Analysis of the packing density of Hunter-Schreger bands (HSB): a systematic review and meta-analysis

Abstract

Objective

This review examines the different configuration of Hunter-Schreger bands (HSBs), focusing on the packing density of HSBs in various regions of the tooth crown, while also considering the clinical implications. This was achieved through a systematic analysis of the literature and by performing a meta-analysis and TSA (Trial Sequential Analysis).The systematic review was conducted following the PRISMA guidelines using electronic databases such as PubMed, Scopus, Cochrane Central Register of Controlled Trials, and ScienceDirect, with the use of combinations of keywords such as 'Hunter–Schreger bands' AND 'packing densities'. The meta-analysis was performed using RevMan 5.4 software. Results: During the selection process, 190 records were identified, of which only 2 studies fully met the inclusion criteria, one of which presented data on packing densities in different areas of the tooth crown. The extracted data were included in the meta-analysis and TSA. The meta-analysis reported a mean difference (MD) of 4.95 [4.96, 5.22] for HSB packing density between the middle third and the cervical third of the enamel. In conclusion, HSB packing density is significantly higher in the middle region of the tooth crown, where greater occlusal forces occur. In contrast, the cervical regions, which have lower HSB density, are more susceptible to erosive and abrasive forces. These findings suggest that the spatial distribution of HSBs influences enamel resistance to mechanical stress and plays a role in the development of NCCLs.

Introduction

The Hunter-Schreger Bands (HSB) are an optical phenomenon that occurs when a enamel surface, following cutting, is observed under reflected light, revealing the presence of an alternating pattern of darker bands. Furthermore, HSB can be seen in uncut enamel and used as a biometric parameter for personal identification in automated systems [1, 2]. These bands are the result of the synchronous crossing of groups of enamel prisms in the horizontal plane and are caused by slight undulations of the prism axis. Depending on the angle of incidence with polarized light, they can appear dark (single-reflective) or bright (birefringent) [3].

From an evolutionary and anthropological perspective, HSBs are present in the enamel of mammalian teeth and seem to have the function of increasing enamel resistance, preventing the formation of cracks [4]. HSB packing density decreases from incisal/occlusal to cervical region of the tooth crown could also have secondary implications in the formation of Non-Caries Cervical Lesions (NCCLs), attributed to abfraction [5].

Among the main causes of NCCLs development, we find abrasive and erosive mechanisms, as well as those defined as "abfraction" [6]. However, it should be noted that the term "abfraction" is not recognized by the scientific community [7], despite being widely used to describe wedge-shaped lesions that typically involve the cervical and vestibular surfaces of premolars, canines, and molars [8]. The distinctive feature of these lesions is the microstructural loss of mineral tissue in areas subjected to high levels of stress [9].

For clarity, it is added that the authors' position regarding the abfraction theory and, more generally, the role of occlusal loading in the formation of wedge-shaped enamel lesions is neutral, as previously stated in earlier publications, which suggest that occlusal loading may be a contributing factor, but not the triggering cause of such lesions [7].

The Hunter-Schreger bands and their spatial distribution on the enamel could therefore play an important aetiopathogenetic as well as clinical role in determining the onset and localisation of certain NCCLs. A special role may be played by the packing density of the HSB, which refers to the number of HSB bands that are found along a given length of the interface between the enamel and the dentin. The packing density of the HSB varies depending on the chewing functions of the teeth. Posterior teeth (molars and premolars) that withstand greater compressive and grinding forces have a higher HSB density than anterior teeth, as do the occlusal and lateral surfaces [10].

Although the varying distribution of HSBs in dental enamel is well documented in in vitro studies, the lack of studies directly correlating HSB packing density with the formation of NCCLs highlights a clear clinical problem. The current scientific literature does not provide adequate answers, leaving a significant gap in the understanding of this phenomenon. Therefore, there is an urgent need for further research to explore the potential link between HSBs and NCCLs.

Although some narrative reviews and individual studies have examined Hunter-Schreger Bands (HSBs), no comprehensive systematic review has synthesised the available evidence on their packing density in different regions of the tooth and secondarily related it to NCCLs. This review aims to address this gap by systematically evaluating the literature to compare the packing density of HSBs in the cervical and middle regions of posterior teeth (primary outcome) and to secondarily correlate NCCLs with packing density (secondary outcome) [10].

This study stands out for the methodological approach adopted, which combines a systematic review and meta-analysis to examine the packing density of HSBs and their potential role in the onset of NCCLs. Unlike previous research that primarily focused on the qualitative observation of HSBs, this work employed a rigorous quantitative analysis to compare the HSB densities in different regions of the tooth crown, particularly between the cervical and middle regions of the molars.

Moreover, another innovative methodological aspect is the use of advanced statistical evaluation techniques, such as sensitivity analysis and Trial Sequential Analysis (TSA), which allowed for a more robust estimation of the results and minimized the risk of type I and II errors. This approach also enabled the identification and reduction of potential sources of heterogeneity in data from in vitro studies, thereby improving the reliability of the conclusions.

The results will contribute to a deeper understanding of enamel structure and its clinical implications, particularly in restorative dentistry and the prevention of enamel-related pathologies and the aetiology of NCCLs.

Materials and methods

Protocol and registration

The systematic review was conducted following a thorough analysis of the literature on the topics of packing density of HSBs in different regions of the tooth crown and, secondarily, in relation to NCCLs. Initially, a narrative review was carried out based on the guidelines provided by the SANRA (Scale for the Assessment of Narrative Review Articles) [11]. This preliminary step led to the decision to perform a systematic review of the literature in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [12]. Consequently, the processes of searching, selecting, and extracting data adhered to the recommendations outlined in the Cochrane Handbook. Furthermore, the review protocol was submitted to and registered on Open Science Framework (OSF), https://osf.io/se3hj.

Eligibility criteria

All studies examining HSBs were considered potentially eligible. Particular attention was given to studies investigating and reporting data on the packing density of HSBs in the cervical regions of teeth.

The formulated PICO question was as follows: What is the mean difference in the packing density of HSBs in posterior teeth (molars) between the cervical region and the middle region of the buccal crown surface?

(P) Participants: Mandibular or maxillary molar teeth.

(I) Intervention: Packing density of HSBs in the cervical third of the mandibular and maxillary molars.

(C) Control: Packing density of HSBs in the middle third of the buccal crown of the mandibular and maxillary molars.

(O) Outcome: Mean difference in packing density between cervical and middle regions.

The inclusion criteria involved including all reports on in vitro studies conducted on extracted human teeth that provided data on the packing density of HSBs, specifically including data from the cervical third and the middle third of the buccal surface of the tooth crown, in order to perform a comparison between the two areas.

The decision to include only in vitro studies was based on the need for controlled experimental conditions that allow for precise measurement and analysis of HSB packing density. In vitro studies provide a more consistent and replicable environment for observing the structural properties of enamel, such as HSB distribution, without the variability introduced by in vivo factors.

Additionally, the following exclusion criteria were applied:

• Manuscripts or reports related to systematic reviews, scoping reviews, mapping reviews, narrative reviews, case reports, case series, in silico studies and clinical studies.

• Studies published in languages other than English or for which a clear translation was unavailable or not possible.

• Reports considered at high risk of bias.

• No exclusion criteria were applied based on the year of publication

Sources of information, research and selection

The search for articles and reports was conducted using online search engines by two reviewers, who are also the authors of the manuscript (M.D. and L.S.). Preliminary exclusion criteria included language restrictions: reports without at least an English abstract were excluded using automated tools available in the databases [13].

The search engines and databases used were PubMed, Scopus, and the Cochrane Central Register of Controlled Trials. Additionally, a search for grey literature was performed using Google Scholar (https://scholar.google.com/scholar?hl=it&as_sdt=0%2C5&q=Hunter%E2%80%93Schreger+bands&btnG =), ScienceDirect (https://www.sciencedirect.com/search?qs=Hunter%E2%80%93Schreger%20bands), and DANS archivie (Data Station Life Sciences) (https://ssh.datastations.nl/dataverse/root?q=Hunter%E2%80%93Schreger++bands&types=datasets&sort=score&order=desc&page=1). To further minimize publication bias, references from previous reviews on HSBs were examined. The search, including the most recent update of identified records, was completed on 10 January 2025.

A latest updated literature search was conducted on February 24, 2025.

Search terms were chosen to include the broadest range of studies focusing on HSBs and their packing density.

For database searches, the following terms were used in combination: Hunter–Schreger bands, packing densities.

On PubMed, the following search terms were used:

Search: Hunter–Schreger bands Sort by: Most Recent.

"Hunter-Schreger"[All Fields] AND ("band s"[All Fields] OR "bands"[All Fields]).

Translations.

bands: "band's"[All Fields] OR "bands"[All Fields].

On scopus: TITLE-ABS-KEY ( hunter–schreger AND bands).

On Cochrane Central Register of Controlled Trials: Hunter–Schreger bands in Title Abstract Keyword.

For greater clarity, a table has been provided and added, listing the web addresses of the databases, the date of the last search, and the records identified, excluding grey literature (Table 1).

Table 1  Table showing the addresses of the databases where the records were searched, to access Scopus you need to be logged in

The identified records were imported into EndNote, and duplicates were automatically detected and removed using the software's "find duplicates" function. Any additional duplicates not recognized by the software were manually removed by the reviewers responsible for article selection.

The selection of articles was conducted independently by two reviewers (M.D. and L.S.). Initially, they listed potentially eligible studies and subsequently included studies in two separate tables, which were later compared. Potentially eligible studies were selected through title screening, while included studies were selected through full-text analysis and reading.

The agreement between the two reviewers was also assessed, and any disagreements were resolved by a third reviewer (A.B.).

Data collection process and data characteristics

The data intended for inclusion in the tables summarizing the selected studies were established during the initial development of the protocol. As with the records evaluated in the selection and screening phases, the data were extracted independently by both reviewers and then cross-checked to minimize any potential reporting inaccuracies. Following this step, one reviewer compiled the verified information into a unified table.

Any inconsistencies identified during the data extraction process were first recorded and acknowledged. To address these issues, the reviewers engaged in thorough discussions to reconcile differences, clarify misunderstandings, and resolve potential errors.

In cases where an agreement could not be reached after these discussions, the unresolved items were escalated to a third reviewer. This reviewer carefully examined the disputed data and provided a final determination on the accuracy of the extraction. This procedure ensured that all conflicts were resolved systematically and transparently, grounded in the evidence available.

The data extracted from the included articles encompassed the first author, year of publication, study design, country of origin, number of teeth analyzed, type of teeth, mean packing density of HSBs, standard deviation (SD) in the cervical and middle thirds, and the number of measured segments.

Risk of bias in individual and across studies, synthesis measures and results, publication bias

Particular emphasis was placed on assessing the risk of bias using Checklist for Reporting In vitro Studies (CRIS) [14, 15]. Studies identified as having a high risk of bias were excluded from the meta-analysis. As with other stages of the research process, the risk of bias assessment was independently performed by two reviewers, A.B. and M.D. (even in this case, discrepancies are resolved by a third reviewer who decides in situations of disagreement), followed by a comparison and consolidation of findings into a single table.

The data and review results were presented in tables, with aggregated data displayed both numerically and graphically. Graphical representations included funnel plots, TSA plots, forest plots, and heterogeneity indices such as Higgins' I². Bias across studies was visually evaluated by examining the overlap of confidence intervals (CI) and quantified using the I² inconsistency index. An I² value greater than 50% was considered moderate, and in specific cases, a random-effects analysis was applied.

For meta-analyses demonstrating high levels of heterogeneity, sensitivity analyses were planned by excluding studies with low CI overlap or disproportionately high study weights. The aggregated mean difference (MD) was calculated using a fixed-effects model, assuming a shared true effect across all studies. The combined MD was computed as a weighted average of the mean differences reported in individual studies, with weights determined by the inverse variance of the mean difference. A 95% confidence interval (CI) was calculated to reflect the uncertainty around the combined effect estimate.

Heterogeneity among studies was assessed using Higgins' I² index and Cochran's Q test. All analyses were conducted using Review Manager (RevMan) version 5.4 (Cochrane Collaboration, Copenhagen, Denmark).

GRADE, and trial sequential analysis (TSA)

The quality of evidence was evaluated using online tools such as GRADEpro Guideline Development Tool (GRADEpro-GDT, Evidence Prime), accessible at https://gdt.gradepro.org/ (last accessed 12 January 2025). The results were summarized in a table [16].

For the power analysis of meta-analytic results, a free software tool for Trial Sequential Analysis (TSA) was utilized. This Java-based software, known as TSA software (Copenhagen Trial Unit, Centre for Clinical Intervention Research, Copenhagen, Denmark), was employed in version 0.9.5.10 Beta.

Results

Study selection

The searches conducted on SCOPUS, PubMed, and Cochrane Central Trial resulted in 190 bibliographic sources. After removing duplicates, 143 potentially eligible articles remained. Of these, only 22 articles underwent full-text screening, and only 2 fully met the inclusion and exclusion criteria. Consequently, the relevant data extracted from these 2 studies were included in the meta-analysis.

The k agreement between the two reviewers who performed the study selection was calculated, and the result was 0.64 (substantial agreement). The K agreement was based on the formulas in the Cochrane Handbook for Systematic Reviews [17]. All data related to the measurement of the agreement have been reported in Table 2

Table 2 K-agreement calculation, Po (Proportion of agreement) = 0.773, Pe (Agreement expected) = 0,370, K-agreement = 0.64 (< 0 no agreement, 0.0–0.20 slight agreement, 0.21–0.40 fair agreement, 0.41–0.60 moderate agreement, 0.61–0.80 substantial agreement, 0.81–1.00 almost perfect agreement)

The results of the selection and inclusion process highlight the lack of studies directly correlating HSB packing density with the development of NCCLs (secondary outcome), suggesting that further experimental investigations, particularly longitudinal clinical studies, are required to draw consistent results.

Additionally, searches of grey literature databases such as DANS archives (Data Station Life Sciences) (https://ssh.datastations.nl/), ScienceDirect, Google Scholar, and previous systematic reviews using "Hunter–Schreger bands" as a keyword did not identify any further studies eligible for inclusion in the meta-analysis. The complete procedure for identifying, selecting, and including studies is outlined in the flowchart shown in Fig. 1; The excluded reports with the relative reasons have been reported in Table 3.

Fig. 1

flow chart of the study selection and inclusion process

Table 3 Excluded reports

Data characteristics

The articles included in the review are two, as follows: Lynch et al. [10], Yang et al., 2022 [38].

The extracted data are presented in a single table. Table 4 represents the data concerning the first author, year of publication, study design, country of origin, number of teeth analysed, type of teeth, mean packing density of HSBs, standard deviation (SD) in the cervical and middle thirds, and the number of measured segments.

Table 4 Data extraction table. Packing density HSB, PD; N, number; SD Standar deviation

The data extracted and used in the meta-analysis specifically relate to the mean packing density of Hunter-Schreger bands (in each dental segment on the buccal surfaces of molar teeth). These were taken and processed from the first table of the study by Lynch et al. [10]. In particular, a weighted average was performed between the values of the lower middle quarter and the upper middle quarter to obtain a single weighted average and weighted standard deviation for the middle portion of the buccal surface of the molars.

The data from the study by Yang et al. [38] come from the supplementary materials attached to their study. In this case, packing density data for each sample and segment analysed were available in an Excel spreadsheet, and the mean and standard deviation were calculated. For Lynch et al. [10], four data sets were extracted: Maxillary first molar (mesial cusps), Maxillary first molar (distal cusps), Mandibular first molar (mesial cusps), Mandibular first molar (distal cusps); whereas for the study by Yang et al. [38], six data sets were extracted: Mandibular first molar, Mandibular second molar, Mandibular third molar, Maxillary first molar, Maxillary second molar, Maxillary third molar.

The total number of segments included in the meta-analysis was 121 for the middle third and 81 for the cervical third, with a total of 81 teeth included in the meta-analysis.

The extracted data from the vestibular surfaces clearly demonstrate a marked difference between the middle and cervical thirds. For example, Lynch et al. [10] reported that, for the maxillary first molar (mesial cusps), the mean packing density was 11.4 (SD 2.26) in the middle third compared with 5.5 (SD 1.5) in the cervical third. Similarly, for the mandibular first molar, the mesial cusps exhibited a mean of 10.85 (SD 1.40) versus 4.1 (SD 0.7) in the cervical region, while the distal cusps recorded values of 10.05 (SD 1.30) and 3.8 (SD 0.6), respectively. Furthermore, Yang et al. [38] provided data showing that the mandibular first molar had a packing density of 7.74 (SD 0.236) in the middle third compared with 4.07 (SD 0.264) in the cervical third.

The simple extraction of the data, as presented in Table 4, already reveals an approximate trend towards a lower packing density of HSBs in the cervical region, a finding that is further explored through meta-analysis and TSA.

Risk of bias

The risk of bias was assessed according to the guidelines provided in the Checklist for Reporting In vitro Studies (CRIS) [14], which are recommended for evaluating in vitro dental studies. The results are presented in Table 5, where each column was assigned a score from one to five (with one = low quality and five = high quality). The questions to which the reviewers responded by assigning scores were as follows:

• Sample size calculation: "Is the sample size adequate to achieve statistically significant results?"

• Significant difference between groups: "Was the measurement of the 'significant difference' correctly set within the groups, considering the sample size and type of measurement?"

• Sample preparation and handling: "Does the study provide information on the production or handling of the samples to be tested?"

• Allocation sequence, randomisation, and blinding: "Did the samples have an equal and independent chance of being assigned to any group?"

• Statistical analysis: "Are the statistical methods described?"

Table 5 Risk of bias

Meta-analysis, sensitivity analysis, publication bias

The meta-analysis of the data was conducted using the Review Manager 5.4 software (Cochrane Collaboration, Copenhagen, Denmark), which was also used to generate the images of the forest plot and funnel plot.

The meta-analysis focused on the mean difference in the packing density (PD) of HSBs between the middle third and cervical third buccal areas in posterior teeth (molars). Fixed effects were applied, and the mean difference and standard deviation (SD) between the two groups were calculated. The aggregated MD value was 4.95 [4.69, 5.22], indicating a significantly higher packing density in the middle regions of the teeth. The black diamond in the forest plot, representing the effect size, clearly deviates from the no-effect line (Fig. 2). The studies included in the meta-analysis were: Lynch et al. [10], and Yang et al. [38]. For Lynch et al. [10], MD values were included from 4 data sets; for Yang et al. [38], MD values were included from 6 data sets. All individual MD values did not exceed the no-effect line within their confidence intervals (Fig. 2).

Fig. 2

Forest Plot of the Fixed-Effects Model of the Meta-analysis. MD = 4.95, 95% CI [4.96, 5.22]; df degrees of freedom; I² Higgins' heterogeneity index (I² < 50% indicates low heterogeneity; I² > 75% indicates high heterogeneity); C.I. confidence intervals; P p-value; SD standard deviation. The graph for each dataset shows the lead author, year of publication, segment measurement (teeth), the mean with SD, and the total number of segments examined, the mean difference with confidence intervals, and the weight of each study expressed as a percentage. The final value is shown in bold with the corresponding confidence intervals. The black line indicates the position of the average value, and the light black diamond represents the measure of the average effect. Specifically, this figure illustrates the individual and cumulative effect sizes for the mean difference in the packing density of Hunter-Schreger bands (HSB) between the middle and cervical regions of tooth enamel. Each dataset is represented by a square, the area of which reflects its relative weight (based on the inverse variance of its effect estimate), while the horizontal line extending from the square represents its 95% confidence interval (CI). The vertical dashed line indicates the null value (MD = 0), serving as a reference for no difference between the regions. The black diamond at the bottom of the plot represents the overall cumulative effect, which is a mean difference of 4.95, thereby demonstrating a statistically significant higher packing density in the middle region. Additionally, the plot provides information on the heterogeneity of the included datasets, quantified by Higgins’ I² statistic (I² 89%), which informs on the consistency of the effect sizes among the studies

Additionally, a sensitivity analysis was conducted by excluding the MD difference in the study by Yang et al. [38] that had the highest weight (44.4%), which had an excessive influence on the meta-analysis due to narrower confidence intervals compared to the other studies. This was likely a source of heterogeneity. Indeed, the inconsistency index (I²) decreased from 89 to 45% when excluding the data from Yang et al. [38] for the Mandibular first molar (Fig. 3).

Fig. 3

Sensitivity Analysis.Excluding the MD with the highest weight (44.4%) from the study by Yang et al. [4], which carried excessive weight in the analysis. The I² index decreased to 45%

Importantly, despite this reduction in heterogeneity, the pooled mean difference remained statistically significant. This indicates that although the Yang et al. 2022 [38] study was a major contributor to the observed heterogeneity, its influence did not drive the overall result. The sensitivity analysis, therefore, reinforces the robustness of our main finding: the packing density of Hunter-Schreger bands is significantly higher in the middle region of the tooth crown compared to the cervical region.

A publication bias assessment was also conducted through a visual analysis of the distribution of studies on the funnel plot, where asymmetry in the data distribution was observed (Fig. 3). Furthermore, no signs of heterogeneity were identified, as indicated by the overlapping confidence intervals (Fig. 3) and the distribution on the funnel plot (Fig. 4).

Fig. 4

Funnel Plot (Review Manager 5.4) SE(MD) = Standard Error (Mean Difference). Graphically, there are no sources of heterogeneity. The presence of partial symmetry suggests the potential absence of publication bias. In detail, this funnel plot illustrates the distribution of the included data sets by plotting the standard error of the mean difference (SE[MD]) against the effect size. Each point corresponds to a single data set. A symmetric distribution of points around the pooled effect estimate implies that publication bias is unlikely, while any asymmetry could suggest the presence of small study effects or other biases. This visual tool helps us assess the robustness and consistency of meta-analytic results

Sequential trial analysis, grade

A sequential trial analysis (TSA) was conducted to assess the strength of the results from the third meta-analysis and to adjust the findings to avoid type I and II errors [39]. The software used was the free Java software for TSA (software version: TSA 0.9.5.10 Beta). The O'Brien-Fleming spending function was applied using fixed effects.

The TSA results showed that, despite only 10 datasets being included, the mean difference was statistically significant. We applied a power of 99% (using a maximum type I error of 10% and a maximum type II error of 1%). In the TSA, it was clearly shown that the blue line (Z curve) crossed the sequential trial boundary (red inclined line) from the first dataset of the initial study and reached the optimal number of samples examined starting from the second dataset, which was determined to be 36 (Fig. 5).

Fig. 5

TSA; On the left, the inward-sloping red lines represent the boundaries of sequential trial monitoring. On the right, the outward-sloping red lines constitute the futility region. The continuous blue line is the cumulative Z-curve. Dark red line (Z = 1.98). In detail this TSA graph shows the cumulative Z-curve (blue line) as datasets are sequentially added, plotted against the accrued sample size. The red inclined lines represent the trial sequential monitoring boundaries, which adjust for the risk of type I and II errors in cumulative meta-analysis. The inward-sloping boundaries indicate the threshold for statistical significance, whereas the outward-sloping lines delineate the futility area. The graph demonstrates that the cumulative Z-curve crosses the significance boundary from the first dataset and reaches the required information size (RIS) of 36 from the second dataset onward, thereby confirming that the evidence is robust and sufficient to support the observed effect

The threshold of 36 was determined empirically, as set by the TSA 0.9.5.10 Beta software. Specifically, this threshold level of 36 is the result of calculating the RIS (Required Information Size), which indicates that until this cumulative number is reached, the meta‐analysis results may not be considered definitive. The TSA software utilises the entered data (group sizes, means, and standard deviations) and the parameters set (α = 0.10 and β = 0.01) to estimate the variance and to define the minimum relevant effect. A power of 99% (β = 0.01) was applied, meaning that the probability of detecting a true effect is extremely high (as evidenced by Table 4, the data extractions on Packing Density HSB, and the meta‐analysis in Fig. 2). Therefore, a maximum type I error of 10% (α = 0.10) and a maximum type II error of 1% (β = 0.01) were established, from which the software determined the threshold of 36. Although an α value of 10% is higher than the conventional 5%, the extremely high power contributes to reinforcing the reliability of the result.

The results of the TSA and the meta‐analysis provide sufficient evidence that the obtained data yield adequate statistical power, particularly in a context where the literature is limited by the presence of only two in vitro studies on extracted teeth.

The authors also used GRADE pro-GDT to assess the quality of the evidence. The results suggest that the quality of evidence is moderate Table 6.

Table 6 Evaluation of GRADE pro GDT: ⊕ ◯◯◯ Very low, ⊕  ⊕ ◯◯ Low, ⊕  ⊕  ⊕ ◯ Moderate, ⊕  ⊕  ⊕  ⊕ High, CI (confidence Interval)

Discussion

This systematic review aimed to critically evaluate the packing density of HSBs on the buccal surface of the tooth crown by calculating the mean aggregate difference between the middle third and the cervical third of the surface, and secondarily focusing on how variations in HSB density across different regions of posterior teeth may contribute to the onset and progression of lesions such as NCCLs. For the primary outcome, which aimed to identify different distributions of HSB density, both qualitative and quantitative assessments were possible through the use of meta-analysis, highlighting only indirect evidence of factors that may contribute to lesion onset. For the secondary outcome, which aimed to establish a relationship between the packing density of HSBs and NCCLs through clinical studies, no clinical studies were found in the literature, and therefore, a quantitative and qualitative assessment of the studies was not possible. As a result, we were unable to establish a clear correlation between HSB distribution and the onset of NCCLs [40].

Primary outcome: HSB packing density analysis

For the primary outcome of the review, two studies were included: Lynch et al. [10], and Yang et al. [38]. The main characteristics and results of these studies, as well as the extracted data, are presented in Table 4 and described as follows.

Yang et al. [38] study examined the configuration HSB in the lateral enamel of human molars, focusing on the differences between "functional" cusps (involved in crushing and grinding) and "guiding" cusps (involved in guiding the teeth into occlusion). His results show that functional cusps have a higher HSB packing density and thicker decussated enamel in the cuspal segment compared to guiding cusps. This suggests that functional cusps are better adapted to withstand masticatory stress, likely due to their greater enamel decussation, which helps prevent crack propagation. Moreover, mandibular molars exhibit weaker decussation than maxillary molars, and the enamel on guiding cusps is more prone to chipping than that on functional cusps due to their weaker decussation [38].

The study by Lynch et al. [10], explored HSB patterns in various types of human teeth (incisors, canines, premolars, molars) and found regional variations in HSB packing densities. The densest HSBs were located in areas where functional and occlusal loads are greatest, such as the occlusal surfaces of posterior teeth and the incisal regions of incisors and canines [10]. The findings support the idea that HSB packing density plays a crucial role in increasing enamel's fracture and wear resistance. Specifically, regions with higher HSB density are more resistant to stress, which is beneficial for clinical dental treatments, such as bonding adhesive restorations and understanding conditions like cracked tooth syndrome. However, the relationship between HSB density and the vulnerability of dental surfaces to NCCL is more complex [18].

Our systematic review, through the analysis of the literature in the included studies, extracted data on the packing density in cervical regions and observed that in these areas, where HSB density is lower, the enamel may be more susceptible to mechanical stresses and erosive effects, contributing to the formation of NCCL. HSBs appear to be more effective in protecting enamel from occlusal loading forces, but their absence in cervical regions may facilitate the propagation of microcracks due to erosion or abrasion, such as in cracked tooth syndrome [41].

The included studies clearly demonstrate that the packing density of HSBs is higher in areas most exposed to functional loads, such as the occlusal surfaces of molars and the incisal regions of incisors and canines, with differences observed between the vestibular surfaces of mandibular teeth compared to those of maxillary teeth. Conversely, the cervical regions of the teeth, where the load is lower, exhibit lower HSB packing densities. These findings are consistent with the data derived from the meta-analysis, which strongly indicate a highly significant mean difference in HSB packing density between the middle third and the cervical third of the enamel (Fig. 2, MD, 4.95 [4.96, 5.22]). The strength of these results is further confirmed by the TSA, which shows that the data from the first study are sufficient to establish certainty in the results, suggesting no need for additional studies with further data.

The clinical implications of this significant difference may provide additional explanations for why teeth subjected to abrasion (primarily brushing trauma) or occlusal forces show lower resistance in the cervical regions, leading to wedge-shaped lesions in NCCLs. These regions, characterized by a lower packing density, are closely correlated with the tooth's resistance to mechanical and physical forces.

Although the results of the present study indicated a significant difference in the packing density of HSBs between the cervical and middle regions of the tooth crown, it should be emphasized that the observed phenomena may be influenced by variables not yet thoroughly explored. For instance, while the lower HSB density in the cervical region seems to be correlated with increased susceptibility to erosion and abrasion, there is evidence suggesting that other factors, such as individual biological variables, patient behavior, or diet—particularly the ingestion of acidic foods—could play a crucial role in the formation of NCCLs. Further investigation is needed to explore how the individual variability of enamel characteristics affects resistance to mechanical and erosive forces.

Secondary outcome

The absence of studies on the packing density of HSBs in relation to the formation of NCCLs makes it impossible to perform a meta-analysis of the data for the secondary outcome. Therefore, only direct evidence emerges regarding how the different distribution of packing density may influence the presence of NCCLs. Data from the studies by Lynch et al. [10] and Yang et al. [38] suggest that areas with lower packing density of HSBs, such as the cervical region (Table 4, range 1.5–5 µm² for the cervical third of the buccal surface of posterior tooth crowns, compared to the range 6–11.4 µm² for the middle third), may be more susceptible to mechanical stresses and damage from wear or masticatory occlusal trauma, and thus compatible with abrasive phenomena (brushing trauma), supporting the formation of NCCLs.

Our results indicate that the packing density of HSB is higher in regions subject to greater functional demands, such as the occlusal surfaces of posterior teeth involved in mastication. The lower HSB packing density observed in the cervical areas of the enamel can be explained by the fact that these regions, which feature thinner enamel, are not exposed to occlusal loads. Consequently, this variation in packing density may lead to different patterns of progression and development of clinical pathological conditions that affect the enamel initially and the dentine subsequently, including abrasive phenomena, abfraction, enamel fractures, and the adhesion of restorative adhesive systems to enamel.

Indeed, the reduced HSB packing density in the cervical region may render this area more vulnerable to erosive, abrasive and occlusal stresses, thereby predisposing it to the development of NCCL.

From a clinical perspective, this insight underscores the need for targeted preventive measures. Dentists might consider implementing personalised oral hygiene strategies, such as recommending less abrasive toothpastes and modified brushing techniques, to mitigate the risk of cervical enamel degradation. Furthermore, these findings, together with the higher failure rate of NCCL restorations compared with other clinical conditions—as evidenced by retrospective studies [42]—may influence both the selection and choice of composite and adhesive materials in the design of restorative treatments. For instance, in cases where NCCLs are present, the use of adhesive materials that can effectively bond to less robust enamel may improve marginal adaptation as well as the longevity of restorations [42].

In light of the findings from this review, our evidence strongly recommends the need for further studies to provide additional data and substantiate the observed trends. While this review presents valuable insights into the packing density of HSBs and their potential role in the development of NCCLs, the current literature lacks conclusive evidence linking these factors directly. Therefore, it is essential that future experimental and clinical studies, particularly longitudinal investigations, are conducted to deepen our understanding and validate these findings.

Limitations of the review

The limitations are primarily due to the absence of trials investigating the onset of NCCLs in relation to packing density, which prevents the execution of a quantitative evaluation of the data for the secondary outcome. The current body of knowledge only provides indirect evidence regarding how HSBs might influence the onset of NCCLs.

Despite the progress, our understanding of the relationship between HSBs and NCCLs remains incomplete. In vitro data are useful, but they do not always accurately reflect the physiological conditions of a live tooth, which is exposed to cycles of occlusion, erosion, and abrasion.

Indeed, the inclusion of only two in vitro studies may limit the generalisability of our findings. This limited evidence base raises concerns that the observed differences in HSB packing density may not be reliably extrapolated to a broader population or clinical settings. Moreover, reliance on in vitro data means that factors present in vivo—such as variations in the oral environment and patient-specific characteristics—are not fully taken into account.

The asymmetric appearance of the funnel plot suggests that there may be a certain degree of publication bias, potentially due to the preferential publication of studies reporting significant results. However, it is important to note that the small number of studies included (only two), with merely 10 datasets incorporated in the meta-analysis, renders the funnel plot inherently unstable and its interpretation problematic. With such a limited dataset, even modest variations can give rise to apparent asymmetry. Therefore, while the observed asymmetry raises concerns, we caution that it may not definitively indicate the presence or absence of publication bias.

The impact of sample preparation on HSB density measurement is a critical factor that warrants thorough discussion. Both Lynch et al. [10] and Yang et al. [38] provide detailed descriptions of their sample preparation protocols, which underscore how methodological differences can influence HSB observations. For example, in Lynch et al. [10], teeth were sectioned using a low‐speed diamond saw, then cleaned in Histolene (CellPath Ltd, Powys, Wales, UK) for three hours and carefully dried with Velin tissue (Koch-Light Laboratories Ltd, Buckinghamshire, UK) prior to imaging under reflected light. Such procedures, including the cleaning and drying steps, may affect the contrast and integrity of the Hunter-Schreger bands, potentially leading to variations in the measured packing density. Similarly, Yang et al. [38], after sectioning the teeth using a Buehler Isomet diamond blade and subsequently polishing and lapping them, employed a combination of scanning electron microscopy (SEM) and digital stereomicroscopy, taking great care to ensure that the sectioning plane was parallel to the focal plane. Variations in these techniques—such as the angle of sectioning, the type of polishing or cleaning agents used, and even the imaging conditions—can introduce artefacts or alter the enamel microstructure, thereby affecting the quantification of HSB density.

Given these considerations, it is essential that sample preparation protocols are standardised across studies to minimise potential artefacts and ensure that HSB density measurements are both accurate and reproducible. Future research should aim to systematically investigate the extent to which different preparation methods affect HSB observations, thereby strengthening the validity of in vitro assessments and facilitating more reliable comparisons between studies.

Future studies should focus on more comprehensive clinical models, possibly using advanced imaging techniques that allow for in vivo observation of HSBs. Additionally, integrating longitudinal studies that monitor the evolution of NCCLs over time could provide a clearer understanding of the dynamics of lesion development. Finally, it would be important to investigate how restorative treatments or the management of occlusal loads may influence the evolution of NCCLs, and whether there are methods to strengthen or preserve the density of HSBs throughout the life of the tooth.

Conclusions

In conclusion the results of the meta-analysis conducted in this study highlight a significant difference in the packing density of HSBs between the cervical and middle regions of the tooth crown.

These findings suggest that the packing density of HSBs may play an important role in enamel resistance to mechanical forces, influencing the susceptibility of dental surfaces to the onset of NCCLs. In particular, the lower HSB density in the cervical regions may explain why these areas are more exposed to erosive and abrasive stresses, which are key factors in the development of non-carious lesions.

Consequently, these data can guide restorative treatment planning and the development of preventive strategies for NCCLs by informing the selection of specific adhesive composite materials or etching techniques designed to bond effectively to enamel with a lower HSB packing density, as well as by supporting preventive measures such as the use of low-abrasivity toothpastes and the adoption of gentle brushing techniques to minimise enamel wear.

Our evidence strongly recommends the need for further studies to collect additional data and provide a more robust understanding of the relationship between HSB packing density and NCCL development.

Imaging techniques must also be refined and standardised, for instance by utilising adjunct methods such as Optical Coherence Tomography or micro-CT, which can aid in evaluating the clinical forms of NCCL. Furthermore, in vivo assessments have also been carried out.

The results of the meta-analysis should be considered as a first step toward a deeper understanding of the relationship between HSB density and NCCLs. Future research should focus on the analysis of clinical and longitudinal data to confirm these findings and further investigate the role of HSBs in preventing the development of NCCLs.