Increased Susceptibility to Lung Dysfunction in Female Smokers1- 3

VUE CHEN,4 SANDRA L. HORNE, and JAMES A. DOSMAN5

Introduction

The relationship between cigarette smoking and chronic obstructive pulmonary disease (COPD) has been well documented. The 1984 Surgeon General's Report on the Health Consequences of Smoking indicates that despite more than 30 yr of intensive investigation, only cigarette smoking and aI-antitrypsin deficiency are established causes of clinically significant COPD in the absence of other agents (1). Data from longitudinal, cross-sectional, and case-control studies show that in comparison with nonsmokers, cigarette smokers have higher death rates for chronic bronchitis and emphysema; higher prevalence and incidence rates for chronic bronchitis, emphysema, and obstructive airways disease; and higher frequencies of respiratory symptoms and lung function abnormalities (2). The results hereinreported suggestthat sex may modify the influence of smoking on the development of COPD. Although many studies have confirmed the detrimental effect of cigarette smoking on lung function, male-female differences in the responses of the small airwaysto smoking are still controversial issues (1). A three-city study reported by Buist and coworkers (3) shows that the prevalence of FEVI/FVC abnormalities was more than twofold higher in female smokers than in male smokers, but almost the same in nonsmokers. Similar sex-related differencesin smoking effects on the prevalence of abnormalities of closing volume (CVlVC) and the slope of phase III of the single-breath nitrogen test (L\N2/L) were also observed in that study (3) as well as in an earlier report (4). Other studies have suggested that the reversibleeffect of smoking was larger for men than for women (5-8), lacked significance (9), or showed no difference between the sexes (10). This study presents the findings of a cross-sectional investigation, focusing on the interaction of sex and smoking on pulmonary function. 1224

SUMMARY The Interaction between sex and smoking habits on pulmonary function was examined among 1,149 adults 25 to 59 yr of age In a rural community In Saskatchewan. Pulmonary function tests Included FVC, FEY" maximal mldexplratory flow rate (MMFR), the slope of phase III of the single-breath nitrogen test (Ll.N,IL), and closing volume as a percent of vital capacity (CVNC). The data show that after fixing the effects of age, height, and weight by analysis of covariance, the adjusted means of Ll.N,/LIn nonsmokers, ex-smokers, and current smokers were 0.92, 1.10,and 1.60% In women and 0.97, 1.05, and 1.23% In men, respectively. The difference In the adjusted means for Ll.N,/L between smokers and nonsmokers was larger In women than In men, 0.67% versus 0.26%, respectively. MUltiple multivariate analyses show that the regression slopes for the residuals of FEY" MMFR, and Ll.N,/Lversus pack-years were significantly different between men and women. The regressions of FEY, and MMFR decreased and the regression of Ll.N,/L Increased.wlth Increasing packyears more rapidly In women than In men. The combined effect of sex and pack-years on pUlmonary function was not significant for ex-smokers. These data suggest that cigarette smoking may be more detrimental In Its effects on lung function In women than In men. AM REV RESPIR DIS 1991; 143:1224-1230

Methods Study Population All adults (1,274)25 to 59 yr of age who were living in the town of Humboldt, 110kilometers east of Saskatoon, Saskatchewan, were selectedfor study in 1977. Humboldt is a commercial and distribution center for a mixed agricultural area. The town was selected because of its stable population, lack of industrial air pollution, and history of cooperation in previous community health surveys sponsored by the Saskatchewan AntiThberculosis League.A more detailed description has been published elsewhere (11).

Pulmonary Function Tests The tests included spirometry and the singlebreath nitrogen test, and were carried out according to criteria established by the National Heart and Lung Institute (12, 13). The forced expired maneuver was performed with a Godart 8-L water spirometer until three acceptable tracings to a maximum of five attempts were obtained in which the FVC was within 50,70 of the maximal value. The best FVC and FEV I, and maximal midexpiratory flow rate (MMFR), which came from the tracing with largest FVC, wereused in this analysis. The single-breath nitrogen test, which was performed prior to the forced expired maneuvers, was performed until three satisfactory tests were obtained. Tracings wereconsidered acceptable if within tracing vital capacities agreed within 5% and between tracings vital capacities agreed within 10%. The slope of phase III (LlN2/L) and closing volume

(CV/VC) were determined as previously described in detail (11, 14). Values were corrected to body temperature and pressure saturated with water vapor (BTPS). Standing height and weight were measured at the same time.

Questionnaire Administration A modified National Heart, Lung, and Blood Institute questionnaire was self-administered in each study subject's home after delivery by an interviewer recruited by the Saskatchewan Anti-Thberculosis League. The questionnaire included information on demographic factors, smoking habits, occupational history, respiratory symptoms, and previous chest illness. In this analysis, subjects who werecurrently smoking cigarettes, pipes, or cigars were defined as smokers. Subjects who had for-

(Received in original form May 22, 1990 and in revised form December 27, 1990) I From the Centre for Agricultural Medicine and the Division of Respiratory Medicine, Department of Medicine, University of Saskatchewan, Saskatoon, Saskatchewan, Canada. 2 Supported by the Saskatchewan Lung Association and the Department of National Health and Welfare, Government of Canada. 3 Correspondence and requests for reprints should be addressed to Yue Chen, M.D., Ph.D., Centre for Agricultural Medicine, Wing 3E, University Hospital, Saskatoon, Saskatchewan, Canada S7N OXO. 4 Recipient of a Research Fellowship from the Saskatchewan Health Research Board. S Scholar of Health and Welfare Canada.

INCREASED SUSCEPTIBILITY 10 WNG DYSFUNCTION IN FEMALE SMOKERS

merly smoked regularly but had quit smoking for at least 6 months at the time of the study were defined as ex-smokers. Subjects who had never smoked any kind of tobacco regularly or had smoked a total of less than one-half pack-year (i.e.• less than one pack daily for 6 months) were defined as nonsmokers.

Statistical Methods The difference in the smoking effect on pulmonary function between men and women is of primary interest in this analysis. Multivariate analysis of variance (MANOVA) was used to examine the interaction of smoking habits and sex on pulmonary function tests. Five pulmonary function indices were considered simultaneously. MANOVAis preferable to multiple univariate analyses under these circumstances because multiple univariate analyses ignore the interdependencies among the pulmonary function tests, and also do not control for Type I error (15). The subjects were classified into three groups according to their smoking status: nonsmokers, ex-smokers, and smokers. Analysisof covariance (ANCOVA) was carried out to test the significance of the main effects of smoking status and sexand the combined effect of these two factors on an additive scale and to adjust for the effects of covariates including age, standing height. and weight. After this, the number of pack-years was calculated for each person to estimate his or her total tobacco consumption. One pack-year is equal to 20 cigarettes smoked per day for a year. Pack-yearswerecalculated for the pipe and cigar smokers by equating one pipeful of tobacco or one cigar to one cigarette.There were no cigar or pipe smokers among the female smokers, and less than 3070 of the male smokerswerecigaror pipesmokers. MANOVA was used to do multivariate multiple regression analysis. In order to compare and clarify the difference in effect of pack-years on pulmonary function between women and men, the residual FVC, FEV" MMFR, AN,/L, and CYIVCof the two sexeswerecalculated. The prediction equations for pulmonary function of each sex were derived from the data of the entire population. From these equations, the proportion of the total variation, which is explained by age, height, and weight, was determined. A residual is the difference between an observed value and the value predicted. For the multivariate multiple regression analysis, there werefive dependent variables for residuals of pulmonary function: rFVC, rFEV" rMMFR, rAN,/L, and rCYIVC. We used interaction of packyears with sex to assess whether there were sex differences in pulmonary function in response to tobacco smoking. Results Ofthe 1,274eligible study subjects, 1,202 participated, giving a compliance rate of 94.3070. Among them, 1,181 (98.3%) had satisfactory spirometry measurements,

1225 TABLE 1 MEAN VALUES OF PULMONARY FUNCTION BY SEX AND AGE

FVC (L)

Age (yr)

MMFR (Lis)

FEV, (L)

AN,/L (%/L)

CVNC (%)

n

Mean

SO

Mean

SO

Mean

SO

Mean

SO

Mean

SO

138 57 84 63 79 62 61

5.52 5.13 5.25 4.94 4.68 4.52 4.31

0.75 0.73 0.72 0.76 0.77 0.81 0.80

4.49 4.12 4.18 3.90 3.72 3.52 3.32

0.61 0.55 0.63 0.58 0.58 0.64 0.72

4.90 4.83 4.58 4.16 4.11 3.61 3.40

1.40 1.74 1.31 1.41 1.26 1.14 1.29

0.69 0.77 0.82 0.95 1.16 1.14 1.61

0.28 0.27 0.42 0.60 0.77 0.67 1.58

12.45 14.20 16.26 18.35 19.70 20.43 21.98

4.63 4.76 4.61 7.51 3.98 4.71 5.14

128 63 94 64 85 82 89

3.81 3.82 3.75 3.70 3.43 3.35 3.01

0.67 0.57 0.55 0.56 0.48 0.58 0.50

3.22 3.17 3.08 2.99 2.71 2.65 2.39

0.53 0.48 0.45 0.48 0.41 0.45 0.43

3.80 3.75 3.56 3.25 2.79 2.79 2.63

0.99 1.04 1.21 0.91 0.92 0.83 0.99

0.98 1.09 1.05

0.40 0.49 0.44 0.86 1.18 1.21 1.93

11.83 12.21 15.34 15.99 17.65 18.69 19.29

5.62 5.75 5.67 4.24 6.41 5.17 7.47

Men

25-29 30-34 35-39 40-44 45-49 50-54 55.,.59 Women

25-29 30-34 35-39 40-44 45-49 50-54 55-59

1.18

1.64 1.51 2.04

Definition of abbreviations: MMFR = maximal midaxpiratory flow rate; l'IN,IL = the slopa of phase III 01the singla-breath nitrogen test; CVNC = closing volume as a parcent 01 vital capacity.

and 1,149 (95.6%) had satisfactory singlebreath nitrogen tests. This analysis is based on the data of 1,149subjects who had satisfactory measurements for all these tests. Among the 1,149subjects, 544 (47.3%) were men and 605 (52.7%) were women. The mean ages were similar, 40.2 (± 10.3) yrin men and 41.5 (± 10.5)yrin women. It can be seen in table 1 that the mean values of FVC, FEV 1 , and MMFR decreased and the mean values of AN 2/L and CVlVC increased with increasing age

both in men and women. There was no significant difference in age distribution between men and women (X2 = 6.270, df = 8, p >0.30). The pattern of decrease or increase in pulmonary function versus age did not change after adjustment for the effects of height, weight, and smoking by analysis of covariance (figure 1). The prevalence of tobacco smoking was higher in men (44.30/0 current smokers and 30.4% ex-smokers) than in women (35.5 and 18.2%, respectively). Among

MMFR

•

LIe

L

.... - .. ----

...........

....

.+--~-~--~-~-~----,

H-

II·

40-

...

....

.+--~-~--~-~--~----,

H-

Age group (yN")

H-

4'~

40-

Age group (YN,,")

CN/VC

..

- -' -

II

t.I

...

---

_. - ... -_._."-"

I.

_

~.

w

__

+ - - - - -,.,. .....

.+----~--~-~-~----, 364046so1066-"

-

Ago group (years)

,.

.

...

.+---r--~--~-~-~-~

-

Fig. 1. Comparison of trend of height-weight-smoking-adjusted means of lung function with increasing age between men (dotted line) and women (solid line).

1226

CHEN, HORNE, AND DOSMAN

TABLE 2 SIGNIFICANCE TEST OF EFFECTS OF SEX AND SMOKING STATUS AND THEIR INTERACTION ON PULMONARY FUNCTION TESTS· Sex

FVC. L FEV" L MMFR. LIs ~N"L, %/L CVNC. %

Smoking Status

Sex and Smoking Statust

F

P

F

P

F

P

245.320 220.202 64.396 2.751 2.542

0.000 0.000 0.000 0.097 0.111

3.680 7.417 1.267 31.658 22.839

0.026 0.001 0.282 0.000 0.000

1.909 1.817 1.312 6.183 0.166

0.149 0.163 0.298 0.002 0.647

For definition of abbreviations. see table 1. • Analysis of covariance. Covariates included age. height, and weight. t Interaction term for sex and smoking status.

female subjects, smokers were younger and nonsmokers were older than among male subjects (average age in nonsmokers: 37.3 yr in men, 42.6 yr in women; ex-smokers: 39.2 yr in men, 39.6 yr in women; current smokers: 43.6 yr in men; 41.3 yr in women). The effects of sex and smoking status and their interaction were tested by ANCOVA. The results of significance tests after adjustment for age, height, and weight are listed in table 2. The combined effect of sex and smoking status on AN2/L was statistically significant (F = 6.18, p = 0.002). The adjusted means of AN2/L (OJo/L) in nonsmokers, exsmokers, and current smokers were 0.92, 1.10, and 1.60 in women and 0.97, 1.05, and 1.23 in men, respectively (figure 2). The difference in adjusted means of AN2/L between smokers and nonsmokers was much larger in women than in men (0.67 versus 0.26), suggesting that smoking is more detrimental in its effects on pulmonary function, as reflected by AN2/L, in women than in men. In this study, the female smokers smoked less (current smokers: 17.1 ± 11.6 pack-years; ex-smokers: 10.6 ± 9.7 packyears) than did the male smokers (current smokers: 22.3 ± 16.1pack-years; ex-

smokers: 20.0 ± 16.4 pack-years). For each person in the study population, the predicted pulmonary function values, FVC, FEV.. MMFR, AN 2/L, and CVlVC, were calculated from the prediction equations, which explained the variations of age, height, and weight, derived from the data of the total population. The predicted values were then subtracted from the observed values to produce pulmonary function residuals. Multivariate multiple regression analysis shows that the interaction of sex and pack-years of smoking had a significant effect on the residual of FEV 1 (t = -1.916, P = 0.0556), MMFR (t = - 2.041, p = -0.0414), and AN 2/L (t = 5.912, P < 0.0001) (table 3), further indicating that smoking affects men and women differently. Sex-specific regression analysis showed the same slopes for residual pulmonary function versus pack-years as those from the sex-pack-year interactive model. Pack-years of smoking were significantly related to FVC, FEV .. MMFR, AN 2/L, and CVlVC in women, but only to AN 2/L and CVlVC in men (table 4). The regressions for the residual values of FEVland AN2/L plotted as a function of pack-years are shown in figure 3. The

Adjusted mean ('lIo/LI

1.8 1.6

Fig. 2. Combined effect of sex and smoking status on the slope of Phase III of the single·breath nitrogen test.

1.4

1.2 1

0.8 0.8 0 .4 0.2

o Males

Females

regression of FEV 1 decreased, and the regression of AN2/L increased with increasing pack-years more rapidly in women than in men. The effects of each packyear of smoking on age-height-weightadjusted FEVland AN2/L was - 0.0062 Land 0.0302 % for women, and - 0.0020 Land 0.0078% for men, respectively. The difference in the interaction of sex and number of pack-years of smoking between current smokers and ex-smokers was also examined by multivariate multiple regression. The smoking status factor produced two dummy variables: current smoking status (CSS) and exsmoking status (ESS). In the analysis, two three-way interactions: "sex-pack-yearCSS" and "sex-pack-year-ESS" were included, which indicated the effect of pack-years related to sex in current smokers and in ex-smokers separately. The analysis shows that the "sex-pack-yearCSS" interaction term was significantly related to reduction of the rFEV , (t = -1.681, p = 0.0931) and rMMFR (t = - 2.131, P = 0.0333), and the increment of rAN2/L (t = 5.288, p < 0.0001), but the "sex-pack-year-ESS" term was not. The results (table 5) signify that the ordinal interaction of sex and pack-years on pulmonary function is significant for smokers but not for ex-smokers. The regression slopes for residual values of FEV , and AN2/L versus pack-years by smoking status are shown in figure 4. It can be seen that the lines for female smokers are steeper than those for male smokers. Although there is a somewhat steeper slope for rAN2/L for female ex-smokers than for male ex-smokers, the sex-pack-yearinteraction is not significant. Discussion Our results suggest that cigarette smoking may be more harmful in its effects on pulmonary function in women than in men. The data show a significant combined effect of sex and pack-years on FEV.. MMFR, and AN2 / L. The interaction of sex and smoking habits on pulmonary function is independent of the effects of age, height, and weight. These findings cannot be attributed to the presence of respiratory symptoms including cough, sputum, and wheeze. When these factors were examined with multivariate analyses, none of them were found to change the pattern of results. . These findings are consistent with the results of a three-city study by Buist and coworkers (3) that revealed that there was an apparent sex difference in the prevalence of abnormalities in smokers but not

INCREASED SUSCEPTIBILITY TO WNG DYSFUNCTION IN FEMALE SMOKERS

1227 TABLE 3

COMBINED EFFECT OF SEX AND NUMBER OF PACK-YEARS ON THE RESIDUALS (r) OF PULMONARY FUNCTION" rFEV,

rFVC

Sex Pack-years Sex x pack-years Multiple R

rMMFR

raN,IL

CVNC

b

Sb

b

Sb

b

Sb

b

Sb

b

Sb

0.0165 -0.0017 -0.0028 0.0723

0.0442 0.0015 0.0025

0.0169 -0.0020 - 0.0041 0.1129

0.0375 0.0012 0.0022

0.0421 -0.0034 -0.0106 0.3552

0.0907 0.0030 0.0052

-0.1190 0.0078 0.0226 0.2921

0.0665 0.0022 0.0038

0.1572 0.0502 0.0300 0.1625

0.4097 0.0135 0.0235

Definition of abbreviations: b = partial regression coefficient;

Sb = standard error of partial regression coefficient. For other definitions, see table 1.

• Multivariate mUltiple analysis.

TABLE 4

in nonsmokers; ~N2/L and FEV1/FVC were abnormal more than twice as often in women than in men as shown by analysis of cross-sectional data of 681 adults Women 25 to 54 yr of age. The UCLA study of b Sb P two areas based on information from -0.0045 0.0018 -2.458 0.0142 5,776 residents 25 to 59 yr of age who 0.0015 -0.0062 -4.016 0.0001 were current smokers or never smokers 0.0036 -0.0140 -3.923 Q.OOOl showed that the age-adjusted prevalence 0.0036 0.0304 8.481 0.0000 of participants with an FEF25-75 below 0.0207 3.877 0.0802 0.0001 35% of expected was greater among female smokers than among male smokers. For example, in the Lancaster study area, 4.60/0 of female heavy smokers had a FEF25-75 below 35% compared with 2.2% of male heavy smokers, but among FEV 1 never smokers, only 0.1% of either sex fell below this cutoff (16). The ageadjusted prevalence of FEV 1 below 50% showed a similar pattern in the Lancaster study area. Another cross-sectional study based on information from 4,690 Cau-------- M casians 7 yr of age and older from three separate communities by Beck and coworkers (9) found that significant differences for the residuals of FEV to maximal expiratory flow at 50% of vital caF pacity (V!11ax50), and 25% of vital capacity (Vmax25) among smoking cate20 30 40 150 60 gories (non-, ex-, light, and heavy smokPackyears ers) began in female subjects at a younger age (15 to 24 yr) than in male subjects (40 to 45 yr). In a study of 504 men ll.~/L and women 24 to 55 yr of age by Manfreda and coworkers (17) in western Canada, the data show that the regression F coefficient for ~N2/L versus numbers of cigarettes smoked daily appeared greater for female smokers than for male smokers (0.0326 versus 0.0185). The prevM alence of abnormalities in tests of small airways function among female smokr. ers in this study was slightly lower than that among the male smokers, but that in nonsmokers was not mentioned. Thus, the contribution of smoking could not 20 30 40 80 be evaluated. Packyear. In contrast, Dockery and coworkers (5)

SEX·SPECIFIC REGRESSION ANALYSIS OF RELATIONSHIP BETWEEN RESIDUAL (r) PULMONARY FUNCTION AND PACK-YEARS Men

rFVC rFEV, rMMFR raN,IL rCVNC

b

Sb

-0.0017 -0.0020 -0.0034 0.0078 0.0502

0.0016 0.0014 0.0035 0.0018 0.0123

p

-1.023 -1.442 -0.970 4.401 4.067

0.3066 0.1499 0.3323 0.0000 0.0001

For definition of abbreviations, see tables 1 and 3.

Residual 0.1

-0.015 -0.1 -0.115 -0.2 -0.215

Fig. 3. Slopes of the residuals of FEY, (L) and the slope of Phase III of singlebreath nitrogen test (aN,/L) (%/L) versus pack-years in men (M) and in wornen (F).

-0.3

10

0

Residual l

0.8 0.8 0.4 0.2 0 -0.2 -0.4 0

10

CHEN, HORNE, AND DOSMAN

1228 TABLE 5 INTERACTION OF SEX AND NUMBER OF PACK-YEARS ON PULMONARY FUNCTION BY SMOKING STATUS'

Sex Pack-years CSS ESS Sex x pack-years x CSS Sex x pack-years x ESS MUltiple R

Sb

b

Sb

b

Sb

b

Sb

b

Sb

0.0105 - 0.0014 -0.0355 0.0161 -0.0031 0.0017 0.0925

0.0454 0.0016 0.0510 0.0542 0.0047 0.0047

0.0075 -0.0015 -0.0529 -0.0004 -0.0040 0.0002 0.1316

0.0384 0.0014 0.0432 0.0458 0.0024 0.0040

0.0625 -0.0050 0.1061 0.1245 -0.0137 -0.0054 0.1119

0.0933 0.0033 0.1048 0.1112 0.0058 0.0097

-0.0870 0.0060 0.1709 0.0165 0.0224 0,0081 0.3121

0.0679 0.0024 0.0763 0.0810 0.0042 0.0070

0.4202 0.0343 1.4909 0.2188 0.0211 -0.0616 0.2140

0.4172 0.0147 0.3794 0.4687 0.0260 0.0433

current smoking status; ESS : ax-smoktnq status. For other definitions, see tables 1 and 3.

recently reported a lesser susceptibility of women to adverse effects of smoking on lung function based on the data from a random sample of 8,191 adults between 25 and 74 yr of age in six U.S. cities. It was estimated that the loss of heightadjusted Fli'V, was7.4mlforeachpackyear smoked for men and 4.4 ml per packyear for women. The estimated effect of cumulative smoking in women was about 60% of that for men. Cook and coworkers (7), studying 3,812 subjects 65 yr of age or older, found a significant interac-

tion between previous smoking and sex on peak expiratory flow rate. The effect of previous smoking was stronger in men, 32 L/min greater than for women. The effect of current smoking was diminished in men, with a difference of 20 L/min from that for women. Years smoked and' years since stopping smoking could not account for the differences between men and women. Burr and colleagues (6)studied 418 subjects older than 70 yr of age and observed that the nonsmokers had the highest FEV 1 at all ages both in men

Residual 0.'

MES ' - - , FES -0.06 -0.1 -0.16

-0.2 -0.26

FS

-0.3f----t----+---+---+---"i---, 60 o 60 20 30 40 10

Packyeara

'.4

Residual FS

1.2

0.6 0.6

MES

0.4

.,' MS _._._ FES

0.2

-0.2

----

..

-0.4 f - - - - - - i r - - - - - j - - - - + - - - - + - - - - + - - - - - - . 60 60 20 30 40 o 10

Packyeara

CVNC

b

ess :

Definition of abbreviations: • Multivariate multiple analysis.

raN,/l

rMMFR

rFEV,

rFVC

Fig. 4. Slopes of the residuals of FE\t; (l) and the slope of Phase III of singlebreath nitrogen test (aN,IL) (%/l) versus pack-years for smokers and exsmokers (MS = male smokers; FS = female smokers; MES = male exsmokers; FES = female ex-smokers).

and women compared with that in exsmokers or current smokers, but the differences for women wereless than for men. This study did not consider the possible differences in smoking amount and duration between men and women. From the Framingham cohort study, Sorlie and coworkers (8) reponed that the mean percent predicted value of FEV 1 was 92.9070 for both men and women who never smoked, but dropped to 80.8% for heavy smoking men and 83.1% for heavy smoking women. Another two studies, however, found no significant sex differences in smoking effects on lung function. Burrows and coworkers (10) studied 2,369 subjects 14 yr of age or older and found that the regression slopes for the percent predicted values ofFEV 1 and Vmaxz 5 on pack -years were nearly identical for men and women after controlling for "respiratory trouble" before 16 yr of age. In the study of Beck and coworkers (9), there was a slight difference in loss for the residual of FEV 1 among adult smokers between the sexes, 10.8 ml per packyear for men and 8.3 ml per pack-year for women, but the difference between these rates of change was not statistically significant. Although convincing explanations for these controversial results are difficult to construct, differences in study design, data collection, pulmonary function measurements, and data processing may be responsible, at least in part, for these apparent discrepancies. Our study used virtually the entire population of the town of Humboldt as study subjects, thus eliminating, or at least considerably reducing, the possibility of selection bias. In our study, the analysis is based on the data of 1,149 subjects who had satisfactory pulmonary function measurements. Compared with the remaining 93 subjects in the community who did not have pulmonary function measurements for all

1229

INCREASED SUSCEPTIBILITY TO LUNG DYSFUNCTION IN FEMALE SMOKERS

the tests, no significant differences were found in distribution of sex and mean values of age, height, weight, and packyears of smoking, nor in prevalences of respiratory symptoms. Therefore, these missing data would not likely distort the present results. The current standards for choosing FVC, FEV hand MMFR are (1) the best FVC and FEV I, not necessarily from the same tracing, and (2) the MMFR comes from the tracing with best sum of FVC and FEV I (18). In this study, we used the same criteria for choosing FVC and FEV 1, but the MMFR was read from the tracing with the largest FVC. Different procedures for pulmonary function testing may cause difficulties in comparisons of studies dealing with the same issues, but a possible systematic bias in MMFR measurement is not pertinent to our findings of sex-related differences in the responses of airways to smoking in this study. Another concern is the validity of statistical means. In the analysis, the prediction equations for lung function of each sex were derived from the data of all the study subjects. Bias might arise since both prevalence of smoking and magnitude of exposure (pack-years) were strikingly different betweenthe sexes. But when we used the prediction equations based only on the data of the asymptomatic "healthy" nonsmoking subjects (11) instead of on those of the whole population, the results wereunchanged and continued to show an interaction of sex and pack-years on lung function. It could be argued that if the female smokers or the male nonsmokers were more likely exposed to other pollutants that also cause pulmonary dysfunction, the observed interaction could be explained in a different way. But the "pollutants" are hard to identify in this study. Outdoor air pollution is virtually nonexistent in the town of Humboldt. In addition, the cooking appliances and heating fuel used were essentially the same for all households, and almost all ranges and ovens are electric. It is possible that women's smoking habits might be correlated with their husbands' smoking habits, and that this extra environmental exposure to tobacco smoke could affect lung function, but this is insufficient to explain the remarkable larger effect of active smoking for women than for men observed in this study. The data demonstrate that the combined effect of sex and pack-years is imperceptible among ex-smokers. This may

reflect the evidence from cross-sectional (19-21) and longitudinal (22-27) studies that cessation of smoking results in significant improvement in small airways function. In those smokers without evidence of chronic airflow obstruction, quitting smoking may lead to a return toward normal of some of the test variables (1). This study shows that L\N2/L is the most sensitive measurement for detecting the combined effect of sexand smoking on lung function. This coincides with results from other studies on the relationship between smoking and lung dysfunction (3, 28-31). In a study on the association between lung function and total mortality, Menkes and coworkers (32) found that L\N2/L was a strong predictor of mortality. Other studies have reported significant relationships between decreased forced expiratory volumes and mortality (33-36). Menkes and Beckett (37) suggested that chronic environmental exposures that produce lung dysfunction may cause an increase in mortality from all causes. If that hypothesis is correct, we might expect sex to modify the relationship between lung dysfunction and mortality, at least in this rural community. However, mortality studies have not been carried out. How sex modifies the influence of smoking on lung function is unclear. It has been suggested that COPD is a disease of multifactorial etiology, and both environmental and genetic factors have been found to be associated with the development of COPD (38). Besides severe ai-antitrypsin deficiency, which is well documented, other genetic factors include the presence of A antigen of the ABO blood groups, nonsecretor status, absence of haptoglobbin IS, absence of Lewis antigen, and presence of HLA B7 (38).Whether the modification bysexhas, a genetic basis is worthy of consideration. References 1. u.s. Department of Health and Human Services.The health consequences of smoking: chronic obstructive lung disease.Washington, DC: USGPO, 1984. (DHHS [PHS] 84-50205). 2. Higgins M. Epidemiology of COPD. Chest 1984; 6(Suppl:3-8). 3. Buist AS, Ghezzo H, Anthonisen NR, et 0/. Relationship between the single-breath N, test and age, sex, and smoking habit in three North American cities. Am Rev Respir Dis 1979; 120:305-18. 4. Buist AS, Ress BB. Quantitative analysis of the alveolar plateau in the diagnosis of early airway obstruction. Am Rev Respir Dis 1973;108:1078-87. 5. Dockery DW, Speizer FE, Ferris BG Jr, Ware JH, Louis TA, Spito A III. Cumulative and reversible effects of lifetime smoking on simple tests of

lung function in adults. Am Rev Respir Dis 1988; 137:286-92. 6. Burr ML, Phillips KM, Hurst DN. Lung function in the elderly. Thorax 1985; 40:54-9. 7. Cook NR, Evans DA, Scherr PA, et 01. Peak expiratory flow rate in an elderly population. Am J Epidemiol 1989; 130:66-76. 8. Sorlie P, Lakatos E, Kannel WB, Celli B. Influence of cigarette smoking on lung function at baseline and at follow-up in 14years: the Framingham study. J Chronic Dis 1987; 40:849-56. 9. Beck GJ, Doyle CA, Schachter EN. Smoking and lung function. Am Rev Respir Dis 1981; 123:149-55. 10. Burrows B, Knudson RJ, Cline MG, Lebowitz MD. Quantitative relationships between cigarette smoking and ventilatory function. Am Rev Respir Dis 1977; 115:195-205. 11. Dosman JA, Cotton DJ, Graham BL, et 0/. Sensitivity and specificity of early diagnostic tests of lung function in smokers. Chest 1981; 79:6-11. 12. National Heart and Lung Institute. Recommended standardized procedures-NHLI Division of Lung Diseases, epidemiology studies. Bethesda: National Institutes of Health, 1971 (revised 1973). 13. National Heart and Lung Institute. Suggested standardized procedures for closing volume determinations (nitrogen method). Bethesda: National Institutes of Health, Division of Lung Diseases, 1973. 14. Buist AS, Ross BB. Predicted values for volumes using a modified single-breath nitrogen test. Am Rev Respir Dis 1973; 107:744-52. 15. SPSS Inc. SPSS-XTM introductory statistics guide. Chicago: SPSS Inc., 1988. 16. Detels R, Sayre JW, Coulson AH, et 0/. The UCLA population studies of chronic obstructive respiratory disease: IV. Respiratory effect of longterm exposure to photochemical oxidants, nitrogen dioxide, and sulfates on current and never smokers. Am Rev Respir Dis 1981; 124:673-80. 17. Manfreda J, Nelson N, Cherniack RM. Prevalence of respiratory abnormalities in a rural and an urban community. Am Rev Respir Dis 1978; 117:215-26. 18. Ferris BG. Epidemiology standardization project. Am Rev Respir Dis 1978; 118(Suppl:55-88). 19. WeissW, Boucot KR, Copper DA, Carnahan WJ. Smoking and the health of older men. Arch Environ Health 1963; 7:538-47. 20. Grimnes CA, Hanes B. Influence of cigarette smoking on the spirometric evaluation of employees c;>f a large insurance company. Am Rev Respir Dis 1973; 108:273-82. 21. Nemery B, Moavero NE, Brasseur L, Stanescu DC. Changes in lung function after smoking cessation: an assessment from a cross-sectional survey. Am Rev Respir Dis 1982; 125:122-4. 22. Bode FR, Dosman JA, Martin RR, Macklem PT. Reversibility of pulmonary function abnormalities in smokers: a prospective study of early disease. Am J Med 1975; 59:43-52. 23. McCarthy DS, Craig DB, Cherniack RM. Effect of modification of the smoking habit on lung function. Am Rev Respir Dis 1976; 114:103-13. 24. Buist AS, Sexton GJ, Nagy JM, Ross BB.The effect of smoking cessation and modification on lung function. Am Rev Respir Dis 1976;114:115-22. 25. Buist AS, Nagy JM, Sexton GJ. The effect of smoking cessation on pulmonary function: a 3D-monthfollow-up of two smoking cessation clinics. Am Rev Respir Dis 1979; 120:953-7. 26. Tashkin DP, Clark VA, Coulson AH, et 0/. The UCLA population studies of chronic obstructive respiratory disease. VIII. Effects of smoking cessation on lung function: a prospective study of

1230

a free-living population. Am Rev Respir Dis 1984; 130:707-15. 27. Bosse R, Sparrow D, Rose CL, WeissST. Longitudinal effect of age and smoking cessation on pulmonary function. Am Rev Respir Dis 1981; 123:378-81. 28. Oxhoj H, Bake B, Wilhelmsen L. Ability of spirometry, flow-volume curve and the nitrogen closing volume test to detect smokers. A population study. Scand J Respir Dis 1977; 58:80-96. 29. Viegi G, Paoletti P, Pede FD, et al. Singlebreath nitrogen test in an epidemiologic study in north Italy. Reliability, reference values and relationships with symptoms. Chest 1988; 93:1213-20. 30. RasmssenFV.Occupational dust exposureand smoking. Different effects on forced expiration and slope of the alveolar plateau. Eur J Respir Dis 1985;

CHEN, HORNE, AND DOSMAN

66:119-27. 31. Ruppel G, ed. Gas distribution tests. In: Man-

ual of pulmonary function testing. S1. Louis: The CY. Mosby Company, 1986. 32. Menkes HA, Beaty TH, Cohen BH, Weinmann. Nitrogen washout and mortality. Am Rev Respir Dis 1985; 132:115-9. 33. Ashley F, Kannel WB, Sorlie PD, Masson R. Pulmonary function: relation to aging, cigarette habit, and mortality. Ann Intern Med 1975; 82: 739-45. 34. Tibblin G, Wilhelmsen L, Werks L. Risk factors for myocardial infarction and death due to ischemic heart disease and other causes. Am J Cardiol 1975; 35:514-22. 35. Freedman GD, Klalsky AL, Siegellaub AB. Lung function and risk of myocardial infarction

and sudden cardiac death. N Engl J Med 1976; 294:1071-5. 36. Beaty TH, Cohen BH, Newill CA, Menkes HA, Diamond EL, Chen CH. Impaired pulmonary function as a risk factor for mortality. Am J Epidemiol 1982; 116:102-13. 37. MenkesH, BeckettW.Airborne environmental exposure, lung function and mortality. In: Dosman JA, Cockcroft DW, eds. Principles of health and safety in agriculture. Boca Raton, FL: CRC Press, Inc., 1989; 8-11. 38. Horne SL, Cockcroft DW, LovegroveA, Dosman JA. ABO, Lewis and secretor status and relative incidence of air flow obstruction. Dis Markers 1985; 3:55-62.