RESEARCH AND PRACTICE

Secondhand Smoke and Periodontal Disease: Atherosclerosis

Risk in Communities Study
Anne E. Sanders, PhD, Gary D. Slade, PhD, James D. Beck, PhD, and Helga Ágústsdóttir, PhD

Objectives. We investigated the relationship between secondhand smoke and

Periodontitis is a chronic condition character-
periodontal disease in nonsmokers.
ized by inflammation of the supporting tissues
Methods. We undertook a cross-sectional analysis of the Atherosclerosis Risk
of the teeth, resulting in breakdown of the in Communities study with 2739 lifetime nonsmokers aged 53–74 years, un-
connective tissue attaching the teeth to the exposed to other sources of tobacco, who received a complete periodontal
alveolar bone and eventually to irreversible examination at visit 4. Exposure was reported as average hours per week in close
loss of that bone. Diagnosis is based on signs contact with a smoker in the preceding year. We defined severe periodontitis as 5
of destruction of the connective tissues attach- or more periodontal sites with probing pocket depth of 5 millimeters or more and
ing the tooth root to alveolar bone, which is clinical attachment levels of 3 millimeters or more in those sites. Other outcomes
assessed clinically by measuring attachment were extent of periodontal probing depths of 4 millimeters or more and extent of
level. Attachment level is determined by com- clinical attachment levels of 3 millimeters or more.
Results. In a binary logistic regression model, adjusted odds of severe
bined parameters of probing depth and gingi-
periodontitis for those exposed to secondhand smoke 1 to 25 hours per week
val recession at numerous sites in the mouth.
increased 29% (95% confidence interval = 1.0, 1.7); for those exposed to second-
Ideally, measurements are made at 6 sites per
hand smoke 26 hours per week, the odds were twice as high (95% confidence
tooth in a full-mouth assessment of 28 teeth. interval = 1.2, 3.4) as for those who were unexposed.
Periodontitis is a leading cause of tooth loss and Conclusions. Exposure to secondhand smoke and severe periodontitis among
an important entity in its own right. Also, the nonsmokers had a dose-dependent relationship. (Am J Public Health. 2011;101:
underlying infection and complex host im- S339–S346. doi:10.2105/AJPH.2010.300069)
mune-modulatory and inflammatory responses
that destroy periodontal tissues contribute to
several systemic conditions.1 Systematic reviews
of the evidence have supported a relationship Nonsmokers exposed to secondhand smoke severity of periodontal disease, and measure-
between periodontitis and cardiovascular dis- are recognized to be at increased risk of ment protocol.12 Both studies used a partial-
ease2,3 and type 2 diabetes.4 periodontitis. On reviewing updated evidence mouth measurement protocol limited to sites in
In 2004, the US Surgeon General concluded on involuntary exposure to tobacco smoke, the 2 quadrants. Compared with full-mouth exam-
that the scientific evidence was sufficient to US Surgeon General concluded in 2006 that inations, half-mouth protocols have severely
infer a causal relationship between tobacco there is no risk-free level of exposure to underestimated the prevalence of periodontitis
smoking and periodontitis.5 The etiologic frac- secondhand smoke.9 To date, only 2 studies and produced biased findings, especially in
tion, that is, the fraction of severe periodontal have examined the association of secondhand populations with low prevalence of severe dis-
disease cases in which cigarette smoking expo- smoke exposure and periodontal disease in ease.13---15 Whether such bias alters the presence
sure plays an etiologic role,6 has been estimated adults. One analyzed data from the 3rd National and strength of a relationship between second-
to be 52.8%.7 This percentage indicates that Health and Nutrition Examination Survey hand smoke and periodontal disease is not
approximately one half of periodontitis cases (NHANES III) and found that the odds of known.
could be prevented if cigarette smoking were periodontitis for exposed adults who had never To address this limitation, we used a rigor-
eliminated, with most of that reduction occurring actively smoked were 1.6 times (95% confidence ous protocol, examining 6 sites per tooth in
among people who quit smoking, rather than interval [CI]=1.3, 2.0) as much as those of a full-mouth survey, and applied a stringent
among nonsmokers exposed to secondhand unexposed adults.10 The second study, of 273 case definition of periodontitis to ensure that
smoke. In the US dentate adult population, predominantly male Japanese workers, reported we correctly identified all cases that met the
prevalence of periodontitis, defined as 2 or more higher odds of periodontal disease in passive case definition. The specific study aim was to
sites with clinical attachment loss of at least 4 smokers relative to nonsmokers (odds ratio determine whether exposure to secondhand
millimeters and 1 or more sites with probing [OR] =2.9; 95% CI =1.1, 7.8).11 smoke was associated with periodontal disease
depth of 4 millimeters or deeper, was estimated Although these findings are informative, in lifetime nonsmokers of cigarettes who were
to be 3.6%.8 The extent of attachment loss and some caution is required in their interpretation. unexposed to other sources of tobacco or
prevalence of the disease increases dramatically Estimates of effect size vary considerably nicotine. We hypothesized that nonsmokers
with age. depending on case definition, extent and exposed to secondhand smoke would have

Supplement 1, 2011, Vol 101, No. S1 | American Journal of Public Health Sanders et al. | Peer Reviewed | Research and Practice | S339
 RESEARCH AND PRACTICE

greater odds of severe periodontitis and greater smoked fewer than 400 cigarettes during (CAL3, the proportion of sites probed with
extent of periodontal disease than those who their lifetime as a lifetime nonsmoker. To CAL ‡ 3 mm), and (3) PD4, (the proportion of
were unexposed. determine secondhand smoke exposure par- sites probed with PD‡ 4 mm).
ticipants were asked, ‘‘During the past year, Covariates were known or hypothesized
METHODS about how many hours per week, on the risk indicators for periodontal disease. We
average, were you in close contact with people examined age (in years); education in 3 cate-
Informed consent was obtained from all when they were smoking? For example, in gories (£11 years, 12---16 years, ‡17 years) as
eligible study participants before the dental your home, in a car, at work or other close a marker of socioeconomic position; oral hy-
examination. quarters?’’ Although this measure falls well giene, defined by tooth-brushing frequency
short of obtaining a cumulative lifetime ex- in the previous day (never, once, twice, or 3
Study Participants posure, the duration of exposure is substan- times) and frequency of flossing in the week
Study participants were enrolled in the tially longer than measures of acute exposure before the examination (never, once, twice, or
Atherosclerosis Risk in Communities (ARIC) obtained from serum cotinine. 3 times); and proportion of sites with plaque
study, a multicenter prospective epidemiologic scores of 2 or more. We also considered
cohort study conducted in 4 communities in Examiner Training and Standardization diabetes (fasting glucose level ‡126 mg/dL,
the United States. At baseline (1987---1989), and Collection of Periodontal Clinical nonfasting glucose level‡ 200 mg/dL, taking
15 792 adults aged 45 to 64 years were Data medications for hyperglycemia, or having
selected by probability sampling from Forsyth Dental examiners assessed each tooth for a physician’s diagnosis of diabetes); body mass
County, North Carolina; Jackson, Mississippi; dental plaque using the Silness & Löe Plaque index (weight in kilograms divided by height
suburban Minneapolis, Minnesota; and Wash- Index.16 Probing depth (PD) was determined in meters squared); use of nonsteroidal anti-
ington County, Maryland. The cross-sectional with a UNC-15 periodontal probe at 6 sites per inflammatory drugs (other than aspirin);
data for this analysis were collected during the tooth and recorded in millimeters, with fractions alcohol use (grams of ethanol/wk) or drinking
comprehensive dental examination conducted of millimeters rounded to the next lower unit. As status (current, former, never); coffee consump-
from 1996 through 1998 (visit 4 of the longi- many as 28 teeth were examined for each tion (cups per day), and race (White or Black).
tudinal ARIC study). Of the baseline sample, person; third molars were excluded. At the same Because women were postmenopausal, we
74% participated at visit 4 (n =11656). Of sites, gingival recession was measured as the derived a categorical variable with 3 levels:
these, 4860 did not take part in the periodontal distance from the cemento-enamel junction to female current hormone replacement
assessment because they had no remaining the free gingival margin and recorded in milli- therapy (HRT) users, female non-HRT users,
teeth (n =1651), had medical contraindications meters, with fractions of millimeters rounded to and men.
(n =1621), refused (n =1317), or had another the next lower unit. We computed clinical When we analyzed data as an overall sum-
reason (n = 271). Of the 6796 who had a peri- attachment level (CAL) during data analysis by mary of responses from all probed sites (i.e.,
odontal assessment, 4057 were omitted from adding the PD to the gingival recession. Exam- severe periodontitis case status), multivariate
this analysis because they were a current or iners assessed the presence or absence of bleed- models included the number of remaining
former smoker (n = 3640), had used another ing on probing after each quadrant of probing at teeth to include the dimension of teeth that
form of nicotine (n = 381), self-identified as 6 sites on all teeth. were at risk. We did not include the number of
non-White and non-Black (n = 25), or had All dental examiners received the same remaining teeth for the CAL3 or PD4 models
another reason (n =11). Hence, the total sample training and calibration. During calibration, because the number of probing sites available
consisted of 2739 ARIC study participants. each examiner was matched with the gold- was included as the denominator for the pro-
standard examiner and another examiner on at portion.
Main Exposure least 5 occasions. The weighted j was 0.90 for As a result of the method of sampling, race
Interviewers administered a questionnaire at PD and 0.82 for CAL within 1 millimeter. Over was incompletely distributed within the
visit 4 to collect self-report information about the 2-year course of examinations, we con- study center locations. Therefore, we created
health status, medication usage, and health ducted quality assurance through conference a combined variable of center---race with 5
behavior. We obtained detailed information calls, site visits, and recalibration to maintain levels: (1) Forsyth County, North Carolina––
about active and past cigarette smoking, as well standardization of examiners. Blacks, (2) Forsyth County, North Carolina––
as lifetime use of pipes, cigars, cigarillos, Whites, (3) Jackson, Mississippi––Blacks, (4)
chewing tobacco, snuff, nicotine gum, and Dependent Variable and Covariates Minneapolis, Minnesota––Whites and (5)
nicotine patches. We omitted people exposed We selected thresholds for 3 periodontal Washington County, Maryland––Whites. The
to any of these sources of tobacco from the disease outcome measures that have been used combined center---race variable therefore
analysis to eliminate possible bias from other previously17---19: (1) case definition of severe adjusted for both race and center differences.
sources of tobacco. periodontitis as 5 or more sites with CAL 3 The center---race variable also reflected any
Using the ARIC study classification, we millimeters or more and PD 5 mm or more in the effects resulting from different examiners
categorized participants who indicated having same sites and (2) extent of periodontal disease across centers.

S340 | Research and Practice | Peer Reviewed | Sanders et al. American Journal of Public Health | Supplement 1, 2011, Vol 101, No. S1
 RESEARCH AND PRACTICE

Statistical Analysis assumes independence of the k dichotomous Covariate Associations With Case Status
We tested our hypothesis of an association responses (0 or 1) at each site. It is reason- and Secondhand Smoke
between periodontal disease and secondhand able to expect that individuals who have 1 The covariates significantly associated with
smoke exposure among lifetime nonsmokers diseased site (CAL ‡ 3 mm or PD ‡ 4) are severe periodontitis after adjustment for the
with no exposure to other tobacco products. more likely to have other diseased sites. This key factors of age, gender---HRT, education,
We coded all variables that had more than 2 intraindividual correlation contributes to and center---race (P < .5) were body mass index
levels as indicator variables and binary vari- extrabinomial variation in the data. To allow and dental visiting pattern (Table 1). When
ables as 0 or 1. We created categorical vari- for this potential overdispersion, we assumed adjusting for key factors, alcohol and coffee
ables for the continuous variables for an that explanatory factors influenced the consumption were significantly associated with
adjusted bivariate analysis. We calculated the proportion of diseased sites, pi = Zn / n, through secondhand smoke exposure (P < .5), along
means and proportions of covariates, adjusting a logistic link function and estimated model with body mass index (Table 1). Periodontal
for age, gender---HRT use, and education and parameters using generalized estimating assessment parameters of bleeding on probing,
center---race, stratified on both severe peri- equations methods. The appeal of this ap- CAL, probing pocket depth, and dental plaque
odontitis case status and exposure to second- proach is that the empirical or robust standard were significantly associated with case status
hand smoke (coded as 1 for ‡1 h/wk and 0 for errors for the parameter estimates are consis- (Table 2), and, with the exception of extent
unexposed). tent, even if the representation of the variance of probing pocket depths of 4 millimeters or
We tested the covariates related to both in the estimating equations is misspecified. more, all periodontal parameters were signifi-
severe periodontitis case status and second- When we tested the main exposure variable as cantly associated with secondhand smoke
hand smoke exposure (P < .1), adjusting for age, a continuous variable, we tested transforma- exposure.
education, and center---race, in multivariable tion––such as quadratic terms, exponential
models. Age, gender---HRT use, education, and terms, logarithmic and exponential transfor- Multivariable Analysis
center---race were forced into all models irre- mation of the main exposure variable––in the Severe Periodontitis. Mean secondhand
spective of their statistical significance. We binomial multivariate models to capture the smoke exposure (average hours per week over
built the models by adding 1 covariate at a time shape of the association better than with the past year) was significantly higher for cases
and evaluated each variable using type III tests a straight-line model. We conducted all analy- (mean = 4.3; 95% CI = 3.0, 5.6) than for non-
(i.e., as though each variable were the last ses using SAS, version 9.1 (SAS Institute, cases (mean = 3.2; 95% CI = 2.8, 3.6). Among
added). Cary, NC). those exposed for 1 to 25 hours per week,
If the parameter estimates for each level of mean secondhand smoke exposure was 4.5
secondhand smoke exposure were changed by RESULTS hours (95% CI = 4.1, 4.8). Among those ex-
more than 10% or if the coefficient for that posed for more than 25 hours per week, mean
covariate was statistically significant at P < .05, Study participants were aged between 53 secondhand smoke exposure was 48.0 hours
we kept the covariate in the model and entered and 74 years (mean = 62.4). The sample was (95% CI = 44.3, 51.6). We observed a signifi-
the next covariate. predominantly female (74.7%), and Blacks cant dose-dependent relationship of second-
We used a binary logistic regression model made up 20.2%. Exposure to secondhand hand smoke exposure and severe periodontitis
to estimate the risk of secondhand smoke smoke for 1 hour or more a week was reported prevalence (Table 3). The odds for people
exposure on severe periodontitis (case vs. by 33.7% (n = 923) of adults. exposed to 1 to 25 hours per week was 29%
noncase), controlling for potential con- Of all participants, 4.3% reported weekly higher (95% CI =1.0, 1.7) and the odds for
founders. The decision to set the threshold exposure exceeding 25 hours (n =117). The people exposed to 26 hours per week or more
for high secondhand smoke exposure at mean exposure, adjusted for age, gender---HRT, was twice as high (95% CI =1.2, 3.4) than that
more than 25 hours per week was based on education, and center---race for the participants of people with less than 1 hour per week of
the distribution, because we know of no who reported secondhand smoke exposure, secondhand smoke exposure, adjusting for age,
preexisting threshold. In this study, 6.4% of all was 9.97 hours a week (SE = 0.6), and the education, center---race, gender---HRT use,
participants (exposed and unexposed to range of exposure was 1 to 108 hours per week. pattern of dental visits, dental plaque, and
secondhand smoke) and 12.7% of all partici- Severe periodontitis was found in 16.0% number of teeth remaining.
pants exposed to secondhand smoke were (n = 438) of participants. On average, 5.2% Extent of Attachment Loss and Probing Depth.
exposed to high levels of secondhand smoke (SE = 0.2; range = 0---97) of participants had The final models for extent of CAL3 and PD4
(> 25 h/wk). periodontal sites with probing depths deeper contained the same covariates: age, gender---
For analyses of extent variables, where the than or equal to 4 mm. The mean extent HRT use, education, center---race, extent of
response was Zn, the number of diseased of CAL 3 millimeters or more was 16.5% dental plaque scores, and pattern of dental
sites (CAL ‡ 3 mm or PD ‡ 4) among n probed (SE = 0.3; range = 0---100). The mean number visits (Table 4).
sites, we treated Zn as a binomially distributed of retained teeth was 22.6 (SE = 0.13; The adjusted odds of periodontal sites
random variable. The binomial distribution range = 2---32). having clinical attachment levels 3 millimeters

Supplement 1, 2011, Vol 101, No. S1 | American Journal of Public Health Sanders et al. | Peer Reviewed | Research and Practice | S341
 RESEARCH AND PRACTICE

or more increased by a factor of 1.1 (95%

TABLE 1—Relationship of Covariates With Severe Periodontitis and Secondhand Smoke CI =1.0, 1.2) for people with 1 to 25 hours per
Exposure, Adjusted for Age, Gender–HRT, Education, and Center–Race: Atherosclerosis week of secondhand smoke exposure, which
Risk in Communities Study; Forsyth County, NC; Jackson, MS; suburban Minneapolis, was not significantly greater than that for un-
MN; and Washington County, MD; 1996–1998 exposed people. However for people with 26
Periodontitis Exposed to hours per week or more of secondhand smoke
Variable No. Cases, % P Secondhand Smoke, % P exposure, the risk increased significantly by
a factor of 1.3 (95% CI =1.0, ---1.6), compared
Severe periodontitis case status .08
with people with less than 1 hour per week of
Cases 438 ... 37.6
secondhand smoke exposure.
Noncases 2301 ... ... 33.0
For periodontal pocket depths of 4 millime-
Secondhand smoke categories .099
ters or more (Table 4), we observed a trend of
No exposure 1816 15.1 ...
increasing extent of deep pockets with greater
1–25 h/wk 806 17.2 ...
exposure, but the confidence intervals of the
‡ 26 h/wk 117 22.4 ...
odds ratios enclosed unity and hence did not
Gender–HRTa,b .001 .004
differ significantly from those unexposed to
Women using HRT 817 9.7 30.4
secondhand smoke.
Women not using HRT 1229 16.0 33.3
Men 692 23.5 38.4
DISCUSSION
Age,a y .350 .013
53–61 1313 15.3 36.0
In this large community-dwelling sample of
62–74 1426 16.6 31.6
lifetime nonsmokers unexposed to other to-
Race/ethnicitya .107 .101
bacco products, those exposed to secondhand
Black 554 18.2 36.6
smoke had a higher prevalence of severe
White 2185 15.4 33.0
periodontitis, after controlling for known risk
Center-racea .001 .001
indicators for periodontal disease. Exposure to
Forsyth—Blacks 52 30.8 59.6
secondhand smoke also showed a dose-de-
Forsyth—Whites 638 6.1 39.5
pendent increase in extent of periodontal dis-
Jackson—Blacks 502 16.9 34.2
ease, but the association was statistically sig-
Minneapolis—Whites 793 15.6 30.3
nificant only for CAL and only at the higher
Washington—Whites 754 23.1 30.2
level of exposure. Our findings were consistent
Education,a,b y .102 .001
with the 2 earlier studies10,11 that used the
£ 11 337 19.6 37.4
partial-mouth examination. They also build on
12–16 1266 16.2 37.9
evidence of lower CALs in children exposed to
‡ 17 1132 14.8 28.0
secondhand smoke in the home.20 Previous
Alcohol .359 .025
studies13---15,21 have shown that bias in prevalence
Current drinker 1134 17.3 35.4
estimates is reduced by using a full-mouth
Former drinker 642 15.6 36.3
assessment of 6 sites per tooth, as we did. If the
Never drinker 963 14.7 30.2
degree of bias in disease measurement is equal
Alcohol consumption .453 .012
in exposed and unexposed people (i.e., the
< 54 g/wk (5 glasses of wine) 2492 15.8 33.0
assumption of nondifferential misclassification),
‡ 54 g/wk 247 17.8 41.4
ORs for the association between exposure and
Coffee (cups) .628 .001
disease are usually biased toward the null, with
> 4/d 277 14.8 34.4
the amount of bias proportional to the degree of
1–3/d 1102 18.4 41.0
bias in disease measurement. Methods that re-
1–24/mo 457 15.6 36.7
duce bias in estimating prevalence might there-
Almost never 903 16.4 27.6
fore be expected to yield larger exposure---disease
Body mass index (kg/m2)b .021 .001
ORs. Yet, we found that the effect sizes at 2
< 27 1148 14.0 29.9
thresholds of secondhand smoke exposure
‡ 27 1584 17.5 36.7
(adjusted OR [AOR] =1.3 and 2.0, respectively)
Continued tended to be smaller than those previously
reported in nonsmokers (OR=1.6 in Arbes
et al.10; OR= 2.9 in Yamamoto et al.11). One

S342 | Research and Practice | Peer Reviewed | Sanders et al. American Journal of Public Health | Supplement 1, 2011, Vol 101, No. S1
 RESEARCH AND PRACTICE

however, seem to differ enough among smok-

TABLE 1—Continued ing groups to explain the strong association of
Diabetesb .848 .232 smoking with periodontal disease.28,29 The
Yes 336 16.1 36.8 effect of active cigarette smoking on the patho-
No 2394 15.7 33.3 genesis of periodontal disease is now believed
Current use of aspirinb .362 .663 to be exerted through both local and systemic
Yes 684 14.9 34.5 pathways.30 The local effects are thought to be
No 2050 16.4 33.5 mediated by the chemical stimuli and include
Current use of other NSAIDsb .369 .573 local vasoconstriction by nicotine and decreased
Yes 361 14.3 35.1 oxygen tension as well as hyperkeratosis of the
No 2374 16.3 33.6 gingival tissues. Local effects possibly include
Times brushed teeth yesterdayb .644 .211 effects of the physical heat from cigarette smoke,
Not at all 26 9.3 17.1 although no studies have been conducted to
1 642 16.8 34.0 confirm that type of effect. The more important
2 1654 15.7 34.5 pathway of the 2 is believed to be the systemic
‡3 407 16.5 31.5 alteration of the host response.29,31 Several
Times used floss last weekb .377 .058 studies have noted impaired chemotaxis and
Not at all 819 17.3 35.7 phagocytosis32 of both oral and peripheral neu-
1 225 15.7 31.2 trophils of smokers as well as reduced antibody
2 267 18.5 39.6 production.33,34 Smoking is associated with sup-
‡3 1415 14.9 31.9 pressed salivary osteocalcin levels, implying a
Frequency of dental visitsb .036 .096 pathogenic mechanism via reduced bone min-
Other 25 17.7 49.2 eral density.35 Very few studies have assessed
Don’t go 32 33.3 29.9 the host response in periodontal tissues of
Only when in discomfort 176 22.9 40.0 passive smokers. One study comparing non-
When something needs fixing 400 15.9 37.5 smokers and passive smokers found elevated
Regular basis 2094 15.2 32.4 levels of salivary biomarkers for periodontal
disease in passive smokers compared with non-
Note. HRT = hormone replacement therapy; NSAIDs = nonsteroidal anti-inflammatory drugs. smokers.36
a
Not adjusted.
b
Numbers do not sum to 2739 because of missing values.
Strengths and Limitations
This study was limited to individuals whose
explanation is that bias in periodontal disease is wide range of covariates used in multivariable only exposure to tobacco products was sec-
differential with respect to smoking, for example, regression models adds to the argument that ondhand smoke. Elimination of other sources
if partial-mouth recordings produce greater un- this association is not a spurious one caused by of smoke and nicotine makes these findings
derestimation of disease in smokers than in residual confounding. even more compelling and establishes that this
nonsmokers. However, differences in effect esti- There is reason to believe that passive excess risk is the result of secondhand smoke
mates may be because of other conspicuous smoking exerts similar systemic effects on the exposure alone and not of active smoking by
differences among these studies, including their periodontal tissues as observed in active the individuals themselves. The magnitude of
age distributions and measures of exposure and smokers, based on studies that have found that this public health problem is considerable,
outcome. Nevertheless, the direction of effect was active and passive smoking have effects in the considering that one third of nonsmoking in-
consistent across all 3 studies, and the magnitude same direction, although not the same magni- dividuals were exposed. The large number of
of effect was approximately equal, establishing tude, on other health outcomes.22---24 Because participants permitted comparisons and ad-
consistency of evidence. secondhand smoke is inhaled, 1 likely biological justments for other risk factors than a smaller
mechanism of effect is through systemic effects sample would have allowed. The extensive
Possible Mechanisms and Explanations of toxic constituents in the tobacco smoke.25 amount of clinical and lifestyle information
A possible explanation for why the associa- These effects may be mediated through injury collected made it possible to examine this
tion of exposure to secondhand smoke and inflicted by proinflammatory agents, such as cyto- association more thoroughly than otherwise
severe periodontitis was statistically significant kines26 or smoking-induced oxidative stress.27 possible.
whereas the association with extent scores was Some authors have attributed the difference One reason for the small crude effect is that
weaker may be that our periodontitis case in periodontal health among active cigarette adults who were less likely to be cases were
definition was more stringent and recognized smoking groups to better oral hygiene among more likely to be exposed to secondhand
only truly serious periodontal disease. The nonsmokers. Plaque accumulation does not, smoke. For example, at the Forsythe study site,

Supplement 1, 2011, Vol 101, No. S1 | American Journal of Public Health Sanders et al. | Peer Reviewed | Research and Practice | S343
 RESEARCH AND PRACTICE

relationship of secondhand smoke and peri-

TABLE 2—Relationship of Periodontal Examination Parameters With Severe Periodontitis odontal disease. Therefore, periodontal
and Secondhand Smoke Exposure, Adjusted for Age, Gender–Hormone Replacement disease may possibly have existed before
Therapy, Education, and Center–Race: Atherosclerosis Risk in Communities Study; Forsyth the exposure. Only longitudinal data could
County, NC; Jackson, MS; suburban Minneapolis, MN; and Washington County, MD; confirm that exposure preceded the onset
1996–1998 of disease. However, reverse causation is
Periodontitis Exposure to implausible: There is no reason to believe that
Examination Parameters No. case, % P Secondhand Smoke, % P people would become exposed to secondhand
smoke as a consequence of developing peri-
Bleeding on probing .001 .005
odontitis. Exposure was self-reported, which is
< 15% of sites 1155 3.4 30.4
also a limitation, but self-report is the most
‡ 15% of sites 1584 25.3 36.2
common way to collect information on life-
Clinical attachment levels ‡ 3 mm .001 .017
style factors. No biomarkers, such as cotinine
< 60% of sites 2626 13.7 33.3
levels, were available for this study to confirm
‡ 60% of sites 113 68.9 44.8
the reported exposure, but in an analysis of
Clinical attachment levels ‡ 3 mm .001 .018
a study population representative of the US
< 25% of sites 2156 7.9 33.3
population10 (NHANES III), serum cotinine
‡ 25% of sites 583 46.0 44.7
levels generally confirmed the self-reported
Probing pocket depths ‡ 4 mm NA .204
exposures. Using a threshold of 10 nanograms
< 2% of sites 1428 0.0 32.6
per milliliter of serum cotinine as an indicator
‡ 2% of sites 1311 33.4 35.1
of current cigarette smoking, only 3.4% of non-
Dental plaque, PI ‡ 1a .001 .009
smokers reportedly exposed to secondhand
< 15% of sites 949 8.3 30.5
smoke and 0.8% of people not exposed to
‡ 15% of sites 1639 19.5 36.0
secondhand smoke were likely current smokers
Dental plaque, PI ‡ 2a .001 .055
or users of other tobacco products.
< 15% of sites 2399 14.1 33.4
A related concern is this study’s measure of
‡ 15% of sites 189 32.1 40.9
secondhand smoke during the preceding
Note. NA = not applicable; PI = increase in dental plaque. 12 months, creating the potential for mis-
a
Numbers do not total 2739 due to missing values. classifying exposure over the longer period
in which harmful compounds in tobacco
6% of White adults were cases compared with We found a stronger effect when secondhand smoke probably contribute to destruction of
16% to 23% of adults at other study sites. Yet smoke was defined over 3 levels. Here, the periodontal tissues (possibly a decade or
secondhand smoke exposure tended to be odds for disease in those exposed for 26 hours more). For 2 reasons, we believe the degree
higher in Whites at the Forsythe site than at the per week were elevated 45% relative to those of misclassification is probably less in this
other study sites, biasing estimates toward the unexposed, and the effect became significant study than in other studies. First, studies such
null. As a result, cases and controls did not after adjustment for covariates including study as NHANES ask only about current second-
differ in secondhand smoke exposure when site (AOR = 2.0). The cross-sectional design hand smoke exposure. For example, during
secondhand smoke was dichotomized (P = .08). did not permit us to establish the temporal the in-home interview for NHANES III, in-
terviewers asked, ‘‘Does anyone who lives
here smoke cigarettes in the home?’’10 In
TABLE 3—Binary Logistic Regression Analysis of the Relationship Between Secondhand
contrast, we asked about exposures during
Smoke and Severe Periodontitis: Atherosclerosis Risk in Communities Study; Forsyth County,
preceding 12 months. Second, people in this
NC; Jackson, MS; suburban Minneapolis, MN; and Washington County, MD; 1996–1998
study were aged 53---74 years, ages at which
Level of Secondhand Smoke Exposure Unadjusted OR (95% CI) AOR (95% CI)a most of those exposed to secondhand smoke
experience that exposure in the home, not the
‡ 26 h/wk 1.5 (0.9, 2.3) 2.0 (1.2, 3.4)
workplace. Moreover, domestic living arrange-
1–25 h/wk 1.2 (0.9, 1.9) 1.3 (1.0, 1.7)
ments tend to be quite stable in this age.
Not exposed to secondhand smoke (Ref) 1.0 1.0
Therefore, most people’s exposure during the
Note. AOR = adjusted odds ratio; CI = confidence interval; OR = odds ratio. Severe periodontitis case status was defined as ‡ 5 preceding 12 months is probably a good proxy
sites with clinical attachment level ‡ 3 mm and probing depth ‡ 5 mm. The pseudo-R2 statistic = 0.002 for the unadjusted for their exposure during several of the pre-
model and 0.16 for the adjusted model. ORs and 95% CI are unadjusted and adjusted for covariates.
a
Adjusted for age, education, center and race, gender and hormone replacement therapy, remaining teeth, extent of plaque ceding years.
scores ‡ 2, and dental visiting pattern. In summary, we found statistically signifi-
cantly higher prevalence of severe periodontitis

S344 | Research and Practice | Peer Reviewed | Sanders et al. American Journal of Public Health | Supplement 1, 2011, Vol 101, No. S1
 RESEARCH AND PRACTICE

destructive periodontal disease: a meta-analysis. Oral

TABLE 4—Multivariate Binomial Regression Model of the Relationship Between Health Prev Dent. 2009;7(2):107---127.
Secondhand Smoke and 2 Periodontal Disease Parameters: Atherosclerosis Risk in 5. Health Consequences of Smoking: A Report of the
Surgeon General. Atlanta, GA: Centers for Disease Con-
Communities Study; Forsyth County, NC; Jackson, MS; Suburban Minneapolis, MN; and
trol and Prevention; 2004.
Washington County, MD; 1996–1998
6. Greenland S, Robins JM. Conceptual problems in the
definition and interpretation of attributable fractions. Am
Level of Secondhand Clinical attachment level ‡ 3 mm Periodontal pockets ‡ 4 mm
J Epidemiol. 1988;128(6):1185---1197.
Smoke Exposure Unadjusted OR (95% CI) AOR (95% CI)a Unadjusted OR (95% CI) AOR (95% CI)a 7. Tomar SL, Asma S. Smoking-attributable periodon-
titis in the United States: findings from NHANES III.
‡ 26 h/wk 1.4 (1.1, 1.8) 1.3 (1.0, 1.6) 1.2 (0.9, 1.7) 1.3 (0.9, 1.8) National Health and Nutrition Examination Survey. J
1–25 h/wk 1.1 (1.0, 1.2) 1.1 (1.0, 1.2) 1.2 (1.0, 1.3) 1.1 (1.0, 1.9) Periodontol. 2000;71(5):743---751.
Unexposed (Ref) 1.0 1.0 1.0 1.0 8. Borrell LN, Crawford ND. Social disparities in
periodontitis among United States adults 1999-2004.
Note. AOR = adjusted odds ratio; CI = confidence interval; OR = odds ratios. Community Dent Oral Epidemiol. 2008;36(5):383---391.
a
Adjusted for age, education, center and race, gender and hormone replacement therapy, remaining teeth, extent of plaque 9. Health Consequences of Involuntary Exposure to To-
scores ‡ 2, and dental visiting pattern. bacco Smoke: A Report of the Surgeon General. Atlanta,
GA: Centers for Disease Control and Prevention; 2006.
10. Arbes SJ Jr, Agustsdottir H, Slade GD. Environ-
mental tobacco smoke and periodontal disease in the
United States. Am J Public Health. 2001;91(2):253---257.
11. Yamamoto Y, Nishida N, Tanaka M, et al. Associa-
but not greater extent of periodontal disease in Contributors tion between passive and active smoking evaluated by
lifetime nonsmokers exposed to secondhand A. E. Sanders wrote the article, which was based on the
salivary cotinine and periodontitis. J Clin Periodontol.
findings of H. Ágústsdóttir’s doctoral research. H.
smoke, compared with those not exposed to Ágústsdóttir conducted the research, including the
2005;32(10):1041---1046.
secondhand smoke. The results build on ear- statistical analysis. J. D. Beck was principal investigator of 12. Page RC, Eke PI. Case definitions for use in
the dental Atherosclerosis Risk in Communities (ARIC) population-based surveillance of periodontitis. J Peri-
lier reports of a relationship between exposure
study that conducted oral examinations of ARIC study odontol. 2007;78(7, suppl):1387---1399.
to secondhand smoke and periodontal disease. participants at visit 4, and he also reviewed the article. 13. Hunt RJ, Fann SJ. Effect of examining half the teeth
Longitudinal studies are needed to evaluate G. D. Slade was mentor to H. Ágústsdóttir and guided the in a partial periodontal recording of older adults. J Dent
the exposure---outcome sequence, as are ex- statistical analysis. He also analyzed data in response to Res. 1991;70(10):1380---1385.
reviewers’ comments.
perimental studies to provide evidence of 14. Susin C, Kingman A, Albandar JM. Effect of partial
recording protocols on estimates of prevalence of peri-
biological mechanisms. Our findings suggest
that secondhand smoke exposure should be
Acknowledgments odontal disease. J Periodontol. 2005;76(2):262---267.
The ARIC study is carried out as a collaborative study 15. Kingman A, Susin C, Albandar JM. Effect of partial
taken into account in future studies of peri- supported by National Heart, Lung and Blood Institute recording protocols on severity estimates of periodontal
odontal disease. (contracts N01-HC-55015, N01-HC-55016, N01-HC- disease. J Clin Periodontol. 2008;35(8):659---667.
55018, N01-HC-55019, N01-HC-55020, N01-HC-
Secondhand smoke remains a serious public 55021, and N01-HC-55022). The collection and analy-
16. Silness J, Loe H. Periodontal disease in pregnancy. II.
Correlation between oral hygiene and periodontal con-
health hazard, with 40% of the nonsmoking sis of dental data were supported by National Institute
dition. Acta Odontol Scand. 1964;22:121---135.
population aged 3 years or older being exposed of Dental and Craniofacial Research (grants DE13807-
01A1 and DE11551). 17. Arbes SJ Jr, Slade GD, Beck JD. Association between
in the United States during 2007---2008.37 We thank the staff and participants of the ARIC study extent of periodontal attachment loss and self-reported
Although understanding the pathogenesis of for their important contributions. history of heart attack: an analysis of NHANES III data.
J Dent Res. 1999;78(12):1777---1782.
secondhand smoke in periodontal disease has
18. Oliver RC, Brown LJ, Loe H. Variations in the
scientific value, the health of the nation is best Human Participant Protection prevalence and extent of periodontitis. J Am Dent Assoc.
served through initiatives with demonstrated No protocol approval was necessary because this study
1991;122(6):43---48.
involved the analysis of secondary data only.
effectiveness in limiting secondhand smoke ex- 19. Brown LJ, Brunelle JA, Kingman A. Periodontal
posure. j status in the United States, 1988-1991: prevalence,
References extent, and demographic variation. J Dent Res. 1996;
1. Beck JD, Offenbacher S. Relationships among clinical 75(spec no):672---683.
measures of periodontal disease and their associations
20. Erdemir EO, Sönmez IS, Oba AA, Bergstrom J,
About the Authors with systemic markers. Ann Periodontol. 2002;7(1): Caglayan O. Periodontal health in children exposed
At the time of the study, Anne E. Sanders, PhD, Gary D. 79---89. to passive smoking. J Clin Periodontol. 2010;37(2):
Slade, and James D. Beck were with the Department of 2. Humphrey LL, Fu R, Buckley DI, Freeman M, 160---164.
Dental Ecology, School of Dentistry, University of North Helfand M. Periodontal disease and coronary heart 21. Eke PI, Thornton-Evans GO, Wei L, Borgnakke WS,
Carolina at Chapel Hill. Helga A´gu´stsdo´ttir was with the disease incidence: a systematic review and meta-analysis. Dye BA. Accuracy of NHANES Periodontal Examination
Ministry of Health and Social Security, Reykjavik, Iceland. J Gen Intern Med. 2008;23(12):2079---2086. Protocols. J Dent Res. 2010;89(11):1208---1213.
Correspondence should be sent to Anne E. Sanders, PhD,
3. Blaizot A, Vergnes JN, Nuwwareh S, Amar J, Sixou M. 22. Howard G, Wagenknecht LE, Burke GL, et al.
Department of Dental Ecology, UNC School of Dentistry,
Periodontal diseases and cardiovascular events: meta- Cigarette smoking and progression of atherosclerosis:
Campus Box 7450, Chapel Hill, NC 27599 (e-mail:
analysis of observational studies. Int Dent J. 2009;59(4): The Atherosclerosis Risk in Communities (ARIC) Study.
[email protected]). Reprints can be ordered
at http://www.ajph.org by clicking the ‘‘Reprints/Eprints’’ 197---209. JAMA. 1998;279(2):119---124.
link. 4. Chavarry NG, Vettore MV, Sansone C, Sheiham A. 23. Howard G, Wagenknecht LE, Cai J, Cooper L, Kraut
This article was accepted November 9, 2010. The relationship between diabetes mellitus and MA, Toole JF. Cigarette smoking and other risk factors for

Supplement 1, 2011, Vol 101, No. S1 | American Journal of Public Health Sanders et al. | Peer Reviewed | Research and Practice | S345
 RESEARCH AND PRACTICE

silent cerebral infarction in the general population. Stroke.

1998;29(5):913---917.
24. Chen LC, Quan C, Hwang JS, et al. Atherosclerosis
lesion progression during inhalation exposure to envi-
ronmental tobacco smoke: a comparison to concentrated
ambient air fine particles exposure. Inhal Toxicol. 2010;
22(6):449---459.
25. Vardavas CI, Panagiotakos DB. The causal rela-
tionship between passive smoking and inflammation on
the development of cardiovascular disease: a review of
the evidence. Inflamm Allergy Drug Targets. 2009;8(5):
328---333.
26. Nishida N, Yamamoto Y, Tanaka M, et al. Associa-
tion between involuntary smoking and salivary markers
related to periodontitis: a 2-year longitudinal study. J
Periodontol. 2008;79(12):2233---2240.
27. Stangherlin EC, Luchese C, Ardais AP, Nogueira
CW. Passive smoke exposure induces oxidative damage
in brains of rat pups: protective role of diphenyl dis-
elenide. Inhal Toxicol. 2009;21(10):868---874.
28. Bergstrom J, Eliasson S, Preber H. Cigarette smoking
and periodontal bone loss. J Periodontol. 1991;62(4):
242---246.
29. Bergstrom J, Preber H. Tobacco use as a risk factor.
J Periodontol. 1994;65(5, suppl):545---550.
30. Salvi GE, Brown CE, Fujihashi K, et al. Inflammatory
mediators of the terminal dentition in adult and early
onset periodontitis. J Periodontal Res. 1998;33(4):
212---225.
31. Pabst MJ, Pabst KM, Collier JA, et al. Inhibition of
neutrophil and monocyte defensive functions by nicotine.
J Periodontol. 1995;66(12):1047---1055.
32. MacFarlane GD, Herzberg MC, Wolff LF, Hardie
NA. Refractory periodontitis associated with abnormal
polymorphonuclear leukocyte phagocytosis and cigarette
smoking. J Periodontol. 1992;63(11):908---913.
33. Gunsolley JC, Pandey JP, Quinn SM, Tew J, Schenkein
HA. The effect of race, smoking and immunoglobulin
allotypes on IgG subclass concentrations. J Periodontal
Res. 1997;32(4):381---387.
34. Tangada SD, Califano JV, Nakashima K, et al. The
effect of smoking on serum IgG2 reactive with Actino-
bacillus actinomycetemcomitans in early-onset peri-
odontitis patients. J Periodontol. 1997;68(9):842---850.
35. Gurlek O, Lappin DF, Buduneli N. Effects of smoking
on salivary C-telopeptide pyridinoline cross-links of type I
collagen and osteocalcin levels. Arch Oral Biol. 2009;
54(12):1099---1104.
36. Nishida N, Yamamoto Y, Tanaka M, et al. Associa-
tion between passive smoking and salivary markers
related to periodontitis. J Clin Periodontol. 2006;33(10):
717---723.
37. Vital signs: nonsmokers’ exposure to secondhand
smoke––United States, 1999-2008. MMWR Morb Mortal
Wkly Rep. 2010;59(35):1141---1146.

S346 | Research and Practice | Peer Reviewed | Sanders et al. American Journal of Public Health | Supplement 1, 2011, Vol 101, No. S1