3.0. Summary of studies
The search strategy through the databases, registers, and manual searches returned 1560 articles. Following deduplication, 869 articles were removed; 645 articles were further removed following a screening of titles, citations, and abstracts. A total of 46 articles were considered relevant for selection however, 14 studies were eliminated for not having a full-text manuscript. Of the remaining 32 studies for inclusion, 9 articles were excluded after full-text screening with reasons (secondary studies, cohort studies without relevant information on study sample and cases, review articles). PRISMA flowchart for selection of included studies is in Fig. 1.
3.1. Study Characteristics
The included studies were published between 1983 and 2023 and included studies from all the geopolitical zones of Nigeria except the North Central zone. There were 16 studies from the Southwest zone; Ile-Ife,15,20–21,29−32 Osun State, and Lagos State had seven studies each.17,19,22,25,33,35–36 The other two Southwest studies were from Ibadan.16,23 The South-south,26–27 Southeast,18,28 and Northwestern24,34 zones have two articles each, while the Northeast37 has an article. The sample size of included studies ranged from 13 to 1549 participants; case counts ranged from 5 to 183 children.
Study locations were mostly school-based (11)16,18,25–28, 30, 33–35,37 followed by hospital-based studies (8);17,19,21–24,31,36 four included studies were conducted in the community or household setting.15,20,27,32 All the studies utilized the decayed, missing, and filled teeth index (DMF) as the diagnostic index for caries. The studies of Folayan et al. in 202120 and Oziegbe et al.30 in 2013 utilized the PUFA index alongside deft criteria. PUFA stands for Pulpal involvement (p), ulceration due to trauma from sharp tooth pieces (u), fistula (f) formation from a pulpal-involved tooth, and abscess (a) from a carious tooth.
The 23 articles in this systematic review and meta-analysis include Folayan et al. (2015),15 Iyun et al. (2014),16 Sowole et al. (2007),17 Onyejaka et al.(2016),18 Olatosi et al. (2015),19, Folayan et al. (2021),20 Adekoya-Sofowora et al. (2006),21 Omotuyole et al.(2023),22 Popoola et al. (2013),23 Oziegbe et al. (2013),24 Sofola et al.(2013),25 Eigbobo et al. (2017),26 Alakija et al.27 (1983), Onyejaka et al.(2021),28 Folayan et al.(2019),29 Oziegbe et al. (2023),30 Folayan et al. (2008),31 Kolawole et al. (2016),32 Folayan et al. (2012),33 Osuhor et al. (1983),34 Adeniyi et al. (2016),35 Oredugba et al. (2008),36 and Ligali et al. (2023).37 Table 1 shows the list of the articles and the data extracted from the studies.
Table 1
Summary of the main characteristics of the eligible studies.
Author | Year | Location | Zone | Total | Cases | Index | Mean_dmft | SD_dmft | Study_site |
Folayan et al15 | 2015 | Ife | SW | 497 | 33 | dmft | 0.15 | 0.66 | Community Based |
Iyun et al16 | 2014 | Ibadan | SW | 540 | 127 | dmft | 0.65 | 1.49 | School Based |
Sowole et al17 | 2007 | Lagos | SW | 389 | 41 | dmft | 2.49 | 2.9 | Hospital Based |
Onyejaka et al18 | 2016 | Enugu | SE | 429 | 42 | dmft | 0.37 | 0.12 | School Based |
Olatosi et al19 | 2015 | Lagos | SW | 302 | 64 | dmft | 0.735 | 2.07 | Hospital Based |
Folayan et al20 | 2021 | Ife | SW | 1549 | 73 | dmft, PUFA, ICDAS | 0.13 | 0.92 | Community Based |
Adekoya-Sofowora et al21 | 2006 | Ife | SW | 423 | 47 | dmft | 0.3 | | Hospital Based |
Omotuyole et al. 22 | 2023 | Lagos | SW | 30 | 5 | dmft | | | Hospital Based |
Popoola et al. 23 | 2013 | Ibadan | SW | 39 | 19 | dmft | | | Hospital Based |
Oziegbe et al. 24 | 2023 | Kano | NW | 346 | 11 | dmft | 0.08 | 0.5 | Hospital Based |
Sofola et al25 | 2013 | Lagos | SW | 182 | 15 | dmft | 0.19 | 0.13 | School Based |
Eigbobo et al26 | 2017 | Port-Harcourt | SS | 160 | 19 | dmft | 0.25 | 0.82 | School Based |
Alakija et al27 | 1983 | Benin | SS | 83 | 10 | dmft | 1.97 | 0.03 | School Based |
Onyejaka et al28 | 2021 | Enugu | SE | 48 | 9 | dmft | | | School Based |
Folayan et al29 | 2019 | Ife | SW | 918 | 43 | dmft | | | Community Based |
Oziegbe et al. 30 | 2013 | Ife | SW | 195 | 33 | dmft, PUFA | 0.58 | 0.14 | School Based |
Folayan et al31 | 2008 | Ife | SW | 106 | 42 | dmft | | | Hospital Based |
Kolawole et al32 | 2016 | Ife | SW | 497 | 30 | dmft | | | Community Based |
Folayan et al33 | 2012 | Lagos | SW | 104 | 15 | dmft | 0.34 | 1.31 | School Based |
Osuhor et al34 | 1983 | Zaria | NW | 56 | 25 | dmft | 1.3 | | School Based |
Adeniyi et al35 | 2016 | Lagos | SW | 217 | 183 | dmft | 0.34 | 1.01 | School Based |
Oredugba et al36 | 2008 | Lagos | SW | 13 | 5 | dmft | 1.46 | 2.06 | Hospital Based |
Ligali et al37 | 2023 | Maiduguri | NE | 239 | 71 | dmft | 0.72 | 1.39 | School Based |
3.2. Risk of Bias Assessment for included studies.
Individual studies were critically appraised for risk of bias using the (Joanna Briggs Institute) JBI checklist for prevalence studies which comprised nine (9) questions with four likely potential responses. According to the JBI checklist, the studies ranged from low risk of bias (19 studies) to moderate risk of bias (4 studies). All studies were thus included in the meta-analysis for downstream computation of pooled estimate of prevalence and pooled mean dmft of ECC in Nigeria. Table 2 displays the JBI checklist for included studies.
Table 2
Assessment of included studies using the JBI criteria.
Studies | Q1 | Q2 | Q3 | Q4 | Q5 | Q6 | Q7 | Q8 | Q9 | %Yes | Risk |
Folayan et al. 201515 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 100 | Low |
Iyun et al. 201416 | Yes | Yes | No | Yes | Yes | Yes | Yes | Yes | No | 77.8 | Low |
Sowole et al. 200717 | Yes | No | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 88.9 | Low |
Onyejaka et al. 201618 | No | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 88.9 | Low |
Olatosi et al. 201519 | No | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 88.9 | Low |
Folayan et al. 202120 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 100 | Low |
Adekoya-Sofowora et al. 200621 | No | No | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 77.8 | Low |
Omotuyole et al. 202322 | No | No | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 77.8 | Low |
Popoola et al. 201323 | No | No | Yes | Yes | Yes | Yes | Yes | Yes | No | 66.7 | Moderate |
Oziegbe et al. 201324 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 100 | Low |
Sofola. 201325 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 100 | Low |
Eigbobo et al. 201726 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 100 | Low |
Alakija et al. 198327 | No | Yes | Yes | Yes | No | Yes | Yes | No | Yes | 66.7 | Moderate |
Onyejaka et al. 202128 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 100 | Low |
Folayan et al. 201929 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 100 | Low |
Oziegbe et al. 202330 | No | Yes | No | Yes | Yes | Yes | Yes | Yes | Yes | 77.8 | Low |
Folayan et al. 200831 | No | No | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 77.8 | Low |
Kolawole et al. 201632 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 100 | Low |
Folayan et al. 201233 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 100 | Low |
Osuhor et al. 198334 | No | Yes | No | No | Yes | Yes | Yes | Yes | Yes | 66.7 | Moderate |
Adeniyi et al. 201635 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 100 | Low |
Oredugba et al. 200836 | No | No | No | Yes | Yes | Yes | Yes | Yes | Yes | 66.7 | Moderate |
Ligali et al. 202337 | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | 100 | Low |
3.3. Prevalence of Early Childhood Caries in Nigeria
The 23 studies pooled a total of 962 cases from a study population of 7362 participants. Meta-analysis revealed the pooled estimated prevalence of ECC in Nigerian children from included studies is 17% (95% CI: 11–24, p < 0.001). Significant heterogeneity was observed between the studies with I2 = 97% (r2 = 1.2759) as shown in Fig. 2.Figure 2. Forest plot showing pooled prevalence estimate of Early Childhood Caries
3.3.1. Gender predilection
Of the included studies in this systematic review, only 7 articles reported gender distribution of affected ECC cases.15–19,21,30 The male prevalence proportion in these studies is 21.4% (Iyun et al.),16 33.3% (Folayan et al.)15, 39.1% (Olatosi et al.)19, 65.9% (Sowole et al.)17, 54.8% (Onyejaka et al.),18 50% (Adekoya-Sofowora et al.)21 and 18.2% (Oziegbe et al.)30. The number of affected males from these studies is 111 and the affected female children are 126 with a male: female ratio of 1:1.14.
3.3.2. Age groups
Fifteen of the included studies reported prevalence for both the younger age group and the children over 3 years of age while the remaining studies only reported prevalence for children over 3 years. Of all included studies, information on age group delineation was only available in 7 studies with a pooled number of 198 children. Of the 198 children, 33 were children under the age of 3 years while the older age group (> 3 years) was 165 in a 1:5 prevalence ratio.
3.4. Subgroup (moderator) analysis
3.4.0. Year of publication
The heterogeneity of the year of publication was regressed against the prevalence of ECC in children in Nigeria. Meta-regression with year showed nil moderating effect of year of publication on the prevalence of ECC in Nigeria (p = 0.30).
3.4.1. Age groups
No moderating effect of age group of children on the prevalence of ECC following subgroup analysis with age group (p = 0.07). The pooled prevalence of studies on older children was however higher at 27% while the pooled prevalence for studies involving both groups was 12%. Age accounted for 12.1% of the study’s heterogeneity.
3.4.2. Geopolitical Zones
Moderator analysis with geopolitical zones showed a moderating effect of zones on the prevalence of ECC in Nigeria (p < 0.01). The pooled prevalence estimate reveals the North-east has the highest prevalence rate at 30% while the South-south has the lowest at 11%. This is shown in Fig. 3.
3.4.3. Study locations
The study settings or location had a moderating effect on the study’s heterogeneity (p = 0.01), accounting for 33.9% of the study’s heterogeneity. School-based study designs had the highest pooled prevalence rate of 22%; community-based population studies had a low prevalence rate of 5% (Fig. 4).
3.4.4. Mean Diagnostic Index of Early Childhood Caries.
The prevalence of ECC in Nigeria was also regressed against the mean diagnostic index in caries (dmft). The mean dmft had nil moderating effect on the study heterogeneity observed in this meta-analysis (p = 0.24)
Figure 5. Bubble plot showing prevalence of ECC against mean dmft scores
3.5. Sensitivity analysis and influential studies
Studentized residuals of included studies revealed the study of Adeniyi et al.35 to be an outlier with its studentized residual (3.79). Random effects modelled sensitivity analysis also showed Adeniyi et al.35 was an influential study. The Baujat plot and leave-one-out sensitivity analysis plot in Figs. 6a and b also showed this. Following the removal of the study of Adeniyi et al.35, the recomputed pooled prevalence estimate of ECC in Nigeria was 14% (95% CI: 10–20) (Fig. 6b). The study heterogeneity also reduced from an I2 of 97% (95% CI: 96.7% − 97.8%, p < 0.0001) to an I2 of 95% (95% CI: 94.2% − 96.5%, p < 0.0001).
3.6. Publication Bias and Small Study Effect Bias
There was no evidence of publication bias in this meta-analysis and no obvious funnel plot symmetry (Fig. 7). In addition, Egger’s regression test (p = 0.696) and Peter’s regression test (0.126) also indicated nil significant publication bias. Egger’s regression test also showed no significant effect of small study effect bias (Table 3).
Figure 6a. Baujat plot showing the influential study in the prevalence of ECC.
Due to the absence of evidence for publication and small study effect bias, a trim and fill bias corrected method was not necessary.
Table 3
Linear regression test of funnel plot asymmetry |
Egger’s Test | Intercept | Confidence Interval | t | p |
1.376 | 0-5.44–8.19 | 0.396 | 0.696 |
Bias estimate: 1.3762 (SE = 3.4776) |
3.7. Mean DMFT of Early Childhood Caries in Nigeria.
Only 17 of the 23 included studies reported the mean dmft of ECC in their individual treatise. These 17 studies include a total number of 814 cases from 5724 total participants. The pooled mean dmft of ECC from the included studies is 0.44 (95% CI: 0.24–0.74). Similar to the pooled prevalence estimate, there is also significant heterogeneity with I2 at 100% (r2 = 0.9617) (Fig. 8).
Subgroup analysis for the mean dmft across geopolitical zones and study settings revealed that both factors had a moderating effect on mean dmft scores (Fig. 9a and b). Northeast and South south zones had the highest mean dmft at 0.72 (95% CI: 0.56: 0.921) and 0.71 (95% CI:0.09: 5.39) respectively. Community-based studies had the lowest mean dmft at 0.14 (95% CI: 0.11–0.18) while school-based studies had the highest mean dmft at 0.48 (95% CI: 0.30–0.76)
Meta-regression across year of publication also revealed significant moderating effect of this factor on mean dmft of ECC (p = 0.0024) with year of publication accounting for 38.2% of the study variability on mean dmft. Age group of children, however, had no moderating effect on the mean dmft of ECC in this meta-analysis (p = 0.16).
Figure 9a: Forest plot showing subgroup analysis with geopolitical zones
Figure 9b: Forest plot showing subgroup analysis with study settings
3.8. Patient and public involvement statement
Neither patients nor the public was involved in the design, conduct, reporting, or dissemination of research plans.