Caries in children with and without orofacial clefting: A systematic review and meta‐analysis

Abstract This systematic review compared children's primary dentition caries experience for those with cleft lip and/or palate (CL/P) and without. Four databases were searched without date restriction for; cross‐sectional studies comparing caries experience for children with CL/P to those without. Screening, data extraction and risk assessment were carried out independently (in duplicate). Meta‐analyses used a random‐effects model. Twenty studies (21 reports) fitting the inclusion criteria comprised 4647 children in primary dentition from 12 countries. For dmft (n = 3016 children; 15 groups), CL/P mean = 3.2; standard deviation = 2.22 and no CL/P mean dmft = 2.5; sd 1.53. For dmfs (n = 1095 children; 6 groups), CL/P mean = 4; sd = 3.5 and no CL/P mean = 3; sd = 2.8. For % caries experience (n = 1094 children; 7 groups), CL/P mean = 65%; sd = 20.8 and no CL/P mean = 52%; sd = 28.1. Meta‐analysis showed higher caries experience in children with CL/P, standardised mean difference = 0.46; 95% CI = 0.15, 0.77. Studies' risk of bias was high (n = 7), medium (n−10) and low (n = 3). Children with CL/P had higher caries experience compared to those without CLP.

suggested. Not only are the reasons for a burden of additional dental caries experience for children with CL/P unclear, but the additional disease experience has not been quantified. Furthermore, it is understood that caries experience in the primary dentition leads to an undesirable dental disease trajectory throughout life (Hall-Scullin et al., 2017), which may further compound oral health inequality in this group.
A systematic review and meta-analysis on caries in CL/P patients was conducted by Antonarakis et al., 2013, nearly a decade ago (Antonarakis et al., 2013) and Worth et al. conducted a further systematic review and meta-analysis on the same topic in 2017.
Since 2017, there have been several primary research papers published on caries in CL/P warranting an update to the secondary literature (Chaudhari et al., 2021;Howe et al., 2017;Malay et al., 2021;Nagappan et al., 2019;Sunderji et al., 2017). Worth's review included 24 studies, two of which used national data for control groups that, while allowing for estimation of prevalence and direction of effect, may reduce confidence in the accuracy of the extent of any differences between groups. In addition, there has been some indication in the literature that the type and side of cleft might, respectively, influence the extent and location of caries.
Systematic review and meta-analysis methodology is constantly improving and becoming more rigorous. Several more primary studies allowing for isolation of primary dentition caries experience have been carried out since the previous reviews' publication. In addition, no paper has looked at the care index (CI) and restorative index (RI) differences for children with and with no CL/P. This means that it is now timely to look at the question of quantification of dental caries disease burden and treatment carried out in the primary dentition of children with CL/P, compared with children with no CL/P using a systematic review of the scientific epidemiological and clinical literature, with meta-analyses of the data where possible.

| AIM AND OBJEC TIVE S
The aim was to investigate whether dental caries experience in the primary dentition of children with cleft lip and/or palate was different to that of children with no cleft lip and/or palate. The primary objective was to compare differences in caries rates in the primary dentition of children with cleft lip and/or palate (CL/P) to children without CL/P using dmft, or % caries experience, or other outcome measure, while the secondary objectives were to investigate whether there is an association between type of cleft and caries experience, and investigate the CI and RI between the CL/P and non--CL/P groups.

| ME THODS
The protocol for this review was developed in accordance with Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) (Page et al., 2021). The review protocol was published in PROSPERO (https://www.crd.york.ac.uk/prosp ero/ displ ay_record.php?ID=CRD42 02128 9287).

| Searches
No date restrictions were imposed on database searches; however, papers published in languages other than English were excluded due to insufficient translation resources.

| Participants/population
Children in the primary dentition with non-syndromic CL/P and similar groups of children without CL/P.

| Study design
Observational and epidemiological primary research with data on caries experience collected through clinical examination were included. Studies that included permanent, primary and mixed dentitions were included; however, only primary dentition data were extracted. If it was not possible to extract data for the primary dentition in isolation, the study was excluded. Studies using national data as a comparator for the caries experience of children with CL/P were excluded.

| Study selection
Literature search results were entered into Microsoft Endnote software before completion of deduplication. Titles and abstracts were screened independently and in duplicate in Rayyan™ QCRI. Where there was disagreement on a study's eligibility for inclusion, consensus was achieved through discussion with a third reviewer. Following this, full texts were obtained and screened independently and in duplicate. Reviewers were not blinded to journal titles, study authors or institutions. The screening process was reported according to the Preferred Reported Items for Systematic Reviews and Meta-Analysis (PRISMA) flow diagram (Figure 1) in addition to reasons for exclusion.

| Data extraction
Data extraction forms were developed and piloted by two authors.
Piloting and subsequent discussion resulted in identification of required revisions and amendments, before extraction forms were re-piloted.
The forms following consensus and agreement between three authors (RG/MR/NI) were used to collect data independently and in duplicate by two investigators for data fields relating to the primary and secondary outcomes (RG/MR) and checked by a third (NI). A single investigator extracted data for the remaining fields relating to study characteristics, and a 10% randomly selected sample was extracted independently by one other investigator and cross-checked to ensure there were no systematic data extraction errors. There were no systematic errors. The following data fields were extracted from included primary studies:

| Data synthesis
Random-effects meta-analysis presented using standardised mean differences (SMD) was carried out for studies with adequate F I G U R E 1 PRISMA flow diagram quantitative data with limited heterogeneity between collected caries experience data.
Mean dmft/deft/dmfs/dfs/ds caries experience for each study for children with CL/P and children with no CL/P was extracted or, where possible, calculated from the data presented. Studies that presented caries experience as a percentage of subjects within groups were excluded from meta-analysis; instead, these data were extracted and described narratively and are presented in tables. Subgroup analyses were carried out for dentition types investigated (primary, primary and mixed together and mixed) where these data were available. Variance estimates were included in the meta-analyses.
Standardised mean difference was used for caries experience across the groups which allowed the difference in size of caries experience between CL/P and no CL/P groups to be compared for different outcome measures (e.g. dmft and dmfs). Publication bias was evaluated using funnel plots. Asymmetric plots were considered to be an indication of publication bias.

| Risk of bias assessment
An adapted Newcastle-Ottawa Scale (NOS) was used to assess and grade the quality of studies with a maximum score of ten points (five for 'Selection', two for 'Comparability' and three for 'Outcome') spread across seven domains (Appendix S1): The adapted NOS assessed the studies for sample representativeness of underlying populations; sample size; non-respondents; comparability of outcome groups with controlled confounding factors; outcome assessment; and statistical testing. Scoring was conducted in duplicate, with a third reviewer to resolve any conflict. Studies were considered at low ROB when the overall scores are 7-8; medium ROB when scores are 4-6; and high ROB when they are 0-3 (Losilla et al., 2018).
Following title and abstract screening, 30 full texts were checked.
All potentially included studies, and all systematic reviews found during searching had references screened. Twenty studies met the inclusion criteria (Appendix S4).

| Size of the participant groups
There were 4647 included in the studies, with data relating to the primary dentition. This comprised 2091 children in the CL/P group and 2,556 in the control, no CL/P group, with 3863 in the metaanalyses (1655 CL/P; 2208 no CL/P). In the CL/P group, the number of participants ranged from 5 to 295 per study (mean = 103.95 sd = 89.80; median = 79) and for the no CL/P group participants per study ranged from 5 to 548 (mean = 123.65; sd = 135.31; median = 77.5).

| Participant ages
All studies reported on the caries status or primary teeth as this was one of the criteria for inclusion in the review.
Although only primary dentition data (or in some cases mixed dentition data) were collected, the participants were aged between 0 and 18 with a wide variety of age groups within this age bracket. There were nine studies that presented data from children within the age range from birth to 6 years and likely to have only primary teeth (Chopra et al., 2014;King et al., 2013;Kirchberg et al., 2014;Mutarai et al., 2008;Sunderji et al., 2017;Dahllof et al., 1989;Zhu et al., 2010;Sundell et al., 2016;Bokhout et al., 1996), six that presented data for children aged 5 years or over and likely to have caries data related to primary teeth as part of the mixed dentition phase of development (Ahluwalia et al., 2004;Chaudhari et al., 2021;King et al., 2013;Nagappan et al., 2019;Sundell et al., 2016;Zhu et al., 2010) and eight where dmft was presented although the age group crossed the primary and mixed dentitions (Hewson et al., 2001;Howe et al., 2017;Lucas et al., 2000;Malay et al., 2021;Parapanisiou et al., 2009;Sunderji et al., 2017;Tannure et al., 2012;Zhu et al., 2010) (Table 1). Three studies presented separate datasets for children in the primary dentition and mixed dentition stages (King et al., 2013;Sundell et al., 2016;Zhu et al., 2010).

| Non-participation and representatives of the participant groups
Caries experience for children with CL/P compared to children with no CL/P Table 4 shows summary data for each outcome measure with the underlying data from individual studies in Appendix S2. Caries experience scores were higher for all outcome measures apart from %dft.
Meta-analysis found no evidence of a difference in the combined primary and mixed dentitions (SMD = 0.09 95% CI = −0.59, 0.76).
The overall effect across all dentitions combined demonstrated a statistically significant difference in caries experience between CL/P and non-CL/P with CL/P children experiencing a higher caries burden (SMD = 0.46 95% CI = 0.15-0.77), p = 0.004 ( Figure 2).

Assessment of publication bias
Publication bias was assessed as part of meta-analysis using a funnel plot ( Figure 3) and Egger's regression intercept test. Asymmetry with respect to the x-axis' overall effect indicates a possible presence of publication bias across primary studies.

Care index (CI) and restorative index (RI)
Five studies provided some component data of caries measurement indices (Dahllof et al., 1989;Howe, 2017;King et al., 2013;Nagappan, 2019, Veiga, 2016. This was only available for the primary dentition. Three studies reported decayed, missing and filled data across groups allowing for calculation of the CI (f/d+m+f) and RI (f/d+f) (King et al., 2013;Nagappan, 2019;Veiga, 2016), while two studies reported decayed and filled data allowing for calculation of the RI only (Dahllof et al., 1989;Howe, 2017) (Table 5).
Of the five studies for which the RI could be calculated, three of them (Dahllof et al., 1989;Nagappan, 2019;Veiga, 2016) showed a lower RI for children with CL/P compared to children with no CL/P demonstrating a lower proportion of restoratively managed carious teeth. Two out of three studies for which the CI could be calculated (Nagappan, 2019;Veiga, 2016) showed a higher CI for non-CL/P children demonstrating a higher proportion of extraction-based treatment in CL/P children compared to non-CL/P children.
The standardised mean difference (SMD) was used in the metaanalysis as this allows different scales used to measure caries (dmft, dmfs, etc.) to be compared directly. SMD uses a ratio between the outcomes to allow the populations (children with CL/P and those without) to be compared. While SMDs are not intuitive to interpret and are limited in their ability to tell us the actual value of the difference between two populations, they allow us to measure the strength of evidence against the null hypothesis of no difference between them.
The linear scale forest plot shows the sample estimates for each trial, the subgroup analyses and overall analyses (together with their 95% confidence intervals) symmetrically around the line of no difference.
We only extracted data that specified the primary dentition so dmfs/ dmft/% caries experience in the primary detention (or the corresponding age group). We subgrouped the unique primary tooth caries experience data, by the child's age, into the 10 studies relating only to the primary dentition which showed a difference between the two populations and higher caries incidence in the CL/P group (SMD = 0.27; 95% CI = 0.06, 0.49), the four covering both the primary and mixed dentition which showed no evidence of a difference (0.09; −0.59, 0.76) and the five with data on primary tooth caries experience in the mixed dentition which again showed a difference with caries experience being higher again the CL/P group (1.74; 0.24, 3.25).
There have been several previous systematic reviews which have also shown a positive relationship between increased caries rates and children with orofacial clefts.
The most recent review was by Worth et al, 2017 where 17 studies with primary teeth were included. However, we excluded five of the studies they included because they were considered to be at high risk of overlapping data with another included study (Hazza'a et al., 2011), compared with national data rather than a directly comparable group (Britton and Welbury, 2010) were not in English language (Hochstein and Hochstein, 1970;Bethmann et al., 1967;Bethmann et al., 1968) or included children likely to be in the permanent dentition (Pisek et al., 2014).
Only five studies allowed the CI and/or the RI for children with CL/P and those with no CL/P to be calculated. Although there was limited data, and this was disproportionately available for the few studies that found children with CL/P to have less caries, it was still apparent that children with no CL/P were more likely to receive interventive management of dental caries. In addition, when that intervention was delivered, for children with CL/P, they were more likely to receive extraction-based treatment compared to children with no CL/P who were more likely to have restorative-based treatment. This finding is similar to a systematic review of children with and without learning disabilities (Robertson et al., 2019). However, the majority of studies in this review did not provide component data of caries measurement indices; therefore, caution should be exercised in the interpretation of these data as it is not representative of all included studies.

F I G U R E 2
Forrest plot as part of random-effects meta-analysis portraying caries experience of CL/P children compared to non-CL/P children in: 1.1.1 the primary dentition; 1.1.2 the primary and mixed dentition; and 1.1.3 the mixed dentition presented as standardised mean differences for dmft/DMFT and 95% confidence intervals (CI). Each subgroup has a black diamond to illustrate the point estimate and CI of effect, with the final diamond illustrating the overall difference in caries experience between the CL/P and non-CL/P children There are a number of putative factors that might contribute to the increased caries risk. Across global epidemiological data sets, children with CL/P are generally born into more socioeconomically deprived areas than those children without CL/P (Chung et al., 2019;Swanson et al., 2017). Many parents/caregivers of children with CL/P have a reduced oral health literacy and poor engagement with healthcare professionals, inadequate home-delivered oral hygiene practices and increased consumption of dietary fermentable carbohydrates (Baskaradoss, 2018). Furthermore, certain risk factors may contribute to the correlation between CL/P and increased caries experience including feeding practice differing between children with CL/P and those without, with CL/P often having extended periods of bottle-feeding; with a greater sugar content compared to breast-feeding (Lin and Tsai, 1999).
Individuals with CL/P have been found to have a reduced salivary flow with normal salivary flow rates in 55% children with CL/P (0.7 ml/min) compared to 66% of non-cleft children (Parapanisiou et al., 2009). Mouth breathing is common in those with CL/P (Halitchi et al., 2017;Hazza'a et al., 2011;Tuaño-Cabrera et al., 2017) and considered a risk modifier to dental caries due to the drying effect it has on the oral cavity and reducing salivary flow.
Finally, there may be an association between the anatomical features of the repaired cleft where scar tissue and dental irregularity can affect access to parts of the oral cavity and result in prolonged retention of residual food.
The possibility that cleft areas are associated with areas of stagnation, and the promotion of dysbiotic plaque biofilms, is supported by several observations: the presence of a fistula (Richards et al., 2015); the severity of the cleft (Lehtonen et al., 2015;Mian et al., 2005); and untreated clefts (Kamble et al., 2017) have been positively associated with increased caries. As such, oral health promotion measures in these patients must be prioritised by clinical teams to mitigate against the increased caries risk, and caries incidence, experienced by those patients with CL/P; early access to dental care may help to preclude dental caries, and associated treatments, in childhood and into adulthood.
Twenty studies were included in this systematic review and meta-analysis. There was significant inter-study heterogeneity of data across domains, especially with respect to outcome reporting and outcome measures. Variation existed among study methodologies and measurements criteria used to assess caries experience.
Three studies used radiographs in addition to clinical examination to detect caries, improving the accuracy of diagnosis (Parapanisiou et al., 2009;Dahllof et al., 1989;Ahluwalia et al., 2004).
A limitation of this review is that only studies published in the English language were included, due to a lack of resources available F I G U R E 3 Funnel plot as part of random-effects meta-analysis to investigate publication bias across primary studies Assessing studies' risk of bias, using a modified Newcastle-Ottawa Scale (NOS), it was found that the majority of studies were of medium risk of bias, followed by high risk of bias, with only 3 studies having a low risk of bias (Howe, 2017;Kirchberg et al., 2014;Sundell et al., 2016). The quality of the studies included should be considered and the influence this has on the concluding outcomes of this systematic review.
The methodology used in this systematic review and metaanalysis is a widely accepted, and standard practice. However, it is important to appraise the methodology used in this systematic review for its overarching strengths and limitations.

| CON CLUS IONS
The overall findings of the meta-analyses align with previous global literature, showing higher caries experience in children with CL/P compared to those without, for children's primary and mixed dentition caries experience. This translates into additional burden of morbidity and need for care, over and above the care needs as a result of the orofacial clefting. Additionally, the data indicate that children with CL/P may receive less restorative treatment for dental caries than non-CL/P children, and be provided with more extractionbased treatment.

| IMPLI C ATI ON S FOR FUTURE RE S E ARCH
Following this thorough systematic review of caries in CLP which included five studies in the last 3 years, some notable gaps in the literature were detected, and this will help inform future work in this field. A subgroup analysis comparing CP and CL/P as different entities was not possible due to the dearth of data on isolated CP.
Likewise additional evidence for the extent to which caries experience in the permanent dentition in those with CL/P would provide a more complete picture of residual morbidity and oral health-related quality of life due to dental caries.
Another area for future research attention is further enquiry into the aetiology of caries in CL/P with an explanation of why there is a consistently greater caries experience among this marginalised group, and this would inform the preventative strategies to combat the problem. From current evidence, it is likely that the causes are a combination of biological and environmental factors, and there are newly emerging methods for behaviour change that could be employed to address the problem.
Future clinical trials in this field would benefit from the provision of a core set of outcomes informing risk factors for the initiation and progression of dental caries in CL/P children, while also defining these outcomes. There are now three systematic reviews in this field, and defined outcome sets would assist trialists and systematic reviewers in pooling outcomes with respect to methodological design and homogeneity of data. Lastly, the literature would benefit from future work in which the contributing components of dmft/ DMFT are presented, facilitating derivation of more comprehensive and representative CI and RI for children with CL/P and control groups. This may help in future training of dental professionals in terms of their clinical management of children with CL/P and with resource allocation.

ACK N OWLED G EM ENTS
The authors acknowledge, and are grateful to, Mr Scott McGregor, Library and learning Technologist at the University of Dundee's Library and Learning Centre, for his expertise in assisting with all aspects of database searching and acquisition of relevant literature.

CO N FLI C T O F I NTE R E S T
No conflicts to declare.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/odi.14183.