701901

research-article2017
JDRXXX10.1177/0022034517701901Journal of Dental ResearchPeriodontal Treatment and Future Cardiovascular Disease

Research Reports: Clinical

Journal of Dental Research
1­–6
Poor Response to Periodontal © International & American Associations
for Dental Research 2017

Treatment May Predict Future Reprints and permissions:

sagepub.com/journalsPermissions.nav

Cardiovascular Disease DOI: 10.1177/0022034517701901

https://doi.org/10.1177/0022034517701901
journals.sagepub.com/home/jdr

A. Holmlund1, E. Lampa2, and L. Lind2

Abstract
Periodontal disease has been associated with cardiovascular disease (CVD), but whether the response to the treatment of periodontal
disease affects this association has not been investigated in any large prospective study. Periodontal data obtained at baseline and 1 y
after treatment were available in 5,297 individuals with remaining teeth who were treated at a specialized clinic for periodontal disease.
Poor response to treatment was defined as having >10% sites with probing pocket depth >4 mm deep and bleeding on probing at ≥20%
of the sites 1 y after active treatment. Fatal/nonfatal incidence rate of CVD (composite end point of myocardial infarction, stroke, and
heart failure) was obtained from the Swedish cause-of-death and hospital discharge registers. Poisson regression analysis was performed
to analyze future risk of CVD. During a median follow-up of 16.8 y (89,719 person-years at risk), those individuals who did not respond
well to treatment (13.8% of the sample) had an increased incidence of CVD (n = 870) when compared with responders (23.6 vs. 15.3%,
P < 0.001). When adjusting for calendar time, age, sex, educational level, smoking, and baseline values for bleeding on probing, probing
pocket depth >4 mm, and number of teeth, the incidence rate ratio for CVD among poor responders was 1.28 (95% CI, 1.07 to 1.53;
P = 0.007) as opposed to good responders. The incidence rate ratio among poor responders increased to 1.39 (95% CI, 1.13 to 1.73; P
= 0.002) for those with the most remaining teeth. Individuals who did not respond well to periodontal treatment had an increased risk
for future CVD, indicating that successful periodontal treatment might influence progression of subclinical CVD.

Keywords: periodontal disease(s)/periodontitis, patient outcome, inflammation, epidemiology, oral-systemic disease(s), cardiovascular
disease(s)

Introduction Heart failure (HF) has a multifactorial etiology. A major

Periodontal disease (PD) has been associated with cardiovas-
cular disease (CVD) in several studies (Wu, Trevisan, Genco,
Dorn, et al. 2000; Desvarieux et al. 2003; Joshipura et al. 2003;
Grau et al. 2004; Gotsman et al. 2007). Despite the difficulty of
comparing studies due to the lack of consensus regarding when
PD is present, a few meta-analysis and systematic reviews
have been performed showing a moderate increased risk for
individuals with PD regarding incident CVD (relative risk,
1.19 to 1.50; Janket et al. 2003; Humphrey et al. 2008).
Although it is clear that an association between PD and
CVD exists, we still do not know if this association is causal.
Treatment studies of PD have shown that inflammatory mark-
ers related to increased CVD risk, such as C-reactive protein
levels and levels of fibrinogen, can be reduced (D’Aiuto et al.
2004; Taylor et al. 2006); it has also been shown that endothe-
lial function could be improved (Tonetti et al. 2007), indicating
that periodontal treatment could be of value for reducing the
risk for CVD. To date, no randomized studies have investi-
gated whether treatment of PD could affect the risk for future
CVD, because it is unethical to withhold treatment from the
control group in such trials. However, observational studies
highlighting how periodontal treatment outcome would affect
future risk for CVD would be of value, yet none have been
published.
cause of HF is damage to the myocardial tissue following an
acute myocardial infarction (MI). Other important contributors
to HF are diabetes, hypertension, and atrial fibrillation (Dhingra
and Vasan 2012; Kalogeropoulos et al. 2012; Emdin et al.
2016). The multifactorial etiology for HF is probably a main
reason why it has, to our knowledge, sparsely been studied in
relation to oral health conditions. However, in the last decade,
researchers have suggested that systemic low-grade inflamma-
tion is a possible cause not only for MI but also for HF (Wrigley
et al. 2011; Kalogeropoulos et al. 2012) and therefore might be
influenced by PD.
The aim of this study was to investigate whether incidence
of CVD (a composite variable of MI, stroke, and HF) differs
between those who respond well to treatment of PD and those
who do not respond well. For this task, we used data from a
1
Department of Periodontology, County Hospital of Gävle; Center for
Research and Development, Uppsala University/Region of Gävleborg,
Gävle, Sweden
2
Department of Medical Sciences, Uppsala University, Uppsala, Sweden
Corresponding Author:
A. Holmlund, Department of Periodontology, Gävle County Hospital,
801 87 Gävle, Sweden.
Email: anders.holmlund@regiongavleborg.se
2 Journal of Dental Research 
specialized clinic for periodontology with >5,000 patients fol-
lowed for more than a decade with no known CVD at baseline.
Methods
Study Population
The cohort consisted of 8,999 subjects aged 20 to 85 y who
were referred to the specialized clinic for periodontal treatment
at Gävle County Hospital between 1979 and 2012. Before
examination, all participants filled in a questionnaire, includ-
ing questions regarding their medical health, medication, and
smoking habits. None of the subjects were edentulous.
Included in this study were individuals for whom data on 1-y
follow-up after active treatment were available before the cen-
sor date. Excluded were edentulous people and those with CVD
at baseline examination (according to questionnaire and/or data
from Swedish hospital discharge register) or no available data
on treatment outcome 1 y after conclusion of active treatment.
After exclusion, data from 5,297 subjects were available.
Exposure and Outcomes Definition
Oral Examination and Definition.  All participants received a
dental examination, with assessment of number of teeth (maxi-
mum, n = 32), presence or absence of dental plaque, and bleed-
ing on probing (BOP). Bleeding was recorded 20 s after
probing the pocket depth at 4 sites per tooth.
Additionally, involvement of furcations, tooth mobility, and
the number of deepened pockets for all present teeth were
recorded. The probing pocket depth (PPD) was measured par-
allel to the long axis of the tooth and rounded up to the nearest
millimeter. Probing was done at 6 sites around the tooth with a
UNC 12 manual probe (Hu-Friedy; mesiobuccal, midbuccal,
distobuccal, mesiolingual, midlingual, and distolingual). In
some sites probing could be done along the proximal area, but
for each side of the tooth, only the deepest site was recorded;
therefore, only 4 sites were registered in the patient’s record.
They were considered to be deepened if PPD were >4 mm. At
follow-up 1 y after treatment, plaque index (%), BOP (%), and
pocket depth around all remaining teeth were recorded.
Two groups were defined: responders and poor responders.
Poor responders were defined as having >10% PPD >4 mm
and ≥20% sites with BOP 1 y after active treatment was
concluded.
Periodontal Treatment. All patients received a thorough case
presentation and instruction in oral hygiene before treatment.
After this, nonsurgical treatment consisting of scaling of the
root surfaces with ultrasound and hand instruments was per-
formed, usually 1 quadrant at a time, 1 wk apart. Reevaluation
was performed after 6 to 8 wk, and when regarded as neces-
sary, residual pockets were rescaled or periodontal surgery was
performed. After active treatment, all patients were put on a
maintenance program performed by a dental hygienist, first
every 2 wk, up to 3 mo; then, at 6, 9, and 12 mo (1-y
follow-up), the outcome of the treatment was evaluated. When
needed, the patients were given repeated instruction on oral
hygiene during the recall program.
Active treatment was completed after the last operation if
periodontal surgery was performed. When the treatment was
nonsurgical, active treatment was completed when the special-
ist judged that further treatment would not add anything. The
treatment goal at the department has been no deep pockets
around 1-rooted teeth. Few residual pockets in the molar region
can be accepted; however, the ambition is that they should not
bleed on probing. As none of the groups were constituted
before the study was initiated, both groups were treated accord-
ing to the same concept.
Medical End Points
The rational for using MI, stroke, and HF in a combined end
point in the present study is that all 3 diseases are associated
with inflammation, a hallmark of PD and a possible link
between PD and those 3 CVDs. Combining fatal and nonfatal
MI, stroke, and HF into a combined end point is commonly
used (Arnlov et al. 2010).
The unique Swedish personal number was merged against
the Swedish cause-of-death register and the National Hospital
Discharge Register to obtain cause of death and hospital treat-
ment of CVD via International Classification of Disease (ICD)
codes. As the patients were enrolled between 1979 and 2013,
ICD-8, ICD-9, and ICD-10 were used, as the codes have
changed over time. Based on ICD-8 and ICD-9, codes 390-459
denoted CVDs. Based on the ICD-10 classification, the corre-
sponding codes were I00-I99. There was no loss at follow-up.
Socioeconomic Factors
The educational level of the participants was not initially
recorded, but as educational level could be a confounder for
PD (Borrell and Crawford 2012), data regarding it was obtained
from the Swedish national register for education. Educational
levels were categorized into 3 categories: 1, ≤9 y; 2, 10 to 12 y;
3, university studies.
Ethical Approval
The Ethics Committee of Uppsala University approved the
study (DNR 2009/259). This article was structured according
to the STROBE (Strengthening the Reporting of Observational
Studies in Epidemiology) guidelines for cohort studies.
Statistical Analysis
Follow-up time for each subject was split into small intervals
(1 y) to accommodate the effects of aging and calendar time.
Rates were assumed constant within each year and calendar
year, rounded to the nearest integer values. The effects of age
and calendar year were modeled through restricted cubic splines,
with 3 knots placed at the 10th, 50th, and 90th percentiles of
Periodontal Treatment and Future Cardiovascular Disease 3

Table 1.  Baseline Characteristics for the Total Study Population and by Periodontal Therapy Response Group (n = 5,297).

Good Responders Poor Responders

Total Sample (N = 5,297) (n = 4,561, 86.2%) (n = 732, 13.8%) P Value

Male:female, % 41:59 42:58 51:49 <0.001

Age, mean ± SD 50.5 ± 11.5 50.8 ± 11.6 48.6 ± 10.6 <0.001
Smoking, % 46 38.2 62 <0.001
No. of teeth, mean ± SD 24.4 ± 4.6 24.4 ± 4.6 24.0 ± 4.8 0.26
PLI, %a 35.6 ± 29.4 33.3 ± 28.5 49.9 ± 31 <0.001
BOP, %b 40.8 ± 29.2 37.8 ± 28 59.5 ± 29.8 <0.001
PPD >4 mm  
  Mean ± SD 24 ± 17 22 ± 16 37 ± 18 <0.001
 % 25 ± 18 23 ± 17 39 ± 18 <0.001
Education level, mean ± SD 1.8 ± 0.73 1.83 ± 0.73 1.71 ± 0.69 <0.001
CVD cases, n 870 697 173  
Person-years 89,719.937 75,707.19 14,012.747  
IR/1,000 person years 9.70 9.21 12.35  

BOP, bleeding on probing; CVD, cardiovascular disease; IR, incidence rate; PLI, Plaque Index; PPD, probing pocket depth.
a
Percentage of surfaces with plaque.
b
Percentage of surfaces with bleeding on probing.
c
Education level was divided into 3 levels: 1, ≤9 y; 2, 10 to 12 y; 3, university studies.

the distribution of age and calendar year, conditioned on an event.
Incidence rate ratios (IRRs) were then obtained via Poisson
regression with the logarithm of person-years used as an offset
(Carstensen 2005). Apart from treatment response/nonresponse,
age, sex, smoking, education, and baseline values for BOP, per-
centage of PPD >4 mm, and number of teeth were included in the
analyses. All analyses were conducted with STATA/SE 14.0
(StataCorp LP). P = 0.05 was regarded as significant.
Results
This study had a median follow-up time of 16.8 y, correspond-
ing to 89,719 person-years at risk, during which 870 incident
cases of fatal/nonfatal CVD occurred.
In total, 13.8% of the population was classified as poor
responders. The mean ages of good responders and poor
responders were 50.8 ± 11.6 and 48.6 ± 10 y, respectively. At
baseline, the poor responder group had significantly more PPD
>4 mm, BOP, frequent smoking, and a lower educational level
than good responders. Basic characteristics for responders and
poor responders are given in Table 1.
No major differences in oral health parameters were found
between those who were included in this study and the remain-
ing persons in the database. Compared with the total sample (N =
8,999), the subsample used in the present study showed a
higher proportion of PPD >4 mm (25 ± 18 vs. 20 ± 18) and a
higher proportion of BOP (41 ± 29 vs. 37 ± 29) but a similar
number of teeth (both 24 ± 5).
A larger proportion of poor responders showed an incidence
of CVD as compared with the good responders (23.6% vs.
15.3%, P < 0.001). After adjustment for age, sex, smoking,
education, and calendar time, as well as for baseline values for
BOP, percentage of PPD >4 mm, and number of teeth, poor
responders had an IRR for CVD of 1.28 (95% CI, 1.07 to 1.53;
P = 0.007) when compared with good responders. No interaction
was observed between smoking and the response to treatment
(P = 0.80). In a sensitivity analysis regarding smoking, we
evaluated the response to treatment separately for smokers and
nonsmokers, and the IRR for poor responding was 1.23 (95%
CI, 0.94 to 1.63) among nonsmokers and 1.30 (95% CI, 1.03 to
1.65) among smokers.
When only the 1-y follow-up data for “percentage PPD
>4 mm” and “BOP” were included in the model instead of
good/poor responder, neither of these 2 variables was signifi-
cantly related to future CVD (for percentage PPD >4 mm: IRR,
1.0063; 95% CI, 0.99 to 1.012; P = 0.059; for BOP: IRR,
1.0026; 95% CI, 0.99 to 1.0062; P = 0.15).
From baseline to 1-y follow-up (including the treatment
phase), the mean of percentage PPD >4 mm changed from 26.0
to 9.8 and the mean of BOP, from 44 to 17. When the changes
in percentage PPD >4 mm and BOP were used in the model,
instead of good/poor responder, the change in percentage PPD
>4 mm, but not the change in BOP, was significantly related to
future CVD (for percentage PPD >4 mm: IRR, 0.995; 95% CI,
0.991 to 0.999; P = 0.023; for BOP: IRR, 0.999; 95% CI, 0.997
to 1.0015; P = 0.55).
When the model was run without including the initial val-
ues for percentage PPD >4 mm, BOP, and number of teeth, the
IRR for being a poor responder was 1.33 (95% CI, 1.12 to
1.58; P = 0.001).
Because the number of teeth could influence the degree of
systemic inflammation associated with PD, the analysis was
performed separately in those with >10, >15, and >20 remain-
ing teeth. IRR for CVD for the poor responders increased with
the number of teeth as compared with good responders. There
was an increase in IRR, from 1.28 (95% CI, 1.07 to 1.53) to
1.39 (95% CI, 1.13 to 1.73), between those having >0 teeth and
those with >20 teeth (Table 2).
An interaction term between the number of remaining teeth
and treatment response was included in the regression model.
The Figure presents the log IRRs for poor versus good response
as a function of the number of remaining teeth. The risk for
4 Journal of Dental Research 
Table 2.  IRR for Fatal/Nonfatal CVD by Groups of Teeth Remaining in
Poor Responders (n = 732) versus Good Responders (n = 4,565).
IRR P Value 95% CI
Total 1.28 0.007 1.07 to 1.53
>10 teeth 1.29 0.005 1.08 to 1.55
>15 teeth 1.32 0.004 1.09 to 1.60
>20 teeth 1.39 0.002 1.13 to 1.73
CVD, cardiovascular disease; IRR, incidence rate ratio.
Adjusted for calendar time, age, sex, education level, smoking, and
baseline values for bleeding on probing and percentage of probing
pocket depth >4 mm.
CVD was clearly elevated for poor responders, as the number
of remaining teeth increased (P = 0.024), further showing the
dependence of the number of teeth in poor responders.
In contrast, when the number of teeth in the good responder
group was divided into quintiles, those subjects with few
remaining teeth were at higher risk for CVD when compared
with those who had the most remaining teeth (odds ratio, 1.08;
95% CI, 1.00 to 1.17).
Discussion
In this study, those who did not respond well to periodontal
treatment had a higher risk of future CVD when compared with
those who did respond well.
It is well known that PD is associated with CVD, but
whether successful periodontal treatment would influence the
risk for future CVD had not yet been studied. The PAVE study
aimed to investigate if periodontal treatment would influence
the risk of new cardiovascular events (Beck et al. 2008).
Participants were randomized to receive periodontal treatment
provided by the study or were referred to their ordinary dentist,
but no difference was found between the groups regarding the
risk for new CVD events. An obvious problem in that study
was that 48% of the subjects in the group who were referred to
their ordinary dentist also received periodontal treatment
(Offenbacher et al. 2009). As it is not possible for ethical rea-
sons to perform a randomized clinical trial among PD patients
in which 1 group is left without PD treatment, we are left to
find other ways to explore whether the relationship between
PD and CVD is causal. This study is 1 example of how to study
the link between PD and CVD, although it does not provide a
definite answer regarding causality.
The number of deepened pockets measured with a probe
and the loss of attachment—either clinically measured with a
probe or assessed on X-ray as bone loss around the teeth—are
usually used to grade the severity or presence of PD. In a num-
ber of studies concerning oral health and CVD, tooth loss was
used as a surrogate marker for oral health (Abnet et al. 2005;
Holmlund and Lind 2012; Vedin et al. 2015). Loss of attach-
ment and tooth loss primarily reflect what has happened in the
past. To clinically assess ongoing inflammation in the peri-
odontal tissue, evaluation for the presence of BOP or pus is
preferred in the clinical setting. Persisting deep pockets and
BOP have been associated with progression of PD (Matuliene
Figure.  Log rate ratio for poor versus good responders according
to number of teeth. Predictive margins were used for the calculations.
Whiskers indicate 95% CI.
et al. 2008). It is therefore essential to achieve pocket reduction
and BOP reduction during PD treatment; therefore, these 2
characteristics were chosen to define response to treatment.
In this study, being a good/poor responder was defined
according to data on percentage PPD >4 mm and BOP at 1-y
follow-up after treatment. Further analysis of the impact of the
change in these 2 variables during the treatment phase dis-
closed that it was mainly the improvement in percentage PPD
>4 mm that was related to future CVD. The improvement in
BOP was less important in this respect.
The biological rationale behind the link between PD and
future CVD is not clear. One suggested explanation is that
locally released inflammatory mediators in the tooth’s sur-
rounding tissues caused by PD could enter systemic circulation
and mediate a systemic low-grade inflammation, thus possibly
affecting the development of atherosclerosis. We know from
previous studies that those with PD have higher levels of sys-
temic inflammatory mediators, such as white blood cell count,
fibrinogen, and C-reactive protein (Loos et al. 2000; Wu,
Trevisan, Genco, Falkner, et al. 2000), and that treatment of
periodontitis could reduce these levels (Bokhari et al. 2012;
Caúla et al. 2014). These studies indicate that periodontal treat-
ment could influence systemic inflammatory mediators and
thereby possibly affect the atherosclerotic process.
What further strengthens the hypothesis in this study—that
reduction of inflammation in the periodontal tissues might be
involved with the risk for CVD—is that poor responders with
more teeth were at greater risk than those with less teeth. This
may sound odd because tooth loss has been postulated to be a
possible marker for cumulative exposure to oral inflammation
and because previous studies have shown increased risk for fatal
and nonfatal CVD in those with few remaining teeth (Holmlund
et al. 2010; Liljestrand et al. 2015; Vedin et al. 2015). In accor-
dance with these studies, having a few teeth was associated, in
this study, with an increased risk for CVD among the good
responders to treatment. However, a possible explanation to
Periodontal Treatment and Future Cardiovascular Disease 5
these opposing results could be that, among the poor respond-
ers, an increasing number of teeth leave more sites where
inflammation in the periodontal tissues could persist, releasing
more inflammatory mediators that could enter the bloodstream
and enhance the atherosclerotic process.
Another theory is that bacteria considered to be involved in
the pathogenesis of PD enter the bloodstream and penetrate the
endothelium of the vessel wall. Inside the endothelium, those
bacteria increase the inflammatory process in the vessel wall,
leading to recruitment of immune-competent cells, an increase
in foam cell development, migration of smooth muscles cells,
and thinning of the fibrous cap; the inflammatory process
would thereby increase the risk for rupture of the fibrous cap,
leading to thrombus formation causing myocardial or brain infarc-
tion. Bacteria, such as Porphyromonas gingivalis, Aggregobacter
actinomycetemcomitans, and Fusobacterium nucleatum, have
been extracted from atherosclerotic plaques of humans (Dorn
et al. 1999). A possibility is that these bacteria are only passive
passengers inside monocytes that enter atherosclerotic plaques,
but this hypothesis is contradicted by the fact that viable
P. gingivalis has been cultured from atherosclerotic plaques
(Kozarov et al. 2005). It is reasonable to assume that the reduc-
tion of pocket depth and BOP among responders in this study
also reduces the risk for the aforementioned bacteria to enter
systemic circulation.
There is also the possibility that the poor responders to peri-
odontal treatment in this study had a more proinflammatory
constitution, which could be the reason for the increased risk
for CVD and an explanation why they did not respond so well
to the treatment. High doses of atorvastatin have been shown to
reduce not only the risk for MI but also periodontal inflamma-
tion (Subramanian et al. 2013), and it is likely that a combination
between a more proinflammatory constitution and increased
inflammatory burden from periodontal inflammation lies
behind the association between PD and CVD. We see that the
poor responders at baseline had more severe PD versus the
good responders, but this difference was adjusted in the statis-
tical analyses by including the baseline values of BOP, deep-
ened pockets, and number of teeth as possible confounders in
the models.
Strength and Limitations
The strength of this study is that it contains a large number of
subjects who received a full-mouth investigation, and all were
treated at a specialized clinic for periodontology with a substan-
tial follow-up time. Furthermore, the Swedish registers have a
high quality and are regarded to be reliable. However, the study
also has some limitations. For instance, the design is prospective
in the sense that a baseline examination was performed and,
thereafter, patients were followed over a time regarding CVD.
Yet, the cohort was not initially recruited for that purpose; there-
fore, we do not know what periodontal treatment the participants
received between the 1-y follow-ups and the censor date.
For obvious reasons, the examiners were not calibrated over
a period of 33 y, but all examiners are experienced and used the
same examination methodology. Any possible differences
among the examiners would therefore be represented in both
groups, and any potential differences among examiners would
likely drive the results toward null. Unfortunately, the partici-
pants were not medically examined at baseline; as such, we
could not adjust for all known cardiovascular risk factors, such
as hypertension, hypercholesterolemia, and diabetes, because
this information was lacking. The sample was also limited to
mainly Caucasians.
Conclusion
Individuals who did not respond well to periodontal treatment
had an increased risk for future CVD, indicating that success-
ful periodontal treatment might influence progression of sub-
clinical CVD.
Author Contributions
A. Holmlund, contributed to design, data acquisition, analysis, and
interpretation, drafted the manuscript; E. Lampa, contributed to
the data analysis and interpretation, critically revised the manu-
script; L. Lind, contributed to design, data analysis, and interpreta-
tion, critically revised the manuscript. All authors gave final
approval and agree to be accountable for all aspects of the work.
Acknowledgments
A special thanks to Dr. Gunnar Holm, who initiated the database
used in the present study. The study was supported by grants from
the Centre for Research and Development, Uppsala University/
Region Gävleborg, Sweden, and Public Dental Health Region
Gävleborg. The funders had no influence of the design and con-
duct of the study. The authors declare no potential conflicts of
interest with respect to the authorship and/or publication of this
article.
References
Abnet CC, Qiao YL, Dawsey SM, Dong ZW, Taylor PR, Mark SD. 2005. Tooth
loss is associated with increased risk of total death and death from upper
gastrointestinal cancer, heart disease, and stroke in a Chinese population-
based cohort. Int J Epidemiol. 34(2):467–474.
Arnlov J, Ingelsson E, Sundström J, Lind L. 2010. Impact of body mass index
and the metabolic syndrome on the risk of cardiovascular disease and death
in middle-aged men. Circulation. 121(2):230–236.
Beck JD, Couper DJ, Falkner KL, Graham SP, Grossi SG, Gunsolley JC,
Madden T, Maupome G, Offenbacher S, Stewart DD, et al. 2008. The
Periodontitis and Vascular Events (PAVE) pilot study: adverse events.
J Periodontol. 79(1):90–96.
Bokhari SA, Khan AA, Butt AK, Azhar M, Hanif M, Izhar M, Tatakis DN.
2012. Non-surgical periodontal therapy reduces coronary heart disease risk
markers: a randomized controlled trial. J Clin Periodontol. 39(11):1065–
1074.
Borrell LN, Crawford ND. 2012. Socioeconomic position indicators and peri-
odontitis: examining the evidence. Periodontol 2000. 58(1):69–83.
Carstensen B. 2005. Demography and epidemiology: practical use of the Lexis
diagram in the computer age. Or: who needs the Cox-model anyway?
Technical report 06.2. Copenhagen (Denmark): Department of Biostatistics,
University of Copenhagen [accessed 14 Mar 2017]. http://publichealth
.ku.dk/sections/biostatistics/reports/2006/rr-06-2.pdf.
Caúla AL, Lira-Junior R, Tinoco EM, Fischer RG. 2014. The effect of peri-
odontal therapy on cardiovascular risk markers: a 6-month randomized
clinical trial. J Clin Periodontol. 41(9):875–882.
D’Aiuto F, Parkar M, Andreou G, Suvan J, Brett PM, Ready D, Tonetti MS.
2004. Periodontitis and systemic inflammation: control of the local infection
6 Journal of Dental Research 
is associated with a reduction in serum inflammatory markers. J Dent Res.
83(2):156–160.
Desvarieux M, Demmer RT, Rundek T, Boden-Albala B, Jacobs DR
Jr, Papapanou PN, Sacco RL; Oral Infections and Vascular Disease
Epidemiology Study (INVEST). 2003. Relationship between periodon-
tal disease, tooth loss, and carotid artery plaque: the Oral Infections and
Vascular Disease Epidemiology Study (INVEST). Stroke. 34(9):2120–2125.
Dhingra R, Vasan RS. 2012. Diabetes and the risk of heart failure. Heart Fail
Clin. 8(1):125–133.
Dorn BR, Dunn WA Jr, Progulske-Fox A. 1999. Invasion of human coronary
artery cells by periodontal pathogens. Infect Immun. 67(11):5792–5798.
Emdin CA, Anderson SG, Salimi-Khorshidi G, Woodward M, MacMahon S,
Dwyer T, Rahimi K. 2016. Usual blood pressure, atrial fibrillation and vas-
cular risk: evidence from 4.3 million adults. Int J Epidemiol [epub ahead of
print 3 Mar 2016] in press. doi:10.1093/ije/dyw053
Gotsman I, Lotan C, Soskolne WA, Rassovsky S, Pugatsch T, Lapidus L,
Novikov Y, Masrawa S, Stabholz A. 2007. Periodontal destruction is asso-
ciated with coronary artery disease and periodontal infection with acute
coronary syndrome. J Periodontol. 78(5):849–858.
Grau AJ, Becher H, Ziegler CM, Lichy C, Buggle F, Kaiser C, Lutz R,
Bultmann S, Preusch M, Dörfer CE. 2004. Periodontal disease as a risk
factor for ischemic stroke. Stroke. 35(2):496–501.
Holmlund A, Holm G, Lind L. 2010. Number of teeth as a predictor of car-
diovascular mortality in a cohort of 7,674 subjects followed for 12 years.
J Periodontol. 81(6):870–876.
Holmlund A, Lind L. 2012. Number of teeth is related to atherosclerotic plaque
in the carotid arteries in an elderly population. J Periodontol. 83(3):287–291.
Humphrey LL, Fu R, Buckley DI, Freeman M, Helfand M. 2008. Periodontal
disease and coronary heart disease incidence: a systematic review and meta-
analysis. J Gen Intern Med. 23(12):2079–2086.
Janket SJ, Baird AE, Chuang SK, Jones JA. 2003. Meta-analysis of periodontal
disease and risk of coronary heart disease and stroke. Oral Surg Oral Med
Oral Pathol Oral Radiol Endod. 95(5):559–569.
Joshipura KJ, Hung HC, Rimm EB, Willett WC, Ascherio A. 2003. Periodontal
disease, tooth loss, and incidence of ischemic stroke. Stroke. 34(1):47–52.
Kalogeropoulos AP, Georgiopoulou VV, Butler J. 2012. From risk factors to struc-
tural heart disease: the role of inflammation. Heart Fail Clin. 8(1):113–123.
Kozarov EV, Dorn BR, Shelburne CE, Dunn WA, Progulske-Fox A. 2005.
Human atherosclerotic plaque contains viable invasive Actinobacillus acti-
nomycetemcomitans and Porphyromonas gingivalis. Arterioscler Thromb
Vasc Biol. 25(3):e17–e18.
Liljestrand JM, Havulinna AS, Paju S, Mannisto S, Salomaa V, Pussinen PJ.
2015. Missing teeth predict incident cardiovascular events, diabetes, and
death. J Dent Res. 94(8):1055–1062.
Loos BG, Craandijk J, Hoek FJ, Wertheim-van Dillen PM, van der Velden U.
2000. Elevation of systemic markers related to cardiovascular diseases in the
peripheral blood of periodontitis patients. J Periodontol. 71(10):1528–1534.
Matuliene G, Pjetursson BE, Salvi GE, Schmidlin K, Bragger U, Zwahlen M,
Lang NP. 2008. Influence of residual pockets on progression of periodonti-
tis and tooth loss: results after 11 years of maintenance. J Clin Periodontol.
35(8):685–695.
Offenbacher S, Beck JD, Moss K, Mendoza L, Paquette DW, Barrow DA,
Couper DJ, Stewart DD, Falkner KL, Graham SP, et al. 2009. Results from
the Periodontitis and Vascular Events (PAVE) study: a pilot multicentered,
randomized, controlled trial to study effects of periodontal therapy in a
secondary prevention model of cardiovascular disease. J Periodontol.
80(2):190–201.
Subramanian S, Emami H, Vucic E, Singh P, Vijayakumar J, Fifer KM, Alon
A, Shankar SS, Farkouh M, Rudd JH, et al. 2013. High-dose atorvastatin
reduces periodontal inflammation: a novel pleiotropic effect of statins.
J Am Coll Cardiol. 62(25):2382–2391.
Taylor BA, Tofler GH, Carey HM, Morel-Kopp MC, Philcox S, Carter TR,
Elliott MJ, Kull AD, Ward C, Schenck K. 2006. Full-mouth tooth extrac-
tion lowers systemic inflammatory and thrombotic markers of cardiovascu-
lar risk. J Dent Res. 85(1):74–78.
Tonetti MS, D’Aiuto F, Nibali L, Donald A, Storry C, Parkar M, Suvan J,
Hingorani AD, Vallance P, Deanfield J. 2007. Treatment of periodontitis
and endothelial function. N Engl J Med. 356(9):911–920.
Vedin O, Hagström E, Gallup D, Neely ML, Stewart R, Koenig W, Budaj
A, Sritara P, Wallentin L, White HD, et al. 2015. Periodontal disease in
patients with chronic coronary heart disease: prevalence and association
with cardiovascular risk factors. Eur J Prev Cardiol. 22(6):771–778.
Wrigley BJ, Lip GY, Shantsila E. 2011. The role of monocytes and inflammation
in the pathophysiology of heart failure. Eur J Heart Fail. 13(11):1161–1171.
Wu T, Trevisan M, Genco RJ, Dorn JP, Falkner KL, Sempos CT. 2000.
Periodontal disease and risk of cerebrovascular disease: the first national
health and nutrition examination survey and its follow-up study. Arch
Intern Med. 160(18):2749–2755.
Wu T, Trevisan M, Genco RJ, Falkner KL, Dorn JP, Sempos CT. 2000.
Examination of the relation between periodontal health status and cardio-
vascular risk factors: serum total and high density lipoprotein cholesterol,
C-reactive protein, and plasma fibrinogen. Am J Epidemiol. 151(3):273–282.