Update on Prevalence of Periodontitis in Adults in the United States: NHANES 2009 to 2012.

BACKGROUND: This report describes prevalence, severity, and extent of periodontitis in the US adult population using combined data from the 2009 to 2010 and 2011 to 2012 cycles of the National Health and Nutrition Examination Survey (NHANES). METHODS: Estimates were derived for dentate adults, aged ≥30 years, from the US civilian non-institutionalized population. Periodontitis was defined by combinations of clinical attachment loss (AL) and periodontal probing depth (PD) from six sites per tooth on all teeth, except third molars, using standard surveillance case definitions. For the first time in NHANES history, sufficient numbers of non-Hispanic Asians were sampled in 2011 to 2012 to provide reliable estimates of their periodontitis prevalence. RESULTS: In 2009 to 2012, 46% of US adults, representing 64.7 million people, had periodontitis, with 8.9% having severe periodontitis. Overall, 3.8% of all periodontal sites (10.6% of all teeth) had PD ≥4 mm, and 19.3% of sites (37.4% teeth) had AL ≥3 mm. Periodontitis prevalence was positively associated with increasing age and was higher among males. Periodontitis prevalence was highest in Hispanics (63.5%) and non-Hispanic blacks (59.1%), followed by non-Hispanic Asian Americans (50.0%), and lowest in non-Hispanic whites (40.8%). Prevalence varied two-fold between the lowest and highest levels of socioeconomic status, whether defined by poverty or education. CONCLUSIONS: This study confirms a high prevalence of periodontitis in US adults aged ≥30 years, with almost fifty-percent affected. The prevalence was greater in non-Hispanic Asians than non-Hispanic whites, although lower than other minorities. The distribution provides valuable information for population-based action to prevent or manage periodontitis in US adults.

Incorporating severity and risk as factors to the Fardal cost-effectiveness model to create a cost-benefit model for periodontal treatment.

BACKGROUND: A previously described economic model was based on average values for patients diagnosed with chronic periodontitis (CP). However, tooth loss varies among treated patients and factors for tooth loss include CP severity and risk. The model was refined to incorporate CP severity and risk to determine the cost of treating a specific level of CP severity and risk that is associated with the benefit of tooth preservation. METHODS: A population that received and another that did not receive periodontal treatment were used to determine treatment costs and tooth loss. The number of teeth preserved was the difference of the number of teeth lost between the two populations. The cost of periodontal treatment was divided by the number of teeth preserved for combinations of CP severity and risk. RESULTS: The cost of periodontal treatment divided by the number of teeth preserved ranged from (US) $1,405 to $4,895 for high or moderate risk combined with any severity of CP and was more than $8,639 for low risk combined with mild CP. The cost of a three-unit bridge was $3,416, and the cost of a single-tooth replacement was $4,787. CONCLUSION: Periodontal treatment could be justified on the sole basis of tooth preservation when CP risk is moderate or high regardless of disease severity.

Update of the case definitions for population-based surveillance of periodontitis.

BACKGROUND: This report adds a new definition for mild periodontitis that allows for better descriptions of the overall prevalence of periodontitis in populations. In 2007, the Centers for Disease Control and Prevention in partnership with the American Academy of Periodontology developed and reported standard case definitions for surveillance of moderate and severe periodontitis based on measurements of probing depth (PD) and clinical attachment loss (AL) at interproximal sites. However, combined cases of moderate and severe periodontitis are insufficient to determine the total prevalence of periodontitis in populations. METHODS: The authors proposed a definition for mild periodontitis as ≥ 2 interproximal sites with AL ≥ 3 mm and ≥ 2 interproximal sites with PD ≥ 4 mm (not on the same tooth) or one site with PD ≥ 5 mm . The effect of the proposed definition on the total burden of periodontitis was assessed in a convenience sample of 456 adults ≥ 35 years old and compared with other previously reported definitions for similar categories of periodontitis. RESULTS: Addition of mild periodontitis increases the total prevalence of periodontitis by ≈31% in this sample when compared with the prevalence of severe and moderate disease. CONCLUSION: Total periodontitis using the case definitions in this study should be based on the sum of mild, moderate, and severe periodontitis.

Reduction of tooth loss associated with periodontal treatment.

Comparing tooth loss for populations comprising subjects with periodontal disease has been limited by broad and different definitions of disease severity. Numeric scores for periodontal disease severity and risk were used to enhance the precision of comparing tooth loss for two populations. Both populations received routine dental care, but only one received comprehensive periodontal treatment. The analysis provides evidence that adding periodontal treatment to routine dental care is associated with less tooth loss and more patients who do not lose any teeth. Furthermore, it may be possible to nearly eliminate tooth loss associated with periodontal disease.

Tooth loss in 776 treated periodontal patients.

BACKGROUND: The most common form of periodontitis is a variably progressive dynamic pathologic process that causes attachment loss, destroys the alveolar bone supporting a tooth, and terminates with tooth loss. We evaluated the loss of teeth of treated periodontal patients categorized by severity and risk. METHODS: Each of nine periodontists evaluated 100 consecutive periodontal maintenance patients. The disease severity and risk level were determined from data at the initial examination. The number of teeth lost was determined from data at the initial and maintenance visits. RESULTS: A stepwise regression analysis showed that disease (P = 0.0000478) and risk (P = 0.00129) scores predicted the mean tooth loss rate. The adjusted R(2) statistic was 88.56%. The ordinal logistic regression model showed that only the disease score (P <0.0005) was significantly associated with the probability of patients losing a specific number of teeth. CONCLUSIONS: Categorizing a patient by severity may be beneficial in the management of the periodontal patient. The disease score can be used to establish a criterion and target for care. For example, treatment can result in nearly no lost teeth when the severity is low, and this benefit is lost when the severity is high. The disease score provides an objective means to quickly determine severity. An increase in the disease score provides evidence that a new treatment plan is needed. Therefore, the effect of the routine use of the disease score could result in fewer patients with severe disease and reduce the number of teeth lost.

Periodontitis severity plus risk as a tooth loss predictor.

BACKGROUND: Tooth loss can be a consequence of the natural history of periodontitis. Stratification of periodontitis severity, risk, and tooth loss exists within the United States adult population, and tooth loss correlates to severity and risk. We evaluated the loss of teeth for a periodontitis-affected population categorized by the combination of severity and risk in which the subjects predominantly did not receive periodontal treatment. METHODS: The clinical records of 523 subjects enrolled in the Veterans Affairs Dental Longitudinal Study, covering a period of 15 years, were used. Disease severity, risk level, and the number of teeth lost for each subject were determined. RESULTS: A stepwise regression analysis showed that disease and risk scores predicted mean tooth loss rate. The P value for the disease score was <0.0005, and the P value for the risk score was 0.001. The ordinal logistic regression model showed that disease (P = 0.002) and risk scores (P = 0.000) were significantly associated with the probability of subjects losing a specific number of teeth. CONCLUSIONS: Tooth loss is more precisely and accurately predicted by the combination of risk score and periodontal disease score than by either score alone. The combined scores may be a surrogate variable for periodontal status. Because the scores are derived from routine clinical measurements, they may be useful for population surveillance and dynamics, practice management, patient care decisions, practice-based research, and the determination of treatment effectiveness and the factors required for successful treatment, resulting in improved oral health and higher clinician productivity and income.

Case definitions for use in population-based surveillance of periodontitis.

Many definitions of periodontitis have been used in the literature for population-based studies, but there is no accepted standard. In early epidemiologic studies, the two major periodontal diseases, gingivitis and periodontitis, were combined and considered to be a continuum. National United States surveys were conducted in 1960 to 1962, 1971 to 1974, 1981, 1985 to 1986, 1988 to 1994, and 1999 to 2000. The case definitions and protocols used in the six national surveys reflect a continuing evolution and improvement over time. Generally, the clinical diagnosis of periodontitis is based on measures of probing depth (PD), clinical attachment level (CAL), the radiographic pattern and extent of alveolar bone loss, gingival inflammation measured as bleeding on probing, or a combination of these measures. Several other patient characteristics are considered, and several factors, such as age, can affect measurements of PD and CAL. Accuracy and reproducibility of measurements of PD and CAL are important because case definitions for periodontitis are based largely on either or both measurements, and relatively small changes in these values can result in large changes in disease prevalence. The classification currently accepted by the American Academy of Periodontology (AAP) was devised by the 1999 International Workshop for a Classification of Periodontal Diseases and Conditions. However, in 2003 the Centers for Disease Control and Prevention and the AAP appointed a working group to develop further standardized clinical case definitions for population-based studies of periodontitis. This classification defines severe periodontitis and moderate periodontitis in terms of PD and CAL to enhance case definitions and further demonstrates the importance of thresholds of PD and CAL and the number of affected sites when determining prevalence.

Case Definitions for Use in Population-Based Surveillance of Periodontitis.

Many definitions of periodontitis have been used in the literature for population-based studies, but there is no accepted standard. In early epidemiologic studies, the two major periodontal diseases, gingivitis and periodontitis, were combined and considered to be a continuum. National United States surveys were conducted in 1960 to 1962, 1971 to 1974, 1981, 1985 to 1986, 1988 to 1994, and 1999 to 2000. The case definitions and protocols used in the six national surveys reflect a continuing evolution and improvement over time. Generally, the clinical diagnosis of periodontitis is based on measures of probing depth (PD), clinical attachment level (CAL), the radiographic pattern and extent of alveolar bone loss, gingival inflammation measured as bleeding on probing, or a combination of these measures. Several other patient characteristics are considered, and several factors, such as age, can affect measurements of PD and CAL. Accuracy and reproducibility of measurements of PD and CAL are important because case definitions for periodontitis are based largely on either or both measurements, and relatively small changes in these values can result in large changes in disease prevalence. The classification currently accepted by the American Academy of Periodontology (AAP) was devised by the 1999 International Workshop for a Classification of Periodontal Diseases and Conditions. However, in 2003 the Centers for Disease Control and Prevention and the AAP appointed a working group to develop further standardized clinical case definitions for population-based studies of periodontitis. This classification defines severe periodontitis and moderate periodontitis in terms of PD and CAL to enhance case definitions and further demonstrates the importance of thresholds of PD and CAL and the number of affected sites when determining prevalence.

Overview of efficacy outcome variables for the evaluation of periodontal disease treatment.

Since the late 1980s, numerous new products for the prevention, diagnosis, and treatment of periodontal disease have been developed. These products have required randomized clinical trials to provide proof of efficacy and safety prior to marketing. Originally, no consensus existed in the scientific community, regulatory agencies, or industry as to the appropriate design, conduct, and methods of analyses of clinical trials in periodontics. However, in 1967, the Task Force on Design and Analysis in Dental and Oral Research was formed. This task force held three conferences to address specific issues in clinical oral health research; these meetings provided the foundation for the development of guidelines for clinical trials by the Subcommittee on Clinical Trials in Periodontics of the American Academy of Periodontology. Members of the subcommittee held an international symposium in 1996. Following this meeting, 4 working groups were convened and charged with the development of guidelines for the design and conduct of clinical trials in periodontics. These working groups developed guidelines for periodontal clinical trials in the areas of chemotherapeutic agents to slow or arrest periodontitis, products and methods for diagnosis and/or management of periodontitis, products designed to regenerate periodontal tissues, and clinical trials on endosseous dental implants. These guidelines have been published and approved or accepted by the American Academy of Periodontology and the Council on Scientific Affairs of the American Dental Association.

The Oral Health Information Suite (OHIS): its use in the management of periodontal disease.

Health care costs continue to increase at a rapid rate. Dental costs alone have risen from $31.5 billion in 1990 to $70.3 billion in 2002, outpacing inflation by 160 percent. Payers for health care services have no means to evaluate the value of these large expenditures. Quantified information is not available regarding a patient's condition prior to and after treatment nor on the probability of future disease. The absence of this information prevents dentists from responding effectively to challenges by payers and patients, and specifically prevents dentists from effectively influencing the quality of periodontal care. We have developed a user-friendly Internet-based technology that quantifies risk for periodontitis and periodontal disease severity and extent and generates recommended treatments and interventions. A caries risk assessment tool has also been developed, and an oral cancer assessment tool is being developed. This technology, designated the Oral Health Information Suite (OHIS), provides quantitative information to the clinician and patient as an aid to diagnosis and to facilitate individual, needs-based treatment planning. OHIS enables successful application of the wellness model of oral health care, which may be expected to result in more uniform and accurate clinical decision making, improved oral health, reduction in the need for complex periodontal therapy, reduction in oral health care costs, and improved clinician productivity and income. It also will permit patients to become more involved in their oral health care, payers to quantify and predict their health care expenditures, dentists to experience an increase in trust and respect, and periodontists to be more properly consulted regarding periodontal care.

Clinical effects of scaling and root planing on untreated teeth.

OBJECTIVE: The aim of this report is to examine whether scaling and root planing (SRP) in one area of the mouth may affect periodontal improvement in untreated areas in the same patient, possibly through systemic effects of treatment. MATERIAL AND METHODS: Twenty patients diagnosed with generalized aggressive periodontitis were randomized into treatment (n=11) and no treatment (n=9) groups. Within the treatment group, three quadrants were treated by SRP at week 0, 3, 12, and 24, while a single experimental quadrant remained untreated throughout the study. The outcome for all teeth was assessed using clinical parameters, subtraction radiography, and pathogenic bacteria levels in the subgingival flora over the 24-week study period. RESULTS: Compared with sites in no treatment patients, the treated sites in the treated patients showed a 1 mm decrease in probing depth (PD) (p<0.01) and a 0.5 mm increase in bone height (p<0.01) by 24 weeks. In untreated sites within treated subjects, however, PDs tended to improve (p=0.09) but at a reduced rate compared with treated sites. The levels of Actinobacillus actinomycetemcomitans, Porphyromonas gingivalis, and Tannerella forsythensis (Bacteroides forsythus) remained unchanged in untreated sites while levels of Prevetolla intermedia and Treponema denticola tended to decrease as compared with controls but did not reach significance. CONCLUSIONS: This study indicates that untreated sites in treated periodontitis patients show a trend towards clinical improvement and exhibit reductions in some but not all periodontopathic bacterial species tested.

The microbiological case for adjunctive therapy for periodontitis.

That chronic periodontitis is an infectious disease is now firmly established, and the primary role of Porphyromonas gingivalis, Tannerella forsythensis and Treponema denticola is generally accepted. Treatment by mechanical means such as scaling and root planing or surgery generally results in significant clinical improvement but may not arrest the progress or recurrence of disease. Several studies have shown that the probability of achieving lasting stability as measured by the arrest of progressive attachment loss and bone loss by primary mechanical therapy is a function, in major part, of whether pathogenic microorganisms are still present at local subgingival sites at the completion of active therapy. The infecting bacterial species are susceptible to killing by several antibiotics including, among others, tetracycline-class drugs, amoxicillin and metronidazole as well as by local exposure to chlorhexidine. Randomized clinical trials have shown that use of systemically administered antibiotics as an adjunct to mechanical therapies significantly enhances clinical outcomes and stability. Several slow-release devices that deliver anti-microbial drugs directly into periodontal pockets have been developed and are now on the market. Use of these devices permits local delivery of long-lasting, high concentrations of doxycycline (Atridox) minocycline (Arestin), and chlorhexidine (PerioChip) directly into periodontal pockets. Although these devices differ with regard to ease of use, concentration of drug delivered and length of time high drug concentrations can be maintained, randomized clinical trials have shown that their use as an adjunctive treatment to scaling and root planing results in a significantly greater reduction of periodontal pocket depth and an average increase in clinical periodontal attachment level of about 0.8 mm. Gain in clinical attachment is greater in deeper pockets than in shallower pockets. Locally delivered adjunctive anti-microbial therapy is an effective means to enhance therapeutic outcomes.

Periodontitis vaccine decreases local prostaglandin E2 levels in a primate model.

Interleukin-1beta, tumor necrosis factor alpha, prostaglandin E2 (PGE2), and Porphyromonas gingivalis-specific immunoglobulin G levels in gingival crevicular fluid were measured in primates immunized with a P. gingivalis vaccine followed by ligature-induced periodontitis. Only PGE2 levels were dramatically suppressed (P < 0.0001) in immunized animals versus controls. A significant correlation (P < 0.027) was also found between PGE2 levels and decreased bone loss scores. This study presents the first evidence of a potential mechanism involved in periodontitis vaccine-induced suppression of bone loss in a nonhuman primate model and offers insight into the role of PGE2 in periodontal destruction.

Use of risk assessment in attaining and maintaining oral health.

The cost of health care is increasing rapidly. A transition from the "repair" to a "wellness" model of oral health care that minimizes the occurrence of disease and the need for complex treatment may reduce costs and improve oral health. Quantification of risk is essential for successful use of the wellness model. Subjective risk assessment by expert clinicians is too variable to be useful in clinical decision making. However, user-friendly Internet-based technology has been developed that provides a consistently accurate and valid quantified risk assessment for periodontitis, as well as a quantified measure of disease severity and extent. The numeric information helps clinicians and patients make diagnoses and generate individual, needs-based treatment plans. This technology enables successful application of the wellness model of care in day-to-day dental practice. Use of the wellness model may result in more uniform and accurate periodontal clinical decision making, improved oral health, less need for complex periodontal therapy, lower oral health care costs, and improved clinician productivity and income.

Longitudinal validation of a risk calculator for periodontal disease.

BACKGROUND: Risk assessment and utilization of the results are important components of prevention, diagnosis and treatment of periodontal diseases. Risk assessment is relatively new to dentistry. Currently risk is assessed by subjective evaluation and results vary widely among clinicians. We have developed a computer-based risk assessment tool, the Periodontal Risk Calculator (PRC), for objective, quantitative assessment of risk. The purpose of the study reported here was to evaluate the accuracy and validity of this tool. METHODS: Clinical records and radiographs of 523 subjects enrolled in the VA Dental Longitudinal Study of Oral Health and Disease, covering a period of 15 years, were used. Information from baseline examinations was entered into the risk calculator and a risk score on a scale of l-5 for periodontal deterioration was calculated for each subject. Actual periodontal status in terms of alveolar bone loss determined using digitized radiographs, and tooth loss determined from the clinical records, was assessed at years 3, 9 and 15. The strength of the association between risk prediction and actual outcome was determined statistically. RESULTS: The risk scores were strong predictors of future periodontal status measured as worsening severity and extent of alveolar bone loss and tooth loss, especially loss of periodontally affected teeth. Over the entire 15-year period, risk scores consistently ranked groups from least to most bone loss and tooth loss. Risk groups differed greatly from one another. By year 3, the incidence rate of bone loss of group 5 was 3.7-fold greater than for group 2, and by year 15, the loss of periodontally affected teeth was 22.7-fold greater than for group 2 (p<0.001). By year 15, 83.7% of subjects in risk group 5 had lost one or more periodontally affected teeth compared to 20.2% of subjects in group 2. CONCLUSIONS: Risk scores calculated using the PRC and information gathered during a standard periodontal examination predict future periodontal status with a high level of accuracy and validity. Use of the risk assessment tool over time may be expected to result in more uniform and accurate periodontal clinical decision-making, improved oral health, reduction in the need for complex therapy and reduction in health-care cost.

Assessing periodontal disease risk: a comparison of clinicians' assessment versus a computerized tool.

BACKGROUND: The authors conducted a study to compare risk scores assigned by subjective expert clinician opinion with quantitative scores generated for the same subjects using the Periodontal Risk Calculator, or PRC. METHODS: The authors assembled a group of 107 subjects and performed standard periodontal examinations. The authors entered the resulting information into the PRC and calculated risk scores for two and four years, assuming no treatment would be performed. Using the same subject records, three groups of expert clinicians assigned risk scores for years 2 and 4. The authors analyzed the data to reveal the extent of interevaluator variation and the level of agreement between expert clinician scores and PRC scores. RESULTS: The extent of variation among scores assigned by individual expert clinicians was greater than the authors had expected. Expert clinicians consistently assigned more subjects to PRC risk group 2 and fewer to risk group 5 than did the PRC. The authors observed very high heterogeneity in the risk scores expert clinicians assigned to patients in each of the PRC-assigned groups. Thus, expert clinicians varied greatly in evaluating risk and, relative to the PRC, they appeared to underestimate periodontitis risk, especially for high-risk patients. CONCLUSIONS AND PRACTICE IMPLICATIONS: The authors' observations suggest that use of risk scores generated for individual patients by subjective expert clinician opinion about risk in periodontal clinical decision making could result in the misapplication of treatment for some patients and support the use of an objective tool such as the PRC. Use of the PRC over time may be expected to result in more uniform and accurate periodontal clinical decision making, improved oral health, reduction in the need for complex therapy and reduction in health care costs.

Perceived risk of deteriorating periodontal conditions.

BACKGROUND: Interpretation of risk for periodontitis is critical for treatment planning. How periodontists assess risk for periodontitis is unclear. PURPOSE: To study (1) what factors periodontists use when assessing the risks for worsening periodontal conditions anticipating that no treatment would be provided, and (2) if risk assessment is consistent and independent of specialty background training. MATERIAL AND METHODS: Medical history, clinical dental data, full-mouth intra-oral radiographs, and slide pictures were obtained from each of 51 subjects, and the information was provided to 23 examiners. RESULTS: The mean age of the study subjects was 51.5 years (SD +/- 17.7, range 23-81), with 28 women included. In 10 of the subjects, only gingivitis was identified, while 22 subjects had advanced chronic periodontitis. Risk scores assigned for 2 and 4 years differed significantly between European- and US-trained periodontists (p < 0.001) and between graduate students in training and periodontists from either the US or Europe (p < 0.01) (Wilcoxon n-pair test), with European periodontists scoring the lowest risks. Risk scores were correlated between groups (p < 0.01 with rho range 0.82-0.89) (Spearman's rank correlation). The best-fit model (r2 = 0.86) to assess perceived risk for worsening periodontal conditions based on data from all examiners combined included the following variables: (1) overall horizontal alveolar bone loss (p < 0.000), (2) age-adjusted proportional radiographic bone height score for the worst site (p < 0.000), and (3) proportion of pocket probing depths > or = 6.0 mm. CONCLUSIONS: Differences exist on the scale of risk values based on specialty training. Consistency in scoring patterns exists. The examiners based their assigned risk scores almost exclusively on measures of existing disease severity, including radiographic bone loss and numbers of periodontal pockets > or = 6.0 mm, and excluding most known risk factors such as smoking, diabetes, and poor oral hygiene.

Serum IgG to heat shock proteins and Porphyromonas gingivalis antigens in diabetic patients with periodontitis.

BACKGROUND: Past studies have reported a correlation between the presence and severity of periodontitis and serum antibody titers to species-specific antigens of Porphyromonas gingivalis or to cross-reactive antigens, such as lipopolysaccharide (LPS) and heat shock proteins (HSP), shared between P. gingivalis and other bacteria. Our recent study of periodontal treatment outcome in insulin-dependent (type 1) diabetes mellitus patients with severe periodontitis (IDDMI/periodontitis) resulted in two key findings: 1. serum glutamic acid decarboxylase autoantibody (GAD65 Ab) levels were significantly associated with periodontal pocket depth change (PDC) and 2. serum IgG titers to P. gingivalis cells were positively associated with GAD65 Ab level in seropositive (GAD65 Ab +) patients. We have therefore hypothesized that profiles of serum autoantibody levels and IgG titers, to P. gingivalis-specific antigens may be useful in assessing risk for refractory periodontitis in such patients. AIM: To determine whether PDC resulting from non-surgical periodontal treatment can be predicted using profiles of baseline IgG titers to P. gingivalis-specific antigens, human HSP, and GAD65. METHODS: PDC was assessed two months after non-surgical periodontal treatment of 7 GAD65 Ab + and 11 GAD65 AbIDDM/periodontitis patients. Pretreatment titers to GAD65, recombinant human heat shock proteins (HSP90, HSP70, and HSP60), and various P. gingivalis antigens were measured using radioligand precipitation or enzyme-linked immunosorbent (ELISA) assays and compared to the same measurements for 154 recent-onset IDDM patients and 46 non-diabetic controls. RESULTS: Median titers (ELISA units) to HSP90 and HSP70 were significantly higher than non-diabetic controls for GAD65 Ab + (p degrees = 0.002) and GAD65 Ab- (p = 0.034) IDDM/periodontitis patients, respectively. Multivariate regression analysis indicated significant partial correlation of PDC with log-transformed titers to HSP90 (r = - 0.62, p = 0.008), HSP70 (r = + 0.62, p = 0.009), GAD65 (r = - 0.60, p = 0.01) and P. gingivalis LPS (r = - 0.5 1, p = 0.04). Furthermore, hierarchical clustering of baseline profiles of log-transformed HSP90, HSP70, and GAD65 Ab titers sorted patients into two distinct clusters with significantly different median PDC (1.45 min, n = 10 vs. 0.65 min, n = 8; p = 0.016, Mann-Whitney). CONCLUSION: Pretreatment profiles of serum antibody titers to HSP90, HSP70, GAD65, and P. gingivalis LPS may be useful for predicting which patients with IDDM/periodontitis will have a poor response to non-surgical periodontal therapy.