Professional Documents
Culture Documents
The Panorama of Familial Hypercholesterolemia in Latin America: A Systematic Review
The Panorama of Familial Hypercholesterolemia in Latin America: A Systematic Review
Abstract The burden caused by familial hypercholesterol- penetrance. Autosomal dominant FH is attributed to mu-
emia (FH) varies among countries and ethnic groups. The tations in three different genes: LDL receptor (LDLR),
prevalence and characteristics of FH in Latin American (LA) APOB, and proprotein convertase subtilisin/kexin type 9
countries is largely unknown. We present a systematic review (PCSK9) (1, 3–5). Other FH genes have been searched for
(following the PRISMA statement) of FH in LA countries.
using exome sequencing without success (2). FH caused by
The epidemiology, genetics, screening, management, and
unique challenges encountered in these countries are dis- mutations in LDLR adaptor protein (LDLRAP) is known as
cussed. Published reports discussing FH in Hispanic or LA autosomal recessive FH (2). FH is the most common mono-
groups was considered for analysis. Thirty studies were in- genic disorder leading to premature CHD; despite this
cluded representing 10 countries. The bulk of the data was fact, it is notoriously underdiagnosed and undertreated
generated in Brazil and Mexico. Few countries have regis- worldwide (6).
tries and there was little commonality in FH mutations be- Homozygous FH (HoFH) is characterized by extremely
tween LA countries. LDL receptor mutations predominate; high levels of LDL-C (460–1,160 mg/dl) and early onset
APOB and PCSK9 mutations are rare. No mutation was found coronary artery disease (typically by the second decade of
in an FH gene in nearly 50% of cases. In addition, some
country-specific mutations have been reported. Scant infor- life) (7). Mean LDL-C concentration in untreated patients
mation exists regarding models of care, cascade screening, is close to 615 mg/dl (7, 8). Patients are classified into two
cost, treatment effectiveness, morbidity, and mortality. In groups based on the level of LDLR activity, either <2% (re-
conclusion, FH is largely underdiagnosed and undertreated ceptor negative) or 2–25% (receptor defective). Receptor
in the LA region. The genetic admixture with indigenous defective patients have a better prognosis than receptor
populations, producing mestizo’s groups, may influence the negative cases (9–11).
mutational findings in Latin America. Potential opportunities Heterozygous FH (HeFH) is caused by a single inherited
to close gaps in knowledge and health care are identified.— copy of a mutation. The frequency of a heterozygous muta-
Mehta, R., R. Zubirán, A. J. Martagón, A. Vazquez-Cárdenas,
tion is >90, 5, and <1% in the LDLR, APOB, and PCSK9
Y. Segura-Kato, M. T. Tusié-Luna, and C. A. Aguilar-Salinas. The
panorama of familial hypercholesterolemia in Latin America: genes, respectively (5). A causal mutation in one of these
a systematic review. J. Lipid Res. 2016. 57: 2115–2129. genes is identified in 60–80% of cases. Affected individuals
are characterized by LDL-C levels two to three times greater
Supplementary key words heterozygous familial hypercholesterol- than normal (190–400 mg/dl). The mean untreated LDL-
emia • homozygous familial hypercholesterolemia • autosomal recessive C concentration is 199.9 mg/dl (12). HeFH is suspected
hypercholesterolemia
Familial hypercholesterolemia (FH) is a syndrome that Abbreviations: ARH, autosomal recessive hypercholesterolemia;
causes defective clearance of LDL cholesterol (LDL-C) and CHD, coronary heart disease; DALY, disability-adjusted life year; DLCN,
Dutch Lipid Clinics Network; FH, familial hypercholesterolemia; HeFH,
results in premature coronary heart disease (CHD) (1–3).
heterozygous familial hypercholesterolemia; HoFH, homozygous famil-
Two forms have been reported: autosomal dominant ial hypercholesterolemia; LA, Latin American; LDL-C, LDL cholesterol;
and autosomal recessive. The vast majority of cases have LDLR, LDL receptor; LDLRAP, Low density lipoprotein receptor adap-
the autosomal dominant pattern of inheritance with 90% tor protein; NICE, National Institute for Health and Clinical Excel-
lence; PCSK9, proprotein convertase subtilisin/kexin type 9; QALY,
quality-adjusted life year.
1
To whom correspondence should be addressed.
Manuscript received 25 September 2016 and in revised form 23 October 2016. e-mail: caguilarsalinas@yahoo.com
Published, JLR Papers in Press, October 24, 2016 The online version of this article (available at http://www.jlr.org)
DOI 10.1194/jlr.R072231 contains a supplement.
Copyright © 2016 by the American Society for Biochemistry and Molecular Biology, Inc.
This article is available online at http://www.jlr.org Journal of Lipid Research Volume 57, 2016 2115
with an elevated LDL-C (LDL-C >190 mg/dl), the presence hypercholesterolemia, severe hypercholesterolemia, stroke and
of xanthomas (tendinous/tuberous), corneal arcus (<45 hypercholesterolemia, premature cardiac disease and hypercho-
years old), family history of FH, and either a personal or lesterolemia, cascade screening, and founder effect. These terms
were crossed with the keywords and MeSH terms for LA ethnicity:
family history of premature coronary artery disease.
minority groups, ethnic groups, Hispanic American, Native Amer-
Traditionally, the prevalence of HeFH is considered to ican, continental ancestry, American Indian, and all countries
be 1 in 500 individuals. This figure is based on the results that constitute Latin America.
of a European survey of familial lipoprotein disorders in All literature in Spanish, Portuguese, and English was included.
myocardial infarction survivors (13, 14). However, recent In addition, the reference lists of review articles and conference
work in a Dutch population estimated a prevalence of 1 in abstracts were also considered. Abstracts were assessed indepen-
137 persons (1, 10, 11). In Australia, 1 in 229–353 inhabit- dently by two research associates to identify eligible research
ants may be affected, while a figure of 1 in 213 has been reports. If an abstract was not disqualified on the basis of the study
proposed in a Chinese population (15, 16). Such disparities methods or data presented in the abstract, a full copy of the re-
port was obtained.
in prevalence estimates are due to differences in popula-
tion ancestry and in diagnostic criteria used to define the
disease. In populations with founder effect, such as those
of South Africa, Quebec, and Lebanon, prevalence estimates RESULTS
are expected to be higher. For example, the prevalence of Our search strategy retrieved a total of 19,658 publications.
HeFH in French-Canadians (Quebec) is approximately After removing any duplicate documents, a total of 19,508
1:270; in the rest of Canada, where the population is more abstracts were reviewed. Of these, 19,454 articles were ex-
genetically heterogeneous, this figure may be closer to cluded as they did not complete the inclusion criteria of
1:500 (14). the systematic analysis. Finally, 63 relevant full-text articles
The burden caused by FH varies among countries and were assessed in detail for eligibility. Of these, 33 publications
ethnic groups. The existing information concerning FH were excluded due to incomplete information. As a result,
has been gathered from European cohorts or countries a total of 30 studies were included for the purposes of this
with national registries. The Netherlands has the best record, article. The PRISMA algorithm is presented in Fig. 1. The
with an estimated 71% of cases identified. However, the characteristics of the 30 articles included in this systematic
situation is very different in other parts of the world. Glob- analysis are shown in Table 1.
ally, less than 5% of FH individuals have been identified
(5). In Latin America, the prevalence of FH is largely un- Characteristics of individuals with FH in Latin America:
known. In this part of the world, there is a lack of public summary
awareness regarding FH. In this article, we present a sys- Information regarding the scenario of FH in Latin
tematic review of FH in LA countries. The epidemiology, America was available from 30 published articles (Table 1)
genetics, screening, management, and unique challenges (11, 18–48). In total, 10 countries are represented, namely
encountered in this region are discussed. Armed with this Argentina, Brazil, Chile, Costa Rica, Cuba, Honduras, Mex-
knowledge, LA health systems may implement policies, ico, Panama, Paraguay, and Uruguay. Of these, Brazil and
including national registries, enabling the opportune diag- Mexico provided the bulk of the information with nine and
nosis and management of this disease. seven publications, respectively. The types of publications
included 18 case reports (often with cascade screening of
the family) and 12 reviews of FH cohorts. Data on both
METHODS
homozygous FH (HoFH) and HeFH have been published.
A systematic review was undertaken to identify all published With regard to FH case reports (n = 18), 10 articles iden-
data regarding FH in LA populations. LA populations are any tified HoFH subjects (those that report mutational analysis
Spanish or Portuguese speaking nations south of the United States. confirm homozygous LDLR mutations), 6 discuss HeFH
A systematic search was conducted following the PRISMA state- cases (4 caused by LDLR mutations and 3 due to APOB
ment (17) in PubMed and SciELO from 1982 to 2016. The aim of mutations), and 1 reports the characteristics of a Mexican
the PRISMA statement is to help authors improve the reporting of
individual with autosomal recessive hypercholesterolemia
systematic reviews and meta-analyses. PRISMA is an evidence-based
minimum set of items for reporting such documents; it consists
(ARH) (due to a LDLRAP mutation). Five case reports do
of a 27 item checklist and a flowchart explaining the flow of infor- not have mutational analysis (Argentina, 2; Honduras, 1;
mation through the phases of the review. Any type of published Panama, 1; and Brazil, 1). In the HoFH case reports (n =
report discussing FH epidemiology, FH mutations, and FH public 10: Brazil, 2; Mexico, 2; Argentina, 2; Honduras, 1; Pan-
health policy in Hispanic or LA groups was considered for ama, 1; Paraguay, 1; and Uruguay, 1), the mean age of
analysis. Articles providing clinical phenotype and/or mutational adult subjects (n = 9) was 32.9 years. The mean total choles-
analysis in FH subjects were selected. Thus, all epidemiological, terol and LDL-C levels were 702.2 mg/dl and 521.1 mg/dl,
cross-sectional, cohort, retrospective, longitudinal, observational, respectively. The majority of the HoFH cases (83%) had
comparative, case-control, and case reports were considered con-
evidence of atherosclerotic disease.
taining the following keywords or MeSH terms: familial hyper-
cholesterolemia, heterozygous familial hypercholesterolemia, In the HeFH case reports (n = 6: Chile, 2; Brazil, 1; Mex-
homozygous familial hypercholesterolemia, type II hyperlipo- ico, 1; and Uruguay, 2), the mean age of subjects was 51.3
proteinemia, LDL receptor mutation, apoB mutation, PCSK9 years and the mean total cholesterol level was 415 mg/dl
mutation, xanthomas, arcus senilis, aortic valvular lesions and (n = 6). Tendon xanthomas and premature coronary artery
disease were reported in four and four cases, respectively. most frequent FH mutations encountered in Latin America
Mutational analysis was available in four case reports, three are on LDLR (21–54). However, autosomal recessive FH and
found mutations in the LDLR (Chile and Costa Rica), and APOB mutations have also been reported in this region
one article reported an APOB mutation (Uruguay). (55–58).
The only case of ARH was reported in two Mexican sib- In Brazil, in the early 1990s, a common mutation in the
lings (aged 20 and 27 years, respectively). The mean total LDLR gene was found in FH probands. This was called the
cholesterol and LDL-C concentrations were 677 and 655 “Lebanese allele” (exon 14 Cys681); it is a cause of FH in
mg/dl, respectively. They both had tendinous and tuberous Arab populations and produces a shortened receptor lacking
xanthomata, but the presence of atherosclerotic lesions three domains, resulting in an abnormally slow exit rate
was not recorded. from the endoplasmic reticulum (55). Figueiredo et al.
In addition, we found 11 HeFH cohorts {n = 11: Argen- (21) searched for the presence of this mutation in 18 FH
tina [1 pediatric cohort without genetic testing (n = 33)]; subjects from 10 unrelated families living in Ribeiro Preto,
Brazil [5, of which only 2 are pure HeFH cohorts (n = 35 Brazil. They were able to identify the mutant allele in nine
and n = 202)], Cuba [1 (n = 16)]; Mexico [3, of which 2 are patients from five families (eight heterozygous cases and
pure HeFH cohorts (n = 46 and n = 29)]; and Uruguay [1 one homozygous case), all five families confirmed Leba-
(n = 71)]}. The mean age of adult participants was 46.0 nese ancestry. This was the first report of this mutation in a
years (range 37.7–52.7 years) and the mean total choles- non-Arab population, indicating an important contribu-
terol and LDL-C levels were 347 and 303.3 mg/dl, respec- tion of this ethnic group to FH in Brazil. Alberto et al. (22)
tively. In the pediatric cohort (aged 2–15.5 years), the studied 59 FH subjects from the same region of Brazil and
mean levels were 305 ± 47.9 mg/dl and 232 ± 54.2 mg/dl, confirmed the presence of the Lebanese allele in 30% of
respectively. The percentage of adult HeFH subjects with the cases. Several years later, Salazar et al. (24) discovered
xanthomas was recorded in nine studies and ranged from seven novel LDLR mutations in families living in Sao Paolo;
6.4% to 100%; the presence of CHD was mentioned in however, the Lebanese mutation was not found. They spec-
nine cohorts and ranged from 6% to 59% of the population. ulated that the frequency of FH mutations may vary from
The age of diagnosis was late (mean 46 years). Finally, with state to state depending on the ethnic background of the
regard to genetic analysis, the majority of cohorts reported inhabitants. Recently, Jannes et al. (27) described 248 in-
LDLR mutations, with the Lebanese allele predominant in dex cases; an FH-causing mutation (LDLR, APOB, or PCSK9
Brazil and an Afrikaner mutation reported in a Cuban family. genes) was found in 125 subjects (50.4%). After analyzing
Mutations in APOB were published in one Brazilian, one index cases and mutation-positive relatives (n = 359), 70
Uruguayan, and one Mexican cohort, and one PCSK9 mu- different causal mutations in the LDLR (97.2%) and 2 in
tation was recorded in a Mexican cohort. APOB (2.8%) were found. Only 10 homozygote cases (8
The cohort studies did not contain information to assess index cases and 2 relatives) and 1 compound heterozygote
the prevalence of atherosclerotic disease. There were no case were identified. The most common mutations were in
prospective studies with long-term follow up. Information exons 14 and 4, with 8.5% of the mutation-positive popula-
regarding statin use was also incomplete. The diagnostic tion carrying the Lebanese allele.
criteria utilized also varied, underscoring the lack of uni- In Mexico City, Robles-Osorio et al. (42) assessed the
form criteria. contributions of the LDLR, APOB, and PCSK9 mutations as
causes of FH. Three reference centers participated in this
Genetics of FH in Latin America analysis and 46 index cases and 68 affected relatives were
LDLR mutations. Table 2 shows the most common identified. All index cases were diagnosed as having hetero-
LDLR mutations associated with FH in Latin America. The zygous autosomal dominant FH. LDLR mutations were
Mean total
HeFH/ Study Premature cholesterola Mean LDL-Ca
a
Country Reference HoFH Type Sample Size Age Male (%) CHD (%) Xanthoma (%) (mg/dl) (mg/dl) FH Mutation FH Diagnosis
Argentina 18 HeFH Retrospective 33 9.5 12% (n = 4) NR 0 305 ± 47.9 232 ± 54.2 NR LDL-C levels and
pediatric family history
cohort
19 HoFH Case report 2 (brother 8 and 12 50% (n = 1) Sister with 50% 521 409 NR LDL-C levels and
and sister) bicuspid (n = 1) family history
aortic valves
and thickened
intima of
carotid artery
20 HoFH Case report 1 22 100% Bicuspid aortic NR 486 NR NR LDL-C levels and
valve and family history
ischemic
cardiomyopathy
Brazil (30) 21 HeFH FH cohort 18 cases from HeFH 48.6 NR 38.9%HeFH 17.6% HeFH HeFH 350.3 HeFH 271.2 9 subjects from Total cholesterol
and HoFH (probands) 10 families (n = 3) 5 families with and LDL-C
(Ribeirao 17 HeFH HoFH 34 100% HoFH 100% HoFH HoFH 610.2 HoFH 545.2 Lebanese >95th percentile
Preto) 1 HoFH allele (LDLR) for age and sex
Mean total
HeFH/ Study Premature cholesterola Mean LDL-Ca
Country Reference HoFH Type Sample Size Agea Male (%) CHD (%) Xanthoma (%) (mg/dl) (mg/dl) FH Mutation FH Diagnosis
2119
positive = 3
2120
TABLE 1. Continued.
Mean total
HeFH/ Study Premature cholesterola Mean LDL-Ca
a
Country Reference HoFH Type Sample Size Age Male (%) CHD (%) Xanthoma (%) (mg/dl) (mg/dl) FH Mutation FH Diagnosis
41 HeFH Case report Proband and 93 0% Atherosclerotic Proband and 395 NR NR Total cholesterol
and cascade 9 family lesions in 4 family and LDL-C >95th
screening of members carotid members percentile for age
family arteries and sex
42 HeFH FH cohort 46 ICs 43.4 ± 18 NR NR 81% (n = 37) 360 ± 63 Mutation+ = 287 ± 62 17 in LDLR Total cholesterol
(46 IC and Mutation + Mutation + = and LDL-C >95th
68 relatives) 17 LDLR 39.2 ± 15.4 390 ± 48 Mutation = 315 ± 46 percentile for
mutation+ Mutation– Mutation = age and sex and
25 mutation 44.3 ± 13.4 325 ± 64 258 ± 65 clinical phenotype
NR, not reported; IC, index case; HeFH, heterozygous familial hypercholesterolemia; HoFH, homozygous familial hypercholesterolemia; CHD, coronary heart disease.
a
Mean or range.
TABLE 2. Most common LDLR mutations causing FH in Latin America (mutations reported in at least 3 subjects
for Brazil and Mexico; or mutations published in case reports for other countries
Mutations reported in at least three subjects for Brazil and Mexico or mutations published in case reports for
other countries. Mutations unique to a country (according to the University College of London database or article
cited) are FH-Hondurus, FH Mexico 2 and 3. Mutations which do not appear in the University College of London
database are shown in italic (we assume they are unique to the country in which they are reported). Nucleotide
numbering as suggested by Yamamoto et al. (54). In accordance with LDLR variants reported in LOVD-Select
Database (http://www.ucl.ac.uk/ldlr).