INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Measurements of lung volumes are important components of pulmonary assessment in children and adolescents with lung disease. They are invaluable for providing objective documentation of the presence and severity of an abnormality, particularly for diseases that are associated with air-trapping and restrictive changes of the lung. Availability of appropriate reference values is essential for meaningful interpretation of the results of any pulmonary function test (1). There have been few such data for lung volume measurements in children. The published data are based almost entirely on studies of white children (2- 4). The reference equations obtained from white children may not necessarily be applicable to those who are nonwhite (5). Moreover, there are some differences in the body proportion in children of different ethnic background, particularly the difference in trunk length between the African and the white children (6, 7). It has been shown that prediction of lung function indices from sitting height rather than standing height would virtually eliminate the differences between the African and white children (8). Trunk length or sitting height, being the closest approximation of chest size of all the commonly used anthropometric parameters, may be a better predictor of static lung volumes than standing height. Confirmation of their usefulness as predictors of lung volumes would help to resolve the current difficulties in interpreting and following lung function in children whose standing height cannot be reliably recorded because of neuromuscular disease or limb and spine deformities.

The body proportion of Chinese children is unique compared with both the white and the black counterparts (9), therefore reference values derived from white children may not be appropriate for the Chinese. There is also extensive evidence to show that somatic development and hence body size of children changes over time, determined by exogenous factors that reflect the socioeconomic status of a population rather than its ethnic genetic constitution (10). Hence lung function values, which are dependent on body size, may demonstrate significant secular trends in developing societies. Lung volume measurements in healthy Chinese children have not been reported previously. As part of a larger study on lung function reference values, the purpose of the present project is to obtain reference prediction equations for static lung volume measurements in Chinese children and adolescents. We would also evaluate the usefulness of trunk length and sitting height as predictors of lung volumes.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The process of recruitment of subjects for lung function tests and the anthropometric measurements have been described in the accompanying study. Lung subdivision indices were recorded using a body plethysmograph (SensorMedics 6200 Automated Bodybox) according to the published standards criteria (11). After obtaining spirometric measurements, the subject was allowed to rest for 5 min before proceeding to the measurement of lung subdivisions. Briefly, the subject was taught the panting technique while breathing through the mouthpiece of the body plethysmograph. The door was then closed for a few minutes and the measurements including total lung capacity (TLC), residual volume (RV), and functional residual capacity (FRC) were made. The results, corrected for body temperature pressure saturation (BTPS), were scrutinized qualitatively by a respiratory physician (M.S.M.I.) before inclusion for statistical analysis. The study was approved by the institutional ethics committee.

Statistical Analysis

Regression analysis was applied to each of the lung volume indices and for each sex separately. The relationship between lung volumes and anthropometric variables was examined first. Various regression models including quadratic and power function, log-transformed and linear relationship were compared. For all the lung indices examined, the log-transformed models provided the best fit to the data, with the log-log model being marginally better than the log-linear model. Natural log-log models were therefore chosen as the basic format for evaluating the relationship between the dependent lung volume variables and the various independent variables.

The usefulness of various predictors of lung volumes was then compared. Age, standing height, weight, arm span, sitting height, suprasternal height, trunk length, body surface area (BSA) according to Du Bois (12), and body mass index (BMI) were considered as potential predictor variables. The dependent lung volume variables were first regressed individually against the independent variables. Stepwise multiple regression analyses were then used to determine which combination of parameters would fit the model best. Predictor variables were retained in the regression model only if their addition significantly improved the explained variance of the dependent variable. Wherever possible the most parsimonious model was chosen.

The final equations for the lung volumes were further studied by plotting the standardized residuals against predicted values for evidence of a curvilinear relationship. For regression models to be considered forming a satisfactory fit to the data, the residuals needed to be Gaussian in distribution and the equations were symmetrical and homoscedastic.

To allow the reference data to be used in situations in which standing height could not be recorded, regression equations with other anthropometric variables such as sitting height and arm span as the independent variables were compared with those using conventional standing height as the predictor variable.

All analyses were performed using Statistical Analysis System (SAS) Version 6.08 (13).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Children who met the study criteria are detailed in the accompanying study. This study summarizes the results of those who had satisfactorily completed the lung volume measurements. Of the 1,030 students who took part in lung function testing, 139 children (83 males) were excluded due to a multitude of reasons including symptoms of colds on the day of testing, scoliosis, non-Chinese parents, smoking history, and technical difficulties in performing lung function maneuver. The 112 children who were excluded because of technical difficulties were significantly younger (age 11.4 ± 2.9 yr) and smaller (standing height 144.5 ± 16.3 cm) than the remainder of the group (age 13.5 ± 2.9 yr, height 153.2 ± 13.5 cm, p < 0.001 for both parameters). A further 340 children did not have lung subdivision indices recorded because of time limitation during school lessons. Omission or recruitment of subjects for lung function tests was random apart from consideration of time limitation and recruitment of adequate number of subjects in all the sex and age groups. Of the lung subdivision data available, 551 test results were considered to be acceptable by the respiratory physician. Among them there were 219 boys and 332 girls.

Details of the anthropometric and lung volume data are shown in Table 1. The frequency distribution of age, height, and weight of these children is shown in Figure 1. In this study, boys were younger but taller than girls. Despite differences in body lengths, boys and girls had equivalent sitting height and trunk length. Much of the difference in height between the two sexes was the result of a difference in lower segment length. Boys and girls also had similar BMI.

View larger version (41K): [in this window] [in a new window]   Figure 1.   Frequency distribution of age and height, for the boys and girls who had complete static lung volume measurements.

For the lung volume measurements, boys had higher TLC (Figure 2) and RV (Figure 3) but similar FRC (Figure 4) compared with the girls. In this study, the difference in TLC and RV in Chinese children between boys and girls was not due to a difference in sitting height or trunk length between them.

View larger version (19K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Figure 2.   Predicted total lung capacity, TLC (liters), according to standing height of (A) boys and (B) girls. Curves show mean of study population and up to three standard deviations about the mean.

View larger version (20K): [in this window] [in a new window]   View larger version (20K): [in this window] [in a new window]   Figure 3.   Residual volume, RV (liters), according to standing height of (A) boys and (B) girls. Curves show mean of study population and up to three standard deviations about the mean.

View larger version (19K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Figure 4.   Functional residual capacity, FRC (liters), according to standing height of (A) boys and (B) girls. Curves show mean of study population and up to three standard deviations about the mean.

Using explained variance as an indicator of model fitting to the lung volume data, various anthropometric variables (standing height, suprasternal height, sitting height, trunk length) were compared. The values of coefficient of determination (R²) for standing height and sitting height were broadly similar (Table 2), suggesting that both measurements were satisfactory predictors of static lung volumes. Using multiple regression analysis, after allowing for standing height or sitting height, age was the next best parameter. Its inclusion to the regression models contributed to less than 3% of deviance for girls and 1% for boys. In this paper we opted for using standing height as the independent variable in our prediction equations for static lung volumes so that the findings could be compared with other populations of children.

Several regression models including various forms of data transformation had been examined. Natural log transformation of both the dependent lung volume indices and the independent predictor (log-log) provided the best fit to the data. The regression equations for the lung volume measurements are shown in Table 3. Plots of the mean predicted lung volumes together with their standard deviations about the mean are shown in Figures 2-4.

As sitting height and arm span are reasonably good predictors of plethysmographic lung volumes, prediction equations based on these variables are also included here (Table 4). The explained variance for these variables is either equivalent or close to that of standing height. They could be handy for deriving reference values of static lung volumes in situations in which standing height cannot be reliably obtained.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Compared with spirometric indices, data on static lung volumes in healthy children have been infrequently reported. Static lung volumes are technically more demanding and more time consuming to measure than spirometric indices. In addition, whether by gas dilution or body plethysmograph, these measurements require cumbersome equipment. Hence collection of reference values for these indices is logistically much more difficult to undertake than those of spirometry, which require only portable recording devices. To our knowledge this study is the first systematic evaluation of lung subdivisions in healthy Chinese children, from which precise reference equations are derived. Similar to spirometric measurements, the numbers of subjects at both ends of the age spectrum were small and there was also some variability in sample size over the ages, although, as stated previously, this should not significantly affect the curve fitting procedure and the reference values thus obtained.

Children of different ethnic background have different body proportions (6). It is well recognized that sex and ethnically appropriate reference values are essential for interpreting lung function results both clinically and in research studies (1). The findings in this study will serve to fill in the deficiency in reference values of static lung volumes for Chinese children and adolescents.

It is generally assumed that static lung volumes are related to chest size, with trunk length being its closest approximation among all commonly used anthropometric measurements. Using sitting height as a predictor, it has been found to eliminate the difference in lung function between black and white children (8). However, in our study subjects, despite the fact that boys had significantly higher TLC and RV compared with girls, their trunk length and sitting height were similar. The findings suggest that factors other than chest size may be important determinants of static lung volumes. In this study, only FRC was similar for the two sexes. As in other studies on white subjects, we find that standing height is a satisfactory predictor of lung volumes, equivalent to sitting height in the R² values for lung volumes. However, there were striking size-corrected differences in certain lung volumes between white and Chinese children.

Compared with the data on white children aged 4-19 yr reported by Rosenthal and coworkers (14), the residual volumes were generally lower, particularly for girls (Figure 5), whereas the TLC and FRC were more comparable except at the taller height ranges in boys, when TLC and FRC markedly increased in the white individuals (Figures 6 and 7). This pattern of TLC and FRC trend between Chinese and white children resembles our findings on dynamic lung volumes reported in the accompanying study. However, our findings are markedly different from that of Haluszka and coworkers (15) in white children over two decades ago, and from a previous study that showed that Chinese values of TLC and FRC in both children and adults were some 10-15% lower than those of white individuals (16). Although there is no recent published information on lung function in healthy Chinese adults, it is possible that the lung function of current generations of adults is still lower than that of white individuals.

View larger version (15K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   Figure 5.   Comparison of predicted RV curves derived from this study in (A) boys and (B) girls with those of studies of white individuals. I. Hong Kong 1996, II. Haluszka and Branski (15), III. Rosenthal and coworkers (14).

View larger version (14K): [in this window] [in a new window]   View larger version (13K): [in this window] [in a new window]   Figure 6.   Comparison of predicted TLC curves derived from this study in (A) boys and (B) girls with those of studies of white individuals. I. Hong Kong (1996), II. Haluszka and Branski (15), and III. Rosenthal and coworkers (14).

View larger version (15K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   Figure 7.   Comparison of predicted FRC curves derived from this study in (A) boys and (B) girls with those of studies of white individuals. I. Hong Kong (1996), II. Haluszka and Branski (15), and III. Rosenthal and coworkers (14).

The difference in residual volumes reported between our Chinese children and white children requires explanation. Measurement of RV requires the subject to exhale maximally. Healthy children have low values. High values may indicate airway disease or inadequate effort of the part of the subject during lung measurements. The difference observed between studies may be due to a number of reasons. First, all our children had been vigorously screened for respiratory symptoms and disease. The low residual volume values may be an indication of the success of our screening procedure. Second, the difference may be physiological in that Chinese children may have constitutionally "better" airway function. However, we can find no evidence to support this conjecture both clinically or in the medical literature. We are inclined to think that the difference reflects the vigorousness of our screening procedures and the exactness of our investigating team in adhering to the stringent standards of lung volume measurements.

Contrary to the findings of Rosenthal and coworkers (14) on white children and Connett and coworkers (17) on Singaporean Chinese subjects, we did not find an abrupt upward shift in the relationship between TLC and height. The latter was purported to occur as a result of developmental changes during puberty in boys (18, 19). Similarly, we have not observed this pattern in our spirometric data (reported in the accompanying study). We have not specifically examined pubertal stage in our subjects. A recent report from Hong Kong suggested that puberty, as indicated by a testicular volume of > 12 ml in boys, occurs at 14 (95% confidence interval, 11- 17.5) yr, whereas in girls the age of menarche occurs at 12.5 (95% confidence interval, 10-15) yr (20). Our study subjects spanning 7-19 yr old in age in boys and girls would therefore have included adolescents in various stages of puberty, which corresponded to a height of 145 to 170 cm in Chinese boys and 140 to 160 cm in girls. The shift in lung volumes observed in boys in the other studies has been attributed to a disproportionate increase in thoracic dimensions compared with standing height occurring at around puberty. A disproportionate age- related increase in trunk length or sitting height (the nearest indication for chest size) has not been found in our study subjects or in a large study on growth in Hong Kong children (9).

Using explained variance as an indicator of model fitting, the sitting height and standing height are equivalent predictors for lung volumes, with arm span being the next best. The R² values of the latter are generally no more than 10% below those for the standing height. Hence application of these anthropometric measurements would help to resolve the current difficulties in interpreting and following up lung function in children whose standing height cannot be obtained reliably due to neuromuscular disease or limb and spine deformities. The comparison of the lung volume values obtained with our prediction equations and those obtained with the Cotton Dust Standard (Knudson's race-adjusted) data set, as available in the lung function system (Operator's Manual, Sensor Medics 6200), showed significant difference (Table 5) and highlights the importance of using originally derived population-specific update data sets.

In summary, Chinese children have size-corrected static lung volumes (TLC and FRC) initially equivalent to their white counterparts but the values diverge at taller height ranges in boys. Chinese boys have larger lung volume (TLC) than girls but their sitting height and trunk length are comparable. Standing height and sitting height are equivalent predictors of static lung volumes. In situations in which standing height cannot be measured, sitting height or arm span is an adequate alternative. The reference lung volume values for Chinese children are presented.