30015.

qxd 2/2/04 4:04 PM Page 16

Volume 75 • Number 1

The Effect of Cigarette Smoking

on Gingival Bleeding
Thomas Dietrich,*†‡ Jean-Pierre Bernimoulin,* and Robert J. Glynn§

Background: The purpose of this study was to investigate

the dose-dependent effect of cigarette smoking upon gingival
bleeding on probing (BOP) in a large representative sample of
the United States population (National Health and Nutrition
Examination Survey III).
Methods: Weighted multiple logistic regression was used to
model bleeding on probing of 141,967 mesio-buccal sites in

G
ingival bleeding after probing is
12,385 individuals with complete case records on all covari- widely used as a clinical marker of
ates. Adjustments were made for age, gender, race/ethnicity, gingival inflammation, both on a
number of missing teeth, tooth type/jaw, root caries, full crown subject and site level.1 Numerous studies
coverage, socioeconomic status (poverty/income ratio), and have shown that current cigarette smok-
survey characteristics. The model stratified by presence of cal- ing suppresses the gingival inflammatory
culus (CALC) and increased probing depth (PD ≥4 mm). Gen- response to a given amount of plaque as
eralized estimating equations were used to account for measured by bleeding on probing. This
dependence of sites within subjects. phenomenon has been demonstrated both
Results: Smoking had a strong suppressive effect on gingi- in observational and experimental gin-
val bleeding. The effect was strongest in heavier smokers (>10 givitis studies.2-9
cigarettes/day) and smallest in former smokers. In healthy sites Almost all of these studies used subject
(no CALC, PD ≤3 mm), the odds ratio (OR) of bleeding for sites level summary measures of gingival bleed-
in heavier smokers compared to never-smokers was 0.56 (95% ing as the outcome variable. Therefore,
CI: 0.45-0.70). Sites with CALC and/or PD ≥4 mm were more little is known about the magnitude of the
likely to bleed in never-smokers (OR: 5.7; 95% CI: 4.3-7.6). effect that smoking has on gingival bleed-
This relationship was less evident among heavier smokers (OR: ing at the site level. Furthermore, findings
1.3; 95% CI: 0.8-1.9). The effect of smoking did not differ from studies on mean bleeding scores
between maxillary and mandibular molars, premolars, or incisors. may not necessarily hold on the site level,
Conclusion: Smoking exerts a strong, chronic, and dose- since many local factors may influence
dependent suppressive effect on gingival bleeding on probing. the probability of a site to bleed on prob-
J Periodontol 2004;75:16-22. ing, or modify the effect of smoking.9
KEY WORDS The mechanisms by which smoking
suppresses gingival bleeding are not
Bleeding, gingival; gingival diseases/etiology; National Health
understood. It is well established that nico-
and Nutrition Examination Survey III; periodontal index;
tine causes an acute vasoconstriction in
smoking/adverse effects; smoking/physiology.
human skin.10 However, gingival blood
flow was found to be increased during
* Department of Periodontology, Charité, Humboldt University, Berlin, Germany. smoking.11,12 A lowered bleeding response
† Department of Oral Surgery and Radiology, Charité, Humboldt University.
‡ Department of Epidemiology, German Institute of Human Nutrition, Potsdam-Rehbrücke, in former smokers would indicate a more
Germany. “chronic” effect of smoking. However,
§ Division of Preventive Medicine, Department of Medicine, Brigham and Women’s
Hospital and Harvard Medical School, Boston, MA. few studies have examined former smok-
ers in this respect, and the findings were
inconsistent.5,13 The distribution of attach-
ment loss and periodontal pockets in
smokers versus non-smokers has sug-
gested a local effect of cigarette smoking
on the periodontium.14-16 Whether there
is a similar pattern for the effect of smok-
ing on gingival bleeding has not been
investigated.

16
30015.qxd 2/2/04 4:04 PM Page 17
J Periodontol • January 2004
The aim of the present analysis was to investigate
the association between smoking and bleeding on
probing in a large representative sample of the United
States population (National Health and Nutrition Exam-
ination Survey III) on a site level.
MATERIALS AND METHODS
Data Source
Data were derived from the Third National Health and
Nutrition Examination Survey (NHANES III) as avail-
able from the National Center for Health Statistics.
NHANES III was conducted between 1988 and 1994 to
assess the health and nutrition of a large representative
sample of the civilian non-institutionalized U.S. popu-
lation. The survey was designed as a complex, multi-
stage, stratified, clustered sample survey. Interviews
were conducted at respondents’ homes, where detailed
information on smoking habits was collected. Stan-
dardized examinations were conducted in mobile exam-
ination centers, including extensive dental examinations
with periodontal assessments, where appropriate.
A detailed description of the survey and the oral health
component can be found elsewhere.17,18 Furthermore,
detailed analyses of destructive periodontal disease, gin-
gival recession, gingival bleeding, dental calculus, and
smoking-attributable destructive periodontitis in NHANES
III have been published previously.19-21
Cigarette Smoking Data
Several questions regarding cigarette smoking were
administered in the household interview. Respondents
were classified as: never smokers (if they had smoked
less than 100 cigarettes in their lifetime), former smok-
ers (≥100 lifetime cigarettes, not currently smoking),
or current smokers (≥100 lifetime cigarettes, currently
smoking). Current smokers were further stratified by
the number of cigarettes smoked per day (up to 10,
11-20, 21-30, >30 cigarettes/day). Self-reported smok-
ing measures of NHANES III have been validated pre-
viously with serum cotinine measurements.22
Dental Assessments
Periodontal assessments were made in two randomly
selected upper and lower quadrants on fully erupted
teeth (excluding third molars) by trained examiners.
For assessment of gingival bleeding, teeth were dried
with air from the buccal aspect and an NIDR probe
was inserted no more than 2 mm into the gingival sul-
cus, starting just distal to the midpoint of the buccal
surface and then moved gently into the mesial inter-
proximal area. After all sites from the facial or buccal
aspect of a single quadrant had been examined in this
fashion, bleeding was scored as a dichotomous vari-
able at each site (bleeding present/absent). For assess-
ment of calculus, teeth were dried with air from the
buccal aspect, and the sites were observed to deter-
Dietrich, Bernimoulin, Glynn
mine the presence or absence of supragingival calcu-
lus (defined as calculus extending up to 1 mm below
the free gingival margin); the NIDR probe was then
used to determine the absence or presence of sub-
gingival calculus. For each site, a score was assigned
as follows: 0, no calculus present; 1, supragingival cal-
culus only; 2, subgingival calculus or both sub- and
supragingival calculus. Periodontal pockets were mea-
sured using the NIDR probe with a probing force of no
more than 25 ponds (∼0.25 N), and measurements
were rounded to the lowest whole millimeter. Root
caries was assessed using a #23 explorer after removal
of plaque and soft accumulations.
Data on Other Covariates
The poverty income ratio (PIR) was computed as the
ratio of family income versus the poverty threshold as
determined annually by the U.S. Census Bureau. These
threshold values adjust for changes caused by inflation
between calendar years. Hence, the higher the family
income relative to the poverty threshold, the higher
the PIR value.
Statistical Analysis
We conducted a complete case analysis on the site and
subject level, limited to mesio-buccal sites in individuals
aged 18 years and older. The analyses accounted for
the NHANES III sampling weights.23 A multiple logistic
regression model was fit to examine the association
between gingival bleeding as the (dichotomous) depen-
dent variable and cigarette smoking status (never, for-
mer, current: ≤10, 11-20, 21-30, >30 cigarettes/day)
as the independent variable. Covariates included age
(decade), gender, calculus, probing depth (0-3 mm,
≥4 mm), poverty/income ratio (continuous), race/
ethnicity (non-Hispanic whites, non-Hispanic blacks,
Mexican Americans, other), tooth/jaw (upper central
incisor, lower central incisor, . . . , upper second molar,
lower second molar), number of missing teeth, full crown
coverage (yes/no), and root caries (yes/no). Further-
more, adjustments for dental examiner, survey phase,
survey stratum, and survey primary sampling unit (PSU)
were made.24 To account for the dependence of obser-
vations (sites) within subjects, generalized estimating
equation (GEE) marginal models with an unstructured
working correlation were used. All analyses were per-
formed using a statistical software package.

RESULTS
There were 13,693 individuals aged 18 years and older
who had at least partial assessment of gingival bleed-
ing, 13,663 of whom had at least partial assessment of
calculus and probing depths. The final sample consisted
of 141,967 mesio-buccal sites with complete records of

STATA version 7.0, Stata Corp., College Station, TX.
17
30015.qxd 2/2/04 4:04 PM Page 18

Smoking and Gingival Bleeding Volume 75 • Number 1

Table 1.
Distribution of Final Sample (12,385 subjects) by Smoking Category*

Smoking Status

Never Former Light Heavier

N % N % N % N %

Gender
Male 2369 40.7% 1719 59.5% 873 51.1% 913 56.1%
Female 4110 59.3% 1004 40.5% 728 48.9% 669 43.9%

Age
18-30 2478 36.1% 334 13.7% 634 43.9% 444 35.4%
31-40 1405 26.0% 522 22.4% 429 30.6% 447 29.4%
41-50 845 15.2% 487 23.1% 247 13.9% 328 19.7%
51-60 548 8.7% 436 17.4% 126 6.0% 184 9.3%
61-70 599 7.7% 481 13.4% 118 3.8% 134 4.9%
71-80 386 4.4% 349 8.2% 42 1.7% 36 1.1%
81+ 218 1.8% 114 1.6% 5 0.1% 9 0.2%

Race/Ethnicity
Non-Hispanic white 2167 70.4% 1347 83.2% 294 59.6% 840 86.2%
Non-Hispanic black 1895 12.1% 541 6.8% 674 22.5% 485 8.1%
Mexican American 2082 6.7% 749 4.5% 583 9.1% 215 1.9%
Other 335 10.8% 86 5.6% 50 8.7% 42 3.8%

Mean 95% CI Mean 95% CI Mean 95% CI Mean 95% CI

Mean probing depth 1.74 1.67-1.80 1.82 1.75-1.90 1.94 1.87-2.02 2.07 1.96-2.18

% sites with calculus

Supragingival 19.9% 15.7-24.0% 21.4% 16.6-26.2% 18.0% 14.4-21.6% 19.0% 14.2-23.8%
Supra- and/or subgingival 17.7% 15.0-20.5% 22.7% 19.5-25.9% 30.6% 26.4-34.8% 39.9% 35.6-44.1%

% sites bleeding 8.5% 7.3-9.6% 7.7% 6.7-8.7% 6.3% 5.3-7.3% 5.6% 4.7-6.5%

Number of missing teeth 3.1 2.9-3.2 5.2 4.7-5.6 3.2 2.7-3.7 4.6 4.3-5.0
% sites with root caries 0.2% 0.2-0.3% 0.5% 0.3-0.7% 0.6% 0.3-0.9% 0.9% 0.6-1.1%

% teeth with full crowns 4.8% 4.1-5.4% 8.8% 7.4-10.3% 2.6% 1.9-3.4% 4.0% 3.0-4.9%
* Unweighted N and weighted percentages and means with 95% CI. Light smoker: ≤10 cigarettes/day; heavier smoker: >10 cigarettes/day.

all site-specific covariates (calculus, probing depths, Table 2.

root caries, full crown coverage) in 12,385 subjects who
Weighted Odds Ratios (OR) and 95%
had complete records on all subject-specific covariates.
Table 1 gives an overview of the distribution of the Confidence Intervals (CI) from the Main
sample by smoking categories. Compared to never Effects Model*
smokers, heavier smokers were more likely to be male,
had higher mean probing depths, more root caries, Smoking Status OR 95% CI
less remaining teeth, and more subgingival calculus. Never 1.00 Reference
In a preliminary main-effects multiple logistic regres-
Former 0.78 0.69-0.89
sion model, all site- and subject-specific covariates were
significant predictors of gingival bleeding. Smoking was Current (cigarettes/day)
negatively associated with gingival bleeding; i.e., sites in ≤10 0.53 0.45-0.62
smokers were less likely to bleed upon gentle probing 11-20 0.40 0.34-0.48
21-30 0.33 0.25-0.44
than sites in non-smokers. However, there was a thresh-
>30 0.44 0.32-0.60
old effect insofar as gingival bleeding did not vary among
* Adjusted for age, gender, race/ethnicity, calculus, probing depths, poverty/
heavier smokers (>10 cigarettes/day, Table 2). There- income ratio, tooth/jaw, number of missing teeth, full crown coverage, root
fore, the upper three smoking categories were collapsed caries, dental examiner, survey phase, stratum, and PSU.

18
30015.qxd 2/2/04 4:04 PM Page 19

J Periodontol • January 2004 Dietrich, Bernimoulin, Glynn

in a single category (current heavier smok- Table 3.

ers). We included interaction terms for age,
Weighted Odds Ratios (OR) and 95% Confidence
tooth group/jaw (upper and lower front, pre-
molar, molar region), calculus, and probing Intervals (CI) by Probing Depth Calculus Category*
depth categories. There was no interaction with
age, and no difference in the effect of smok- Probing Depth 0-3 mm Probing Depth ≥4 mm
ing on gingival bleeding between different tooth Smoking Status OR 95% CI OR 95% CI
groups/jaws. However, we found significant
interactions between cigarette smoking status No calculus
and calculus and probing depth categories; Never 1.00 – 1.84 1.43-2.37
i.e., the effect of cigarette smoking on gingi- Former 0.81 0.69-0.95 1.51 1.08-2.13
Light 0.60 0.48-0.75 0.92 0.57-1.49
val bleeding varied across calculus and prob-
Heavier 0.56 0.45-0.70 0.60 0.40-0.90
ing depth levels. Hence, the final model
included these interaction terms, and odds ratio Supragingival calculus
estimates for the effect of smoking on gingi- Never 1.60 1.39-1.83 2.94 2.19-3.95
val bleeding are presented by calculus and Former 1.06 0.83-1.36 1.98 1.32-2.98
probing depth category (Table 3, Figs. 1 Light 0.76 0.52-1.12 1.16 0.66-2.06
and 2). Unweighted analyses and weighted Heavier 0.56 0.41-0.77 0.60 0.36-0.99
analyses that accounted for NHANES III sam- Subgingival calculus or supra- and subgingival calculus
pling weights yielded similar results. Never 3.10 2.75-3.49 5.72 4.29-7.63
Sites in smokers were less likely to bleed Former 2.49 2.04-3.04 4.65 3.31-6.54
on gentle probing than sites in never smok- Light 1.60 1.30-1.96 2.44 1.51-3.94
ers. This effect was strongest in heavier Heavier 1.20 0.98-1.46 1.28 0.84-1.94
smokers (>10 cigarettes per day) and weak- * Reference: sites with PD 0-3 mm and no calculus in never smokers, adjusted for age, gender,
est in former smokers and consistently evi- race/ethnicity, probing depths, poverty/income ratio, tooth/jaw, number of missing teeth, full
crown coverage, root caries, dental examiner, survey phase, stratum, and PSU.
dent across all calculus and probing depth
categories. For example, in shallow sites
without any calculus, heavy smoking was asso-
ciated with a 44% (95% CI: 30% to 55%) reduc-
tion in gingival bleeding. The smoking effect was
modified by periodontal pockets and calculus.
While calculus and increased probing depths were
associated with an increased risk of gingival bleed-
ing, the effect of smoking tended to be more pro-
nounced if either or both of these conditions were
present. As a result, the differences in gingival
bleeding across probing depth and calculus cat-
egories tended to level off in heavier smokers. In
a subsample of 11,911 individuals with informa-
tion on pack years of smoking, pack years were
associated with less gingival bleeding in addition
to the effects of current/former smoking. How-
ever, this effect was small and did not reach sta-
tistical significance (OR for increment of 10 pack
years: 0.97; 95% CI: 0.94-1.00; P = 0.09). Effect Figure 1.
Weighted OR estimates showing the effect of smoking status on gingival
estimates of other clinical covariates that were bleeding by calculus for sites with probing depths (PD) of 0 to 3 mm (reference:
predictors of gingival bleeding in the final model sites with PD 0 to 3 mm without calculus in never-smokers).
are given in Table 4. Sites in males, Mexican
Americans, and people with lower incomes, and
sites with full crown coverage and root caries were of several previous cross-sectional and experimental
more likely to show gingival bleeding. studies.2-9
The present analysis gives new insight into the mag-
DISCUSSION nitude of the suppressive effect on a site level under var-
In the present analysis, smoking showed a strong and ious clinical conditions. In healthy sites, heavy smoking
consistent suppressive effect on gingival bleeding upon reduces the likelihood of bleeding on probing by almost
gentle probing. These findings corroborate the results 50%. This suppressive effect appears to be relatively

19
30015.qxd 2/2/04 4:04 PM Page 20

Smoking and Gingival Bleeding Volume 75 • Number 1

among never-smokers. For example, in never-

smokers, sites with subgingival calculus and
increased probing depths were almost six times as
likely to bleed compared to the reference group
(Table 3). However, due to the strong suppressive
effect of smoking, these differences in the bleeding
response are not evident among heavier smokers.
For example, sites with increased probing depth
and supragingival calculus in heavier smokers
showed less bleeding than healthy sites in never-
smokers.
The association between smoking and gingi-
val bleeding appears to be dose dependent; how-
ever, the smoking effect reaches a plateau at
approximately 10 to 20 cigarettes per day. This
is in agreement with previous cross-sectional
studies that investigated the dose-response func-
Figure 2. tion among current smokers.5
Weighted OR estimates showing the effect of smoking status on gingival The fact that a suppressive effect was evident
bleeding by calculus for sites with probing depths (PD) of ≥4 mm (reference:
among former smokers provides further evidence
sites with PD 0 to 3 mm without calculus in never-smokers).
against an “acute” effect of smoking.25 The only
previous study that examined cumulative measures
Table 4. of smoking did not adjust for current smoking and
reported inconsistent findings.5 Although the pre-
Weighted Odds Ratio Estimates (OR)
sent data suggest that lifetime exposure to smok-
and 95% Confidence Intervals (CI) of ing has some suppressive effect, it is likely to be
Clinical Covariates That Were Significant small compared to the effect of current smoking.
Predictors of Gingival Bleeding in the Final In a study on periodontal attachment loss in smok-
Model (independent of age, smoking ers versus non-smokers, Haffajee and Socransky
status, probing depth, and calculus) observed a differential pattern of attachment loss. The
difference between smokers and non-smokers was
Variable OR 95% CI
greatest on mandibular front teeth and maxillary palatal
sites, suggesting that local effects of smoking play a
Gender (Female versus Male) 0.85 0.78-0.94 major role in the pathogenesis of smoking-associated
Race/ethnicity
periodontal disease.14 Similar findings for the distrib-
Non-Hispanic white 1.00 Reference ution of periodontal pockets have been reported.15 We
Non-Hispanic black 0.99 0.88-1.11 did not observe a similar pattern in the present study;
Mexican American 1.49 1.30-1.72 i.e., the effect of smoking did not differ between front
Other 1.13 0.92-1.39 and molar teeth in either jaw. However, this study might
not have captured such a pattern since periodontal
Poverty income ratio* 0.88 0.86-0.91
assessments were not made on lingual/palatal sites.
Number of missing teeth 1.05 1.04-1.06 The strengths of the present analysis are its popu-
lation-based design: the weighted estimates can be
Full crown coverage 1.95 1.66-2.29
generalized to the non-institutionalized civilian U.S.
Root caries 1.53 1.00-2.36 population at the time of NHANES III. Furthermore,
* Poverty income ratio as continuous variable (range 0 to 11.9).
the large sample provided sufficient power to perform
Model adjusted for dental examiner, survey phase, PSU, and stratum. valid analyses within subgroups. The results are con-
sistent across all subgroups, thus lending credibility to
the overall findings.
stronger in sites with calculus and/or increased probing However, the study has several important limitations.
depths (≥4 mm). In this cross-sectional sample of a pop- Plaque levels were not recorded during the dental exam
ulation not under comprehensive periodontal care, cal- in NHANES III. Consequently, adjustments for plaque
culus and increased probing depths can reasonably be could not be made in the present analysis. However, we
viewed as indicators of gingival inflammation. Accord- do not believe that the association between smoking
ingly, bleeding on probing occurs much more frequently and gingival bleeding reported here is due to con-
at sites with calculus and/or increased probing depth founding by plaque. Confounding by plaque could bias

20
30015.qxd 2/2/04 4:04 PM Page 21
J Periodontol • January 2004
the association between smoking and gingival bleed-
ing only if plaque levels differed between smokers and
non-smokers. However, most studies reported similar
plaque levels for smokers and non-smokers.2,4-6,14,26-29
In experimental gingivitis studies, no difference between
smokers and non-smokers with regard to plaque accu-
mulation could be observed.7,8,30 Furthermore, com-
pliance with oral hygiene instructions was shown to be
similar between smokers and non-smokers.4 However,
some authors found higher plaque levels in smokers.9,31
If smokers had more plaque than non-smokers in the
population studied here, our odds ratio estimates would
underestimate the apparent inhibitory effect of smok-
ing on gingival bleeding; i.e., adjustment for plaque
levels would yield an even higher suppressive effect of
smoking. However, for the other predictors of gingival
bleeding shown in Table 4, with the exception of the
number of missing teeth, it is reasonable to assume
that plaque is a mediator of the association.
Bleeding on probing was assessed as a dichoto-
mous measure in NHANES III. Dichotomous bleeding
scores are frequently used both in practice and
research.1 To assess gingival bleeding more precisely,
numerous semiquantitative bleeding scores have been
proposed.32 Such scores are more sensitive to assess
small changes in bleeding tendency; however, quan-
tification of bleeding by the examiner is susceptible to
subjective interpretation, which may cause bias. In the
present analysis, the effect of smoking on gingival
bleeding was strong and consistent; the interpretation
of the results is therefore not limited by the use of a
dichotomous measure.
In the present study, bleeding on probing was pro-
voked by inserting the probe no more than 2 mm into
the gingival sulcus; i.e., only the marginal gingiva was
probed. However, when used as a clinical diagnostic
parameter at the site level, bleeding is frequently
assessed after probing to the bottom of the
pocket.1,33,34 Lie et al. have shown that the latter pro-
cedure consistently yields higher average bleeding
scores.8 The results of the present analysis may there-
fore not be generalizable to bleeding on probing to the
bottom of the pocket. However, the study also showed
that smoking yielded a similar suppressive effect on
gingival bleeding as provoked by either technique.
Given these findings and the fact that smoking is a
strong risk factor for periodontal disease and adversely
affects periodontal treatment outcome, the predictive
value of bleeding on probing at the site level is likely
to differ between smokers and non-smokers.
In conclusion, the present analysis shows that smok-
ing exerts a strong, chronic, and dose-dependent sup-
pressive effect on gingival bleeding on probing. Further
research is necessary to determine the significance of
bleeding on probing as a prognostic marker in smok-
ers and non-smokers with periodontitis.
Dietrich, Bernimoulin, Glynn
REFERENCES
1. Lang NP, Tonetti MS. Periodontal diagnosis in treated
periodontitis. Why, when and how to use clinical para-
meters. J Clin Periodontol 1996;23:240-250.
2. Bergstrom J, Floderus-Myrhed B. Co-twin control study
of the relationship between smoking and some peri-
odontal disease factors. Community Dent Oral Epidemiol
1983;11:113-116.
3. Bergstrom J, Preber H. The influence of cigarette smok-
ing on the development of experimental gingivitis. J Peri-
odont Res 1986;21:668-676.
4. Bergstrom J. Oral hygiene compliance and gingivitis
expression in cigarette smokers. Scand J Dent Res 1990;
98:497-503.
5. Bergstrom J, Bostrom L. Tobacco smoking and peri-
odontal hemorrhagic responsiveness. J Clin Periodontol
2001;28:680-685.
6. Calsina G, Ramon JM, Echeverria JJ. Effects of smoking
on periodontal tissues. J Clin Periodontol 2002;29:771-776.
7. Danielsen B, Manji F, Nagelkerke N, Fejerskov O, Baelum
V. Effect of cigarette smoking on the transition dynam-
ics in experimental gingivitis. J Clin Periodontol 1990;17:
159-164.
8. Lie MA, Timmerman MF, van der Velden U, van der
Weijden GA. Evaluation of 2 methods to assess gingival
bleeding in smokers and non-smokers in natural and exper-
imental gingivitis. J Clin Periodontol 1998;25:695-700.
9. Muller HP, Stadermann S, Heinecke A. Longitudinal asso-
ciation between plaque and gingival bleeding in smokers
and non-smokers. J Clin Periodontol 2002;29:287-294.
10. Black CE, Huang N, Neligan PC, et al. Effect of nicotine
on vasoconstrictor and vasodilator responses in human
skin vasculature. Am J Physiol Regul Integr Comp Phys-
iol 2001;281:R1097-R1104.
11. Baab DA, Oberg PA. The effect of cigarette smoking on
gingival blood flow in humans. J Clin Periodontol 1987;
14:418-424.
12. Mavropoulos A, Aars H, Brodin P. Hyperaemic response
to cigarette smoking in healthy gingiva. J Clin Peri-
odontol 2003;30:214-221.
13. Bergstrom J, Eliasson S, Dock J. Exposure to tobacco
smoking and periodontal health. J Clin Periodontol
2000;27:61-68.
14. Haffajee AD, Socransky SS. Relationship of cigarette
smoking to attachment level profiles. J Clin Periodontol
2001;28:283-295.
15. van der Weijden GA, de Slegte C, Timmerman MF, van
der Velden U. Periodontitis in smokers and non-smokers:
Intra-oral distribution of pockets. J Clin Periodontol 2001;
28:955-960.
16. Preber H, Bergstrom J. Effect of non-surgical treatment
on gingival bleeding in smokers and non-smokers. Acta
Odontol Scand 1986;44:85-89.
17. Drury TF, Winn DM, Snowden CB, Kingman A, Kleinman
DV, Lewis B. An overview of the oral health component
of the 1988-1991 National Health and Nutrition Exam-
ination Survey (NHANES III-Phase 1). J Dent Res 1996;
75(Spec. Issue):620-630.
18. U.S. Department of Health and Human Services and
National Center for Health Statistics. NHANES III Refer-
ence Manuals and Reports [CD-ROM]. Hyattsville, MD:
Centers for Disease Control and Prevention; 1996.
19. Tomar SL, Asma S. Smoking-attributable periodontitis
in the United States: Findings from NHANES III. National
Health and Nutrition Examination Survey. J Periodontol
2000;71:743-751.
20. Albandar JM, Kingman A. Gingival recession, gingival
21
30015.qxd 2/2/04 4:04 PM Page 22

Smoking and Gingival Bleeding Volume 75 • Number 1

bleeding, and dental calculus in adults 30 years of age 34. Claffey N, Nylund K, Kiger R, Garrett S, Egelberg J. Diag-
and older in the United States, 1988-1994. J Periodon- nostic predictability of scores of plaque, bleeding, sup-
tol 1999;70:30-43. puration and probing depth for probing attachment loss.
21. Albandar JM, Brunelle JA, Kingman A. Destructive peri- 3 1/2 years of observation following initial periodontal
odontal disease in adults 30 years of age and older in the therapy. J Clin Periodontol 1990;17:108-114.
United States, 1988-1994. J Periodontol 1999;70:13-29.
22. Caraballo RS, Giovino GA, Pechacek TF, et al. Racial Correspondence: Dr. Thomas Dietrich, Department of
and ethnic differences in serum cotinine levels of ciga- Periodontology, Charité, Humboldt University of Berlin,
rette smokers: Third National Health and Nutrition Exam- Augustenburger Platz 1, 13353 Berlin, Germany. Fax: 49-30-
ination Survey, 1988-1991. JAMA 1998;280:135-139. 4505-62931; e-mail: tommi@tomsoft.com.
23. Korn EL, Graubard BI. Epidemiologic studies utilizing
surveys: Accounting for the sampling design. Am J Pub- Accepted for publication May 23, 2003.
lic Health 1991;81:1166-1173.
24. Slade GD, Beck JD. Plausibility of periodontal disease
estimates from NHANES III. J Public Health Dent 1999;59:
67-72.
25. Meekin TN, Wilson RF, Scott DA, Ide M, Palmer RM.
Laser Doppler flowmeter measurement of relative gin-
gival and forehead skin blood flow in light and heavy
smokers during and after smoking. J Clin Periodontol
2000;27:236-242.
26. Axelsson P, Paulander J, Lindhe J. Relationship between
smoking and dental status in 35-, 50-, 65-, and 75-year-
old individuals. J Clin Periodontol 1998;25:297-305.
27. Linden GJ, Mullally BH. Cigarette smoking and peri-
odontal destruction in young adults. J Periodontol 1994;
65:718-723.
28. Monteiro da Silva AM, Newman HN, Oakley DA, O’Leary
R. Psychosocial factors, dental plaque levels and smok-
ing in periodontitis patients. J Clin Periodontol 1998;25:
517-523.
29. Preber H, Bergstrom J. Cigarette smoking in patients
referred for periodontal treatment. Scand J Dent Res
1986;94:102-108.
30. Bergstrom J, Persson L, Preber H. Influence of cigarette
smoking on vascular reaction during experimental gin-
givitis. Scand J Dent Res 1988;96:34-39.
31. MacGregor ID. Toothbrushing efficiency in smokers and
non-smokers. J Clin Periodontol 1984;11:313-320.
32. Newbrun E. Indices to measure gingival bleeding. J Peri-
odontol 1996;67:555-561.
33. Badersten A, Nilveus R, Egelberg J. Scores of plaque,
bleeding, suppuration and probing depth to predict prob-
ing attachment loss. 5 years of observation following
nonsurgical periodontal therapy. J Clin Periodontol 1990;
17:102-107.

22