TABLE 2.-Continued

Author, year,
country

Population          cough

Other

Commenta

Higeins
1977.
U.S.

et al.,

4,699 men and
women, aged
B-74, Team&

Chronic bmnchitii
     Men
NS         5.1
Ex         2.6
SM
<m/day    13.4
2 2Lvday    29.9
Pipekigtk?   8.4

    Women
NS         3.5
Ex        4.7
SM
<2o/dav    4.8

Chronic bronchitis knugh and
phlegm 23 mo/yr); chronic
bronchitis increaeed with age
for male smokers; no age trend
apparent for male or female
nonsmokers; doee-response
relationship between chronic
bronchitis and cigarette smoking
(age adjusted)

z'xvda;    15.3


TABLE 2.-Continued


Author, year,
wuntry         Population          cough          Phlesm             Other              Commenta


Lhwitz and     2,857 men and          Chronic wu6h and/or phlegm                     Male prevaienax am8&tently
BIllTow&       women, aged 14-66,                 Men
1977, U.S.        Tucson                   NS          10.3                      greater than female only in older

                          SM (pack years)                          w grow; fresuencr of

                                 <6         29.0                       Bymptoms in& with age;


                                     6-20        35.8                     impossible to distinquiah effecta
                                                          of ag@ and increaeed duration
                                   21-40        47.9                       of 8mokin6
                                   41+        61.1

                                     Women
                                 NS          12.1
                         SM back-years)
                               <6         21.0
                                   6-20        33.1
                                  21-40        40.5
                                  41+        60.4


TABLE 2.-Continued

Author, year,
wuntry

Population         couch

Other

Commenta

Bland et al.,
1976,
Great Britain

6,320 f-year
secondary echo01
Children,
Derby&in?

Morning cough
    Boys
NS        3.1
SM once     2.9
oaaa.     4.0
1 per wk.   19.2


     Girl8
NS        1.8
SM one     4.5
occaa     6.0
1 wk.
  per     13.5
Cough at other timee
    fm
NS       20
SM once    27
0a-a.Y.    34
1 wk.
  per     47

    Gil8
NS        19
SM once    30
oa9e.    35
1 wk.
  per     47

  Breathkmness
    Boye
NS        12
SM once     14
occaa     22
1 wk.
  per     35


     Girls
NS        17
SM once     22
oaxla     29
1 wk.
  per     40

No questions on phlm; child's
smoking habita more important
than parer&' smoking habite;
parents' smoking hahita
independently related to mu&
eymptom frequencim for
boyaandgirla


$   TABLE 2.-Continued


   Author, year,
   country          Population          bush            Phlegm              Other               Comment.3

Huhti et al.,      1,162 men,
1978,         asd 25-65,
Finland         Hankaealmi

All day in winter
Age        NS
2x39        2
40-49        2
5049        5
m-69        4

Age       Jzx
25-39        -
40-49        -
50-59        4
60-69        12

Age       SM
B-39        9
40-49        19
50-59        19
60-69        30

AU day in winter
   NS
    5
    2
    8
    7

Ex
11
10
8
18

Chronic bronchitis
Age      NS
25-39 9
40-49 2
M-59      15
6043 19

        Ex
25-39 11
4049 14
50-59 11
60-69 26

SM          SM 1-14 g/day
14           25-39 29
29           a49 39
24           5049 31
27           60-69 39

For total group, aign&ant
increase with age of cough,
phlegm, and severe breathleas
neea; for nonsmokers, sign&ant
inc- with age for phk?gm
and breathleesnees only

SM 16-24 g/day
2639 13
40-49 45
50-59 46
60-69 46

SM 225 g/day
25-39 50
4049 63
50-59 57
al-69 53


TABLE 2.-Continued

Author, year,
-try

Population         coush

Phlegm

Other

Comment8

Manfreda et al..    273 men and 229      Most daye 23 molyr
1978         women, aged
          2455, two             CH' PLP"
           communitlea         NS   8.3    4.0
           inMnnitoha         Ex   8.1    2.9
                         SM  25.4   31.5

NS - 4.0
E!i -    10.0
SM 20.3   31.7

Moat day 2.3 mo/yr
    Men
    CH   PLP
NS   -    4.0
Ex   10.8    5.7
SM   16.9   24.7

   Women
NS   -    4.0
Ex   -    5.0
SM   10.2   26.4

Wheeze apart from colds

   CH   PLP
NS   4.2    8.0
M   10.8   14.3
SM   26.8   31.5

NS   3.5    8.0
Ex   12.1   20.0
SM   25.4   30.2

No consistent difference in
symptom prevalence for two
communities; higher prevalence
of mugh, phlegm, and wheeze
among emokem than
nonsmokers or exanokers;
`CH=ChdfSWOOd
`*PLP=Portage la Prairie


g TABLE 2.-Continued

Author, year,
anlntry

Population          cough

Other

Comments

Rawbone et al.,
1678,
Great Britain

10,498 secondary
school children,
aged 11-17,
Hounelow

A little mast dayn
Age      NS
11      26.4
13      16.7
15      13.4
      EXPEB'
11      20.6
13      17.3
15      11.3
       SM
11      24.3
13      25.9
16      27.4

A little every day
       NS
11       7.3
13       4.2
15       4.7
      EXPEB
11       7.0
13       4.7
15       3.5
       SM
11      26.2
13      17.8
16      13.4

With colds        Frequent adds
  NS               NS
22.8                36.0
26.3                32.3
24.6                34.3
EXPEB             EXPEB
32.3                42.2
31.2               36.1
26.6                36.3
SM               SM
56.1                51.1
49.6                60.4
63.6                39.8

' EXF%B (examokere end
experimental smokers combined);
sex differences not -cant;
nonstandard questions; higher
symptom prevalence in younger
children not explained


TABLE 2.-Continued

Author, year,
colmtry

Population          coush           Phlegm             Other               Commenta

Bouhuye et al..
1979,
U.S.

7,203 reaidente,
43-d 7-65+,
three ccxnmunitim

  Usual cough
   LB' AN"
     Men
NS    8   10
SM   27   32
   Women
NS    7   12
SM   13   17

WV

15
34


9
24

Smoking eigniiicantiy amociated
with cough, phlegm, wheeze, and
dyqmq prevalence increased
signiflcnntly with age, slightly
higher in urban oxnmunity;
women had lower prevalence of
phkgm and higher prevalence of
dppnea than men
o ???????*
???????

twl=wiiro

Burghard et al.,
1979,
France

29.136 &dents,        Morning                         Breathleumeas     Prevalence of 8ympt.onu increased
aged M-18.         NS        13.7                      NS        14.1     with incwasing cigarette
Badthin Department SM       25.7                       SM        22.2     consumption; girln had higher
             Day or night                             prevalenw of respiratory
             NS        16.9                             aymptanu for each emoking
              SM        29.1                                  --aw
                  chronic
              NS        2.7
              SM        6.6


TABLE 2.-Continued

Author, year,
country

Population          `hsh

Other

Comment8

culavik,       19,988 people,        Morning cough
1979,        aged 16-70.        NS         11
Norway         GE10             Ex         16
                          SM         36
                       Daytime cough
                          NS          4
                          Ex          7
                         SM         16
                       Cough 3 mmlyr
                          NS          3
                          Ex          5
                         SM         14

When awghing     Cough and phlegm perioda   Cough and phlegm increasedmore
NS         10      NS        6      with age for smokem (10-19
Ex         18      Ex         8      c&/day) than nonsmokers; no
SM         26      SM        16      significant relation&p between
                           age and prevalence of periods of
                           cough and/or phlegm;
                           dose-response relationship
                           between number of c@rettea and
                          symptoma reported; data not
                             given

Lid et al.,
l=t
France

899 men and women,
ased -t,
(av. 39), Paris

NS
SM


NS
SM

Men                           Respiratory aymptcuna (cough and/or
      16.0                      phlegm 3 moe/yr. for 2 years); not a
      25.4                       random sample; male and female
Women                          amokem had eimihr symptom
      8.1                       prevalence; female nonsmokera had
      26.6                      lower prevalence


TABLE 2.-Continued

Author, year,
country

Population         h?h           Phlegm              Other              Commente

Park,
1961,
KOll?Il

856 university
men and women,
seed 18-w
seoul

NS
Ex
SM


NS
Ex
SM

NS
Ex
SM

Morning
    16.0
    31.3
    34.0
Daytime
     4.2
     8.3
    10.9
Nightime
    13.5
    18.8
    19.9

NS
Ex
SM

NS
Ex
SM

NS
Ex
SM

Morning
    20.1
    22.9
    26.1
Daytime
     2.1
    14.6
    12.0
Niihtime
     7.3
    10.4
    13.2

Breathlessneea on exertion   Symptom prevalencea apparently
NS        23.2     not ageadjusted; eiguificance
Ex        25.0     levels not reported; nonetand-
SM        29.7     ard questions; eymptome current

Neukirch et al.,
1982.
FreIW

23% Becondary
school etudente,
43ed<ll-218
(me3n 14.9).
Peri.9

U.sual cough and/or phlegm
     Boys
NS           26.1
SM          34.9
      &la
NS           26.9
SM           44.7

21.8% of boys, 32.2% of girls
were current smokers; girl
smbkers had higher symptom
prevalence than boys if total
lifetime cigarette8 >4,000; resulta
probably not age adjusted

Schenker et al.,
1962,
U.S.

5,686 women,
aged 17-74 (mean
44.6), western
Pennsylvania,
telephone interviewn

Chronic cough       chronic phlegm    Wheeze moat daye or nigh&   Cough and phlegm meet strongly
NS        5.6      NS        4.5      NS         7.2     related to current cigarettes/day;
w        7.5       Ftx        6.7       Ex         8.3     tar content had independent
SM               SM              SM            effecte; age effect Been for
l-14     9.1         1-14      7.2         1-14     14.4     nonsmokera, but not current
15-24     17.0        15-24     16.7        15-24      16.5     smokers; symptom
25t     31.8        25+     24.8        St     28.0     prevalencee age adjusted


TABLE a.-Prevalence (percent of cough, phlegm, and other symptoms for nonsmokers (NS), smokers
     (SM), and ex-smokers (MI, longitudinal studies


Author, year,
country         Population       Smoking habits               SYmptoms                      Comments


Ferris et al.,      1,201 men and                       Cwrh           Phlegm        72.3% of men, 78.4% of
1976,        women, aged 25-74          1973           1967       1973      1967      1973     women followed up; 1973,
U.S.          in 1961, Berlin                              Men                      symptom prevalencea, ege
          New Hamphire         NS               6.0        8.5       8.9        7.6    @usted to compare with 1967,
                            Ex             20.5        9.7       23.3       15.9     showed little change
                           SM
                                 l-14          22.2       25.5       17.9       27.5
                                 15-24          35.4       26.5       31.6       30.0
                               25-34          26.1       26.7       33.8       32.4
                            St           50.6       56.4       37.1       51.9

                                             Women
                            NS               4.4        6.2        8.1        7.4
                            Ex               3.2       5.2        7.3       10.1
                            SM
                                  1-14          10.7       10.0       11.6        9.8
                               15-24          19.5       16.3       21.8        9.8
                               2544          27.2       16.1       22.5       21.8
                             St           44.7       31.0       43.1       41.2

Kieman et al.,    2,736 men and                        Cough day or night in winter         Effecta of cheat illnem before age
1976.        women, aged 26, born    1966       1971           1966               1971        2, father's vocation, and current
Great Britain     ill1946,eUIlllin       NS       NS            5.5                4.9        emokhg r$nikant; air pollution
           1966 and 1971        NS       SM            7.2                9.6 '        effect not aignif~cant; current
                           SM       SM            14.3               18.5 '        smoking had lanpst effects
                           SM       Ex            9.2                5.8         `Prevalence, 1966 vs. 1972
                                                                   P <0.05


TABLE 3.-Continued

Author, year,
cmntly

Population

Smoking hahits

Comments

Leeder et al.,     2,130 fathers,                        Cough/phlegm prevalence range        In male ersmokers, prevalence
1977,         mean age 31.0rt6.1,                         Men                      of cough/phlegm decreased
Great Britain     2,146 mothers,      let period   2nd period     1st syr period        2nd 3yr period      over time; no signiicant
          mean age 27.9f5.3,     NS       NS           8.69.6              9.2-11.1       change in prevalence in female
            children horn         NS       SM          4.6-16.9             13.3-20.5        er+mokera. but numhera
          19634965,          SM       SM          25.6-30.2            30.W4.0        were amall
         London, 6 year        SM       Ex          21.6-25.3             5.b20.7
         r0110wup                               Women
                          NS       NS           4% 6.8             5% 7.3
                            NS       SM          8.2-10.2            13.3-18.4
                           SM       SM          16.3-22.4             23.0-28.4
                           SM       Ex           4.1-22.5            12.2-14.3

Woolf and Zamel.   302 women, wed                   Cough and/or phlegm       Breathleesneas      60% followcxl up; all subjects
1960.          `E-54 at initial                        let exam   Findexam    let exam   Finalexam   maintained con&ent smoking
Canada        study. &year             NS            10        14        10         5      habita for 5 years
          followup                J%x             3        13        18         8
                                SM            56        54        25        21
---   ---_.-. -.  ___.-__. .~_..


TABLE 3.-Continued

Author, year,
muntry

Population

Smoking habits

Comments

Beck et al..
1982,
U.S.

1,262 white
residents, aged 7-55 t ,
Lebanon, Connect
icut, exams in
1972 and 1978

1972


NS
NS
SM
SM
FX

1978

NS
SM
SM
Ex
Bx

NS       NS
NS       SM
SM       SM
SM       Ex
Ex       Ex

Usual cowh
1972       1978
    Men
5        2
0        0
23       21
25        2'
7        6

7         4
0         0
20        14
26        12
10        3

usual phlegm
1972       1978


7         3
0         4
22        26
18         8
12        15


5         6
0         9
15        11
8        16
4         1

65% followed up; health indiaza
of respondents and non-
respondents similar; symptom
prevalence tended to decline, but
few changes were. significant;
`Prevalence, 1972 vs. 1978,
p <CL5


FIGURE Il.-Prevalence of chronic cough and/or chronic
       sputum among samples of men and women in
       Tucson, Arizona
SOURCE: Lebxvitz and Burrowa (1977).

Dose-Response Relationships

The most common measures of dose are the number of cigarettes
currently smoked per day and the pack-years of cigarette consump
tion; the latter estimates lifetime exposure by integrating the
number of cigarettes smoked (by pack) and the duration of cigarette
use. Errors of memory compromise the accuracy of retrospective
information, which may also be biased by differential recall in those

63


50.0 -





40.0 -





30.0 -





20.0 -




10.0 -

Ex-smokers

Nonsmokers

220

lo-19 \ smokers
    c4Qaretles~day

l-9

19-23    24-27   29-32   233

Tar (rtq/c@arette)

FIGURE 12.-Percentage of smokers with phlegm
      production (adjusted for age), according to tar
       yield of cigarettes
SOURCE: Higenbvttam et al (1980).

with and without smoking-related symptoms or diseases. Even
accurate reports of current smoking habits fail to take into account
all the sources of variation in exposure associated with the material
used in cigarette manufacture or generated in the burning of
cigarettes, The dose of noxious materials received is also influenced
by human behavior, including the number, volume, and timing of
puffs taken with each cigarette; retention of smoke in the mouth;
depth of inhalation; disposition of the cigarette between puffs; and
other aspects of smoking style that are not reproduced by the
smoking-machines used to measure tar and nicotine yield.

Prevalence rates of cough or phlegm, or both, generally increase as
the number of cigarettes smoked per day increases. The trends
illustrated in Figures 11 and 12 were present in both sexes and all
age groups (Lebowitz and Burrows 1977; Bean et al. 1978; Higgins et
al. 1977; Huhti et al. 1978; Higenbottam et al. 1980; Schenker et al.
1982). Bland et al. (1978) found a d-response relationship in
secondary school children, among whom rates of reporting cough
were higher in those who smoked most, even though levels of
cigarette consumption were generally reported to be low. At the
other extreme of the age range the trend is also apparent, even
though symptomatic smokers are more likely than asymptomatic
smokers to stop smoking or to reduce their cigarette consumption


(Higgins 1974;  Fl t h
                e c er 1976). Symptoms were more prevalent
among heavier smokers of filter cigarettes as well as of nonfilter
cigarettes (Dean et al. 1971). Prevalence rates of cough, phlegm,
chronic bronchitis, and mucus hypersecretion show a similar pattern
of association with pack-years of exposure (Tager and Speizer 1976;
Lebowitz and Burrows 1977). Pates of incidence and remission
observed in longitudinal studies add further support to the strong
evidence that respiratory symptoms increase as exposure to cigarette
smoke increases (Table 3).

In their study of more than 18,000 male civil servants in London,
Higenbottam and colleagues (1980) found that the percentage of
smokers who produced phlegm increased with increased daily
cigarette consumption and also with increasing tar content of
cigarettes among those who smoked less than 20 cigarettes a day.
Symptoms were prevalent about equally among smokers of 20 or
more cigarettes per day, regardless of the tar yield of the brands they
used (Figure 12).

Schenker et al. (1982) reported the relationship of tar content of
cigarettes to respiratory symptoms in a cross-sectional telephone
survey of 5,686 adult women in rural Pennsylvania. The risk of
chronic cough and phlegm was more strongly affected by the number
of cigarettes smoked per day than by tar content. Cough and phlegm
were reported least often by never smokers and with increasing
frequency as the number of cigarettes smoked per day increased. Tar
content of cigarettes was significantly associated with symptoms of
chronic cough and phlegm--especially cough-and its effects were
independent of the number of cigarettes smoked per day in a
multiple logistic analysis. The risk (relative odds) of chronic cough
for smokers of high tar cigarettes (20 or more mg) was approximately
twice that for smokers of an equivalent number of low tar cigarettes
(10 or less mg). A limitation of this cross-sectional study was the
determination of tar content for current cigarettes only, rather than
for lifetime smoking habits. Although the apparent relationship
between tar content and symptoms could have been caused by
changes in smoking habits, this was considered unlikely because
symptomatic smokers tend to reduce their consumption of cigarettes
more than asymptomatic smokers (Fletcher et al. 1976) and may also
switch to low yield cigarettes. In this situation, any reported effect of
tar content on symptoms would be an underestimate.

In summary, the prevalence of symptoms increases with dose of
smoke exposure, when dose is measured by number of cigarettes
smoked per day or tar content of the cigarette smoked.

Relationship of Cough and Phlegm to Sex and Age
Prevalence rates of cough and phlegm ascertained in epidemiologi-
cal studies generally increase with age and are higher in men than

65


in women, as shown in Figure 11 and Tables 2 and 3. Rates also vary
with smoking habits. Rates in nonsmokers better clarify associations
of symptoms with age and sex than do rates in smokers, since they
are less confounded by variations in exposure to cigarette smoke.
However, recent evidence linking passive smoking with increased
prevalence of respiratory symptoms suggests that rates in nonsmok-
ers may be in excess of those that would be found in a population
completely free of exposure to cigarette smoke (Lefcoe et al. 1983;
Weiss et al. 1983).

Rates of reporting cough or phlegm or both were roughly equal in
nonsmoking men and women in several cross-sectional studies
(Bland et al. 1978; Higgins et al. 1977; Lebowitz and Burrows 1977;
Manfreda et al. 1978; Neukirch et al. 1982; Rawbone et al. 1978).
Rates were higher in nonsmoking men in some populations (Dean et
al. 1977; Liard et al. 1980; Tager and Speizer 1976). Bouhuys et al.
(1979) found no sex difference in the prevalence of cough, but a
higher rate of reporting phlegm in male nonsmokers (Table 2). In
most of these studies, the rates were not corrected for exposures to
other respiratory irritants in the workplace or in the general
environment.

In general, symptoms are more prevalent in male smokers than in
female smokers (Table 3). However, differences in prevalence rates
between the sexes are generally smaller or absent when comparisons
are made between men and women who smoke similar numbers of
cigarettes. Lebowitz and Burrows (1977) found that the excess
prevalence of symptoms in male smokers compared with female
smokers tended to be greatest at older ages, where there are also the
greatest differences in smoking behavior. Men in these birth cohorts
tend to have begun smoking earlier in life, smoke more cigarettes
per day, inhale more deeply, and smoke higher tar and nicotine or
unfiltered cigarettes. Two studies from France, Burghard et al.
(1979) and Neukirch et al. (1982), concentrated on high school
students. In general, the prevalence of smoking was similar for both
boys and girls for the two studies, although the Neukirch group
found a somewhat higher rate among the girls (46 percent versus 39
percent). Slightly more boys than girls, however, smoked more than
10 cigarettes per day. In these two studies, the prevalence of
symptoms was higher among female smokers than among male
smokers. These data suggest that the past differences in prevalence
of symptoms between the sexes is largely attributable to differences
in cigarette consumption. These differences were substantial in the
past, and are still present among older adults, whereas current
smoking practices are about the same in male and female adoles-
cents and young adults.

Prevalence rates of cough, phlegm, and chronic bronchitis in-
creased with increasing age in the U.S. population samples studied

66


by the National Center for Health Statistics (1981) and in several of
the cross-sectional studies cited in Table 2. However, differences in
rates of reporting symptoms among people of different ages may
relate to effects of aging, differences in current exposures, or
differences in exposure to cigarette smoke or other noxious agents in
the past. It is therefore difficult or impossible to use cross-sectional
data to separate effects of aging from effects of duration, dose, and
nature of cigarette smoke exposure throughout life. Longitudinal
studies provide information on time trends, both in exposure and in
onset and course of disease. Nevertheless, conclusions may be
incorrect if people who drop out of longitudinal studies are different
from those who continue to participate.

Prevalence of symptoms increased with increasing age among men
in cross-sectional data from Tucson (Figure ll), but the trend was
more apparent among smokers and ex-smokers than among non-
smokers. However, Lebowitz and Burrows (1977) could not distin-
guish between an association caused by increasing age and an
association due to increasing duration of exposure to cigarette smoke
in smokers because the two were so highly correlated. Among
women, symptoms were reported more frequently at ages 30 to 44
than at ages 15 to 29 (except by ex-smokers), but prevalence rates
were essentially the same for the three groups over age 30. Higgins
et al. (1977) found that there was no increase in cough and phlegm
with increasing age in male or female nonsmokers in Tecumseh
(Michigan), whereas prevalence rates increased with increasing age
in male smokers. The pattern in female smokers was similar to that
in Tucson and showed an increase with age up to age 30 or 40, but
rates declined with increasing age after age 50. The extent to which
these patterns related to amount smoked or duration of smoking was
not reported, but these older birth cohorts of women probably began
to smoke later in life and smoked fewer cigarettes per day, according
to national smoking survey data.

In other cross-sectional studies cited in Table 2, symptom preva-
lence increased with age in the populations studied (Bouhuys et al.
1979; Dean et al. 1977; Gulsvik 1979; Huhti et al. 1978; Tager and
Speizer 1976), but the trend was noted by Gulsvik to be less in
nonsmokers. Huhti et al. found a significant increase with age
among nonsmokers for phlegm and dyspnea only. Schenker et al.
(1982) observed a trend for nonsmokers but not for smokers, and
Tager and Speizer found that adjusting for smoking eliminated the
trend with age.

Prevalence rates of cough and phlegm on two occasions 3 to 6 years
apart are shown in Table 3 for five recent longitudinal studies of
populations in the United States, Canada, and Great Britain.
Kiernan et al. (1976), Leeder et al. (1977), Woolf and Zamel (1980),
and Beck et al. (1982) found little change in the prevalence of

67


symptoms among continuing nonsmokers during the followup inter-
val of up to 6 years. The rates among nonsmokers reported by Ferris
et al. (1976) are similar on the two occasions, but symptoms were
presented by smoking habits at followup only, and any effect of age
was deliberately adjusted out because the authors' purpose was to
evaluate effects of changes in smoking, changes in pollution, and
trends over time independent of changes in age. Cough and phlegm
appeared to be more frequent at followup in the persistent smokers
studied by Kiernan et al. (1976) and Leeder et al. (1977), and about
the same in women studied by Woolf and Zamel(1980) and in men
studied by Reck et al. (1982). However, rates were slightly lower at
followup in the female smokers followed by Reck. Even though
starting to smoke or quitting can be eliminated as the explanation
for increases or decreases in symptom prevalence over the course of
these studies, it is possible that changes in the number or type of
cigarettes smoked by persistent smokers influenced the prevalence
of symptoms. The duration of followup in all these studies was
relatively brief, and it is still difficult to distinguish between effects
of aging and effects of duration, amount, and nature of exposure to
cigarette smoke in smokers, even when major changes in smoking
behavior are controlled. However, available data suggest that age
itself is not the major factor responsible for differences in the
frequency or distribution of symptoms in populations of nonsmokers
and smokers.

Relationship of Cough and Phlegm to Airflow Obstruction

Many cross-sectional studies have found associations between
cough, phlegm, chronic bronchitis, or mucus hypersecretion and
reduced levels of pulmonary function. The forced expiratory volume
at 1 second (FEVI) has been measured in most clinical studies and in
nearly all epidemiological studies, and mean levels of FEVI are
generally slightly lower in groups of people who report respiratory
symptoms (USPHS 1964, 1971; USDHEW 1979; USDHHS 1980a,
1981). Recent studies have compared other measures of pulmonary
function in people with and without symptoms and have provided
longitudinal data on pulmonary function for symptomatic and
asymptomatic smokers and nonsmokers. Attention has been given to
understanding the natural history of chronic airways obstruction
and the interrelationships of respiratory symptoms, levels and rates
of decline of pulmonary function, and their independent and
interrelated associations with cigarette smoking. Several investiga-
tors have emphasized the desirability of identifying in advance those
smokers who will develop severe COLD; symptoms and other
characteristics have been evaluated as potential predictors of
morbidity or mortality from COLD.

68


Fletcher and colleagues (1976) found that the age-height standard-
ized FEVl at the initial survey of their population of working men in
London was inversely related to the volume of sputum produced in
the first hour after getting up. The regression of FEVl on age, given
height, was steeper for symptomatic cigarette smokers than for
asymptomatic smokers or nonsmokers. However, the authors cau-
tion that men may develop symptoms as they age and change from
one regression slope to the other.

Burrows et al. (1977a) found that an index of cough or sputum was
related to FEVl percent predicted when pack-years of smoking were
controlled in a multiple regression analysis. Regressions of FEVl
percent predicted on pack-years are shown for people with and
without chronic cough and sputum in Figure 13; the intercept at 0
pack-years was lower and the decline in FEVl with increasing pack-
years was significantly greater for those with chronic cough and
sputum than for those with no cough or sputum. The authors
calculated that values of FEVl were lower by about 10 percent in
people with cough and sputum, regardless of smoking habits, and
that values declined by about 4 percent of predicted for each 10 pack-
years of smoking in people with cough and sputum and by about 2
percent in subjects without productive cough. There was a signifi-
cant relationship between FEVI and pack-years of smoking in
asymptomatic smokers in this population. A weaker relationship
between cough and sputum and Vmax~% was also found to be
independent of pack-years of smoking; however, prediction equations
for flow rates have been revised substantially (Knudsen 19831, and
the extent to which relationships between the revised flow rates and
pack-years of smoking differ in symptomatic and asymptomatic
subjects has not been reported.

Dosman et al. (1976) found poor correlations between respiratory
symptoms and dynamic lung compliance, closing volume, closing
capacity, slope of phase III, and helium flow-volume curves in a
study of 49 smokers and 60 nonsmokers who were recruited from a
smoking cessation clinic, a personnel department, and the staff of a
laboratory.

In their community-based studies of children and adults, Bouhuys
and colleagues (1977) studied relationships between respiratory
symptoms and loss of lung function in smokers and nonsmokers.
They found that residual values (observed-predicted) of FVC, FEVl,
PEF, MEFsoa, and MEFzsz, were not significantly different in people
with no symptoms or only one symptom when analyses were done
separately for adult white male smokers and nonsmokers. When a
symptom score was used to combine information on usual cough,
usual phlegm, wheeze, and dyspnea, decrements in lung function
were greatest among those with most symptoms.

69


   1M)









   90


s
5
e
2    90









    70

Subjects with chronic
Subjects with chronic
cough and chronic
cough and chronic
Sputum (n = 247)
Sputum (n = 247)

chrome sputum (n = 1.803)

oL--.---
            20          40          60          80

Pack-years

Figure 13.-Percentage distribution of pred&&d forced
     expiratory volume in 1 second (FEVI) versus
      pack-year of cigarettes smoked, by cough and
      sputum history
SOURCE: Burrows et al. (1977a).

In a study (Detels et al. 1982) designed to assess the relative
sensitivity and specificity of symptoms, the flow-volume curve (FV),
the single breath nitrogen test (SBNT), and specific airway conduc-
tance (SGaw) for identifying COLD were compared with the
FEVl/FVC ratio and with one another in 1,201 residents of Los
Angeles 25 to 29 years old. The tests were done in 1978-1979 at a
followup examination of a previously defined cohort. Prevalence
rates of cough and sputum were 9 percent in never smokers, 26

70


percent in current smokers, and 33 percent in smokers of 20 or more
cigarettes a day. Prevalence rates of an abnormal FEVl/FVC ratio in
these groups were 8,23, and 33 percent, respectively (the FEVAWC
ratio was considered abnormal if it was below the 95th percentile for
never smokers without a history of respiratory illness). The research-
ers found that there was very little overlap between the presence of
productive cough and abnormal tests, and that none of the tests of
function showed reasonable concordance with this symptom. Lack of
reasonable concordance meant that none of the other tests were
abnormal in 50 percent or more of the individuals with productive
cough. In this study, the FEVl/FVC ratio was used as the standard
against which the sensitivities of the other tests were judged; the
sensitivity of the FEVAWC itself was evaluated by its agreement
with those tests found to be sensitive in the study. The lack of an
independent method for identifying COLD, the cross-sectional na-
ture of these data, and the way in which analyses were done restrict
the ability to make biological inferences about the independence of
the effects of cigarette smoking that lead to cough and sputum or to
chronic airflow limitation. However, the authors note their findings
are consistent with the hypothesis that effects of smoking on cough
and sputum are independent of effects on airflow limitation.

Insights into the course and pathogenesis of COLD have been
developed by Fletcher and his colleagues from observations made
during their 8-year longitudinal studies of levels and rates of decline
in lung function in middle-aged working men in London (Fletcher
1976; Fletcher et al. 1976). These investigators found that various
measures of sputum production were correlated with FEVl standard-
ized for height and age, and that this correlation was weakened only
slightly by adjusting for smoking habits. The researchers maintained
that the association between sputum production and pulmonary
function could be due entirely to a common causation. Some men
with mucus hypersecretion had normal FEVl; conversely, some men
with airflow obstruction did not report phlegm. Nevertheless, the
relationship between phlegm and reduced FEVl was strong enough
to give rise to an estimated reduction in FEVl of about 0.1 liters for
every ml of sputum expectorated in the first hour after getting up.
However, because decline in FEVl (FEV1 slope) was not related to
measures of sputum production when level of FEVl and smoking
habits were controlled, the researchers concluded that mucus
hypersecretion is not a cause of accelerated decline in FEV1.
Furthermore, there was no evidence that short-term changes in
sputum production were associated with short-term changes in
FEVl. The researchers concluded that the association between
expectoration and reduced FEVl is caused by the increased suscepti-
bility of some people to both expectoration and excessive loss of
FEVl when they are exposed to cigarette smoke or, presumably, to

71


other noxious materials. This study has made important contribu-
tions to understanding the natural history of chronic bronchitis and
emphysema, but the duration of followup was only 8 years, the men
were 30 to 59 years of age at the start of the study, and their mean
age was 51 years at the midpoint. Similar studies of younger men
and women and observations over longer periods of time are needed
to extend these findings.

Johnston et al. (1976) found that sputum volume was not related to
decline in FEVl in a lO-year followup study of chronic bronchitic
patients. There was no difference in sputum volume between
patients whose FEVl fell by more than 33 percent and controls
(matched on initial FEVl 1 whose FEVl did not fall. Furthermore,
sputum volume was reduced in response to stopping smoking or to
antibiotic treatment, whereas rate of decline of FEVi was unaffected.
In this and other studies (Higgins et al. 1970; Fletcher et al. 1976;
Peto et al. 1983) FEVl was strongly predictive of morbidity and
mortality, whereas respiratory symptoms were not.

Woolf and Zamel (1980) studied "normal" employed women aged
25 to 54 in a longitudinal study designed to identify smokers at
increased risk of developing COLD. Ventilatory function was mea-
sured at the beginning and at the end of a 5-year period during
which smoking habits and symptoms were ascertained annually.
Differences between initial and followup values of pulmonary
function tests were expressed as a percentage of the initial value and
compared in persistent nonsmokers and persistent smokers who
either consistently reported or consistently denied cough or sputum.
The decline in FEVi, FEVl/FVC, and FEFw7s% was greater in
symptomatic smokers than in asymptomatic nonsmokers, but not
significantly different in asymptomatic smokers compared with
either nonsmokers or symptomatic smokers. The average number of
cigarettes smoked during the course of the study was greater for
smokers with cough and sputum. Change in FEFs75% was evaluated
in individual smokers, and no association was detected between
cough and sputum and percentage change in this measure of lung
function. The investigators identified one group of smokers whose
decline in FEFZSSS was similar to that in nonsmoking women and
another group with a greater decline; cigarette consumption was
similar in the two groups. The investigators concluded that individu-
al susceptibility is an important determinant of the effect of
cigarette smoking, because some women develop symptoms and
others remain symptomless but experience rapid worsening of
ventilator-y function. However, they noted a tendency for both cough
and sputum and rapid worsening of ventilatory function to coexist.
The number of women in some groups was very small, and the
measure of decline in lung function used by these researchers does
not take into account regression to the mean or assess absolute

72


reduction; those with smaller initial values will have greater
percentage reductions for a constant absolute reduction in function.

Followup studies at 10 and 15 years of the Tecumseh, Michigan,
population showed that incidence rates of obstructive airways
disease were higher in men and women who reported cough, phlegm,
or both symptoms (chronic bronchitis) at entry compared with those
who denied these symptoms (Figure 14) (Higgins et al. 1982). Both
cough and chronic bronchitis were significant predictors of obstruc-
tive airways disease in men even when smoking habits were
controlled in multiple logistic analyses. However, respiratory symp-
toms were poorer predictors of impaired pulmonary function at
followup than were smoking habits and baseline levels of lung
function. In a multiple logistic model with age, smoking habits, and
level of lung function as risk factors, over 60 percent of the lo-year
incidence cases developed among men and women in the top 10
percent of the risk distribution, whereas only 36 percent of incidence
cases were in the top decile of risk when cough, rather than FEVl,
was used as a risk factor (Higgins 1984).

summary

Cigarette smoking is associated with respiratory symptoms, in-
cluding mucus hypersecretion, and with prevalence and incidence of
COLD manifested by irreversibly impaired pulmonary function.
While some smokers develop both conditions, and those with cough
and phlegm are at increased risk of developing airways obstruction,
the conditions can occur separately by mechanisms that are imper-
fectly understood but appear to be different. The excess risk of
reduced FEVl or COLD in symptomatic smokers compared with
asymptomatic smokers may be a reflection of increased susceptibili-
ty in some individuals. However, it may also be a measure of
increased dose of cigarette smoke, in that smokers who report cough
and phlegm tend to smoke more heavily than smokers who deny
these symptoms, and measures such as numbers of cigarettes smoked
per day are not precise enough to control adequately for the amount
of smoke exposure. The rate, number, and volume of puffing as well
as the depth of inhalation can vary substantially between smokers
and are important additional measures of cigarette smoke exposure
dose.

73

480-144 0 - 85 - 4


15

2
2
a
5

ru
3
3
R 5
I
t

0

Cough     Phlegm

Chronfc bronchlbs

Men

15 -









10 --









5 --

0  Absent

Cough

Phlegm   Chrorw bronchltls


CHRONIC AIRFLOW OBSTRUCTION

Introduction

Airflow obstruction is the physiol,ogicnl consequence of disease
processes that narrow the airway. ln asthma the obstruction is
reversible with pharmacologic bronchodilation, whereas the obstruc-
tion associated with airways damage and emphysema is often not
reversible. The terminology with regard to permanent airflow
obstruction has varied. The 1958 Ciba Foundation Guest Symposium
proposed "generalized obstructive lung disease," which was subdivid-
ed into "asthma" and "irreversible or persistent obstructive lung
disease" (1959); in the 1962 recommendations of the American
Thoracic Society, "chronic obstructive bronchitis" was the only
definition that mentioned abnormality of expiratory flow (American
Thoracic Society 1962). In 1975, a joint committee of the American
College of Chest Physicians and the American Thoracic Society
recommended the term "chronic obstructive pulmonary disease"
(American College of Chest Physicians and American Thoracic
Society 1975). Thurlbeck (1976, 1977) has advocated the use of
"chronic airflow obstruction," a functionally based definition that
does not specify the underlying disease processes. Previous Reports
of the Surgeon General have used varying terminology, including
"chronic bronchopulmonary disease" in 1964, "chronic obstructive
bronchopulmonary disease" in 1971, and "chronic obstructive lung
disease" in 1979 (USPHS 1964,197l; USDHEW 1979).

These definitions, however, cannot be readily applied to identify
specific populations. Physiologists, epidemiologists, and clinicians
often use differing approaches in determining whether airflow
obstruction is present (Fletcher 1978). Physiologists, with the
capability for making sophisticated laboratory measurements of
airflow obstruction, may regard subtle early abnormalities of flow as
definitive. In the community, epidemiologists have generally used
spirometry as the primary method for assessing airflow obstruction.
For epidemiologic purposes, airflow obstruction is usually defined by
a forced expiratory volume in 1 second (FEV1) less than a particular
level after standardization for sex, age, and height, or by a ratio of
the FEVl to the forced vital capacity (FVC) below a specified value.
Tests of forced exhalation, such as the FEVl, have the advantage of
sensitivity to abnormalities of both the lung parenchyma and the
airways (Mead 1979). Clinicians are more likely to detect and
diagnose airflow obstruction when it is advanced and symptomatic.
As would be anticipated, the differing approaches of physiologists,
epidemiologists, and clinicians may lead to differing estimates of the
frequency of airflow obstruction.

The natural history of chronic airflow obstruction in adults has
been partially described by several recent prospective investigations:

76