Association of Molar Incisor Hypomineralization with Hypomineralized Second Primary Molars: An Updated Systematic Review with a Meta-Analysis and Trial Sequential Analysis

Molar incisor hypomineralization (MIH) clinically manifests as demarcated opacities on the first permanent molars/incisors [1, 2], which in severe cases can lead to post-eruptive enamel breakdown accompanied by dental sensitivity, atypical caries/restorations, and extractions due to MIH [3, 4]. The detection of demarcated opacities in both the primary and permanent dentitions has been reported. This condition in second primary molars (SPMs) is known as hypomineralized second primary molars (HSPM) [5]. This condition may develop because SPMs and first permanent molars (FPMs) share a period of amelogenesis and mineralization, so risk factors arising during this period could affect both of the teeth simultaneously [6, 7].

HSPM and MIH share similar clinical presentations, structural properties, and putative etiologies [6, 8]. Some studies have shown that HSPM increases the risk of developing MIH, although the absence of HSPM does not exclude an MIH diagnosis [6‒8]. Consequently, concerns have emerged that HSPM is not only a phenomenon related to MIH but also represents a predictive factor for MIH [6, 9]. However, few studies have confirmed this association with statistical significance [10, 11].

Teeth affected by MIH/HSPM appear to be susceptible to similar problems, including increased susceptibility to carious lesion development, increased sensitivity, and an increased need for restorations and extractions. MIH is a multifactorial disorder with the possible influence of local, systemic, genetic, and environmental factors [12]. Unfortunately, no effective preventive measures can currently be taken against MIH. Therefore, an improved understanding of the association between HSPM and MIH is needed. If HSPM represents a unique predictive factor for MIH, MIH could reward attention and be implemented as a preventive program to minimize post-eruptive enamel breakdown and severe caries. Therefore, the aim of this systematic review was to evaluate and summarize the relationship between HSPM and MIH to provide clinical recommendations for predicting the health status of permanent teeth.

This systematic review and meta-analysis was designed and reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [13]. We prospectively registered this systematic review in the International Prospective Register of Systematic Reviews (PROSPERO database) under the protocol CRD42024497019 (https://www.crd.york.ac.uk/PROSPERO/).

The review question was “is there an association between MIH and HSPM?”. This question was based on the Population, Exposure, Comparators, Outcomes, and Study design (PECOS) format:

• P: the population: appropriately aged children with HSPM and MIH examination data;

• E: the exposure: children with HSPM;

• C: the comparators: non-HSPM children in the studies;

• O: the outcomes: prevalence of MIH in HSPM children and non-HSPM children;

• S: the study design: cohort studies, case-control studies, and cross-sectional studies.

A literature search of the PubMed, Embase, Web of Science, and Cochrane Library databases was performed to assess the scientific evidence supporting this association (shown in Table 1). The survey covered the period up to December 2022. In addition, the gray literature using the System for Information Grey Literature in Europe (SIGLE) through OpenGrey was also searched for additional relevant studies. To ensure literature saturation, manual searches of the reference lists of the selected studies were performed.

The inclusion criteria for the review included the following:

• 1.

  Study design: to identify all the studies that examined the relationship between MIH and HSPM, this study included cohort studies, case-control studies, and cross-sectional studies.

• 2.

  Child participants: children who participated in the epidemiological survey of HSPM and MIH were enrolled. Studies involving participants of younger than 18 years with MIH were included.

• 3.

  Enamel defect description: title and abstract were related to the presence of any developmental defects of enamel including both quantitative and qualitative enamel defects such as enamel hypoplasia, enamel hypomineralization, diffuse opacities (fluorosis), and amelogenesis imperfecta.

• 4.

  Report characteristics: no restrictions on setting or geographical location were applied. The search was limited to English and Chinese language articles.

The exclusion criteria were as follows:

• 1.

  Studies that did not include records of HSPM/MIH. Studies with illegible descriptions of enamel defects or otherwise not in accordance with the MIH/HSPM diagnostic criteria were excluded.

• 2.

  Studies containing insufficient data to allow us to calculate the respective prevalence and number of MIH/HSPM patients and the number of patients affected by both MIH and HSPM were excluded. For articles with incomplete or unclear information, the corresponding authors were contacted by e-mail to confirm the ambiguous data.

• 3.

  For studies with duplicate data based on the same population, only the study that reported the most detailed data was included, and the rest were excluded.

• 4.

  Redundant studies that were identified in more than one database were excluded.

All the literature in the four databases was searched and imported into the literature management software EndNote (EndNote 20; Thomson Reuters, Toronto, Canada). After removing the duplicates from the four databases, two reviewers (Zhaoxin Zhang and Yueying Liu) independently screened the titles and abstracts of the identified studies. Next, the full texts of the selected articles were evaluated, and studies were eliminated according to the exclusion criteria to screen the final included articles. Disagreements were resolved by direct discussion, or by consultation with one other reviewer (Jie Jia) until a consensus agreement was reached among all the authors.

Data extraction was performed using Excel spreadsheets by two reviewers (Zhaoxin Zhang and Yueying Liu) and checked by two other reviewers (Jie Jia and Yaxin Zhu). Disagreements were discussed until a consensus was reached. The following standardized data were extracted: study characteristics (author, publication year, title, publication journal), study design, geographic region, patient selection criteria (inclusion and exclusion criteria), MIH/HSPM diagnostic criteria (European Archives of Pediatric Dentistry [EAPD] diagnostic criteria and/or self-devised criteria), sample size of each eligible study, number of patients included (the number of MIH patients, the number of HSPM patients, and the number of patients with both MIH and HSPM), patient demographics (age, sex, country of residence, ethnicity), and defect characteristics.

For calibration purposes, random selection of papers (10% of the articles identified by initial search; n = 1,020) were assessed. Kappa analysis was carried out to determine the inter-examiner reliability for the inclusion of the selected papers (inclusion criteria 1–4) and for the reasons why the article was not included (exclusion criteria 1–4).

Risk-of-bias assessment was determined independently by two reviewers (Zhaoxin Zhang and Yueying Liu) for each study, using the Newcastle-Ottawa Scale (http://www.ohri.ca/programs/clinical-epidemiology/pxford.asp) tool for cohort studies and case-control studies [14, 15]. Using this tool, each study was judged on 8 items and categorized into 3 groups: sample selection (four stars), comparability of groups (two stars), and measures of exposure/outcomes (three stars). The maximum score was 9 stars. Studies that received ≥7 stars were considered to be of high quality or to have a low risk of bias. Those with <3 stars were rejected for low methodological quality or a high risk of bias.

The Agency for Healthcare Research and Quality (AHRQ) methodology checklist (https://www.ncbi.nlm.nih.gov/books/NBK35156/) for observational studies was used to assess the quality of the cross-sectional studies [15]. The tool consists of an 11-item checklist. An item is scored 0 if it is answered “no” or “unclear”; if it is answered “yes,” then the item is given a score of 1. Studies were further classified as high risk (0–3 points), moderate risk (4–6 points), or low risk (≥8 points) of bias. Studies with <4 points were eliminated.

RevMan 5.4 software (Review Manager 5.4, the Cochrane Collaboration) was used to perform the meta-analysis. The pooled data of all the outcomes were subjected to the meta-analysis to estimate the odds ratios (ORs) and 95% confidence intervals (CIs).

Cochran’s Q test was applied to analyze the heterogeneity between the included studies, and no heterogeneity was detected if the p value was greater than 0.1. Otherwise, the I² test was used to assess the statistical heterogeneity, and the following threshold was used as recommended by the Cochrane Collaboration: 0–25% might not be important; 25–50% may represent moderate heterogeneity; 50–75% may represent substantial heterogeneity; and 75–100% may represent considerably high heterogeneity. A random-effects model was utilized if the study results showed high heterogeneity (p < 0.10, I² > 50%). High heterogeneity means that the results need to be explored by subgroup and sensitivity analyses. The risk of publication bias was assessed by funnel plots and Egger’s test [16].

Then, we conducted a trial sequence analysis (TSA) using TSA software, Copenhagen Trial Unit, version 0.9.5.10 Beta (Centre for Clinical Intervention Research, Copenhagen, Denmark), to determine whether our findings were conclusive, to reduce the frequency of “type-1 errors,” and to ascertain the effect of random errors in the cumulative meta-analysis. The parameters were set as follows: boundary type, two sided; type 1 error α = 5%; type 2 error β = 20%; and statistical power 1 − β = 80%. The information axis was the sample size. The cumulative Z-score, trial sequential monitoring boundary, futility boundary, the conventional boundary of benefit (z = 1.96), and required information size (RIS) are presented in the TSA graph. If the included sample size reached the RIS or the cumulative Z curve crossed the trial sequential monitoring boundary and the conventional boundary of benefit (Z = 1.96), the result was considered reliable.

The details of the literature search and the article selection process are shown in Figure 1. The initial electronic search yielded 10,183 records (2,690 from PubMed, 3,180 from Embase, 298 from Cochrane Library, 4,015 from Web of Science, and 0 from additional records identified through other sources), and 5,735 records remained after removing the duplicates. From these 5,735 records, 43 potentially pertinent records were selected after screening the titles and abstracts. Full-text articles were retrieved for eligibility assessment, and 31 articles were excluded due to the lack of prevalence data for MIH or HSPM (n = 15), diagnostic criteria other than that from the EAPD (n = 3), overlapping data (n = 5), small sample size (n = 1), or different age-groups (n = 7). Ultimately, only 12 articles [6, 8‒10, 17‒24] met the inclusion criteria, including 11 cross-sectional studies and 1 cohort study, and were included in the qualitative analysis [6].

The included studies provided data from 327 to 1,562 participants, for a total of 8,944 included children. The main characteristics of the included studies are shown in Table 2. The articles were published in countries in Asia [8, 9, 22], Europe [6, 17, 23], Oceania [20], Africa [10, 18], and South America [19, 21, 24]. Most studies recorded MIH-related hypomineralization according to the European Academy of Paediatric Dentistry (EAPD) diagnostic criteria [2] (n = 7), and other indices also used the integrated EAPD criteria with the modified developmental defects of enamel index [3, 25] (n = 5). Studies were conducted across various settings, including schools, stomatology departments, and medical centers. Operator training is an essential methodological step for ensuring the reliability of clinical studies and reducing the risk of bias. Ten studies performed intra-examiner reliability tests, with results ranging from 0.61 to 1 (shown in Table 2), highlighting almost perfect intra-examiner agreement.

In terms of the risk-of-bias assessment, the NOS score of the cohort study was 8 stars (shown in Table 2; online suppl. material 1; for all online suppl. material, see https://doi.org/10.1159/000540752). Using the AHRQ tool (shown in Table 2; online suppl. material 2), 2 of the cross-sectional studies had a moderate risk, and 0 had a low risk of bias. All the included studies were of moderate-to-high quality. And the results of sensitivity analysis showed that excluding any study did not change the direction of the outcome, which indicated that the study was relatively stable (shown in online suppl. material 3).

Significant heterogeneity was detected among the studies (I² = 94%, p < 0.05) (shown in Fig. 2). We explored the potential sources of heterogeneity and performed subgroup analyses according to the type of study, oral examination site, geographic region variability, and diagnostic criteria (shown in Fig. 3-6), but none of the examined factors was the source of heterogeneity. No significant publication bias was detected in this meta-analysis according to the funnel plot (shown in Fig. 7). The shapes of the funnel plots for the 12 included studies on HSPM and MIH did not seem to have obvious asymmetry. Moreover, the results of the Egger’s test also showed that publication bias did not materially affect the overall results of the included studies (t = −1.54, p_Egger’s = 0.155, 95% CI = −7.43 to 1.36).

A total of 12 studies were included in the meta-analysis to assess the association between HSPM and MIH incidence. The study population was divided into an HSPM group and a non-HSPM group, and the number of MIH patients in the two groups was investigated separately. Twelve studies including 666 children with HSPM and 8,278 controls were pooled for the present meta-analysis. Compared with patients in the non-HSPM group, HSPM patients were approximately 11 times more likely to have MIH (OR = 10.90, 95% CI = 4.59–25.89, p < 0.05). The overall pooled prevalence of co-occurrence of MIH and HSPM was 4.74%.

The TSA results showed that the Z curve crossed both the conventional boundary of benefit and the trial sequential monitoring boundary. Although the accumulated information did not reach the RIS, the Z curve crossed the trial sequential monitoring boundary in the 8th study (shown in Fig. 8), indicating that the results obtained by the current information were conclusive in advance. Therefore, there was a statistically significant association between HSPM and MIH.

There is a growing interest in addressing the association between HSPM and MIH. Unfortunately, only one systematic review, including 5 studies, compared the incidence of MIH in patients affected by HSPM versus healthy patients [7]. In comparison, a total of 12 studies were included in this meta-analysis, thus confirming the recent increase in the number of studies and expanding interest in the effect of HSPM on MIH. Furthermore, not only did the total number of studies more than double, but there was also a wider geographic spread compared to those included in the previous review. This demonstrates that HSPM/MIH is a global health problem that affects the oral health of children. The results from the 12 included studies revealed an overall positive association between MIH and HSPM (OR = 10.90, 95% CI = 4.59–25.89, p < 0.05), indicating that HSPM is associated with increased odds of MIH. This result is in alignment with that of a previous review [7].

This systematic review included only studies that reported the prevalence of both HSPM and MIH. Furthermore, Elfrink et al. [3] recommended a standardized protocol that included the selection of a random sample of at least 300 children for prevalence studies. Consequently, 12 studies were eligible for this review, and the sample sizes of the included studies ranged from 327 to 1,562 participants.

According to the I² test, significant heterogeneity was observed among studies (I² = 94%), and we used a random-effect model to merge the ORs and 95% CIs. To investigate the source of heterogeneity, we conducted a sensitivity analysis and showed that the exclusion of any single trial did not markedly affect the overall measures of the relationship between HSPM and MIH. In addition, the funnel plot for the 12 included studies did not seem to have obvious asymmetry, and the results of the Egger’s test also showed that publication bias did not materially affect the overall results, suggesting that the results obtained were stable and reliable.

In the subgroup analyses, we observed variability in the geographic region, oral examination site, diagnostic criteria, and type of study. The included studies were from Asia, Africa, Oceania, South America, and Europe. The incidence of HSPM and MIH in each geographic region varied greatly, which could be explained by the characteristics of the different populations. The subgroup analysis revealed that the variability between geographic regions did not drive the heterogeneity, and HSPM increased the risk of MIH in Europe, South America, and Asia. However, there was no correlation between HSPM and MIH in Africa, probably due to the small number of cases of MIH and HSPM. These findings confirmed that geographic region variability drives differences in the incidence of MIH and HSPM but does not affect the association between them. Further multiarea epidemiology studies of HSPM and MIH are still needed to confirm this conclusion.

Of the 12 included studies, 11 were cross-sectional studies, and only 1 was a cohort study. The subgroup analysis of the 11 cross-sectional studies indicated that the type of study did not induce heterogeneity (OR = 12.10, 95% CI = 4.82–30.35), and the presence of HSPM was associated with MIH. In the only cohort study, the rate of developing MIH was four-fold higher in patients affected by HSPM than in normal controls [6]. In the 11 included cross-sectional studies, most of the SPMs were damaged by hypomineralization or other defects, such as dental caries. Almuallem et al. [22] reported that in their study, 18.6% of the SPMs had a clinical status recorded as “unscored” due to severe damage. Therefore, there may be diagnostic difficulties and underestimation of the prevalence of HSPM if the examination occurs too late. Elfrink et al. [3] suggested that the best age for examining for the presence of HSPM is 5 years, owing to the complete eruption of the SPMs and greater cooperation by the child. In contrast, to assess for MIH, the best time is approximately 8 years due to the presence of the FPMs as well as the permanent incisors [3]. Molars with severe MIH are at increased risk for substantial restoration or extraction when children are older than 10 years, which may lead to an underestimation of the prevalence of MIH. In this review, the included studies had a wide age range, some of which exceeded the ideal age for HSPM and MIH diagnosis, which could mask underlying hypomineralization defects and influence the reporting of the true defect prevalence. A longitudinal prospective cohort design is recommended with more accurate data collection, but such studies have methodological challenges due to their duration.

Of the 12 included studies, two involved medical centers (OR = 8.15, 95% CI = 1.60–41.50) [6, 17], and 10 involved schools (OR = 11.57, 95% CI = 4.15–32.24). According to the subgroup analysis, the heterogeneity was not based on the oral examination site, and HSPM was strongly correlated with the incidence of MIH regardless of the oral examination site. Medical centers provided standardized conditions for examination according to specific guidelines. However, examinations performed at schools require additional artificial light to stabilize the natural light source as natural light is susceptible to the underestimation of defects. Nonetheless, some of the included studies [10, 17, 18] still relied on natural light, which may have influenced the reported prevalence of HSPM and MIH. It is essential that there is a standardization of the oral examination site used, which may contribute to overcoming the inconsistencies of examination conditions.

Regarding the diagnostic criteria, we confirmed that the diagnostic criteria of the included studies did not cause heterogeneity, and HSPM and MIH were highly correlated. Among the 12 studies, 7 were conducted according to the EAPD diagnostic criteria of 2003 [2], and 5 studies followed EAPD practice guidelines, which recommend integrating the EAPD criteria with the modified developmental defects of enamel index [25, 26]. All studies were based on the 2003 EAPD criteria for MIH diagnosis, with some variation regarding adaptation of the diagnostic criteria, examination protocols, and data recording. For example, out of 12 studies, 7 studies [6, 8, 10, 17‒19, 22] were conducted with wet tooth surfaces, and 3 other studies [9, 20, 21] were conducted with dry tooth surfaces. Additionally, teeth were viewed as sound when the lesion was less than 2 mm in diameter in 6 studies [8‒10, 18, 20, 21] and 1 mm in four other studies [17, 19, 22, 23]. All these differences and characteristics lead to difficulty in combining data from various studies, which may cause heterogeneity. Elfrink et al. [3] recommended using standardized assessment criteria in the epidemiological study of MIH/HSPM. Standardized assessment criteria were proposed by Ghanim et al. [25, 26], and they are the core components of current guidelines. The guidelines consider methodological parameters related to the diagnostic criteria, differential diagnosis, standardized assessment criteria, kappa value calculations, and operator training, which make the epidemiological study of MIH/HSPM more standardized and uniform [27]. Adherence to these guidelines are recommended to enable a more accurate diagnosis and decrease the variation between studies, eventually improving the quality of studies.

The present systematic review failed to find the source of heterogeneity in the subgroup analyses, which might be driven in part by the small number of included studies and small sample sizes of the included studies. Furthermore, the potential heterogeneity may come from some unavoidable factors, such as the design of the included studies, the characteristics of the included population, and variations in the understanding of the diagnostic criteria, which all affect the inference and accuracy of the final results. Hence, further studies with larger sample sizes and multiple geographic regions are required to compliment this review.

We further used TSA to confirm the adequacy of the accumulated evidence. The use of TSA can reduce random error and systematic bias, improve the reliability of results, and overcome the shortcomings of classical systematic reviews and meta-analyses. TSA is capable of estimating the RIS that reaches a stable conclusion. When the number of included studies is insufficient, TSA can minimize the number of false positives resulting from random error. In this study, the Z curve crossed the conventional boundary of benefit and monitoring boundary, which affirmed the association between MIH and HSPM, although the Z curve did not reach the RIS. The TSA analysis confirmed that HSPM can predict the occurrence of MIH through the 12 included studies, and no additional related studies are needed.

The present systematic review indicated that HSPM is predictive of MIH, but not all MIH cases are accompanied by HSPM. Notably, individuals with MIH but no observable HSPM could have a very mild form of HSPM that is not detectable with the naked eye [23]. However, the severity of HSPM linked with MIH needs to be further investigated.

Our systematic review and meta-analysis could provide evidence for positive associations between HSPM and an increased risk of MIH, and HSPM could play a predictive role in the occurrence of MIH. For HSPM patients, dentists should pay more attention to the FPMs and take early preventive measures to protect the affected teeth. Nevertheless, there are also several limitations. First, the publication languages of the included studies were limited to English, which may cause language bias. Second, the 11 cross-sectional studies included patients who exceeded the ideal age for HSPM diagnosis, which may have affected the diagnosis and identification of HSPM. Consequently, the conclusions of this study still need to be verified by high-quality, multicenter, and large-sample prospective studies.

A statement of ethics is not applicable because this study is based exclusively on published literature.

The authors have no conflicts of interest to declare.

This study was supported by a scientific research project of the Public Health Department of Henan (SBGJ202102197) and a scientific research project of the Science and Technology Department of Kaifeng (2207012).

Zhaoxin Zhang, Yueying Liu, and Jie Jia designed the research and wrote the manuscript. Yaxin Zhu, Jingya Guo, Mingzhen Yang, Yang Lu, and Yimeng Zhang contributed to the data acquisition, analysis, and interpretation and critically revised the manuscript. All authors reviewed the results, commented on previous versions of the manuscript, and approved the final version of the manuscript.

All the data generated or analyzed during this study are included in this article and its online supplementary materials. Further inquiries can be directed to the corresponding author.