Social, Psychological, and Behavioral Predictors of Salivary Bacteria, Yeast in Caries-Free Children

Abstract

Introduction:

Early childhood caries (ECC) is a complex, multifactorial oral disease that is a major public health concern because it is prevalent, profoundly alters a child’s quality of life, is difficult to treat effectively, and has a distressing tendency to recur following treatment.

Objectives:

The purpose of the study was to examine social, psychological, and behavioral predictors of salivary bacteria and yeast in young children at risk for ECC.

Methods:

A sample of 189 initially caries-free preschool children was assessed for child stress physiology from salivary cortisol, child and family stress exposure, diet, oral health behaviors, and sociodemographic risks. Multiple logistic regression analysis was implemented to examine the associations between these risk factors and cariogenic microorganisms: mutans streptococci (MS), lactobacilli (LB), and Candida species.

Results:

Higher baseline salivary cortisol (odds ratio [OR] = 6.26; 95% confidence level [CL], 1.69–23.16) and a blunted response to an acute laboratory stressor (OR = .56; 95% CL, .37–.83) were associated with an increased likelihood of elevated salivary MS (≥10⁵ colony-forming units/mL) in caries-free children. Sociodemographic risk for cariogenic microorganisms was also found. Specifically, lower education attainment of the parent/primary caregiver was associated with children being more likely to carry salivary Candida species and elevated salivary MS; in addition, children from households with an unemployed parent/primary caregiver were more likely (OR = 3.13; 95% CL, 1.2–8.05) to carry salivary Candida species and more likely (OR = 3.03; 95% CL, 1.25–7.33) to carry elevated levels of MS and/or salivary Candida and/or LB.

Conclusions:

The impact of sociodemographic risk and stress physiology on cariogenic disease processes are evident prior to ECC onset. The findings provide novel data on the early onset of cariogenic processes in children and the importance of considering sociodemographic, psychosocial, and behavioral factors when judging ECC risk.

Knowledge Transfer Statement:

The findings provide valuable and novel findings that, pre-ECC onset, the caries disease process is explicable from a detailed assessment of behavioral, sociodemographic, and psychosocial stress variables.

Introduction

Early childhood caries (ECC) is a complex, multifactorial oral disease that is a major public health concern because it is prevalent (Schroth et al. 2005; Dye et al. 2007; Kopycka-Kedzierawski et al. 2008), profoundly alters a child’s quality of life, is difficult to treat effectively, and has a distressing tendency to recur following treatment. Epidemiological and clinical studies indicate that ECC has a strong social class gradient. This association has often been interpreted as suggesting a role for psychosocial stress in the disease process, but it may also reflect overlapping risks associated with, for example, diet and oral health behaviors (Finlayson et al. 2007a, 2007b; Seow et al. 2009; Fontana 2015; Kirthiga et al. 2019). The degree to which these different sources of risk are causally associated with ECC remains unclear due to methodological limitations of existing studies, such as cross-sectional designs and lack of attention to confounding. One novel approach for advancing the field is to identify predictors of candidate risks for the cariogenic disease process underlying ECC. Specifically, numerous studies indicate that ECC results from tooth-specific bacteria (mainly mutans streptococci [MS] and lactobacilli [LB]) that metabolize dietary sugars to produce acid. Over time and with increased acid exposure, the tooth structure demineralizes, resulting in caries formation (Loesche 1986; Law et al. 2007; Plonka et al. 2012; Weber-Gasparoni et al. 2012; Kirthiga et al. 2019). High MS and LB counts predict caries onset in young children (Loesche 1986; Laitala et al. 2012; Plonka et al. 2012) and may be interpreted as part of the causal pathway. In addition to MS and LB species, classic cariogenic microorganisms that are responsible for dental caries onset and progression, Candida, a diploid fungus, may further exacerbate caries process, as suggested in numerous animal and human studies (Signoretto et al. 2009; Raja et al. 2010; Klinke et al. 2011; Xiao et al. 2016; Xiao et al. 2018; de Jesus et al. 2020). We capitalize on this established microbiological model to advance our understanding of the sociodemographic, psychosocial, diet, and oral health behavior sources of cariogenic risk. Importantly, we do this in a sample of caries-free children, aged 1 to 3 y, thereby ensuring that risk exposure precedes caries disease onset.

The purpose of the study was to examine social, psychological, and behavioral predictors of cariogenic salivary bacteria and yeast in young children at risk for ECC. In addition to assessing psychosocial stress exposure of the child, we assessed children’s cortisol reactivity, a measure of stress physiology and a candidate biological mediator of stress exposure on ECC development (Boyce et al. 2010; Caruso et al. 2018).

Methods

A cohort study sample of 189 initially caries-free, Medicaid- and Child Health Plus–eligible preschool children and their primary caregivers was drawn from patients, aged 1 to 3 y, who presented to the Division of Pediatric Dentistry in the Eastman Institute for Oral Health (EIOH) for regular dental care. Healthy children and their parents/primary caregivers or legal guardians 18 y of age or older, irrespective of gender, ethnic origin, or race, were eligible to participate in the study. They were derived from approximately 800 parents/primary caregivers who were approached to participate in the study over the course of 3 y; most individuals who opted not to participate in the study stated lack of interest, had limited free time due to schoolwork or employment, or had language barriers. Approximately 100 of the approached participants scheduled a study visit but did not present for the study visit and were not enrolled. The study was approved by the University of Rochester Medical Center Research Subject Review Board (RSRB#57726). Written informed consent was obtained from parents/guardians of participating children prior to the study. The International Caries Detection and Assessment System (ICDAS) was used to assess dental caries status (Ismail et al. 2007). Only children with ICDAS = 0, who were screened and examined by a calibrated pediatric dental examiner, were enrolled in the study.

Approximately 2 mL of saliva was collected from study participants to examine their microbiological profile, including MS, LB, and Candida species. Whole, stimulated saliva samples were obtained through a disposable saliva ejector attached to a 50-mL sterile centrifuge tube, which in turn was attached to a vacuum pump (Mundorff et al. 1990). All salivary samples were processed the same day using an Autoplate Spiral Plating System (Advanced Instruments, Inc.). Saliva samples were evaluated for MS, LB, and Candida species using Mitis Salivarius agar plus bacitracin (MSB), Rogosa, and CHROMagar, respectively. For dispersion of cell clumps and dechaining of streptococci, a 2-mL aliquot of saliva was sonicated (three 10-s sonic bursts, 100 W of peak power). The suspension was serially diluted (10-fold dilutions) in phosphate buffer and 50-µL aliquots uniformly plated on 1) Mitis Salivarius (MS) agar (Beckton-Dickinson) supplemented with 20% sucrose and bacitracin (0.2 U/mL) to determine the presence of MS, 2) Rogosa agar (Oxoid) to evaluate the presence of lactobacilli, 3) CHROMagar (BBL) supplemented with 0.1 mg/mL chloramphenicol to determine the presence of Candida spp., and 4) tryptic soy agar supplemented with 5% sheep blood to enumerate the total microflora. MS plates were incubated at 37°C in a 5% CO₂ atmosphere, Rogosa plates were incubated anaerobically at 37°C, CHROMagar plates were incubated aerobically at 37°C, and duplicated blood agar plates were incubated at 37°C under aerobic and anaerobic conditions. All dilutions were plated in triplicates and the plates incubated under the described conditions for 72 h before colonies were counted. The number of MS, LB, Candida spp., and total microflora was expressed as colony-forming units (CFUs) per milliliter of saliva. Detailed procedures for collection, storage, and analysis of oral microflora are well established and described elsewhere (Mundorff et al. 1990). Salivary samples were processed between 1 and 4 h after collection, as there is no significant loss of numbers of total viable flora during the first 24 h (Mundorff et al. 1990).

Child stress physiology, based on salivary cortisol, was measured from 3 salivary collections separate from those described for microbiology. Salimetrics Children’s Swabs (Salimetrics) were used to collect saliva samples prior to a standard laboratory stressor that has been widely used in behavioral science studies (Spangler and Grossmann 1993; O’Connor et al. 2013) and again 15 and 30 min following completion of the procedure. Following collection, the samples were transported to the laboratory on ice and stored at –20°C until processed in triplicate using enzyme-linked immunosorbent assay (Salimetrics). Standard curves were fit to a 5-parameter logistic for determination of cortisol concentration in saliva (elisaanalysis.com). We took considerable care in collecting saliva samples using established procedures to avoid confounds and contamination (e.g., saliva cortisol was obtained at the beginning of the visit and was not confounded by the subsequent dental examination; no food or beverages were consumed within 30 min of saliva collection). We considered the baseline and reactivity measures of cortisol as well as the area under the curve in analyses below. Time of awakening and time of first sample collection were considered covariates because of the strong diurnal influence on cortisol levels.

Child and family stress exposure was based on parent-reported measures of stress from multiple sources; each is widely used in studies of psychosocial stress. Specific measures included parental depression using the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff 1977), anxiety/worry based on the Penn State Worry Questionnaire (PSWQ; Meyer et al. 1990), alcohol use from the Alcohol Use Disorders Identification Test (AUDIT; Babor et al. 2001), and stressful life events from a list of standard high-stress conditions (e.g., losses of income, health problems; Compas et al. 1989). Household disorganization and confusion were derived from the CHAOS (Confusion, Hubbub and Order Scale) scale (Matheny et al. 1995), violence exposure was based on the psychological aggression and physical assault subscales of the Conflict Tactics Scale (Straus et al. 1996), and caregiver social support was based on the Interpersonal Support and Evaluation List (ISEL; Cohen and Hoberman 1983).

Diet, Oral Health Behaviors, and Sociodemographic Variables

Favorite snacks/drinks and oral health behaviors were identified via parental questionnaire, which included questions related to the child’s eating and drinking habits, snacking choices, sippy cup use, and the type and amount of beverages and snacks consumed by the children; we also assessed current breastfeeding, formula feeding practices, and oral hygiene regimen (Kressin et al. 2009). Sociodemographic variables included the child’s and primary caregivers’/legal guardians’ age, race, ethnicity, gender, employment status, income, education, insurance status, daycare attendance, and the number of individuals who resided in the household.

This article reports the analysis of baseline data of a prospective cohort study. A sample size of 201 was calculated based on the number of children who were projected to convert to caries status in the course of a 2-y follow-up period. Of the 201 children enrolled, 11 were screen failures and were discontinued from the study, resulting in a baseline study sample of 189 children.

This article also reports the descriptive data for the sociodemographic, psychosocial, and oral hygiene behaviors of initially caries-free children and their parents/primary caregivers. Mean and standard deviation are reported for each continuous variable; frequencies are reported for categorical variables. Oral microbiological factors were dichotomized in the following manner (Epstein et al. 1980; Zoitopoulos et al. 1996; Seow 1998): MS <10⁵ versus MS ≥10⁵, LB = 0 versus LB >0, and Candida = 0 versus Candida >0. In addition, a composite measure of microbiology was defined as MS <10⁵ and LB = 0 and Candida = 0 versus MS ≥10⁵ or LB >0 or Candida >0. Children were categorized as having low levels of MS or high levels of MS in saliva. A high level of salivary MS was defined as 10⁵ CFU/mL or greater, as 18% of the children fell into the higher MS category. Approximately 84% of children had undetectable LB in saliva, and 67% of children had undetectable Candida species. Given these observations, the cutoff point of 10⁵ CFU/mL of saliva between high and low salivary MS levels was chosen, and carriage or not of LB and Candida species was chosen. A 2-sample t test was used to compare the mean values of continuous variables of 2 groups. For each dichotomized outcome variable, the Cohen’s d effect size was calculated for each continuous covariate between 2 subgroups.

Pearson’s χ² test (or Fisher’s exact test) was used to compare the frequencies of categorical variables of 2 groups.

Bivariate analyses were used to study the association between the 4 dichotomized microbiological outcomes (MS <10⁵ versus MS ≥10⁵; LB = 0 versus LB >0; Candida = 0 versus Candida >0; MS <10⁵ and LB = 0 and Candida = 0 versus MS ≥10⁵ or LB >0 or Candida >0; i.e., any elevation in the microbiological measures) and sociodemographic, psychosocial, and oral hygiene variables of interests. Multiple logistic regression analysis was implemented (with backward stepwise model selection) to examine the combined effects of covariates selected from bivariate analyses. We also include, on an a priori basis, child gender and race/ethnicity; other variables are retained based on the backward elimination method.

The significance level of each analysis was set at 0.05. All analyses were implemented with SAS 9.4 (SAS Institute).

Results

Descriptive Statistics and Bivariate Associations

The demographic characteristics of 189 caries-free children are presented in Table 1. The mean (SD) age of the children was 29.5 (9.1) mo. Almost 75% of the families lived in an urban setting; 64.5% of the parents/primary caregivers’ educational attainment was less than high school, high school equivalent, or General Educational Development (GED). The findings make clear that the study sample is at high sociodemographic risk, based on measures of income, education, and insurance status.

Table 1.

Demographic Characteristics of the Study Participants.

Participant Characteristics	Value
Parent’s gender
Male	15 (7.9)
Female	174 (92.1)
Parent’s race
Black	87 (46.0)
White	50 (26.5)
Mixed	37 (19.6)
Other	15 (7.9)
Parent’s ethnicity
Hispanic	40 (21.2)
Not Hispanic	145 (76.7)
Parent’s dental insurance
Medicaid	151 (79.9)
Private	23 (12.2)
No insurance	15 (7.9)
Education
Less than high school	17 (9.0)
High school or GED	105 (55.5)
College	62 (32.8)
Marital status
Single	104 (55.0)
Married or cohabitating	85 (45.0)
Residence
Urban	140 (74.1)
Suburban	40 (21.2)
Other	9 (4.8)
No. of people in household, mean ± SD	4.1 ± 1.5
Parent’s age, mean ± SD, y	30.2 ± 6.6
Child’s gender
Male	98 (51.9)
Female	91 (48.2)
Child’s race
Black	76 (40.2)
White	35 (18.5)
Mixed	65 (34.4)
Other	13 (6.9)
Child’s ethnicity
Hispanic	50 (26.5)
Not Hispanic	136 (72.0)
Child’s dental insurance
Medicaid	143 (75.7)
Child Health Plus	35 (18.5)
Private	10 (5.3)
No insurance	1 (0.5)
Current work status
Employed	108 (57.1)
Unemployed	80 (42.3)
Parental dental health status
Excellent	25 (13.2)
Very good	34 (18.0)
Good	67 (35.5)
Fair	41 (21.7)
Poor	21 (11.1)
Household income per week
Less than $200	42 (22.2)
$200–$400	42 (22.2)
$401 and above	103 (54.5)
Child’s age, mean ± SD, mo	29.5 ± 9.1

Values are presented as n (%) unless otherwise indicated.

GED, General Educational Development.

In these caries-free children, 73% carried MS, 21% were positive for the yeast Candida albicans and/or non-albicans Candida, and 16% were positive for lactobacilli. Out of the 24 children who carried non-albicans Candida, 42% carried Candida tropicalis, 63% Candida glabrata, 8% Candida dubliniensis, and 4% Candida krusei. Data on bacteria and Candida species are presented in the Appendix.

Bivariate analyses for categorical variables are presented separately by microbiological target and the combined targets (data shown in the Table 2). Among sociodemographic risks, associations with microbiological outcomes were most consistent for education: children in families headed by parents with low education exhibited elevated MS and presence of Candida and LB. Less consistent but notable were the associations between work status, income, education, and daycare attendance with Candida species. Notably absent, for all microbiological outcomes, was an association with child sex, race, or ethnicity. Frequency of snacks was very high but did not differentiate between those with positive or elevated levels of microbiological outcomes.

Table 2.

Bivariate Analyses by Salivary Mutans Streptococci Levels, Lactobacilli, and Candida Species Status.

	Mutans Streptococci			Lactobacilli			Candida Species			MS (Cutoff 105 CFU) or Lactobacilli or Candida Species
Variables	MS <105 CFU, n (%)	MS ≥105 CFU, n (%)	P Value	Absent, n (%)	Present, n (%)	P Value	Absent, n (%)	Present, n (%)	P Value	Absent, n (%)	Present, n (%)	P Value
Child’s age, mo
12–24	52 (34.9)	8 (24.24)	0.02 a	46 (30.87)	12 (42.86)	0.03 a	35 (29.41)	23 (38.98)	0.02 b	27 (28.13)	32 (38.55)	0.009 b
25–36	55 (36.91)	21 (63.64)		60 (40.27)	14 (50)		46 (38.66)	29 (49.15)		36 (37.5)	39 (46.99)
37–48	42 (28.19)	4 (12.12)		43 (28.86)	2 (7.14)		38 (31.93)	7 (11.86)		33 (34.38)	12 (14.46)
Child’s gender
Female	70 (46.98)	15 (45.45)	0.87 b	70 (46.98)	11 (39.29)	0.45 b	60 (50.42)	21 (35.59)	0.06 b	46 (47.92)	36 (43.37)	0.54 b
Male	79 (53.02)	18 (54.55)		79 (53.02)	17 (60.71)		59 (49.58)	38 (64.41)		50 (52.08)	47 (56.63)
Child’s ethnicity
Hispanic	42 (28.38)	5 (15.63)	0.14 b	39 (26.35)	6 (22.22)	0.65 b	33 (27.73)	12 (21.05)	0.34 b	25 (26.04)	20 (24.69)	0.84 b
Non-Hispanic	106 (71.62)	27 (84.38)		109 (73.65)	21 (77.78)		86 (72.27)	45 (78.95)		71 (73.96)	61 (75.31)
Child’s race
African American	59 (39.6)	15 (45.45)	0.38 b	59 (39.6)	12 (42.86)	0.8 b	45 (37.82)	27 (45.76)	0. 41 b	40 (41.67)	33 (39.76)	0.78 b
Caucasian	26 (17.45)	8 (24.24)		26 (17.45)	7 (25)		25 (21.01)	8 (13.56)		19 (19.79)	14 (16.87)
Mixed	64 (42.95)	10 (30.3)		64 (42.95)	9 (32.14)		49 (41.18)	24 (40.68)		37 (38.54)	36 (43.37)
Is your child currently attending daycare?
No	90 (61.64)	22 (68.75)	0.45 b	88 (60.27)	22 (81.48)	0.04 b	68 (58.12)	43 (75.44)	0.03 b	55 (57.89)	56 (70)	0.10 b
Yes	56 (38.36)	10 (31.25)		58 (39.73)	5 (18.52)		49 (41.88)	14 (24.56)		40 (42.11)	24 (30)
Education status
Less than high school	13 (8.84)	4 (12.9)	0.08 b	13 (8.97)	3 (10.71)	0.71 a	8 (6.96)	9 (15.25)	0.03 a	5 (5.32)	12 (14.81)	0.02 b
High school or GED	80 (54.42)	22 (70.97)		83 (57.24)	18 (64.29)		63 (54.78)	38 (64.41)		51 (54.26)	50 (61.73)
College or graduate	54 (36.73)	5 (16.13)		49 (33.79)	7 (25)		44 (38.26)	12 (20.34)		38 (40.43)	19 (23.46)
Income per week
Less than $200	31 (21.09)	10 (30.3)	0.67 a	33 (22.45)	6 (21.43)	0.07 a	24 (20.17)	16 (28.07)	0.01 a	17 (17.71)	24 (29.63)	0.01 b
$200–$400	31 (21.09)	9 (27.27)		27 (18.37)	12 (42.86)		21 (17.65)	18 (31.58)		16 (16.67)	23 (28.4)
$401–$600	28 (19.05)	5 (15.15)		28 (19.05)	3 (10.71)		26 (21.85)	5 (8.77)		22 (22.92)	9 (11.11)
$601–$800	22 (14.97)	3 (9.09)		24 (16.33)	1 (3.57)		18 (15.13)	7 (12.28)		16 (16.67)	9 (11.11)
$801–$1,000	16 (10.88)	4 (12.12)		16 (10.88)	4 (14.29)		11 (9.24)	9 (15.79)		9 (9.38)	11 (13.58)
More than $1,000	19 (12.93)	2 (6.06)		19 (12.93)	2 (7.14)		19 (15.97)	2 (3.51)		16 (16.67)	5 (6.17)
Current work status
Employed	89 (59.73)	16 (48.48)	0.24 b	91 (61.07)	11 (39.29)	0.03 b	78 (65.55)	24 (40.68)	0.002 b	63 (65.63)	39 (46.99)	0.01 b
Unemployed	60 (40.27)	17 (51.52)		58 (38.93)	17 (60.71)		41 (34.45)	35 (59.32)		33 (34.38)	44 (53.01)
Marital status
Married	40 (26.85)	4 (12.12)	0.04 a	36 (24.16)	8 (28.57)	0.94 a	32 (26.89)	12 (20.34)	0.85 a	24 (25)	20 (24.1)	0.76 a
Single	81 (54.36)	22 (66.67)		83 (55.7)	16 (57.14)		64 (53.78)	36 (61.02)		51 (53.13)	50 (60.24)
Divorced	13 (8.72)	0		11 (7.38)	1 (3.57)		8 (6.72)	4 (6.78)		8 (8.33)	4 (4.82)
Cohabitating	14 (9.4)	7 (21.21)		18 (12.08)	3 (10.71)		14 (11.76)	7 (11.86)		12 (12.5)	9 (10.84)
Widowed	1 (0.67)	0		1 (0.67)	0		1 (0.84)	0		1 (1.04)	0
Residence
Urban	108 (72.48)	27 (81.82)	0.67 a	108 (72.48)	22 (78.57)	0.51 a	86 (72.27)	45 (76.27)	0.88 a	71 (73.96)	61 (73.49)	1.0 a
Nonurban	33 (22.15)	5 (15.15)		34 (22.82)	4 (14.29)		27 (22.69)	11 (18.64)		20 (20.83)	18 (21.69)
Other	8 (5.37)	1 (3.03)		7 (4.7)	2 (7.14)		6 (5.04)	3 (5.08)		5 (5.21)	4 (4.82)
Caregiver’s dental health
Excellent	21 (14.19)	4 (12.12)	0.39 a	22 (14.86)	3 (10.71)	0.99 a	15 (12.71)	10 (16.95)	0.33 b	14 (14.74)	11 (13.25)	0.39 a
Very good	27 (18.24)	5 (15.15)		27 (18.24)	5 (17.86)		24 (20.34)	8 (13.56)		20 (21.05)	12 (14.46)
Good	48 (32.43)	17 (51.52)		51 (34.46)	11 (39.29)		37 (31.36)	26 (44.07)		28 (29.47)	36 (43.37)
Fair	34 (22.97)	5 (15.15)		31 (20.95)	6 (21.43)		27 (22.88)	10 (16.95)		22 (23.16)	15 (18.07)
Poor	18 (12.16)	2 (6.06)		17 (11.49)	3 (10.71)		15 (12.71)	5 (8.47)		11 (11.58)	9 (10.84)
Does your child take liquids from a sippy cup?
Yes	101 68.24)	25 (75.76)	0.40 a	98 (66.22)	24 (85.71)	0.04 a	75 (63.03)	48 (82.76)	0.01 b	59 (61.46)	65 (79.27)	0.39 a
No	47 (31.76)	8 (24.24)		50 (33.78)	4 (14.29)		44 (36.97)	10 (17.24)		37 (38.54)	17 (20.73)
How often does your child breastfeed at night?
More than once a night	6 (4.05)	0	0.67 a	2 (1.35)	4 (14.29)	0.007 a	5 (4.24)	1 (1.69)	0.77 a	2 (2.11)	4 (4.82)	0.42 a
A few times a year	1 (0.68)	0		1 (0.68)	0		1 (0.85)	0		1 (1.05)	0
Never	141 (95.27)	33 (100)		145 (97.97)	24 (85.71)		112 (94.92)	58 (98.31)		92 (96.84)	79 (95.18)
How often does your child have milk formula at night?
More than once a night	3 (2.01)	1 (3.03)	0.92 a	3 (2.01)	1 (3.57)	0.73 a	3 (2.52)	1 (1.69)	0.48 a	1 (1.04)	3 (3.61)	0.16 a
Once every night	22 (14.77)	5 (15.15)		23 (15.44)	3 (10.71)		20 (16.81)	6 (10.17)		19 (19.79)	8 (9.64)
Several times a week	2 (1.34)	0		2 (1.34)	0		1 (0.84)	1 (1.69)		1 (1.04)	1 (1.2)
Several times a month	3 (2.01)	1 (3.03)		3 (2.01)	1 (3.57)		2 (1.68)	2 (3.39)		1 (1.04)	3 (3.61)
A few times a year	1 (0.67)	0		1 (0.67)	0		0	1 (1.69)		0	1 (1.2)
Never	118 (79.19)	26 (78.79)		117 (78.52)	23 (82.14)		93 (78.15)	48 (81.36)		74 (77.08)	67 (80.72)
Does your child have snacks most days?
Yes	148 (99.33)	33 (100)	1.0 a	148 (99.33)	28 (100)	1.0 a	119 (100)	58 (98.31)	0.33 a	96 (100)	82 (98.8)	0.46 a
No	1 (0.67)	0		1 (0.67)	0		0	1 (1.69)		0	1 (1.2)
Does your child share a toothbrush with anyone?
Yes	3 (2.01)	1 (3.03)	0.55 a	2 (1.34)	2 (7.14)	0.12 a	2 (1.68)	2 (3.39)	0.60 a	1 (1.04)	3 (3.61)	0.34 a
No	146 (97.99)	32 (96.97)		147 (98.66)	26 (92.86)		117 (98.32)	57 (96.61)		95 (98.96)	80 (96.39)
Do you or other adult brush your child’s teeth?
Yes	141 (95.27)	32 (96.97)	1.0 a	140 (94.59)	28 (100)	0.36 a	112 (94.12)	57 (98.28)	0.27 a	89 (92.71)	81 (98.78)	0.07 a
No	7 (4.73)	1 (3.03)		8 (5.41)	0		7 (5.88)	1 (1.72)		7 (7.29)	1 (1.22)
How often do you or another adult clean or brush your child’s teeth?
More than once a day	76 (51.01)	15 (45.45)	0.73 a	70 (46.98)	18 (64.29)	0.11 a	53 (44.54)	35 (59.32)	0.25 a	42 (43.75)	47 (56.63)	0.39 a
Every day	62 (41.61)	15 (45.45)		68 (45.64)	7 (25)		55 (46.22)	21 (35.59)		45 (46.88)	31 (37.35)
Several times a week	9 (6.04)	2 (6.06)		9 (6.04)	2 (7.14)		9 (7.56)	2 (3.39)		7 (7.29)	4 (4.82)
Less than several times a week	2 (1.34)	1 (3.03)		2 (1.34)	1 (3.57)		2 (1.68)	1 (1.69)		2 (2.08)	1 (1.2)
Does anyone in your household smoke?
Yes	30 (20.13)	8 (24.24)	0.60 b	30 (20.13)	8 (28.57)	0.32 b	26 (21.85)	12 (20.34)	0.82 b	19 (19.79)	19 (22.89)	0.61 b
No	119 (79.87)	25 (75.76)		119 (79.87)	20 (71.43)		93 (78.15)	47 (79.66)		77 (80.21)	64 (77.11)

CFU, colony-forming units; GED, General Educational Development; MS, mutans streptococci.

^aFisher’s exact test.

^bPearson’s χ² test.

The results of the bivariate analyses of psychosocial variables, salivary cortisol, and the 3 dichotomized microbiological outcomes are included in the Appendix. The Cohen’s d effect sizes varied widely across categories of risk and within risk categories. For psychosocial variables, associations across the sources of child psychosocial stress (parent symptoms, family conflict, and disruption) were small and nonsignificant.

Multivariable Regression Analyses

Regression models are depicted in Table 3 for each microbiological outcome variable. As shown in Table 3, high initial salivary cortisol (measured prior to the stressor; odds ratio [OR] = 6.26, P = 0.006) and blunted cortisol response were associated with an increased likelihood (OR = 6.26) of elevated (MS ≥10⁵) salivary MS. Elevated salivary MS was associated with lower education of the parent/primary caregiver, independently of other factors in the model.

Table 3.

Results of the multiple regression with MS < 10⁵ versus MS ≥10⁵; LB = 0 versus LB >0; Candida = 0 versus Candida >0 and MS <10⁵, LB = 0, Candida = 0 versus MS ≥10⁵ or LB >0 or Candida >0 as outcome variables.

Odds Ratio Estimates
	MS <105 vs. MS ≥105				LB = 0 vs. LB >0				Candida = 0 vs. Candida >0				MS <105, LB = 0, Candida = 0 vs. MS ≥105 or LB >0 or Candida >0
Effect	Point Estimate	95% CI		P Value	Point Estimate	95% CI		P Value	Point Estimate	95% CI		P Value	Point Estimate	95% CI		P Value
Age of the child	0.995	0.935	1.059	0.876	0.939	0.874	1.009	0.088	0.961	0.914	1.011	0.125	0.965	0.921	1.012	0.138
White vs. African American	2.824	0.657	12.130	0.121	1.431	0.306	6.698	0.442	0.786	0.225	2.753	0.664	1.389	0.443	4.358	0.883
Mixed vs. African American	0.870	0.237	3.190	0.313	0.643	0.146	2.825	0.394	1.047	0.370	2.960	0.750	1.645	0.598	4.521	0.493
Not Hispanic vs. Hispanic	3.709	0.758	18.163	0.106	1.240	0.277	5.551	0.778	2.223	0.699	7.070	0.176	2.083	0.697	6.226	0.189
Male vs. female	1.220	0.436	3.415	0.705	1.501	0.486	4.639	0.480	2.004	0.860	4.673	0.107	1.482	0.675	3.255	0.327
Attending day care center	0.770	0.222	2.674	0.681	0.260	0.059	1.133	0.073	0.692	0.269	1.780	0.445	0.773	0.328	1.823	0.556
Initial salivary cortisol	6.260	1.692	23.158	0.006	0.182	0.015	2.226	0.183	0.652	0.159	2.676	0.553	1.431	0.502	4.075	0.503
Area under the curve	0.558	0.374	0.833	0.004	0.855	0.561	1.302	0.466	0.778	0.564	1.075	0.128	0.734	0.537	1.002	0.052
Depressive symptoms	1.022	0.951	1.098	0.560	1.048	0.978	1.124	0.183	0.988	0.935	1.044	0.670	1.021	0.969	1.076	0.438
High school graduate vs. less than high school	0.626	0.128	3.065	0.235	0.765	0.130	4.495	0.507	0.499	0.122	2.036	0.911	0.224	0.048	1.047	0.595
College/graduate vs. less than high school	0.096	0.014	0.672	0.007	0.252	0.037	1.731	0.083	0.224	0.050	0.997	0.031	0.084	0.016	0.431	0.002
Unemployed vs. employed	2.614	0.772	8.850	0.123	2.149	0.611	7.563	0.233	3.129	1.217	8.046	0.018	3.033	1.255	7.330	0.014
Sippy cup use (yes vs. no)	0.372	0.093	1.482	0.161	0.595	0.133	2.666	0.498	0.693	0.251	1.912	0.479	0.593	0.225	1.562	0.290

Models were adjusted for time of awakening and time of the first cortisol collection.

LB, lactobacilli; MS, mutans streptococci.

The results of the regression model with LB as an outcome are shown in Table 3; none of the predictors was significant at P < 0.05.

For Candida outcome, lower education attainment of the parent/primary caregiver was associated with increased likelihood of carrying salivary Candida species. In addition, children from households in which the parent/primary caregiver was unemployed were over 3 times more likely to carry salivary Candida species (OR = 3.129, P = 0.018).

Table 3 considers also each of the 3 microbiological measures, combined as a single composite outcome. As anticipated from the findings from individual microbiological markers, lower education attainment of the parent/primary caregiver was associated with greater microbiological risk of any kind in caries-free children. In addition, unemployment status was associated with an OR of more than 3 for predicting the composite of any microbiological risk (OR = 3.03, P = 0.018).

A final series of exploratory analyses were carried out to test the robustness of the findings. In particular, analyses (not tabled) indicated that the prediction from education to microbiological risk (individually or the composite)—the most reliable predictor in the multivariable analyses—did not vary significantly by child sex or race (contact authors for additional information).

Discussion

In a comparatively large sample of caries-free children at elevated risk for ECC, we found that sociodemographic risk—and in particular educational status—was robustly associated with multiple markers of microbiological risk for ECC, including elevated MS and the presence of Candida. In contrast, associations with oral health behaviors were weak and inconsistent, and associations with a wealth of psychosocial risk variables were nonsignificant. We did, however, find that one physiological measure of stress, salivary cortisol, was associated with salivary MS: children with elevated salivary MS exhibited a higher initial cortisol level and a decreased response to the stressor. The findings provide valuable and novel data on the cariogenic microbiology underlying ECC.

The current study includes 2 particularly novel features. The first is that we examine predictors of individual microbiological markers (MS, LB, and Candida separately) as well as combined microbiological risk (MS, LB, and Candida jointly). Few studies have identified risks for established oral microbiological markers of ECC, and many included caries-affected children (Ollila et al. 1997; Weber-Gasparoni et al. 2012; Zhou et al. 2013; de Jesus et al. 2020). It may be that it is the combined influence of microbiological risks that is most closely associated with ECC onset; accordingly, we examine predictors of individual and composite measures of cariogenic microbiology. The second novel feature of the current study is the consideration of multiple types of predictors, with a particular focus on psychosocial stressors. The hypothesis that psychosocial stress may be causally related to ECC is notable insofar as it may help explain the social class gradient of ECC and may offer additional prevention and treatment strategies (Seow et al. 2009; Plonka et al. 2012).

It is notable that, even in this selected sample of high psychosocial risk families, we found that variation in education, a leading index of socioeconomic status, independently predicted multiple individual microbiological risks for ECC and a composite measure of risk. These findings extend prior work on the social determinants of oral health in young children (Kirthiga et al. 2019) by demonstrating that a social class gradient is apparent in microbiological risk prior to disease onset. These findings do not allow us to infer what it is about lower educational status that may increase microbiological risk for ECC, but the prediction from educational status was independent of leading explanations, including income, insurance, oral health behaviors, and psychosocial risk—all of which were also considered as covariates. Educational status is a proxy for many kinds of risk mechanisms, from diet to exposure to environmental (chemical) risk; not all of these were measured and included in our analyses. The results concerning educational status may not yet imply a causal process, but they do suggest the value of continued attention to prioritizing the oral health needs of children in these circumstances.

In response to a number of studies suggesting a potential role of psychosocial stress on ECC (Finlayson et al. 2007b; Seow et al. 2009), we designed a very detailed assessment of psychosocial risk exposures based on studies of oral health and leading approaches in developmental and clinical psychology. Specific measures assessed caregiver psychological symptoms, major life stresses affecting the family, and family violence and disorganization. Given the intensity of the measurement, it is especially notable that none of these factors had a significant, independent association with oral microbiological risk. The effects sizes (see Appendix) indicate small effects—although largely in the expected direction. These nonfindings may not be inconsistent with past studies, as the current study targeted specific microorganisms, whereas other studies examined ECC. Follow-up of this sample, when children have either converted to ECC or remained ECC free, will provide a more directly comparable analysis. However, it may be that prior studies reporting associations between psychosocial risk and ECC did not account for confounding. For example, Burgette et al. (2019) found that parent-reported social support was associated with ECC, but analyses did not adjust for education, oral health behaviors, and other factors that may have confounded the association with social support.

We did not find that psychosocial stress predicted oral microbiological risk for ECC but did find that a measure of initial salivary cortisol (measured prior to the stressor) was associated with increased likelihood to carry high levels of salivary MS; decreased cortisol reactivity was associated with high levels of salivary MS. There is limited animal and very limited human research on cortisol and ECC risk and particularly cariogenic microbiology. Apart from its association with stress exposure (Koss and Gunnar 2018), cortisol may have a distinct biological role in ECC disease progress via MS in particular. The limited studies in this area suggest that cortisol may also play a role in ECC in relation to enamel surface quality (Boyce et al. 2010; Caruso et al. 2018). Our finding that cortisol is associated with microbiological risk—independent from confounders—is novel. Further follow-up of the sample is needed to establish cortisol as part of a causal cascade leading to oral disease and to establish cortisol as having a mediating role of stress on ECC onset. The finding that lower or blunted response to the laboratory stressor was associated with MS may reflect a stress effect insofar as a blunted stress response may imply a downregulation of stress physiology, a pattern reported in chronic stress-exposed children (e.g., Koss and Gunnar 2018). A blunted stress response pattern would also be found among children with comparatively elevated prestress levels; for MS, both cortisol measures were significantly predictive. Why salivary cortisol but not specific measures of psychological stress exposure associated with microbiological risk is unclear. It may be that salivary cortisol was indexing nonmeasured sources of stress in the child or that cortisol confers cariogenic disease risk other than through stress-mediated pathways. Replications and extensions of this line of research are needed to understand if and how children’s oral biology may be a marker of stress exposure.

The lack of association between oral microbiology risk and child sex, age, and race and ethnicity may be notable given that some previous studies signal differences in ECC risk regarding these child characteristics; neither did we find that race moderated the robust association between educational status and oral microbiological risk. That contrasts with some previous evidence of increased rates of ECC in minority children (Dye et al. 2007); it may be that differences according to microbiological risk—in caries-free children—are premature and that race or other sociodemographic risks emerge once there is conversion from a clinically caries-free state to ECC.

The study has several limitations. The first is that the study targets microbiological risk rather than ECC status per se. This had sizable benefits for our study of risk processes, but it does mean that the findings may not be comparable to those targeting ECC onset as the main outcome. In addition, although the study was comparatively large, it may not have been powered sufficiently to detect very small effects, particularly concerning psychosocial factors (although the study is comparable in size, or larger, than some studies reporting associations between psychosocial stress and ECC). We have focused on MS, LB, and salivary yeast as cariogenic organisms and have not focused on microbiome analyses. Third, the study purposely targeted a high psychosocial risk group; the findings may or may not generalize to other populations. Set against these limitations were several strengths of the study, including a detailed battery of measurements, careful oral examination, adjustment for time of awakening and time of first sample salivary cortisol collection, and assessment of multiple microbiological markers of ECC.

The clinical implications of the findings suggest the need for an assessment of sociodemographic, psychosocial, and behavioral variables when evaluating risk for ECC in young children.

Author Contributions

D.T. Kopycka-Kedzierawski, T.G. O’Connor, contributed to conception, design, data acquisition, analysis, and interpretation, drafted and critically revised the manuscript; K. Scott-Anne, P.G. Ragusa, M. Cvetanovska, contributed to conception, data acquisition, and interpretation, drafted and critically revised the manuscript; K. Flint, C.L. Wong, contributed to conception, data acquisition, and interpretation, critically revised the manuscript; C. Feng, G.E. Watson, contributed to conception, design, data analysis, and interpretation, drafted and critically revised the manuscript; R.J. Billings, contributed to conception, design, data analysis, and interpretation, critically revised the manuscript; R.J. Quivey, contributed to conception, design, data acquisition, and interpretation, critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.