Saliva Changes in Pediatric Patients with Eating Disorders.

Objectives: To determine the differences between pediatric patients with eating disorders (ED) and the control group in the amount of saliva and the concentration of total amylase and electrolytes in saliva, and to evaluate the correlation between the saliva changes and nutritional status. Material and methods: The study included 101 participants (14.34 ±1.99 years), out of which 50 participants with ED subgroups and 51 participants in the control group. Data were statistically analyzed (Mann-Whitney, Kruskal-Wallis, chi-square, Spearman rank correlation test, α=0.05).
Results: No significant differences in salivary volume between the groups were found. A significant difference in the volume of saliva secreted in the 5th and 15th minute was found between the anorexia nervosa and bulimia nervosa subgroups. The examined anthropometric parameters were marginally or significantly positively associated with saliva volume at 5 and 15 minutes, noting a more significant correlation of the same at 15 than at 5 minutes. The patients with ED had a significantly higher concentration of inorganic phosphates in saliva while the concentrations of other electrolytes and total amylase in saliva did not differ significantly. Conclusions: Nutritional status affects salivation. There is a difference in saliva volume in pediatric patients with different ED disorders. Variations in saliva electrolytes in pediatric patients with ED are possible.

Introduction

Eating disorders (ED) are mental health conditions which are, according to the Diagnostic and Statistical Manual of Mental Disorders (DSM V), classified as Anorexia nervosa (AN) restrictive type (AN-RT) or binge/purge type (AN-BP), Bulimia nervosa (BN), Binge eating (BE), Other Specified Feeding or Eating Disorder (OSFED) and Unspecified Feeding or Eating Disorder (UFED) (). The last two disorders, that is, OSFED and UFED, came from what used to be called eating disorders not otherwise specified (EDNOS) in DSM IV (). EDS is a relatively rare disorder occurring relatively rarely in the general population. Lifetime prevalence estimates for anorexia nervosa, bulimia nervosa, and binge-eating disorder in USA were 0.3%, 0.9%, and 1.6%, respectively (). Among European women, the prevalence of AN is <1-4%, BN <1-2%, overeating disorders 1-4%, other eating disorders 3% with significant differences depending on the country/region, age groups and ethnicity. Among men, 0.3% of them met the criteria for one of the eating disorders (). There is no accurate data on the prevalence of eating disorders in Croatia. However, there are a few studies showing that a significant number of adolescents exhibit abnormal eating habits. And even though ED usually occurs in early puberty, we are witnessing a shift towards younger age groups (–).

Numerous changes involving all organ systems are developing as a result of dysfunctional eating behavior in patients with eating disorders (ED). The changes also affect the oral cavity. The most commonly described oral findings are dental plaque, marginal gingivitis, calculus, caries, exfoliative cheilitis, angular cheilitis and burning tongue (–). Some changes in the volume and composition of saliva are also described (). Serum amylase levels are often elevated in patients with BN and AN-BP, most often due to an increase in salivary amylase due to frequent and intense vomiting (). Patients with ED often complain of a feeling of reduced salivation and consequently dry mouth ().

Although numerous studies about composition and volume of saliva, electrolytes and total amylase in saliva in patients with eating disorders (, , –) were performed, the existing data about these parameters in pediatric patients with ED are scarce.

The aim of this study was to determine differences between the group of pediatric patients with ED and the control group in the amount of saliva and the concentration of total amylase and electrolytes in saliva and to evaluate the correlation between saliva changes and nutritional status.

Material and methods

This cross-sectional study was carried out from September 2016 until November 2017 at the Department of Pediatric Gastroenterology in University Hospital Center Sestre milosrdnice Zagreb and School of Dental Medicine, Zagreb, Croatia in accordance with the Declaration of Helsinki and approved by the Hospital Ethics Committee and by the Ethics Committee of the School of Dental Medicine, University of Zagreb (05-PA-26-1/2016). The study included 101 female participants with a mean age of 14.34 ±1.99 years, of which 50 participants had eating disorders and 51 participants were included in the control group. The patients included in the study as well as their parents received both written and oral information about the study and a written informed consent was signed by the parents. The study group consisted of 50 patients with ED who were hospitalized in the Department of Gastroenterology, Hepatology and Eating Disorders of the Pediatric Clinic (University Hospital Center Sestre milosrdnice, Zagreb, Croatia). The diagnosis of ED was made according to DSM V criteria. Participants were classified into subgroups of ED: AN-RT, AN-BP, BN, BE, UFED and OSFED. Participants with AN-RT and AN-BP, due to the small number of participants, were merged into one group (AN). Also, for the same reason, BE, UFED and OSFED disorders were merged into one group named EDNOS (the name is taken from the DSM IV criteria). The control group consisted of 51 participants of the same age who were also hospitalized in the Pediatric Clinic where they were admitted for treatment due to symptoms of diseases outside the gastrointestinal tract, i.e. due to conditions and diseases in which no changes in the oral cavity were expected. Each participant underwent anthropometric measurements, determination of body height in centimeters and body weight in kg. The measurements were performed with a scale with an altimeter and measuring scale, in the morning on an empty stomach after waking up, wearing only underwear. The Body Mass Index (BMI) was calculated by dividing weight and height squared (kg/m2). The nutritional status was expressed in z-score for BMI, since the values of BMI change with age in children. Whole unstimulated saliva samples were collected between 10 and 12 o'clock in the morning, at least 2 hours after the last meal or fluid intake. Participants sat with their head tilted forward and, after initially swallowing saliva, they expelled unstimulated saliva every 60 seconds into calibrated tubes (0.1 ml) for 15 minutes. Salivary flow, expressed in milliliters in the fifth and fifteenth minutes, was determined for each participant. Samples of collected unstimulated saliva were submitted for processing to the Clinical Department of Chemistry of the University Hospital Centre Sisters of Mercy Zagreb within one hour of sampling, frozen at -60°C and stored until used for measuring electrolyte concentration and total amylase. The determination of electrolytes in saliva was performed using the method of indirect potentiometric on the analytical platform Architect c8000 (Abbott Laboratories, Chicago, Il, USA). The determination of a total salivary amylase was performed using the spectrophotometric enzymatic method on the same analyzer.

Statistical analysis was performed by Statistical software package version 13.5 (TIBCO Inc., Palo Alto, CA, USA). The normality of data distribution was tested by the Shapiro-Wilks W test and homogeneity of variance by Leven test. The results of descriptive analyses were expressed as median (min-max) for continuous data and percentages for categorical data. The differences in distributions of continuous data were evaluated by nonparametric tests (Mann-Whitney, Kruskal-Wallis) since the assumption of homogeneity of variance for tested variables had not been proven. The multiple comparison tests were used for post-hoc analysis. The differences in distributions of categorical data were assessed by the Chi-square test. The Spearman rank correlation test was used to evaluate the direction and the strength of the associations between the tested variables; p < 0.05 was considered statistically significant.

Results

The study included 101 participants with a mean age of 14.34 ±1.99 years, of which 50 participants were patients with ED and 51 participants were included in the control group. Participants from both groups did not significantly differ by age (p=0.261). Patients with ED were significantly malnourished. They had significantly lower BMI z-value (t=9.822, p<0.001) than the participants in the control group. Table 1 shows subjective feeling of saliva volume and volume of secreted saliva in examined groups. Table 2 shows subjective feeling of saliva volume and volume of saliva secretion in patients with ED divided into subgroups according to the type of eating disorder. A post hoc multiple comparison test showed a significant difference in the volume of saliva secreted in the 5th (p=0.007) and 15th minute (p=0.028) only between the AN and BN subgroups.

Table 1

Subjective sensation of saliva volume and volume of saliva secreted in participants divided into two groups according to the presence/absence of ED.

	ED patients (n=50)	Control group (n=51)	Chia Zb	p
Subjective sensation of saliva volume *	20/72/8	12/86/2	3.590a	0.166a
Saliva volumein the 5th minute (mL)**	1, 0-3.3	1.4, 0.1-8.5	1.609b	0.107b
Saliva volumein the 15th minute (mL)**	2.5, 0.5-8	3.5, 0.6-15	1.434b	0.143b

* percentages ** med, min-max a Chi-square test df=2 b Mann-Whitney test

Subjective sensation of saliva volume: decreased/normal/increased.

ED - eating disorders.

Table 2

Subjective sensation of saliva volume and volume of saliva secretion in patients with ED divided into subgroups according to the type of eating disorder.

	AN(n=27)	BN(n=6)	EDNOS(n=17)	Chia Hb	p
Subjective sensation of saliva volume *	22/70/8	33/50/17	12/82/6	2.513a	0.642b
Saliva volumein the 5th minute (ml)**	0.8, 0-2.3	1.9, 1.2-2.8	1.2, 0.5-3.3	10.449b	0.005b
Saliva volumein the 15th minute (ml)**	2.2, 0.5-7	4.7, 2.5-8	3.6, 0.7-7	9.274b	0.009b

* percentages ** med, min-max a Chi-square test df=4 b Kruskal-Wallis test

Subjective sensation of saliva volume: decreased/normal/increased.

AN - anorexia nervosa; BN - Bulimia nervosa; EDNOS - eating disorders not otherwise specified.

All examined anthropometric parameters were marginally or significantly positively associated with saliva volume at 5th and 15th minute (Table 3, Figures 1 and 2), noting a more significant correlation of the same at 15th than at 5th minute.

Table 3

Correlations of anthropometric parameters with saliva volume at 5 and 15 minutes (all participants).

		BMI	BMI z-score	BMI percentile
Saliva volume in the 5th minute	Spearman Rt(N-2)p	0.2252.2950.024	0.1971.9990.048	0.1911.9330.056
Saliva volume in the 15th minute	Spearman Rt(N-2)p	0.2382.4350.017	0.2132.1680.033	0.2042.0700.041

BMI - body mass index.

Figure 1

Relationship of BMI z-score with saliva volume at 5 minutes (all participants).

Figure 2

Relationship of BMI z-score with saliva volume at 15 minutes (all participants).

Total amylase in saliva did not differ significantly between patients with ED and the control group (med. 82802 vs. 77891 U/L; p=0.666). Table 4 shows concentration of total salivary amylase in patients with ED divided into subgroups according to the type of eating disorder. A marginally higher concentration of total serum amylase was found in patients with ED than in the control group (med. 56.5 vs. 50 U/L, p=0.052) but no association between serum amylase and salivary amylase concentrations has been demonstrated (p>0.05).

Table 4

Concentration of total salivary amylase in patients with ED divided into subgroups according to the type of eating disorder.

	AN(n=27)	BN(n=6)	EDNOS(n=17)	H	p
Total amylasein saliva (U/L)*	71030, 11007-357796	86212, 60658-383484	91907, 2998-309260	1.554	0.459

* med, min-max Kruskal-Wallis test

AN - anorexia nervosa; BN - Bulimia nervosa; EDNOS - eating disorders not otherwise specified.

Patients with ED had a significantly higher concentration of inorganic phosphates in saliva than participants in the control group (p=0.025), while the concentrations of other electrolytes in saliva did not differ significantly (Table 5). There was no significant difference in salivary electrolyte concentration among patients with different types of ED (p>0.05) (data not shown). BMI z-score and salivary volume at 5 and 15 minutes were marginally or significantly negatively associated with the concentrations of most of the tested electrolytes in saliva (Table 6), except for the concentrations of inorganic phosphates and a total salivary amylase with which they were not bounded, hence these results are not shown in any tables.

Table 5

Salivary electrolyte concentration in participants divided into two groups according to the presence/absence of ED.

	ED patients(n=50)	Control group(n=51)	Z	p
Total calcium (mmol/L)*	1.21, 0.92-2.94	1.11, 0.66-1.85	-1.446	0.147
Inorganicphosphates(mmol/L)*	6.08, 3.41-14.69	5.11, 2.89-19.2	-2.234	0.025
Magnesium(mmol/L)*	0.30, 0.25-0.92	0.27, 0.25-0.58	-1.653	0.098
Potassium(mmol/L)*	24.4, 15.9-47.8	23.2, 15.5-44.6	-1.120	0.262
Sodium(mmol/L)*	20, 20-28	20, 20-24	-0.173	0.862
Chlorides(mmol/L)*	22, 20-53	21, 20-34	-1.300	0.193

* med, min-max Mann-Whitney test

ED - eating disorders.

Table 6

Correlations of BMI z-score and saliva volume at 5 and 15 minutes with salivary electrolyte concentration.

ELECTROLYTES IN SALIVA		BMI z-score	Saliva volume in the 5th minute	Saliva volume in the 15th minute
Total calcium	Spearman Rt(N-2)p	-0.098-0.9830.328	-0.333-1.7670.074	-0.280-1.4590.157
Magnesium	Spearman Rt(N-2)p	-0.167-1.6880.094	-0.279-1.4500.159	-0.324-1.8170.061
Potassium	Spearman Rt(N-2)p	-0.070-0.7010.485	-0.375-2.0220.054	-0.336-1.7630.066
Sodium	Spearman Rt(N-2)p	-0.025-0.2520.801	-0.329-1.7290.079	-0.303-1.5890.125
Chlorides	Spearman Rt(N-2)p	-0.165-1.6640.099	-0.392-2.1310.043	-0.354-1.8900.059

BMI - body mass index.

Discussion

Eating disorders are often diagnosed late. They are most commonly developed during adolescent age (15-19 years) (–). Complications of ED develop in all organ systems, and numerous changes in the oral cavity have been described (–). So far, changes in saliva have been described in patients with eating disorders in adulthood, with the exception of research conducted by Paszynska et al (), which likewise in the present study included younger girls.

The results of a series of studies showed that patients with ED are more likely to report a subjective feeling of decreased saliva and have a smaller volume of unstimulated saliva at 5 and 15 minutes than participants in the control group (, –), which could be considered similar to the results of this study (Table 1). However, the differences found between the groups did not reach statistical significance (Table 1), which coincides with the results published by Johansson et al (). Decreased salivary flow (in the ED group in this study) was not due to antidepressant therapy and/or the use of appetite-suppressing drugs, which have reduced salivation as a side effect, because none of the participants used them. The results of a study by Paszynska et al (), which included 23 girls with AN RT with an average age of 15.5 ±2.1 years, showed a statistically significant difference in the flow of unstimulated saliva between the group with AN RT and the control group. The girls in Paszynska’s study (), as well as the patients with AN RT in this study, were significantly malnourished, which, along with dehydration, is a known cause of decreased saliva, which could explain the lower saliva flow in AN patients compared to BN patients (Table 2). This conclusion () was confirmed by a determined borderline or statistically significant positive correlation between the degree of nutrition and the volume of non-stimulating saliva, which was more pronounced in the 15th minute compared to the 5th minute of the study (Table 3, Figure 1 and 2).

The literature review has revealed different results for total amylase in saliva. Riad et al () found elevated amylase values in unstimulated saliva in bulimic patients. Elevated amylases in 25–60% of patients with BN were also observed by a number of other researchers with the assumption that it was salivary amylase (, , -). Contrary to these studies (, , -), Paszynska et al () found no differences in amylase concentration between the group with AN RT and the control group, nor did Johansson et al (), which is close to the present study results. Paszynska et al () explained these findings by the shorter duration of the disease with respect to the age of the patients. They have concluded that further research is needed to determine the effects of nutritional status on amylase excretion.

There are very few studies describing salivary electrolyte concentrations in ED patients. In the first studies examining salivary electrolyte levels, no significant differences were found between patients and control groups (, ). In the study by Johansson et al (), significantly higher concentrations of phosphate in unstimulated saliva were determined which is similar to the results of the present study (Table 5). A higher concentration of phosphate which was obtained in unstimulated saliva could be linked with a decreased salivary flow in patients with ED (–) (Table 1). Yet, other electrolyte concentration did not differ between ED patients and the control group despite lower salivary flow in ED patients (Table 1). Johansson et al () also found significantly higher concentrations of chloride and magnesium (). Blazer et al () found significantly higher concentrations of magnesium, and there was no significant difference in the concentrations of sodium, calcium, phosphate, potassium, and chloride. Statistically significantly higher calcium concentrations in unstimulated saliva in adult BN patients (23.8 ±4 years) were noted by Dynesen et al (), which is contrary to the results obtained in the present study (Table 5). However, some limitations in the present study should be pointed to such as the use of combined subgroups with distinct clinical characteristics (AN, BN and EDNOS within ED group). Based on previous investigations (, –, ), and on the results of the present study (Table 5), the effect of ED on the state of electrolytes in saliva remains unclear. Nevertheless, some studies showed common electrolyte (serum) disturbances in children with malnutrition (, ). With this in mind, along with determined different concentrations of saliva electrolytes (Table 5) (, , ), variations in saliva electrolytes in ED can be considered possible. To confirm or dispute these findings, further studies are needed, especially with phosphate in saliva in pediatric patients with different types of ED.

Conclusions

Nutritional status affects salivation. There is a difference in saliva volume in pediatric patients with different ED disorders. Variations in saliva electrolytes in pediatric patients with ED are possible.