ArticlePDF Available

The influence of smoking on vitamin C status in adults

Authors:

Abstract and Figures

To further define the relation between smoking and vitamin C status, the dietary and serum vitamin C levels of 11,592 respondents in the second National Health and Nutrition Examination Survey (NHANES II) were analyzed. Smokers of 20 cigarettes daily had the lowest vitamin C dietary intake (79 mg, 95% CI:73, 84) and serum levels (0.82 mg/dl, 95% CI: 0.77, 0.86; 46.6 mumol/L, 95% CI: 43.7, 48.8), while smokers of 1-19 cigarettes daily had decreased vitamin C intake (97 mg; 95% CI: 90, 104 mg) and serum levels (0.97 mg/dl, 95% CI: 0.92, 1.03; 55.1 mumol/L, 95% CI: 52.2, 58.5) compared to respondents who had never smoked (109 mg, 95% CI: 105, 113 and 1.15 mg/dl, 95% CI: 1.11, 1.18; 65.3 mumol/L, 95% CI: 63.0, 67.0, respectively). This inverse association between both intake and serum levels of vitamin C and smoking was independent of age, sex, body weight, race, and alcoholic beverage consumption. Following further adjustment for dietary vitamin C intake, the negative correlation between cigarette smoking and serum vitamin C levels persisted. The risk of severe hypovitaminosis C (serum levels less than or equal to 0.2 mg/dl; 11.4 mumol/L) was increased in smokers, particularly when not accompanied by vitamin supplementation (odds ratio 3.0, 95% CI: 2.5, 3.6). These data suggest that even though smoking adversely affects preferences for vitamin C rich foods, the inverse association between smoking and serum vitamin C levels occurs independently of dietary intake.
Content may be subject to copyright.
The
Influence
of
Smoking
on
Vitamin
C
Status
in
Adults
GORDON
SCHECTMAN,
MD,
JAMES
C.
BYRD,
MD,
AND
HARVEY
W.
GRUCHOW,
PHD
Abstract:
To
further
define
the
relation
between
smoking
and
vitamin
C
status,
the
dietary
and
serum
vitamin
C
levels
of
11,592
respondents
in
the
second
National
Health
and
Nutrition
Examina-
tion
Survey
(NHANES
II)
were
analyzed.
Smokers
of
20
cigarettes
daily
had
the
lowest
vitamin
C
dietary
intake
(79
mg,
95%
CI:73,
84)
and
serum
levels
(0.82
mg/dl,
95%
CI:
0.77,
0.86;
46.6
jimol/L,
95%
CI:
43.7,
48.8),
while
smokers
of
1-19
cigarettes
daily
had
decreased
vitamin
C
intake
(97
mg;
95%
CI:
90,
104
mg)
and
serum
levels
(0.97
mg/dl,
95%
CI:
0.92,
1.03;
55.1
,umol/L,
95%
CI:
52.2,
58.5)
compared
to
respondents
who
had
never
smoked
(109
mg,
95%
CI:
105,
113
and
1.15
mg/dl,
95%
CI:
1.11,
1.18;
65.3
,iLmol/L,
95%
CI:
63.0,
67.0,
Introduction
Preliminary
reports
have
indicated
adverse
affects
of
smoking
on
vitamin
C
metabolism"2
and
more
recent
studies
have
confirmed
that
serum
vitamin
C
levels
are
lower
in
cigarette
smokers
than
non-smokers.
3
These
observations
suggest
that
smoking
directly
lowers
serum
vitamin
C
levels,8
although
other
factors
including
gender,4'9
age,4"10
and
ethanol
consumption'
113
have
also
been
reported
to
influence
serum
vitamin
C
concentrations.
However,
since
cigarette
smoking
is
not
independent
of
age,
sex,
or
alcohol
use,14
the
relation
between
smoking
and
vitamin
C
status
requires
evaluation
controlling
for
these
potentially
confounding
variables.
Sev-
eral
previous
reports
have
partially
addressed
these
concerns
by
stratifying
participants
according
to
age,7'15"16
sex,15"16
and
level
of
vitamin
C
intake."'
Within
each
stratum,
vitamin
C
levels
were
decreased
in
smokers,
suggesting
that
the
rela-
tionship
may
indeed
be
independent
of
these
factors.
A
univariate
analysis
of
the
United
States
second
National
Health
and
Nutrition
Examination
Survey
(NHANES
II)
was
consistent
with
these
findings,
as
serum
vitamin
C
levels
were
lower
in
smokers
for
each
level
of
dietary
vitamin
C
intake.
17
The
purpose
of
this
study
was
to
better
define
the
independent
relationship
between
cigarette
smoking
and
vita-
min
C
status,
while
controlling
for
potentially
confounding
variables
using
multivariate
statistical
techniques
on
data
from
NHANES
II,
conducted
from
1976-80.
NHANES
II
is
a
comprehensive
population
survey
which
included
serum
vi-
tamin
C
measurements
and
a
detailed
dietary
questionnaire
from
which
nutrient
intakes
were
calculated
in
a
sample
of
over
11,000
adults.
The
association
of
cigarette
smoking
with
serum
vitamin
C
levels
was
analyzed
while
simultaneously
adjusting
for
variables
which
significantly
influenced
serum
vitamin
C
concentrations
including
age,
sex,
race,
dietary
vitamin
C
intake,
ethanol
consumption,
and
body
mass
index
(BMI).
In
addition,
the
impact
of
smoking
on
the
risk
of
marginal
or
frankly
deficient
serum
vitamin
C
levels
was
determined.
From
the
Divisions
of
General
Internal
Medicine
and
of
Biostatistics/
Clinical
Epidemiology,
Medical
College
of
Wisconsin.
Address
reprint
re-
quests
to
Gordon
Schectman,
MD,
Assistant
Professor,
Division
of
General
Internal
Medicine,
Medical
College
of
Wisconsin,
8700
West
Wisconsin
Avenue,
Milwaukee,
WI
53226.
This
paper,
submitted
to
the
Journal
February
18,
1988,
was
revised
and
accepted
for
publication
June
13,
1988.
©
1989
American
Journal
of
Public
Health
0090-0036/89$1.50
respectively).
This
inverse
association
between
both
intake
and
serum
levels
of
vitamin
C
and
smoking
was
independent
of
age,
sex,
body
weight,
race,
and
alcoholic
beverage
consumption.
Following
further
adjustment
for
dietary
vitamin
C
intake,
the
negative
corre-
lation
between
cigarette
smoking
and
serum
vitamin
C
levels
per-
sisted.
The
risk
of
severe
hypovitaminosis
C
(serum
levels
<
0.2
mg/dl;
11.4
,umol/L)
was
increased
in
smokers,
particularly
when
not
accompanied
by
vitamin
supplementation
(odds
ratio
3.0,
95%
CI:
2.5,
3.6).
These
data
suggest
that
even
though
smoking
adversely
affects
preferences
for
vitamin
C
rich
foods,
the
inverse
association
between
smoking
and
serum
vitamin
C
levels
occurs
independently
of
dietary
intake.
(Am
J
Public
Health
1989;
79:158-162.)
Methods
The
NHANES
II
probability
sample
was
selected
so
that
population
subgroups
of
special
interest
for
nutritional
as-
sessment
were
oversampled.
Among
adults,
the
oversampled
groups
included
persons
over
60
years
old
and
persons
below
poverty
level.
Sample
weights
were
then
computed
for
these
subgroups
so
that
the
total
sample
closely
estimated
the
noninstitutionalized
civilian
population.
This
sample
weight-
ing
reflects
the
individual
selection
probability,
adjustments
for
non-response,
and
poststratification
adjustments.
A
com-
plete
description
of
the
survey
methods
has
been
reported.'8
Adults
ages
18 to
74
years
who
attended
a
second
examination
visit
following
a
preliminary
interview
were
studied.
However,
respondents
who
did
not
have
serum
vitamin
C
concentrations
measured
(n=866),
who
did
not
record
alcoholic
beverage
consumption
(n=40),
and
who
did
not
complete
smoking
histories
(n=6)
were
excluded
from
the
analysis.
After
these
exclusions,
11,592
respondents
re-
mained
eligible
for
the
study.
Appropriate
sample
weighting
was
used
to
adjust
for
possible
non-response
bias
introduced
because
of
failure
to
attend
the
second
examination
visit.'9
Those
who
attended
the
second
visit
but
who
were
subse-
quently
excluded
(n=9
12)
were
similar
to
included
subjects
in
the
per
cent
who
were
male
(46.6
vs
47.4
per
cent),
White
(82
vs
89
per
cent),
currently
smoking
(32
vs
34
per
cent),
and
in
their
mean
age
(46.7
vs
46.2
years).
Nutrient
intakes
were
estimated
from
24-hour
recall
interviews.
Three-month
food
frequency
reports
provided
recent
intake
of
individual
food
items,
including
estimates
for
beer,
wine,
and
liquor.
Weekly
frequencies
of
beer,
wine,
and
liquor
consumption
were
summed
to
estimate
ethanol
con-
sumption.
The
use
of
vitamin
and
mineral
supplements
was
determined
by
the
question
"Are
you
taking
vitamins
or
supplements?"
The
vitamin
content
of
supplements
was
excluded
from
the
nutrient
intake
summaries
and
therefore
the
calculated
vitamin
C
intake
underestimates
actual
vitamin
C
consumption
in
those
respondents
taking
supplements.
For
this
reason,
the
use
of
supplements
was
included
as
a
separate
variable
in
the
multivariate
analysis
assessing
serum
vitamin
C
levels.
NHANES
II
participants
were
originally
divided
into
racial
categories
of
"White",
"Black",
and
"other".
Indi-
viduals
of
Oriental
extraction,
American
Indians,
and
Eski-
mos
were
listed
in
the
"other"
category
and
accounted
for
less
than
1
per
cent
of
the
total
study
cohort;
they
were
AJPH
February
1989,
Vol.
79,
No.
2
158
SMOKING
AND
VITAMIN
C
STATUS
included
with
Whites
in
this
study.
Body
mass
index
(BMI)
was
measured
as
weight/height2
(kg/M2).
Serum
vitamin
C
levels
were
measured
in
a
central
laboratory
at
the
Centers
for
Disease
Control
(CDC)
by
the
dinitrophenylhydrazine
method.20
Criteria
from
the
Nutri-
tion
Canada
Survey
were
used
to
classify
risk
of
clinical
scurvy.2'
Respondents
with
serum
vitamin
C
levels
of
0.2
mg/dl
(11.4
,umoVL)
or
less
were
considered
to
be
at
high
risk
for
clinical
vitamin
C
deficiency
(severe
hypovitaminosis
C),
while
those
with
serum
concentrations
of
0.4
mg/dl
(22.7
,umoVL)
or
less
were
considered
to
be
at
marginal
risk
(marginal
hypovitaminosis
C).
Variance
estimates
were
corrected
for
the
complex
sample
design
using
the
SUPERCARP
program.22
Multiple
regression
was
used
to
determine
the
independent
contribu-
tion
of
a
variable
to
serum
vitamin
C
levels,
while
simulta-
neously
adjusting
for
all
other
variables.23
An
order
of
entry
was
determined
using
a
stepwise
regression
model,
however,
regression
coefficients
were
computed
following
the
insertion
of
all
variables
into
the
model.
For
the
regression
equations,
smoking
status
was
categorized
into
six
levels:
never
smoked,
stopped
for
the
previous
year
or
longer
(DC
>
1
year),
stopped
within
the
past
year
(DC
<
1
year),
smoking
less
than
20
cigarettes daily
(<1
PPD;
mean
9.3
cigarettes),
smoking
20
cigarettes
daily
(1
PPD),
and
smoking
more
than
20
cigarettes
daily
(>1
PPD;
mean
35.1
cigarettes).
This
was
done
to
compare
serum
vitamin
C
levels
in
cohorts
of
subjects
who
were
currently
smoking,
had
stopped
smoking,
or
who
had
never
smoked.
In
additional
analyses,
smoking
status
was
also
recorded
as
a
continuous
variable
(i.e.,
number
of
cigarettes
smoked
daily).
Odds
ratios
for
smoking
and
nutrient
supplements
on
the
risk
of
hypovitaminosis
C
while
controlling
for
possible
confounding
variables
were
determined
by
the
Mantel-Haenzel
procedure.2425
Results
Table
1
shows
univariate
analyses
for
variables
associ-
ated
with
both
serum
vitamin
C
levels
and
cigarette
smoking
prevalence.
Serum
vitamin
C
levels
were
significantly
asso-
ciated
with
gender,
race,
age,
and
body
mass
index.
Males,
when
compared
to
females,
Blacks
when
compared
to
Whites,
and
adults
under
the
age
of
60
years
when
compared
with
those
over
60
had
lower
serum
vitamin
C
levels
and a
higher
prevalence
of
smoking.
Heavier
individuals
(BMI
>
27),
on
the
other hand,
had
lower
serum
vitamin
C
concen-
trations
and
a
lower
prevalence
of
smoking.
Higher
ethanol
intakes
(>5
drinks
daily)
were
associated
with
a
higher
likelihood
of
smoking,
while
a
non-significant
increase
in
vitamin
C
levels
(p=0.09)
was
observed
in
those
with
more
moderate
consumption
(1
to
5
drinks
daily).
An
inverse
relation
between
cigarette
smoking
and
dietary
vitamin
C
intake
is
demonstrated
in
Table
2.
The
mean
vitamin
C
intake
for
all
smokers
was
20%
lower
compared
with
non-smokers
(86
mg,
95%
CI:
83,
90
vs
108
mg,
95%
CI:
106,
111
[not
shown]).
This
decrease
was
particularly
marked
in
individuals
smoking
20
cigarettes
daily.
Smokers
of
less
than
20
cigarettes
daily
had
an
intermediate
vitamin
C
intake
while
individuals
who
stopped
smoking
for
more
than
one
year
had
a
vitamin
C
intake
(108
mg,
95%
CI:
103,
113)
similar
to
never
smokers.
This
inverse
association
between
smoking
and
dietary
vitamin
C
intake
persisted
following
adjustment
for
age,
sex,
race,
ethanol
intake,
income
and
body
weight.
TABLE
1-Dietary
Vitamin
C
Intake,
Serum
Vitamin
C
Levels,
and
Smoking
Frequency
in
Individuals
by
Category
of
Gender,
Race,
Age
Tertiles,
Body
Mass
Tertiles,
and
Alcohol
Consump-
tion
Serum
Vitamin
C
mg/dl*
Smokers
N
(95%
Cl)
(per
cent)
Gender
Male
5520
0.94
(0.90,
0.97)
40.5
Female
6072
1.12
(1.09,
1.16)
33.4
Race
White
10430
1.05
(1.03,
1.08)
36.3
Black
1162
0.86
(0.80,
0.91)
41.2
Age
(years)
18-35
5178
1.14
(1.11,
1.18)
40.4
36-59
4614
1.01
(0.98,
1.04)
38.3
60-74
1800
1.02
(0.98,
1.06)
22.6
Body
mass
(BMI)
Thin
(<23)
4255
1.08
(1.05,
1.12)
43.1
Medium
(23-27)
4090
1.05
(1.02,
1.08)
34.9
Heavy
(>27)
3245
0.95
(0.91,
0.98)
30.9
Alcohol
(drinks/weekly)
None
3875
1.02
(0.98,
1.05)
25.5
1-5
4509
1.06
(1.02,
1.10)
39.5
>5
1947
1.01
(0.96,1.06)
50.1
*To
convert
to
rmol/L,
multiply
by
56.78
In
addition
to
decreased
vitamin
C
consumption,
smokers
had
serum
vitamin
C
levels
which
were
24
per
cent
lower
than
non-smokers
(0.87
mg/dl,
95%
CI:
0.86,
0.89
vs
1.13
mg/dl,
95%
CI:
1.12,
1.14;
49.4
,umol/L,
95%
CI:
48.8,
50.5
vs
64.2
,umol/L,
95%
CI:
63.6,
64.7
[not
shown]).
As
shown
in
Table
3,
smokers
of
20
cigarettes
daily
had
the
lowest
serum
vitamin
C
concen-
trations
(0.82
mg/dl,
46.6
,umol/L),
while
serum
vitamin
C
levels
were
highest
in
individuals
who
had
either
stopped
smoking
for
longer
than
one
year
or
who
had
never
smoked.
The
decrease
in
serum
vitamin
C
levels
observed
in
smokers
persisted,
even
after
adjusting
for
differences
in
dietary
intake,
age,
gender,
body
weight,
race,
and
vitamin
supplementation.
In
multiple
regression
analysis,
smoking
status
was
the
most
important
non-nutritive
predictor
of
serum
vitamin
C
levels,
explaining
more
of
the
total
variance
than
did
gender,
race,
BMI,
or
age
(Table
4).
In
additional
analyses,
smoking
status
was
entered
into
the
regression
equation
as
a
continuous
variable
defined
as
the
number
of
cigarettes
smoked
daily.
The
order
of
entry
and
regression
coefficients
in
these
analyses
were
similar
to
those
shown
in
Table
4
(i.e.,
when
smoking
was
entered
as
a
non-continuous
variable).
The
regression
curves
plotting
dietary
vitamin
C
intake
TABLE
2-Effect
of
Smoking
Status
on
Dietary
Vitamin
C
Intake
Dietary
Vitamin
Intake
(mg)
N
Mean
(95%
Cl)
Adjusted
Mean*
Smoker
>1
PPD
1288
83
(76,
89)
79
1
PPD
1396
79
(73,
84)
78
<1
PPD
1498
97
(90,
104)
97
Non-smoker
DC <
1
yr
335
97
(79,116)
92
DC
>
1
yr
1870
108
(103,113)
105
Never
smoked
4815
109
(105,
113)
111
NOTES:
>1
PPD-Greater
than
1
pack
per
day;
1
PPD-One
pack
per
day;
<1
PPD-
Less
than
one
pack
per
day;
DC
<
1
yr-Stopped
smoking
within
the
past
year;
DC
>
1
yr-
Stopped
for
the
previous
year
or
longer.
*Mean
vitamin
C
intake
adjusted
for
age,
sex,
race,
ethanol
intake,
monetary
income,
and
body
weight.
AJPH
February
1989,
Vol.
79,
No.
2
159
SCHECTMAN,
ET
AL.
TABLE
3-Effect
of
Smoking
Status
on
Serum
Vitamin
C
Levels
Serum
Vitamin
C
Concentration
(mg/dl)t
N
Mean
(95%
Cl)
Adjusted
Mean*
Smoker
>1
PPD
1288
0.83
(0.79,
0.88)
0.89
1
PPD
1454
0.82
(0.77,
0.86)
0.88
<1
PPD
1553
0.97
(0.92,
1.03)
0.99
Non-Smoker
DC
<
1
yr
346
0.99
(0.92,
1.06)
1.03
DC
>
1
yr
1942
1.10
(1.06,1.14)
1.11
Never
Smoked
5009
1.15
(1.15,
1.18)
1.12
NOTES:
>1
PPD-Greater
than
1
pack
per
day;
1
PPD-Smoker
of
one
pack
per
day;
<1
PPD-Less
than
one
pack
per
day;
DC
<
1
yr-Stopped
smoking
within
the
past
year;
DC
>
1
yr-Stopped
for
the previous
year
or
longer.
Vitamin
C
serum
level
adjusted
for
age,
sex,
race,
ethanol
intake,
body
weight,
dietary
vitamin
C
intake,
and
supplementation
status.
tTo
covert
to
Fmol/L,
multiply
by
56.78
with
serum
vitamin
C
levels
for
smokers
and
non-smokers
are
shown
in
Figure
1.
The
slope
of
the
two
lines
is
similar,
with
a
consistent
difference
in
serum
vitamin
C
levels
between
smokers
and
non-smokers
of
approximately
0.2
mg/dl
(11.4
p,mol/L),
independent
of
the
level
of
dietary
vitamin
C
intake.
Using
the
regression
equations,
an
additional
133
mg
of
dietary
vitamin
C
was
calculated
to
be
necessary
to
increase
the
serum
vitamin
C
level
in
smokers
to
the
level
found
in
non-smokers.
To
test
the
hypothesis
that
smoking
increases
the
risk
of
hypovitaminosis
C,
the
frequency
of
low
serum
vitamin
C
levels
occurring
in
smokers
was
compared
with
non-smokers.
Of
all
smokers,
7.4
per
cent
had
serum
vitamin
C
levels
of
0.2
mg/dl
(11.4
,umol/L)
or
less
compared
with
1.9
per
cent
of
non-smokers
who
had
similar
vitamin
C
levels.
An
additional
19.7
per
cent
of
smokers
had
serum
vitamin
C
levels
between
0.2
mg/dl
(1
1.4
t±molIL)
and
0.4
mg/dl
(22.7
,umol/L),
compared
with
8.2
per
cent
of
non-smokers.
Since
further
adjusting
for
BMI
and
age
did
not
affect
mean
serum
vitamin
C
levels,
only
gender,
race,
and
dietary
vitamin
C
intake
were
used
as
controlling
variables
to
determine
the
risk
of
smoking
on
marginal
or
severe
hypovitaminosis
C.
After
adjusting
for
gender
and
race,
risk
ratios
for
smoking
of
3.5
(95%
CI:
2.9,
4.2)
for
severe,
and
3.1
(95%
CI:
2.8,
3.5)
for
marginal
hypovitaminosis
C
were
obtained
(Table
5).
Further
adjust-
ment
for
dietary
vitamin
C
intake
resulted
in
a
minor
reduction
in
the
odds
ratio.
Of
the
total
sample,
37
per
cent
reported
taking
vitamin
TABLE
4-Stepwise
Regression
of
Variables
Influencing
Serum
Vitamin
C
Levels
Independent
Order
of
Regression
95%
Confidence
Variable
Entry
Coefficients
Intervals
Supplements
1
-0.3254
-.3005,-.3503
Smoking
status
2
0.0515
.0467,
.0564
Vit.
C
intake
3
0.0013
.0012,
.0015
Sex
4
0.1379
.1200,
.1558
BMI
5
-0.0098
.0078,
.0118
Race
6
-0.1601
.1181,
.2022
Age
7
0.0018
.0010,
.0026
ETOH
0.0016
-.0009,
.0042
Total
R2:
0.29.
Vitamin
and
mineral
supplement
use,
coded
as
1
=
Yes,
2
=
No.
Smoking
status
coded
as
non-continuous
variable-see
text.
Sex
coded
as
1
=
male,
2
=
female.
Race
coded
as
1
=
White
(and
"other"),
2
=
Black.
ETOH
is
estimate
of
alcoholic
drinks
consumed
weekly.
1.4
1.2-
o
1.0
z
--
0.8
>'4
>
a-
m
E
0.6
c"
0.4
0.2
0
NW,
....!.."''
........
1.2~~~~~~~~~..
...............
o
25
50
75
100
125
150
175
200
225
VITAMIN
C
INTAKE
(mg)
FIGURE
1-Linear
Expression
Curves
for
Smokers and
Non-smokers
Showing
Relation
between
Dietary
Vitamin
C
Intake
and
Serum
Levels.
(Adjusted
for
age,
gender,
race,
body
mass
and
vitaniin
supplementation.
To
convert
serum
levels
to
Imol/L,
multiply
by
56.78.)
or
mineral
supplements.
The
frequency
of
marginal
hypovi-
taminosis
C
according
to
smoking
and
supplementation
status
is
shown
in
Table
6.
Smokers
who
did
not
take
vitamin
supplements
had
an
11.3
fold
increased
risk
(95%
CI:
9.6,
13.3)
of
marginal
hypovitaminosis
C
compared
with
non-
smokers
who
took
supplements.
Even
smokers
who
took
vitamin
supplements
were
at
increased
risk
of
marginal
hypovitaminosis
C
(odds
ratio
2.5;
95%
CI:
2.0,
3.1).
Discussion
Our
analysis
of
the
NHANES
II
data
confirms
that
cigarette
smoking
is
associated
with
decreased
serum
vitamin
C
levels.
This
association
persisted
despite
correction
for
factors
which
independently
affected
serum
vitamin
C
levels
such
as
age,
gender,
race,
and
BMI.
Dietary
vitamin
C
intakes
were
lower
in
smokers
than
non-smokers,
a
finding
which
might
potentially
explain
this
inverse
correlation.
However,
further
adjustment
for
vitamin
C
consumption
revealed
that
the
association
between
smoking
and
serum
levels
was
independent
of
dietary
intake
(Table
4).
In
fact,
smoking
status
explained
more
of
the
total
variance
of
serum
vitamin
C
levels
than
did the
self-reported
dietary
intake
of
vitamin
C.
Moreover,
the
increased
risk
of
hypovitaminosis
C
in
smokers
was
not
substantially
lowered
when
correcting
for
the
decreased
dietary
vitamin
C
intake
associated
with
smoking
(Table
5).
These
observations
suggest
that
cigarette
smoking
has
an
important
influence
on
serum
vitamin
C
levels
which
occurs
predominantly
via
a
mechanism
inde-
pendent
of
decreased
dietary
vitamin
C
consumption.
TABLE
5-Risk
of
Hypovitaminosis
C
In
Smokers
versus
Non-smokers
Unadjusted
for
Adjusted
for
Vitamin
C
Intake
Vitamin
C
Intake
Vitamin
C
Smoking
Smoking
Deficiency
Odds
Ratio
(95%
Cl)
Odds
Ratio
(95%
Cl)
Severe
3.5
(2.9,
4.2)
3.0
(2.5,
3.6)
Marginal
3.1
(2.8,
3.5)
2.8
(2.6,
3.1)
Unsupplemented
subjects
only.
AJPH
February
1989,
Vol.
79,
No.
2
160
SMOKING
AND
VITAMIN
C
STATUS
TABLE
6-Combined
Effect
of
Smoking
and
Vitamin
Supplementation
on
Risk
of
Marginal
Hypovitaminosis
C
Risk
of
Hypovitaminosis
C
Hypovitaminosis
C
Prevalence
(%)
Odds
Ratio
(95%
Cl)*
Supplemented
non-smokers
3.5
1.0
Supplemented
smokers
9.8
2.5
(2.0,
3.1)
Unsupplemented
non-smokers
14.5
3.9
(3.3,
4.7)
Unsupplemented
smokers
35.7
11.3
(9.6,
13.3)
Inverse
associations
between
smoking
and
vitamin
C
status
have
been
previously
reported.
Most
studies,
however,
have
employed
small
sample
sizes3'5
or
have
been
unable
to
control
potentially
confounding
variables.4'6'7
Pelletier
strati-
fied
4,500
Canadians
participating
in
the
Nutrition
Canada
National
Survey
into
three
categories
of
dietary
vitamin
C
intake
(0-30
mg,
30-100
mg, and
over
100
mg)
and
analyzed
serum
vitamin
C
levels
according
to
gender
and
age;
an
average
30
per
cent
reduction
in
serum
levels
were
observed
in
smokers
compared
with
non-smokers.
15
Ritzel,
et
al,
determined
that
mean
serum
vitamin
C
levels
were
lower
in
smokers,
independent
of
weekly
citrus
fruit
consumption,
in
4,053
healthy
Swiss
employees.'6
The
risk
of
severe
or
marginal
hypovitaminosis
C
was
increased
approximately
threefold
in
smokers.
Univariate
analysis
of
NHANES
II
demonstrated
a
difference
in
serum
vitamin
C
levels
of
0.2
mg/dl
(11.4
,umol/L)
between
smokers
and
non-smokers
at
each
of
six
levels
of
dietary
vitamin
C
intake.
17
These
results
are
similar
to
the
findings
in
this
study
using
stepwise
multiple
regression
to
analyze
the
NHANES
II
data.
The
lowered
serum
vitamin
C
levels
in
smokers
could
be
due
to
either
impaired
vitamin
C
absorption
or
increased
turnover.
On
the
basis
of
studies
measuring
urinary
vitamin
C
excretion
in
conjunction
with
the
administration
of
known
vitamin
C
intakes,
Pelletier
suggested
the
presence
of
impaired
bioavailability
but
normal
turnover
of
vitamin
C
in
smokers.26'27
On
the
other
hand,
Kallner,
et
al,
measured
vitamin
C
kinetics
using
radio-labeled
ascorbic
acid
and
demonstrated
increased
turnover
in
smokers
but
only
small
differences
in
absorption
when
compared
to
non-smokers.28
Others
have
reported
that
smoking
acutely
increases
urinary
excretion
of
vitamin
C,
29
also
suggesting
an
accelerated
metabolism
in
smokers.
Inter-
actions
between
vitamin
C
metabolism
and
a
number
of
drugs
have
been
described.303'
Identification
of
the
responsible
compound
in
tobacco
smoke
and
the
elaboration
of
its
mech-
anism
will
require
further
study.
In
this
study,
serum
vitamin
C
levels
were
approximately
0.2
mg/dl
(11.4
,umol/L)
lower
in
smokers
compared
with
non-smokers
regardless
of
dietary
vitamin
C
intake
(Figure
1).
Therefore,
the
impact
of
smoking
to
decrease
serum
vitamin
C
levels
was
proportionately
greater
in
participants
with
initially
low
serum
vitamin
C
levels.
This
resulted
in
a
higher
proportion
of
smokers
with
serum
levels
below
the
0.2
mg/dl
(11.4
,umol/L)
or
0.4
mg/dl
(22.7
,umol/L)
range
cutoffs
for
severe
or
marginal
hypovitaminosis
C.
In
this
regard,
the
risk
ratio
for
hypovitaminosis
C
was
threefold
greater
in
smokers
when
compared
with
non-smokers.
Therefore,
although
the
prevalence
of
severe
and
marginal
hypovitaminosis
C
in
NHANES
II
was
low,
certain
subgroups,
such
as
smokers
not
taking
nutritional
supplements,
may
be
at
substantial
risk.
The
significance
of
the
increased
risk
of
hypovitaminosis
C
produced
by
the
interaction
between
smoking
and
serum
vitamin
C
levels
is
not
currently
known.
Although
guidelines
developed
by
the
National
Survey
of
Canada
in
1973
sug-
gested
categories
of
high
risk
for
serum
vitamin
C
levels
below
0.2
mg/dl
(1
1.4
,umol/L),
and
of
moderate
risk
for
levels
between
0.2
and
0.4
mg/dl
[11.4
and
22.7
pLmol/L],2'