Screening for Familial Hypercholesterolemia in Small Towns: Experience from 11 Brazilian Cities at the HipercolBrasil Program

Background: Familial hypercholesterolemia is an autosomal dominant disease clinically characterized by elevated serum levels of low density lipoprotein-cholesterol (LDL-C) and associated with the occurrence of early atherosclerotic cardiovascular disease. In Brazil, HipercolBrasil, currently the largest screening program underway, exists since 2012 and has already identied more than 2000 individuals with causal genetic variants for FH. The standard approach of HipercolBrasil is based on cascade-screening of refereed index cases, hypercholesterolemic individuals with a clinical suspicion of FH. Methodology: to assess a new methodology for identifying new individuals with genetic alterations for FH we performed a comprehensive city-wide screening in 11 small Brazilian cities (up to 60 thousand inhabitants). The selection of cities occurred in 3 ways: 1) cities with suspicious founder effect (4 cities); 2) Cities in a region with high rates of dyslipidemic individuals as described by the National Health System database (DATASUS) (2 cities); and 3) Cities geographically close to other cities with a high incidence of individuals with FH (5 cities). Results: One-hundred and ve (105) index cases and 409 rst-degree relatives were enrolled. The yield of such approach was signicantly better than the general HipercolBrasil positivity rate in molecular screening. We identied 36 IC with a pathogenic or likely-pathogenic variants for FH and 240 affected rst-degree relatives. Conclusion: our data suggest that once detected, specic geographical regions warrant a target approach for the identication of clusters of FH individuals.


Introduction
Familial hypercholesterolemia (FH) is an autosomal dominant disease clinically characterized by elevated blood levels of low density lipoprotein-cholesterol (LDL-C) and associated with the occurrence of early atherosclerotic cardiovascular disease (ASCVD). 1 2 The prevalence of FH in the world is estimated to be approximately 1:250 in the heterozygous form and 1: 600.000 in homozygosity 3  However, although relatively frequent, FH is still an underdiagnosed disease 5 . To assist in the identi cation of individuals with the disease, genetic cascade screening has been used in several countries, such as the Netherlands 6 , United Kingdon 7 and Spain 8 . This method has already been recognized as cost effective for the identi cation as well as for the prevention of early ASCVD in individuals with FH 9 .
In Brazil, HipercolBrasil, currently the largest screening program underway, exists since 2012 10 and has already identi ed more than 2000 individuals with causal genetic variants for FH. The program currently performs genetic testing on any index-cases (ICs) individual with LDL-cholesterol (LDL-C) ≥ 230 mg/dL 11 and in rst-degree relatives of those with pathogenic or likely-pathogenic variants. In these cases, after receiving the genetic report, the patients are referred to an outpatient clinic for FH treatment.
Between July 2017 and July 2019, we tested another methodology for identifying new individuals with genetic alterations for FH. In this new model, the HipercolBrasil program performed comprehensive city-wide screening in 11 small Brazilian cities (up to 60 thousand inhabitants) that showed signs of a higher prevalence of people with FH. Here we describe the methodology and rst results of a new FH screening strategy based on small cities with some evidence of increased FH prevalence.

Aim
Between July 2017 and July 2019, we tested another methodology for identifying new individuals with genetic alterations for FH. In this new model, the HipercolBrasil program performed comprehensive city-wide screening in 11 small Brazilian cities (up to 60 thousand inhabitants) that showed signs of a higher prevalence of people with FH. Here we describe the methodology and rst results of a new FH screening strategy based on small cities with some evidence of increased FH prevalence.

Study Design
This study is an open prospective study evaluating a cohort of patients who were ascertained by the FH genetic screening cascade program (HipercolBrasil).
The study was conducted at the Laboratory Genetics and Molecular Cardiology of the Heart Institute (InCor), University of São Paulo Medical School Hospital, São Paulo, Brazil. The protocol was approved by the Institutional Ethics Committee (CAPPesq protocol l00594212.0.1001.0068). Figure 1 shows inclusion criteria and study design. We enrolled individuals from 11 selected cities with up to 60,000 inhabitants throughout the Brazilian territory. The selection of cities ( Fig. 1) occurred in 3 ways: 1) cities with a suspicious founder effect, i.e. occurrence of homozygous individuals, but with nonrelated parents and from distinct geographic origins -Major Vieira, Papanduva, Lagoa do Mato and Passagem Franca; 2) Cities in a region with high rates of myocardial infarction as described by the National Health System database (DATASUS) -Bom Despacho and Moema 12 ; and 3) Cities geographically close to other cities with a high incidence of individuals with FH -Bambuí, Pimenta, Luz, Colinas and Buriti Bravo.

Study population
In all cities, initial contact was made with the local secretary of health to explain the project and agreement on the partnership. Contact was made via telephone before visiting each city and an agreement was set by both parties via e-mail. Once in the city, the team was assisted by a health agent appointed by the health secretary. In the cities where there were evidence of a founder effect and in the ones where there were reports of high myocardial infarction incidence, the sample collection started from family members of these ICs. In these cities, there was also an active search for new ICs from medical records and cholesterol tests carried out in the clinical analysis laboratories of the local healthcare units. Individuals were considered as ICs when total cholesterol (TC) > 300 mg/dL and/or LDL-c ≥ 210 mg/dL with triglycerides < 300 mg/dL. In these cases, a blood sample was collected to perform a second measurement of cholesterol fractions in our laboratory, since many cities only performed TC exams. Those with LDL-c ≥ 210 mg/dl were selected for genetic sequencing.
Individuals who did not reach this value received a report with the values of total cholesterol and fractions and were excluded from the study.

Data Analysis
For continuous variables, the mean and standard deviation were calculated. Categorical variables are shown as frequencies. The differences between frequencies were compared using the chi-square test. The differences between means were compared with Student's T-test or ANOVA, if necessary. Statistical signi cance was considered at a p-value < 0.05. Table 1 shows characteristics of the 11 visited cities, state of location at the Brazilian federation, number of inhabitants and date of each visit. The city with the lowest number of total inhabitants was Moema with 7,028 and the largest was Bom Despacho with 45,624 inhabitants, both in the state of Minas Gerais.

Results
The rst cities to be visited were Major Vieira and Papanduva (Sep/2017) and the last were Buriti Bravo and Colinas (Feb/2019). Table 2 Table 3 shows the three ICs groups (negative, positive or VUS) and their clinical and biochemical data. In total, 105 ICs were sequenced, and pathogenic or likely-pathogenic variants were found in 36 (37.8%) and VUS in 5 (5.25%) individuals. Most of ICs were females (67.6%) and when the clinical and biochemical characteristics were evaluated among the three groups, there was, as expected, a statistically signi cant difference regarding baseline (untreated) TC and LDLc, with the positive group presenting the highest values of TC and LDL-c respectively 382 ± 150 mg/dL and 287 ± 148 mg/dL.    Table 3: CAD-coronary artery disease ; early CAD de ned as ASCVD event < 55 and 60 years of age respectively in males and females; lipids in mg/dL; baseline lipids = untreated ;  Table 4: CAD-coronary artery disease ; early CAD de ned as ASCVD event < 55 and 60 years of age respectively in males and females; lipids in mg/dL; baseline lipids = untreated Table 5 shows all the encountered variants and the location where they were identi ed. In total, 21 different variants were identi ed with 3 variants appearing more frequently: LDLR duplication from promoter to exon 6 in Passagem Franca (49) and Lagoa do Mato (29); LDLR duplication from exon 4 to exon 8 in Major Vieira (45) and Papanduva (41); and the variant p.Asp224Asn in Bom Despacho (39) and Moema (34). These frequencies suggest that these variants have founder effects in these localities. Six homozygous patients and one compound heterozygous in trans were found.

Discussion
This study describes the result of implementing a genetic cascade screening system for FH in 11 small Brazilian cities. FH is a high frequency disease worldwide and in Brazil 4 . A Spanish study pointed that 55% of men and 24% of women with FH in their 50 s present manifestations of coronary heart disease, such as myocardial infarction and angina pectoris, which makes FH a public health problem and the proper identi cation and treatment are of primary importance 1415 . Genetic cascade screening is considered a cost-effective method for identifying undiagnosed cases of FH 9 .
Despite the known cost bene ts of cascade screening in FH, its implementation worldwide has been suboptimal. Different local barriers and implementation hurdles have to be identi ed and overcome. How to implement cascade screening in small localities, for example, has been mainly overlooked. This challenge is greater in a continental country like Brazil, where in addition to the enormous geographic distances there is inequality in access to health services. Here we describe the HipercolBrasil experience in conducting comprehensive cascade screening in small towns in Brazil. In this new model, genetic cascade screening was carried out in cities that showed evidence of a higher prevalence of FH due to previous nding of individuals with the HoFH phenotype from the same city, or because those regions had elevated reported frequency of myocardial infarction or if the cities were geographically close to either.
Cities that had evidence of a founder effect were the ones that presented a higher identi cation of individuals affected per each IC analyzed (respectively Major Vieira, Papanduva, Lagoa do Mato and Passagem Franca). In these cities, we started from homozygous individuals whose parents were non-related and were born in different geographic regions.
Of importance, the rate of family members per IC was higher in cities with suspected founder effects: Major Vieira, Papanduva, Moema, Passagem Franca and Lagoa do Mato. This probably occurred because these cities had a small number of inhabitants and since most relatives had some degree of familial rapport there was a low detection of new ICs. This did not occur in Bom Despacho, a city considerably larger than the others (45,624 inhabitants) and, although the number of family members collected was similar to that of other cities, there was a higher number of ICs collected (67) decreasing the rate of relatives/IC to 2.3.
Cities visited geographically close to cities with suspected founder effects, (Bambuí, Buriti Bravo, Colinas, Pimenta and Luz) had a low uptake of ICs and consequently a low number of identi ed relatives. Bambuí was the city with the lowest number of IC uptake and no pathogenic or likely-pathogenic variants were identi ed. Particularly for Bambuí there was a low adherence from the local health care professionals and most medical reports lacked the biochemical pro les of patients. In the other cities, although we could detect ICs with FH associated variants, the genetic cascade screening was not continued by trained health agents. However, it is noteworthy that these were the last cities visited by the program and that it remains open for the genotyping of new cases.
Initially, some cities had high IC uptake from exams with TC > 300 mg/dL. However, a second and more speci c determination of LDL-c in our laboratory discarded the genotyping of many individuals that were below the LDL-c cutoff (≥ 210 mg/dL), decreasing the actual number of cases.

Conclusion
Cascade screening in small cities (less than 60 thousand inhabitants) proved to be effective in those with a suspected founder effect. However, some points might be of great importance for the cascade screening to be effective, and those might be executed before the decision of what city to track: establishment of a formal partnership and explicit interest of the local health department in receiving the program and performing the cascade screening; availability of clinical analysis laboratories datasets to carry out a retrospective survey of cholesterol tests; dissemination via radio stations and social media about the disease and the program for greater adherence by the inhabitants.
This study is limited by the relative number of cities evaluated considering the continental size of Brazil. However, it suggests that the designed approach may be useful for detecting individuals with FH. In conclusion, our data suggest that once detected, speci c geographical regions warrant a targeted approach for the identi cation of clusters of FH individuals.