Abstract
Introduction: Root caries (RC) remains a global health problem leading to negative impacts on an elderly person’s well-being causing oral health-related quality of life issues, such as inadequate nutrition and detrimental oral functionality. The present systematic review with meta-analysis is designed to synthesize existing research findings on the prevalence and experience of root caries globally over the past 30 years. It aims to describe its distribution by country and explore its links with various socioeconomic indicators. Methods: Selection criteria: Three electronic databases (PubMed, Embase, and Scopus) were screened for observational epidemiological studies (cross-sectional and cohort studies) reporting the prevalence of RC and/or mean RC experience between 1990 and 2023. No languages were applied. Study selection, data extraction, and quality assessment were done in duplicate. Risk of bias was graded with customized quality assessment tools (Study Quality Assessment Tools NHLBI, NIH). Data collection and analysis: studies reporting on (1) root-caries experience (mean and SD) and (2) root-caries prevalence (%) were used to synthesize the results. It was assessed as decayed, missing, and filled teeth (RDMFT). Results: A total of 91 publications were included in the descriptive analysis; the estimated overall pooled mean RC was 2.87 teeth and the global estimated random-effects pooled RC prevalence was 41%. Low gross national income (GNI) countries reported a low mean number of RC (1.35 GNI <USD 5,000), while high GNI countries reported a higher mean number of RC (3.45 GNI USD 10,000–USD 19,999). Countries of higher inequalities (Gini index) reported lower means medium RC (1.98 teeth) than countries with no inequalities (4.90 teeth). Conclusion: This study highlights the high caries burden among adult population globally by estimating overall trends and comparing against factors including area, GNI, and Gini index. The large magnitude of these inequities indicates that oral health equity can only be achieved taking into account socioeconomic factors on a global scale. The lack of uniform data collecting among studies as well as knowledge gap regarding the incidence and experience of RC in different countries.
Introduction
In many countries throughout the world, both the number of elderly people and life expectancy have grown dramatically in recent decades [1]. Root caries (RC) is one of the most common dental public health concerns associated with aging [2]. An increase in the number of retained teeth raises the risk of periodontal disease, which eventually leads to gingival recession and RC [3]. Although RC is avoidable, it is one of the main sources of reduced oral health-related quality of life, which may lead to tooth loss and may have a negative impact on adults' quality of life in terms of oral health [4]. RC remains a global health problem leading to negative impacts on a person’s well-being as well as causing oral health-related quality of life issues, such as inadequate nutrition and detrimental oral functionality [5]. RC is an age-dependent disease [3]. An increase of RC prevalence was reported by many authors [6‒8], as the average lifespan of the population is increasing and more people retain their natural teeth throughout their lives. In numerous surveys, it is reported that nearly half of the examined subjects are affected by RC [6‒11]. Several factors are known to be associated with RC, such as being older individuals with a lower socioeconomic status (SES), tobacco use, as well as gingival recession and poorer oral hygiene [12, 13]. Three peaks in (root)-caries activity in a 70-year period were described [3]. The most common risk factors were identified as the plaque index (biofilm), the number of retained teeth, the previous experience of RC and recession and closeness to dentures [2, 14]. SES significantly affects a nation's health status and it is influenced by factors such as per capita income, educational attainment, diet, and lifestyle. This can vary widely between countries and may impact the prevalence of RC [10, 15‒18]. However, few studies evaluated the experience and prevalence of RC in countries at different levels of economic and human development [19, 20]. Furthermore, only fragmentary knowledge exists regarding the correlation between factors such as the economic inequality index (Gini coefficient) and/or gross national income per capita/year (GNI) of various countries and RC [14, 21, 22].
As over past 30 years dental status of the elderly changes dramatically as they now retain more teeth due to better oral health care in more and more countries worldwide [6‒8], the present meta-analysis is designed to synthesize existing research findings on the prevalence and experience of RC globally over the past 30 years. It aims to describe its distribution by country and explore its links with various socioeconomic indicators. This study seeks to fill this gap in the scientific literature by providing a comprehensive and updated overview of the global RC situation and highlighting the role that socioeconomic and geographical factors may play in determining this phenomenon.
Materials and Methods
This systematic review was conducted following the Preferred Reporting Items for Systematic reviews and Meta-analysis (PRISMA 2020, online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000542783) guidelines (Page et al. 2021). Protocol was registered at the international Prospective Register of Systematic Reviews (PROSPERO; registration: CRD-42022290418).
PECO Question and Eligibility Criteria
The research question was taking into account following the PECO structure:
P = General population older than 45 years old who live independently or with the help of other in their own homes.
E = RC prevalence and experience.
C = Socioeconomic indicators of the countries where the subjects lived (GNI, Gini index, geographic area, life expectancy).
O = Prevalence of RC (root decayed, mean, filled surface [RDMFS] ≠ 0); mean root caries experience values (RDMFS).
The eligibility criteria were observational epidemiological studies (cross-sectional and cohort studies) reporting the prevalence of RC and/or mean RC experience between 1990 and 2023. In line with this, intervention studies (both randomized and nonrandomized) were considered only if they had a baseline inclusion of participants with an epidemiological approach and the data were gathered only from baseline reports. Moreover, studies reporting the prevalence of RC of a specific population (institutionalized patients, hospitalized subjects, and those, living in specialized institutions [congenital anomalies, dento-facial anomalies]), studies based on self-reported case definitions of RC, and preclinical studies were excluded.
Data Sources and Strategy
Detailed search strategies and search strings were appropriately created using Boolean terms. The search strategy was initially developed for PubMed using keywords and MeSH terms and adapted to the other databases. Electronic databases such as PubMed, Embase, Scopus, and Open Grey literature (http://www.opengrey.eu) were searched (January 11, 2023). The search strings used in the search strategy were defined for each electronic base (online suppl. Table 2). Cross-referencing was also performed using the references lists of full-text papers and gray literature was retrieved via opengrey.eu (http://www.opengrey.eu).
Study Selection
Three authors (A.M., M.E.-O., and G.C.) independently reviewed titles and abstracts, excluding those not meeting the inclusion criteria. The reviewers were not blinded to the identity of the journal names, article authors, institutions, or the results of the research. The full texts of the selected papers were then assessed by 10 authors (A.M., M.E.-O., G.C., R.A.G., M.K.R., C.S., A.R., R.J.W., R.M., and R.B.-B.) and agreement concerning study inclusion was made by discussion between the authors. Endnote software was used for study selection.
Data Extraction
Two authors performed data extraction independently and in duplicate (A.M. and M.E.-O.), and a third author decided disagreements (G.C.). Extracted data were saved in an Excel file. The following data were collected in pilot-tested xcel files: author/title/year of study, study affiliation, study type and setting, design of the study, number/age/gender of patients as well as RC prevalence (percentage) and experience (RDMFS index). Studies without sufficient data for meta-analyses (e.g., reporting relative indices (ΔDMRS/T, changes in surface textures, and RCI) instead of absolute indices (DMFRS/T, surface texture) were kept in the systematic review and findings reported in the descriptive analysis.
Quality Assessment
At all phases, the reviewers were trained by G.C. and R.J.W., and a pilot test was conducted to audit the eligibility criteria. In phase 1 (title and abstract reading), the studies were assessed for the eligibility by two independent reviewers (A.M. and M.E.-O.). In phase 2 (full-text reading), papers included after title and abstract reading. Disagreements were solved in a consensus meeting, and if any disagreement persisted, another reviewer (R.G.) was involved to steer the decision. After selecting full-text papers, they were distributed equally between the authors (A.M., M.E.-O., G.C., R.A.G., M.K.R., C.S., A.R., R.J.W., R.M., and R.B.-B.) and the data were independently in a self-designed excel spreadsheet (Microsoft Excel®). Data which include author name(s), year of publication, country of the study, sampling method, sample size, dental or RC prevalence, and dental or RC experience in terms of RDMFT or RDMFS, mean number of decayed teeth (DT) and diagnostic criteria were extracted on a predefined spreadsheet. Data were recorded separately for RC and grouped according to the continents. Only observational cohort and cross-sectional studies have been included. Randomized/nonrandomized studies have not been accepted.
Reviewers independently assessed the quality of the included studies using a customized quality assessment tool developed by The National Heart, Lung and Blood Institute for Observational Cohort and Cross-sectional studies, Case-Control studies, and Controlled-Intervention studies (Study Quality Assessment Tools NHLBI, NIH) (https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tool). It consisted of 14 questions with a yes, no, and other (CD, cannot determine; NA, not applicable; NR, not reported). The authors allocated the papers into three overall risk of bias categories: Positive – 0–6 answers poor quality, 7–10 fair quality, and 11–14 good quality.
Risk of Bias Assessment
The risk of bias assessment was carried out by three reviewers independently (M.E.-O., A.M., and G.C.), using a customized quality assessment tool developed by The National Heart, Lung and Blood Institute for Observational Cohort and Cross-sectional studies, Case-Control studies, and Controlled-Intervention studies (Study Quality Assessment Tools NHLBI, NIH) (https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tool). Authors answered to signaling questions in each domain and then estimated the overall risk of bias as: 0–6 poor quality, 7–10 fair quality, and 11–14 good quality. The assessment was carried out by all authors (A.M., M.E.-O., G.C., R.A.G., M.R., C.S., A.R., R.J.W., R.M., and R.B.-B.). For this, the authors were trained in one session, and afterward, the inter-reviewer assessment was carried out (online suppl. Table 4).
The Grading of Recommendation, Assessment, Development, and Evaluation (GRADE) tool was used to evaluate the certainty of the evidence, using the GRADE-PRO website (https://www.gradepro.org/, accessed on 28 October 2023) (online suppl. Table 5). In the absence of a formal procedure, the framework developed for incidence estimates in the context of prognostic studies was adopted. The largest body of evidence for a prevalence meta-analysis comes from cross-sectional studies or baseline assessments in cohort studies. Therefore, the assessment of evidence from these types of studies starts with “high certainty of evidence” and is downgraded according to the risk of bias, inconsistency, indirectness, imprecision, and publication bias. Finally, the level of certainty of the selected aspects of the evidence can be rated as high, moderate, low, or extremely low. The assessment was carried out independently by two authors (A.M. and M.E.-O.). In cases of disagreement, a third reviewer was involved (G.C.).
Data Synthesis
As the studies reported on (1) RC experience (mean and SD) and (2) on RC prevalence (%), two separate meta-analyses were conducted. In the first study, studies reporting on RC experience, i.e., the mean number of teeth with RC or filled roots, and the corresponding standard deviation, were included. In the second study, studies that reported on RC prevalence, i.e., the proportion of individuals with RC, were included. Studies that did not report sample size, standard deviation, prevalence, or experience of RC were excluded. Where prevalence for 2 or more subgroups was reported, the overall prevalence or combined mean was used in the meta-analysis. Forest plots were constructed to display the results of each meta-analysis.
Countries were divided in 5 different regions: Africa, North America, Central/South America, Asia-Oceania, and Europe. The Gross National Income (GNI per capita), life expectancy, and Gini index were used as measures of the SES of a country. The GNI was reported in US dollars as per June 13, 2023, according to the year of publication of the study included in the meta-analysis (https://www.worldbank.org/en/home), and divided into five categories: <USD 5,000, USD 5,000–9,999, USD 10,000–19,999, USD 20,000–39,999, and >USD 40,000. Life expectancy was divided into three categories: 70–75 years, 76–80 years, and >80 years; the Gini index was divided into four categories: <32, 32.1–35, 35.1–40, and >40.
Heterogeneity was estimated using the I2 statistic as the percentage of variation across studies that are due to heterogeneity and not due to chance, as well as 95% prediction interval. I2 value >50% was considered indicative of substantial heterogeneity, and a random-effects model with a 95% confidence interval was chosen for the meta-analysis. To investigate heterogeneity meta-regression using the DerSimonian-Laird estimator with year of publication, regions of the world, sample size, and quality of the included studies, where the studies were subdivided into two groups; good quality and fair/poor quality, were carried out (online suppl. Fig. 2a–d). To test the robustness of the results, sensitivity analyses were carried out. The studies were divided into 2 subgroups: RC prevalence or RC mean from studies conducted on a national level and those conducted on a regional/local level. Sensitivity analyses were further performed by excluding outlier data for RC prevalence [23‒26] and RC experience [16, 27‒30]. The leave-out-one meta-analysis technique was employed to detect changes in the overall pooled mean RC and pooled RC prevalence should one study be excluded from the meta-analysis [31]. Once these studies were identified, meta-analyses were performed with the studies excluded (online suppl. Fig. 1). Statistical analysis was performed using StataSE18®. Two funnel plots were constructed to assess for publication bias for studies conducted on caries experience and caries prevalence (online suppl. Fig. 3a, b).
Results
Studies Selection and Characteristics of the Included Studies
Overall, 794 papers were retrieved, and then 103 duplicates were excluded and 691 were selected. After title and abstract evaluation, 208 reports assessed for eligibility and 483 were excluded. After full-text reading further 118 were excluded (online suppl. Table 3). Remaining 91 were obtained in the full-text format (Fig. 1).
PRISMA 2020 flowchart of the search strategy and identification process of the papers.
PRISMA 2020 flowchart of the search strategy and identification process of the papers.
According to the year of publication, 26 studies were published between 1990 and 2000, 33 studies from 2001 to 2010, 30 from 2011 to 2020, and 2 from 2021 to 2023. One paper reported data from 3 different countries [32]; thus, data were entered separately for each country reported in the study.
Quality Assessment
Most of the studies (n = 57) were classified as fair quality, while 27 papers and 7 studies were ranked as high and poor quality, respectively. Quality assessment relied on sample size justification or power analysis and lacking those criteria affected the quality outcome of the papers. However, in almost all publications, research questions, study populations, and outcome measures were clearly defined (online suppl. Table 4).
Synthesis of Results
From the included papers (n = 91), RC prevalence was reported in 66 of them. One paper reported data from 3 different countries [32] (online suppl. Table 6). Worldwide highest RC prevalence was reported in Mexico (96.5%) [23], while the lowest was in Denmark (4%) [33]. The highest RC prevalence for North America was observed in Canada (19.7%) [34], for South America in Mexico (96.5%) [23], for Europe in Sweden (54%) [35], and for Asia in Sri Lanka (89.7%) [24] (Fig. 2).
Worldwide RC prevalence map. Prevalence was color coded as follows: (a) <29% – low prevalence (dark green), (b) 29.1–39% – medium-low prevalence (light green), (c) 40–49.9% – medium prevalence (yellow), (d) 50–60% – medium-high prevalence (orange) and 60–80% – high prevalence (light red), and (e) >80% – very high prevalence (dark red).
Worldwide RC prevalence map. Prevalence was color coded as follows: (a) <29% – low prevalence (dark green), (b) 29.1–39% – medium-low prevalence (light green), (c) 40–49.9% – medium prevalence (yellow), (d) 50–60% – medium-high prevalence (orange) and 60–80% – high prevalence (light red), and (e) >80% – very high prevalence (dark red).
The highest RDMFS mean was reported in Sweden (14.6 ± 13.8) [16] and the lowest in Australia (0.89 ± NR) [15]. The highest for North America was reported in the USA (6.5 ± 7.6) [27], for South America in Colombia (2.03 ± 2.78), and for Asia in Sri Lanka (mean 3.8 ± 3.6) [24].
The global estimated random-effects pooled RC prevalence was 41% (95% CI = 0.36, 0.47), 37% (95% CI = 0.28, 0.47) (North America), 39% (95% CI = 0.30, 0.48) (Europe), 41% (95% CI = 0.23, 0.60) (Central/South America), and 47% (95% CI = 0.36, 0.57) (Asia-Oceania) (Fig. 3). When stratified by GNI, the random-effects pooled prevalence ranged from 37% (GNI USD 5,000–9,999) to 43% (GNI <USD 5,000, GNI USD 10,000–19,999) (Fig. 4); Gini index 37% (high inequality) to 49% (low inequality) (Fig. 5) and life expectancy 38% (life expectancy >80 years) to 42% (life expectancy 76–80 years). Heterogeneity between the studies was high, with an I2 of 99.7%, 95% prediction interval of (0.01, 0.88) (Fig. 6). Sensitivity analysis did not change the overall pooled RC prevalence.
Forest plot of the pooled RC prevalence stratified by geographical area.
Forest plot of the pooled RC prevalence stratified by the Gini index.
Forest plot of the pooled RC prevalence stratified by life expectancy.
The estimated overall pooled mean RC was 2.87 teeth (95% CI = 1.85, 3.88); 1.87 teeth (95% CI = 1.05, 2.68) (Asia-Oceania), 3.0 teeth (95% CI = 1.35, 4.64) (North America), and 4.36 teeth (95% CI = 1.63, 7.10) (Europe) (Fig. 7) [11, 14, 36‒65]. Areas with low GNI reported an overall low mean number of RC (1.35; GNI <USD 5,000), while areas with higher GNI reported a higher mean number of root carious teeth (3.45; GNI USD 10,000–19,999) (Fig. 8) [11, 14, 24, 38‒55, 66‒70]. When stratified by the Gini index countries of higher inequalities reported lower means (medium inequality = 1.98 teeth) than countries with no inequalities (4.90 teeth) (Fig. 9) [11, 14, 24, 37, 39, 41‒55, 71‒74]. The oldest age group reported the lowest overall mean RC 1.89 (life expectancy >80 years) (Fig. 10) [11, 14, 18, 24, 37, 39‒55, 75‒84]. Heterogeneity between studies was high: I2 statistic 99% and 95% prediction interval of 2.08–7.81. The year of publication explained 44% of the heterogeneity reported in the meta-analysis. The variables area, sample size, and quality of the included studies did not explain any of the heterogeneity. The overall pooled mean RC was lower after excluding outlier data (1.81) and for studies conducted on a national level (2.00) (online suppl. Fig. 1). The funnel plots are asymmetrical, suggesting publication bias (online suppl. Fig. 3a,b). However, the assumption of positive results being more often published is not necessarily true for proportional studies, since there is no clear definition or consensus about what a positive result in a meta-analysis of proportion is [85].
Forest plot of the pooled RC experience stratified by geographical area.
Forest plot of the pooled RC experience stratified by life expectancy.
Discussion
Almost half of the senior population worldwide suffers with RC. However, as most teeth were lost before RC lesions formed, research on caries preventive techniques has primarily focused on coronal caries [19]. Longer tooth retention in the elderly population has been documented recently, leading some to speculate that RC experience is likewise trending upward [86, 87]. Thus, an investigation of RC prevalence will be useful in assessing the reach and availability of dental disease preventive and treatment programs, as well as in improving the quality of life for the elderly population.
In light of this, the objective of this systematic review was to locate and examine global data on the prevalence and experience of RC. This is the first officially recognized systematic review that includes a meta-analysis of the combined prevalence and experience of RC that has been assessed using data from the previous 30 years.
This review’s overall RC estimates are in accordance with recent estimates [8, 10]. Our result also indicates that RC prevalence is higher in more economically advanced countries. This can be explained by the fact that in more economically advanced countries, life expectancy is longer as well as people tend to retain more teeth in their old age due to better health care and education level. A wide variation between countries is generally grouped within similar economic levels. The fact that certain studies are carried out on subgroups of the population may help to explain this. It appears that the distribution of RC is not uniform due to differences in the pooled prevalence of RC between continents and even within particular countries.
The outcomes showed that RC was still widespread globally in adults. Most of the included studies reported a prevalence of RC of more than 50%. The prevalence ranged widely between continents with the highest prevalence reported in South America (96.5%) [23] and the lowest in Europe (4%) [33].
The estimated overall pooled mean RC was 2.87 teeth per older adult globally and the global estimated random-effects pooled RC prevalence was 41%. Low GNI countries reported a low mean number of RC (1.35 GNI <USD 5,000), while high GNI countries reported a higher mean number of RC (3.45 GNI USD 10,000–19,999). Countries of higher inequalities (Gini index) reported lower means medium RC (1.98 teeth) than countries with no inequalities (4.90 teeth). This study highlights the high caries burden among adult population globally by estimating overall trends and comparing against factors including area GNI and inequality index (Gini index). The large magnitude of these inequities indicates that oral health equity can only be achieved taking into account socioeconomic factors on a global scale. A large variation of RC prevalence was reported within some countries such as USA, Brazil, and Canada [75, 86, 88‒94]. These differences indicate high variation in methods and area of data collection. Considering the variation and heterogeneity among the included studies, the presented estimates must be interpreted with caution. High heterogeneity among the studies could be due to the geographical variations or different criteria used to assess the prevalence of RC. Altogether, it can be concluded that the prevalence of RC was high, with an uneven distribution in geographical areas. However, the status of RC in older adults has not been widely investigated. Few studies reported the RC status of older adults, and sometimes, only the prevalence was reported [95]. For example, it was not possible to report the pooled prevalence in the African continent as no studies were retrieved through the search. Less than one-third of our included studies reported RC status in older adults and was limited to prevalence only. More studies are needed to provide information of RC status in older adults with details such as the number of sites and affected surfaces so that effective prevention and treatment strategies can be planned in the preparation of the increasing number of untreated RC in the foreseeable future.
This study has several strengths. First, the studies allowed identifying socioeconomic risk indicators for RC. Second, combined data allow grasping the changes to the prevention programs which can be made by policy makers on the national level taking into consideration population-level risk factors.
There are, however, several limitations, such as lack of data on RC prevalence and experience for several countries as well as nonconcordant data acquisition among studies. More than two-thirds of the studies have not reported the prevalence figures and/or caries experience. While many publications were available from the North American, no data were available from the African continent. It is also important to bear in mind that RC data were not available for all the countries, which clearly influences the overall estimation for the prevalence worldwide. The included studies showed that a wide variation in RC prevalence exists across geographical areas and remains high in all of them. Furthermore, as the data in the studies were collected in regional or local populations, the results might poorly represent the overall situation on a country level.
The systematic review found high heterogeneity among studies, which limits the current results and highlights the need for standardized RC assessment protocols, including activity assessment. This heterogeneity, partly due to variations in study locations and times, could also stem from the nature of proportional data [85]. Sensitivity analysis showed that while RC prevalence was unaffected, the pooled mean RC decreased when outliers were excluded and when studies were national. Therefore, the present results should be interpreted with caution. To the best of our knowledge, this is the only current systematic review on RC prevalence and experience focusing on social determinants in dental caries and offering empirical evidence of social inequality in oral health across national boundaries and oral health care systems. It highlights the need for more standardized research to monitor individual oral health, taking into account cultural and socioeconomic differences and to identify priority oral health needs of these populations [41].
In the present manuscript, studies involving institutionalized patients, hospitalized individuals, and those residing in specialized institutions (such as those with congenital or dento-facial anomalies) were excluded. This exclusion was based on the distinct oral health conditions and varying abilities to perform oral hygiene independently within these groups. Ultimately, only three studies met these exclusion criteria. Consequently, while the analysis focuses on individuals capable of performing their own oral hygiene, this approach does introduce a limitation to the generalizability of the findings.
The etiology of diseases is likely to be a complicated web of behavioral and environmental factors influenced by broader socioeconomic factors. The majority of research on sociobehavioral risk factors for dental caries has been done in developed nations, although, in recent years, findings from low- and middle-income nations have also been released. International collaborative studies of the World Health Organization and other international studies using the same technique, focusing on social determinants in dental caries, offer empirical evidence of social inequality in oral health across national boundaries and oral health care systems [96]. The issues facing dental public health practice are emphasized in the study, with a focus on the significance of risk assessment in determining the possibility for prevention. Therefore, systematic risk factor evaluation may play a key role in the design and monitoring of oral health promotion and oral disease prevention programs in the future. In summary, the present paper evaluates for the first time the associations between global RC prevalence and experience and socioeconomic indicators (geographical area, GNI, inequality distribution). The obtained results outline the scientific rationale to monitor and address the existing oral health problems, but their connection to the individual socioeconomic challenges within each country or even region. Consequently, it is of a paramount importance to detect potential links between socioeconomic indicators and successful implementation of oral health problems exist in order to better address weaknesses of currently implemented programs. The challenges of persisting high RC prevalence may be resolved through addressing the needs of the caregivers within community, since the success of RC prevention programs depends on the caregivers who are responsible for safeguarding the oral health of elderly. This will help to redirect the focus of research strategies in the direction of personalized and society-based approach in order to create more effective preventive programs to reduce RC prevalence and experience worldwide.
Acknowledgments
The authors would like to thank Bernadette Rawyler and Ines Badertscher from the Graphics Department of School of Dental Medicine, University of Bern, Switzerland for their help to create the figures.
Statement of Ethics
An ethics statement is not applicable because this study is based exclusively on the published literature.
Conflict of Interest Statement
All other authors declare no competing interests.
Funding Sources
This study was not supported by any sponsor or funder. This study received no funding.
Author Contributions
Anastasia Maklennan and Roberta Borg-Bartolo: contributed to design, data acquisition, analysis, and interpretation, and drafted and critically revised the manuscript; Andrea Roccuzzo, Claudia Salerno, Maria Katharina Raabe, Riccardo Monterubbianesi, and Richard Johannes Wierichs: contributed to design, data acquisition, analysis, and interpretation and critically revised the manuscript; Marcela Esteves-Oliveira: contributed to conception and design, data acquisition, analysis, and interpretation, and drafted and critically revised the manuscript; and Rodrigo A. Giacaman and G. Campus: contributed to conception and design, performed all statistical analyses, data acquisition, and interpretation, and drafted and critically revised the manuscript. All authors gave their final approval and agree to be accountable for all aspects of the work.
Data Availability Statement
All data generated or analyzed during this study are included in this article and its supplementary material. Further inquiries can be directed to the corresponding author.