Association between diabetes mellitus and pre-malignant oral diseases: A cross sectional study in Kerala, India
Recent studies have suggested role of diabetes mellitus in the development of periodontal diseases, oral inflammatory and neoplastic lesions. We analysed data from randomised oral cancer screening trial in Kerala, India, to study the association between chronic diseases in general, diabetes mellitus in particular, and pre-malignant oral lesions and conditions. A total of 927 subjects with oral leukoplakia, 170 with oral submucous fibrosis, 100 with erythroplakia lesions and 47,773 apparently healthy subjects constituted the cases and controls respectively for our present study. The odds ratio (OR) for oral leukoplakia among women with a history of diabetes mellitus was found to be 2.0 (95% confidence interval (CI) = 1.4–2.9) whereas that for erythroplakia was 3.2 (95% CI = 1.3–7.9), after adjusting for potential confounders. No statistically significant association between diabetes mellitus and oral submucous fibrosis and pre-cancerous lesions/conditions in men was found. An association between diabetes mellitus and pre-malignant oral lesions among women has been observed in our study, although the underlying mechanisms are not clear. This needs to be further evaluated in other settings. © 2005 Wiley-Liss, Inc.
Betel quid chewing with or without tobacco, smoking and alcohol drinking are well-established risk factors for pre-malignant oral diseases.1 Recently attempts have also been made to investigate an association between clinical variables and risk of pre-malignant oral diseases. Past history of diabetes has been linked with development of pre-malignant oral diseases in a few studies.2, 3 An association between diabetes mellitus and periodontal and oral diseases, and various inflammatory lesions in oral cavity has also been demonstrated.4 An association between diabetes mellitus and cancers of the colon,5 endometrium,6 kidney,7 pancreas8 and stomach9 has been reported in recent literature. Our study addresses the association between chronic diseases, including diabetes mellitus and oral leukoplakia, erythroplakia and submucous fibrosis.
Our present study is based on the data collected from the first round of intervention from a randomised oral cancer screening trial in Kerala, India.10 We have reported previously the role of chewing betel quid with and without tobacco, alcohol, body mass index (BMI) and socio-economic status in development of pre-malignant oral lesions using the same data set.11, 12, 13, 14, 15 We considered a total of 927 leukoplakia, 100 erythroplakia and 170 submucous fibrosis cases and 47,773 controls to assess the role of various chronic diseases, diabetes mellitus in particular, in the development of pre-malignant oral lesions. To our knowledge, this is the first large study on this topic.
Material and methods
A cluster randomised, controlled oral cancer screening trial was conducted in the Trivandrum district, India, to evaluate the efficacy of screening in reducing oral cancer mortality. Details of study design and population for this trial are described previously.10 A case control study design was used to evaluate the role of chronic diseases in the development of pre-malignant oral lesions. Oral pre-malignant lesions considered in this study were oral leukoplakia, oral submucous fibrosis and erythroplakia. Oral leukoplakia, is a non-removable white patch in the oral mucosa and is the most common type of oral pre-malignant lesion. Oral submucous fibrosis is an irreversible, pre-malignant condition with a moderate malignant potential in between that of oral leukoplakia and erythroplakia. Erythroplakia is a red patch lesion that is considered the most severe oral patch lesion with high malignant potential. Data from the first round of intervention of the ongoing screening trial was used. The analysis included only the intervention group because subjects from this group had been screened with oral visual inspections and their disease status was known. The intervention group included a total of 59,894 eligible subjects who lived in one of 7 clusters, where the total resident population was 172,567. In the first round of intervention, 49,174 subjects were interviewed and screened in their houses by trained health workers. A total of 927 oral leukoplakia, 100 erythroplakia, 170 submucous fibrosis cases and 47,773 controls were identified from this intervention group, for our study.
The trained health workers interviewed the subjects about demographic information, tobacco chewing, smoking, alcohol drinking and history of chronic diseases before undertaking screening. After the interview, blood pressure, body weight and height were measured. Data on past medical history were also obtained. The interviewer asked about the history of diabetes, chronic bronchitis, bronchial asthma, tuberculosis, high blood pressure, heart attack, stroke and cancer. The health workers then carried out oral visual screening to identify positive subjects with lesions suggestive of oral leukoplakia, erythroplakia, submucous fibrosis or oral cancer by inspecting the buccal and labial mucosa, gingivae, bucco alveolar sulci, tongue, palate and the floor of mouth. Subjects with positive or suspicious findings were referred to dentists and oncologists, who made the final diagnosis.
Logistic regression was used to estimate odds ratios (OR) and their 95% confidence intervals (95% CI). History of chronic diseases was dichotomised into no history of disease and positive history of disease under investigation. We adjusted the risk estimates for the following factors: age (continuous), gender, education (categorical: none and illiterate, none and literate, primary, middle, ≥ high school), BMI (continuous, kg/m2), smoking (continuous, pack years), chewing betel quid with or without tobacco (continuous, duration in years) and alcohol drinking (continuous, duration in years). BMI and education was shown to be related with pre-malignant diseases in previous studies11, 12 using the data of the same trial and as they may also be related with history of chronic diseases we treated them as a potential confounder in the analysis. To further investigate the role of diabetes in the development of leukoplakia and erythroplakia according to obesity, we divided obesity into underweight (<18.5 kg/m2), normal range (18.5–24.9 kg/m2) and overweight (≥ 25.0 kg/m2). Statistical analyses were carried out with the STATA 8.0 statistical software package.
The distribution of demographic and lifestyle factors among the oral pre-malignant cases and controls have been published previously.11, 12, 13, 14, 15 Table I shows the distribution of history of chronic diseases among pre-malignant cases and controls stratified by gender. Prevalence of high blood pressure was observed to be 15.3% followed by bronchial asthma (11.8%) and diabetes (6.5%), in the control population. Number of pre-malignant cases with history of tuberculosis, heart attack and stroke was <5. Further analysis for individuals with a positive history of aforesaid conditions could not be done.
Table I. Distribution of Pre-Malignant Cases and Controls According to History of Chronic Diseases
| Yes||34||8.3||31||6.0||3||2.2||0|| ||6||12.2||6||11.8||1,590||5.3||1,472||8.2|
|High blood pressure|
| Yes||0||–||3||0.6||0||–||0|| ||1||2.0||0|| ||49||0.2||94||0.5|
Table II presents OR for pre-malignant oral diseases according to history of diabetes mellitus. The OR for developing leukoplakia was 2-fold whereas that for developing erythroplakia was observed to be more than 3 times among women with a past history of diabetes in comparison to those without having a history of diabetes, after adjustment for age, pack-years of smoking, duration of chewing, alcohol drinking, education and BMI. No statistically significant excess risk, however, could be observed for men. Similarly no increase risk for submucous fibrosis among men and women was observed for individuals with positive history of diabetes. When the analysis was restricted to non-smokers, odds ratio for leukoplakia and erythroplakia among women with history of diabetes was still very high (data not shown). When we repeated the analysis for non-chewers, OR for leukoplakia was observed to be 3.1 (95% CI = 1.1–8.4) whereas that for erythroplakia was observed to be 8.0 (95% CI = 1.4–45.0) for women with positive history of diabetes (data not tabulated).
Table II. History of Diabetes Mellitus and Risk of Pre-Malignant Oral Diseases
|Yes||0.7 (0.5–1.0)||0.9 (0.6–1.4)||1.3 (0.9–1.9)||2.0 (1.4–2.9)||1.0 (0.7–1.2)||1.4 (1.0–1.8)|
|Yes||1.5 (0.6–3.4)||2.1 (0.9–5.3)||2.2 (0.9–5.2)||3.2 (1.3–7.9)||1.8 (1.0–3.4)||2.7 (1.4–5.1)|
|Yes||NE4||NE4||0.3 (0.1–1.1)||0.5 (0.2–1.7)||0.2 (.07–0.7)||0.4 (0.1–1.3)|
Risk of leukoplakia and erythroplakia according to history of diabetes, stratified by obesity is presented in Table III. The OR were significantly high for underweight and overweight women with a positive history of diabetes. No statistically significant increase in risk associated with a positive history of diabetes was, however, observed for women with normal weight.
Table III. Risk of Pre-Malignant Oral Diseases According to History of Diabetes Mellitus, Stratified by Obesity
|Underweight||No||4931/7310||169||1.0|| ||155||1.0|| ||9||1.0|| ||18||1.0|| |
|Normal range||No||9250/14938||274||1.0|| ||187||1.0|| ||30||1.0|| ||22||1.0|| |
|(BMI = 18.5–24.9 kg/m2)||Yes||1013/911||18||0.7||0.5–1.2||16||1.3||0.8–2.3||3||1.8||0.5–6.1||2||1.5||0.3–6.5|
|Overweight||No||2083/5996||38||1.0|| ||35||1.0|| ||6||1.0|| ||3||1.0|| |
|(BMI ≥ 25.0 kg/m2)||Yes||296/557||5||0.8||0.3–2.2||12||3.6||1.8–7.1||2||3.0||0.5–16.8||2||13.8||1.5–129.6|
When we stratified with pack-years of smoking, the OR for erythroplakia were observed to be 0.8 (95% CI = 0.1–6.1), 4.7 (95% CI = 1.2–17.8) and 2.1 (95% CI = 0.5–9.5), for nonsmokers, for 1–19 pack-years of smoking and more than 19 pack-years of smoking respectively, among males with history of diabetes compared to those without any such history. On stratifying by duration of betel quid use, observed OR for erythroplakia among individuals with history of diabetes were 1.7 (95% CI = 0.2–14.1), 2.4 (95% CI = 0.5–10.9), and 2.8 (95% CI = 0.8–9.7) among males for non-quid users, up to 19 years of quid use and for use of quid for more than 19 years respectively. For females OR could not be estimated for first 2 strata of duration of quid use because one of the cell contained zero.; However, OR were 7.1 (95% CI = 2.7–18.9) for erythroplakia among women with duration of quid use for more than 19 years and a history of diabetes compared to those with no history of diabetes. All these estimates were adjusted for potential confounders (data not tabulated).
Table IV shows risk of pre-malignant oral diseases according to positive history of other chronic diseases. No statistically significant excess risk for erythroplakia and submucous fibrosis among men and women was associated with a positive history of chronic bronchitis, bronchial asthma or high blood pressure. Statistically significant increased risk for leukoplakia was observed for women with positive history of chronic bronchitis and high blood pressure. When we further adjusted for history of diabetes, however, we could not observe the excess risk of leukoplakia among women with a history of chronic bronchitis (OR = 0.9; 95% CI = 0.1–7.3) and high blood pressure (OR = 1.3; 95% CI = 0.9–1.9). Risk of leukoplakia among women was still significantly high for women with a positive history of diabetes even after including chronic bronchitis or high blood pressure in the model (data not shown). Further, when we restricted analysis to non-chewers or non-smokers, risk of leukoplakia for women with a positive history of high blood pressure, was not observed (data not shown).
Table IV. History of Other Chronic Diseases and Risk of Pre-Malignant Oral Diseases
|Chronic bronchitis4||1.5 (1.2–1.9)||1.1 (0.8–1.4)||3.1 (2.0–5.0)||1.8 (1.0–3.1)||0.9 (0.3–2.3)||0.9 (0.3–2.3)||1.3 (0.2–9.8)||0.9 (0.1–7.3)||NE3||NE3||0.5 (0.06–3.3)||0.3 (0.03–2.2)|
|Bronchial asthma5||1.0 (0.8–1.3)||0.8 (0.6–1.1)||1.4 (1.0–1.8)||1.2 (0.9–1.6)||1.7 (0.8–3.5)||1.2 (0.6–2.6)||1.8 (0.8–3.7)||1.6 (0.8–3.4)||0.9 (0.3–3.0)||0.7 (0.2–2.4)||1.0 (0.6–1.7)||0.9(0.5–1.5)|
|High blood-pressure6||0.8 (0.6–1.1)||1.0 (0.7–1.3)||1.1 (0.9–1.5)||1.5 (1.1–1.9)||0.9 (0.3–2.3)||1.2 (0.4–3.1)||0.9 (0.5–2.0)||1.2 (0.6–2.6)||1.4 (0.5–4.1)||2.3 (0.8–7.3)||0.6 (0.3–1.0)||0.8 (0.4–1.3)|
We found diabetes as an independent risk factor of oral leukoplakia and erythroplakia for women. The OR were >2-fold and statistically significant at 5% level after adjustment for potential confounders. The risk was consistently high among women, even when we restricted analysis among non-smokers or non-chewers. Dietrich et al.2 also observed a >2-fold increase in risk of oral leukoplakia among individuals with positive history of diabetes in a study conducted among participants in the United States National Health and Nutrition Examination survey2 after adjusting for potential confounders. Another study by Albrecht et al.16 also reported prevalence of oral leukoplakia to be 6.2% among diabetics as compared to 2.2% in control group. A marginal excess of mouth and pharynx cancer was observed in a cohort of patients hospitalised with diabetes mellitus in Denmark. The result of this study reflects mainly excess associated with non-insulin dependent diabetes mellitus. The estimates, however, were not adjusted for potential confounders.17
The strengths of our present study are large sample size, ascertainment of cases by dentist/oncologist and population based design from a randomised oral cancer screening trial. Bias due to differential selection of cases and control is therefore very unlikely. The data on socio-demographic factors, personal habits and history of chronic diseases were obtained before the workers carried out oral visual inspection thus avoiding potential biases in data between cases and controls. Trained health worker conducted interviews with utmost care for interviewing on all study variables. Possibility of bias due to inability to recall past chronic events such as diabetes, chronic bronchitis, high blood pressure, tuberculosis, heart attack and stroke is minimal as these conditions required constant medical supervision and treatment for a long period of time.
Our results of an independent effect of diabetes on risk of leukoplakia and erythroplakia, do not conflict with present knowledge and are biologically plausible. Many possible mechanisms for an association between diabetes and pre-malignant oral lesions may be proposed. Diabetes (type II) is usually associated with insulin resistance and increased pancreatic secretion. Chronically increased levels of insulin resulting in hyperinsulinaemia, have been associated with colon cancer pathogenesis and with cancer of breast, pancreas and endometrium. These tumourogenic effects of insulin could be directly mediated by insulin receptors in the pre-neoplastic target cells or might be due to related changes in endogenous hormone metabolism. For example insulin promotes the synthesis and biological activity of insulin like growth factors 1 (IGF1). Insulin and IGF1 act as growth factors that promote cell proliferation and inhibits apoptosis.18 Prospective and retrospective studies have demonstrated that serum IGF-1 levels are associated with increased risk of various epithelial tumours including prostate, breast, colorectal, lung19 and oesophagus.20 Positive association between IGF1 in primary oral cancer and stage of disease has also been observed.21 Limited evidence suggests that the effect of IGF-1 might also be related to p53 mutations, which are quite common in head and neck tumours. Certain type of p53 mutation will lead to p53 overexpression, which in turn will upregulate IGF receptor I (IGF-IR) expression.22 The increased expression of IGF-IR will allow for more IGF-1/IGF-IR binding, thereby modulating cell proliferation.
It is also possible that a positive association between diabetes mellitus and pre-malignant lesion observed in the present study might occur due to shared risk factors. In particular, both may reflect some common exposure such as diet high in fat and energy and low in dietary fibres. It has been suggested that increase consumption of saturated fat increases the risk of oral cancer.23 Further, association between betel quid chewing, areca nut chewing and diabetes24, 25 has been demonstrated. Smokeless tobacco use may also affect BMI,26 which is a strong predictor of diabetes. We have adjusted for smoking, quid chewing (with and without tobacco) and BMI in the analysis. To rule out residual confounding due to quid chewing and smoking, we repeated the analysis separately for non-chewers and non-smokers. The OR for leukoplakia and erythroplakia were still statistically significantly higher among women with history of diabetes compared to those without any diabetic history. When we stratified by obesity, we could observe the effect of diabetes among underweight and overweight women but not in women with weight within normal range (based on BMI).
In our study positive association with diabetes was statistically significant only for women. Among men the OR were >2 for erythroplakia, although not statistically significant (OR = 2.1, 95% CI = 0.9–5.3). Further, when stratified by obesity the OR were much higher in underweight and overweight men, a trend similar to that observed for females. When we stratified by duration of quid chewing the odds ratios for erythroplakia were similarly higher in the highest categories among males and females. The higher OR among males and females in highest category of duration of quid chewing could be explained, as increased level of IGF-1 inhibits apoptosis of damaged cells leading to clonal outgrowth of genetically damaged populations.18
Absence of statistically significant association in men, in contrast to females, for the development of pre-malignant oral lesions is difficult to explain. It is possible that males and females reported history of diabetes differently. It may be speculated that the management of diabetes among males and females is different. Women are likely to represent patients with poorer metabolic control, leading to higher levels of insulin, and probably more oxidative damage to DNA. It has been proposed that poor diabetic control is associated with an increased cancer risk due to enhanced oxidative damage to DNA.27 Few studies have similarly observed effect of diabetes on kidney cancer, non-melanoma skin cancer28 and colon cancer29 only among women. The gender difference observed for the effect of diabetes mellitus on risk of pre-malignant oral diseases in our present study and for other cancer sites in other studies remains unexplained.
Our findings on risk of pre-malignant diseases in relation to positive history of diabetes have certain limitations. We do not have information on age at onset of diabetes. Because this was a cross sectional study, we do not have date of diagnosis for all pre-malignant oral lesions. This precluded quantification of risk stratified by age at onset of diabetes and according to duration between onset of diabetes and development of pre-malignant oral lesions. Oral lesions and diabetes are chronic and because of the cross sectional nature of the study we could not establish that diabetes preceded the development of oral lesions. We do not have any reason, however, to believe that pre-malignant oral lesions will lead to development of diabetes. Further, we do not have information on the type of diabetes (insulin dependent or non-insulin dependent) or treatment for diabetes (only dietary advice, oral agents or insulin). We cannot determine whether the present association is related to characteristics of the diabetic state or is related to treatment. We are planning a case-control design nested within the cohort of ongoing randomised oral cancer screening trial to address these issues.
We also observed a statistically significant increase risk of oral leukoplakia among women with a history of high blood pressure. When analysis was restricted to non-smokers or non-chewers, however, we could not observe the similar effect. This suggests that observed risk could be due possibly to residual confounding of smoking.
In summary, we demonstrated an increased risk of developing leukoplakia and erythroplakia but not for submucous fibrosis, in relation to positive history of diabetes among women. Positive history of diabetes was also associated with statistically insignificant increase risk for developing erythroplakia among men. Because leukoplakia and erythroplakia are considered to be pre-malignant with a high malignant potential, association between positive history of diabetes and oral cancer could also be hypothesized. Further studies with additional details about diabetes history or conducted in other settings are required to confirm this observed association as a causal.
The Trivandrum oral cancer-screening trial is supported by a generous grant from the Association for International Cancer Research (AICR), St. Andrews, Scotland. Additional data collection in our study is supported by Imperial Cancer Research Fund (ICRF), United Kingdom. Dr. R. Dikshit worked on our study under the tenure of an American Cancer Society Beginning Investigators award from UICC.