Type 2 Diabetes as a Risk Indicator for Dental Caries in Korean Adults: The 2011-2012 Korea National Health and Nutrition Examination Survey

Abstract

Objectives:

The associations between type 2 diabetes (T2D) and dental caries was examined. This study hypothesized that there would be a positive association between T2D and the prevalence of decayed permanent teeth (DT) in representative Korean adults.

Methods:

The information was derived from the Korea National Health and Nutrition Examination Survey conducted in 2011–2012. Sociodemographic and lifestyle variables, anthropometric and biochemical status, metabolic health and glucose tolerance status, oral health behaviors, and dental caries index were evaluated.

Results:

The number of DT had positive association with degree of fasting plasma glucose (FPG) level, and glycated hemoglobin (HbA1c) (p-value = 0.045 and 0.007, respectively). The levels of FPG and HbA1c showed increasing pattern according to the number of DT (p for trend = 0.009 and 0.004, respectively). The prevalence of caries experience of uncontrolled T2D participants was about 26% higher with significance than those with normal glucose tolerance levels after adjustment of potential confounders including diets and socioeconomic status (OR [95% CI] = 1.26 [1.02, 1.56]).

Conclusions:

T2D is an independent risk indicator for dental caries in Korean adults.

Introduction

Diabetes mellitus (DM) has been a social and economic burden worldwide. International Diabetes Federation estimated that about 415 million people, or one in 11 adults worldwide suffer from diabetes mellitus in 2015, and this will rise to 642 million, or one in 10 adults by 2040. The burden of diabetic cost worldwide reached to $673 billion USD, 12% of the whole global health expenditure in 2015 (IDF, 2015). WHO reported that about 1.5 million deaths were directly caused by DM in 2012, and predicted that DM would be the 7^th top reason of mortality in 2030 (Alwan, 2011).

Dental caries is explained as the local breakdown of vulnerable tooth structure by acidic by-products originating from dietary carbohydrates fermented by bacterial species (Selwitz et al., 2007). The process of tooth decay begins within the bacterial biofilm (plaque), and proceeds to cavitation due to physiological imbalance between mineralized tooth portion and oral bacterial films. Dental caries is triggered by endogenous microbial components like Streptococcus mutans or lactobacilli, which was suggested as strong casual factors of dental caries (Selwitz, Ismail et al., 2007). Dental caries is a complex condition that is affected by other factors including a low socioeconomic status (SES), dietary habits, salivary flow rate, and fluoride exposure as well as bacteria (Schwendicke et al., 2015). Dental caries was reported to be associated with various systemic diseases like cardiovascular disease (Taubman and Nash, 2006), cerebrovascular disease (Ylöstalo et al., 2006), metabolic syndrome (Ojima et al., 2015), and so on. Over the past decade several reports have tried to explain the relationship between dental caries and DM; however, their findings have been controversial because of the opposite associations seen. Some researchers showed positive correlation between dental caries and DM (Arheiam and Omar, 2014; Biesbrock et al., 2003; Gæde et al., 2008), whereas others reported no association (Collin, H.-L. et al., 1998; Collin, H.L. et al., 1998), or even inverse association (Gupta et al., 2014).

In this study, we tried to identify a clear association between T2D and dental caries with a representative sample of Korean adults. This study hypothesized that T2D would increase the risk of dental caries.

Methods

Overview of the survey and participants

The records for the study was collected from the 2011–2012 Korea National Health and Nutrition Examination Survey, cross-sectional and a nationwide survey supervised by the Ministry of Health and Welfare of South Korea. Specially trained investigators inspected the representative population of South Korean adults with well-designed questionnaires including physical inspections, health interviews, and nutritional examination.

In this study, 10,439 participants aged≥30 and maintained fasting at least 8 hours prior to blood sample collection were included, and the 290 participants who had missing values in dental caries and 692 participants who had missing values in diabetes were removed from the study. Finally, 9,457 participants were selected for this study. All participants got their written informed consent. This study was permitted by the Institutional Review Board (IRB) of the Korean Center for Disease Control and Prevention, and was accomplished according to the Ethical Principles for Medical Research Involving Human Subjects based on Helsinki Declaration. This study was confirmed according to the STROBE guidelines, and presented as a flowchart (Fig 1).

Fig 1.

Flowchart of the study according to STROBE guideline.

Sociodemographic and lifestyle variables

Sociodemographic and lifestyle variables of the participants were collected with a self-administered questionnaire including those for physical exercise, cigarette smoking, household income, alcohol drinking, and education level. Smokers were categorized as non-smokers, ex-smokers, or current smokers. Alcohol drinking was categorized as heavy drinkers (> 30 g/day), mild to moderate drinkers (1–30 g/day), and non-drinkers. Education level was classified as either having graduated from high school (≥ 13 years) or not. Physical exercise rate was measured based on the International Physical Activity Questionnaire. Participants who do body exercise for 30 minutes/session and at least 5 times/week, or those who do physical exercise actively for 20 minutes/session and at least 3 times/week were demarcated as regular exercisers. Household income was divided into quartiles by the quantity of the family members.

Anthropometric measurements

Qualified trained examiners attained the data. Height was calculated to the nearest 0.1 cm, and body weight was recorded using a digital scale to the nearest 0.1 kg in bare feet and lightweight clothing. Waist circumference was estimated to the nearest 0.1 cm at the slenderest mid-point between the costal and the iliac crest margin, with loose clothing at the end of a normal expiration. BMI was measured by dividing body weight (kg) by the square of height (m²). These measurements were accomplished according to the recommendations of World Health Organization. Hypertension was well-defined as a systolic and/or a diastolic blood pressure steadily higher than 140/90 mm Hg. Prehypertension was demarcated as elevated blood pressure above normal but below hypertension as follows; a diastolic pressure 80–89 mm Hg or a systolic blood pressure 120–139 mm Hg. Neither hypertension nor prehypertension was designated as normotension.

Biochemical measurements

Trained staffs gathered biochemical samples from the participants. A standard mercury sphygmomanometer (Baumanometer, W. A. Baum Co., Copiague, NY, USA) was used to measure blood pressure. Gradations of diastolic and systolic blood pressures were read three times at five minute intervals, and averageing out were determined. Each participant was taken blood samples, which were assembled from the antecubital vein after fasting for at least eight hours. Samples were stored immediately at a low temperature, were then conveyed to the central testing institute (NeoDin Medical Institute, Seoul, South Korea). Low and high-density lipoprotein cholesterol, serum fasting plasma glucose (FPG), total cholesterol, and triglycerides were admeasured by an automated enzymatic analyzer (Hitachi 7600; Hitachi, Ltd., Tokyo, Japan). To evaluate liver function, serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) enzyme levels were admeasured with UV light by the aforementioned chemical analyzer (Hitachi 7600), and were classified as follows; abnormal (ALT≥35 unit/L, AST≥40 unit/L) or normal (ALT<35 unit/L, AST<40 unit/L).

Descriptions of metabolic syndrome and T2D

Data regarding T2D were collected from each participant based on blood samples of at least 8 h fasting. Participants with 100–125 mg/dL FPG were classified as having impaired fasting glucose, and participants with FPG level < 100 mg/dL were defined as normal glucose tolerance. Those previously diagnosed with T2D by a physician, or those who fulfilled the diabetic criteria provided from American Diabetic Association were defined as participants with T2D; FPG ≥ 126 mg/dL (ADVIA® 1650; Siemens, Deerfield, IL, USA) or ≥6.5% glycated hemoglobin (HbA1c) admeasured by liquid chromatography (Bio-Rad Varian™ II, Bio-Rad, Hercules, CA, USA) . Additionally, T2D being under control was defined as having HbA1c levels below 6.5% (48 mmol/mol) among the type 2 diabetic patients.

Metabolic syndrome in Asians was defined according to the scientific statement criteria of the American Heart Association/National Heart, Lung, and Blood Institute. The participants with at least three of the following 5 medical status were defined as having metabolic syndrome: waist circumference ≥ 80 cm for women and ≥ 90 cm for men, current use of an anti-hypertensive drug or blood pressure ≥ 130/85 mmHg, use of an anti-diabetic drug or FPG ≥ 100 mg/dL, use of an anti-dyslipidemic drug or fasting triglycerides ≥ 150 mg/dL, and use of an anti-dyslipidemic drug or high-density lipoprotein cholesterol < 50 mg/dL in women and < 40 mg/dL in men.

Oral health behaviors and dental caries index

Participants reported the times of day when they brushed their teeth from the choices before bedtime, after snacks, and before or after breakfast, lunch, and dinner. This study also determined whether secondary oral products such as dental gargle, dental floss, interdental brushes, and electric toothbrushes were used. The frequency of daily toothbrushing was defined as total number of toothbrushings per day. The self-reported oral health status and the number of dental visits within a year were also determined. The self-reported oral health status was categorized as “very good,” “good,” “normal,” “bad,” or “very bad.” We defined positive oral health status belonging to the “very good” and “good” categories.

Total caries experience, or decayed-missing-filled teeth (DMFT) index was defined by WHO criteria, as lifetime experience of dental caries in permanent dentition. DMFT index includes decayed teeth, missing teeth, and filled teeth. In this study, we used only decayed teeth because of these limitations of the DMFT index. The number of dental caries indicated the number of decayed teeth in one’s permanent dentition at the time of examination. The DT was detailed by the surface that affected by dental caries as being a buccal, distal, occlusal, mesial, or lingual surface.

Statistical analyses

The data are offered as percentage (standard error) for categorical variables and mean ± standard error for continuous variables. Rao–Scott chi-square tests for categorical variables and Student’s t-tests for continuous variables were used to compare the difference between those with decayed teeth and those without decayed teeth. Multiple linear regression test between the amount of DT and biochemical parameters were performed after adjustment of covariates including age and sex. We explored whether parameters of metabolic health status including FPG and HbA1c would differ according to the number of DT. In addition, multiple logistic regression tests were conducted to determine the odd ratios for experience of permanent dental caries according to the diabetic status (normal glucose tolerance, impaired fasting glucose, or T2D). Model 1 was a model without any adjustment. Model 2 was adjusted for sex and age. Model 3 was adjusted for components of Model 2 and smoking, drinking, and physical activity, education status, and household income. Model 4 was adjusted for the components of model 3 plus hypertension, metabolic syndrome, BMI, and number of daily toothbrushing. The SAS statistical software version 9.3 (SAS institute, Cary, NC, USA) was used for the analysis. All the data were statistically significant when the p- values were < 0.05.

Results

Table 1 displays the general features of including participants. The participants were subdivided by the presence or absence of dental caries. The prevalence of dental caries was significantly higher in male, heavy drinkers, present smokers, people who did not exercise regularly, and participants with household income in the lowest quartile. Individuals with annual dental visits, positive self-awareness of oral health status, and no mastication problems had significantly fewer DT (all p-value < 0.05). Having one DT was most common among participants with tooth decay, and those who had DT were more likely to belong to the higher age, weight, height, or waist circumference category (all p-value < 0.05).

Table 1.

General characteristics of the participants

	Presence of Decayed Permanent Teeth
	No	Yes	p-value*
No.	6,786	2,671
Sex (male)	46.8(0.6)	54.2(1.1)	<0.001
Present smoker (yes)	20.2(0.6)	29.4(1.1)	<0.001
Heavy drinker (yes)	9.6(0.5)	12.5(0.8)	0.001
Regular exerciser (yes)	18.8(0.7)	16(0.9)	0.01
Income (lowest quartile, yes)	15.2(0.7)	17.3(1)	0.04
Education level (≥ 13 years)	65.8(1)	67(1.3)	0.39
DT (n)
0	100(0)	.
1	.	47.8(1.3)
2	.	23.2(1)
3	.	12.3(0.8)
4 or more	.	16.8(0.9)
Annual dental visits (yes)	27.9(0.8)	22.3(1.1)	<0.001
Positive SROH (yes)	15.3(0.7)	8.3(0.7)	<0.001
Mastication problem (yes)	23(0.7)	33.6(1.2)	<0.001
Speech problem (yes)	8(0.4)	8.5(0.7)	0.54
Daily toothbrushing (n)			0.001
1	11.2(0.6)	13.6(0.8)
2	43.9(0.9)	46.3(1.2)
3 or more	44.8(0.9)	40.1(1.2)
T2D MED (yes)	6.7(0.4)	6.8(0.5)	0.83
MetS (yes)	31.5(0.8)	30.2(1.1)	0.34
Hypertension (yes)			0.63
Normotension	43(0.9)	41.7(1.3)
Prehypertension	26.4(0.7)	27.4(1.1)
Hypertension	30.6(0.8)	30.9(1.2)
Age (year)	50.9±0.3	48.7±0.4	<0.001
Height (cm)	162.8±0.1	163.8±0.2	<0.001
Weight (kg)	63.6±0.2	64.9±0.3	<0.001
WC (cm)	82±0.2	82.7±0.3	0.02
BMI (kg/m2)	23.9±0.1	24.1±0.1	0.01

The data are shown as percentage (standard error) for categorical variables and mean ± standard error for continuous variables.

^*P-values were gained by chi-square exam for categorical variables or independent t-test for continuous variables. Positive self-reported oral health includes “very good” and “good.” Hypertension was defined as >140/90 mmHg, prehypertension as 120–139 mm Hg of a systolic blood pressure, or 80–89 mm Hg of a diastolic pressure. Neither hypertension nor prehypertension was designated as normotension.

Abbreviation: DT; decayed permanent teeth individually, SROR; self-reported oral health, MetS; metabolic syndrome, WC; waist circumference. T2D MED; Present taking T2D medication or insulin injection

Table 2 shows the associations between the numbers of DT and biochemical / anthropometric parameters by multiple linear regression analysis. The number of DT per individual showed positive association with FPG level and HbA1c (p-value = 0.045 and 0.007, respectively) following adjustment for age and sex. However, other variables did not show significance.

Table 2.

Association between number of decayed permanent teeth and biochemical/ anthropometric parameters

	Individual decayed permanent teeth (n)
	Beta	SE	p-value
WC (cm)	0.00204	0.00286	0.48
BMI (kg/m2)	0.00391	0.00804	0.63
DBP (mm/Hg)	−0.00160	0.00215	0.46
SBP (mm/Hg)	0.00252	0.00148	0.09
FPG (mg/dL)	0.00276	0.00137	0.045
HbA1c (%, mmol/mol)	0.09229	0.03414	0.007
TC (mg/dL)	0.00009	0.00059	0.88
LDL-C (mg/dL)	0.00002	0.00070	0.98
HDL-C (mg/dL)	−0.00106	0.00168	0.53
TG (mg/dL)	0.01029	0.03282	0.75
ALT (unit/L)	−0.02124	0.04910	0.67
AST (unit/L)	0.00253	0.06509	0.97

Data were analyzed by multiple regression model, and adjusted for the covariates including age and sex.

^*A p-value < 0.05 designated statistical significance.

Abbreviations: Beta; beta coefficient, SE; standard error, WC; waist circumference, BMI; body mass index, DBP; diastolic blood pressure, SBP; systolic blood pressure, FPG; fasting plasma glucose, HbA1c; glycated hemoglobin, TC; total cholesterol, LDL-C; low density lipoprotein cholesterol, HDL-C; high density lipoprotein cholesterol, TG; serum triglyceride, ALT; alanine transaminase, AST; aspartate aminotransferase.

The patterns of glucose intolerance by the number of DT are shown in Fig 2. The levels of FPG and HbA1c displayed increasing pattern according to the number of DT (p for trend = 0.009 and 0.004, respectively).

Fig 2.

The patterns of glucose intolerance according to the amount of permanent teeth with decay (DT). The values of fasting plasma glucose (A) and glycated hemoglobin (HbA1c) (B) were elevated gradually according to the increasing quantity of DT (p for trend = 0.009 and 0.004, respectively).

Table 3 presents data of the prevalence of caries (DT) experience according to T2D control and glucose tolerance. The prevalence of caries experience of uncontrolled T2D participants was about 26% higher with significance than those with normal glucose tolerance levels after adjustment of potential confounders including diets and SES (OR [95% CI] = 1.26 [1.02, 1.56]), and this prevalence was higher than controlled T2D (1.2 [0.84, 1.72]).

Table 3.

Prevalence of caries experience according to the values of fasting plasma glucose and T2D control

	Prevalence of caries experience
	MODEL1	MODEL2	MODEL3	MODEL4
Glucose tolerance status
Normal	1	1	1	1
IFG	1.00(0.88,1.13)	1.1(0.97,1.25)	1.08(0.95,1.23)	1.03(0.89,1.19)
T2D	1.24(1.04,1.47)	1.49(1.22,1.81)	1.42(1.17,1.73)	1.26(1.02,1.56)
T2D control (yes)	1.17(0.82,1.67)	1.23(0.86,1.76)	1.19(0.84,1.7)	1.2(0.84,1.72)

Abbreviations: T2D; T2D. IFG; impaired fasting glucose, FPG; fasting plasma glucose level for at least 8 h of fasting. Data are designated as odds ratio (95% confidence intervals). Control of T2D was defined by glycated hemoglobin (HbA1c), as below 6.5% (48 mmol/mol) was controlled T2D among diabetic patients.

MODEL1 was non-adjusted, MODEL2 was adjusted for sex and age. MODEL3 was adjusted for covariates of Model 2 plus drinking, smoking, household income, physical exercise, place of residence, and education level. MODEL4 was adjusted for covariates of Model 3 plus BMI, number of daily toothbrushings, hypertension, total energy intake, and carbohydrate intake (%).

Discussion

This study revealed that tooth decay is positively associated with T2D. The number of DT was significantly correlated with FPG and HbA1c levels. Prevalence of caries experience was much higher in T2D participants than those of normal and participants with impaired fasting glucose. Moreover, increasing pattern of T2D diagnostic parameters (FPG and HbA1c) according to the number of DT were clearly demonstrated.

The link between dental caries and diabetes has been sought in several animal studies (Kodama et al., 2011; Selwitz, Ismail et al., 2007). One in vivo study revealed that the occurrence of dental caries increased in diabetic rats compared to healthy group, and insulin-mediated glycemic control decreased the caries-related periodontal infection and progression of dental caries (Nakahara et al., 2013). Diabetes was also closely associated with numerous infectious diseases, supported by the fact that diabetic animals were prone to fungal and bacterial infection (Nakahara, Sano et al., 2013). The present study demonstrated that odds ratio for the prevalence of caries experience slightly decreased in case of T2D control compared to T2D without control. Because dental caries is caused by oral cariogenic bacteria (Kleinberg, 2002), we could assume that diabetic control might diminish the incidence of tooth decay.

The associations between dental caries and diabetes, however, are still controversial from clinical studies. Some reports showed no difference in tooth decay between non-diabetic and T2D individuals (Cherry-Peppers and Ship, 1993; Collin, H.L., Uusitupa, M. et al., 1998). However, higher prevalence of dental caries was shown in type 1 diabetic patients (Miko et al., 2010; Siudikiene et al., 2006), and type 2 diabetic patients (Biesbrock, Bartizek et al., 2003). This cross-sectional study also found that significantly high prevalence of caries on the crown portion in the T2D patients.

The potential confounders which may affect both T2D and dental caries should be considered, because these confounders would cloud the exact relationship between two factors. Dental caries and T2D may share common risk factors including diet and SES.

Dental caries and T2D may increase in prevalence if excessive carbohydrate consumed. A previous report showed that frequent and/or excessive sugar intake may be attributed a key role in caries propagation, and streptococcus mutans has a principle role in demineralizing enamel by metabolizing carbohydrate to lactic acid (Bradshaw and Lynch, 2013). WHO suggested a guideline that consumption of free sugar should not exceed 5% of total energy intake to prevent dental caries both in adults and children (Rugg-Gunn, 2013). Similarly, T2D may increase if excessive carbohydrate intake. A randomized clinical trial indicated that a very low carbohydrate improved glucose control in T2D (Saslow et al., 2014). Recently, American Diabetes Association recommended that a carbohydrate reduction less than 130g/day to lower postprandial glucose, saturated fat less than 7% of total calories intake, and protein intake up to 20% of the energy intake (Khazrai et al., 2014).

Both risk of dental caries and T2D may increase if low SES experienced. A systematic review demonstrated that the prevalence of caries experience was significantly higher in those with low educational or occupational status, or low income level. (Schwendicke, Dorfer et al., 2015). Also, low SES may increase cariogenic bacteria and secretion of stress-evoked salivary cortisol, which would results in dental caries susceptibility (Boyce et al., 2010). As for T2D, a systematic review of prospective cohorts showed close association between T2D and low SES group (Kivimaki et al., 2015). Likewise, a women cohort study showed that both income and education level was inversely associated with risk of T2D. Several cross-sectional study revealed that low income and low education level are the independent risk factor for T2D (Asadi-Lari et al., 2016; Bird et al., 2015; Hwang and Shon, 2014). Accordingly, risk of both T2D and dental caries increased in case of excessive consumption of carbohydrate, and/or low SES. However, this study clearly showed that dental caries is an independent risk indicator for T2D after adjusting the effect of diet as well as SES, and exact casual factors are remained to elucidate.

Salivary dysfunction can be a clue of association between dental caries and T2D. Saliva has a protective role against dental caries through its unique components of pH, calcium, phosphate, fluoride, and the rinsing effect by salivary flow (Stookey, 2008). Elevation of mineral contents in saliva inhibited caries incidence (Pearce et al., 2002), and saturation of saliva with calcium and phosphate reduced the decay of the tooth surface by a continuous ion exchange mechanism (Ten Cate, 2008). As such, an altered salivary function in T2D was suggested as the mechanism of caries susceptibility. One study reported that high blood glucose level led to a high glucose content in saliva (Tenovuo et al., 1986). Also, decreased pH and salivary flow rate, and altered protein composition were suggested as causes for caries susceptibility in diabetic individuals (Sampaio et al., 2011). One comparative clinical study between non-diabetic and T2D participants revealed that DMFT index was significantly higher in T2D participants, and the salivary flow rate and mineral components including fluoride, phosphate, and calcium were significantly lower in the T2D patients (Jawed et al., 2011). Similarly, lack of salivary flow with acidic change could aggravate the tooth decay (Chu et al., 2008), whereas participants with controlled diabetes had lower DT, better salivary flow rate, and pH (Kumar and Clark, 2002). Collectively, these results indicated that salivary dysfunction caused by T2D may lead to the onset and/or propagation of dental caries.

A previous large population-based longitudinal study revealed that HbA1c was associated with diabetic risk as well as FPG, and was more powerfully related to the mortality (Selvin et al., 2010). The 2015 ADA suggested that HbA1c of 5.7 to 6.4% (39–46 mmol/mol) was suitable as a gauge for pre-diabetes, and that of > 6.0% (42 mmol/mol) should be of concern for a “very high” risk of diabetes (ADA, 2015). Similarly, other reports have shown that adults with HbA1c >13% (119 mmol/mol) had significantly higher DMFT index than those with HbA1c <10% (86 mmol/mol) (Gæde, Lund-Andersen et al., 2008). Another study reported that individuals with HbA1c >8.5% (69 mmol/mol) was significantly associated with higher level of DMFT index than individuals with HbA1c <8.5% (69 mmol/mol) (Syrjala et al., 2003). The present study found that the whole set of surveyed participants had an HbA1c of 5.7% (39 mmol/mol) or more, and this value reached up to almost 6.0% (42 mmol/mol) in those who had more than 4 DT. These results emphasized that participants with a higher number of DT should concern more about “very high” risk of diabetes.

There were some limitations in this study. First, this study didn’t elucidate the causal relationship between dental caries and T2D as the present study was based on cross-sectional observational samples. Additional longitudinal cohort would be required to reveal the exact role of caries risk in T2D. Second, this study didn’t address some influencing factors such as changes in salivary protein composition or flow rate, bacterial/yeast components on saliva or plaque, and outcome variables including root caries incidence. Third, this study couldn’t differentiate between type 1 and 2 participants, because records from KNHANES didn’t include the diagnostic age which is the criteria differentiating between the two, although majority of the patients would be T2D patients. However, this study also have confidences as follows. First, this study showed close association between T2D and dental caries in adults as well as infants or children. Most of the previous studies focused on caries risk of young individuals, therefore lack of adult data. Second, the present study showed that control of T2D would benefit from low caries risk. Second, this study firstly showed that dental caries is an independent risk indicator of T2D after adjustment of both diet and SES. Our findings would suggest a role of diabetes control in maintain healthy oral health status without dental caries.

Conclusion

T2D is an independent risk indicator for dental caries in Korean adults. These findings emphasize an importance of diabetes control to prevent dental caries and maintain oral health.

Abbreviations:

FPG

fasting plasma glucose

HbA1c

glycated hemoglobin

DT

the number of decayed permanent teeth

HbA1c

glycated hemoglobin

IFG

impaired fasting glucose

KNHANES

Korean National Health and Nutrition Examination Survey