Paola Primatesta, Anne McMunn, Marion Brookes
|
SUMMARY
|
This chapter deals with lung function measurements obtained from informants aged 7 to 74. Lung function values play an important role in population screening and classification of lung disease. Measurements are based on spirometric tests: spirometers are devices that make possible accurate measurements of inspiratory and expiratory airflow. They are routinely used in clinical settings to reveal impairment and to monitor changes - such as improvements as a result of interventions. These tests are now applied more widely in the assessment of respiratory health of the general population.1 Respiratory function indices are known to predict future mortality: this is not restricted to mortality from respiratory disease, but also applies to cardiovascular disease and all-cause mortality.2,3,4,5
Various tests may be used in the quantitative evaluation of lung function. Measurements are made of the airflow rate (expressed as the volume expelled over a certain period of time) and changes in volume of the respiratory system. People with current airflow limitation (for example due to asthma or emphysema) nearly always show some reduction. Nevertheless, since in asthma the airways obstruction is by definition reversible, lung function measurements can be close to normal between attacks, especially at the onset of the disease or as the patients respond to treatment.
As in the 1995 survey, three measures were selected for use in this survey:
These measurements are determined largely by the size of the lungs, which in turn depends on age, sex, stature, body mass and composition, ethnic group, and personal habits such as smoking. In both sexes, the variation of respiratory function with age and anthropometric measurements has been extensively studied, and it is generally agreed that standing height and age are the characteristics that correlate best with lung volume measurements.
Estimates of FEV1, FVC and PEF are presented in this report by age and height groups, with different cut-offs in men (less than 170 cm, 170 but less than 180 cm, 180 cm and over), women (less than 160 cm, 160 but less than 170 cm, 170 cm and over), and children (subdivided into height tertiles). For adults only, FEV1 and FVC estimates are reported by region, social class, smoking status and respiratory symptoms. Relative lung function levels, calculated in relation to predicted values (see section 4.2.2) are also presented in this chapter. The results for FEV1 are also compared with those obtained in the Health Survey for England 1996.6
4.2 METHODS AND INTERPRETATION OF THE RESULTS
4.2.1 Equipment and performance
All informants aged 7-74 were eligible for the spirometric tests, except those who were pregnant, had abdominal or chest surgery in the preceding three weeks, or had been admitted to hospital with heart disease or stroke in the preceding six weeks.
The equipment used was the Vitalograph 'Escort Spirometer', a portable spirometer with a calibration device. The manufacturer reports the accuracy of volume measurement as ± 3% or ± 0.05 litres, whichever is greater, and of flow measurement as ± 5%. These standards comply with those of the ECSC (European Community for Steel and Coal) and the ATS (American Thoracic Society).7
The detailed measurement protocol is contained in Appendix B. Briefly, the measurement technique requires a maximum inspiration followed by a forced expiration (for as long as possible) into the instrument, which then calculates and displays FEV1, FVC and PEF.8 The test procedures were demonstrated by the nurse using a detached mouthpiece. While in a standing position (unless chairbound), informants were asked to perform a forced inspiration, then an expiration with maximal effort without hesitation. A test was considered technically satisfactory if it was without the following: 1) an unsatisfactory start of expiration; 2) laughing or coughing, especially during the first second; 3) holding the breath in; 4) a leak in the system; 5) an obstructed mouthpiece.
Lung function tests require the informant to learn the technique and to apply considerable effort to the task. In order to allow for invalid attempts and to give informants an opportunity to practise, they were asked to carry out the task up to five times in total. The data presented here are based on the highest values for each of the three parameters (FEV1, FVC, and PEF) from any of the five technically satisfactory tests. The values used for a particular informant may not all come from the same test.
The data presented here should be interpreted with caution. Lung function measurements are affected by the way the informant performs the task. This in turn is affected by the extent to which the nurse was able to explain the task and to motivate informants to perform it as well as they possibly could. In spite of the careful development of protocols, extensive briefing and training of nurses, and the use of repetitions of the measurement to allow familiarisation, there remains a risk that a suboptimal ability to perform the task may result in a possible underestimate of the readings. However, there is no reason to believe that this possible underestimation varies between sub-groups of population, and internal comparisons, such as between regions, are still likely to be valid. In comparisons of these results with those from other studies, possible differences due to choice of equipment, measurement procedures and study settings should be taken into account.
4.2.2 Predicted values and relative lung function levels
The interpretation of the results of lung function tests usually relies on comparison with reference values. This is required in order to determine whether or not a test result can be considered 'normal'. Unfortunately, unlike (for example) blood pressure measurements, there is no consensus as to normal values for groups of individuals. Instead, it is customary to use reference values, predicted for a normal 'healthy', non-smoker individual with similar characteristics. Usually these values are derived from a reference population by means of multiple regression equations and their residual standard deviation, age and height being the key predictor variables.
By comparing the predicted values for healthy adults with the observed values for adults (of the same sex) who have the same set of predictor characteristics (age and height), a measure of relative lung function is obtained. A reading substantially below the predicted value in principle indicates a potential impairment of the lung function. An aggregate comparison of predicted and observed readings enables the lung functions of sub-groups to be compared: groups with collective values substantially below those predicted would thus appear to have a propensity to impairment of lung function.
It is clearly important that the reference values used should be appropriate to the particular population being studied. As in the 1995 report, the values chosen were the ECSC (European Community for Steel and Coal) reference values for persons of European descent, as recommended by the European Respiratory Society in 1993.9
Relative lung function levels presented in this report compare the values observed in the sample with predicted values. A four-level classification of relative lung function has been used, as follows:
4.2.3 Response to lung function measurements
Adult response
Valid readings were obtained from 97% of men and 95% of women visited by a nurse. The remainder either did not co-operate, were not eligible, or managed no technically satisfactory tests (see 4.2.1). Of those seen by a nurse slightly fewer women (in most age groups) completed lung function tests, partly due to the exclusion of pregnant women. Men aged 65-74 had a slightly lower measurement rate than younger informants.
Child response
Among the 1,997 children aged 7-15 who were visited by a nurse 97% of boys and 96% of girls provided a valid result. The main reason for the absence of a valid result was that all tests performed were technically unsatisfactory (3% in girls, 1% in boys). In girls a slightly higher response was achieved among those aged 10-15 than among those aged 7-9.
4.3 LUNG FUNCTION BY HEIGHT, AGE AND SEX
4.3.1 FEV1
Longitudinal studies have shown that ventilatory function reaches a maximum at around 12 years of age in girls and 14 in boys.10 In this survey, mean FEV1 was higher in males than in females, with differences increasing with age up to the late teens. The peak was reached in late teens in boys and early teens in girls. Afterwards FEV1 decreased steadily with increasing age; the decrease was greater for men than for women, so that the gap between the sexes was reduced in the elderly.
Figure 4A
Figure 4A Mean FEV1, by age and sex
Overall, mean FEV1 (in litres) was 3.8 in men and 2.7 in women; and 2.5 in boys and 2.3 in girls aged 7-15.
Mean FEV1 was greatly influenced by height: both adults and children in the tallest groups had higher FEV1 than people of corresponding age in the middle or shortest group. Among men of all ages mean FEV1 increased from 3.1 litres in the shortest group to 4.3 litres in the tallest group. Similarly, among women it increased from 2.3 to 3.3 litres.
4.3.2 FVC
Like FEV1, FVC increased with age during childhood and adolescence, reached a peak in young adulthood and then declined with increasing age. Mean FVC (in litres) was 4.7 in adult men and 3.3 in women. It was 2.9 litres in boys and 2.7 litres in girls. It increased with increasing height: among men of all ages from 3.9 litres in the shortest group to 5.4 litres in the tallest, and in women from 2.9 litres to 4.0 litres.
4.3.3 PEF
Variations in PEF by age, sex and height were somewhat different from those for FEV1 and FVC. Men had higher PEF (554 l/min) than women (376 l/min), but the differences between boys and girls were less marked: mean PEF was 319 l/min in boys and 311 l/min in girls. Values were higher for boys except at age 10-12 where they were higher for girls. Mean PEF peaked at a later age than FEV1 and FVC; it was highest in both sexes at age 25-34 (615 l/min in men and 421 l/min in women) and it decreased thereafter. Like FEV1 and FVC, PEF increased with height: in men it rose from 473 l/min in the shortest group to 611 l/min in the tallest group, and in women from 337 to 434 l/min respectively.
4.4 ADULTS' FEV1 AND FVC BY REGION, SOCIAL CLASS, SMOKING STATUS AND RESPIRATORY SYMPTOMS
4.4.1 By region
Some differences were seen between regions, for both FEV1 and FVC. Among men and women aged 16-44 and in the shorter group of those aged 45-74, those living in Greater Glasgow had the lowest mean values, while in both age groups those living in Highland & Islands and Grampian and Tayside had the highest values.
4.4.2 By social class
Mean FEV1 did not show marked differences by social class in the younger age group (16-44). This was true in both sexes. Among those aged 45-74 on the other hand, mean FEV1 was generally higher in non-manual than in manual social classes, with a general decrease from those in Social Classes I & II to those in Social Classes IV & V: among men it decreased from 3.1 litres in Social Classes I & II to 2.7 litres in Social Classes IV & V in the shorter group (<175 cm) and from 3.6 to 3.2 litres in the taller group; and in women it decreased from 2.3 to 2.0 litres (height group <165 cm) and from 2.7 to 2.4 litres respectively.
Table 4.11, Figure 4B
A similar pattern to that for FEV1 was seen for social class variations in FVC: mean FVC was higher among non-manual than manual social classes in both sexes among those aged 45-74, while the difference were smaller among those aged 16-44.
4.4.3 By cigarette smoking status
There was a general tendency for FEV1 to decrease progressively from those who had never regularly smoked to those who currently smoked 20 or more cigarettes a day. This was true for both sexes and both age groups, although the differences were most marked in men aged 45-74 (mean FEV1 decreased from 3.2 to 2.8 litres in the shorter age group and from 3.8 to 3.2 litres in the taller age group). The fact that the negative effect of smoking on FEV1 was more marked among older people supports the theory of a cumulative effect of lifetime smoking on lung function.
FVC showed a similar pattern to FEV1, but the differences between cigarette smoking groups tended to be small.
Figure 4B Adults' mean FEV1 by age group, height group, social class of chief income earner and sex
4.4.4 By respiratory symptoms
Tables 4.15 and 4.16 show mean FEV1 and FVC for selected respiratory symptoms (ever wheezed, wheezed in the last 12 months, cough/phlegm for at least 3 months per year) and doctor-diagnosed asthma. In general, those who had wheezing symptoms (ever or in the last 12 months) and cough and phlegm showed lower mean FEV1 and FVC than those without the symptoms. This was especially true for those aged 45-74, while the association was weaker in the younger age group. Among men aged 45-74 mean FEV1 was 2.4 litres in those who wheezed in the last 12 months and 3.0 litres in those who did not in the shorter group, and 2.9 and 3.6 litres respectively in the taller group. In women the corresponding figures were 1.8 and 2.2 litres; and 2.2 and 2.6 litres respectively. Similar differences were observed for FVC.
Doctor-diagnosed asthma was also associated with a lower mean FEV1, especially in the older age group. Among those aged 16-44 the differences in FEV1 were smaller.
To quantify the association of respiratory symptoms and smoking status with lung function, adjusting for age and height, a linear regression was performed. The variables corresponding to the questions on wheezing in the last 12 months, cough/phlegm production and doctor-diagnosed asthma were included in the analysis, as well as smoking status and social class. FEV1 was the dependent variable, chosen as the best single indicator of airflow limitation.
In both sexes, subjects with respiratory symptoms and those with asthma had significantly lower FEV1 than those without the symptoms. For example, FEV1 among men with wheezing symptoms in the last 12 months was estimated to be 0.25 litres lower than among those without the symptoms; and women with wheezing symptoms in the last 12 months had FEV1 0.17 litres lower than those without the symptoms. Both smoking and social class also showed an independent effect on FEV1. Men who smoked more than 20 cigarettes a day had FEV1 0.23 litres lower than those who did not smoke, while in women it was 0.17 litres lower. Men in Social Classes IV and V had an estimated reduction in FEV1 of 0.15 litres compared to men in Social Classes I and II and women had a reduction of 0.11 litres.
4.5 ADULTS' RELATIVE LUNG FUNCTION
As already noted (see 4.2.2) relative lung function levels presented here compare the values observed in the sample for FEV1 with the predicted values as recommended by the ECSC.9 The ECSC predictive values were derived for persons of European descent, non-smokers with no reported respiratory disease. Therefore the results presented in Table 4.18 are presented, as well as for the whole population, for the group of 'healthy non-smokers'. In this context these are people who have never smoked cigarettes regularly and who did not report any respiratory disease that could compromise their ventilatory function, that is they did not wheeze in the last 12 months, did not have cough/phlegm up to 3 months every year and did not report doctor-diagnosed asthma. Using this definition, those classified as healthy non-smokers were 34% of the (unweighted) sample.
Since the reference curves are for people of European descent, informants from ethnic minority groups should be not be included in the analysis. It is possible to apply conversion factors and derive reference values for different ethnic groups, but this was not attempted here given the small number of informants in any such group. The relative lung function was computed including and excluding those of origin other than 'white European': given the very small number of informants from these ethnic groups (65 in total) the results were very similar and Table 18 therefore refers to all informants with a valid height and lung function measurement, irrespective of their ethnic origin.
Looking at the healthy population, given that their characteristics should be comparable with those of the healthy population for which the ECSC values were derived, one would expect to find half of the cases at or above the reference values and 5% of them at more than 1.64 standard deviations below the reference value. In the healthy group the proportions of men and women at or above the reference value were in fact higher than expected: 62% of men and 62% of women were at or above the reference value. These proportions varied considerably with age, being higher in the middle of the age range. Overall the proportion with 'low' levels was close to the expected, but differences by age were observed.
Considering the whole population, overall over half of the informants had results equal to, or in excess of, predicted values: 53% of men and 55% of women were in this group. This proportion increased with age up to age 35-44; afterwards it decreased continuously to 37% of men and 42% of women aged 65-74. The proportion of the total population with FEV1 levels more than 1.64 below the predicted value ('low' levels) was 9% in men and 8% in women. These proportions increased markedly with age: in the oldest age group there were 22% of men and 12% of women with 'low' FEV1 levels.
4.6 COMPARISON OF FEV1 BETWEEN SCOTLAND AND ENGLAND
No substantial differences were observed between Scotland and England and Northern England in mean FEV1 in both sexes and all age groups including children. This was also true for the proportion of the population with 'low' levels of FEV1 (data not shown).