Diagnosis and Management of Classical Homocystinuria in Brazil : A Summary of 72 Late-Diagnosed Patients

This study described a broad clinical characterization of classical homocystinuria (HCU) in Brazil. This was a cross-sectional, observational study including clinical and biochemical data from 72 patients (60 families) from Brazil (South, n1⁄4 13; Southeast, n1⁄4 1 Post-Graduation Program in Medical Sciences, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil 2 Laboratory of Basic Research and Advanced Investigations in Neurosciences (BRAIN), Hospital de Clı́nicas de Porto Alegre, Porto Alegre, Brazil 3 Post-Graduation Program in Genetics and Molecular Biology, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil 4 Complexo Hospitalar Materno-Infantil do Maranhão, São Luis, Brazil 5 Complexo Hospitalar Professor Edgard Santos, Salvador, Brazil 6 Universidade do Estado da Bahia, Salvador, Brazil 7 Hospital das Clı́nicas de Ribeirão Preto, Ribeirão Preto, Brazil 8 Centro Universitário Estácio de Ribeirão Preto, Ribeirão Preto, Brazil 9 Universidade de São Paulo, São Paulo, Brazil 10 Medical Genetics Service, Hospital de Clı́nicas de Porto Alegre, Porto Alegre, Brazil 11 Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil 12 Universidade Estadual de Campinas, Campinas, Brazil 13 Universidade Federal de Santa Catarina, Florianópolis, Brazil 14 Universidade Federal de Minas Gerais, Belo Horizonte, Brazil 15 Hospital da Criança Conceição, GHC, Porto Alegre, Brazil 16 Hospital Materno-Infantil Presidente Vargas, Porto Alegre, Brazil 17 Hospital de Apoio de Brası́lia, Brası́lia, Brazil 18 Escola Superior de Ciências da Santa Casa de Misericórdia de Vitória, Vitória, Brazil 19 Department of Psychobiology, Universidade Federal de São Paulo, São Paulo, Brazil 20 Universidade Federal do Pará, Belém, Brazil 21 Laboratory for Clinical Biochemistry and Metabolism, University Medical Center, Freiburg, Germany Received April 16, 2018, and in revised form June 07, 2018. Accepted for publication June 12, 2018. Corresponding Author: Giovana W. Hoss, Ms, Universidade Federal do Rio Grande do Sul, R. Ramiro Barcelos, 2350, Porto Alegre, 90040, Brazil. Email: giovana.weber@gmail.com Journal of Inborn Errors of Metabolism & Screening 2018, Volume 6: 1–6 a The Author(s) 2018 DOI: 10.1177/2326409818788900 journals.sagepub.com/home/iem This article is distributed under the terms of the Creative Commons Attribution 4.0 License (http://www.creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage). 37; Northeast, n 1⁄4 8; North, n 1⁄4 1; and Midwest, n 1⁄4 1). Parental consanguinity was reported in 42% of families. Ocular manifestations were the earliest detected symptom (53% of cases), the main reason for diagnostic suspicion (63% of cases), and the most prevalent manifestation at diagnosis (67% of cases). Pyridoxine responsiveness was observed in 14% of patients. Only 22% of nonresponsive patients on treatment had total homocysteine levels <100 mmol/L. Most commonly used treatment strategies were pyridoxine (93% of patients), folic acid (90%), betaine (74%), vitamin B12 (27%), and low-methionine diet þ metabolic formula (17%). Most patients diagnosed with HCU in Brazil are late diagnosed, express a severe phenotype, and poor metabolic control. Milder forms of HCU are likely underrepresented due to underdiagnosis.


Introduction
Classical homocystinuria (HCU; OMIM 236200) is an inborn error of methionine metabolism caused by deficient activity of cystathionine b-synthase (CbS; EC 4.2.1.22).Cystathionine b-synthase deficiency leads to massive accumulation of homocysteine and methionine and low levels of cysteine. 1,25][6] However, diagnosis and management of HCU is still a major challenge in developing countries, where newborn screening is unavailable and access to health-care services is often poor and unequal. 7rom a clinical standpoint, the classic signs of HCU are lens dislocation, thromboembolism, mental retardation, psychiatric disorders, osteoporosis, and marfanoid features. 8stablished treatment strategies include supplementation of pyridoxine (CbS cofactor), folic/folinic acid, betaine, and a methionine-restricted diet supplemented with an essential amino acids admixture free of methionine. 91][12] Usually, patients who respond to pyridoxine supplementation exhibit a milder phenotype and have a better prognosis. 8he worldwide prevalence of HCU is estimated to be 1:100 000. 13However, several mild and late presentation phenotypes have been described, [14][15][16] raising the issue of whether HCU is largely underdiagnosed.Newborn screening for HCU is performed in some countries with high incidence of the disease, such as Ireland and Qatar. 5,17Usually, newborn screening is performed by the determination of methionine on filter paper.However, this method has a large percentage of false negatives (up to 50%), since pyridoxine-responsive HCU may not present with hypermethioninemia in the first days of life. 18,19In Brazil, the prevalence of HCU is unknown, and the disease is not included in the National Neonatal Screening Program.
Brazil is a very large country with over 200 million inhabitants, and the country is characterized by intense admixture. 20,21The country also has one of the world's largest publicly funded health-care systems, the Unified Health System (Sistema U ´nico de Sau´de), which was established to provide equitable and comprehensive care to all users.However, betaine and the methionine-free amino acid formulation are not available through Unified Health System, and few centers across the country offer biochemical testing for diagnosis and management of this condition.Furthermore, several factors make diagnosis and management of HCU within the Brazilian health system a major challenge; hence, a substantial number of patients with this treatable condition are believed to remain undiagnosed and thus untreated.Within this context, the present study sought to establish a broad clinical characterization of HCU in Brazil through a survey on diagnosis and management of a representative patient population that is being followed at several centers nationwide.

Materials and Methods
The present study was approved by the local research ethics committee.Collection procedures for the study were conducted only after participants or their caregivers had agreed to take part in the investigation and provided written informed consent.

Patients
The study sample comprised 72 Brazilian patients with a diagnosis of HCU, from 60 different families.Diagnosis was in general made on clinical symptoms followed by clear elevations in homocysteine and methionine.Mutation analyses was performed in 35 of the 72 patients. 22Families from all 5 regions of Brazil were represented: South (n ¼ 13), Southeast (n ¼ 37), Northeast (n ¼ 8), North (n ¼ 1), and Midwest (n ¼ 1).Patients were recruited through contact with physicians involved in care and/or research activities at medical genetics centers across the country.A structured questionnaire containing queries regarding diagnosis, consanguinity, treatment strategies, metabolic control, and current health condition of patients with HCU was sent to 15 medical centers that had agreed to participate in the study.Clinical data regarding diagnosis were available only for 28 patients.
All 72 patients had delayed diagnoses: 62 patients had been diagnosed after clinical suspicion and biochemical findings consistent with HCU (hyperhomocysteinemia and hypermethioninemia), and the remaining 10 had been diagnosed on family screening.Each of the participating centers used a different protocol for determination of pyridoxine responsiveness.For the purposes of this study, patients were classified as responsive if they achieved homocysteine levels <50 mmol/L on pyridoxine alone or pyridoxine þ folic acid (regardless of the number of weeks since testing).All other patients were classified as nonresponsive to pyridoxine.
Target total homocysteine levels on treatment were set according HCU guidelines, 9 which were <50 mmol/L for pyridoxine-responsive patients and <100 mmol/L for nonresponsive patients.Treatment adherence was determined by the subjective impressions of the care team at each medical center.

Statistical Analysis
Statistical analysis was performed using SPSS for Windows, version 18.0 (SPSS Inc, Chicago, Illinois).Asymmetrically distributed variables were expressed as the median (range).The Mann-Whitney U test (continuous variables) or w 2 test (categorical variables) was used to assess between-group differences.Values of P < .05were considered significant.

Results
Of the 72 patients included, 55% were male.Two patients were already deceased at the time of inclusion in the study (presumably due to thromboembolic events).The patients belonged to 60 families; parental consanguinity was reported in 25 (42%) families.Median age at assessment 19 years.The youngest patient was aged 5, and the oldest was 45.Regarding pyridoxine responsiveness, 61 (85%) patients were classified as nonresponsive and 10 (13.8%) as responsive.In 1 patient, pyridoxine responsiveness was not reported/determined.

Journey to Diagnosis
The very first symptom noticed by families and/or physicians was visual impairment (mainly due to ectopia lentis) in 53% of the cases, followed by developmental delay (22% of cases), seizures (11% of cases), and isolated thromboembolic episodes (9.5% of cases).Table 1 reports clinical features at diagnosis for the sample as a whole and stratified by pyridoxine responsiveness.
The median time elapsed between symptom onset and diagnosis was 5 years (maximum, 34 years).One-third of the patients had 3 or more systems already affected at the time of diagnosis.The main clinical findings leading to investigation of HCU are represented in Figure 1.Eye disease, the most prevalent symptom at diagnosis (67% of cases), accounted for 63% of referrals for HCU investigation.

Management
Current clinical and biochemical data were available for 44 patients, of whom 7 were responsive and 37 were nonresponsive to pyridoxine.The median length of follow-up was 6 years (range, 0-27 years).Table 2 describes clinical manifestations and biochemical control in this group of patients.Ocular manifestations were more prevalent among pyridoxinenonresponsive patients (71% vs 97%, P ¼ .01).Ectopia lentis was the most common complication in our sample, affecting 91% of patients at the time of study inclusion.
Pyridoxine-responsive patients had significantly lower total homocysteine levels at study inclusion (P < .001).Only 22% of nonresponsive patients achieved target total homocysteine levels (<100 mmol/L) on treatment, while all responsive patients (n ¼ 7) had total homocysteine <50 mmol/L.Treatment adherence was reported as appropriate in 44% of patients.

Discussion
The present report provides the largest clinical profile of patients with HCU ever studied in Brazil to date.Clinical data of 72 patients (60 unrelated) from 15 medical genetics centers across Brazil were analyzed.Most families lived in the South and Southeast regions of the country.These regions are home to 57% of the country's population (http://www.ibge.gov.br/)and, compared to other regions of Brazil, have higher rates of access to health-care services and procedures.The fact that patients from other regions, particularly the North (n ¼ 1) and Midwest (n ¼ 1), were underrepresented suggests high rates of HCU underdiagnosis and/or limited access to care in these regions. 7 high proportion of pyridoxine-nonresponsive patients with HCU was found in our study (85%).Nonresponsive patients usually present a more severe phenotype, have more complications, and younger ages . 6,8,23In our study, this proportion exceeded rates described worldwide of approximately 50%. 8In countries where the proportion of nonresponsive patients is disproportionately high, such as Qatar and Ireland, specific genotypes are highly prevalent and account for these discrepancies . 24,25In a previous study of our group, we explored in more depth the genotype of 35 patients with HCU who are also included in the current study. 22ur results show there is no single molecular basis for the high prevalence of HCU nonresponsive phenotypes in Brazil.In fact, the most prevalent mutation in that study (p.Ile278Thr) is a pyridoxine-responsive mutation.Another genetic study of 14 Brazilian patients with HCU showed similar results. 26These findings rather suggest underdiagnosis of pyridoxineresponsive patients in Brazil who express milder phenotypes with fewer symptoms and later onset of clinical presentation. 6,8n our study, no significant differences in age at symptom onset or age at diagnosis were found between pyridoxine-responsive and nonresponsive patients; however, we believe our analysis was underpowered because of the low number of pyridoxineresponsive patients (n ¼ 10 vs n ¼ 61, respectively).
The high prevalence of ocular manifestations at diagnosis and their predominant contribution to diagnostic suspicion reinforce the importance of eye disease in HCU.Ectopia lentis is usually the earliest manifestation of HCU, occurring in half of all untreated patients by age 10 years and in over 90% by age 24 years. 8No significant difference in the prevalence of ocular manifestations at diagnosis was found between responsive and nonresponsive patients (75% vs 50%, P ¼ .10),although the lack of significance could also be explained by the small sample size.However, this finding is consistent with previous reports in the literature. 8keletal and neurologic manifestations were also highly prevalent at diagnosis, affecting more than half of patients of our cohort.The lower prevalence of vascular events at diagnosis is consistent with the natural history of HCU, in which such manifestations usually occur in general at a later age. 8owever, vascular disease may have been underrepresented due to the high lethality of thromboembolic events.The relative large time gap between symptom onset to diagnosis (median, 5 years) and the presence of multiple clinical manifestations attest to the difficulty in establishing a definitive diagnosis in these patients.
The main strategies and goals of HCU treatment have been recently established in the first guidelines for the diagnosis and management of CbS deficiency. 9Early diagnosis and early treatment are the key to prevent clinical manifestations and improve prognosis. 4,18In the present study, current clinical and biochemical data were obtained from 44 patients in treatment.There was a clear difference in metabolic control between the groups: Responsive patients had low and even near-normal homocysteine and methionine levels, whereas most nonresponsive patients had persistently high homocysteine levels (>100 mmol/L) despite multiple treatment strategies.This difficulty in achieving metabolic control in pyridoxinenonresponsive patients has been reported elsewhere. 18,27,28The high rate of betaine supplementation and comparatively low use of methionine-restricted diet in the nonresponsive patients may be attributed to several factors: (1)  difficulties in obtaining the metabolic formula, which is expensive and not provided by the Unified Health System in Brazil; 29 (2) low adherence to dietary methionine restriction, particularly in patients with a late diagnosis; and (3) limited training of health-care professionals in dietary prescription.In a European survey of 181 patients with pyridoxinenonresponsive HCU, 66% were on dietary treatment, that is, twice as many as in the present study. 30Homocystinuria guidelines clearly state that betaine should not be considered a first-line treatment for HCU-nonresponsive patients but used as adjunct treatment in those who cannot achieve target levels of homocysteine by other means. 9While dietary therapy dramatically reduces methionine and homocysteine and normalizes cysteine, betaine supplementation reduces homocysteine but increases methionine levels. 31,32In animal models, betaine is less effective at preventing clinical manifestations, 33 and its efficacy declines over time. 34In humans, there are no studies of the long-term efficacy of betaine supplementation alone in HCU.
In conclusion, this study provides the first broad clinical characterization of HCU in Brazil.All patients described here were late diagnosed, and most expressed a severe phenotype associated with nonresponsiveness to pyridoxine, early and multisystem clinical manifestations, and poor metabolic control.Limitations of this study include the underrepresented number of patients responsive to pyridoxine, and the number of patients coming from some regions of the country.We believe that our findings will contribute to the development of protocols and guidelines to improve diagnosis and management of HCU in Brazil.

Table 1 .
Classical Homocystinuria in Brazil: A Summary of Clinical Findings at Diagnosis.a a N ¼ 72, data expressed as the median (range) or percentage.In 1 patient, pyridoxine responsiveness could not be determined, data from this patient were used alone in the whole-group analysis.

Table 2 .
Classical Homocystinuria in Brazil: Clinical and Biochemical Profile of Patients on Treatment at the Time of Study Inclusion.a a n ¼ 44, data expressed as the median (range) or percentage; reference ranges: homocysteine, 5-15 mmol/L; methionine, 5-30 mmol/L.