40018.

qxd 4/4/05 1:27 PM Page 418

Volume 76 • Number 3

Diabetes and Periodontal Disease:

A Case-Control Study
Guglielmo Campus,* Abeer Salem,* Sergio Uzzau,† Edoardo Baldoni,* and Giancarlo Tonolo‡

Background: Periodontitis is often associated with diabetes

and might be considered one of the chronic complications of dia-
betes mellitus, both in Type 1 (T1DM) and Type 2 (T2DM). This
case-control study was designed to evaluate the possible asso-
ciation between non–insulin-dependent diabetes (T2DM) and
clinical and microbiological periodontal disease among adult
Sardinians.

T
he criteria for diagnosing diabetes
Methods: A total of 212 individuals participated in this study:
have undergone significant changes
71 T2DM patients aged 61.0 ± 11.0 years and 141 non-diabetic
since the early 1960s; consequently,
controls in good general health aged 59.1 ± 9.2 years. All sub-
the diagnosis of periodontal diseases
jects were given a clinical periodontal examination for probing
has been better defined.1-3 Using refined
depth, attachment level, presence of calculus, bleeding on prob-
standards for diagnosing these two dis-
ing, and assessment of plaque. Subgingival plaque samples
ease states, several general trends are
were obtained, and P. gingivalis, P. intermedia, and T. forsythensis
apparent.
were identified using multiplex polymerase chain reaction.
Diabetes prevalence is increasing
Results: T2DM patients showed a significantly lower number
worldwide and it is estimated that more
of teeth present (P = 0.002); a significant increase in number of
than 300 million subjects will be affected
probing depths >4 mm, and percent of pocket depths >4 mm
by the year 2025;4 thus, all diabetes
(P = 0.04 and P = 0.05, respectively); periodontitis (P = 0.046);
complications will increase. Uncontrolled
bleeding on probing (P = 0.02); and plaque index (P = 0.01). A
or poorly controlled diabetes is associ-
significant association with diabetes was detected for plaque (χ2 =
ated with increased susceptibility to oral
4.46; P <0.05) and bleeding on probing (χ2 = 3.60; P <0.05). Con-
infections, including periodontitis. The
cerning bacteria prevalence, a positive association was detected
incidence of periodontitis increases
for P. gingivalis (χ2 = 2.80; P <0.05) and T. forsythensis (χ2 = 3.87;
with age among diabetic subjects after
P <0.05). Presence of plaque was positively associated with case
puberty.5,6 Periodontal disease may be
status (odds ratio [OR] = 1.3; 95% confidence interval [CI]: 1.2,
more frequent and severe in diabetic indi-
3.6) and with prevalence of P. gingivalis and T. forsythensis (OR =
viduals with more advanced systemic
1.2, 95% CI: 1.3, 2.2; and 1.2, 95% CI: 1.2, 1.8, respectively).
complications.5 The increased suscepti-
Conclusion: Patients with T2DM undoubtedly have a suscep-
bility does not correlate with increased
tibility for more severe periodontal disease. J Periodontol 2005;76:
levels of plaque and calculus. Collec-
418-425.
tively, the data support the hypothesis
KEY WORDS that periodontal disease could affect dia-
Diabetes, non-insulin dependent; periodontal diseases; risk betics, especially those with poorly con-
factors; Sardinia. trolled disease.5,7-9 Since type 2 diabetes
mellitus (T2DM) is debuting earlier in
patients, increasing their length of expo-
* Dental Institute, University of Sassari, Sassari, Italy.
† Diabetology Unit, Clinica Medica and Patologia Medica, University of Sassari.
sure to the disease, periodontal disease
‡ Department of Biomedical Sciences, Center for Biotechnology Development and might become a serious health and social
Biodiversity Research, University of Sassari.
problem. Type 2 diabetes mellitus (T2DM)
patients had a higher prevalence of perio-
dontal disease as determined by using
either periodontal attachment loss or
radiographic bone loss parameters, indi-
cating that T2DM is a risk factor for
periodontal disease.10 The United States
Adult National Survey11 found signifi-
cantly more missing teeth and sextants

418
40018.qxd 4/4/05 1:27 PM Page 419

J Periodontol • March 2005 Campus, Saalem, Uzzau, Baldoni, Tonolo

with deep pockets (>4 mm) in T2DM groups than in estrogen deficiency, gender (males having more dis-
controls using the Community Periodontal Index of ease), age (more disease seen in the elderly), and
Treatment Needs (CPITN). genetic factors.6,19 Smoking and the presence of cer-
Diabetes mellitus can be divided into two main, broad tain subgingival microorganisms have also been proven
categories: Type 1 (insulin-dependent) and Type 2 to be true risk factors. The study of risk in periodon-
(non–insulin-dependent). Although Sardinian popula- tal disease is a rapidly emerging field, and much is
tions have the second highest prevalence of Type 1 dia- yet to be learned.
betes in the world after the Finnish,12 the prevalence The aim of the present study was to evaluate the
of Type 2 diabetes is more than 90% of diabetes cases. possible association between adult T2DM and partic-
The diagnosis of Type 1 diabetes mellitus in young ular clinical and microbiological periodontal disease
patients is relatively easy, but in patients with adult- parameters, not present in the non-diabetic popula-
onset diabetes mellitus, the classification into Type 1 tion, among Sardinians.
or 2 is sometimes difficult.2,13 Indeed, adult Type 1
may masquerade as Type 2 at presentation, with a MATERIALS AND METHODS
slow progression to insulin dependency14 after several Selection of Samples
years of good metabolic control without insulin. Type 1 This study was designed as a case-control study. Cases
diabetes is mainly an autoimmune disease character- were identified as T2DM patients from another study20
ized by the specific destruction of pancreatic beta cells who were referred by the Diabetic Unit of the Univer-
and the presence of specific autoantibodies. These sity of Sassari. Diabetes was diagnosed according to
autoantibodies are also detected in a subgroup of the following criteria:
patients misclassified as Type 2 diabetics at diagno- 1) Fasting plasma glucose concentration more than
sis.15 This autoantibodies-positive Type 2 diabetes is 7.8 mmol/l at two different occasions.
referred to as Latent Autoimmune Diabetes in Adults 2) In patients with a fasting plasma glucose con-
(LADA)15 or Type 1.5 diabetes,16 and interest in LADA centration less than 7.8 mmol, a 75 g oral glucose tol-
patients is increasing in the wake of an ensuing world- erance test (OGTT) was performed. Plasma glucose
wide epidemic of Type 2 diabetes.17 higher than 11.1 mmol/l after 120 minutes during the
We previously reported18 that frequency of these OGTT was defined as diabetes. OGTT was performed
autoantibodies (mainly GAD65Ab) is high among Type after 14 days withdrawal of any drug known to inter-
2 Sardinian diabetic patients. These patients did not fere with glucose homeostasis.
show any other clinical characteristic which distin- 3) In insulin-treated patients, insulin was not
guished autoimmunity since we avoided enrolling required during the first 4 years after diagnosis, and
patients who had been treated with insulin during the GAD65Ab (65kDa glutamic acid decarboxylase
first 4 years of disease as well as those who belonged autoantibody, marker of autoimmune diabetes also in
to Type 1 diabetes multiplex families. LADA or Type 1.5 adult subjects) was measured.18
diabetes recognizes an autoimmunity mechanism while 4) Known age at onset of diabetes varied between
“classic” T2DM does not. Indeed, T2DM, where insulin 30 and 69 years.
resistance plays a major role together with a decrease At the time of enrollment, height and weight were
in the ability of the beta cell to secrete insulin in measured and body mass index (BMI) was calculated
response to a meal, is mainly part of the metabolic (weight in kg/height in meters).21 Hypertension was
syndrome more than being an autoimmune disease defined according to WHO guidelines.21 Fasting serum
such as T1DM or LADA. Thus, since a substantial pro- was obtained for total cholesterol – high-density lipopro-
portion of T2DM patients might have been misclassified tein (HDL), triglycerides, and creatinine – and at least
at diagnosis and might indeed be LADA patients, it three 24-hour urine collections obtained on different
remains to be evaluated whether the severity of perio- days over the period of 1 year were analyzed for albu-
dontal disease is due to different etiopathogenesis of min excretion rate (AER).22 Microalbuminuria was deter-
diabetic diseases (autoimmune or not) or simply to mined when the AER was 30 to 300 mg in at least two
the duration of the disease and metabolic control. Evi- of three collections in a 24-hour period; macroal-
dence also suggests that periodontal infection and buminuria was determined when AER was constantly
periodontal treatment have the potential to alter >300 mg in a 24-hour period. Glycated hemoglobin
glycemic control in diabetic patients.19 The presence (HbA1c) was measured routinely with high performance
of systemic disease in patients requiring periodontal liquid chromatography (HPLC). All biochemical and
therapy creates challenges for management and alter- hormonal measurements were performed at the labo-
ation of treatment plans, with emphasis on physician ratory of the Diabetic Unit, University of Sassari, using
consultation and preventive periodontal care. Several standard methods22 as previously described.
potentially important periodontal risk indicators include: Non-diabetic controls were selected from subjects
stress, coping behaviors, osteopenia associated with referred to the Dental Institute at the University of

419
40018.qxd 4/4/05 1:27 PM Page 420

Type 2 Diabetes Mellitus and Periodontitis Among Adult Sardinians Volume 76 • Number 3

Sassari. Individuals with a history of infectious disease Table 1.

and/or patients subjected to immunosuppressant drugs
Clinical Criteria* of Periodontitis
or chemotherapy at the time of admittance to the Dia-
betologic Center were withdrawn from the study. Categorized by Authors
Individuals with less than six teeth present were not
included in either group (T2DM or non-diabetic) in the Presence of Periodontal
study. Disease
The total sample consisted of 212 individuals. A None (score 0) 3 mm probing depth
total of 71 diabetics aged 61.0 ± 11.0 years (range 36 No furcation involvement of posterior
to 75 years), 31 males and 40 females, were included teeth
in the study. Due to the small frequency of GAD65Ab+ Bleeding on probing may be present
in T2DM patients,18 22 subjects were selected among in the active phase
the known GAD65Ab+ LADA patients. No bias in Mild (score 1) Pocket depths or attachment loss of
selection occurred since patients were selected in the 3 to 4 mm
order of random follow-up at the diabetic clinic (the first Localized areas of recession
22 LADA and the first 49 GAD65Ab- T2DM patients). Possible Class I furcation invasion areas
T2DM and LADA patients were considered in good
metabolic control (gmc) if HbA1c was consistently Moderate (score 2) Pocket depths or attachment loss of
4 to 6 mm
≤8.0% in the last 2 years or in poor metabolic control
Bleeding on probing
(bmc) if HbA1c was consistently >8.1% in the last Grade I and/or Grade II furcation
2 years. A total of 141 subjects in good general health invasion areas
aged 59.1 ± 9.2 years (range 35 to 75 years), 61 males Tooth mobility of Class I
and 80 females, were included as controls.
All subjects selected (diabetics and controls) had Advanced (score 3) Bleeding on probing
at least six teeth in their mouths. Pocket depths or attachment loss over
6 mm
Subjects were designated as either current smok-
Grade II, Grade III furcation invasion areas
ers, non-smokers (never smoked), or ex-smokers Mobility of Class II or Class III
(quit smoking at least 2 years prior to study com-
* Based on American Dental Association criteria.24
mencement). Among the 212 participants, 59 were
current smokers, 114 were non-smokers, and 39 were
ex-smokers. based on the severity of probing depth. The clinician
uses clinical and radiographic data gathered and classi-
Methods fies the patient into one of four case types. These case
All patients recruited into the study were asked to com- types are commonly required for insurance billing. In
plete a questionnaire that included vital statistics (age, addition, the ADA provides treatment recommenda-
weight, etc.), address, employment information, mar- tions for each case type.24 Each patient was assigned
ital status, medical history, and their major complaint. to a specific group based on at least one of the corre-
Questions concerning present and past oral hygiene sponding criteria.
compliance were also included.
All subjects were given a clinical periodontal exami- Subgingival Plaque Samples
nation to determine probing depth and attachment level Supragingival plaque around the deepest site was
for each tooth present, presence of calculus, bleeding removed with a sterile curet and discarded. To obtain
on probing, assessment of plaque, and gingivitis scores a subgingival plaque sample, a second sterile curet
using the modified Löe and Silness gingival index.23 was introduced into the periodontal pocket and
The intraexaminer reproducibility of clinical diagnoses extended apically as far as possible. Samples were
was provided prior to the study by examining and reex- collected in sterile tubes containing 40% glycerol and
amining, after 72 hours, 30 subjects and calculating stored at −80°C. A total of 212 samples were collected;
the Cohen’s kappa for presence/absence of gingival however, bacterial DNA was extracted from only 199
bleeding/calculus restorations and probing depth. The subgingival plaque samples as described by Jaulhac;25
resulting kappa values ranged between 0.87 and 0.93, 13 samples were discharged since not enough material
suggesting a good intraexaminer agreement. The indi- was collected. Briefly, 30 to 50 µl of clinical specimens
viduals were grouped according to extent and severity were centrifuged to collect bacteria, and the pellet
of periodontal disease using the criteria described in obtained was washed twice and resuspended in 150 µl
Table 1. The criteria were derived from the American of phosphate buffered saline.
Dental Association (ADA), and subjects were catego- Suspensions were boiled for 10 minutes to release
rized by the authors. This classification is primarily the bacterial DNA and centrifuged for 3 minutes at

420
40018.qxd 4/4/05 1:27 PM Page 421

J Periodontol • March 2005 Campus, Saalem, Uzzau, Baldoni, Tonolo

12,000 × g. DNA was extracted from the supernatant variance (ANOVA) was performed for means com-
with a 5% solution of chelating resin in water for a parison among T2DM subjects in good metabolic con-
30-minute incubation at 55°C. Finally, samples were trol, T2DM in bad metabolic control, and non-diabetics.
centrifuged for 3 minutes at 12,000 × g to eliminate Second, a two-way table analysis was conducted to
the resin, and the supernatant was saved and stored determine the crude odds ratio for the relationship
at −20°C. PCR amplifications were performed in a vol- between diabetics and controls. The presence of bleed-
ume of 25 µl containing 0.2 mM of each deoxynucle- ing and calculus was then dichotomized, and a pres-
oside triphosphate,§ 2 U of dynazime termopolymerase ence of <20% of sites affected was considered 0, while
and 1X dynazime buffer, and 2.5 µl of template. Iden- the presence of >20% of sites affected was considered
tification of T. forsythensis and P. intermedia chromo- 1. Third, the possible association in diabetes cases (D)
somal DNA was carried out with a multiplex PCR26 and controls (C) between the dental parameters mean
with primers PAU (5′-AGA GTT TGA TCC TGG CTC attachment level (continuous variable, mm), mean
AG-3′), BFV530 (5′-GTA GAG CTT ACA GTA GAG probing depth (continuous variable, mm), percentage
CTT ACA-3′), and Pi (5′-GTT GCG TGC ACT CAA of probing depth number of teeth present (continuous
GTC CGC C-3′). Amplicons of 840 bp and 660 bp variable), and case status was analyzed using uncon-
were expected for T. forsythensis and P. intermedia, ditional logistic regression analysis.29
respectively. The mixture was overlaid with two drops
of light mineral oil. PCR amplification was performed RESULTS
in an automated thermal cycler
with initial denatura- The 22 LADA subjects were otherwise similar to the
tion (94°C, 5 minutes), followed by 30 cycles of de- GAD65Ab- T2DM patients in relation to periodontal
naturation (94°C, 1 minute), annealing (55°C, 1 minute), status. Therefore, all diabetics were considered a sin-
and extension (72°C, 2 minutes), with a single final gle group, called T2DM. No significant differences in
extension at 72°C for 7 minutes. After the reaction, terms of urinary albumin excretion were present
10 µl of the product was separated on a 1% agarose between T2DM in gmc and in bmc (data not shown).
gel, stained with ethidium bromide, visualized on an Similar results about BMI were observed in T2DM
ultraviolet transilluminator, and photographed with patients and controls. T2DM patients clearly showed
instant print film. The DNA molecular marker was 1 Kb a significantly lower number of teeth present (P =
DNA ladder.§ For identification of P. gingivalis DNA, a 0.002), and significantly increased number of probing
nested PCR was performed according to Leys.27 The depths >4 mm, percentage of probing depths >4 mm
first round of amplification was run with primer 785 (P = 0.04 and P = 0.05, respectively), periodontitis (P =
(5′-GGA TTA GAT ACC CTG GTA GTC-3′) and primer 0.046), bleeding on probing (P = 0.02), and plaque
422 (5′-GGA GTA TTT AGC CTT-3′). After initial index (P = 0.01) compared to non-diabetic controls
denaturation at 94°C for 5 minutes, amplification was (Table 2).
obtained with 27 cycles of denaturation (92°C, 1 min- Table 3 displays the comparison between good
ute), annealing (42°C, 1 minute), and extension (72°C, metabolic control (gmc) diabetics and poorly con-
3 minutes), with a single final extension at 72°C for trolled (bmc) diabetic subjects. Gmc subjects had a
7 minutes. A total of 0.25 µl of reaction was subse- significantly better periodontal condition than bmc sub-
quently used in a second round of amplification with jects: 17.9 ± 15.0 versus 27.8 ± 18.6 for the number
primers PG8R (TGTATATGACTGATGGTGAAAACC) of pocket depths >4 mm; 20.9% versus 32.9% for
and L189 (5′-GGT ACT TAG ATG TTT CAG TTC-3′). the percentage of pocket depths >4 mm; and 1.1 ± 0.7
Amplification was performed with initial denaturation and 1.6 ± 0.8 for the level of periodontitis. The com-
(94°C, 5 minutes), followed by 30 cycles of denatu- parison between the control group and gmc demon-
ration (94°C, 1 minute), annealing (54°C, 1 minute), strates overlapping results for level of periodontitis and
and extension (72°C, 2.5 minutes), with a single final number of pocket depths >4 mm; moreover, no statis-
extension at 72°C for 7 minutes. After the reaction, tical difference was observed in the percentage of
10 µl of the product was analyzed as described above. pocket depths >4 mm. No significant differences in
BMI were observed between gmc and bmc. Notably,
Data Analysis the prevalence of smokers was significantly lower
Initially, clinical condition parameters and potential risk (χ2 = 10.34, P <0.05) in T2DM patients (11 subjects,
indicators were analyzed univariately to describe the 15.5%) than in controls (52 subjects, 36.9%) (data not
variables and distributions. Student t test between the shown).
two groups was calculated, and P <0.05 was considered Table 4 describes the association between diabetics
a significant level. To avoid the attenuating effect of (D) and controls (C) regarding presence of periodontitis,
unequal variability among groups on the value of t, a
square root transformation was performed when the § Gibco, Carlsbad, CA.
response variable was a count.28 One-way analysis of
Hybaid, Omnigene, Cambridge, MA.

421
40018.qxd 4/4/05 1:27 PM Page 422

Type 2 Diabetes Mellitus and Periodontitis Among Adult Sardinians Volume 76 • Number 3

gingival condition (plaque, blood, and calculus pres- ratio (OR) estimates and the associated 95% confidence
ence), and bacteria prevalence. No significant differ- intervals for the association between case status (dia-
ence between T2DM and controls was shown regarding betes) and covariates examined. Presence of plaque
the presence of periodontal disease (χ2 = 1.53, P >0.05) was positively associated with case status (OR = 1.3,
and presence of calculus (χ2 = 1.76, P >0.05), while a 95% CI: 1.2, 3.6) and with prevalence of P. gingivalis
significant association was detected regarding plaque and T. forsythensis (OR = 1.2, 95% CI: 1.3, 2.2; and
presence (χ2 = 4.46, P <0.05) and bleeding on probing 1.2, 95% CI: 1.2, 1.8, respectively).
(χ2 = 3.60, P <0.05). Concerning bacteria prevalence,
a positive association was noticed for P. gingivalis (χ2 = DISCUSSION
2.80, P <0.05) and T. forsythensis (χ2 = 3.87, P <0.05). We took advantage of an isolated population (Sardinians),
It is noteworthy that the prevalence of T. forsythensis was which allowed us to examine T2DM and non-diabetic
higher in the control group than in the T2DM group subjects with a very similar genetic background. In
(32.6% versus 19.7%). Table 5 presents the crude odds this way, we reduced interference due to differences in
genetic background, commonly
found in mixed populations, that
Table 2. might influence the analysis of
prevalence of periodontitis when
Descriptive Statistics of Clinical Condition Parameters in the
diabetic and control groups are
Two Groups not genetically homogeneous.
First, our data speak against
Diabetics Controls the hypothesis that severity of
(N = 71) (N = 141) periodontal disease is due to
Characteristics Mean ± SD (range) Mean ± SD (range) P Value* the different etiopathogenesis of
BMI 28.7 ± 6.8 (22.1-45.0) 29.1 ± 6.7 (21.4-45.0) P = 0.23 the diabetic diseases, autoim-
mune (LADA) or not (“classic”
N teeth present 17.6 ± 6.9 (6-28) 21.9 ± 4.3 (6-28) P = 0.002 T2DM). Our results corroborate
N pocket depth >4 mm 21.3 ± 17.0 (0.0-72.0) 18.4 ± 20:7 (0.0-10.8) P = 0.04 the findings of several studies
reporting a positive association
% of pocket depth >4 mm 24.7 ± 23.3 (0.0-100) 14.5 ± 15.8 (0.0-72.0) P = 0.005 between diabetes and perio-
Level of periodontitis 1.2 ± 0.8 (0.0-3.0) 1.1 ± 0.6 (0.0-3.0) P = 0.046 dontal disease.5,11,30,31 T2DM
patients were more likely to
Presence of bleeding 0.7 ± 0.4 (0.0-1.0) 0.5 ± 0.5 (0.0-1.0) P = 0.02 have periodontal disease de-
Presence of calculus 0.6 ± 0.5 (0.0-1.0) 0.6 ± 0.5 (0.0-1.0) P = 0.35 fined by pocket depth. The
comparison of T2DM and non-
Plaque index 0.9 ± 0.3 (0.0-1.0) 0.7 ± 0.4 (0.0-1.0) P = 0.01 diabetic samples allows us to
* Student t test. observe that the periodontal

Table 3.
Descriptive Statistics of Clinical Condition Parameters in Diabetic Patients With Good
Metabolic Control (gmc) and Bad Metabolic Control (bmc), and Controls

gmc bmc Controls

(N = 49) (N = 22) (N = 141)
Characteristics Mean ± SD (range) Mean ± SD (range) Mean ± SD (range) P Value*

HbA1c 6.7 ± 1.2 (4.8-8.0) 9.3 ± 1.4 (8.1-11.9) NA P = 0.04

Max GLY 158.0 ± 43.1 (97.0-309.0) 239.0 ± 48.0 (164.0-340.0) NA P = 0.02

BMI 27.3 ± 6.4 (20.8-43.7) 30.2 ± 6.3 (20.4-46.7) 29.1 ± 6.7 (21.4-45.0) P = 0.09

N pocket depth >4 mm 17.9 ± 15.0 (0.0-52.0) 27.8 ± 18.6 (0.0-72.0) 18.4 ± 20.7 (0.0-10.8) P = 0.03

% of pocket depth >4 mm 20.9 ± 21.5 (0.0-95.8) 32.9 ± 25.2 (0.0-100) 14.5 ± 15.8 (0.0-72.0) P = 0.005
Level of periodontitis 1.1 ± 0.7 (0.0-3.0) 1.6 ± 0.8 (0.0-3.0) 1.1 ± 0.6 (0.0-3.0) P = 0.046
* One-way ANOVA.

422
40018.qxd 4/4/05 1:27 PM Page 423

J Periodontol • March 2005 Campus, Saalem, Uzzau, Baldoni, Tonolo

Table 4. that prevalence of periodontitis is higher in diabetic

patients than in non-diabetics. In general, efforts to com-
Two-way Analysis for the Relationship
pare the degree of periodontitis with the degree of
between Diabetics (D) and Controls (C) metabolic control have yielded conflicting results.27,32,33
Our microbiological analysis was focused on three
D C pathogens generally associated with periodontal dis-
N (%) N (%) ease: P. gingivalis, P. intermedia, and T. forsythensis.
Periodontitis An analysis of the subgingival plaque showed that,
No 10 (14.1) 16 (11.3) overall, P. gingivalis, P. intermedia, and T. forsythensis
Low 40 (56.3) 96 (68.1) were frequently detected in the gingival pockets of
Medium 16 (22.6) 22 (15.6) both T2DM and control patients, demonstrating that
Severe 5 (7.0) 7 (5.0) these pathogens are not specific to T2DM patients,
χ2 = 1.53, P >0.05 and suggesting that the subgingival microbiota of
Plaque T2DM individuals might be similar to those found in
No 8 (11.3) 33 (23.4) non-diabetic periodontitis patients. We found that the
Yes 63 (88.7) 108 (76.6) presence of P. intermedia is dominant over P. gingi-
χ2 = 4.46, P <0.05, OR = 2.40, 95% CI = 1.04-5.53 valis and T. forsythensis in the subgingival pockets of
the Sardinian population of both T2DM and control
Bleeding
patients. This agrees with the results of other studies
No 17 (23.9) 52 (36.9)
Yes 54 (76.1) 89 (63.1)
done in Sardinia.34 Thus, none of the laboratory data
χ2 = 3.60, P <0.05, OR = 1.85, 95% CI = 0.98-3.53 obtained suggested any role in the development of
periodontal disease in T2DM. However, it is notewor-
Calculus thy that T. forsythensis was less frequently associated
No 24 (33.8) 61 (43.3) with periodontal disease in T2DM patients. T. forsythen-
Yes 47 (66.2) 80 (65.7) sis has been clearly associated with periodontal dis-
χ2 = 1.76, P >0.05, OR = 1.49, 95% CI = 0.82-2.70
ease35-37 and, specifically, it does not occur frequently
T. forsythensis in generalized aggressive periodontitis. Our findings
No 57 (80.3) 95 (67.4) that T. forsythensis prevalence was higher in the con-
Yes 14 (19.7) 46 (32.6) trol group than in the T2DM group suggests that T2DM
χ2 = 3.87, P <0.05, OR = 0.50, 95% CI = 0.25-0.993 patients may reach as significant an extent of perio-
P. intermedia dontal disease as the control subjects, even in the
No 28 (39.4) 50 (35.5) absence of bacterial growth. T2DM patients have an
Yes 43 (60.6) 91 (64.5) increased prevalence of periodontal disease; however,
χ2 = 0.32, P >0.05, OR = 0.84, 95% CI = 0.47-1.51 this was significant only in T2DM patients with poor
metabolic control. Yet, the differences in periodontal
P. gingivalis
status were minimal between T2DM patients with good
No 49 (69.0) 112 (79.4)
Yes 22 (31.0) 29 (20.6)
metabolic control and non-diabetic subjects. For T2DM
χ2 = 2.80, P <0.05, OR = 1.73, 95% CI = 0.90-3.31 patients, periodontal status must be carefully evalu-
ated on an individual basis, since it can be anticipated
that many or most well-controlled T2DM patients
respond favorably to periodontal therapy. A small
Table 5. group, however, may manifest a persistent life-long
Unconditional Logistic Regression tendency to increased incidence and severity of perio-
dontal destruction.
OR P Value 95% CI The severity of periodontal disease, even taking into
account the age correction, could bring attention to the
Presence of plaque 1.3 <0.01 1.2-3.6 very poor oral hygiene level in our population, T2DM
T. forsythensis 1.2 <0.01 1.2-1.8 or not, smokers or not. Patients in our population did
not follow doctors’ suggestions concerning oral hygiene
P. gingivalis 1.2 0.02 1.3-2.2 instructions, giving up smoking, changing their eating
habits, etc.
Age is a very important factor for periodontal dis-
disease level was significantly higher in the T2DM ease,38 and the prevalence of periodontal disease
group, confirming the data of Pima Indians.10,30 Despite increases rapidly with age. Patient ages in this study
the fact that the etiological factors and the level of ranged between 36 and 75 years, and the mean dur-
dental health behavior were similar, our results indicate ation of diabetes since diagnosis was 18.2 ± 5 years.

423
40018.qxd 4/4/05 1:27 PM Page 424

Type 2 Diabetes Mellitus and Periodontitis Among Adult Sardinians Volume 76 • Number 3

All T2DM patients had diabetes longer than 8 years, Sardinia. Diabetes Care 1992;15:1317-1322.
long enough for chronic complications of the disease 13. Harris MI, Zimmet P. Classification of diabetes mellitus
to appear. Since the “known” diabetes duration in and other categories of glucose intolerance. In: Keen H,
Defronzo R, Alberti KGMM, Zimmet P, eds. The Internatio-
T2DM does not correspond to the actual beginning of nal Textbook of Diabetes Mellitus. London: John Wiley
the disease,39 although it is possible to estimate that & Sons; 1992:3-18.
in these patients the disease begins some 10 to 12 years 14. Niskanen LK, Tuomi T, Karjalainen J, Groop LC, Uusitupa
before diagnosis, no attempt to differentiate T2DM for MIJ. GAD antibodies in NIDDM. Diabetes Care 1995;18:
diabetes duration has been done. 1557-1565.
15. Zimmet PZ, Tuomi T, Mackay IR, et al. Latent autoim-
CONCLUSION mune diabetes mellitus in adults (LADA). The role of
antibodies to glutamic acid decarboxylase in diagnosis
Patients with diabetes undoubtedly have more plaque and prediction of insulin dependency. Diabet Med 1994;
and thus may have more periodontitis. Poorly controlled 11:299-303.
diabetic patients (bmc) have a worse periodontal status 16. Juneja R, Hirsch IB, Naik RG, Brooks-Worrell BM,
than control subjects, but well-controlled diabetic pa- Greenbaum CJ, Palmer JP. Islet cell antibodies and glu-
tamic acid decarboxylase antibodies, but not the clinical
tients (gmc) have a periodontal status similar to control phenotype, help to identify type 1(1/2) diabetes in pa-
subjects. Thus, based on the current observations, tients presenting with type 2 diabetes. Metabolism 2001;
periodontal disease should be considered another 50:1008-1013.
important complication of diabetes. However, our obser- 17. Zimmet P, Alberti KG, Shaw J. Global and societal
vation is that the positive presence of plaque and severe implications of the diabetes epidemic. Nature 2001;41:
782-787.
periodontal infection can exist independently of the 18. Maioli M, Tonolo G, Bekris L, et al. GAD65 and IA-2
presence of diabetes, even in the absence of such autoantibodies are common in a subset of siblings of
etiologic cofactors as specific bacterial pathogens. Sardinian type 2 diabetic families. Diabetes Res Clin Prac
2002;56:41-47.
REFERENCES 19. Paulander J, Wennstrom JL, Axelsson P, Lindhe J. Some
1. Bell SS, Bowden GI, Halter DW, Polonsky JB. Maturity- risk factors for periodontal bone loss in 50-year-old indi-
onset diabetes of the young. Life Science 1994;55:413- viduals. J Clin Periodontol 2004;31:489-496.
422. 20. Tonolo G, Ciccarese M, Melis MG, et al. A resource for
2. Çolbak AC, Christau B, Marner B, Boech-Johnsen K, studying gene-environment interrelations in type 2
Nerup J. Incidence of insulin dependent diabetes melli- diabetes. Diabetes Nutr Metab 1998;11:157-168.
tus in age groups over 30 years in Denmark. Diabetic 21. Guidelines Subcommittee of the WHO/ISH Mild Hyper-
Med 1994;11:650-655. tension Liaison Committee. J Hypertens 1993;11:905-
3. Zimmet PZ. The pathogenesis and prevention of dia- 918.
betes in adults. Genes, autoimmunity and demography. 22. Tonolo G, Ciccarese M, Brizzi P, et al. Reduction of albu-
Diabetes Care 1995;18:1050-1064. min excretion rate in normotensive microalbuminuric
4. King H, Aubert RE, Herman W. Global burden of diabetes type 2 diabetic patients during long term simvastatin
1995-2025. Prevalence, numerical estimates and pro- treatment. Diabetes Care 1997;20:1891-1895.
jections. Diabetes Care 1998;21:1414-1431. 23. Löe H, Silness J. Periodontal disease in pregnancy:
5. Seppálá B, Seppálá M, Ainamo J. A longitudinal study Prevalence and severity. Acta Odontol Scand 1963;21:
on insulin-dependent diabetes mellitus and periodontal 533-551.
disease. J Clin Periodontol 1993;20:161-165. 24. American Dental Association. Risk Management Series:
6. Arrieta-Blanco JJ, Bartolome-Villar B, Jimenez-Martinez Diagnosing and Managing the Periodontal Patient.
E, Saavedra-Vallejo P, Arrieta-Blanco FJ. Dental prob- Chicago: American Dental Association; 1986:12-16.
lems in patients with diabetes mellitus (II): Gingival 25. Jaulhac B, Reyrolle M, Sodahlon YK, et al. Compari-
index and periodontal disease. Med Oral 2003;8:233- son of sample preparation methods for detection of
247. Legionella pneumophila in culture-positive bronchoalveo-
7. Miller ME, Baker L. Leukocyte functions in juvenile dia- lar lavage fluids by PCR. J Clin Microbiol 1998;36:2120-
betes mellitus: Humoral and cellular aspects. J Pediatr 2122.
1972;81:979-982. 26. Conrads G, Flemmig TF, Seyfarth I, Lampert F,
8. McMullen JA, Van Dyke TE, Horoszewicz HU, Genco Lutticken R. Simultaneous detection of Bacteroides
RJ. Neutrophil chemotaxis in individuals with advanced forsythus and Prevotella intermedia by 16S rRNA gene-
periodontal disease and a genetic predisposition to dia- directed multiplex PCR. J Clin Microbiol 1999;37:1621-
betes mellitus. J Periodontol 1981;52:167-173. 1624.
9. Akintewe TA, Kulasekara B, Adetuyibi A. Periodontitis 27. Leys EJ, Smith JH, Lyons SR, Griffen AL. Identification
diabetica. A case report from Nigeria. Trop Geogr Med of Porphyromonas gingivalis strains by heteroduplex
1984;36:85-86. analysis and detection of multiple strains. J Clin Micro-
10. Shlossman M, Knowler WC, Pettitt DJ, Genco RJ. Type biol 1999;37:3906-3911.
2 diabetes mellitus and periodontal disease. J Am Dent 28. Fleiss JL. The Design and Analysis of Clinical Experi-
Assoc 1990;121:532-536. ments. New York: John Wiley & Sons; 1986:62-63.
11. Oliver RC, Tervonen T. Periodontitis and tooth loss: Com- 29. Breslow NE. Covariance adjustment of relative-risk esti-
paring diabetics with the general population. J Am Dent mates in matched studies. Biometrics 1982;38:661-672.
Assoc 1993;124:71-76. 30. Bacic M, Plancak D, Granic M. CPITN assessment of
12. Muntoni S, Songini M. Sardinian collaborative group for periodontal status in diabetics patients. J Periodontol
epidemiology of IDDM: High incidence rate of IDDM in 1988;59:816-822.

424
40018.qxd 4/4/05 1:27 PM Page 425

J Periodontol • March 2005 Campus, Saalem, Uzzau, Baldoni, Tonolo

31. Nelson RG, Shlossman M, Budding LM, et al. Periodon- Correspondence: Dr. Guglielmo Campus, Dental Institute,
tal disease and NIDDM in Pima Indians. Diabetes Care University of Sassari, Viale San Pietro 43/C, I-07100 Sassari
1990;13:836-840. (Sardinia), Italy. Fax: 39-079-228541; e-mail: gcampus@
32. Ervasti T, Knuuttila M, Pohjamo L, Haukipuro K. Relation uniss.it.
between control of diabetes and gingival bleeding.
J Periodontol 1985;56:154-157. Accepted for publication July 1, 2004.
33. Tervonen T, Karjalainen K, Knuuttila M, Huumonen
S. Alveolar bone loss in type 1 diabetic subjects. J Clin
Periodontol 2000;27:567-571.
34. Maida C, Campus G, Piana A, Solinas G, Milia E, Castiglia
P. Periodontal status in an Italian young adult popula-
tion. Prevalence and relationship with periodontopathic
bacteria. New Microbiol 2003;26:47-56.
35. Grossi SG. Treatment of periodontal disease and con-
trol of diabetes: An assessment of the evidence and need
for future research. Ann Periodontol 2001;6:138-145.
36. Dogan B, Antinheimo J, Cetiner D, et al. Subgingival
microflora in Turkish patients with periodontitis.
J Periodontol 2003;74:803-814.
37. Takeuchi Y, Umeda M, Ishizuka M, Huang Y, Ishikawa
I. Prevalence of periodontopathic bacteria in aggres-
sive periodontitis patients in a Japanese population.
J Periodontol 2003;74:1460-1469.
38. Genco RJ. Current view of risk factors for periodontal dis-
eases. J Periodontol 1996;67:1041-1049.
39. Colwell JA. Intensive insulin therapy in type II diabetes:
Rationale and collaborative clinical trial. Diabetes 1996;
45:87-90.

425