Örebro universitet
Institutionen för hälsovetenskap och medicin Enheten klinisk medicin
Kurs: Medicin, avancerad nivå. Examensarbete Datum: 2015-06-07
30-year (1983-2013) trends in saliva flow rate
and saliva buffer capacity. Analyses from
10-year repeated, cross-sectional population
samples in the Jönköping area
Författare: Martin Ågren Handledare: Ingegerd Johansson
Professor, Umeå Universitet
A
BSTRACTOBJECTIVE: To determine trends in saliva flow rate and saliva buffer capacity and to evaluate
associations between saliva flow rate and buffer capacity with potential explanatory factors, including chewing ability, medication and other health variables, including oral health.
MATERIALS AND METHODS: The study group consisted of 2,509 individuals (1,204 men and 1,305 women) randomly selected from the town of Jönköping, Sweden. All participants underwent an oral examination as well as completed a questionnaire.
RESULTS: The highest mean saliva flow rate in men and women was seen at age 50 and 40 years, respectively. Sex and age standardized means for saliva flow rates for all subjects differed significantly between the screening years, but they did not follow a linear trend over time. Through PLS analysis low buffer capacity (pH <5.5), being a woman, xerostomia, less occlusal support zones (Eichner index), and having more daily meals of food, or having less teeth were identified to have a significant
correlation to lower saliva flow rate. The analytic method of buffer capacity evaluation changed between the study years, thus differences between screening years are not taken into consideration. In 8 out of 14 strata mean flow rate was significantly higher in those with buffer pH >5.5 than those with a buffer pH<5.5. Univariate analysis, from variables indicated in the PLS analysis, revealed that low saliva flow rate, moderate degree of periodontal disease, smokers, and those with the poorest Eichner index have low buffer capacity.
CONCLUSIONS: The stimulated salivation is dependent on gender, number of teeth, Eichner index, subjective xerostomia and buffer capacity. Different normal salivation distribution scales could be needed for men and women.
R
EGISTERIntroduction
1
The Jönköping study
4
Aim
4
Subjects and method
4
Study cohort
4
Questionnaire
5
Outcome variables
5
Saliva collection and buffer capacity analyses
5
Potential explanatory or confounding variables
5
Statistical analyses
6
Ethical aspects
6
Results
7
Participants
7
Identification of potential confounders
7
Thirty year time trends in saliva flow rate
8
Identification of factors associated with saliva flow rate
9
Thirty years time trends in buffer capacity
11
Identification of factors associated with buffer capacity
12
Discussion
14
Conclusion
16
Acknowledgements
16
1
30-
YEAR(1983-2013)
TRENDS IN SALIVA FLOW RATE AND SALIVA BUFFER CAPACITY.
A
NALYSES FROM10-
YEAR REPEATED,
CROSS-
SECTIONAL POPULATION SAMPLES IN THEJ
ÖNKÖPING AREAAuthor: Ågren Martin
I
NTRODUCTIONSaliva plays a crucial role in protection of the oral cavity (Edgar et al., 2004). Thus, flushing (clearance and exposure) of the soft and hard tissues and biological effects of saliva constituents, such as pH-regulatory components, inorganic components participating in the de- and remineralisation of tooth tissues, and proteins, peptides, lipids and carbohydrates affecting microbial growth and colonization, are of importance. Saliva flow and oral muscle activities are the key determinants for flushing. Lack of proper amount of saliva may lead to discomfort for the individual, such as a burning sensation, dysphagia, speech impairment, taste disturbances and increase risk for dental caries (Bergdahl, 2000; Närhi, 1994).
Knowledge of population based reference values are important for identification of conditions which need to be observed or treated. For saliva flow rate, the cut off values 0.7 mL/min to 1.0 mL/min and 0.1 mL/min for stimulated and unstimlated saliva, respectively are applied for
hyposalivation (Fujibayashi et al., 2004; Bergdahl 2000). Values above are considered normal. If the flow rate has drifted in the population, such as indicated in Bergdahl (2000), these cut-offs may be misleading, and should eventually be questioned. For buffer capacity the cut off for low capacity is a final pH <5.5 after addition of a standardized amount of acid. This is a reflection of the pKa value and buffer range of the bicarbonate ion. The buffer capacity is crucial for dental erosion, caries, and microbiota ecology as a pH below 5.5 creates an environment of demineralisation and supports a bacteria dysbiosis characterized by acidophilic and aciduric bacteria(Schuurs, 2013; Marsh, 2010).
Saliva is produced by the three paired major salivary glands, the parotid, submandibular and sublingual glands, and numerous minor salivary glands distributed in the oral mucosa. In an unstimulated state the submandibular glands account for two thirds of the secreted saliva. When stimulated, secretion from the parotid glands increases proportionally the most, and accounts for up to 60 % of the secretion (Carpenter, 2013). The parotid glands are comprised of serous and the submandibular glands of both mucous and serous acinar cells, with a majority of the latter. In
comparison to the parotid saliva, saliva from the submandibular glands is more viscous and mucin-rich. The smallest of the major salivary glands are the sublingual glands, which produce highly viscous, mucous saliva that accounts for a few per cent of the total amount of saliva. The minor glands
2
contribute for ≤10 % of the secreted saliva. They produce mainly mucous saliva rich in secretory IgA (Dawes and Wood, 1973).Secretion of saliva is regulated by the parasympathetic and the sympathetic systems. Secretion is initiated by afferent signals from three cranial nerves, the trigeminal (V), facial (VII) and
glossopharyngeal (IX) nerves. The afferent innervation stems from mechanoreceptors in the
periodontal ligament and chemoreceptors in the taste buds. The stimulated salivation is increased by frequency of chewing cycles, force of chewing and number of teeth occluding on the object chewed upon. The increase in salivation from the parotid glands is higher on the ipsilateral side, but the contralateral side also shows an increase in salivation (Jensen Kjeilen et al., 1987; Samnieng et al., 2012). The masticatory-salivary reflex is initiated not only by the periodontal ligament, as it has been shown that edentulous subjects also have an increase in parotid salivation when chewing (Scott et al., 1998). The salivation is also dependent on conditioned reflexes, controlled through the autonomic nervous system. This central control inhibits unstimulated salivation during sleep, fear and mental depression, and excites salivation during fight. The classical parasympathetic and sympathetic
transmitters, i.e. acetylcholine and either noradrenaline (norepinephrine) or adrenaline (epinephrine), respectively, are the main activators of salivation (Ekström, 1989). The parasympathetic system mainly regulates water secretion and the sympathetic protein secretion. The complex regulation of the secretory cells makes secretion sensitive to numerous external and intrinsic factors. Examples of such factors are medications, diseases, stress, water balance, starvation, and lack of chewing stimulation (Bardow et al., 2008).
Water secretion from the acinar cells is driven by intracellular loss of potassium to the
insterstitium and chloride to the lumen. The loss of potassium to the extracellular fluid forces sodium into the lumen causing a salty environment, which drives water, through osmosis, into the lumen of the salivary gland. During its way through the duct of the salivary gland the primary saliva is modified in a flow rate related fashion by selective reabsorption/secretion of electrolytes. Among the
electrolytes that are affected during secretion are sodium and bicarbonate (Bardow et al., 2008). One important aspect of saliva action is its contribution to remineralisation of the teeth, as well as preventing demineralisation by buffering (Selwitz et al., 2007). After each intake of food, the pH in dental plaque drops and remains at that state until soluble carbohydrates are cleared from the oral cavity, and the acids produced from bacteria are neutralized (buffered).
The magnitude and the time period below the tooth tissue critical pH are determined by the amount of acid produced by bacteria, by the buffer capacity of the saliva, and saliva flushing. There are three systems in saliva that contribute in buffering the pH after an acid attack, and striving it above the
3
critical pH for caries development in tooth enamel (around 5.5) or dentin (around 6.2), as well as erosion.Fig. 1. Schematic drawing of the complex
interactions and interplays in the caries process. The figure is redrawn from Selwitz RH, Ismail AI, Pitts NB. 2007. Dental caries. Lancet 369:51–59.
The phosphate buffer system - In stimulated saliva, phosphate is mainly in the form of hydrogen
phosphate, compared to in unstimulated saliva where it is mainly dihydrogen phosphate. Thus stimulated saliva is a bit more alkaline than unstimulated saliva. The acid dissociation constant, or pK, for the equilibrium between these two phosphate forms is around pH 7. At this neutral pH the phosphate buffer system is most efficient, with a buffer span of about ± 1 pH unit around the pK. As the phosphate concentration in saliva decreases with the flow rate of saliva, the contribution from the phosphate buffer system to the overall buffer capacity ranges from around 50% in unstimulated saliva to 10% in highly stimulated saliva (Bardow et al., 2008).
The bicarbonate buffer system - As opposed to the phosphate buffer system, the contribution of
the bicarbonate buffer system to the buffering effect increases when saliva flow rate increases. In unstimulated saliva, this system contributes to less than 50% and in stimulated saliva more than 90% of the buffering effect. The pK value for the bicarbonate buffer system lies around pH 6 with a
buffering effect up to pH 7 (Bardow et al., 2008). The lower limit of a pure bicarbonate buffer system is at pH 5.5, under that the effect of this system is minimal (Garbacz et al., 2013). In addition, CO2 in gas form is undersaturated versus the ambient air leading to loss of CO2, and a so called phase buffering (Bardow et al., 2008; Birkhed and Heintze, 1989).
The protein buffer system - Saliva proteins and peptides can act as buffers if the pH exceeds (both
positively and negatively), their isoelectric points, leading to the proteins accepting or releasing protons. Many of the saliva proteins have their isoelectric points at pH 5 and pH 9, making them effective as buffers at alkaline environments as well as when the pH drops below the critical value for
4
demineralisation. Although generally less effective than the two first buffer system, the localconcentration of proteins/peptides may render local significant buffer effect (Bardow et al., 2008).
The Jönköping study
Jönköping is a middle sized Swedish town with 130,000 inhabitants in in 2013. It is the Capital of the Jönköping County, and is situated on the main transport roads between the three biggest cities in Sweden, namely Stockholm, Gothenburg and Malmö. The Public Dental Service of Jönköping County also manages the Institute for Postgraduate Dental Education in Jönköping. The Jönköping study (Hugoson et al., 2005 a,b), with its 10-year repeated oral screenings with for the most part
standardized methods, is one of few studies worldwide offering the capacity to do a “longitudinal” evaluation of trends in saliva flow and buffer capacity in repeated, cross-sectional population based samples over a long time.
A
IMThe primary aim of the present study was to evaluate time trends for saliva flow rate and saliva buffer capacity using the 10-year screenings in the Jönköping study. A second aim is to evaluate the
associations between saliva flow rate/buffer capacity and potential explanatory factors, including number of teeth, chewing ability, diet intake, and medication.
S
UBJECTS ANDM
ETHODSStudy cohort
In the Jönköping study independent cross-sectional selections of individuals in the age groups 3, 5, 10, 15, 20, 30, 40, 50, 60, 70 and 80 years have been randomly selected from individuals residing in Kristine, Ljungarum, Sofia and Järstorp parishes as defined in 1973 (by 2006 the two first parishes merged into one, and so did the two last in 2010). The examinations are done on a decennial basis since 1973, saliva flow rate and buffer capacity was not measured in 1973 though. The selected individuals have been invited, and in those who agreed to participate, about a 100 in each age group, an extensive oral examination was performed under good clinical conditions (Hugoson et al., 1986, 1995, 2005 a,b). The age groups of 3, 5, 10 and 15 are not included in the present evaluation (Koch et
al., 2009). At all screening occasions 130 randomly selected adult in each of the 10-year age group
were invited to participate (Huguson et al., 2005). Due to a low participation rate in 2013 additionally 40, 40 and 50 subjects were invited in the age groups of 30, 40 and 50 years.
5
Questionnaire
All participants answered a questionnaire in immediate connection to the examination. In the questionnaire the participants answered questions on the number of meals and snacks per day, medication, and graded their subjective experience of xerostomia, health, tobacco use, and ease of chewing.
Outcome variables
The outcome variables in the present study are flow rates of stimulated whole saliva, and buffer capacity of the same saliva from the screenings in 1983, 1993, 2003 and 2013.
Saliva collection and buffer capacity analyses
Whole saliva, stimulated by chewing on a 1 gram piece of paraffin, was collected into graded test tubes for 3 minutes. Flow rates were calculated and buffer capacity analysed. The analysis method of the buffer capacity changed over the years. In 1983 and 1993 it was done according to Ericsson’s laboratory buffer capacity test (Ericsson, 1959). In 2003 and 2013 chair-side simplified methods (Dentobuff Strip, Orion Diagnostica, Espoo, Finland; GC Saliva Check, GC Europe NV, Leuven, Belgium) was used. A study showed best agreement among results provided by strip-type systems in patients with high buffering capacity, all compared to Ericsson’s laboratory buffer capacity test (Cheaib et al., 2012). Therefore, buffer capacity measures were dichotomized into low and normal capacity, i.e. <pH 5.5 and ≥ pH 5.5, respectively.
Fig 2. Correspondence between
Ericsson’s laboratory buffer capacity test, Dentobuff Strip and GC Saliva-Check. The figure is drawn from Cheaib et al. 2012
Potential explanatory or confounding variables
The following variables were evaluated for a potential explanatory effect or included as confounders in statistical models:Sex, age, number of teeth, Eichner index, subjective reported chewing capacity, dietary variables, medication and periodontal health. The Eichner index describes the number of occlusal support zones in the mouth, that is where occlusion (chewing) is supported, with a maximum of four support zones. These are the molar support zone (left and right), and the premolar support zone (left and right). If occlusion exists in all four support zones the occlusion is categorised as class A and sub grouped if the support zones lacks teeth in either one or both of the jaws, if there exists an
6
occlusion, but at least one of the four support zones does not participate in the occlusion it iscategorised as class B which is sub grouped according to how many of the support zones that are lacking, if there are no occlusion or the individual is edentulous it is a class C, sub grouped to whether the individual is edentulous or have non-occluding teeth in either one or both jaws. Tooth or implant supported fixed dentures are considered to participate in the occlusion, but removable dentures are not. (Eichner, 1955). The number of teeth and the Eichner index does not include the wisdom teeth, i.e. maximum number is 28 teeth.
Statistical analyses
Statistical analyses were performed for all subjects together and separately for men and women in 10-year age strata using IBM SPSS version 22 (IBM, Armonk, NY, USA). Statistical tests were two-sided and p-values <0.05 considered statistically significant. Descriptive statistics include frequencies,
proportions, and means with measures of variation. Differences in mean values for normally distributed variables were tested with Student´s unpaired t-test or ANOVA, followed by Bonferroni post hoc tests where applicable. Standardization for age and/or sex was done by general linear regression (glm). Non-normally distributed variables were tested by Chi2 –test.
Multivariate partial least square regression (PLS; SIMCA 14, version 14.0, Umetrics AB, Umeå, Sweden) was used to search for hidden structures in the data. In contrast to traditional regression, PLS is suitable for data where the x variables co-vary and the group number is limited. Variables were autoscaled to unit variance, and cross-validated prediction of Y calculated (Ståhle and Wold, 1988). Cross validation is done by a systematic prediction of 1/7th of the data by the remaining 6/7th of the data. The importance of each x variable in explaining the variation in y is displayed in a PLS loading plot and the correlation coefficients in a bar plot with means and 95% CI. Variables for which the 95% CI does not include zero are statistically significant.
Ethical aspects
7
R
ESULTSParticipants
The total number of participants in the four study years were 2,509 (1,204 men and 1,305 women), with a virtually equal distribution on sex and 10-year age groups at each screening occasions (Table 1). The attendance rate was 77.2% in 1983, 75.4% in 1993, 69.0% in 2003 and 56.4 % in 2013.
Table 1. Numbers (%) of participants per study year in sex and 10-year age strata by study year.
Agea 1983 1993 2003 2013 Men n (%) Women n (%) Men n (%) Women n (%) Men n (%) Women n (%) Men n (%) Women n (%) 20 years 45 (45) 55 (55) 50 (50) 50 (50) 46 (55) 38 (45) 22 (29) 54 (71) 30 years 50 (51) 48 (49) 63 (62) 39 (38) 42 (46) 50 (54) 46 (50) 46 (50) 40 years 47 (47) 52 (53) 54 (58) 39 (42) 47 (57) 36 (43) 48 (52) 45 (48) 50 years 43 (41) 60 (59) 45 (46) 52 (54) 41 (45) 50 (55) 43 (43) 58 (58) 60 years 51 (52) 47 (48) 50 (54) 42 (56) 45 (50) 45 (50) 42 (51) 40 (49) 70 years 48 (48) 51 (52) 36 (36) 64 (64) 41 (46) 48 (54) 38 (41) 54 (59) 80 years 30 (38) 50 (62) 34 (48) 37 (52) 21 (35) 39 (65) 36 (51) 35 (49) Total 314 (47) 363 (53) 332 (51) 323 (49) 283 (48) 306 (52) 275 (47) 312 (53) a) Participants turning the age during the study year.
Identification of potential confounders
As a first step, the associations between saliva flow rate and buffer capacity on the one hand and sex and age on the other were analysed.Univariate comparisons revealed that the mean saliva flow rate was systematically higher in men than women in all age groups (Table 2; Fig 3), and that it differed among the age groups (Table 2; Fig 3). The highest mean saliva flow rate in men and women was seen at age 50 and 40 years, respectively, and thereafter flow rate decreased continuously in both sexes (Fig 3). The proportion with low saliva buffer capacity differed by age in men but not women (Table 2, Fig 4). Hence, measures for screening year comparisons should be standardized for age or sex.
Table 2. Saliva flow rate and buffer capacity by sex and 10-year age groups. Unstandardized means and 95% CI
for flow rate and proportions with a low saliva buffer capacity, i.e. a final buffer pH <5.5.
N
Saliva flow rate mL/min (mean (95% CI)
Proportion with saliva buffer pH <5.5 Age Men n=1,178 Women n=1,271 p-value (sex)a N Men n=1,149 Women n=1,229 p-value (sex)b 20 years 358 1.7 (1.6-1.8) 1.5 (1.4-1.7) 0.009 356 58.9 % 60.6 % 0.412 30 years 376 1.7 (1.6-1.8) 1.6 (1.4-1.7) 0.299 363 41.1 % 54.3 % 0.008 40 years 355 1.9 (1.8-2.0) 1.7 (1.5-1.8) 0.014 338 44.0 % 53.2 % 0.057 50 years 388 1.9 (1.8-2.1) 1.6 (1.5-1.7) 0.002 373 47.6 % 53.6 % 0.147 60 years 351 1.7 (1.5-1.8) 1.5 (1.3-1.6) 0.027 346 50.8 % 52.1 % 0.445 70 years 359 1.7 (1.5-1.8) 1.4 (1.3-1.5) 0.010 349 55.3 % 61.9 % 0.126 80 years 262 1.5 (1.4-1.7) 1.1 (0.9-1.2) <0.001 253 56.6 % 59.3 % 0.383 p-value (age) - 0.002a <0.001a - 0.006b 0.309b
a) Differences between means were tested with unpaired t-test (sex) or ANOVA (age). b) Differences in distributions were tested with a Chi2-test
8
Fig 3. Unstandardized mean saliva flow rate with 95% CI in (A) men, and (B) women).Age 20 30 40 50 60 70 80 m L/ m in 0,0 0,5 1,0 1,5 2,0 Age 20 30 40 50 60 70 80 m L/ m in 0,0 0,5 1,0 1,5 2,0
* for p-0.05 and ** for p<0.01, *** for p<0.001 when compared to all other age groups.
Fig 4. Percentage of study subjects with low buffer capacity, pH <5.5, for (A) men, (B) women).
* for p-0.05 and ** for p<0.01, *** for p<0.001 when compared with 1983.
Thirty year time trends in saliva flow rate
Sex and age standardized means for saliva flow rates for all subjects differed significantly between the screening years, but they did not follow a linear trend over time (Fig 5A). The same pattern was seen when men and women were analyzed separately with standardization for age (Fig 5B), and in sex and age strata (data not shown). Thus, mean flow rates in 1993 were significantly higher than in 1983, and then again lower in 2003 and 2013 compared to 1993. Standardization for sex and age had very limited effect on the mean values, i.e. only the second decimal was affected compared to the means from the univariate analyses.
A
B
A
B
**
*
**
*
9
Fig 5. Mean (95% CI) of saliva flow rate by study year for (A) all subject and with standardization for age and sex,and (B) for men and women standardized for age.
Screening year mL /min 0,0 1,2 1,4 1,6 1,8 2,0 2,2 1983 1993 2003 2013
A
Screening year mL/min 0,0 1,2 1,4 1,6 1,8 2,0 2,2 Men Women 1983 1993 2003 2013
B
*** for p<0.001 compared to all other screening years.
Identification of factors associated with saliva flow rate
As a final step all variables were included in multivariate PLS modeling to identify variables that were associated with the saliva flow rate. The PLS analysis identified that having a low buffer capacity, being a woman, experiencing a feeling of dryness in the mouth, having less occlusal support zones (Eichner index), and having more daily meals of food was significantly associated with having a low saliva flow rate, whereas having more teeth was associated with higher saliva flow rate (Figs 6, 7).
Variables indicated to be influential for flow rate by PLS were followed up in univariate comparisons with standardization for age and sex as appropriate (Table 3). It was confirmed that women had significantly lower flow rate than men, as did subjects with low buffer capacity (pH <5.5.) versus those with “normal” (pH ≥5.5) capacity. In addition, mean flow rate decreased by increasing experience of dry mouth, and decreasing Eichner index, but only those who had the highest number of meals per day had had a significantly lower flow rate than others (Table 3).
***
***
***
10
Fig 6. Scatter plot of the relation between saliva secretion and the other variables of the study populationFig 7. Coefficient plot of the correlation of stimulated salivation and other variables. Where the error bars do not
11
Table 3. Sex and/or age standardized measures for variables identified as influential for saliva flow rate in themultivariate PLS model, all screening years.
Stimulated salivation mL/min, mean (95% CI) p-value between groups Gender <0.001 Men (n=1,178) 1.73 (1.68 – 1.79) Women (n=1,271) 1.48 (1.42 – 1.53) Buffer capacitya <0.001 pH 5.5 and above (n=1,108) 1.79 (1.74 – 1.85) Below pH 5.5 (n=1,267) 1.45 (1.40 – 1.50) Subjective xerostomia <0.001 Never (n=904) 1.72 (1.67 – 1.78) Occasionally (n=1,215) 1.55 (1.50 – 1.61) Often (n=200) 1.45 (1.32 – 1.58) Always (n=31) 1.25 (0.92 – 1.58) Number of teeth <0.001 0 to 4 (n=157) 1.28 (1.12 – 1.44) 5 to 9 (n=81) 1.22 (1.01 – 1.43) 10 to 14 (n=77) 1.24 (1.02 – 1.45) 15 to 19 (n=172) 1.30 (1.15 – 1.44) 20 to 24 (n=446) 1.50 (1.41 – 1.59) 25 to 28 (n=1,516) 1.74 (1.68 – 1.79) Eichner index <0.001 Class A (n=1,777) 1.71 (1.66 – 1.75) Class B (n=438) 1.33 (1.23 – 1.43) Class C (n=227) 1.30 (1.16 – 1.43) Daily meals 0.013 One (n=142) 1.63 (1.47 – 1.78) Two (n=711) 1.57 (1.50 – 1.64) Three (n=1,060) 1.64 (1.58 – 1.69) Four (n=369) 1.59 (1.49 – 1.69) Five or more (n=53) 1.20 (0.95 – 1.45) a
The analysis method changed between study years (see Method section).
Thirty years time trends in buffer capacity
Due to the shift in analysis method for buffer capacity estimation only measures from 1983 and 1993 could be compared, and those from 2003 and 2013. As seen in Table 4, the proportion of individuals with a low buffer capacity (pH<5.5) was significantly higher in 1983 than 1993, and were lower in 2003 compared to 2013. To evaluate if the lower proportion of subjects in 1993 versus 1983 was associated with saliva flow rate, the saliva flow rate was compared in 10-year age groups of men and women with a buffer pH <5.5 and those with a buffer pH>5.5 (Fig. 8 A,B). In 8 out of 14 strata mean flow rate was significantly higher in those with buffer pH >5.5 than those with a buffer pH <5.5. For some of the age groups a significant difference between buffer capacity and years are observed (Fig. 4).
12
Table 4. Proportion (%) with a saliva buffer capacity pH <5.5.Study year 1983 Study year 1993
p-value
1983 vs 1993 Study year 2003 Study year 2013
p-value 2003 vs 2013 Analysis method Ericsson’s laboratory test Ericsson’s
laboratory test Dentobuff strip
GC Saliva Check, Dentobuff strip
Men 59.3 % 78.3 % <0.001 25.4 % 28.8 % 0.218
Women 66.7 % 88.1 % <0.001 32.0 % 35.3 % 0.228
Differences in proportions are tested with a Chi2-test. Significances are tested by t-test.
Fig 8. Mean (95% CI) of saliva flow rate in (A) men (n=1,149) and (B) women (n=1,229) with buffer pH <5.5 or
≥5.5, all screening years.
* for p-0.05 and ** for p<0.01, *** for p<0.001.
Identification of factors associated with buffer capacity
When all measured variables were included in a multivariate PLS model with low or “normal” saliva buffer capacity as the dependent variables, study year, periodontal health, a higher saliva flow rate, not being a smoker, and number of teeth were significantly associated with having a normal buffer pH (Fig. 9), whereas less occlusal support zones (Eichner index) was associated with low buffer capacity. Variables indicated to be influential for buffer capacity by PLS were followed up in univariate
comparisons with standardization for age and sex as appropriate (Table 5). Thus, more subjects had low buffer capacity among those with low saliva flow rate, among those with the moderate degree of periodontal disease, among smokers, and those with the poorest Eichner index (Table 5).
13
Fig 9. Coefficient plot of the correlation of buffer pH. Negative numbers equals to higher buffer pH. Where theerror bars do not pass through the zero the correlation is significant (p <0.05).
Table 5. Sex and/or age standardized measures for variables identified as influential for saliva
buffer pH in the multivariate PLS model.
Proportion of subjects with buffer pH ≥5.5 p-value between groups Saliva secretion <0.001 ≥ 1.0 mL/min (n=1,818) 84.4 % < 1.0 mL/min (n=557) 69.7 %
Periodontal health classification <0.001
Healthy (n=441) 44.4 % Gingivitis (n=508) 41.1 % Periodontitis levis (n=589) 43.0 % Periodontitis gravis (n=228) 41.7 % Periodontitis complicata (n=93) 49.5 % Smoking <0.001 Smoker (n=1,852) 39.1 % Non-smoker (n=445) 49.4 % Number of teeth <0.001 0 to 4 (n=152) 30.9 % 5 to 9 (n=80) 28.8 % 10 to 14 (n=75) 34.7 % 15 to 19 (n=170) 38.8 % 20 to 24 (n=438) 41.8 % 25 to 28 (n=1,463) 52.2 % Eichner index 0.013 Class A (n=1,719) 51.8 % Class B (n=430) 34.9 % Class C (n=223) 29.6 %
14
DISCUSSION
The present study evaluated if saliva flow rate and saliva buffer capacity has changed in the population from 1983 to 2013, and searched for factors associated with saliva flow rate and having a poor or acceptable saliva buffer capacity, respectively. The main findings were that the present data do not support the hypothesis that saliva flow rate has changed over the 30-year period, but the proportion with poor saliva buffer capacity tended to have increased between 1983 and 1993, where the same analytic method was used.
The strengths of the present study include that (i) all clinical assessments and samplings were performed by experienced and calibrated dentists in well-equipped dental offices at all screening rounds, (ii) that the method for saliva collection was identical at each screening occasion, and (iii) that participants were randomly selected from population registers and represented a wide age range. Besides the weaknesses of a change in the buffer analysis method, it is a weakness that stratification for sex and age resulted in small groups in relation to the variation in the population. This is especially true for the higher age groups. Further, it cannot be excluded that the decreasing attendance rate, i.e. 77.2%, 75.4% 69% and 56.4%, has resulted in an increasing selection bias. The examination was free for the participants, but even with free screenings the attendance by those having a low
socio-economical background is lower compared to others (Zarrouk et al., 2013). Another group that is likely to not have attended for time reasons are busy, healthy people, which in fact is supported by that extra subjects had to be invited in 2013 to get an acceptably sized group in middle-aged individuals. For the present study, it can only be speculated on the potential selection bias, but it may be anticipated that health concerned subjects and those with an oral health related problem (including dry mouth) are overrepresented in later years compared to in 1983 and 1993 when >75% attended the screening, but that subjects with a socioeconomically weak background and worse dental status (but not necessarily impaired saliva flow rate and buffer capacity) might be underrepresented (Wamala et
al., 2006).
The incitement for the present longitudinal evaluation of saliva flow rate was that mean saliva flow rates of 2.5 and 2.0 mL/min were reported for men and women, respectively in a screening performed in 1995 (Bergdahl, 2000). The screening was based on approximately 1,400 Swedish adults. Those flow rates were considerably higher than values reported in studies from the 1970ies (Heintze et
al., 1983). It was plausible to hypothesize that the mean flow rate had increased in the population
given that the number of teeth had increased, people were on average taller (an overall larger anatomy and physiology, thus larger glands), and many medications inhibiting saliva secretion were over the years substituted for non-inhibitory alternatives. Notably, the flow rate in 1993 was higher than in 1983, and approached the levels reported by Bergdahl (2000). It was therefore surprising to find that the mean flow rates were again close to the 1983 levels in 2003 and 2013. Collection of
15
chewing stimulated saliva is a very well standardized and easy-to-perform measure, which makes the variations in flow rate hard to explain with anything but a selection bias in later screening years or regression towards the mean. Still, it seems unlikely that two independent screenings performed in the same time window (early 1990ies) and giving a very similar result would be influenced by the same bias or systematic error. However, at present, the conclusion is that the present data do not support that the saliva flow rate has increased in the population over time, but future large scale, population-based studies have to follow up on this. The implication of an increased salivation, paired with a lower buffer capacity of the same saliva, might confer the need of adjusted cut off values when it comes to normal or sub-normal saliva flow rate, or a completely new way of measuring the properties of saliva, as this would imply that even though the saliva flow rate increased, its buffering capacity would have decreased.A common finding in medical studies is that different measures are not independent, i.e. body weight is correlated with height, saliva flow rate is correlated with buffer capacity, number of teeth is correlated with Eichner index, etc. This is an obstacle in traditional linear or logistic regression. In the present study the multivariate projection method PLS was selected. The rational for choosing that method is that it is insensitive to covariation among variables and it works well also in smaller samples. The drawback is that it does not report standardized central measures and variation. Therefore, PLS was followed up with separate analyses where standardization for age and/or sex was applied.
As described the method for saliva buffer capacity was changed in 2003. The rational for this was to save manpower since the Ericsson method need laboratory personal and logistics when it comes to transporting the saliva. When comparing the proportions with low buffer capacity in 1993 with 2003 and 2013 it is evident that the two methods did not perform as well as claimed by Cheaib et al. (2012). The comparisons for this factor were therefore done with greatest restriction, i.e. dichotomized to poor or good.
The factors associated with saliva flow rate, i.e. sex, age in women, subjective xerostomia, number of teeth, occlusal support zones and buffer capacity, are in accordance with what has been reported in previous studies (Ikebe et al., 2012; Heintze et al., 1983). Thus, though the results in the present study are mainly confirmatory, they are of significant clinical importance. Subjective xerostomia is not an explanatory factor, but rather a symptom of possible hyposalivation. Notably, both number of teeth and Eichner index correlated with stimulated saliva secretion, a finding that should be remembered when choosing between the options of extracting a tooth or save it. Thus, quality of life and medical/biological aspects must be considered, and likely more so in patients with impaired saliva secretion, those taking medicine causing hyposalivation, those that have short dental arches or previous extractions or tooth aplasia. One may speculate if fixed prosthodontics should be
16
encouraged in patients with lost teeth, to improve the occlusal support zones and hence salivasecretion. Further studies are needed to evaluate this.
CONCLUSION
The present study found chewing stimulated saliva secretion to be associated with sex, number of teeth, Eichner index, subjective dry mouth and buffer capacity, but it does not support that saliva secretion has increased over time in the population. Saliva secretion, and its buffer components, is of outmost importance for the oral environment. Hyposalivation and an associated low buffer capacity puts a person at risk for both dental diseases and oral infection, including candida infections, and overall health problems as acute respiratory infection (Iwabuchi et al., 2012) and an increase over time might reduce the risk. Given that several factors associated with improved saliva flow rate has taken place in the population, future studies should follow up on the results to confirm or reject that the present results are not due to a selection bias.
ACKNOWLEDGEMENTS
Prof. Ingegerd Johansson DDS PhD, without whom this essay could not have been written. The personal support of Dr Elisabeth Wärnberg Gerdin, DDS PhD.
The board of the Jönköping Studies, especially the chairman Dr Ola Norderyd, DDS PhD. The County Council of Västerbotten who allowed me to take time to write this essay.
17
REFERENCES
Bardow A, Lagerlöf F, Nauntofte B, Tenovuo J. The role of saliva. In: Fejerskov O, Kidd E (eds.). Dental
Caries: The Disease and Its Clinical Management. 2nd ed. Oxford: Wiley-Blackwell: 2008. 189-208.
Bergdahl M. Salivary flow and oral complaints in adult dental patients. Community Dent Oral Epidemiol 2000:28:59-66.
Birkhed D, Heintze U. Salivary secretion rate, buffer capacity, and pH. In: Tenovuo J (ed.). Human Saliva: Clinical Chemistry and Microbiology, Vol I. Boca Raton (FL): CRC Press: 1989. 25-73. Carpenter GH. The secretion, components, and properties of saliva. Annu Rev Food Sci Technol
2013:4:267-276.
Cheaib Z, Ganss C, Lamanda A, Turgut MD, Lussi A. Comparison of three strip-type tests and two laboratory methods for salivary buffering analysis. Odontology 2012:100:67-75.
Dawes C, Wood CM. The contribution of oral minor mucous gland secretions to the volume of whole saliva in man. Arch Oral Biol 1973:18:337-342.
Edgar M, Dawes C, O’Mullane D (eds.). Saliva and Oral Health. 3rd
ed. London: British Dental Association. 2004.
Eichner K. Über eine gruppeneinteilung der lückengebisse für der prothetik. Dtsch Zahnärztl Z 1955: 10:1831-1834.
Ekström J. Autonomic control of salivary secretion. Proc Finn Dent Soc 1989:85:323-331: 361-363. Ericsson Y. Clinical investigations of the salivary buffering action. Acta Odontol Scand 1959:17:131–
165.
Fujibayashi T, Sugai S, Miyasaka N, Hayashi Y, Tsubota K. Revised Japanese criteria for Sjögren's syndrome (1999): availability and validity. Mod Rheumatol 2004:14:425-434
Garbacz G, Kołodziej B, Koziolek M, Weitchies W, Klein S. An automated system for minitoring and regulating the pH of bicarbonate buffers. AAPS PharmSciTech 2013:14:517-522.
Heintze U, Birkhed D, Björn H. Secretion rate and buffer effect of resting and stimulated whole saliva as a function of age and sex. Swed Dent J 1983:7:227–238.
Hugoson A, Koch G, Bergendal T, Hallonsten AL, Laurell L, Lundgren D, et al. Oral health of individuals aged 3-80 years in Jonkoping, Sweden in 1973 and 1983. I. A review of findings on dental care habits and knowledge of oral health. Swed Dent J 1986:10:103-117.
Hugoson A, Koch G, Bergendal T, Hallonsten AL, Slotte C, Thorstensson B, et al. Oral health of individuals aged 3-80 years in Jonkoping, Sweden in 1973, 1983, and 1993. I. Review of findings on dental care habits and knowledge of oral health. Swed Dent J 1995:19:225-241
Hugoson A, Koch G, Göthberg C, Helkimo AN, Lundin SA, Norderyd O, Sjödin B, Sondell K. Oral health of individuals aged 3-80 years in Jönköping, Sweden during 30 years (1973-2003). I. Review of findings on dental care habits and knowledge of oral health. Swed Dent J 2005a:29:125-138. Hugoson A, Koch G, Göthberg C, Helkimo AN, Lundin SA, Norderyd O et al. Oral health of individuals
aged 3–80 years in Jonkoping, Sweden during 30 years (1973–2003). II. Review of clinical and radiographic findings. Swed Dent J 2005b:29:139–155.
18
Ikebe K, Matsuda K, Kagawa R, Enoki K, Okada T, Yoshida M, Maeda Y. Masticatory performance inolder subjects with varying degrees of tooth loss. J Dent 2012:40:71-76.
Iwabuchi H, Fujibayashi T, Yamane GY, Imai H, Nakao H. Relationship between hyposalivation and acute respiratory infection in dental outpatients. Gerontology 2012:58:205-11.
Jensen Kjeilen JC, Brodin P, Aars H, Berg T. Parotid salivary flow in response to mechanical and gustatory stimulation in man. Acta Physiol Scand 1987:131:169-175.
Koch G, Kreiborg S, Andreasen JO. Eruption and shedding of teeth. In: Koch G, Poulsen S (eds.).
Pediatric Dentistry: A clinical approach. 2nd ed. Oxford: Wiley-Blackwell: 2009. 197-211.
Marsh PD. Microbiology of dental plaque biofilms and their role in oral health and caries. Dent Clin
North Am 2010:54:441-454.
Närhi TO. Prevalence of subjective feelings of dry mouth in the elderly. J Dent Res 1994:73:20–25. Österberg T, Birkhed D, Johansson C, Svanborg A. Longitudinal study of stimulated whole saliva in an
elderly population. Scand J Dent Res 1992:100:340-345.
Samnieng P, Ueno M, Shinada K, Zaitsu T, Wright FA, Kawaguchi Y. Association of hyposalivation with oral fuction, nutrition and oral health in community-dwelling elderly Thai. Community Dent Health 2012:29:117-123.
Schuurs A. Pathology of the hard dental tissues. West Sussex: Wiley-Blackwell: 2013.
Scott BJ, Jassamwalia R, Linden RW. The masticatory-parotid salivary reflex in edentulous subjects. J
Oral Rehabil 1998:25:28-33.
Selwitz RH, Ismail AI, Pitts NB. Dental caries. Lancet 2007:369:51-59.
Ståhle L, Wold S. Multivariate data analysis and experimental design in biomedical research. Prog Med
Chem 1988:25:291-338.
Wamala S, Merlo J, Boström G. Inequity in access to dental care services explains current
socioeconomic disparities in oral health: the Swedish National Surveys of Public Health 2004-2005. J
Epidemiol Community Health 2006:60:1027-1033.
Zarrouk M, Holst J, Malina M, Lindblad B, Wann-Hansson C, Rosvall M, Gottsäter A. The importance of socioeconomic factors for compliance and outcome at screening for abdominal aortic aneurysm in 65-year-old men. J Vasc Surg 2013:58:50-55.
