Postauthorization safety study of betaine anhydrous

Patient registries for rare diseases enable systematic data collection and can also be used to facilitate postauthorization safety studies (PASS) for orphan drugs. This study evaluates the PASS for betaine anhydrous (Cystadane), conducted as public private partnership (PPP) between the European network and registry for homocystinurias and methylation defects and the marketing authorization holder (MAH). Data were prospectively collected, 2013–2016, in a noninterventional, international, multicenter, registry study. Putative adverse and severe adverse events were reported to the MAH's pharmacovigilance. In total, 130 individuals with vitamin B6 nonresponsive (N = 54) and partially responsive (N = 7) cystathionine beta‐synthase (CBS) deficiency, as well as 5,10‐methylenetetrahydrofolate reductase (MTHFR; N = 21) deficiency and cobalamin C (N = 48) disease were included. Median (range) duration of treatment with betaine anhydrous was 6.8 (0–9.8) years. The prescribed betaine dose exceeded the recommended maximum (6 g/day) in 49% of individuals older than 10 years because of continued dose adaptation to weight; however, with disease‐specific differences (minimum: 31% in B6 nonresponsive CBS deficiency, maximum: 67% in MTHFR deficiency). Despite dose escalation no new or potential risk was identified. Combined disease‐specific treatment decreased mean ± SD total plasma homocysteine concentrations from 203 ± 116 to 81 ± 51 μmol/L (p < 0.0001), except in MTHFR deficiency. Recommendations for betaine anhydrous dosage were revised for individuals ≥ 10 years. PPPs between MAH and international scientific consortia can be considered a reliable model for implementing a PASS, reutilizing well‐established structures and avoiding data duplication and fragmentation.


| INTRODUCTION
Reliable data on natural history, therapy, and long-term health outcomes of individuals with rare diseases, especially inherited metabolic diseases (IMDs), are often scarce. With more than 1600 distinct genetic diseases (IEMbase, http:// www.iembase.org/), IMDs form the largest group of rare diseases. Data fragmentation has significantly hampered the advancement of epidemiological, clinical, and pharmacological knowledge in this field. 1,2 Specialists on IMDs address this challenge through utilization of uniform accepted standards to achieve syntactic and semantic interoperability. 3 These efforts resulted amongst others in the foundation of the EU-funded European network and registry for homocystinurias and methylation defects (E-HOD).
The challenges for research-oriented studies in the field of IMDs also translate into clinical trials or the structural assessment of safety data for drugs. This is of special significance since medicines with an orphan drug status are common for IMDs and generally require the marketing authorization holder (MAH) to perform postauthorization safety studies (PASS) as specified by the Pharmacovigilance Risk Assessment Committee (PRAC) of the European Medicines Agency (EMA). Data protection compliant reutilization of data collected by observational patient registries is feasible and efficient, helping to fulfill a postauthorization measure in the framework of a public private partnership (PPP) and to avoid data fragmentation. This paper presents the "Cystadane ® (betaine anhydrous) Surveillance Protocol in cooperation with the E-HOD" (CSP) as an example for orphan drug surveillance as PPP between the MAH and an international scientific consortium focusing on rare diseases.  4 The E-HOD registry gathers comprehensive information on individuals with cystathionine beta-synthase (CBS) deficiency (synonym, classic homocystinuria; OMIM #23600), 5,10-methylenetetrahydrofolate reductase (MTHFR) deficiency (OMIM #236250), cobalamin C disease (CblC; OMIM #277400), cobalamin D disease (CblD; OMIM #277410), methionine synthase reductase deficiency (CblE; OMIM #236270), methionine synthase deficiency (CblG; OMIM #250940), cobalamin F disease (OMIM #277380), cobalamin J disease (OMIM #614857), methionine adenosyl transferase I/III deficiency (OMIM #250850), glycine N-methyltransferase deficiency (OMIM #606664); Sadenosylhomocysteine hydrolase deficiency (OMIM #613752), adenosine kinase deficiency (OMIM #102750), methylenetetrahydrofolate dehydrogenase deficiency (OMIM #617780); glutamate formiminotransferase deficiency (OMIM #229100), and intestinal folate carrier 1 deficiency (OMIM #600424). The most common diagnoses are CBS deficiency, CblC disease, and MTHFR deficiency. [5][6][7][8] CBS deficiency is for all patients further subdivided by vitamin B 6 responsiveness using the classification system recently introduced by Kožich et al. 7 E-HOD is implemented as an observational, noninterventional, multicenter registry study (November 2021: 71 participating sites). Upon enrollment patients receive a baseline visit, including detailed information on initial disease manifestation, mode of diagnosis, case confirmation, as well as family and individual medical history. Subsequent regular annual visits map the disease course, with unscheduled admissions at the hospital and fatal disease outcomes being covered by specific emergency and fatal disease course visits tailored to the context. At every visit the clinical and neurological phenotype is systematically recorded, as well as results from technical investigations, neuropsychological tests, quality of life assessments, dietary and drug treatment, and biochemical markers. At present (November 2021), the E-HOD registry contains comprehensive patient health data on 713 baseline visits, 1353 regular (annual) visits, as well as 62 emergency and 3 fatal disease course visits, averaging 3.0 visits per patient.

| Betaine anhydrous
Cystadane ® (active substance: betaine anhydrous) was granted an orphan drug designation on July 9, 2001, subsequently receiving market authorization on February 15, 2007 for the adjunctive treatment of homocystinurias, including CBS deficiency, MTHFR deficiency, and diseases of cobalamin cofactor metabolism. 9 As methyl group donor betaine activates betaine-homocysteinemethyltransferase, converting homocysteine to methionine as a bypass to the cobalamin-dependent methionine synthase, and is used in combinations with vitamin B 6 (pyridoxine), vitamin B 12 (cobalamin), folic acid, folinic acid, and for some diseases a low methionine diet.

| Public private partnership
The company Orphan Europe SARL (acquired by the Recordati Group in 2007 and continued as Recordati Rare Diseases since 2019) was granted the EU marketing authorization for the drug Cystadane ® on February 15, 2007 by the EMA. To guarantee the safe use of Cystadane ® , with special consideration of cerebral edema due to extreme hypermethioninemia as potential adverse drug reaction the company conducted the Registry for Cystadane-Homocystinuria (RoCH 10 ) as part of its Risk Management Plan. RoCH collected retrospective and prospective health data of 125 patients with homocystinuria enrolled at 29 centers in France and Spain (2009)(2010)(2011)(2012)(2013). RoCH was terminated in 2013 and replaced by CSP, a PPP between the newly established E-HOD consortium and Orphan Europe SARL in coordination with the EMA, to continue the postauthorization measure (http://www.encepp.eu/encepp/ viewResource.htm?id=40022) and to act along recommendations of the EMA and European Union Committee of Experts on Rare Diseases.

| Study organization
The study protocol for CSP was developed by the sponsor of the study in collaboration with the principal investigator of the E-HOD study center coordinating the registry (Heidelberg University Hospital [UKHD]) and under oversight of the E-HOD consortium and the PRAC of the EMA.
Putative (severe) adverse events [(S)AEs] were primarily reported directly to the pharmacovigilance of the sponsor to honor legal timelines, but were additionally also entered into the registry. The coordinating investigator at UKHD was responsible for administrative management and communication with the local investigators, providing assistance to participating clinical sites in study management and record keeping, and acted as primary liaison with the sponsor providing the agreed descriptive and analytical reports.

| Ethical and legal aspects
E-HOD was first approved by the local ethics committee of UKHD (Application no. S-525/2010, as the first amendment to E-IMD), followed by approvals of the respective ethics committees of further associated and collaborating partners contributing to the registry. CSP was first approved by the local ethics committee of UKHD (Application no. S-525/2010, as the second amendment to E-IMD), followed by approvals of the respective ethics committees of further associated and collaborating partners contributing to the PASS. All participating E-HOD partners had to obtain an additional ethics vote for the CSP study protocol. The patients' participation in the study was voluntary. Prospective written informed consent of participants and/or their parents and/or legal representatives, respectively, for both E-HOD and CSP were requested before inclusion of all eligible patients.

| Data protection and data management
CSP augmented the E-HOD data model by inclusion of additional variables for the start of Cystadane ® treatment, previous participation in Cystadane ® PASS studies, current pregnancy status, management of Cystadane ® treatment, and putative (S)AE. The data collection for CSP started in June 2013 with the acceptance of the CSP study protocol by the first local ethics committee and ended in December 2016 with the update of the Cystadane Summary of Product Characteristics (SmPC). 11 Data collection was performed using standardized paper-based case report forms which were subsequently entered into the E-HOD registry. E-HOD uses a custom-made registry based on a MySQL database and a Typo3 backend for user administration.
Data transfer between webserver and clients is protected by hypertext transfer protocol secure which enables encrypted communication and secure identification. Access to the patient registry is protected by secure individualized passwords. All entered data is pseudonymized, and users receive access limited to their institutional list of patients. Access to the complete dataset is only granted to the administrators (coordinator, data manager, biostatistical manager) for regular data plausibility checks and statistical analyses agreed by the consortium. Third parties, especially the MAH, do not have access to the primary data. Confidential personal information is handled in accordance to medical confidentiality, and the commitments of the federal and local data protection laws. Participants give written informed consent before data entry.

| Inclusion and exclusion criteria
Inclusion criteria for E-HOD are (1) confirmed diagnosis of a homocystinuria, methylation defect or folate disorder and (2) written informed consent for the E-HOD registry. Exclusion criteria for E-HOD are: (1) Metabolic derangement induced by any IMD not included in this study, congenital or acquired cobalamin (vitamin B 12 ) deficiency due to malnutrition or failure of absorption, or hyperhomocystinemia due to MTHFR gene polymorphism, (2) extreme low birth weight (<1500 g). Inclusion criteria for CSP are (1) participation in E-HOD, (2) being treated with Cystadane ® and (3) giving written informed consent for CSP. CSP does not have additional exclusion criteria to E-HOD.

| Statistical analysis
For the present study data analysis was limited to individuals with vitamin B 6 nonresponsive and partially responsive CBS deficiency, MTHFR deficiency, and CblC disease. CSP data sets of individuals with cobalamin metabolism disorders in small sample sizes [CblD (N = 3), CblE (N = 7), CblG (N = 5), as well as unclassified remethylation defects (N = 3) and CBS patients with unclassified vitamin B 6 responsiveness according to Kožich et al. 7 (N = 3)] and incomplete data sets regarding betaine therapy (N = 2) were excluded from the analysis.
Age in years, diagnostic delay, and duration of treatment were compared with Kruskal-Wallis and Mann-Whitney U test; when appropriate, the Holm method was used to adjust p values. Pearson-product moment coefficient was used to correlate two continuous variables. Multiple regression models were used to analyze the response of betaine anhydrous dose and homocysteine and methionine levels with respect to predictors sex, age group (< and ≥ 10 years) and disease. Variable selection was done with forward and backward step-wise variable selection with AIC. Tukey method was used for contrast analysis in multiple regression. When initial and last betaine and methionine levels were available, these were analyzed with linear mixed-effect (LME) models with the predictors sex and disease. Contrast analyses for LME were computed by estimated marginal means with Bonferroni adjustment, or paired t-test. Due to data protection, only month and year of birth were recorded and the day of birth imputed as the first day of the month. No other data was imputed. Missing data was handled by pairwise deletion. For all statistical analyses and figures R version 4.0.2 was used.

| Model for PASSs in PPP with a rare disease registry
The organization structure established to perform a PASS (CSP) as PPP with a rare disease registry (E-HOD) is shown in detail in Figure 1. CSP was conducted between July 2013 and December 2016, recruiting patients in 25 metabolic centers in 9 European countries (Austria 1, Croatia 1, Czech Republic 1, France 6, Germany 1, Italy 1, Spain 6, Switzerland 2, United Kingdom 6). Data were collected during baseline visits (N = 153), scheduled follow-up visits (N = 233), and emergency visits (N = 2). No fatal disease course was recorded during the observation period. Three CSP participants were recorded to also having been participants of the preceding PASS RoCH. 10
Cohort characteristics are summarized in Table 1

| Safety findings
Apart from supporting direct reporting of putative (S)AEs to the MAH's pharmacovigilance, the CSP also gathered information on drug safety in the registry that was used as an additional data source for the obligatory periodic safety update report compiled by the MAH during the reporting period of CSP. One new case of diarrhea and two follow-up visits to this case, with the events diarrhea and vision blurred were identified. Causality was reported as noncompliance by the patient. Events resolved without sequelae when the patient followed the diet and the dosage prescribed by the caring physician. One event of extreme hypermethioninemia (defined as plasma methionine levels equal to or exceeding 1000 μmol/L) was identified in a 14-year-old adolescent with vitamin B 6 nonresponsive CBS deficiency on 6 g betaine anhydrous (105 mg/kg/day) who showed behavioral problems and a diet noncompliance (total plasma homocysteine 242 μmol/L, plasma methionine 1056 μmol/L). No potential severe complications like cerebral edema were observed and thus the episode was not registered as (S)AE. In total, no new or potential risks were identified after review of safety data from the CSP registry.

| Betaine anhydrous dosage: Diseasespecific variations and nonadherence to recommendations
Median (IQR, range) duration of treatment with betaine anhydrous was 6.8 (3.2-8.3; 0-9.8) years and did not differ between the four disease groups (p = 0.4284; Kruskal-Wallis test; Table 1). Betaine anhydrous treatment was recommended to be adapted to body weight (maximum 150 mg/kg/day) for children below 10 years of age, and from then onwards to be prescribed using fixed doses (maximum 6 g/day). Since in individuals older than 10 years of age the dosage of betaine anhydrous continuously increased with age in parallel to increasing weight (Pearson; R = 0.296; p = 0.0065), we unraveled that age-dependent treatment recommendations were disregarded by participating metabolic centers in a significant number of individuals receiving betaine anhydrous treatment. Furthermore, the absolute betaine dose (g/day) increased by age (Pearson; R = 0.237; p = 0.0265) and exceeded the maximum recommended dose in 49% of participants ≥10 years, indicating that treating physicians continued to adapt doses relative to body weight according to the recommendations for individuals younger than 10 years of age. Noteworthy, for all ages prescribed amounts of betaine anhydrous depended on the diagnosis (Table 1, Figure 2; p = 0.0074; ANOVA): Compared to individuals with vitamin B 6 nonresponsive CBS deficiency doses were higher in MTHFR deficiency (p = 0.0057, Tukey HSD). In analogy, an excess of the betaine anhydrous dose above the maximum recommended dose (44%) also differed between diagnoses (likelihood ratio test; p = 0.0391), ranging from 31% of participants in B 6 nonresponsive CBS deficiency to 67% in MTHFR deficiency (Figure 3).

| Disease-specific biochemical response to specific therapy including betaine anhydrous
Betaine anhydrous was one component of the specific treatment in our study participants. Further therapy components, depended on the diagnosis, were: low protein diet, amino acid supplements, supplementation of micronutrition, and drugs. Therefore, the biochemical response could not be selectively analyzed for betaine anhydrous, but for the combined specific therapy including betaine F I G U R E 1 Organization structure for PASS as PPP with a rare disease registry. Model of organizing a PASS in cooperation with a rare disease registry: Healthcare providers (HCPs) (A) are free to only participate in the rare disease registry and consequently only enter data for the associated variables (1) or to additionally also participate in the PASS (B) and consequently enter data for an expanded set of variables (2) and directly report (S)AEs (3)   anhydrous. To analyze the treatment effect on total homocysteine and methionine concentrations before initiation of treatment (initial) and at last measurement (last) were recorded and analyzed with a LME model with the predictor variables treatment (initial, last), sex, and disease and the interaction treatment with disease. For total plasma homocysteine, we found significant main effects of disease and treatment (initial to last) as well as the interaction disease with treatment (Table 1). Overall, total plasma homocysteine concentration (mean ± SD; N = 79) before the start of specialized therapy (203 ± 116 μmol/L) was higher than at last visit (81 ± 51 μmol/L; p < 0.0001); however, differed between disease groups (p < 0.0001) but not between females and males (p = 0.2246). Noteworthy, total plasma homocysteine concentrations decreased in treated individuals with vitamin B 6 nonresponsive CBS deficiency (N = 39; p < 0.0001), vitamin B 6 partially responsive CBS deficiency (N = 6; p = 0.0038), and CblC disease (N = 27; p = 0.0163), while it remained virtually unchanged in individuals with MTHFR deficiency (N = 7; p = 0.9999, all p values were adjusted with Bonferroni correction; Figure 4). Overall, plasma methionine concentrations (mean ± SD; N = 46) did not reveal significant changes before (354 ± 290 μmol) and after start of treatment (283 ± 303 μmol/L; p = 0.3418); however, they differed between disease groups (p < 0.0004). As expected, the highest initial methionine concentrations (mean ± SD) were found in individuals with vitamin B 6 nonresponsive (436 ± 278 μmol/L) and vitamin B 6 partially responsive CBS deficiency (336 ± 237 μmol/L), while they were low in CblC disease (10 ± 6 μmol/L; Table 1; Figure 4). Although treatment with betaine anhydrous is known to bear the risk of further increasing plasma methionine concentrations to toxic levels in individuals with CBS deficiency, this was not observed during the study interval. Plasma methionine and total homocysteine concentrations of the last study visit did not correlate with the prescribed betaine anhydrous dose, neither for all diseases (Pearson; R = À0.186, p = 0.0864; R = 0.135, p = 0.1684, respectively) nor for single diseases.

| Comparison of RoCH and CSP studies
The RoCH study 10 and, in succession, the CSP study were designed to evaluate betaine anhydrous following market launch. Therefore, the same set of key variables were recorded, mainly retrospectively in RoCH and prospectively in CSP (  The RoCH evaluation 10 demonstrated an overall reduction of total plasma homocysteine levels of 29% following betaine anhydrous introduction and a methionine increase in CBS deficiency without exceeding the safety threshold of 1000 μmol/L. 10 Prospective follow-up of patients treated with betaine anhydrous did not allow to evaluate this effect selectively, but from the perspective of clinical everyday life of a combined and continuously adapted specific therapy.

| DISCUSSION
The presented CSP study serves as a successful model for PPPs between MAH and international scientific consortia focusing on rare diseases for implementing a PASS. Such collaborations allow recruitment of well-established patient cohorts and efficient data collection using already existing structures, thus avoiding data duplication and low enrollment turnouts due to burdening patients with multiple study protocols.
E-HOD highlighted that networking activities and the establishment of patient registries have a relevant impact for improving the health of patients with IMDs and for harmonizing diagnostic algorithms, therapy, long-term follow-up, and care on a European and international level. It described in detail the natural history including disease variants of more than six different IMDs, evaluated the impact of interventional and noninterventional parameters (including evaluation of newborn screening programs), and allowed the development of evidencebased care protocols being the basis for national guidelines and being translated into information brochures for patients, families, and healthcare professionals. [5][6][7][8]12,13 For rare diseases treated with orphan drugs requiring a PASS, the model of implementing a PPP between MAH and international patient registries was demonstrated to be successful in generating the needed cohort size and data quality. Apart from helping to enable the market introduction of critically needed drugs, the reuse of existing data also lowers the impact on patients and health infrastructure posed by data duplication and provides funding for the sustainability of research registries after the end of the initial funding period. Thereby, research registries will help to improve treatment and care in rare diseases like IMDs.
The CSP study presented has a similar cohort size and structure compared to the previously reported RoCH data 10 enrolled in France and Spain. It is representative for the selected diagnoses, but features a greater geographical coverage in Europe. As a valuable extension of the former RoCH study 10  recommendation, our evaluation could now analyze this practice in greater detail. We demonstrated continuous dose adaptation of betaine anhydrous to weight in most patients aged 10 years and older, exceeding the former maximum dosage recommendations by a factor of up to three. This was especially true for individuals with MTHFR deficiency and CblC disease (two diseases with hypo-methioninemia), but also for 30%-50% of the CBS-deficient patients (with hypermethioninemia). Noteworthy, this prescription practice in CBS deficiency did not result in reports of severe adverse effects associated with toxic methionine concentrations such as cerebral edema or signs of cerebral pressure. Plasma methionine concentrations in vitamin B 6 nonresponsive CBS-deficient patients remained stable in all but one patient under betaine anhydrous therapy due to further concomitant treatment, including restricted dietary intake of natural methionine containing protein. 8 Adding this information to the RoCH evaluation and further pharmacovigilance data collected by the sponsor, the CSP study data resulted in an update of the Cystadane SmPC in 2016, 14 introducing the observed clinical practice of weight-adapted dose titration of 100-200 mg/kg/ day depending on the therapeutic goal as new dosage recommendation. 11 Total homocysteine levels in plasma at last visit were lower than pretreatment in all diagnoses; in contrast, methionine levels remained stable in vitamin B 6 nonresponsive CBS-deficient patients under betaine anhydrous-containing specific treatment, but did not increase to toxic levels. Presumably this was supported by regular controls and adaption of dietary treatment if necessary.
In individuals with MTHFR deficiency and CblC disease, observed therapy did not normalize plasma methionine concentrations, supporting the notion of necessary regular laboratory surveillance and continuous adaptation of medication in these patients to maintain methionine in the normal range. 6 Patients with CblC disease having the highest betaine anhydrous dosages ≥10 years, usually as additional treatment to the high effective parenteral hydroxocobalamin, 5,6 might surprise and points on the necessity of regular dosage reevaluation of all treatment components in consideration of the metabolic laboratory to avoid under-and overdosages.

| LIMITATIONS
Not all E-HOD study sites participated in the CSP and as participation is voluntary and requires informed consent not all eligible patients participated in E-HOD and CSP. This might be a source of bias for the presented study. However, the implementation of the study as an observational study with none or very low impact on patients, makes refusal to participate low. Due to the different incidences and the observational design of the study the enrolled disease groups differ in size, with the MTHFR group being especially small. This could hamper the robustness of our results.
Observational studies collect real world data on heterogenous disease groups regarding disease severity, age, sex, and treatment components. Furthermore, often only data on the prescribed treatment is recorded without detailed information on individual treatment compliance. This might be a source of bias for the study. However, PASSs are designed and done to assess the medication (severe) adverse and desired effects in a real-world setting.
Analysis of metabolic laboratory parameters as paired data (initial and last value) was only possible for small cohorts, due to the retrospective nature of data collection of the initial level and the resulting fragmentation. This reduced the power of the analysis. Furthermore, data collection for CSP as extension of RoCH was not long enough to generate a meaningful sequence of longitudinal metabolic laboratory data. However, results from metabolic laboratory tests at last visit were comparable to the follow-up data collected in RoCH.

| CONCLUSION
The presented CSP study serves as a successful model for PPPs between MAH and international scientific consortia focusing on rare diseases for implementing a PASS using well-established structures and avoiding data duplication. The results of this study which observed a continuous dose titration relative to body weight without reported occurrence of severe side effects resulted in adaption of Cystadane dosage recommendations and thereby warrants a safe drug administration in patients with homocystinurias.
(CHAFEA agreement no. 2012 12 02). Ulrike Mütze was supported by Olympia Morata research fellowship of the Medical Faculty of the Ruprecht Karl University of Heidelberg (grant number F.206852). Viktor Kožich and Pavel Ješina received institutional support from the Ministry of Health (General University Hospital in Prague-RVO VFN 64165) and from the Czech Health Research Council (grant NV19-01-00307). The authors confirm independence from the sponsor; the content of the article has not been influenced by the sponsor. ETHICS STATEMENT E-HOD was first approved by the local ethics committee of UKHD (application no. S-525/2010, as the first amendment to E-IMD), followed by approvals of the respective ethics committees of further associated and collaborating partners contributing to the registry. CSP was first approved by the local ethics committee of UKHD (application no. S-525/2010, as the second amendment to E-IMD), followed by approvals of the respective ethics committees of further associated and collaborating partners contributing to the PASS. All participating E-HOD partners had to obtain an additional ethics vote for the CSP study protocol.