In the last decades health professionals have been often organized into many specialties and subspecialties directed to several body organs and systems. The human organism is a unity that is composed by an infinite number of biologic processes so strongly linked that abnormalities in any part of the body and/or its processes may have deep effects in many other body areas, exemplified in this review by two highly prevalent diseases: PD and DM 25 .
PD is a chronic infectious disease, caused by Gram-negative microorganisms. An imbalance between a localized infection and an exaggerated host inflammatory response plays a pivotal role in determining gingival tissue damage. Recent evidence suggests that the effect of PD might not be limited just to the oral cavity but it might have systemic consequences. Indeed, PD has also been associated with a moderate systemic inflammatory response. Although, the mechanisms behind this association remain unclear, PD might represent one distant source of low-grade systemic inflammation. This association could explain the increased risk of impaired metabolic control in diabetes-related complications and the adverse effects of DM on periodontal health 15 . There is strong evidence that the prevalence, severity and progression of PD are significantly higher in people with DM 97,98,101 .
In this review we describe the common processes involved in PD and DM and briefly review the evidence produced to support an assocation between PD, DM and diabetes-related complications.
PD is a chronic bacterial infection that affects both the gingiva and the bone that supports the teeth and is caused by anaerobic Gram-negative microorganisms that are present in the bacterial plaque that adheres to the teeth 69 .
PD is a very prevalent condition. In the United States, over half the population aged 18 years or more have PD in its early stages, increasing to up to 75% after the age of 35 years; its mild to moderate forms are present in 30% to 50%, and the severe generalized form in 5% to 15% of the general adult population 25 . PD has even higher prevalence in minorities, in poor and developing countries and a considerable global variation 77,96 .
The presence of anaerobic Gram-negative bacteria causes a local inflammatory response that becomes chronic and progressive; this inflammation of the gingiva causes alveolar bone destruction and loss of the tissue attachment to the teeth, caused by components of microbial plaque that have the capacity to induce an initial infiltrate of inflammatory cells, such as lymphocytes, macrophages, and polymorphonuclear leukocytes (PMNs) 93 .
Some microbial components, especially lipopolysaccharide (LPS), activate macrophages that synthesize and secrete a great variety and amount of pro-inflammatory molecules, such as the cytokines interleukin-1 (IL-1) and tumor necrosis factor-α (TNF-α); prostaglandins, especially prostaglandin E2 (PGE2); and some other enzymes 93 .
Bacterial toxins can also activate T lymphocytes to produce IL-1 and lymphotoxin (LT), a molecule with properties that are similar to those of TNF-α. These cytokines show potent pro-inflammatory and catabolic activities, and have important roles in periodontal tissue destruction caused by collagenolytic enzymes such as metalloproteinases (MMPs) 93 . These collagenolytic enzymes are activated by reactive oxygen species and elevate the levels of interstitial collagenase in inflamed gingival tissue 56 .
The attachment loss deepens the sulcus, creating a periodontal pocket that contains thousands of millions of bacterial cells. This stage is the transition between gingivitis and periodontitis, the most common PDs 26,50 .
When bacterial biofilms on the teeth are not disrupted on a regular basis, the emergences of Gram-negative anaerobic bacterial species activate several host processes that will interfere in the extent and severity of the disease 25 .
Recently, many advances have occurred in the knowledge of the nature of the infectious agents involved in PD. Approximately 500 different bacterial entities and various human viruses have already been associated with the formation of dental microbial plaque 3 .
The most frequently recognized periodontal pathogens belong to three microaerophilic species (Actinobacillus actinomycetemcomitans, Campylobacter rectus, and Eikenella corrodens) and seven anaerobic species (Porphyromonas gingivalis, Bacteroides forsythus, Treponema denticola, Prevotella intermedia, Fusobacterium nucleatum, Eubacterium, and spirochetes) 92 . Various herpes viruses, such as the human cytomegalovirus (HCMV) and Epstein-Barr virus (EBV-1), have recently also emerged as pathogens in cases of destructive PD 90 .
Many conditions can predispose and/or facilitate the occurrence of PD such as smoking 7,32,35,105 , genetic influences 9,49,62 , estrogen deficiency 30,32,35 , estrogen excess 39 , dyslipidemia 58,66,111 and obesity 2,27 . The prevalence of obesity is increasing worldwide. This epidemic is also associated with an increased occurrence of obesity-related diseases like hypertension, cardiovascular disease, metabolic syndrome and DM that are also linked to PD 2,27 .
DM encompasses a group of genetically and clinically heterogeneous metabolic disorders characterized by hyperglycemia that results from a defective insulin secretion and/or activity 89 .
DM is classified according to its etiology as type 1 (T1D), type 2 (T2D), gestational diabetes (GDM) and other specific types. T1D results from the destruction of beta-cells within the islets of Langerhans of the pancreas, which results in a complete insulin deficiency; it can be immunemediated or have an idiopathic etiology. T2D ranges from an insulin resistance which progresses into an insulin deficiency due to a secondary failure in the pancreatic beta-cells. GDM is defined as any degree of glucose intolerance with onset or first recognition during pregnancy. Lastly, the category “other specific types” comprehends a group of several types of DM with different etiologies 21 .
Developed countries have a higher prevalence of DM than developing countries and more women than men are affected with DM. T2D constitutes 90% of the cases. In 1995 the prevalence of DM in adults all over the world was estimated to be around 4.0% and it was expected to rise to 5.4% by the year 2025. Numerically it means a rise from 135 million in 1995 to 300 million in the year 2025. The majority of this increase will occur in developing countries. An increase of 42% (from 51 to 72 million) is expected in developed countries and an increase of 170% (from 84 to 228 million) in the developing world. Therefore, it is expected that by the year 2025, 75% of people with DM will be living in developing countries where the majority of people with DM are aged between 45-64 years old. In developed countries, the majority of people with DM are older than 65 years. DM will be increasingly occurring in urban areas 47 .
People with diabetes and with chronically poor metabolic control can experience micro-vascular and macro-vascular complications leading to a significant burden for the individual and for the society. This burden includes direct costs of medical care and indirect costs, such as loss of productivity, which result from diabetes-related morbidity and premature mortality 4,38 .
Health care expenses for people with diabetes is more than two-times higher than the expenses for people without diabetes; the direct and indirect expenditures attributable to diabetes in 2007 in the USA were estimated at US$174 billion, with slightly more spent on chronic complications attributable to DM than properly on DM care 4 . The International Diabetes Federation estimated that DM costs account for 5-10% of the total healthcare budget in many countries 40 .
In Brazil, the per capita total, direct medical, direct nonmedical and indirect costs of patients with T1D were US$ 1,741.42, US$ 1,319.15, US$ 61.47 and US$ 360.81, respectively. The total direct non-medical costs were US$ 195,461.54, spent with transportation for the patients and caregivers 29 . Total annual costs for care of outpatients with T2D were US$ 2,108 per patient, out of which US$ 1,335 per patient of direct costs (63.3%) and US$ 773 per patient of indirect costs (36.7%). Patients with both micro-vascular and macro-vascular complications had higher costs (US$ 3,199 per patient) compared to those with either micro-vascular (US$ 2,062 per patient) or macro-vascular (US$ 2,517 per patient) complications only. The greatest amount of direct costs was attributed to medication (48.2%) 6 .
Effects of diabetes mellitus on periodontal disease
The search conducted for this review used the BBO (Bibliografia Brasileira de Odontologia), MEDLINE, LILACS and PubMed for Controlled Trials databases, in English and Portuguese languages published from 1960 to October 2012. Primary research reports on investigations of relationships between DM/DM control, PD/periodontal treatment and PD/DM/diabetes-related complications identified relevant papers and meta-analyses published in this period. This review does not provide an assessment of the quality of the reports. The identified reports are displayed in figures organized according to the following groups: 1- The effects of DM on PD; 2- The effects of glycemic control on PD and 3- The effects of PD on glycemic control and on diabetes-related complications.
The studies listed in Figure 1 compared periodontal status in individuals with and without DM in the majority of the reports. These studies were classified according to the study design, type of diabetes, sample number, age range, evaluation of PD and other diabetes-related variables. The majority of the studies were cross-sectional (21/29) and limited in the possibility of providing a causal-effect relationship.
The link between DM and the adverse effects on PD has been extensively described 61,96 . We have found that 27 in 29 studies showed supportive evidence of the adverse impact of DM on periodontal health.
There were four studies of T1D; one study reported more extensive radiographic bone loss in patients with T1D compared to the controls 102 , Lalla, et al. 54 (2006) in a case-control study found that periodontal destruction can start very early in life in patients with T1D and becomes more prominent as children become adolescents 54 . In a population aged 4 to 33 years, Cianciola, et al. 13 (1982) reported a significantly higher prevalence of PD in T1D than in non-diabetic siblings and nondiabetic unrelated controls. The prevalence of PD among 11- to 18-year-old teenagers with DM was 9.8% as compared to 1.7% in controls without DM. An accelerated periodontal destruction was found in children and teens with DM, with poor metabolic control 13 . In an adult population aged 40-69 years, 58.4% of patients with long standing T1D exhibited severe PD as opposed to 7.1% of controls without DM 103 .
Regarding the relationship between T2D and PD, we identified fourteen reports. Two reports were comprised of patients aged 15 years or older 20,70 , and twelve 8,10,12,13,19,59,60,74,80,100,106,113 included only adults. Twelve of these fourteen studies reported significantly poorer periodontal health in subjects with T2D, whereas a significantly poorer greater prevalence was found in one study 8 and no significant difference was found in another study 113 .
Six reports consist of analyses in which subjects with T1D and T2D were analyzed together without distinction of diabetes type. Four studies included children and adolescents 6,46,53,85 and another two included only adult subjects 17,73 . Five of these six studies reported greater prevalence, extent, or severity of PD in subjects with DM 5,17,46,53,85 . One report did not find significant differences in PD between subjects with and without DM when an adequate metabolic control was found in the former group 73 .
Regarding GDM, five reports were analyzed. One was conducted only with women with GDM that were compared to a control group between the 34-36th gestation weeks. The results of the study suggest that gingivitis seems to be more prevalent in women with GDM compared to healthy pregnant women and the plaque accumulation seems to be the main cause of gingival inflammation 65 . Another study found that all types of DM increase the risk of PD, including GDM 61 . Two other studies conducted in the USA collected data from over 4,000 women with a history of GDM. One report included ages 15-44 112 and the other, ages 20-59 112 . Both reports concluded there is a strong relationship between GDM and PD. PD was found in 45% of pregnant women with GDM vs. 13% in the healthy pregnant women, with an adjusted odds ratio of 9.11. In non-pregnant women, 40% of women with T1D or T2D, 25% of those with a history of GDM, and 14% of healthy women had PD. The odds ratio for those with T1D and T2D was 2.76 63 . Novak, et al. 71 (2006) found the prevalence of PD to be higher in women with a history of GDM and concluded that these women may be at greater risk for developing more severe PD, than women without a history of GDM. Finally, Taylor 96 (2001) and Mealey 61 (2006) in two extensive literature reviews found a bidirectional interrelationship between all types of DM, including GDM and PD.
Effects of glycemic control on periodontal disease
Current evidence also supports poorer glycemic control contributing to poorer periodontal health. We have identified fourteen studies reporting this relationship. Two of these studies included subjects with T1D exclusively, seven studies subjects with T2D exclusively and five a combination of subjects with either T1D, T2D, GDM and others (Figure 2). One prospective study conducted with T1D did not show any association between the degree of glycemic control and PD but a positive association with local oral hygiene measures 82 , and another study, that was cross-sectional, has regarded this association 102 . Five of the seven reports published regarding the association between glycemic control and PD in T2D 10,43,59,75,106 have found this association and two did not 12,80 . We have found five studies providing information on the differences in periodontal health in groups of mixed types of diabetes 5,37,45,61,68 ; three have found this association 37,61,68 and two did not 5,45 . Among these fourteen studies, eleven were cross-sectional that imposes some limitations on the cause-effect inference; two were prospective and one was an extensive literature review. Otherwise, nine of these reports support the evidence of greater prevalence, extent and severity of PD and also provide evidence that glycemic control worsens in parallel with the worsening of PD.
Effects of periodontal disease on glycemic control and on diabetes-related complications
Substantial evidence has been demonstrating DM as a risk factor for the impairment of periodontal health and a growing body of evidence has been supporting PD as having an adverse effect on glycemic control and on the pathophysiology of diabetes-related complications. The inflamed periodontal tissue may serve as a chronic source of bacteria, bacterial products and many inflammatory mediators such as TNF-α, IL6, and IL1 that have been shown to have important effects on lipid and glucose metabolism 24,31,36,57 and have also been reported to be insulin antagonists and related to insulin resistance that is predominantly found in T2D and GDM 23,36,63,76 .
As shown in Figure 3, data interpretation is often confounded by varying definitions of DM and PD and different clinical criteria applied to determine the prevalence, extent, severity of PD, levels of glycemic control and diabetes-related complications; there is also marked heterogeneity in the studies' designs, conduct, length of follow-up, types of participants, and periodontal treatment protocols 61,72,76,96 .
Evidence regarding the effects of PD on glycemic control comes from observational and treatment studies (Figure 3). The treatment studies are a set of reports that include ten randomized clinical trials (RCTs), twelve non-RCTs, four meta-analyses, one literature review, one longitudinal study, one transversal study, one retrospective study, one prospective study and two clinical cases discussions.
The RCTs used control groups that were either treated controls, non-treated controls or controls that did not change their usual dental care. Among the ten RCTs, eight reported a beneficial effect for periodontal therapy 33,34,44,48,51,52,79,88 and two did not 1 . One of the RCTs, recently conducted, showed significant improvement in HbA1c levels but did not result in a statistically significant improvement in serum levels of inflammatory markers such as hs-CRP, d-8-iso, MMP-2 and MMP-9 52 .
An important source of variation in the RCTs is the use of antibiotics with the non-surgical periodontal therapy. This fact brings a lot of confusion in the interpretation of the results of these trials in such a way, that to date, there is no clear evidence to support a requirement for the use of antibiotics in combination with non-surgical periodontal treatment in order to observe an improvement in glycemic control associated with periodontal therapy 98 .
Among the group of twenty-three periodontal treatment studies that were not RCTs, seventeen reported a beneficial effect on glycemic control 14,16,17,22,41−43,55,64,83,84,87,94,99,101,109,110 and six did not 11,78,91,95,108 . Only ten of these studies had controls or comparison groups 11,14,22,43,78,84,94,99,101,108 . Like the RCTs, there was marked variation in the use of adjunctive antibiotics, with six of the eight studies that used systemic antibiotics reporting a beneficial effect on glycemic control 41,42,64,83,87,109 .
Additional evidence to support the effect of severe periodontitis on increased risk for poorer glycemic control comes from two longitudinal observational studies. A longitudinal epidemiological study of the Pima Indians in Arizona, USA, which present the world's highest reported prevalence of DM 16 , found that subjects with T2D in good to moderate control and with severe periodontitis at baseline were approximately six times more likely to have poor glycemic control at a 2-year follow-up than those without severe periodontitis at baseline 96 . Collin, et al. 14 (1998) in another observational study of 25 adults with T2D, aged 58-77 years, also reported an association between advanced periodontal disease and impaired metabolic control 14 .
Recently, some important trials have recognized that poor glycemic control is a major determinant for the development of the chronic complications of DM. The Diabetes Control and Complications Trial, the Epidemiology of Diabetes Interventions and Complications (EDIC) Trial, the long-term follow-up study of the DCCT, both conducted with T1D and the United Kingdom Prospective Diabetes Study (UKPDS) conducted with T2D, demonstrated that attaining and maintaining good glycemic control could reduce the risk for and/or postpone the progression of micro-vascular complications in patients with T1D and T2D 18,67,107 .
Initially, the UKPDS observed a statistically nonsignificant 16% reduction (P=0.052) in the risk of combined fatal or nonfatal myocardial infarction and sudden death. Recently, it was observed that a longstanding good metabolic control can bring significant long-term consequences including the reduction in the risks of fatal or nonfatal myocardial infarction and sudden death. The epidemiological analysis from the UKPDS showed a continuous association between the risk of cardiovascular complications and glycemic control; every percentage point decrease in HbA1c, was associated with a 25% reduction in diabetes-related deaths, 7% reduction in all-cause mortality, and a 18% reduction in combined fatal and nonfatal myocardial infarction 28 .
Some observational studies regarding the association between PD and the risk for DM complications have given strong evidence for this association. In a study conducted in Sweden, with 39 case-control pairs of individuals with T1D and T2D for a median follow-up time of six years, Thorstensson, et al. 104 (1996) observed a significantly higher prevalence of proteinuria and cardiovascular complications such as stroke, transient ischemic attacks, angina, myocardial infarction and intermittent claudication in the case group than in controls. These findings suggest that an association between renal disease, cardiovascular disease and its complications and severe periodontitis seems to exist 104 .
Saremi, et al. 80 (2005), studied the contribution of PD to the mortality associated with T2D in the Gila River Indian Community in Arizona, USA, on behalf of the National Institute of Diabetes and Digestive and Kidney Diseases, addressing nephropathy and cardiovascular disease. This was a prospective longitudinal study with a cohort of 628 individuals, aged approximately 35 years old, for a median follow-up of eleven years (range 0.3 to 16). During the study period 204 subjects died. Individuals with severe PD had 3.2 times greater risk for cardio-renal mortality (i.e., ischemic heart disease and diabetic nephropathy combined) compared with the reference group (no, mild, or moderate PD combined), after adjustment for several major risk factors of cardio-renal mortality such as age, sex, diabetes duration, HbA1c, body mass index (BMI), hypertension, blood glucose, cholesterol, electrocardiographic abnormalities, macro-albuminuria, and smoking 81 .
Another study conducted by Shultis, et al. 86 (2007), in the same community investigated the effect of periodontitis on overt nephropathy and end-stage renal disease (ESRD) in a group of 529 subjects with T2D, aged approximately 25 years old. After adjusting for age, sex, diabetes duration, BMI, and smoking, they found that periodontitis and edentulism were significantly associated with the risk of overt nephropathy and ESRD. The incidence of macro-albuminuria was 2.0, 2.1, and 2.6 times greater in individuals with moderate or severe periodontitis or in those who were edentulous, respectively, than those with none/mild periodontitis. The incidence of ESRD was also 2.3, 3.5, and 4.9 times greater for individuals with moderate or severe periodontitis or for those who were edentulous, respectively, than those with none/mild periodontitis 86 .
Summary and conclusions
The clinical and epidemiological evidence found in the literature we reviewed provides support for the concept that DM can have adverse effects on PD, that PD worsens in parallel with glycemic control and finally that PD is associated with an increase in the risk for diabetes-related complications. However, further prospective, rigorous, controlled trials with a larger number of patients, in ethnically diverse populations are warranted to establish these relationships and that treating PD can positively influence glycemic control and possibly reduce the burden of diabetes-related complications.