208.116)* 0.10 (190.221) 0.07 (1.258) 0.09 (0.738) ICC-ECC 0.22 (198.155)* 0.18 (211.603) 0.04 (1.159) 0.10 (0.801) Data are presented as R (SE). *p<0.05; **p<0.01; ***p<0.001. ICC=inspiration chest circumference; ECC=expiration chest circumference. Table 4. Adjusted predictive equations of lung function Regression equations R R 2 Adjusted R 2 SEE F Male FVC 5,515.566 24.178A+65.190H 2.233W 0.926 0.857 0.857 244.40 2,405.829 FEV 1 3,288.002 23.968A+46.342H 0.873 0.751 0.761 291.33 1,922.791 FEV 1 % 106.902 0.154A 0.134H+0.053W 0.505 0.255 0.253 4.06 137.073 FEF 25 75% 2.491 0.040A+0.055H 0.640 0.410 0.409 0.95 419.059 Female FVC 3,919.855+51.222H 16.396A 2.191W 0.854 0.729 0.728 241.87 1,507.252 FEV 1 2,530.923 17.954A+39.726H 2.443W 0.782 0.612 0.611 292.70 885.965 FEV 1 % 100.528 0.120A 0.072H 0.291 0.085 0.084 5.411 77.900 FEF 25 75% 0.628 0.038A+0.031H 0.538 0.290 0.289 0.929 343.815 H=height in centimeters, A=age in years, W=weight in kilograms. association was found for FEV 1 vs. ICC, FVC vs. ECC, FEF 25 75% vs. ECC and FEF 25 75% vs. ICC-ECC in males and for FEV 1 % vs. ICC, FVC vs. ECC and FVC vs. ICC- ECC in females. Multiple regression equations Stepwise multiple regression analysis showed that the three chest circumference variables made relatively small contributions to the equations. Also, considering that measuring the chest circumference is not practical, we decided to exclude the three variables from the equations. This exclusion had only a mild effect on the R and SEE values. The Adjusted Regression Equations are shown in Table 4. From the statistical analysis, the F values for the Adjusted Regression Equations (chest circumferences excluded in stepwise regression analysis) became larger than those for the Unadjusted Regression Equations (chest circumferences introduced in stepwise regression analysis). The R, R 2, adjusted R 2 and Beta values showed little change. In order to compare the predictive power, we calculated the percentage predicted values obtained from the adjusted equations and unadjusted equations. No significant difference was found for each lung function parameter. The difference in prevalence of subjects with less than 70% of the predicted values between the adjusted and unadjusted equations was also insignificant. Validation of the predictive equations The mean of percentage predicted values and proportion of subjects with normal lung function (actual value is equal to or greater than 70% of the predicted value) are shown in Tables 5 and 6. Almost all values were consistently higher for the lung function parameters

Yonghui WU, et al.: Predictive Equations of Lung Function for Chinese Adults 475 Table 5. Percentage predicted values (% of predicted, mean ± SD) derived from the study and validation groups Gender Group FVC FEV 1 FEV 1 % FEF 25 75% Male Study group 100.0 ± 5.90 100.0 ± 8.64 99.9 ± 5.00 101.2 ± 18.49 Validation group 98.7 ± 13.83 99.7 ± 16.45 97.2 ± 10.24 97.6 ± 22.94 Female Study group 100.0 ± 7.11 100.0 ± 10.14 100.0 ± 6.42 100.7 ± 23.61 Validation group 102.7 ± 13.28 101.2 ± 14.42 98.2 ± 8.18 100.1 ± 25.64 Table 6. Comparison of percentage of subjects with normal lung function between the study group and validation groups (%) Gender Group FVC FEV 1 FEV 1 % FEF 25 75% Male Study group 98.34 97.52 99.25 95.03 Validation group 95.00 90.00 92.50 90.00 Female Study group 98.70 97.51 99.17 97.87 Validation group 97.96 96.94 93.88 92.86 in the equation group except for FVC and FEV 1 in females for the proportion of subjects with normal lung function. However, the observed differences between the two groups remained insignificant statistically. Discussion It is generally agreed that a gross difference in interpretation of lung function would occur if different predictive equations are employed 2, 8). Glindmeyer 9) estimated that the variability introduced by choice of different predictive equations could be as much as 20% for FVC, FEV 1 and FEF 25 75%. Our previous studies showed that some of the reviewed reference equations derived from China and other countries would not be appropriate if applied to the occupational population in the North-East China 5). Another concern is that the predictive equations most widely used at present are based on studies on a small population, and the resulting estimates would be far too variable (i.e, would vary between samples if the survey were repeated) 10). Even if the population that the equations are based on is large enough, cross-sectional study, which is most widely used today, is more likely subject to cohort bias due to a variety of host and environmental factors 11). Therefore, such studies need to be repeated regularly to produce updated predictive values more relevant to the characteristics of the population being tested. The present study strictly followed the American Thoracic Society standardized procedures for sample selection, equipment, technique procedure and statistical analysis criteria to select the optimal equations. A random sample including only asymptomatic, lifelong non-smokers was selected to estimate predictive equations for lung function variables FVC, FEV 1, FEV 1 % and FEF 25 75%. The large sample with a wide age range provides a suitable setting for developing such equations. Other equations published and used in North China did not appear to offer advantages over the sample size, data collection protocol and statistical analysis used in the present study. In this study, we examined the validity of the developed predictive equations by comparing the data between the study and validation groups. We noted that in comparison with the values in the validation group, the developed equations tend toward a higher value of lung function. On the other hand, it seems probable that more subjects could be predicted as normal with the equations developed, although this was not supported by the results of our statistical analysis. A random error in sample selection may be involved, and we have no other explanation to sufficiently account for this discrepancy since the same exclusion procedures were used to recruit our validation subjects. Further work should be done to validate the developed equations. Ideally our equations should be used to calculate lung function for adult individuals and populations as reference values in North China. Results for pulmonary function are influenced by many biological factors including race, gender and age and by anthropometrical indices like height, weight and body size; thus, variation in predictive values is largely due to these factors 12). Aging is associated with a decline in pulmonary function 13, 14). This is the theoretical basis on which most published spirometric reference studies have used age in their models because there is good inverse linear correlation between lung function measurements and age. In healthy subjects, lung function increases with growth, reaches a peak around the age of 20 and then begins to slowly decline. On review of our data from scatter plot and correlation analyses (data not shown),

476 J Occup Health, Vol. 51, 2009 both FVC and FEV 1 reached a maximum in males at age of 21 to 22 yr and then reduced with the increase of age, but in females, no obvious peak values could be seen around these ages. In addition, our data showed that FVC was more affected by age than FEV 1. For FEV 1 %, only a slight, but statistically significant, decrease with increasing age was noted in both genders. Height is the second most commonly used predictor for constructing lung function predictive values. We used height as an independent variable in all our predictive equations since it contributed significantly to prediction of FVC, FEV 1 and FEF 25 75%. Another reason we considered age and height as the best predictors in our models is that adding other variables such as weight and chest circumferences did not increase the significance level. The impact of body weight on pulmonary function has been examined in many previous studies. There are abundant data showing the relationship between weight (weight gain or weight loss) and lung function 15 19). In principle, it seems logical to take changes in weight into account to predict various lung function indices. However, weight has been a controversial independent variable in constructing predictive equations of lung function. It is well recognized that increased body weight can be caused by obesity. Obesity can affect the thorax, diaphragm and abdominal muscles, thereby resulting in altered respiratory function 20). However, in most of popular spirometric predictive equations throughout the world, weight is seldom used, which suggests that retaining weight in the final model may not improve on the predictive power of age and height only. The findings in this study indicate that weight is a significant parameter only in half of our regression equations, including FVC and FEV 1 % for males and FVC and FEV 1 for females. Though regression equations using one or two of independent variables (such as age and height) are sufficient, we subscribe to the opinion that equations containing more variables can render more promising results 21) as long as objective statistical criteria are used to select optimal variables. Another anthropometric parameter, chest circumference, which can be linked to chest dimension, is sometimes used to predict lung function in view of the significantly high correlation of this parameter with lung function 22 24). Although chest circumference might theoretically be good at predicting lung function, the reproducibility of the measurements may be poor 25). This may be the reason why this variable is less popular in routine use. In the present study, chest circumference was not considered in development of our predictive models, although parts of our results support its good predictability. The slight discrepancies between the equations with fewer variables and equations with more variables allow us to choose the former as final equations. It is more practical to select simple and accurately measurable anthropometric indices to develop regression equations for prediction of lung function. We have determined that the present equations give slightly higher predicted values for FVC, FEV 1 and FEF 25 75% in both sexes than the existing equations used for populations in eight regions across China, including Dongbei, Huabei, Xibei, Xinan, Shanghai, Guangdong, Taiwan and Hong Kong (data not included) 26), because of the highest possible quality control in this study. In summary, spirometric predictive equations for FVC, FEV 1, FEV 1 % and FEF 25 75% were derived for both genders from a large and representative population sample in North China and by rigidly following ATS guidelines. This study emphasizes the importance of choosing appropriate reference equations. These newly developed predictive equations can be used for evaluation of lung function for the population in North China in both epidemiological studies and clinical practice. These newly derived predictive equations should be a significant improvement over those currently used in North China. Acknowledgments: We express our sincere appreciation and gratitude to the participants, local government and industry management who encouraged their staff to participate in this study. References 1) Crapo RO, Jensen RL, Oyunchimeg M, Tsh T, DuWayne SC. Differences in spirometry reference values: A statistical comparison of a Mongolian and a Caucasian study. Eur Respir J 1999; 13: 606 9. 2) Quadrelli S, Roncoroni A, Montiel G. Assessment of respiratory function: Influence of spirometry reference values and normality criteria selection. Respir Med 1999; 93: 523 35. 3) Enright P, Skloot G, Herbert R. Standardization of spirometry in assessment of responders following manmade disasters: World Trade Center worker and volunteer medical screening program. Mt Sinai J Med 2008; 75: 109 14. 4) Liu GH, Ma XC. The regression formula of estimated and expected values of lung function in the North-East region of P. R. China. In: Mu HJ, Liu SW, editors. National normal values of pulmonary function tests in P. R. China. Beijing: Chinese Xiehe Medical University and Beijing Medical University Publishing House; 1990. p. 10 2. 5) Zhang ZY, Wu YH, Wang FX, Gang BQ, Love EJ. Do the predicted values of lung function apply in North- East China? Int Med J 1999; 6: 163 6. 6) Ferris BG. Epidemiology Standardization Project (American Thoracic Society). Am Rev Respir Dis 1978; 118: 1 120. 7) American Thoracic Society. Standardization of Spirometry, 1994 update. Am J Respir Crit Care Med 1995; 52: 1107 36. 8) Roca J, Burgos F, Sunyer J, et al. Reference values for

Yonghui WU, et al.: Predictive Equations of Lung Function for Chinese Adults 477 forced spirometry. Group of the European Community Respiratory Health Survey. Eur Respir J 1998; 11: 1354 62. 9) Glindmeyer HW. Predictable confusion. J Occup Med 1981; 23: 845 9. 10) 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 63. 11) American Thoracic Society. Lung function testing: Selection of reference values and interpretative strategies. Am Rev Respir Dis 1991; 144: 1202 18. 12) Ostrowski S, Barud W. Factors influencing lung function: Are the predicted values for spirometry reliable enough? J Physio Pharmacol 2006; 57: 263 71. 13) Dockery DW, Ware JH, Ferris BG Jr, 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 20. 14) Wagner NL, Beckett WS, Steinberg R. Using spirometry results in occupational medicine and research: Common errors and good practice in statistical analysis and reporting. Indian J Occup Environ Med 2006; 10: 5 10. 15) Cotes JE, Gilson JC. Effects of inactivity, weight gain and antitubercular chemotherapy upon lung function in working coal-miners. Ann Occup Hyg 1967; 10: 327 35. 16) Womack CJ, Harris DL, Katzel LI, Hagberg JM, Bleecker ER, Goldberg AP. Weight loss, not aerobic exercise, improves pulmonary function in older obese men. J Gerontol A Biol Sci Med Sci 2000; 55: M453 7. 17) Thomas PS, Cowen ER, Hulands G, Milledge JS. Respiratory function in the morbidly obese before and after weight loss. Thorax 1989; 44: 382 6. 18) Wise RA, Enright PL, Connett JE, et al. Effect of weight gain on pulmonary function after smoking cessation in the Lung Health Study. Am J Respir Crit Care Med 1998; 157: 866 72. 19) Chen Y, Horne SL, Dosman JA. Body weight and weight gain related to pulmonary function decline in adults: A six year follow up study. Thorax 1993; 48: 375 80. 20) Jubber AS. Respiratory complications of obesity. Int J Clin Pract 2004; 58: 573 80. 21) Golshan M, Nemat-Bakhsh M. Normal prediction equations of spirometric parameters in 799 healthy Iranian children and adolescents. [On-line]. [cited 2008 Oct 23]; Available from: URL: http://www.ams.ac.ir/ AIM/0033/golshan0033.htm. 22) Raju PS, Prasad KV, Ramana YV, Ahmed SK, Murthy KJ. Study on lung function tests and prediction equations in Indian male children. Indian Pediatr 2003; 40: 705 11. 23) Raju PS, Prasad KV, Ramana YV, Murthy KJ. Pulmonary function tests in Indian girls. Prediction equations. Indian J Pediatr 2004; 71: 893 7. 24) Pasiyan R. The correlation between chest circumference and forced vital capacity parameters on healthy adolescence with normal lung function. Medika 1997; 23: 359 61. 25) Schrader PC, Quanjer PH, Borsboom G, Wise ME. Evaluating lung function and anthropometric growth data in a longitudinal study on adolescents. Hum Biol 1984; 56: 365 81. 26) Mu HJ, Liu SW. National normal values of pulmonary function tests in P. R. China. Beijing: Chinese Xiehe Medical University and Beijing Medical University Publishing House; 1990. p.10 92.