Age and sex disparities in Latin-American adults with gliomas: a systematic review and meta-analysis

This study aimed to identify if there are ethnic differences in the age and sex distribution of gliomas in the Latino adult population. A systematic review and meta-analysis were conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 recommendations. Databases used were MEDLINE, LILACS, Web of Science, and Scopus. Studies were included if they reported the age and/or sex distribution of gliomas in Latin adults, published in English or Spanish from January 1st, 1985, to December 1st, 2022. The quality of the studies was assessed using the Newcastle—Ottawa Quality Assessment Scale and the NIH Quality Assessment Tool. From 1096 articles, fifteen studies with information on 6,815 patients were selected for the systematic review, and thirteen were selected for the meta-analysis. The mean ages of diagnosis of glioma and glioblastoma were 50.9, 95\%\ CI [47.8–53.9] years and 53.33 years, 95 \% CI [51–55.6], respectively. The male-to-female incidence rate ratio of gliomas was 1.39. Our study found mean ages of glioma and glioblastoma were 6 and 10 years lower than those reported in the CBTRUS. Our study suggests disparities in the age and sex distribution of gliomas in Latin America compared to other regions. CRD42021274423.


Introduction
Gliomas are the most common malignant primary central nervous system (CNS) tumors in adults, accounting for approximately 27% of all brain tumors and 81% of malignant tumors [1].Although they represent less than 1% of all incident cancer cases, they carry considerable morbidity and mortality.Gliomas are classified according to their histologic and molecular features, glioblastoma (GB) being the most common and aggressive subtype [1].
Less aggressive gliomas, like pilocytic astrocytoma, are more common in children [2,3], while more aggressive gliomas, like glioblastomas, are more common in adults [4].In addition, the incidence of glioblastoma increases significantly with age.The reported mean age at presentation varies by ethnicity, with non-Hispanic whites having a higher mean age than Hispanics [4,5].Furthermore, sex differences have been well-established in many brain tumors, including glioblastoma [6].
Epidemiological studies of gliomas in the United States, Canada, Australia, and Western Europe rely on national or regional registries.However, just a few Latin American countries, like Uruguay and Costa Rica, have national registries.Some countries have population-based cancer registries (PBCR), such as Colombia, Argentina, and Brazil.Most Latin American countries do not meet the criteria for high-quality registries, and therefore the frequency of these tumors is determined from series obtained from second or third-level care reference centers [7].Notably, most of these registries classify the tumor type according to the International Disease Classification without considering the histology.Consequently, the information regarding the epidemiology of gliomas is predominantly based on the Hispanic population from developed countries.Therefore, the Hispanic population is frequently assessed as a single ethnic group, even though it is highly diverse [8].Thus, the age and sex distribution of these tumors in Latin American countries is scantly known, and there is a lack of reliable and updated systematic reviews on this matter.Understanding the epidemiology of gliomas can help identify at-risk populations.This study aimed to determine if there are ethnic differences in the age and sex distribution of gliomas in the Latino adult population.

Methods
This systematic review and meta-analysis were conducted in compliance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines and recommendations and were registered in the PROS-PERO database (CRD42021274423) [9].

Eligibility criteria
Studies were included in the systematic review if they reported the age and/or sex distribution of gliomas in Latino adults, published in English or Spanish.Studies published from January 1985 to December 2022 were included.Before this date, Magnetic Resonance Imaging (MRI), an essential tool in diagnosing brain tumors, was not commonly used [10].Publications were excluded if they comprised abstracts only, preclinical studies, case reports, less than 30 participants, randomized clinical trials, reviews, systematic reviews, meta-analyses, letters, editorials, studies with more than 10% of pediatric patients, and studies done in non-Latin American countries.The pediatric population was defined as those subjects under 18 years of age.Grey and non-electronic literature were not included.In addition, studies were included in the meta-analysis if they reported a mean or median age at diagnosis with standard deviation and sample size (or when these could be calculated).

Information sources
Studies were identified by searching MEDLINE´s electronic databases through PubMed, Latin American and Caribbean Center on Health Sciences Information (LILACS), Web of Science, and Scopus.

Selection and data collection process
Titles and abstracts were screened independently and standardized by two reviewers (A.S.P and J.L.O.).Then, these reviewers independently screened the full text of the selected articles and individually retrieved the data.If necessary, a third party resolved disagreements (R.V.H.).A data extraction table was developed with the following information: first author's name, publication date, country, sample size, study design, diagnostic criteria used, tumor type, mean age, age range, number of male participants, and number of female participants.

Study risk of bias assessment
To assess the quality of the included studies, two reviewers (R.V.H. and C.R.) worked independently and evaluated the articles according to the Newcastle-Ottawa Quality Assessment Scale (NOS) for cohort studies and the National Heart, Lung, and Blood Institute (NIH) Quality Assessment Tool for case series studies (14 questions are used to rate studies as "good," "fair," or "poor" quality).Articles with a very high risk of bias were excluded.NOS scores of 1 to 3 indicate low quality, 4 to 6 indicate moderate quality, and 7 to 9 indicate high quality [11][12][13].

Effect measures
The outcomes were measured using mean age in years and male-to-female incidence rate ratio.If the median was reported, the mean was calculated using the formula reported by Hozo et al.X ≈ a+2m+b 4 where x is the mean, a is the smallest value (minimum), b the largest value (maximum), and m the median [14].

Meta-analysis
The male-to-female incidence rate ratio was calculated using a weighted average of the 15 selected articles.Inverse-variance weighting summarized the effect size from each study.We compared the resulting summarized statistics with the median age reported by Ostrom et al. [1,5].Standard deviation was estimated by quartile and age-related range.Random effects model was performed with the Sidik-Jonkman error variance estimator paralleled with Knapp-Hartung standard error adjusted.SPSS® meta-analysis package for raw data was used (version 28) [15,16].Missing sample sizes and standard deviations were calculated when possible.We included 13 studies reporting the mean and median age at glioma diagnosis.Significance was established using confidence intervals at 95% or p < 0.05.

Statistical model
The random-effects model was used for the analysis.The studies included in the analysis were assumed to be a random sample from a universe of potential studies, and the research was used to infer that universe.

Literature research
The search strategy yielded 1096 articles.After removing duplicates, the titles and abstracts of 487 articles were screened.Of these, 66 articles were assessed for eligibility, and 51 were excluded due to lack of data, other types of tumors included, pediatric patients and insufficient sample size.Studies based on The Central Brain Tumor Registry of the United States (CBTRUS) and The Surveillance, Epidemiology, and End Results Program (SEER) were excluded, for they encompass the Hispanic population as a single ethnic group, excluding the geographical, cultural, socioeconomic, and genetic characteristics, which could be a source of bias [8].Thus, a total of 15 studies were selected to be analyzed in the systematic review.Two of these studies did not report the standard deviation, so they were not included in the meta-analysis.The study selection process is presented in Fig. 1.

Characteristics of the included studies
The studies were published between 2000 and 2022.The sample size ranged from 40 to 3346, and the age interval ranged from 3 to 96 years.There were six studies from Mexico, four from Brazil, two from Chile, two from Argentina, and one from Colombia (Table 1).All the studies used histopathology to confirm the diagnosis.One study included molecular tests to make the diagnosis.Ten studies used the 2007 WHO classification of CNS tumors.One study used the 2021 WHO classification of CNS.One study used the 2016 WHO classification of CNS tumors.One study used the St. Anne-Mayo grading system.Finally, two studies did not include the classification they used [17,18].Two studies [19,20] were excluded from the meta-analysis because they did not report the standard deviation.

Risk of bias
The risks of bias in the cohort studies (n = 14) were measured by answering the eight questions from the NOS Assessment Scale; all studies scored ≥ 5 points.Supplementary Table 1 summarizes the NOS Assessment Scale results.The study conducted by Martinez-Muñoz et al. [21] underwent an assessment of bias using the NIH Quality Assessment Tool for case studies; the results of all nine questions were positive, indicating high quality [11] Suplementary info.

Age
Gliomas (mean age at the time of diagnosis): The mean age was 50.88 years, with a 95% confidence interval (95% CI) of 47.83 to 53.94 years.The I 2 statistic was 97%, and the Tau 2 was 29.06 (Tau = 5.39).The estimated prediction interval was 38.53 to 63.24.The actual effect size was 95%

Sex
Male-to-female rate ratio: The estimated incidence ratio from the analysis was 1.39.Table 3 describes the analyzed studies.

Geographic distribution
The age distribution of different geographic regions of the world and the Americas are presented in Fig. 2. The data used for the Latin American countries were based on our systematic review and meta-analysis.The data for non-Latin American countries were retrieved from national and regional cancer registries and from articles that reported the mean or median age at diagnosis of; data used for the US and Canada was collected from the CBTRUS and the Brain Tumor Registry of Canada (BTRC).Although other countries have cancer registries, many report data concerning all CNS tumors and do not report specific data for gliomas.Figures were created using MapChart ® (Fig. 2).

Discussion
After conducting a thorough analysis of the research on gliomas in the Latin American population, we found that patients with glioblastoma and other gliomas tend to be diagnosed at a younger age compared to other regions around the world.This was determined through a systematic review and meta-analysis of the available literature.Neuroepithelial tissue tumors are typically diagnosed in individuals with a median age of 57, according to CBTRUS  [1,5].The Latino mean age of diagnosis found in this study was 6.1 years lower, for a mean age of 50.89 years, 95% CI (47.8-53.9),suggesting that gliomas occur considerably earlier in Latinos than in non-Hispanic whites and is crucial because it represents a health problem and an economic problem since it is affecting an economically active population.Wegman-Ostrosky et al. proposed three possible explanations for the younger age at diagnosis: 1) differences in the Latin American population pyramid compared to those from the US, 2) environmental exposure, and 3) genetic factors, such as germline mutations in TP53, MSH2, MLH1, and MSH6 [22].
Moreover, the reported incidence of glioblastoma increases significantly with age, with a median of 63 years.As stated in the CBTRUS, the median age varies by ethnicity, with non-Hispanic whites having a higher median age (64 years) than Hispanics (60 years) [4,5,23].The mean age at diagnosis of glioblastoma in our systematic review was 53.33 years (95% CI 51.04 to 55.68), a decade earlier than the reported median age in non-Hispanic whites.Nonetheless, the difference we found is considerably larger than the CBTRUS described [5,23].
This fact has been asserted by Walsh et al. [8] in which the Hispanic population from the CBTRUS was divided into two categories: those with Mexican/Central American origin or Caribbean origin.The Mexican/Central American group The data used for the Latin American countries were based on our systematic review and meta-analysis.The data for non-Latin American countries were retrieved from national and regional cancer registries and from articles that reported the mean or median age at diagnosis of; data used for the US and Canada was collected from the CBTRUS and the Brain Tumor Registry of Canada (BTRC).Although other countries have cancer registries, many report data concerning all CNS tumors and do not report specific data for gliomas had a lower median age than the Caribbean group (45 years vs. 52 years).The authors hypothesized that the increased European admixture could explain this phenomenon.However, subsequent analysis implies that this is just a partial explanation and that other variables should be considered.Sex differences have been well-established for many brain tumors, including gliomas [6].The male-to-female incidence rate ratio (IRR) of gliomas reported by the CBTRUS is 1.47.The reported Hispanic male-to-female IRR reported by the CBTRUS is lower (1.35) [5]; in our review, it was 1.39.In 2021, the US had a sex ratio of 97.94 males per 100 females, Chile 97.31 males per 100 females, Brazil 96.51 males per 100 females, Colombia 96.46 males per 100 females, and Mexico had the lowest sex ratio of 95.77 males per 100 females [24].The US has a higher male-to-female IRR than Latin America, which could explain the higher maleto-female incidence ratio reported by the CBTRUS.Additionally, gliomas are more common in men than in women [25].Moreover, the wide variation in the male-to-female IRR present in individual studies can be explained by the small sample sizes included.Histological subtypes included in each study can also affect the male-to-female IRR.For instance, glioblastoma is more frequent in men as opposed to lower grade gliomas, which are relatively more common in women.The studies with the highest male-to-female IRR [26][27][28] include glioblastoma exclusively, while the only study in which the male-to-female IRR favors women [20] does not include glioblastoma.
Age and sex distribution were not the only disparities we discovered in the glioma literature from Latin America.For instance, there needs to be more epidemiologic information.Most of the articles we found were from Brazil [20,27,29,30] and from Mexico [17,19,22,[31][32][33].We found some from Colombia [21] Argentina [18,34], and Chile [26,28].Nonetheless, most Latin American countries still need to report their epidemiologic information on gliomas.Only one of the articles used the 2021 WHO classification [34].In addition, even though many of the articles were published after 2016, only one [31] used the 2016 WHO classification of CNS tumors.The rest used the 2007 WHO CNS tumors classification or the St. Anne-Mayo grading system.The 2016 and 2021 WHO classifications of CNS tumors incorporate molecular parameters with histological characteristics [35,36].Molecular parameters revolutionized CNS tumors' diagnosis and prognostic accuracy; however, these tools are only available to some patients in Latin American countries due to the elevated costs [37].The disparity in the use of molecular tests will most likely cause differences in how gliomas are diagnosed and treated.This will have important repercussions in the patients' outcomes and will make adequate cross-country comparisons very difficult.
Our study presented a number of limitations.The most noteworthy is that the data we retrieved is based on published articles and not on cancer registries, for these registries are lacking in several of Latin American countries.The classification of CNS tumors has changed considerably, and most reviewed manuscripts did not include molecular reports.One study incorporated glioma subtypes not considered in the newest WHO5 classification [19], and only one study incorporated molecular parameters [34].The wide variation in sex ratios can be attributed to the small sample sizes and the preponderance of glioblastoma as the underlaying histology in several of the included studies.The quality score for most studies was low because many did not select a non-exposed cohort.We limited publications to English and Spanish so that we might have missed publications written in other languages in Latin America, i.e., Portuguese or French.Gliomas are rare cancers; therefore, the sample size of most articles is small; this is a possible explanation for the high heterogeneity described in our meta-analysis.Furthermore, the main objectives of these studies were different and not about age and sex distribution; this could be another explanation for the high heterogeneity in our meta-analysis.
Finally, our study's results propose disparities in the age and sex distribution of gliomas in Latin America compared to other continents.There is a need for prospective registration of patients with gliomas in Latin America to consolidate the epidemiology of these CNS tumors and identify at-risk populations.This study represents the first systematic review and meta-analysis of the age and sex distribution of gliomas in Latin American people.

Fig. 1
Fig. 1 PRISMA Flow Diagram.Illustrates the selection process of literature research based on PROSPERO:.1096 articles were identified, 487 articles were screened, 15 articles were included in the review and 13 articles were included in the meta-analysis

Table 1
Table 2 describes the studies used for this analysis.Glioblastoma (mean age at the time of diagnosis): The mean age was 53.36 years, with a 95% CI of 51.04 to 55.68 years.The I 2 was 91%, and Tau 2 was 11.86 (Tau = 3.44), with an estimated prediction interval of 44.96 to 61.76.The actual effect size was 95% of all comparable populations.This analysis was based on the studies mentioned in Table 2. Characteristics of the included studies WHO world health organization, NR not reported, GBM glioblastoma multiforme *These articles were not included in the meta-analysis

Table 2
Forest Plot of observed mean age at the time of diagnosis of gliomas and glioblastomas in years

Table 3
Observed male-to-female incidence rate ratio (IRR) Global mean age distribution of gliomas by country.Illustrates the mean or median age of diagnosis of gliomas worldwide.