The Journal of EVIDENCE-BASED DENTAL PRACTICE

REVIEW

RISK PREDICTORS OF EARLY CHILDHOOD

CARIES INCREMENT—A SYSTEMATIC
REVIEW AND META-ANALYSIS

PHOEBE P.Y. LAM a , HELENE CHUA b , MANIKANDAN EKAMBARAM c , EDWARD C.M. LO d,

AND CYNTHIA K.Y. YIU e
a
Paediatric Dentistry, Faculty of Dentistry, The University of Hong Kong, Pokfulam, Hong Kong SAR Hong Kong
b
Auckland District Health Board, Auckland, New Zealand
c
Paediatric Dentistry, Faculty of Dentistry, University of Otago, Dunedin, New Zealand
d
The University of Hong Kong, Pokfulam, Hong Kong SAR, Hong Kong
e
Faculty of Dentistry, The University of Hong Kong, Pokfulam, Hong Kong SAR Hong Kong

ABSTRACT Corresponding author: Dr. Cynthia KY

This systematic review aimed to investigate the risk predictors of caries in pri- Yiu, Paediatric Dentistry and
mary teeth and evaluate their association with the increment of Early Childhood Orthodontics, 2nd Floor, Prince Philip
Caries (ECC) among preschool children. This systematic review included only co- Dental Hospital, 34-Hospital Road,
hort or case-control studies of at least 2 years duration, over 300 participants and Sai Ying Pun, Hong Kong (SAR), Hong
with English full-text. Potentially eligible studies were retrieved from 4 electronic Kong China.
databases (Ovid Embase, Ovid MEDLINE; PubMed, Web of Science) from incep- E-mail: ckyyiu@hku.hk
tion to March 1, 2021. Independent screening and data extraction by 2 reviewers
to identify factors associated with ECC increment, including family and socioeco- KEYWORDS
nomic factors, dietary and oral health-related habits, and clinical parameters. A Early childhood caries, Social
total of 18 studies from 163 potential reports were included, involving 1,159,226 inequality, Education status, Risk
preschool children. Lower parental education attainment was found associated predictor
with ECC increment (WMD:0.87; 95% CI 0.52, 1.21); whereas immigration status
(WMD:-0.38; 95% CI -1.09, 0.34), gender (WMD:-0.02; 95% CI -0.28, 0.24), and
Source of funding: Supported by the
dental service utilization (WMD:0.35; 95% CI -0.10, 0.79) were not significant fac-
Hong Kong Health and Medical
tors for ECC increment. All included studies consistently suggested positive cor-
Research Fund (HMRF) Grant
relations between ECC increment and baseline caries experience, plaque level,
(16172221).
cariogenic microorganisms, and prenatal and passive smoking, while mixed find-
ings were detected between ECC increment with dietary and oral hygiene prac- DECLARATION OF COMPETING
tices. Preschool children whose parents have low education level are more likely INTEREST: The authors declare no
to have greater increment of ECC over 2 years. Existing caries lesions, increased conflict of interest.
dental plaque level, cariogenic microorganisms, prenatal or passive smoking
were also consistently identified as risk factors for ECC in all reviewed studies.
This systematic review highlights specific risk factors to target for the prevention
of ECC and supports implementing more oral health promotion for preschool
children with parents of lower educational attainment.
Received 1 September 2021; revised
3 April 2022; accepted 25 April 2022
J Evid Base Dent Pract 2022: [101732]
INTRODUCTION 1532-3382/$36.00
Skewed distribution has long been reported across the globe, with particular © 2022 Elsevier Inc.
high-risk groups in the populations being more susceptible to dental caries.1 All rights reserved.
For instance, in the UK, 75% of the caries experience were harbored in approxi- doi: https://doi.org/10.1016/
mately 30% of the highly-deprived Caucasian children, but less than one-fifth of j.jebdp.2022.101732
those with low deprivation.2 Meanwhile, among less-privileged Asian children,

September 2022 1
 The Journal of EVIDENCE-BASED DENTAL PRACTICE

the data reporting their caries experience varied between ily and socioeconomic background, oral hygiene habits, di-
16.7%−26.9%.3 The phenomenon has aroused researchers’ etary, and child factors among children below 71 months old.
interests and attention to search for potential caries indica-
tors to locate the children at high risk. By identifying these
“high-risk” individuals and the associated risk factors, target- MATERIALS AND METHODS
specific preventive measures can be implemented to tackle This systematic review followed the guidelines proposed by
the disease effectively. MOOSE14 and Preferred Reporting Items for Systematic Re-
views and Meta-Analyses (PRISMA) statement.15 The proto-
Caries risk assessment methods have been developed and
col of the review was registered in PROSPERO (registration
proposed by professional bodies and panel of experts.4-9
number: CRD42021227656).
The implementation of caries risk assessment enhanced the
paradigm shift from treating the disease outcome to inter-
rupting the disease process in caries management.10 By un- Information sources and literature search
derstanding the specific caries etiological factors in an in- Systematic searches for relevant cohort or case-control stud-
dividual, targeted preventive and therapeutic measures can ies were conducted from 4 electronic databases (Ovid Em-
be provided to effectively achieve caries prevention and sta- base, Ovid MEDLINE; PubMed, Web of Science) using the
bilization.10 The current caries risk assessment (CRA) models broad keywords and MeSH terms to retrieve studies pub-
mostly include the evaluation of pathological factors, pro- lished from inception to Mar 2021 (Appendix 1). To ensure
tective factors, clinical findings and socio-demographic fac- no relevant study was missed, further hand search was also
tors.4-9 These factors interplay and alter the caries balance conducted by screening the reference lists of the included
between progression or reversal of caries development.10 studies and relevant past systematic reviews. Attempts were
made to contact authors of potentially-eligible manuscripts
Pathological factors contribute to the probability of initia-
for any unclear or missing information. gray literature, and
tion and progression of caries. Hyposalivation, frequent con-
unpublished data without undergoing peer-review process
sumption of fermentable carbohydrates and presence of aci-
was excluded. Studies without full text in English were also
dogenic bacteria, may tip the caries balance to demineraliza-
excluded; as published non-English studies had been shown
tion and caries progression.10 On the other hand, increased
to have little impact on the results of meta-analyses.16
protective factors such as salivary flow, antimicrobials and flu-
oride can restore the balance towards remineralization and
lead to caries arrest.10 Socio-demographic factors including Study selection
poverty, socioeconomic status, parental educational back- Two authors of this review (P.L. and H.C.) independently as-
ground may affect the ability and attitude of parents in taking sessed the articles retrieved from the databases based on
care of their children’s oral health.11 However, such associa- their titles, keywords and abstracts, followed by determina-
tion between inequality and Early Childhood Caries (ECC) tion of the potential eligibility of each article. Agreements
has not been thoroughly investigated in large-scale longitu- between reviewers were determined using Cohen’s kappa
dinal studies. coefficient (κ). Opinion from the third reviewer (M.E.) was
sought to adjudicate on the final eligibility, if required.
Although a number of caries risk assessment models have
been established from existing evidence and panels of ex-
perts, there is still insufficient evidence on whether any caries Types of studies
risk factor or combinations of factors can provide accurate Only prospective and retrospective cohort or case-control
predictive values. Moreover, the specificity of these caries studies with at least 24 months duration, investigating the as-
risk assessment models has been inadequately reported,12 sociation between potential risk factors and ECC increment
with uncertainty on the reliability of each variable to predict in terms of the increase in caries prevalence, caries experi-
future development of caries. Furthermore, initiation and ence and incidence were included. Cross-sectional studies
progression of caries in primary teeth would normally take and observational studies that evaluated participants clini-
around 6 months or more, which would not be accurately re- cally at only one time point, as well as interventional stud-
flected in cross-sectional studies or short-term observational ies were excluded. To ensure there would be sufficient sam-
studies.13 The aim of this systematic review was to evaluate ple size to minimize imprecision for dichotomous outcomes,
the association of different risk predictors of caries in primary only studies with 300 or more participants at the final review
teeth and evaluate their association with the increment of were included.17 , 18
ECC over a 24-month period among preschool children. The
objectives of this review were to investigate the association
Types of participants
between the occurrence of caries lesions or new increment
To assess different risk factors contributing to ECC,
of ECC over 24 months with different clinical conditions, fam-
preschool children between birth and 71 months of age with

2 Volume 22, Number 3

 The Journal of EVIDENCE-BASED DENTAL PRACTICE

primary dentition were included.4 Studies including partic- selection reported. After summarizing the results in all do-
ipants beyond that age range, or children of ASA (Ameri- mains, an overall rating of low, moderate, serious or critical
can Society of Anesthesiologists Physical Status Classifica- risk of bias was given to each study.
tion System) II or above (ie, with medical diseases), or chil-
dren with special health care needs were excluded. Data synthesis and analyses
Stata version 13.1 (StataCorp, College Station, TX, 2013)
Types of exposure and control/comparison was used to perform the meta-analysis. Meta-analysis with
The present review compared the development of new fewer than 5 studies was handled with fixed-effect model,
caries lesions or new increment of ECC among children with while random-effects model were adopted for analysis with
different family and socioeconomic background (such as par- more studies.20 When studies were found to display signif-
ents’ educational level, parents’ occupation, household in- icant heterogeneity or divergent results, results were pre-
come), clinical conditions (such as oral health status, oral sented narratively. For instance, when data pooling for meta-
microbiota, salivary parameters), oral health-related habits analysis was precluded because the included studies de-
(such as toothbrushing, dental service utilization), dietary fined their exposure and control groups very differently from
habits, and other factors deemed relevant to ECC. each other, especially when categorizing different socio-
behavioral and dietary factors. Also, when individual stud-
ies had not considered the potential interdependency effect
Types of outcome measures between factors or used appropriate analytical approach to
The study outcome was development of new caries lesions in
control for confounding factors, individual findings were re-
the children during the follow-up period, in terms of caries in-
ported narratively only. Sensitivity analyses were conducted
cidence or increase in the number of decayed teeth or tooth
as necessary to determine if the effect estimate is dependent
surfaces. Studies recording dental caries by the use of the
of any studies with poor validity.
World Health Organization (WHO) recommended decayed,
missing and filled teeth or surfaces (dmft or dmfs) index; or Assessment of heterogeneity
the International Caries Detection and Assessment System In accordance with the MOOSE Guidelines,14 heterogeneity
(ICDAS) were included. of outcome results was assessed. I2 statistics and Chi square
test were used to calculate the amount of heterogeneity and
Data collection and measurement of treatment level of significance (P < .05). Based on the value of the I2
effect statistics calculated, the outcomes were rated as substantial,
Data were independently extracted by the first and second if I2 > 50% and P < .05.
authors (P.L. and H.C.) using a standardized data extraction
form. The data include study characteristics (study design, Assessment of publication bias
beginning year and duration of follow-up), participants (loca- Publication bias was assessed in accordance to the recom-
tion, inclusion and exclusion criteria, gender, age, and base- mendations from the MOOSE Guidelines.14 Funnel plots
line caries), exposure and control (parental and maternal ed- were used for the assessment if there are more than 10 stud-
ucation background, household income). These were input ies contributed to the outcome.20
into Microsoft Excel (Microsoft, USA, 2010) for initial analy-
sis. Regarding outcome, both absolute effects and adjusted Assessment of quality of evidence
effect estimates were obtained during data extraction, but The quality of evidence for each outcome was evaluated
only adjusted estimates were used in the quantitative syn- independently by adopting the Grading of Recommenda-
thesis. tions Assessment Development and Evaluation (GRADE) ap-
proach.17 As only observational studies were included, the
Risk of bias in individual studies quality of evidence was initially assessed as moderate to low.
Risk of bias of each included study was determined based The body of evidence were further downgraded if it pre-
on the risk of bias in nonrandomised studies tool (ROBINS-I sented with serious issues related to risk of bias, impreci-
tool).19 The tool was designed for use with interventional and sion, inconsistency, indirectness and publication bias, or up-
observational studies for comparison to an ideal randomised graded due to large magnitude of effect, a dose response
controlled trial. Each study was evaluated in 7 domains of or no plausible confounding.
bias guided with respective signaling questions to obtain an
overall final judgment. The 7 domains are: (I) bias due to con- RESULTS
founding, (II) bias in selection of participants into the study,
(III) bias in classification of interventions, (IV) bias due to de- Selection process
viation from intended interventions, (V) bias due to missing A systematic literature search identified 1072 records af-
data, (VI) bias in measurement of outcomes and (VII) bias in ter removal of duplicates. After screening the titles and

September 2022 3
 The Journal of EVIDENCE-BASED DENTAL PRACTICE

Fig. 1. PRISMA flowchart of the systematic review and meta-analysis.

abstracts, 163 records were selected for full-text read- ten23 , 25 , 27 , 29 , 30 , 33-37 were performed between 2001 and 2010,
ing, with 145 were excluded with reasons detailed in Ap- and 224 , 26 studies between 2011 and 2020.
pendix 2. Eighteen reports fulfilled the inclusion criteria
(kappa = 0.860),21-35 but only 3 studies with data that could Risk of bias of included studies
be included in meta-analyses,23 , 34 , 36 due to heterogeneity The risk of bias of all included studies was assessed with
between studies in result reporting. The PRISMA flow dia- ROBINS-I tool (Fig. 2).19 All studies were rated as low risk of
gram was described in Fig. 1. bias in the following domains: domain (III) bias in classifica-
tion of exposures, domain (V) bias due to missing data, and
domain (VI) bias in measurement of outcomes.21-35
Data extraction
The general characteristics of 16 prospective and 2 ret- Most studies were rated as of low risk of bias in domain (IV)
rospective studies from 18 reports were described in bias due to deviation from intended exposures. Ten stud-
Table 1 .21-38 This review included 1198,430 preschool chil- ies re-evaluated their variables with questionnaires and clin-
dren from 11 different countries across Asia (n = 3),33 , 36 , 37 ical parameters at subsequent follow-ups,23 , 24 , 29-31 , 33-37 and
North America (n = 3),26 , 27 , 30 South America (n = 2),23 , 34 4 studies only evaluated variables that would not deviate
Europe (n = 9)21 , 22 , 25 , 28 , 29 , 31 , 32 , 35 , 38 and Australia (n = 1)24 through time.21 , 28 , 38 , 39 Three studies were considered as se-
with 1159,226 children followed up for 24 to 84 months. rious risk as they only measured sociodemographic and oral
Six studies commenced before 2001,21 , 22 , 28 , 31 , 32 , 38 while health-related variables at baseline.22 , 32 , 40

4 Volume 22, Number 3

 Table 1. Characteristics of included studies.

First author (year, Commencement Inclusion (I)/Exclusion (E) criteria Commencement Baseline/ Significant factors
countrya )P/R year/Duration age/age at Completed
follow-ups

Akinkugbe (2021,GBR) P 1992/30m (I) Pregnant women in 3 health districts 31 m/43 m,61m 13,449/1429 dmft Prenatal smoking
in Bristol England and expected to Partner smoking
deliver Apr 1991-Dec 1992 (I) Parents
completed questionnaires

Bernabe (2017,GBR) P 1993/48m (I) Children born in Dundee Scotland 0–12 m/24 m,36 1419/1102 dmfs Low birth weight
between Apr 1993-Mar 1994 m,48m Smoking and less educated mother
Families with neither parent in
employment
Area deprivation
Toothbrushing frequency

de Melo (2019,BRA) P 2006/48m (I) Children from 2006 cross-sectional 18–36 m/60–84m 1045/469 dmft Maternal education
study Maternal participation in labor
market
Consumption of sweets
Attendance at public school
Private clinical services

Gussy (2020,AUS) P 2020/72m (I) Families of disadvantaged and 0–1 m/6 m,12 m,18 467/419 dmft Maternal smoking
culturally diversed m,36 m,48 m,72m Night bottle feeding
(I) Families with newborn children in 6 Salivary Streptococcus mutans

The Journal of EVIDENCE-BASED DENTAL PRACTICE

local government areas

Hultquist (2016,SWE) P 2002/24m (I) Children living in the catchment of 12 m/36 m 1042/779 dmft Beverages other than water
Vidablick Public Dental clinic Caries in sibling
(I) Parents responded to invitation and Higher level of MS
questionnaire via mail Night meal

Ismail (2015,USA) P 2015/24m (I) Low-income African-American 0–60 m/24–84 m 1021/645 Baseline caries
children
(I) Children with signed consent

Jean (2018,USA) R 2009/72m (I) Data available on PEDSnet 0–12 m/25–36 Oral thrush
(I) Children seen as infants between 1 m,37–48 m,49–60 1012,668/1012,668
and 12 m at 6 hospital sites m,61–71 m
September 2022

(continued on next page)

5
6
Volume 22, Number 3

The Journal of EVIDENCE-BASED DENTAL PRACTICE

Table 1 (continued)
First author (year, Commencement Inclusion (I)/Exclusion (E) criteria Commencement Baseline/ Significant factors
countrya )P/R year/Duration age/age at Completed
follow-ups

Julihn (2020,SWE) P 2000/48 m (I) Children born in 2000–2003 and 36 m/84 m 83,147/65,259 deft Vaginally-delivered
resided in Stockholm Gender
(I) Children received dental checkups at Parity
Public Dental Service, with private Country of birth
practitioners, or Karolinska Institutet Age at delivery
Smoking in early pregnancy
BMI in early pregnancy
Educational level
Income

Leroy (2012,BEL) P 2003/24 m (I) Children born in Tielt-Winge and 36 m/60 m 1057/772 Caries experience at age 3
Berlaar, Flanders Plaque accumulation at age 5
(I) Children served as controls in oral
health promotion program

Lim (2015, USA) P 2002/48m (I) Households from the 39 census tracts 0–60 m/48 m 1021/654 Soda intake
with the highest proportion of residents Caregiver’s age
living below 200% of poverty in Detroit Baseline caries experience

Mattila (1998,FIN) P 1985/42m (I) Pregnant women who made their first 18 m/60 m 1292/1003 Snacking frequency of sweets
visit to maternity health care clinic in Dental caries at 3 years
1986 Dental plaque at 3 years
(I) Gave informed consent Mother’s basic education

Meurman (2010, FIN) P 1998/42m (I) Children (n = 545) born between 18 m/42 m 545/347 MS in oral, occupation of caretaker
January (white vs blue collar), nighttime
1, 1998 and June 30, 1999 in one of 4 feeding, sugar use, drinks other than
health care areas water, oral health of father
(I) Received screening test for MS and
had caretakers interviewed

Nishimura (2008,JPN) P 2001/42m (I) Children underwent health 18 m/24 m,42 m 1208/1208 Caries activity test
examinations at 18 m,24 m,30 m at Increased frequency and duration of
Kurashiki-City Public Health Center total sucrose intake
(I) Children receive the caries activity test Failure of parental brushing
at 18 months of age

Reyes (2020,BRA) P 2010/24m (I) Children recruited from primary care 0–60 m/24–84 m 639/467 Dental caries at baseline
units in city’a 8 administrative regions
(I) Children allowed examination and
answered structure questionnaire
(continued on next page)
 Table 1 (continued)
First author (year, Commencement Inclusion (I)/Exclusion (E) criteria Commencement Baseline/ Significant factors
countrya )P/R year/Duration age/age at Completed
follow-ups

Skeie (2008 and 2010 2002/24m (I) Children took part in previous 36 m/60 m 354/304 Parent attitude to diet
NOR) P cross-sectional study in 2002 at 36m Parental indulgence
(I) Parent responded to questionnaires
(I) Children belonged to one of the 7
clinic in Oslo

Tanaka (2015,JPN) R 2004/32m (I) Mother and children resided in Kobe 0 m/4 m,9 m,18 76,920/70,711 Second-hand smoke
from pregnancy to 36 m m,36 m
(I) Children participated in the health
check-up program
(I) Children born between 2004 and 2010
(I) Children with available info on
smoking at age 4m
(I) Dental records available at 18 m and
36m

Wendt (1999, SWE) P 1987/36m (I) Pre-schoolchildren living 12 m/23–26 671/632 Clinical and radiographic findings

The Journal of EVIDENCE-BASED DENTAL PRACTICE

in the community of Jonkoping, Sweden m,34–39 m
(I) Followed from the age of I to 3 yr.

Wong (2012,HKG) P 2007/24m (I) Children enrolled in 67 kindergartens 36–48 m/60–72 m 465/358 Nursing bottle
with more than 60 children Toothbrushing after 12m
(E) Children refused to attend follow-up, Snacking once or more daily
involved in other research, did not Parental education below 9 years
attend baseline survey

a
ISO alpha-3 codes of CountriesP/R: Prospective study/Respective study.
September 2022
7
8
Volume 22, Number 3

The Journal of EVIDENCE-BASED DENTAL PRACTICE

Table 2. GRADE summary of findings.

Comparison Results N patient N studies Risk of biasa Inconsistencyb Indirectnessc Imprecisiond Publication Quality of
age range biase evidence
(m) (GRADE)

I2 (%) Heterogenicity
χ 2 test
(P value)

Parental Children with parents 2055 (0–60) 3 Not serious 0.0% .834 Not serious Not serious N/A O
education having eight years or moderate
less education have
increased risk of ECC
increment

– – – – –

Immigration No significant 827 (0–60) 3 Not serious 0.0% .654 Not serious Not serious N/A O
difference moderate

– – – – –

Gender No significant 1294 (0–60) 3 Not serious 72.2% .027 Not serious Not serious N/A OO low
difference

– ↓ – – –

Dental No significant 988 (0–60) 2 Not serious 0.0% .368 Not serious Not serious N/A O
service difference moderate
utilization

– – – – –

a
Risk of bias: Considered as serious if overall half of the studies included were of serious risk of overall bias.
b
Inconsistency: Considered as serious when I2 statistics ≥70% (∗ ) and p-value of χ 2 test <0.05 (∗ ∗ ).
c
Indirectness: Considered as serious when applicability of findings were restricted in terms of population, intervention, comparator and outcomes.
d
Imprecision: Considered as serious when total number of events was below 300 for dichotomous outcomes or 400 for continuous outcomes (# ), or when the upper and lower limits of 95% CI include both
meaningful benefits and harm.
e
Publications bias: Considered as serious if p-value of Begg’s funnel plot <0.05. Not applicable (N/A) if funnel plot could not be constricted given limited numbers of study. Publication bias was difficult
to detect and thus no downgrading was performed↓: Downgrade by one level in quality of evidence; ↓↓: Downgrade by 2 levels in quality of evidence–: No change in quality of evidence.
 The Journal of EVIDENCE-BASED DENTAL PRACTICE

Fig. 2. Risk of bias assessment with ROBINS-I.

Five studies were assessed as having serious risk of bias in risk of bias; hence, were assessed as overall serious risk of
domain (I) bias due to confounding, as the authors either bias.21 , 26 , 27 , 35 , 37 , 38
just compared the caries experience between exposure and
control group of interests, without evaluating other potential
confounding factors such as socioeconomic status, tooth- Potential factors associated with caries increment
brushing and dietary habits.21 , 27 , 35 , 37 , 38 Whereas 4 studies 1. Family and socioeconomic factors
were considered as of moderate risk of bias in this domain as Seven out of 9 studies identified low parental education as a
these confounders were not well-defined or classified.25 , 32-34 significant risk predictor for caries increment in primary den-
tition.22-24 , 28 , 29 , 31 , 36 Meta-analyses from 3 studies suggested
Almost all studies were considered as of low risk of bias in a significant mean dmft increment of 0.87 (weighted mean
domain (II) Bias in selection of participants into the study, difference (WMD); 95% CI, 0.52, 1.21) among children whose
except 2 studies.21 , 38 Akinkugbe et al. (2021)21 reported that parental education was 9 years or less (Fig. 3).23 , 34 , 36
the mother participants selected into the study were of more
advanced age compared with the mothers who were not in- Four studies investigated prenatal smoking as a risk factor
cluded from the sampling frame. Tanaka et al. (2015)37 only for caries increment.21 , 22 , 28 , 30 Three out of 4 studies sug-
included dental records in which maternal smoking informa- gested a significant association between prenatal smoking
tion were available, which may have systemically excluded and caries occurrence.21 , 22 , 28 A higher proportion of children
mothers who did not smoke. of prenatal smokers developed caries or had a higher dmft
score over time as compared to those of nonsmokers.21 , 22 , 28
Four studies were rated as serious risk of bias in domain (VII) Three included studies suggested that second-hand smok-
selection of the reported results, because they did not in- ing from family members increased the likelihood of caries
clude or report other potential confounding factors in their increment.21 , 30 , 37 Leroy et al. (2012) did not identify any sig-
regression analyses in caries increment.21 , 27 , 35 , 37 nificant difference in caries increment due to parental smok-
Seven studies were rated as of low risk of overall bias ing from 3 to 5 years old, but it is important to note that the
because all 7 domains were evaluated as of low risk of participants had significantly higher prevalence of caries at
bias.23 , 24 , 28-31 , 36 Five studies were considered as of moder- baseline.29
ate risk of overall bias as they have no more than 2 do- A few studies also reported that single-parent22 , 28 , 30 and
mains being considered as of moderate risk of bias.22 , 25 , 32-34 advanced age at delivery24 could play a significant role
Six studies had more than one domain that were of serious in caries increments. Other parental factors including em-

September 2022 9
 The Journal of EVIDENCE-BASED DENTAL PRACTICE

Fig. 3. Meta-analysis: Parental education attainment.

ployment, parental BMI, and attitudes had inconsistent tween studies, only qualitative but not quantitative analyses
results.22 , 28 could be performed. Lactobacilli and Candida levels were
also found to be significant, but were only investigated in
Among all studies that investigated the familial status of
one study.26
child in relation to caries increment over 2 years, the fol-
lowing factors were investigated but were only found signif-
icant in some of the included studies, including siblings hav- 3. Factors related to oral hygiene habits
ing caries (100%, n = 1/1 study),25 immigration status (40%, Only Wong et al. (2012)36 assessed the correlation between
n = 2/5 studies),28 , 41 low socioeconomic status (20%, n = 1/5 onset of toothbrushing and caries increment, and children
studies),22 born second or later (66.7%, n = 2/3 studies),22 , 28 who started brushing before 12 months of age had signifi-
and requiring social welfare (50%, n = 1/2 studies).28 , 35 Only cantly less caries than those who started after.
data regarding immigration status can be pooled for meta- Nishimura et al. (2008)33 found that supervised toothbrush-
analysis, the findings suggested preschool children who ing for 18-month-old children significantly reduced caries in-
were recent immigrants were not predisposed to increased crement over 2 years. Leroy et al. (2012)42 also identified that
caries development in their primary dentition compared to children were less likely to have caries at age 5 when parents
their peers (WMD: −0.38; 95% CI, −1.09, 0.34) (Fig. 4).23 , 36 helped with their toothbrushing at age 3.

Bernabe et al. (2017)22 and Wong et al. (2012)36 found that

2. Clinical conditions
children who brushed twice or more per day were less likely
The clinical parameters at baseline that were associated with
to have caries increment, but the frequency of toothbrushing
caries increment in the reviewed studies included baseline
was not significant in their final regression models.
caries experience,24 , 26 , 30 , 31 , 35 plaque level29 , 31 , 33 , 34 and mu-
tans streptococcus level.24 , 25 , 33 Although all included studies Wong et al. (2012)36 and Reyes et al. (2020)34 did not iden-
consistently reported findings that significantly higher caries tify any difference between children who had regular den-
experience (n = 5),24 , 26 , 30 , 31 , 35 plaque level (n = 4),29 , 31 , 33 , 34 tal visits and those who had not (WMD: 0.35, 95% CI −0.10,
and mutans streptococcus level (n = 4)24 , 25 , 33 at baseline sig- 0.79) (Fig. 5). de Melo et al. (2019)23 did not compare the
nificantly increased the possibility of ECC increment in the caries increment between regular and irregular attendees,
follow-ups, it was not applicable to synthesize the results as but found those who utilized private dental services had sig-
the raw data could not be pooled for analysis. Moreover, as nificantly lower dmft score increment than those who used
the comparison groups and outcome measures varied be- public dental service or had no dental services at all.

10 Volume 22, Number 3

 The Journal of EVIDENCE-BASED DENTAL PRACTICE

Fig. 4. Meta-analysis: Immigration status.

Regarding the use of fluoride, Gussy et al. (2020)24 reported (WMD: −0.02, 95% CI −0.028, 0.24)23 , 34 , 36 (Fig. 6). Other in-
that water fluoridation at baseline could reduce caries de- cluded studies which have performed regression analyses
velopment. Whereas Wong et al. (2012)36 also investigated based on the results obtained from their recruited subjects
the use of low fluoride toothpaste but could not determine also found no significance between boys and girls with re-
a relationship as one-third of the respondents did not know spective to caries increment over 24 months.22 , 25 , 28 , 29 , 32
the fluoride content of their toothpastes.
Only one study, with participants aged between 0 and 60
months, found that older children are more likely to have
4. Dietary factors more caries increment,34 while such difference were not
Inconsistent findings were identified among included stud- found among other studies.23 , 29 , 30
ies regarding different sugary dietary habits. There was sig-
nificant caries increment with night bottle feeding (57.1%,
4/7 studies),24 , 25 , 33 , 36 frequent snacking (42.8%, n = 3/7 stud- GRADE assessment
ies),23 , 33 , 36 sweet beverages (40%, n = 2/5 studies)25 , 30 and The body of evidence was rated moderate for parental edu-
breastfeeding (33.3%, 1/3 studies).33 As the findings of each cation attainment, dental service utilization and immigration
study displayed significant differences, their results were status as we found no serious concerns in relations to risk of
only presented narratively without any quantitative synthe- bias of the included studies, inconsistency, imprecision, and
sis. indirectness.

However, as the outcomes for gender had considerable het-

5. Child factor erogeneity (I2 = 72.2%, P = .027), the quality of evidence
Gender was found not significant regarding caries increment were rated low despite no serious concerns in the other do-
over 24 months, as shown in the results of the meta-analyses mains.

September 2022 11
 The Journal of EVIDENCE-BASED DENTAL PRACTICE

Fig. 5. Meta-analysis: Dental service utilization.

DISCUSSION dental care, professionally-applied caries-preventive mea-

This review looked into a number of factors which were sures,45 materials for oral hygiene maintenance and healthier
hypothesised to increase the risk of ECC increment. How- diet.11
ever, due to heterogeneity of exposure definitions, outcome There is established evidence that shows poverty and lower
measurements and data presentation, only parental educa- socioeconomic status are associated with increased caries
tion attainment, immigration status, gender and dental ser- experience in general.46 However, we could not establish
vice utilization were pooled for meta-analysis. Other familial sufficient evidence to affirm such findings in ECC due to in-
and socioeconomic factors, clinical parameters, oral health- consistency in how the included studies defined, assessed
related habits, dietary habits and other factors were only re- and categorized different socioeconomic factors, precluding
ported narratively. pooling of data for meta-analyses.
Our findings affirm that clinical factors are reliable in predict- Prenatal smoking was identified as a predisposing factor for
ing caries increment in ECC. Increased baseline caries expe- caries increment among preschool children, which is con-
rience, plaque level, mutans streptococcus, and lactobacilli sistent with previous surveys and cross-sectional studies.47
counts are positively associated with increased caries risk. Smoking adversely alters salivary flow and pH and might
High quality evidence suggests parental education level to significantly increase the caries susceptibility of children ex-
be a significant factor for ECC increment. In agreement posed to prenatal or passive smoking.48 , 49 Smoking also in-
with past studies, lower parental educational attainment has creases the likelihood of low birth weight50 and premature
been shown to be significantly associated with greater risk tooth emergence,51 which increases caries susceptibility as
of having any caries experience.43 , 44 One possible reason a result of developmental enamel defects52 and frequent in-
is that education background often determines household fant feeding. However, many included studies that evaluated
income, affecting one’s affordability and access to routine smoking did not assess the effect of confounding factors.21 ,

12 Volume 22, Number 3

 The Journal of EVIDENCE-BASED DENTAL PRACTICE

Fig. 6. Meta-analysis: Gender.

37 ROBINS-I, and evidence assessment with GRADE approach.

Since the prevalence of smoking is higher among disad-
vantaged groups in poverty53 and individuals with low edu- The inclusion of studies with over 24 months follow-up and
cation attainment,54 these factors should be considered be- at least 300 participants minimize bias due to imprecision
fore concrete conclusions could be drawn. and other confounding factors. Major limitations include the
exclusion of gray literature and non-English full texts, even
The association between ECC increment and dietary habits
though studies have shown they only represented a small
was inconsistent among the included studies due to varia-
proportion of the included studies in medical systematic re-
tions in evaluation methods and reporting. In the past, a link
views and seldom affect the effect estimates.16 Other limi-
between dental caries experience and frequent consump-
tations include the inability to conduct subgroup analyses,
tion of fermentable carbohydrates, night or on-demand
sensitivity analyses and publication bias analyses due to lim-
breast or bottle feeding55 was well established. However,
ited number of included studies.
with other modulating factors in the modern days, for ex-
ample fluoride exposure, this relationship was found to be
weaker than expected.56 There was only one included study
CONCLUSIONS
that evaluated the protective effect of water fluoridation.24
More longitudinal studies are required to evaluate how these This systematic review identified lower parental education
pathological and protective factors interplay in the current attainment as a risk factor for development of ECC over
settings. 24 months among preschool children. Preschool children
whose parents received below 9 years of education would
The strengths of this review include close adoption of have greater increment of ECC. Although data cannot be
the recommendations from the MOOSE guidelines14 and pooled for meta-analysis, all included studies consistently
PRISMA guidelines to minimize potential bias, independent identified children with existing caries, plaque accumulation,
screening and data extraction, risk of bias assessment with increased oral mutans streptococcus count were more likely

September 2022 13
 The Journal of EVIDENCE-BASED DENTAL PRACTICE

to develop caries over 24 months, affirming these clinical pa- 10. Young DA, Featherstone JD. Implementing caries risk
rameters as risk predictors for ECC. While prenatal smoking assessment and clinical interventions. Dental Clinics.
might be a potential risk factor for caries increment, more 2010;54(3):495–505.
high quality studies are needed to validate this finding. 11. Drewnowski A, Specter SE. Poverty and obesity: the role of en-
ergy density and energy costs. Am J Clin Nutr. 2004;79(1):6–16.
ACKNOWLEDGMENT doi:10.1093/ajcn/79.1.6.

We are immensely grateful to Ms. Liu Louise, the dental li- 12. Gao X, Di Wu I, Lo EC, Chu CH, Hsu CY, Wong MC. Validity
brarian of the University of Hong Kong, for her assistance in of caries risk assessment programmes in preschool children. J
the study selection and screening process. We would also Dent. 2013;41(9):787–795. doi:10.1016/j.jdent.2013.06.005.
like to express our gratitude to Ms. Li Kar Yan Samantha, 13. Tickotsky N, Petel R, Araki R, Moskovitz M. Caries progression
the statistician of the Faculty of Dentistry, The University of rate in primary teeth: a retrospective study. J Clin Pediatr Dent.
Hong Kong; for her assistance in data syntheses and meta- 2017;41(5):358–361. doi:10.17796/1053- 4628- 41.5.358.
analyses.
14. Brooke BS, Schwartz TA, Pawlik TM. MOOSE reporting guide-
lines for meta-analyses of observational studies. JAMA Surg.
SUPPLEMENTARY MATERIALS 2021;156(8):787–788. doi:10.1001/jamasurg.2021.0522.
Supplementary material associated with this article can be 15. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred report-
found, in the online version, at doi:10.1016/j.jebdp.2022. ing items for systematic reviews and meta-analyses: the PRISMA
101732. statement. Ann Intern Med. 2009;151(4):264–269.

16. Hartling L, Featherstone R, Nuspl M, Shave K, Dryden DM,

REFERENCES Vandermeer B. Grey literature in systematic reviews: a cross-
1. Antunes JL, Narvai PC, Nugent ZJ. Measuring inequalities in the sectional study of the contribution of non-English reports, un-
distribution of dental caries. Community Dent Oral Epidemiol. published studies and dissertations to the results of meta-
2004;32(1):41–48. doi:10.1111/j.1600-0528.2004.00125.x. analyses in child-relevant reviews. BMC Med Res Methodol.
2017;17(1):64. doi:10.1186/s12874- 017- 0347- z.
2. Shackleton N, Broadbent JM, Thornley S, Milne BJ, Crengle S,
Exeter DJ. Inequalities in dental caries experience among 4- 17. Schünemann HJ, Cuello C, Akl EA, et al. GRADE guidelines:
year-old New Zealand children. Community Dent Oral Epi- 18. How ROBINS-I and other tools to assess risk of bias in
demiol. 2018;46(3):288–296. doi:10.1111/cdoe.12364. nonrandomized studies should be used to rate the certainty
of a body of evidence. J Clin Epidemiol. 2019;111:105–114.
3. Dugmore CR, Rock WP. The effect of socio-economic status doi:10.1016/j.jclinepi.2018.01.012.
and ethnicity on the comparative oral health of Asian and
White Caucasian 12-year-old children. Community Dent Health. 18. Daniel JN. Sampling Essentials: Practical Guidelines For Mak-
2005;22(3):162–169. ing Sampling Choices. Los Angeles: SAGE Publications; 2011.

4. American Academy of Pediatric DentistryGuideline on 19. Schünemann HJ, Cuello C, Akl EA, et al. GRADE guidelines:
caries-risk assessment and management for infants, children, 18. How ROBINS-I and other tools to assess risk of bias in
and adolescents. Pediatr Dent. 2013;35(5):E157. nonrandomized studies should be used to rate the certainty
of a body of evidence. J Clin Epidemiol. 2019;111:105–114.
5. Halasa-Rappel YA, Ng MW, Gaumer G, Banks DA. How use- doi:10.1016/j.jclinepi.2018.01.012.
ful are current caries risk assessment tools in informing the
oral health care decision-making process? J Am Dent Assoc. 20. Guyatt G, Rennie D, Meade M, Cook D. Users’ Guides to the
2019;150(2):91–102 e2. doi:10.1016/j.adaj.2018.11.011. Medical Literature: A Manual For Evidence-Based Clinical Prac-
tice. 3rd ed. Chicago: AMA Press; 2015 vol 706.
6. Lee H. Translating Current Evidence Into Practice in Paediatric
Dentistry With Clinical Practice Guidelines. HKU Theses Online 21. Akinkugbe A.A. Does the Trimester of Smoking Matter in the
(HKUTO). 2015; Association between Prenatal Smoking and the Risk of Early
Childhood Caries?Caries Res.. doi:10.1159/000513257
7. Gao XL, Hsu CY, Xu Y, Hwarng HB, Loh T, Koh D. Building caries
risk assessment models for children. J Dent Res. 2010;89(6):637– 22. Bernabé E, MacRitchie H, Longbottom C, Pitts NB, Sab-
643. doi:10.1177/0022034510364489. bah W. Birth weight, breastfeeding, maternal smoking and
caries trajectories. J Dent Res. 2017;96(2):171–178. doi:10.1177/
8. Ramos-Gomez F, Crystal YO, Ng MW, Tinanoff N, Feather- 0022034516678181.
stone JD. Caries risk assessment, prevention, and management
in pediatric dental care. Gen Dent. 2010;58(6):505–517. 23. de Melo M, de Souza WV, de Goes PSA. Increase in dental
caries and change in the socioeconomic profile of families in a
9. Bratthall D, Hansel Petersson G. Cariogram–a multifactorial child cohort of the primary health care in Northeast Brazil. BMC
risk assessment model for a multifactorial disease. Commu- Oral Health. 2019;19(1):183. doi:10.1186/s12903- 019- 0871- 9.
nity Dent Oral Epidemiol. 2005;33(4):256–264. doi:10.1111/j.
1600-0528.2005.00233.x.

14 Volume 22, Number 3

 The Journal of EVIDENCE-BASED DENTAL PRACTICE

24. Gussy M, Mnatzaganian G, Dashper S, et al. Identifying pre- retrospective cohort study. BMJ. 2015;351:h5397. doi:10.1136/
dictors of early childhood caries among Australian children us- bmj.h5397.
ing sequential modelling: findings from the VicGen birth cohort
38. Wendt LK, Hallonsten AL, Koch G. Oral health in pre-school
study. J Dent. 2020;93:103276. doi:10.1016/j.jdent.2020.103276.
children living in Sweden. Part III–A longitudinal study. Risk anal-
25. Hultquist AI, Bågesund M. Dentin caries risk indicators in 1- yses based on caries prevalence at 3 years of age and immigrant
year-olds. A two year follow-up study. Acta Odontol Scand. status. Swed Dent J. 1999;23(1):17–25.
2016;74(8):613–619. doi:10.1080/00016357.2016.1227085.
39. Brandquist E, Dahllof G, Hjern A, Julihn A. Caesarean section
26. Ismail AI, Lim S, Tellez M. Tooth surface level caries progression does not increase the risk of caries in Swedish children. JDR Clin
in the primary dentition among preschool children. Caries Res. Transl Res. 2017;2(4):386–396.
2015;49(4):442–448. doi:10.1159/000434729.
40. Ismail AF, McGrath CP, Yiu CKY. Oral health of children with
27. Jean J, Goldberg S, Khare R, et al. Retrospective analysis type 1 diabetes mellitus: a systematic review. Diabetes Res Clin
of Candida-related conditions in infancy and early childhood Pract. 2015;108(3):369–381.
caries. Pediatr Dent. 2018;40(2):131–135.
41. Skeie MS, Espelid I, Riordan PJ, Klock KS. Caries increment
28. Julihn A, Soares FC, Hammarfjord U, Hjern A, Dahllöf G. Birth in children aged 3-5 years in relation to parents’ dental
order is associated with caries development in young children: a attitudes: oslo, Norway 2002 to 2004. Community Dent Oral
register-based cohort study. BMC Public Health. 2020;20(1):218. Epidemiol. 2008;36(5):441–450. doi:10.1111/j.1600-0528.2008.
doi:10.1186/s12889- 020- 8234- 7. 00430.x.

29. Leroy R, Bogaerts K, Martens L, Declerck D. Risk fac- 42. Leroy R, Bogaerts K, Martens L, Declerck D. Risk factors for
tors for caries incidence in a cohort of Flemish preschool caries incidence in a cohort of Flemish preschool children. Clin
children. Clin Oral Investig. 2012;16(3):805–812. doi:10.1007/ Oral Investig. 2012;16(3):805–812.
s00784- 011- 0579- y.
43. Ju X, Jamieson LM, Mejia GC. Estimating the effects of ma-
30. Lim S, Tellez M, Ismail AI. Dental caries development among ternal education on child dental caries using marginal struc-
African American children: results from a 4-year longitudinal tural models: the Longitudinal Study of Indigenous Australian
study. Community Dent Oral Epidemiol. 2015;43(3):200–207. Children. Community Dent Oral Epidemiol. 2016;44(6):602–610.
doi:10.1111/cdoe.12140. doi:10.1111/cdoe.12259.

31. Mattila ML, Paunio P, Rautava P, Ojanlatva A, Sillanpää M. 44. van der Tas JT, Kragt L, Elfrink MEC, et al. Social inequalities
Changes in dental health and dental health habits from 3 to 5 and dental caries in six-year-old children from the Netherlands.
years of age. J Public Health Dent. 1998;58(4):270–274. doi:10. J Dent. 2017;62:18–24. doi:10.1016/j.jdent.2017.04.008.
1111/j.1752-7325.1998.tb03008.x.
45. Marinho VC, Higgins JP, Logan S, Sheiham A. Topical fluo-
32. Meurman PK, Pienihäkkinen K. Factors associated with caries ride (toothpastes, mouthrinses, gels or varnishes) for preventing
increment: a longitudinal study from 18 months to 5 years of dental caries in children and adolescents. Cochrane Database
age. Caries Res. 2010;44(6):519–524. doi:10.1159/000320717. Syst Rev. 2003(4):Cd002782. doi:10.1002/14651858.Cd002782.

33. Nishimura M, Oda T, Kariya N, Matsumura S, Shimono T. Using 46. Schwendicke F, Dörfer C, Schlattmann P, Page LF, Thomson W,
a caries activity test to predict caries risk in early childhood. J Paris S. Socioeconomic inequality and caries: a systematic re-
Am Dent Assoc. 2008;139(1):63–71. doi:10.14219/jada.archive. view and meta-analysis. J Dent Res. 2015;94(1):10–18.
2008.0022.
47. Kellesarian SV, Malignaggi VR, de Freitas P, Ahmed HB, Javed F.
34. Reyes LT, Knorst JK, Ortiz FR, Mendes FM, Ardenghi TM. Path- Association between prenatal maternal cigarette smoking and
ways influencing dental caries increment among children: a co- early childhood caries. A systematic review. J Clin Exp Dent.
hort study. Int J Paediatr Dent. 2021;31(3):422–432. doi:10.1111/ 2017;9(9):e1141–e1146. doi:10.4317/jced.54064.
ipd.12730.
48. Fujinami Y, Fukui T, Nakano K, et al. The effects of cigarette
35. Skeie MS, Klock KS, Haugejorden O, Riordan PJ, Espelid I. exposure on rat salivary proteins and salivary glands. Oral Dis.
Tracking of parents’ attitudes to their children’s oral health- 2009;15(7):466–471. doi:10.1111/j.1601-0825.2009.01572.x.
related behavior-Oslo, Norway, 2002-04. Acta Odontol Scand.
49. Avşar A, Darka O, Topaloğlu B, Bek Y. Association of pas-
2010;68(1):49–56. doi:10.3109/00016350903364926.
sive smoking with caries and related salivary biomarkers in
36. Wong MC, Lu HX, Lo EC. Caries increment over 2 years in young children. Arch Oral Biol. 2008;53(10):969–974. doi:10.
preschool children: a life course approach. Int J Paediatr Dent. 1016/j.archoralbio.2008.05.007.
2012;22(2):77–84. doi:10.1111/j.1365-263X.2011.01159.x.
50. Nigg JT, Breslau N. Prenatal smoking exposure, low birth
37. Tanaka S, Shinzawa M, Tokumasu H, Seto K, Tanaka S, weight, and disruptive behavior disorders. J Am Acad
Kawakami K. Secondhand smoke and incidence of dental caries Child Adolesc Psychiatry. 2007;46(3):362–369. doi:10.1097/01.
in deciduous teeth among children in Japan: population based chi.0000246054.76167.44.

September 2022 15
 The Journal of EVIDENCE-BASED DENTAL PRACTICE

51. Ntani G, Day PF, Baird J, et al. Maternal and early life factors of 54. Wang Q, Shen JJ, Sotero M, Li CA, Hou Z. Income, occu-
tooth emergence patterns and number of teeth at 1 and 2 years pation and education: are they related to smoking behaviors
of age. J Dev Orig Health Dis. 2015;6(4):299–307. doi:10.1017/ in China? PLoS ONE. 2018;13(2):e0192571. doi:10.1371/journal.
s2040174415001130. pone.0192571.

52. Salanitri S, Seow WK. Developmental enamel defects in the pri- 55. Reisine S, Douglass JM. Psychosocial and behavioral issues
mary dentition: aetiology and clinical management. Aust Dent in early childhood caries. Community Dent Oral Epidemiol.
J. 2013;58(2):133–140 quiz 266. doi:10.1111/adj.12039. 1998;26(1):32–44 Suppl.

53. Belvin C, Britton J, Holmes J, Langley T. Parental smoking and 56. Burt BA, Pai S. Does low birthweight increase the risk of caries?
child poverty in the UK: an analysis of national survey data. BMC A systematic review. J Dent Educ. 2001;65(10):1024–1027.
Public Health. 2015;15:507 ,. doi:10.1186/s12889- 015- 1797- z.

16 Volume 22, Number 3