Effectiveness of Salivary Glucose as a Reliable Alternative in Diagnosis of Type 1 Diabetes Mellitus: A Cross-Sectional Study.

Aims and Objective: The aim of the present study was to ascertain the reliability of using salivary glucose in diagnosing and monitoring the blood glucose levels in type 1 diabetic patients and also to evaluate the salivary glucose in fasting and postprandial states and correlate with the blood glucose levels and hemoglobin A1c values. Materials and
Methods: The study was conducted on 79 Type 1 diabetic patients and 100 healthy controls. The blood and salivary glucose levels were evaluated in both study and control groups using glucose-oxidase peroxidase method.
Results: There was significantly higher level of salivary glucose in diabetics than in controls. There was a positive moderate-to-strong correlation between stimulated and unstimulated salivary glucose to blood glucose in both fasting and postprandial states. The cutoff values for stimulated and unstimulated fasting salivary glucose was 5.8 mg/dl and 5.6 mg/dl, respectively. The cutoff value for stimulated and unstimulated salivary glucose in postprandial state was 10.3 mg/dl and 9.5 mg/dl, respectively.
Conclusion: Salivary glucose appears to be a reliable biofluid to assess the blood glucose levels and can be a reliable alternative to blood glucose in Type 1 diabetes mellitus patients. Copyright:

INTRODUCTION

Loss of insulin-producing beta cells in the pancreas is characteristic of Type 1 diabetes, which is an autoimmune disease. This destruction in beta cells causes insulin insufficiency and patient develops hyperglycemia eventually, leading to weight loss, polyuria, and polydipsia. Children <5 years of age are most commonly affected by Type 1 diabetes. There is a male predilection for Type 1 diabetes, unlike other immune diseases.[1] Various complications have been documented due to increased blood glucose levels; therefore, regular plasma glucose monitoring helps prevent further complications occurring due to poor disease control. The diagnosis and monitoring of the disease is mainly through blood-based investigation involving invasive procedures. Needle anxiety is frequently encountered by the patient, thereby compromising regular sugar monitoring.[2] Saliva transcends the above issues because of its noninvasiveness, ease, and low cost of collection.[3]

Hyperglycemia and associated endothelial dysfunction are responsible for acute and chronic complications in Type I diabetes patients. The blood flow to exocrine glands such as salivary glands primarily determines its secretory activity, which leads to the large amount of glucose concentration in saliva of patients with diabetes mellitus.[4]

Earlier studies have shown a positive correlation between salivary and blood glucose levels.[5] However, all studies done till now in Type I diabetes mellitus to assess the salivary glucose have been done in unstimulated saliva and mostly all studies were performed during fasting conditions. The limitation of all these studies mentioned in the literature was small sample size, no uniformity in the methodology in sample collection, and also no studies showed any values of both stimulated and unstimulated salivary samples in Type I diabetes mellitus.[6] Our study is done with a larger sample size, and also various other parameters were added to overcome all the pitfalls of the previous studies.

The aim of the present study is to assess the reliability of using salivary glucose in diagnosing and monitoring the blood glucose levels in Type I diabetes patients. The salivary glucose levels were evaluated in fasting and postprandial states and were correlated with the blood glucose levels and hemoglobin (HbA1c) values, thereby showing that salivary glucose can be a much reliable alternative to blood glucose in Type I diabetes mellitus patient.

MATERIALS AND METHODS

The present cross-sectional study was conducted in the outpatient department of a premier tertiary care diabetic center in Chennai, Tamil Nadu. All participants signed the informed consent for the study, and the study was approved by the Institutional Ethics Committee (003/09/2017/IEC/SU). A total of 179 participants were included in the study included 100 nondiabetic individuals (sex and age matched) – Group 2 and 79 diagnosed cases of Type 1 diabetes– Group 1. Patients with a history of tobacco and alcohol use, individuals with any other type of diabetes other than Type I Diabetes mellitus, history of salivary gland diseases, previous salivary gland surgery, patients on radiotherapy to the head-and-neck region, and history of xerostomia were excluded from the study.

The blood and salivary samples from each patient in the control and study group were taken. Following an overnight fasting of 6–8 h, the blood samples were first collected. For saliva collection, the subjects were asked to rinse the mouth thoroughly with 150 ml of water and then sit erect with the head slightly down. Standard spitting method was employed to collect 3 ml of unstimulated whole saliva into a sterile container which was then centrifuged for 15 min at 3000 rpm, following which the supernatant was stored at −20°C.

A small amount of 0.1–0.2 mol/L citric acid was applied on either side of the dorsal surface of the tongue following which 3 ml of stimulated whole saliva was collected using a sterile cup. The collected samples were treated in the same way as unstimulated saliva and then stored at 20°C until testing and test was performed immediately without any time delay.

For blood samples, 5 ml of intravenous blood was obtained under aseptic conditions from the median cubital vein; 2 ml of the collected blood was stored in a blood collecting tube containing ethylene diamine tetraacetic acid. The remaining blood was centrifuged at 3500 rpm for 10 min in a sterile glass tube, and the serum was stored at −20°C until analysis. All patients' blood samples collected in Ethylene diamine tetraacetic acid (EDTA) tubes were subjected to HbA1c estimation using ion exchange resin method. The blood and salivary samples were collected in the same way after 2 hafter food intake for postprandial samples. All samples were subjected to glucose estimation using glucose-oxidase peroxidase method, immediately without much time lapse.

Statistical analysis

Statistical package for the social sciences software IBM SPSS Statistics for windows, version 23.0, Sigma plot 13(Systat software, USA) was used for statistical analysis Normality test was done using Kolmogorov–Smirnov test, Sharpiro–Wilk numerical test, and Q. Q plot.

RESULTS

The control group's mean age was 24.5 ± 5.147 and that of the study group was 25.03 ± 5.996. There was no statistically significant difference in the mean age between study and control group (P = 0.585). The total number of males in study group was 38 and in control group was 58; the females were 41 and 42 in study and control groups, respectively. The mean fasting blood glucose levels in Group 2 and Group 1 were 87.98 ± 14.321 and 142.11 ± 48.466, respectively. The mean blood glucose levels in postprandial states were also 119.10 ± 16.095 and 203.78 ± 62.189 in Group 2 and Group 1, respectively. The HbA1C levels in Group 2 and Group 1 were also found to be 5.543 ± 0.461 and 7.768 ± 2.008, respectively. All values were statistically significant with P ≤ 0.001 [Table 1].

Table 1

Mean distribution of blood glucose among study and control group

Blood glucose	Mean±SD		Mean difference	P	95% CI
	Group 1	Group 2
Fasting	142.11±48.466	87.98±14.321	54.134	<0.001	44.061-64.207
Postprandial	203.78±62.189	119.10±16.095	84.685	<0.000	71.911-97.459
HbA1C	7.768±2.008	5.543±0.461	2.2254	<0.000	1.816-2.634

CI: Confidence interval, SD: Standard deviation

The mean fasting stimulated and unstimulated salivary glucose values in Group 2 patients were 1.673 ± 0.654 and 1.456 ± 0.619, respectively. The mean values of Group 1 patients were 6.119 ± 0.344 and 6.043 ± 0.325 for fasting stimulated and unstimulated mean salivary glucose levels, respectively. Under postprandial states, the mean stimulated salivary glucose values were 1.673 ± 0.654 and 11.61 ± 1.123 in Group 2 and Group 1 patients, respectively. The mean unstimulated salivary glucose levels in Group 2 and Group 1 patients were 1.456 ± 0.619 and 10.24 ± 0.562, respectively. All values were statistically significant (P ≤ 0.001) [Table 2]. The correlation between blood and salivary glucose levels was evaluated using Pearson's correlation test. There was a positive moderate correlation between stimulated fasting salivary glucose and fasting blood glucose levels (r = 0.684). The linear regression equation was calculated with a model fit R2 = 0.467 and was y = 64.69 ± 12.98 (stimulated fasting salivary glucose) [Figure 1].

Table 2

Mean distribution of salivary glucose among study and control groups

Salivary glucose	Type of saliva	Mean±SD		Mean difference	P	95% CI
		Group 1	Group 2
Fasting	Stimulated	6.119±0.344	1.673±0.654	4.44539	<0.001	4.284-4.605
	Unstimulated	6.043±0.325	1.456±0.619	4.58704	<0.001	4.435-4.738
Postprandial	Stimulated	11.61±1.123	1.673±0.654	9.94286	<0.001	9.677-10.20
	Unstimulated	10.24±0.562	1.456±0.619	8.78577	<0.001	8.609-8.962

CI: Confidence interval, SD: Standard deviation

Figure 1

Correlation between stimulated fasting salivary and fasting blood glucose

The unstimulated fasting salivary glucose and fasting blood glucose levels also showed a moderate positive correlation (r = 0.678) and obtained a model fit R2 = 0.460. The linear regression equation was calculated to be y = 68.25 ± 12.53 (unstimulated fasting salivary glucose) [Figure 2].

Figure 2

Correlation between unstimulated fasting salivary and fasting blood glucose

The stimulated postprandial salivary glucose and postprandial blood glucose levels correlated well with a positive strong association (r = 0.737). The linear regression equation was calculated with a model fit R2 = 0.629 and y = 79.8 + 10.51 (stimulated postprandial salivary glucose) [Figure 3]. The correlation between unstimulated postprandial salivary glucose and postprandial blood glucose levels were obtained and showed a positive strong correlation (r = 0.742). The linear regression equation was y = 97.89 ± 11.23 (unstimulated postprandial salivary glucose) [Figure 4].

Figure 3

Correlation of stimulated postprandial salivary and postprandial blood glucose

Figure 4

Correlation of unstimulated postprandial salivary and postprandial blood glucose

The HbA1c values also correlated with salivary glucose levels in both stimulated and unstimulated saliva samples and showed a positive moderate correlation in both fasting and postprandial conditions [Table 3]. Receiver operating characteristic (ROC) curves were plotted, and sensitivity and specificity of stimulated and unstimulated saliva were calculated in fasting and postprandial conditions. The sensitivity and specificity for the diagnosis of diabetes using unstimulated fasting salivary glucose were 87% and 95%, respectively. The cutoff values for unstimulated fasting salivary glucose was 5.6 mg/dl, which translates the idea that patients with unstimulated salivary glucose above this value are most likely to be diabetic. ROC curves were also plotted for stimulated fasting salivary glucose and fasting blood glucose levels, and 86% and 92% were obtained, respectively. The cutoff value for stimulated fasting salivary glucose was 5.8 mg/dl which requires a blood glucose estimation to confirm the glycemic status.

Table 3

Correlation between blood and salivary glucose

Salivary glucose	Blood glucose	Correlation coefficient (r)
Stimulated fasting	Fasting	0.684
	HBA1C	0.645
Unstimulated fasting	Fasting	0.678
	HBA1C	0.639
Stimulated postprandial	Postprandial	0.737
	HBA1C	0.639
Unstimulated postprandial	Postprandial	0.742
	HBA1C	0.646

In postprandial states, salivary glucose levels were obtained and ROC curve was plotted and the cutoff values derived were 9.5 mg/dl and 10.3 mg/dl in unstimulated and stimulated salivary samples. The sensitivity and specificity obtained for unstimulated postprandial salivary glucose levels were 81% and 96%, respectively, and that of stimulated were 96% and 90%, respectively. The positive and negative predictive value and likelihood ratio were also derived [Tables 4 and 5].

Table 4

Sensitivity, specificity and area under the curve for salivary glucose and blood glucose

Blood glucose	Salivary glucose	Sensitivity (%)	Specificity (%)	AUC	95% CI	Cut-off value
Fasting	Stimulated fasting	86	92	0.900	0.841-0.958	5.8
	Unstimulated fasting	87	95	0.891	0.832-0.951	5.6
Postprandial	Stimulated postprandial	96	90	0.966	0.942-0.989	10.3
	Unstimulated postprandial	81	96	0.981	0.965-0.996	9.5

AUC: Area under the curve, CI: Confidence interval

Table 5

Predictive value and likelihood ratio

Salivary glucose	Positive predictive value	Negative predictive value	Positive likelihood ratio	Negative likelihood ratio
Stimulated fasting	0.9324	0.9048	17.468	0.1332
Unstimulated fasting	0.8947	0.8932	10.795	0.1513
Stimulated postprandial	0.9412	0.8649	20.253	0.1978
Unstimulated postprandial	0.8837	0.9677	9.6203	0.0422

DISCUSSION

Diabetes mellitus is a potential burden in the patients and the society due to its “silent killer” nature due to complication in various systems. Hence, an early detection and monitoring of diabetic status are extremely important. The gold standard method to monitor the diabetic status is through blood investigation, which has various disadvantages, so different body fluids were used in diagnosis, and one such was the use of saliva. Even though plentiful research has been done with saliva in monitoring the blood glucose levels, the obtained results were inconclusive in establishing the fact that saliva can be a valuable substitute in diagnosing and monitoring the disease in diabetes mellitus.[7] Our study, the male to female ratio was equal, which is in the ratio of 1:1.07 in type 1 diabetic patients. The P value was 0.31 which is not statistically significant. The present study showed the salivary glucose did not vary with gender. Similar observations were also seen in other studies showing gender do not affect the salivary glucose levels in type 1 diabetic patients.[8]

The correlation of blood and salivary glucose levels in type 1 diabetic patients have been shown in various studies. This positive correlation has been attributed to the fact that glucose being a small molecule, easily diffuses through the semi-permeable membrane of the glands.[9] Thus, large amounts of glucose become available to the saliva when blood glucose levels were elevated as in diabetes.[10] Salivary glands act as a filter of blood glucose that would be altered by hormonal or neural regulation. Glucose can also be present in saliva through gingival crevicular fluid.[11]

In the present study, salivary glucose values were higher among type 1 diabetes patients than in controls and were statistically highly significant in fasting and postprandial states. Salivary glucose levels were evaluated both in stimulated and unstimulated saliva, and both correlated positively with a moderate-to-high association with the blood glucose levels. There was a positive correlation between salivary glucose levels and HbA1c. Similar observations were also seen in various other studies showing that saliva can help in assessing the glycemic status of type 1 diabetic patients.[121314] Our study also showed mean salivary glucose levels both in stimulated and unstimulated samples and also in fasting and postprandial conditions to be higher in type 1 diabetics compared to controls, and all values were statistically significant. Similar study also showed a correlation in both fasting and postprandial conditions in salivary and blood glucose levels in type 1 diabetes.[15] This observation of positive correlation between blood and salivary glucose levels before and after food intake showed increased values of blood glucose levels after food intake and is reflected in the saliva. The duration of the entry of glucose into the salivary glands can also play an important role in the increased correlation of salivary glucose levels to blood glucose levels.[16] On the contrary, some studies have not shown good correlation between salivary and blood glucose levels, which can be attributed to the small sample size.[17]

The present study evaluated the mean salivary glucose levels in stimulated and unstimulated salivary samples showed variation in the values, and stimulated mean salivary glucose values were more than unstimulated mean salivary glucose. Another study was performed with 5 sets of experiments which consisted of different groups of individuals and different methods of analysis of the glucose levels in blood and saliva. They reported that salivary glucose concentration increased as blood glucose levels increased and also in normal subjects comparing stimulated and unstimulated saliva; it was found that there was a decrease in salivary glucose concentration, an increase in salivary flow but an unchanged glucose excretion rate in stimulated compared to unstimulated saliva. However, in diabetic patients, there was unchanged salivary glucose concentration and glucose excretion rate in both stimulated and unstimulated saliva. There was also no significant correlation between glycemia and either glucose concentration or glucose excretion rate in diabetes, whether stimulated or unstimulated saliva.[18] Salivary glucose was also estimated in parotid saliva, which showed a statistically significant result and good correlation on comparison of salivary and blood glucose levels.[19] Our study used stimulated and unstimulated whole saliva and showed a correlation between blood glucose levels in fasting and postprandial states.

The study also derived at cutoff values of salivary glucose levels in both stimulated and unstimulated saliva and was 5.8 mg/dl and 5.6 mg/dl in fasting condition and 10.3 mg/dl and 9.5 mg/dl in postprandial condition. Similar studies were also conducted showing cutoff values of 11.6 mg/dl in unstimulated whole saliva and normal range being 7.60–11.6 mg/dl.[20] Another study also showed salivary glucose level as 11.5 mg/dl, which indicated that above this value, the patient can be considered diabetic. The linear regression equation has also been derived in earlier studies with which blood glucose levels can be evaluated, serum glucose = 9.766 (salivary glucose) +7.353.[21] Another study also derived the equation as blood glucose = 9.875 (salivary glucose) +4.93 and R2 = 0.995, which was statistically significant.[20] Our study also gives regression equation of stimulated and unstimulated salivary samples in both fasting and postprandial condition with which blood glucose levels can be evaluated by knowing the value of salivary glucose values.

A metaanalysis and systematic review showed higher levels of salivary glucose concentrations in type 1 diabetes mellitus patients than in the controls, pointing to the fact that saliva could be used to monitor the health status of diabetic individuals. Glucose levels in saliva may increase during infection/inflammation. This meta-analysis also emphasized on the future line of research to provide sensitivity and specificity of salivary glucose levels in monitoring glycemia also other limitations in the meta-analysis were the studies did not show any time-dependent values as the concentrations of glucose varies in various intervals according to the concentration of glucose in the blood. Frthermore, smaller sample size were used in many of the earlier studies.[4]

Hence, keeping all the limitations of the systematic review, our study was performed in a larger sample size, and also standardized methodology was used. Furthermore, patients with underlying comorbid diseases were excluded from the study. Samples were collected in stimulated and unstimulated saliva samples at various time intervals. The specificity and sensitivity of the test with cutoff values were evaluated and interpreted. Thus, our study is unique in overcoming all the limitations and to give cutoff values for both stimulated and unstimulated salivary samples in both fasting and postprandial states.

CONCLUSION

This study shows that salivary glucose levels correlate well with blood glucose levels in Type 1 diabetes. There was not much difference in mean salivary glucose levels or cutoff values of stimulated and unstimulated salivary samples, but a slight increase in cutoff values was observed in stimulated salivary samples. Furthermore, the test had good specificity and sensitivity and a good positive predictive value, thereby showing that salivary glucose can definitely be a reliable alternative to blood glucose in the diagnosis of type 1 DM. Thus, the results of the present study support the use of saliva as a diagnostic fluid and can be used as a simple, noninvasive, easy method to diagnose, and in screening the blood glucose levels in Type 1 DM patients. Future research should aim at a multicentric study and ultimately fabricate a device where glucose assessment is done with a spit instead of a prick.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.