Annals of Oncology original articles

13. Gavaghan DJ, Moore RA, McQuay HJ. An evaluation of homogeneity tests in meta- 17. Applebaum KM, Malloy EJ, Eisen EA. Reducing healthy worker survivor bias by
analyses in pain using simulations of individual patient data. Pain 2000; 85(3): restricting date of hire in a cohort study of Vermont granite workers. Occup Environ
415–424. Med 2007; 64(10): 681–687.
14. Checkoway H, Pearce N, Kriebel D. Cohort studies. Research Methods in 18. Wen CP, Tsai SP. Anatomy of the health worker effect—a critique of summary
Occupational Epidemiology, 2nd edition. New York, USA: Oxford University Press statistics employed in occupational epidemiology. Scand J Work Environ Health
2004; 123–178. 1982; 8(Suppl 1): 48–52.
15. Fox AJ, Collier PF. Low mortality rates in industrial cohort studies due to selection 19. De Vocht F, Straif K, Szeszenia-Dabrowska N et al. A database of exposures in the
for work and survival in the industry. Br J Prev Soc Med 1976; 30(4): 225–230. rubber manufacturing industry: design and quality control. Ann Occup Hyg 2005;
16. Goldblatt P, Fox J, Leon D. Mortality of employed men and women. Am J Ind Med 49(8): 691–701.
1991; 20(3): 285–306.

Annals of Oncology 27: 941–947, 2016

doi:10.1093/annonc/mdw028
Published online 24 January 2016

Periodontal disease and risk of all cancers among male

never smokers: an updated analysis of the Health
Professionals Follow-up Study
D. S. Michaud1,2*, K. T. Kelsey2,3, E. Papathanasiou4, C. A. Genco5 & E. Giovannucci6,7,8
1
Department of Public Health and Community Medicine, Tufts University School of Medicine, Boston; Departments of 2Epidemiology; 3Pathology and Laboratory
Medicine, Brown University School of Medicine, Providence; 4Department of Periodontology, Tufts University School of Dental Medicine, Boston; 5Department of
Integrative Physiology and Pathobiology, Tufts University School of Medicine, Boston; Departments of 6Nutrition; 7Epidemiology, Harvard School of Public Health, Boston;
8
Channing Division of Network Medicine, Department of Medicine, Harvard Medical School and Brigham and Women’s Hospital, Boston, USA

Received 22 October 2015; revised 12 January 2016; accepted 13 January 2016

Background: Periodontal disease has a direct impact on the immune response and has been linked to several chronic
diseases, including atherosclerosis and stroke. Few studies have examined the association between periodontal disease
and cancer.
Patients and methods: A total of 19 933 men reported being never smokers (of cigarette, pipes or cigars) in the Health
Professionals’ Follow-up Study. Periodontal disease status and teeth number were self-reported at baseline and during
follow-up. All cancers were ascertained during 26 years of follow-up. Cox’s proportional hazard models were used to
estimate hazard ratios (HRs) and 95% confidence intervals (95% CIs) adjusting for risk factors.
Results: A 13% increase in total cancer was observed among men reporting periodontitis at baseline, and a 45%
increase in risk was observed among men with advanced periodontitis ( periodontitis with <17 remaining teeth).
Periodontitis was not associated with prostate cancer, colorectal cancer or melanoma, the three most common cancers
in this cohort of never smokers, but a 33% increase in risk was observed for smoking-related cancers (lung, bladder,
oropharnygeal, esophageal, kidney, stomach and liver cancers; HR = 1.33, 95% CI 1.07–1.65). Men with advanced
periodontitis had an HR of 2.57 (95% CI 1.56–4.21; P = 0.0002) for smoking-related cancers, compared with men who
did not have periodontitis and had 17 teeth or more. Advanced periodontitis was associated with elevated risks of
esophageal and head and neck cancers (HR = 6.29, 95% CI 2.13–18.6; based on five cases with advanced periodontitis)
and bladder cancer (HR = 5.06, 95% CI 2.32–11.0; based on nine cases with advanced periodontitis).
Conclusions: Advanced periodontitis was associated with a 2.5-fold increase in smoking-related cancers among never
smokers. Periodontitis may impact cancer risk through system immune dysregulation. Further studies need to examine
the immune impact of advanced periodontitis on cancer, especially for cancers known to be caused by smoking.
Key words: periodontal disease, cancer risk, never smokers

*Correspondence to: Prof. Dominique S. Michaud, Department of Public Health and

Community Medicine, Tufts University, 136 Harrison Avenue, Boston, MA 02111, USA.
Tel: +1-617-636-0482; Fax: +1-617-636-4017; E-mail: dominique.michaud@tufts.edu

© The Author 2016. Published by Oxford University Press on behalf of the European Society for Medical Oncology.
All rights reserved. For permissions, please email: journals.permissions@oup.com.
original articles Annals of Oncology

introduction 19 933 participants. The follow-up rate for the HPFS cohort is >96%. This
study was approved by the Human Subjects Committee of the Harvard School
Many of the known risk factors for cancer, including smoking, of Public Health.
obesity, diabetes and physical inactivity, share some commonal-
ity in that they are all associated with systemic inflammation
and immune perturbations, and likely share common pathways assessment of dental measures
to cancer development. Advanced periodontal disease, through periodontal disease. At baseline, participants of the HPFS were asked
local inflammation, bone destruction and tooth loss, can have a whether they had a history of periodontal disease with bone loss. This
major systemic impact on the body, which extends beyond the question was validated in dentists [14] and non-dentists [15] in the HPFS
cohort by obtaining radio graphs from subsets of individuals with and
oral cavity [1]. Recent advances in identifying and quantifying
without a self-reported history of periodontal disease (for additional details,
bacteria using new sequencing methodology (e.g. metagenomics)
refer to supplementary Material, available at Annals of Oncology online). In
[2] have provided new opportunities to examine the role of
addition, participants updated their periodontal disease status on each
microbiota in cancer development [3, 4]. Poor oral health and de-
biennial questionnaire during follow-up.
velopment of periodontitis are associated with dysbiosis [5], i.e.
changes in bacterial communities in the mouth, and changes in
the immune status [6], which may well be directly involved in the tooth loss. Participants reported a number of natural teeth at baseline,
genesis of cancer. and any tooth loss during the previous 2 years was reported on the follow-up
questionnaires. Self-reported number of teeth is highly correlated with the
In a previous analysis, we reported positive associations between
actual number of teeth on clinical assessment in a general population
periodontal disease and cancer risk in the Health Professionals
(r = 0.97) [16].
Follow-up Study (HPFS) [7]. Three other prospective cohort
studies with data on periodontal disease, the NHANES I study [8],
the NHANES III study [9] and the Women’s Health Initiative identification of participants with cancer
(WHI) [10], have reported positive associations for periodontitis Medical records were obtained from physicians/hospitals for newly diagnosed
and cancers, including pancreatic, lung and orodigestive cancers. cancers as reported by participants on biennial questionnaires (supplementary
In addition, antibodies to a periodontal pathogen (Porphyromonas Material, available at Annals of Oncology online, for more details). Smoking-
gingivalis) have been associated with elevated risk of pancreatic related cancers are defined as those cancers linked to smoking by IARC [17].
and orodigestive cancers in two different prospective studies [9,
11]; to date, there are no null studies on this topic. Retrospective
statistical analysis
studies have examined these associations, especially for head and
We computed person-time of follow-up for each participant from the return
neck cancers [12, 13], but as these are prone to bias, and reverse
date of the baseline questionnaire to the date of cancer diagnosis, death from
causation, they cannot provide strong support for causality. In the
any cause or the end of follow-up (31 January 2012), whichever came first. For
previous HPFS analysis [7], and similarly in the WHI [10], peri-
the periodontal disease analyses, we examined baseline periodontal disease
odontal disease associations with lung cancer were not elevated in
status and updated periodontal disease status to represent chronic and more
never smokers, but did remain elevated for all smoking-related recent periodontal disease development, respectively. We estimated hazard
cancers in the HPFS (not reported in WHI). ratios (HRs) and 95% confidence intervals (CIs) by the use of Cox proportion-
With an additional 8 years of follow-up available in the HPFS al hazard models adjusting for potential confounders. Additional details on
cohort (since our last analysis), and substantially more cancer confounders are provided in supplementary Material, available at Annals of
cases among never smokers only, we herein re-examine the as- Oncology online.
sociation between periodontal disease, tooth number and cancer
risk with a focus on never smokers. This analysis also includes a
new approach to examining periodontal disease; we combined results
self-reported history of periodontal disease with a number of At baseline (1986), the prevalence of periodontitis among never
teeth lost to better capture severity of periodontitis, given that smokers in this population was 9.7% which is slightly lower than
a great number of teeth lost would indicate more advanced the overall HPFS population (16% [7]), as expected, given that
periodontitis smoking is strong risk factor for periodontitis. In this highly edu-
cated population of health professionals, substantial tooth loss
methods was uncommon (fewer than 17 teeth remaining—due to prede-
fined categories on the questionnaire, we could not use 20 teeth
The HPFS is an ongoing prospective cohort study that was initiated in 1986
and is maintained by investigators in the Department of Nutrition, Harvard
remaining as the cut off; 3% never smokers, 5.5% overall popula-
School of Public Health, Boston, MA. Participants in this cohort were recruited
tion). Men with fewer teeth were older and had slightly higher
from all US states by mail questionnaires and were eligible if they were regis- BMI, and those with periodontal disease were more likely to have
tered dentists, veterinarians, pharmacists, optometrists, osteopathic physicians diabetes (Table 1). Alcohol and total caloric intake was similar
or podiatrists, between ages 40 and 75 years and male; no exclusions were across exposure groups (Table 1). Men with advanced periodon-
made in the original cohort. Data on behavioral, medical and lifestyle factors titis (defined as having fewer than 17 teeth and reported peri-
were obtained at baseline and subsequently during follow-up through mailed odontal disease) were more likely to have hypertension (29.7%),
questionnaires (for more details, refer to supplementary Material, available at were more likely to be non-steroidal anti-inflammatory drug
Annals of Oncology online). The current analysis is restricted to men who users (38.9%) and a higher percent were non-White (20.8%),
reported being never cigarette, cigar or pipe smokers at baseline and consists of compared with men in the other categories (Table 1).

 | Michaud et al. Volume 27 | No. 5 | May 2016

Annals of Oncology original articles
Table 1. Age-standardized baseline characteristics (mean or %) of participants in the Health Professionals Follow-up Study by periodontal disease and
number of teeth among never cigarette, pipe or cigar smokers (n = 19 933)a
No periodontal disease No periodontal Periodontal disease Periodontal disease
and 17–32 teeth disease and <17 teeth and 17–32 teeth and <17 teeth

Participants, n 17 554 434 1768 177

Age, yearsb 52.2 (9.56) 61.3 (9.20) 56.5 (9.59) 64.8 (8.06)
Height, inb 70.1 (3.23) 70.2 (2.89) 69.9 (4.05) 70.1 (3.01)
Body mass index, kg/m2b 25.2 (4.81) 26.0 (5.42) 25.3 (5.20) 25.9 (4.48)
Physical activity, MET h/weekb 22.1 (30.1) 17.6 (20.9) 21.6 (28.2) 23.8 (34.6)
Ethnic origin, %
White 94.4 89.3 90.6 79.2
Otherc 5.6 10.7 9.4 20.8
Diabetes (%) 2.1 2.0 3.8 5.6
High cholesterol (%) 10.9 7.5 13.0 8.5
High blood pressure (%) 18.8 20.4 21.0 29.7
NSAID (%) 32.3 35.5 31.0 38.9
Multivitamin (%) 41.3 46.8 37.9 29.4
Daily dietary intakeb
Calories, kcal 1997 (617) 2089 (692) 1987 (642) 2132 (666)
Vitamin D, IU 409 (308) 408 (320) 413 (336) 379 (290)
Alcohol, g 7.57 (11.6) 8.24 (11.8) 7.36 (11.6) 7.06 (12.6)

a
All variables (except age) are age standardized.
b
Mean (SD) for all such values.
c
All non-White categories combined due to small numbers.
MET, metabolic equivalent; NSAID, non-steroidal anti-inflammatory drug.

Table 2. Periodontal disease and risk of total cancer and the three most common cancers among never smokers in Health Professionals Follow-up
Study 1986–2012 (n = 19 933)
Baseline periodontal status Total cancer, 2959 cases Prostate cancera, 696 cases Colorectal cancer, 461 cases Melanoma, 451 cases
Case no. HR (95% CI)$b Case no. HR (95% CI)b$ Case no. HR (95% CI)b$ Case no. HR (95% CI)b$

No history of periodontal disease 2593 1.0 598 1.0 406 1.0 399 1.0
Periodontal disease 366 1.13 (1.01–1.27) 98 1.17 (0.94–1.47) 65 1.03 (0.75–1.39) 52 1.20 (0.89–1.63)
25–32 teeth 2542 1.0 584 1.0 396 1.0 407 1.0
17–24 teeth 285 0.93 (0.81–1.06) 77 0.91 (0.71–1.18) 45 0.89 (0.63–1.23) 31 0.78 (0.53–1.13)
0–16 teeth 132 1.05 (0.87–1.27) 35 0.89 (0.61–1.30) 20 1.02 (0.63–1.66) 12 0.65 (0.34–1.25)

No history periodontal disease 2512 1.0 576 1.0 395 1.0 392 1.0
with 17–32 teeth
No history periodontal disease 81 0.96 (0.76–1.21) 22 0.90 (0.57–1.43) 11 0.90 (0.48–1.66) 7 0.61 (0.27–1.38)
with 0–16 teeth
History of periodontal disease 315 1.09 (0.96–1.23) 85 1.16 (0.92–1.48) 46 0.97 (0.70–1.34) 47 1.15 (0.84–1.58)
with 17–32 teeth
History of periodontal disease 51 1.45 (1.07–1.96) 13 1.09 (0.61–1.95) 9 1.35 (0.64–2.83) 5 0.95 (0.34–2.63)
with 0–16 teeth

a
Prostate cancers include only aggressive cancers.
$b
Hazard ratios (HRs) and 95% confidence intervals (CIs) controlling for age (continuous), race (white, other), alcohol use (0, 0.1–4.9, 5–14.9, 15–29.9,
≥30 g/day), physical activity (quintiles), history of diabetes (yes or no), body mass index (BMI; <22, 22–24.9, 25–29.9, 30+), geographical location
(south, west, northeast, midwest), height (quintiles), NSAID use.

Periodontal disease status at baseline was associated with a number and risk of the most common cancers among never
13% higher risk of developing cancer [95% CI 1.01–1.27] and a smokers in this cohort, i.e. prostate, melanoma and colorectal
44% increase in cancer risk was observed among those with cancer (Table 2).
periodontal disease and fewer teeth number (Table 2). No asso- Strong associations were observed for periodontitis and
ciations were observed with either periodontal disease or teeth smoking-related cancers (lung, bladder, oropharnygeal,

Volume 27 | No. 5 | May 2016 doi:10.1093/annonc/mdw028 | 

 original articles
 | Michaud et al.

Table 3. Periodontal disease and cancer risk among never smokers in HPFS 1986–2012 (n = 19 933)
Smoking-related cancersa, Bladder cancer, 222 cases Lung cancer, 109 cases Kidney cancer, 137 cases Esophageal and Pancreatic cancer, 141
752 cases oropharyngeal cancers, 93 cases
cases
Case no. HR (95% CI)b Case no. HR (95% CI)b Case no. HR (95% CI)b Case no. HR (95% CI)b Case no. HR (95% CI)b Case no. HR (95% CI)b

Baseline exposures status

No periodontal disease 644 1.0 189 1.0 101 1.0 120 1.0 74 1.0 117 1.0
Periodontal disease 108 1.33 (1.07–1.65) 33 1.38 (0.93–2.05) 13 0.92 (0.49–1.71) 17 1.06 (0.61–1.85) 13 2.25 (1.30–3.90) 24 1.57 (0.98–2.50)
25–32 teeth 628 1.0 186 1.0 89 1.0 115 1.0 77 1.0 118 1.0
17–24 teeth 85 1.07 (0.83–1.36) 24 0.97 (0.60–1.54) 19 1.36 (0.76–2.42) 15 1.32 (0.74–2.34) 8 0.96 (0.44–2.06) 16 0.92 (0.52–1.61)
0–16 teeth 39 1.21 (0.85–1.72) 12 1.40 (0.73–2.70) 6 1.47 (0.59–3.68) 7 1.45 (0.64–3.31) 8 1.73 (0.74–4.06) 7 1.10 (0.48–2.51)
No periodontal disease and 17–32 teeth 624 1.0 186 1.0 97 1.0 116 1.0 71 1.0 111 1.0
No periodontal disease and 0–16 teeth 20 0.83 (0.52–1.35) 3 0.37 (0.09–1.53) 4 1.05 (0.34–3.23) 4 1.14 (0.40–3.27) 3 1.05 (0.30–3.74) 6 1.50 (0.62–3.64)
Periodontal disease and 17–32 teeth 89 1.22 (0.96–1.54) 24 1.09 (0.70–1.70) 11 0.86 (0.44–1.67) 14 1.02 (0.56–1.86) 14 1.97 (1.08–3.59) 23 1.68 (1.05–2.70)
Periodontal disease and 0–16 teeth 19 2.57 (1.56–4.21) 9 5.06 (2.32–11.0) 2 2.46 (0.53–11.3) 3 2.04 (0.60–6.91) 5 6.29 (2.13–18.6) 1 0.72 (0.10–5.44)
Updated periodontal status
No periodontal disease 595 1.0 178 1.0 89 1.0 113 1.0 71 1.0 110 1.0
Periodontal disease 157 1.27 (1.05–1.53) 48 1.29 (0.92–1.82) 25 1.27 (0.77–2.10) 24 1.09 (0.68–1.75) 22 1.54 (0.92–2.58) 31 1.26 (0.82–1.93)

a
Includes lung, bladder, kidney, oropharyngeal, esophagus, pancreatic, stomach and liver. Stomach and liver cancers are not shown individually as the number of cases was too small (46 and 24, respectively).
Hazard ratios (HRs) and 95% CIs controlling for age (continuous), race (white, other), alcohol use (0, 0.1–4.9, 5–14.9, 15–29.9, ≥30 g/day), physical activity (quintiles), history of diabetes (yes or no), body
b

mass index (BMI; <22, 22–24.9, 25–29.9, 30+), geographical location (south, west, northeast, midwest), height (quintiles), NSAID use.
Volume 27 | No. 5 | May 2016

Annals of Oncology
Annals of Oncology original articles
esophageal, kidney, stomach and liver [17]). For all smoking- evidence for a link with a number of conditions, including athero-
related cancers combined, a 33% increase in risk was observed sclerosis [1, 18, 19]. Although research on periodontal disease and
for those with periodontal disease at baseline, compared with no cancer lags behind the work on atherosclerosis, case–control
history of periodontal disease (Table 3). The association was studies have reported positive associations on oral health and risk
stronger among those who had advanced periodontitis of head and neck cancer for some time [13]. The real challenge,
(HR = 2.57, 95% CI 1.56–4.21) when compared with those who with cancer as with heart disease, is to evaluate causality in the
had no periodontitis and 17 or more teeth remaining at baseline. associations. Prospective cohort studies with long follow-up can
For bladder cancer, an HR of 5.06 was observed for men with begin to address causality, as exposure is ascertained many years,
advanced periodontitis at baseline, and a greater than sixfold in- sometimes decades, before cancer diagnosis. In our analysis, we
crease in risk was observed for oropharyngeal and esophageal observe persistent elevated risk of cancer 10 years post-periodontal
cancers (combined because of small numbers) (Table 3). While disease, suggesting the association is not likely to be due to reverse
no associations were observed for periodontal disease status and causation.
risk of lung and kidney cancers, elevated risks were noted To date, three large cohorts have reported associations with
among those with advanced periodontitis (associations were not periodontal disease and cancer. In the NHANES I Epidemologic
statistically significant). Updating periodontal disease status Follow-up Study, individuals with periodontitis at baseline had a
during follow-up years were similar to those observed with base- higher risk of fatal cancer compared with those with healthy per-
line periodontal status (Table 3). Similarly, updating tooth loss iodontium [odds ratio (OR) = 1.55, 95% CI 1.25–1.92], control-
did not strengthen associations for the advanced periodontitis ling for age and gender but not smoking [8]. In the NHANES
analyses. III data, a twofold increase in mortality of orodigestive cancers
Given that age was significantly higher among those with was observed among those with moderate (RR = 2.22, 95%
advanced periodontitis at baseline (Table 1), we were interested CI 1.11–4.46) and aggressive periodontitis (RR = 2.64, 95%
in whether the associations would be modified by age. The asso- CI 0.85–8.23; P-trend = 0.01), compared with those with no oral
ciations with advanced periodontitis were stronger in men disease [9]. A positive association was also reported for lung
who had these conditions before they were 60 years old than cancer in the WHI; periodontitis was associated with a 24%
those who were 60 years or older, but there was no statistically increase in risk (HR = 1.24, 95% CI 1.07–1.45), adjusting for
significant interaction (for smoking-related cancers, comparing smoking history [10].
advanced periodontitis with no periodontitis with 17–32 teeth, In addition, two cohorts examined antibodies to periodontal
age <60 years: HR = 3.70, 95% CI 1.45–9.40; age ≥60 years: pathogens and cancer risk [9, 11]. In a large prospective cohort,
HR = 2.22, 95% CI 1.24–3.97, P-interaction = 0.89). Periodontal a greater than twofold increase in risk of pancreatic cancer was
disease associations with smoking-related cancers were not observed among those with high levels of antibodies to a patho-
modified by BMI (data and additional analyses are provided in genic strain of P. gingivalis (OR = 2.38, 95% CI 1.16–4.90, com-
supplementary Material, available at Annals of Oncology paring >200 versus ≤200 ng/ml) after adjusting for known risk
online). factors [11]. In the NHANES III cohort study, elevated anti-
bodies to P. gingivalis (≥69.1 EU, compared with <69.1) were
associated with a threefold increase risk of orodigestive cancer
discussion mortality (RR = 3.03, 95% CI 0.99–9.31) [9].
Among never smokers, periodontal disease was associated with a Porphyromonas gingivalis, a periodontal disease pathogen, has
33% increase in risk of smoking-related cancers, and a 2.5-fold in- been shown to trigger oral inflammation and lead to immune dys-
crease in risk among men who had advanced periodontitis regulation in mice [20]. A number of biological mechanisms
(P = 0.0002), compared with men who had no periodontitis and could explain the link between periodontitis and cancer risk and
more than 16 teeth remaining. The associations were particularly include systemic inflammation [20], immune dysregulation as a
strong for risk of bladder, esophageal and head and neck cancers, result of chronic periodontitis [21, 22] and alteration of cancer sig-
but were also elevated in other smoking-related cancers. No naling pathways by oral pathogens [23]. Elevated systemic markers
associations were noted for prostate cancer, colorectal cancer or of inflammation in patients with chronic periodontitis include
melanoma, the three most common cancers in this population of interleukin 6, tumor necrosis factor-α and fibrinogen [24].
never smokers. Our finding that only smoking-related cancers appear to
Known risk factors for cancer were not, for the most part, asso- increase in risk with periodontitis suggests that alterations in
ciated with smoking-related cancers (including BMI, diabetes, immune pathways may play an important role in mediating
physical activity and height; data not shown) in this population; these associations, given that smoking is well known to affect
the only statistically significant association observed was for the immune response, both the innate and adaptive immune
obesity and kidney cancer. Our findings were strongest among systems [25, 26]. Moreover, smoking has been associated with
men with periodontal disease and few teeth number (16 or less), changes in DNA methylation patterns that are directly linked
i.e. with advanced periodontitis; this is a particularly important to the immune response [27]. Recent studies have shown
finding as it supports a biological mechanism, and may provide that certain bacterial metabolites can modulate the immune
an opportunity to treat patients with advanced periodontitis to response through epigenetic modification, including the regu-
reduce risk of cancer, although confirmation of these findings in latory T cells (Tregs) response [28, 29], which plays an import-
other populations will be important. ant role in carcinogenesis [30]. More work needs to be done to
There is overwhelming evidence that periodontitis results in understand the link between periodontal disease, bacterial
systemic inflammation and immune dysregulation, and growing dysbiosis and the systemic immune response. Obesity and

Volume 27 | No. 5 | May 2016 doi:10.1093/annonc/mdw028 | 

original articles Annals of Oncology

diabetes were not strongly associated with these cancers in this 4. McAllister F, Housseau F, Sears CL. Microbiota and immune responses in colon
population of never smokers (with the exception of kidney cancer: more to learn. Cancer J 2014; 20: 232–236.
cancer) and the periodontal disease associations were inde- 5. Hajishengallis G, Lamont RJ. Beyond the red complex and into more complexity:
the polymicrobial synergy and dysbiosis (PSD) model of periodontal disease
pendent of these factors; given that obesity is well known to
etiology. Mol Oral Microbiol 2012; 27: 409–419.
cause systemic inflammation, the association with periodontal
6. Hajishengallis G, Sahingur SE. Novel inflammatory pathways in periodontitis. Adv
disease may be driven by other immune components. More re- Dent Res 2014; 26: 23–29.
search in this area will likely reveal important new insights into 7. Michaud DS, Liu Y, Meyer M et al. Periodontal disease, tooth loss, and cancer risk
critical immune pathways for carcinogenesis. In addition, it is in male health professionals: a prospective cohort study. Lancet Oncol 2008; 9:
possible that some of the associations observed in this study 550–558.
were due to shared genetic susceptibility among those who 8. Hujoel PP, Drangsholt M, Spiekerman C, Weiss NS. An exploration of the
developed advanced periodontitis [31]. periodontitis–cancer association. Ann Epidemiol 2003; 13: 312–316.
The strength of the current analysis using the HPFS cohort 9. Ahn J, Segers S, Hayes RB. Periodontal disease, Porphyromonas gingivalis serum
antibody levels and orodigestive cancer mortality. Carcinogenesis 2012; 33:
include restriction to never smokers, 26 years of follow-up with
1055–1058.
large number of cancer cases and classification of advanced peri-
10. Mai X, LaMonte MJ, Hovey KM et al. History of periodontal disease diagnosis and
odontitis by combining self-reported periodontitis with tooth lung cancer incidence in the Women’s Health Initiative Observational Study.
number. There were also limitations, including small number Cancer Causes Control 2014; 25: 1045–1053.
of individuals with advanced periodontitis in this population 11. Michaud DS, Izard J, Wilhelm-Benartzi CS et al. Plasma antibodies to oral bacteria
(as defined in this analysis), self-reported periodontal disease and risk of pancreatic cancer in a large European prospective cohort study. Gut
status and number of teeth, no data on pocket depth, gingival 2013; 62: 1764–1770.
attachment loss or periodontal treatment, few numbers of non- 12. Eliot MN, Michaud DS, Langevin SM et al. Periodontal disease and mouthwash
White participants and no females. use are risk factors for head and neck squamous cell carcinoma. Cancer Causes
Control 2013; 24: 1315–1322.
Among never smokers, we observed a 2.5-fold increase in
13. Meyer MS, Joshipura K, Giovannucci E, Michaud DS. A review of the relationship
smoking-related cancers for those with advanced periodontitis.
between tooth loss, periodontal disease, and cancer. Cancer Causes Control
Associations were not modified by age or BMI and remained 2008; 19: 895–907.
elevated with a 10-year lag between periodontal status and 14. Joshipura KJ, Douglass CW, Garcia RI et al. Validity of a self-reported periodontal
cancer diagnosis. The impact of advanced periodontal disease disease measure. J Public Health Dent 1996; 56: 205–212.
on the immune function may be an underlying cause for these 15. Joshipura KJ, Pitiphat W, Douglass CW. Validation of self-reported periodontal
observed associations. Further studies need to examine immune measures among health professionals. J Public Health Dent 2002; 62:
perturbations from periodontal disease and biological pathways 115–121.
to address causality. 16. Douglass CW, Berlin J, Tennstedt S. The validity of self-reported oral health status
in the elderly. J Public Health Dent 1991; 51: 220–222.
17. Vineis P, Alavanja M, Buffler P et al. Tobacco and cancer: recent epidemiological
acknowledgements evidence. J Natl Cancer Inst 2004; 96: 99–106.
18. Lockhart PB, Bolger AF, Papapanou PN et al. Periodontal disease and
We would like to thank the participants and staff of the Health atherosclerotic vascular disease: does the evidence support an independent
Professionals Follow-up Cohort study for their valuable contri- association? A scientific statement from the American Heart Association.
butions as well as the following state cancer registries for their Circulation 2012; 125: 2520–2544.
help: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, 19. Desvarieux M, Demmer RT, Jacobs DR et al. Changes in clinical and microbiological
LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND, OH, OK, OR, periodontal profiles relate to progression of carotid intima-media thickness: the Oral
Infections and Vascular Disease Epidemiology study. J Am Heart Assoc 2013; 2:
PA, RI, SC, TN, TX, VA, WA, WY. The authors assume full
e000254.
responsibility for analyses and interpretation of these data.
20. Hayashi C, Gudino CV, Gibson FC, III, Genco CA. Review: pathogen-induced
inflammation at sites distant from oral infection: bacterial persistence and induction
of cell-specific innate immune inflammatory pathways. Mol Oral Microbiol 2010; 25:
funding 305–316.
This study was funded by UM1 CA167552 and R01 CA166150 21. Schmidt J, Jentsch H, Stingu CS, Sack U. General immune status and oral
from the National Cancer Institute of the National Institutes of microbiology in patients with different forms of periodontitis and healthy control
subjects. PLoS One 2014; 9: e109187.
Health.
22. Cifcibasi E, Ciblak M, Kiran B et al. The role of activated cytotoxic T cells in
etiopathogenesis of periodontal disease: does it harm or does it heal? Sci Rep
2015; 5: 9262.
disclosure
23. Garrett WS. Cancer and the microbiota. Science 2015; 348: 80–86.
The authors have declared no conflicts of interest. 24. Kim J, Amar S. Periodontal disease and systemic conditions: a bidirectional
relationship. Odontology 2006; 94: 10–21.
25. Arnson Y, Shoenfeld Y, Amital H. Effects of tobacco smoke on immunity,
references inflammation and autoimmunity. J Autoimmun 2010; 34: J258–J265.
26. Cozen W, Diaz-Sanchez D, James Gauderman W et al. Th1 and Th2 cytokines and
1. Hajishengallis G. Periodontitis: from microbial immune subversion to systemic
IgE levels in identical twins with varying levels of cigarette consumption. J Clin
inflammation. Nat Rev Immunol 2015; 15: 30–44.
Immunol 2004; 24: 617–622.
2. Fodor A. Utilizing ‘omics’ tools to study the complex gut ecosystem. Adv Exp Med
27. Bauer M, Linsel G, Fink B et al. A varying T cell subtype explains apparent tobacco
Biol 2014; 817: 25–38.
smoking induced single CpG hypomethylation in whole blood. Clin Epigenetics
3. Saggioro A. Leaky gut, microbiota, and cancer: an incoming hypothesis. J Clin
2015; 7: 81.
Gastroenterol 2014; 48 (Suppl. 1): S62–S66.

 | Michaud et al. Volume 27 | No. 5 | May 2016

Annals of Oncology original articles
28. Arpaia N, Campbell C, Fan X et al. Metabolites produced by commensal bacteria 30. Whiteside TL. What are regulatory T cells (Treg) regulating in cancer and why?
promote peripheral regulatory T-cell generation. Nature 2013; 504: 451–455. Semin Cancer Biol 2012; 22: 327–334.
29. Furusawa Y, Obata Y, Fukuda S et al. Commensal microbe-derived butyrate 31. Arora M, Weuve J, Fall K et al. An exploration of shared genetic risk factors
induces the differentiation of colonic regulatory T cells. Nature 2013; 504: between periodontal disease and cancers: a prospective co-twin study. Am J
446–450. Epidemiol 2010; 171: 253–259.

Annals of Oncology 27: 947–952, 2016

doi:10.1093/annonc/mdw044
Published online 21 February 2016

A phase I trial of panobinostat and epirubicin in solid

tumors with a dose expansion in patients with sarcoma
S. Thomas, R. Aggarwal, T. Jahan, C. Ryan, T. Troung, A. M. Cripps, P. Raha, K. T. Thurn, S. Chen,
J. A. Grabowsky, J. Park, J. Hwang, A. Daud & P. N. Munster*
Department of Medicine/Hematology and Oncology, University of California, San Francisco, San Francisco, USA

Received 14 September 2015; revised 12 November 2015 and 14 January 2016; accepted 27 January 2016

Background: Treatment options for sarcoma are limited. Histone deacetylase inhibitors increase the efficacy of topo-
isomerase II inhibitors by promoting access to chromatin and by down-regulating DNA repair. Thus, combined panobino-
stat and epirubicin therapy was evaluated to treat refractory sarcoma.
Patients and methods: Patients with advanced solid tumors were enrolled in a 3 + 3 dose-escalation phase I trial of
panobinostat given on days 1, 3, and 5 followed by 75 mg/m2 of epirubicin on day 5 in 21-day cycles, with a dose expan-
sion at maximum tolerated dose (MTD) in 20 sarcoma patients. Peripheral blood mononucleocyte histone acetylation was
also evaluated.
Results: Forty patients received 20–60 mg panobinostat. Dose-limiting toxicities included thrombocytopenia, febrile
neutropenia, and fatigue at 60 mg, defining a panobinostat MTD at 50 mg. Four responses were seen in 37 assessable
patients, all after progression on prior topoisomerase II inhibitors. For those with sarcoma, 12 of 20 derived clinical
benefit (1 partial response and 11 stable disease, median overall survival 8.3 months), including 8 of 14 previously pro-
gressed on topoisomerase II therapy. Treatment benefits correlated with increased histone acetylation and decreased
neutrophil count on day 5.
Conclusions: Panobinostat and epirubicin treatment is well tolerated and may reverse anthracycline resistance.
Changes in histone acetylation and associated decrease in neutrophil count correlated with clinical benefit and warrant
investigation as predictive biomarkers.
Clinical trial: This trial is registered at www.Clinicaltrials.gov, Identifier: NCT00878904.
Key words: histone deacetylase, panobinostat, sarcoma, epirubicin, topoisomerase

introduction progression-free survival (PFS) from 1.6 to 4.6 months, but not
overall survival (OS) 12.5 versus 10.7 months [5]. Trabectedin
More than half of patients treated for localized soft tissue was recently approved with similar benefit for liposarcoma and
sarcoma will experience relapse. Anthracycline-based chemo- leiomyosarcoma; eribulin showed PFS of 2.6 months and OS of
therapy remains the standard of care in the first-line setting [1], 13.5 months [6].
including topoisomerase II inhibitors doxorubicin and epirubi- Histone deacetylases (HDACs) modulate gene expression and
cin that act by increasing DNA damage and promoting apop- protein activity by regulating protein acetylation. HDAC inhibi-
tosis [2]. Response rates are low and resistance to doxorubicin is tors, vorinostat and romidepsion, have been approved for T-cell
common [3, 4]. Treatment with pazopanib modestly improves lymphoma and panobinostat for myeloma [7]. Preclinical
studies have shown that HDAC inhibitors potentiate DNA-
damaging chemotherapeutics in various cancer types, including
*Correspondence to: Prof. Pamela N. Munster, Department of Medicine/Hematology and
sarcoma [8–11]. Prior clinical studies evaluating HDAC inhibi-
Oncology, Helen Diller Family Comprehensive Cancer Center, University of California,
San Francisco, 1450 3rd street, San Francisco, CA 94158, USA. Tel: +1-415-353-7287; tors in combination with anthracyclines demonstrated efficacy
E-mail: pamela.munster@ucsf.edu [12, 13]. Supportive preclinical studies suggest a role for

Published by Oxford University Press on behalf of the European Society for Medical Oncology 2016. This work is written by (a) US Government employee(s) and is in the
public domain in the US.