Eur Respir J 1997; 10: 1115–1119
Printed in UK - all rights reserved
Copyright ERS Journals Ltd 1997
European Respiratory Journal
ISSN 0903 - 1936
A longitudinal analysis of pulmonary function in rats
during a 12 month cigarette smoke exposure
J.L. Wright*, J-P. Sun*, S. Vedal**
A longitudinal analysis of pulmonary function in rats during a 12 month cigarette smoke
exposure. J.L. Wright, J-P. Sun, S. Vedal. ERS Journals Ltd 1997.
ABSTRACT: We wanted to examine the longitudinal effects of chronic cigarette
smoke exposure, and to determine whether the chronic alterations in pulmonary
function induced by long-term cigarette smoke exposure in an animal model could
be predicted by initial or early alterations in function.
A group of Sprague Dawley rats was exposed to the smoke of 7 cigarettes·day-1
for 5 days·week-1during a total period of 12 months. Lung volume, flow-volume
curves and pressure-volume curves were recorded at baseline, and after 2, 4, 8
and 12 months of smoke exposure. A control group of rats was subjected to the
same regimen of testing, but was not exposed to smoke.
Thirteen rats completed the study in the smoke-exposed group and seven rats
in the control group. We found that chronic exposure to cigarette smoke produced
early abnormalities in pulmonary function, with the forced expiratory volume in
one second/forced vital capacity (FEV1/FVC) ratio showing an acceleration of age-
ing effect, particularly between 4 and 8 months of exposure. In this model,
although the two groups had significantly different airflow after 12 months, the
initial values in each group were remarkably similar, and we could not identify
any pulmonary function test which had predictive value.
We conclude that longitudinal studies of cigarette smoke exposure in this rat
model allow better characterization of the nature and time course of the effects of
smoking on the lung.
Eur Respir J 1997; 10: 1115–1119.
Depts of *Pathology and **Medicine, Uni-
versity of British Columbia, Vancouver,
Correspondence: J.L. Wright
Dept of Pathology
2211 Wesbrook Mall
B.C. V6T 2B5
Keywords: Cigarette smoke
pulmonary function tests
Received: July 16 1996
Accepted after revision January 14 1997
Supported by grants from the B.C. Health
Research Foundation and the B.C. Lung
In our laboratory, we have previously developed a
guinea-pig model of cigarette smoke-induced alterations
in pulmonary function [1, 2]. Although this model has
proved successful in that we have been able to show
clear evidence of airflow obstruction and emphysema-
tous lung destruction in smoke-exposed animals, it has
also illustrated the problems inherent in such kinds of
cross-sectional studies. In the guinea-pig, there was sig-
nificant interanimal variation in pulmonary function,
even when accounting for animal age and weight .
Furthermore, we found that our cross-sectional studies
were biased by a "healthy smoker" effect, whereby the
animals that were most affected by cigarette smoke
tended to die before the end of the experiment; the final
results were, thus, derived from the animals with the
least disease [2, 4]. Finally, the use of a single pul-
monary function end-point does not allow for analysis
of the temporal profile of the development of abnor-
malities, or for analysis of results that might predict
response to cigarette smoke.
It was for these reasons that, in the present study, a
longitudinal study design was chosen, whereby each
animal was used as its own control and function tests
were performed at multiple time-points, thus making it
possible to track each animal's function by calculating
a percentage change from control values. This type of
longitudinal analysis also had the advantage that it was
possible to ascertain whether any baseline or early func-
tion test could be predictive of ultimate function. The
primary objective of this study was to characterize the
long-term effect of cigarette smoke exposure on a large
array of pulmonary function tests that reflect different
pathophysiological derangements in the lung. A sec-
ondary objective was to ascertain whether any pulmon-
ary function test could be predictive of final pulmonary
Although other workers have reported success in intu-
bation of guinea-pigs [5, 6], we found this technique
hampered by pronounced laryngospasm, resulting in an
unacceptable death rate. We therefore performed our
study in the Sprague Dawley rat, an animal which is eas-
ily intubated. Although we had previously shown that
there was little variation in pulmonary function tests
repeated at the same time , we had not been able to
test reproducibility over a period of several days. An
initial objective of the present study was, therefore, to
show this reproducibility, thereby allowing for the long-
itudinal analysis of pulmonary function.
To ascertain whether the pulmonary function tests used
in this model were reproducible, 10 Sprague Dawley
rats weighing 400–510 g (mean±SD 447±30 g) were uti-
lized. In this portion of the study, three sets of pulmo-
nary function tests were performed: baseline, Day two,
and Day three.
In the 12 month study, Sprague Dawley rats that ini-
tially weighed 317±18 g were used. Baseline function
tests were performed, and the animals were then ran-
domly assigned to either control (non-smoke-exposed)
or smoke-exposed groups. Pulmonary function was
analysed at baseline, 2, 4, 8 and 12 months.
Pulmonary function tests
The animals were anaesthetized using a combination
of intraperitoneal diazepam (5 mg·kg-1) and intramus-
cular Innovar Vet (0.4 mL·kg-1), and intubated with a
14 gauge intravenous cannula , using a paediatric
laryngoscope and a tilted table. After intubation, the rats
were placed in a pressure sensitive 7.5 L small animal
plethysmograph, and ventilated at 80 breaths·min-1with
a tidal volume of approximately 1.5 mL. A water-filled
oesophageal tube (PE 240) with a multiholed tip was
used to measure pleural pressure; transpulmonary pres-
sure was calculated as the difference between mouth
and pleural pressure.
Pulmonary function tests were performed according
to our usual protocol . After measurement of func-
tional residual capacity (FRC), the rats were given sup-
plementary doses of Innovar Vet, and were rendered
apnoeic by hyperventilation. Pressure-volume curves
were constructed by deflation to -30 cmH2O (expira-
tory reserve volume (ERV)), inflation to +30 cmH2O,
and deflation to -30 cmH2O (vital capacity (VC)). Two
inflation-deflation procedures were performed prior to
measurement of the curve, and calculation of lung vol-
umes. Static lung compliance (CL,st) was calculated
between FRC and 10 cmH2O transpulmonary pressure.
A flow-volume curve was constructed by inflation to
+30 cmH2O, and rapid deflation to -50 cmH2O pres-
sure. From this curve, forced vital capacity (FVC),
forced mid-expiratory flow (FEF25–75), forced expira-
tory volume in one second (FEV1), and peak expira-
tory flow (PEF) were calculated. The flow volume curve
itself was plotted, using the volumes at 95–30% total
lung capacity (TLC) as the independent variable.
Following the pulmonary function tests, the animals
were kept warm, allowed to recover from the anaes-
thetic, and were extubated. They were housed in a lam-
inar hood in standard rat cages with sterile paper pellets
as bedding, and were allowed free access to rat chow
Cigarette smoke exposure
We utilized our standard smoke exposure system
. In this system, exposure to the smoke of 7 cigarett-
es·day-1causes a chronic carboxyhaemoglobin level of
appro-ximately 4%, a value similar to that found in human
chronic cigarette smokers , and has been shown to
produce emphysema in guinea-pigs . In brief, the sys-
tem consists of a vented nose only chamber, into which
serial aliquots of 20 mL of fresh cigarette smoke were
injected, with each cigarette consumed over a period of
approximately 10 min. The animals received the smoke
of 7 cigarettes·day-1for 5 days·week-1during a 12 month
total time period.
The data were analysed using the SYSTAT  and
Statistical Analysis System (SAS)  statistical pack-
ages. A multiple analysis of variance (MANOVA) was
used both to analyse the reproducibility of the pulmon-
ary function measurements over 3 days, and to test for the
effect of cigarette smoke exposure on the repeated mea-
surements over 12 months. For reproducibility, Hotell-
ing's T2statistic was used to determine whether significant
differences in the three lung function measurements (Day
0, 2 and 3) were present. The Hotelling T2test is the ana-
logue of the paired t-test for the case where more than
two measurements are made on each experimental ani-
mal, and accounts for the correlation between tests with-
in each animal.
For testing the smoke effect, the five pulmonary func-
tion measurements (Month 0, 2, 4, 8 and 12) were mod-
elled as a function of time and group (smoke-exposed
and control). An effect of cigarette smoking was deem-
ed to be present when a statistically significant inter-
action between time and group was present; i.e. there
was an effect of smoking when the effects of time dif-
fered between the smoke-exposed and control groups.
A "profile" contrast analysis  was performed to
identify for which of the four adjacent month pairs (0–2,
2–4, 4–8, and 8–12 months) significant effects for time
and smoking were present. Similar analyses were per-
formed after dividing each pulmonary function mea-
surement by the animal's weight at the time of the
measurement. A Bonferroni correction was applied to
account for the four comparisons made in the profile
analysis. A p-value of less than 0.05 was considered
Table 1 presents the mean±SD values for the lung vol-
umes, which were derived from the pressure-volume cur-
ves (data not shown), whilst table 2 presents the mean±
SD values for the flow data, which were calculated from
the flow-volume curve. Each table also indicates the p-
value corresponding to the Hotelling's T2statistic. No
J.L. WRIGHT ET AL.
Table 1. – Reproducibility study: lung volume analysis
Test Day 1 Day 2 Day 3 p-value#
CL,st mL·cm-1 0.92±0.18
Values are presented as mean±SD. VC: vital capacity; FRC:
functional residual capacity; RV: residual volume; CL,st: sta-
tic lung compliance. #: from repeated measures analysis.
significant intra-animal differences were found for any
of the parameters. The only test in which a near sig-
nificant difference (p=0.07) was found was the residual
volume (RV), which tended to decrease slightly in Days
2 and 3. The flow data were stable over the 3 days, with
p-values ranging from 0.56 (FEV1/FVC) to 0.77
Twelve month longitudinal analysis
At the termination of the study, there were seven rats
remaining in the control group and 13 remaining in the
smoke-exposed group. Table 3 presents the data, ex-
pressed as percentage baseline value, for the control and
smoke-exposed groups at each of the pulmonary func-
tion test times. Based on the raw values for each test,
the MANOVA showed an overall effect of time for each
of the pulmonary function tests (p<0.001), and these
effects remained when the analysis was repeated using
weight as a corrective factor (values divided by weight).
There was a significant overall smoke exposure effect
for FRC, FEV1, and FEV1/FVC, as well as for weight
(all p<0.001), using the raw function values. After
weight correction, an overall smoke exposure effect was
also found for vital capacity (VC), RV, TLC, CL,st, PEF
and FVC (all p<0.001), but statistical significance for
effects on FEV1 and FEV1/FVC were lost.
Significant interactions between time and smoke were
found for VC (p<0.02), FRC (p<0.001), RV (p<0.02),
TLC (p<0.001), weight (p<0.001), FEV1, (p<0.05) and
FEV1/FVC (p<0.05), when the raw data values were
examined. After weight correction, statistically sign-
ificant interactions persisted for VC (p<0.001), FRC
(p<0.001), RV (p<0.05), TLC (p<0.001), and further
interactions were identified for CL,st (p<0.006), PEF
(p<0.002), FEF25–75(p<0.003) and FVC (p<0.001), while
statistical significance for FEV1 and FEV1/FVC was
The profile analysis was performed to identify the
time periods between which effects occurred. Using the
raw values, an overall time effect on lung volumes was
found between baseline and 2 months for VC, RV, TLC,
CL,st (p<0.001) and FRC (p<0.01); after weight correc-
tion, effects remained for FRC (p<0.01), TLC (p<0.05)
and RV (p<0.001). Time effect on flow values between
baseline and 2 months was found on the raw values
for PEF, FEF25–75, FVC, FEV1/FVC (p<0.001); after
weight correction, effects remained (p<0.001) for all
but FVC, and became apparent for FEV1 (p<0.001).
A time effect was found between 2 and 4 months on the
raw values for VC, TLC, CL,st, FRC and FVC (p<0.001);
after weight correction, a time effect remained for FVC
(p<0.05) only, and was joined by a significant effect for
A time effect between 4 and 8 months was found for
VC, TLC, CL,st and FRC (p<0.001), RV (p<0.01) and
FVC (p<0.001), and FEV1/FVC (p<0.02); after weight
correction, the time effect on FVC (p<0.05) and FEV1/
FVC (p<0.001) remained (p<0.01) and was joined by a
significant effect on FEV1 (p<0.01).
A time effect between 8 and 12 months was found for
FRC (p<0.001), which was removed after weight cor-
rection, but replaced by an effect on VC (p<0.05). Time
effects between 8 and 12 months were also found for
FEV1, and FEV1/FVC (p<0.001) and PEF (p<0.05) raw
data, and these effects remained after weight correction
and were joined by a significant effect on FEF25–75
(p<0.001). A significant time effect was found for weight
between all adjacent time pairs (p<0.001).
As shown in table 3, differences in lung volumes and
CL,st between the control and smoke-exposed animals
(evidence of cigarette smoke effect) were found in the
LONGITUDINAL ANALYSIS OF PULMONARY FUNCTION
Table 2. – Reproducibility study: flow rate analysis
Test Day 1 Day 2 Day 3 p-value#
FEF25–75 mL·s-1 122±15
Values are presented as mean±SD. PEF: peak expiratory flow;
FEF25–75: forced mid-expiratory flow; FEV1: forced expira-
tory volume in one second; FVC: forced vital capacity. #: from
repeated measures analysis.
Table 3. – Pulmonary function (as % baseline) over 12 months in control and smoke-exposed rats
2 months 4 months 8 months 12 months
p-value p-value p-value p-value
C S-E 0–2 mo C S-E 2–4 mo C S-E 4–8 mo C S-E 8–12 mo
Values are presented as percentage of baseline value±SD. C: control group; S-E: smoke-exposed group; mo: months; TLC: total
lung capacity; NS: nonsignificant. For further definitions see legends to table 1 and 2. Statistical significance: profile analysis
(analysis of the differences between C and S-E groups in the pulmonary function changes between months). The first symbol
represents analysis using raw values; the second symbol represents analysis using weight correction. *: p<0.05; **: p<0.01; ***:
baseline to 2 month time period, and these were gen-
erally enhanced by weight correction. Differences bet-
ween groups were present in the 2–4 month time period
only for CL,st and FRC, but these were no longer sta-
tistically significant after weight correction. Differences
between groups were present in the 4–8 month time
period for FRC and TLC, which remained after weight
correction. For the 8–12 month time period, a differ-
ence in raw values was present for TLC; after weight
correction this difference was no longer statistically sig-
nificant, but a difference for RV reached statistical sig-
When the flow rates were examined, differences in
the 0–2 time period were found in FVC raw values,
which was joined after weight correction by differences
in PEF and FEF25–75. Significant differences between
control and smoke-exposed animals were found in the
8–12 month time period for FEV1 and FEV1/FVC, both
with and without weight correction. The effect of smoke
on FEV1 and FEV1/FVC seemed to occur between 4 and
8 months (p=0.04 and p=0.05, respectively, before Bon-
ferroni correction), whereas the ageing effect in the
control animals occurred between 8 and 12 months. A
smoke effect on weight was only present between base-
line and 2 months.
Predictive value of tests
To test whether any of the flow tests had predictive
value, we examined whether the initial test value, the
test value at the 2 month time-point, or the slope of the
line between initial and 2 month tests were able to pre-
dict the flow test values at the 12 month time-point. No
such relationships could be found.
During the course of the experiment, five control and
six smoke-exposed rats died, usually because of post-
anaesthetic bronchospasm. There were no differences
with surviving animals either in baseline function or
function tests performed prior to death.
This study utilized a longitudinal format, with pulmo-
nary function tests performed at multiple time-points,
after first confirming that such tests can be reproducibly
performed. We have shown that pulmonary function in
the rat normally changes over a 12 month time period,
most likely in relation to growth of the animals. This is
particularly true of the lung volumes, which increased
over time, as did the CL,st. By contrast, although the
FEF25–75 and FVC tended to increase slightly over time,
the FEV1 and FEV1/FVC decreased during the 8–12
month time period, suggesting an ageing effect. As
discussed below, cigarette smoke appears to accelerate
this ageing effect, much in the same manner as has re-
cently been found in humans .
Although rats have previously been shown to deve-
lop emphysema and airflow obstruction after exposure
to cigarette smoke , no previous study has utilized
a longitudinal design with the extensive pulmonary
function test analysis reported in the present study.
Using this regimen, we have shown that chronic ciga-
rette smoke exposure produced airflow obstruction, albeit
not to the same degree as that seen in the guinea-pig
model . The VC and FVC of the smoke-exposed ani-
mals were increased compared to the control animals,
findings which coincided with an effect of smoke on
increasing the CL,st. The profile analysis showed that
this effect on CL,st was significant between baseline
and 2 months, and between 2 and 4 months. In humans,
VC is usually decreased in patients with chronic ob-
structive pulmonary disease (COPD); however, the para-
doxical increase seen in the animal models  can be
explained by the fact that the animals are inflated to a
certain transpulmonary pressure, rather than by active in-
spiration. Thus, a small increase in lung CL,st will result
in an increased VC.
Analysis of the flow data showed that, overall, FEV1,
decreased between 4 and 12 months, and there was a
significant overall decrease in the smoke-exposed ani-
mals. The decrease in the smoke-exposed animals ap-
peared to occur earlier compared to the control animals,
which showed a decrease between 8 and 12 months.
FEV1/FVC also decreased over the total experimental
time period, and there was a significant overall effect
of smoke exposure, which occurred primarily between
4–8 months. By contrast, in the control animals, the nor-
mal decrease in airflow associated with ageing was
found between 8 and 12 months.
There is an obvious time disparity between the effect
of smoke on the lung volumes and compliance, and the
effect of prolonged smoke exposure on airflow, with the
former occurring early and the latter occurring later in
the exposure time course. Alterations of the lung matrix
could be one explanation for this. MACKLEMand EIDELMAN
 and LAROS and KUYPER  have suggested that
there is a dynamic alteration of the lung collagen in
response to chronic cigarette smoke exposure, which
may be manifested in abnormal pulmonary physiology.
SNIDER et al.  developed a model of emphysema
induced by cadmium chloride, and found decreased for-
ced expiratory flow with an increase in lung collagen
as measured biochemically. SUGIHARA et al.  exam-
ined length-tension curves of lung tissue fragments and
demonstrated that emphysematous alveolar walls had
increased length, but decreased distensibility. MACKLEM
and EIDELMAN  showed that patients with COPD had
an increased specific elastance and, therefore, requir-
ed a greater inflation pressure to produce volume in-
In our guinea-pig model of chronic smoke exposure
, we have previously shown that the lung parenchy-
mal matrix components are bimodally altered by ciga-
rette smoke exposure, with a decrease in collagen after
1 month of exposure, followed by an increase in colla-
gen matrix components after 6 and 12 months .
Physiologically, the animals developed initial altera-
tions in lung volumes, followed in the latter months by
decreased flow rates. The present study also shows
alterations in lung volumes, with an overall decrease in
flow rates. Thus, it seems likely that a similar process
of matrix remodelling is occurring in this rat model.
Matrix is known to be altered throughout life ; the
physiological changes seen in the smoke-exposed rats
may be analogous to a process of accelerated ageing,
such as that found in human cigarette smokers .
J.L. WRIGHT ET AL.
The smoke-exposed animals weighed less than the
control animals at all of the time-points. This pheno-
menon was previously found in our guinea-pig model of
chronic cigarette smoke exposure , and has recently
also been found in rats exposed chronically to nicotine
vapour . In humans, there is a known relationship
between lung function and somatic growth , and we
have previously shown that both lung volumes and flow
rates are related to weight in the guinea-pig model .
When we examined the data using weight as a correc-
tive factor, we found that the majority of the overall
effects of time remained significant, and that the pro-
file analysis again suggested that cigarette smoke exer-
ted its effect on lung volumes early in the experimental
time course, but between the last two time periods for
FEV1 and FEV1/FVC.
Animals, like humans, show a wide degree of sus-
ceptibility to deleterious agents . In humans, this pro-
duces the "healthy worker" effect, i.e. workers who are
particularly susceptible will drop out of the workforce
and, hence, not appear in subsequent cross-sectional
analyses [21, 22]. The healthy worker effect, thus, tends
to underestimate the adverse effects of specific expos-
ures. Although animals do not have the ability to remove
themselves from exposure to test agents, those indi-
viduals who are very susceptible may die before the
completion of the experiment [2, 4], and thus "remove"
themselves from any statistical analysis, producing a
similar situation. In the present study, a comparison of
measured characteristics of the animals who died with
those of the living animals at the same time-point, showed
no significant differences. Thus, unlike the guinea-pig
model, the "healthy smoker" effect is a not a feature of
the rat model.
An additional potential feature of a longitudinal ana-
lysis is the ability to ascertain whether any one test is
able to predict final function . However, one of the
problems in using an inbred animal model is that these
animals tend to have a stereotypical response to injury.
Although there was some variation in the initial pul-
monary function tests, all smoke-exposed animals suf-
fered a similar percentage decrement in function by the
final test time. Thus, we were unable to find any test
of predictive value.
In summary, longitudinal analysis of pulmonary func-
tion in the rat model has shown that cigarette smoke ex-
erts its effects early on lung volumes and compliance,
while flow alterations occur later, consistent with a pic-
ture of accelerated ageing. This type of experimental
design should prove useful in better characterizing the
time course and nature of cigarette smoking effects.
1.Wright JL, Churg A. Cigarette smoke causes physio-
logic and morphologic changes of emphysema in the
guinea-pig. Am Rev Respir Dis 1990; 142: 1422–1428.
Wright JL, Sun J-P. The effect of smoking cessation on
pulmonary and cardiovascular function and structure.
J Appl Physiol 1994; 76: 2163–2168.
Wiggs BR, Churg A, Wright JL. Influence of weight on
pulmonary function in the adult guinea-pig. Respira-
tion 1991; 58: 37–41.
Wright JL. The relationship of increased pulmonary artery
pressure and airflow obstruction to emphysema. J Appl
Physiol 1993; 74: 1320–1324.
Blouin A, Cormier Y. Endotracheal intubation in guinea-
pigs by direct laryngoscopy. Lab Animal Sci 1987; 37:
Kujime K, Natelson BH. A method for endotracheal
intubation of guinea-pigs. Lab Animal Sci 1981; 31:
Prignot J. Quantification and chemical markers of to-
bacco exposure. Eur J Respir Dis 1987; 70: 1–7.
Wilkinson L. In: SYSTAT: the system for statistics.
Evanston, IL, SYSTAT Inc., 1988.
SAS Institute Inc. In: SAS/STAT Users Guide. Cary,
NC, SAS Institute, 1985.
van Ende CN. Repeated measures analysis: growth and
other time-dependent measures. In: Scheiner SM, Gure-
vitch J, eds. Design and Analysis of Ecological Ex-
periments. New York, Chapman and Hall, 1993.
Gold DR, Wang X, Wypij D, Speizer FE, Ware JH,
Dockery DW. Effects of cigarette smoking on lung
function in adolescent boys and girls. N Engl J Med
1996; 335: 931–937.
Huber GL, Davies P, Zwilling GR, et al. A morpho-
logic and physiologic bioassay for quantifying alterations
in the lung following experimental chronic inhalation
of tobacco smoke. Clin Respir Physiol 1981; 17: 269–
Macklem PT, Eidelman D. Re-examination of the elas-
tic properties of emphysematous lungs. Respiration 1990;
Laros CD, Kuyper CMA. The pathogenesis of pulmon-
ary emphysema. Respiration 1976; 33: 325–348.
Snider GL, Lucey EC, Faris B, Jung-Legg Y, Stone PJ,
Franzblau C. Cadmium chloride-induced airspace en-
largement with interstitial pulmonary fibrosis is not
associated with destruction of lung elastin. Am Rev Res-
pir Dis 1988; 137: 918–923.
Sugihara T, Martin CJ, Hildebrandt J. Length-tension
properties of alveolar wall in man. J Appl Physiol 1971;
Wright JL, Churg A. Smoke-induced emphysema in the
guinea-pig is associated with morphometric evidence of
collagen breakdown and repair. Am J Physiol 1995; 268:
Turino GM. The lung parenchyma: a dynamic matrix.
Am Rev Respir Dis 1985; 132: 1324–1334.
Waldum HL, Nilsen T, Rorvik H, et al. Long-term
effects of inhaled nicotine. Life Sci 1996; 58: 1339–
Sherrill DL, Camilli A, Lebowitz MD. On the tempo-
ral relationships between lung function and somatic
growth. Am Rev Respir Dis 1989; 140: 638–644.
Becklake M. Epidemiological evidence of varying sus-
ceptibility to inhaled substances. Pharmacogen 1991; 1:
Becklake MR. Chronic airflow limitation: its relation-
ship to work in dusty occupations. Chest 1985; 88:
Burrows B, Knudson RJ, Camilli AE, Lyle SK, Lebowitz
MD. The "horse-racing effect" and predicting decline in
forced expiratory volume in one second from screening
spirometry. Am Rev Respir Dis 1987; 135: 788–793.
LONGITUDINAL ANALYSIS OF PULMONARY FUNCTION