Jun 23, 2025 Current Resources

Impact of 2014 Japanese practice guidelines on treatment patterns in patients with myasthenia gravis: an insurance claims database study

Shingo Konno1,
Akiyuki Uzawa2,
Makoto Samukawa3,
Hiroshi Todaka4,
Izumi Mishiro4,
Kentaro Taki4,
Céline Quelen5,
Adrianna Czubin6,
Renata Majewska6,
Kunihiko Shiraiwa4,
Yohei Ohashi4

¹Department of Neurology, Toho University Ohashi Medical Center, Tokyo, Japan
²Department of Neurology, Graduate School of Medicine, Chiba University, Chiba, Japan
³Department of Neurology, Kindai University Faculty of Medicine, Osaka, Japan
⁴UCB, Tokyo, Japan
⁵Putnam, London, UK
⁶Putnam, Kraków, Poland

Correspondence to Dr Yohei Ohashi; Yohei.Ohashi{at}ucb.com

Abstract

Objectives To evaluate changes in oral corticosteroid (OCS) use after the publication of the 2014 Japanese clinical practice guidelines for myasthenia gravis (MG).

Design Retrospective cohort study performed in three Japanese health insurance databases: the JMDC database between 2005 and 2021; the DeSC database covering the National Health Insurance (NHI) and the DeSC database covering the Late-Stage Elderly Healthcare Service (LSEHS) between 2014 and 2021. Achievement of OCS ≤5 mg/day was defined as ≥90 days of consecutive OCS ≤5 mg/day during follow-up, without any gap longer than 60 days between two consecutive claims. The time to achieve OCS ≤5 mg/day was estimated using Kaplan-Meier survival analysis.

Setting Real-world treatment setting in Japan.

Participants Patients aged ≥16 years with a record of MG (International Classification of Diseases 10th edition code: G70.0) with a serological test, who had baseline period ≥180 days before inclusion with no MG claim and who started immunotherapy (including OCS) within 90 days of MG diagnosis.

Outcome measures Prescription of OCS and other therapies for MG.

Results Overall, 811 patients were included. The mean age was 49 years in the JMDC, 61 years in the NHI and 80 years in the LSEHS. In the JMDC, the median time to achieve OCS ≤5 mg/day was significantly shorter (p=0.042; log-rank test) in patients included in 2015 or later (11.0 months) than in patients included before 2015 (17.9 months). The median time to achieve OCS ≤5 mg/day was shorter in the LSEHS (6.5 months) than in the JMDC (11.0 months) and the NHI (11.7 months).

Conclusions Faster tapering of the OCS dose was observed in patients starting treatment after the publication of the 2014 guidelines, although use of higher-dose OCS remained widespread after this date. This highlights the need to improve awareness of guidelines by healthcare providers in order to decrease the burden of higher-dose OCS.

Trial registration number Clinical Trials Registry (UMIN-CTR): UMIN000051155; Post-results.

Neuromuscular disease
EPIDEMIOLOGY
Drug Utilization
Electronic Health Records

Data availability statement

Data may be obtained from a third party and are not publicly available. The data that support the findings of this study are the property of JMDC Inc. and DeSC Healthcare. Restrictions apply to their availability, which were used under license for the current study, and so are not publicly available. Reasonable requests for access to the data can be addressed to the corresponding author, who will transfer them to JMDC or DeSC.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Strengths and limitations of this study

Relatively large sample size for a rare disease.
Standardised coding system for entering information into the database.
Absence of information on clinical variables.
Quality of coding unknown.

Introduction

Myasthenia gravis (MG) is a rare neuromuscular autoimmune disease characterised by muscle weakness and abnormal muscular fatigue in response to exertion.1 MG is caused by autoantibodies that bind to proteins involved in synaptic transmission at the neuromuscular junction. In most cases, these autoantibodies are directed against either the nicotinic acetylcholine receptor (AChR) or muscle-specific tyrosine kinase (MuSK).1 A nationwide epidemiological survey conducted in Japan in 2018 estimated the prevalence of MG to be 23.1 per 100 000 population with a male-to-female ratio of 1:1.15 and a median age of onset of 59 years.2 This estimated prevalence is substantially higher than that reported in a survey performed in 2006 (11.8 per 100 000).3 In particular, the proportion of MG patients aged ≥65 years has increased over recent years, from 7.3% in 1987 to 16.8% in 2006.3

If treated adequately, patients with MG can achieve a satisfactory quality of life (QoL) and have near-normal life expectancy.1 4 The current therapeutic strategy typically involves platform treatment with acetylcholinesterase inhibitors (AChEIs) with the addition of corticosteroids or immunosuppressants in patients who continue to present clinically relevant disease activity.5 6 Intravenous immunoglobulins (IVIg) or plasma exchange (PLEX) are used for the management of myasthenic exacerbations or crises, which may also necessitate intubation or non-invasive ventilation. Although the effectiveness of treatment protocols has improved over previous decades, many patients remain difficult to treat.5 6

Historically, higher-dose oral corticosteroids (OCS) have been the standard treatment option in Japan. However, the long-term use of higher-dose OCS may result in adverse events, such as osteoporotic fractures, depression and Cushing’s syndrome, which may impair QoL.7–10 With the emergence of novel immunosuppressants, revised Japanese clinical practice guidelines for MG were published in 2014.11 The treatment target in the Japanese guidelines is to maintain QoL and mental health, and to achieve minimal clinical manifestations12 with an OCS dose of ≤5 mg/day prednisolone equivalents (MM-5 mg).8 13 14 To achieve this target as rapidly as possible, it is recommended to initiate early fast-acting treatment (EFT),15 16 with IVIg, PLEX, intravenous methylprednisolone (IVMP) or a combination of these. Immunosuppressants approved for the treatment of MG in Japan are the calcineurin inhibitors (CNIs: ciclosporin and tacrolimus).17 It is recommended that CNIs should be initiated soon after, or concurrently with, the initiation of OCS.17

A series of studies from the Japan MG registry (JAMG-R) conducted between 2010 and 2021 have reported a steady increase in EFT strategies, with a concomitant increase in the proportion of patients achieving MM-5 mg and a decrease in the use of higher-dose OCS >20 mg/day.18 The most recent of these studies, which included 1710 patients, showed that, when using the recommended EFT strategy, around 60% of patients could achieve a treatment target of OCS ≤5 mg/day.18 However, these studies were performed in 20 centres with experience in treating large numbers of patients with MG, and no data are available regarding whether comparable success rates can be achieved in routine clinical practice across Japan.

Although the publication of the Japanese guidelines may have influenced treatment of MG, there is still no evidence that this has reduced use of higher-dose OCS. Studies comparing prescribing patterns before and after the introduction of the Japanese guidelines are needed. The availability of information from health insurance claims databases has facilitated the evaluation of medical practice in the real-world setting. Using such claims databases, we have now undertaken an analysis of MG treatment patterns in patients in Japan. The objective of this study was to evaluate the potential impact of the publication of the Japanese guidelines on OCS use compared with previous years in the real-world setting.

Materials and methods

Study design and patients

This was a retrospective cohort study performed in three Japanese health insurance claims databases, the JMDC Claims Database (JMDC), the National Health Insurance (NHI) database and the Late-Stage Elderly Healthcare Service (LSEHS) database. For the JMDC, the selection period lasted from 1 July 2005 to 31 December 2019 and was divided into an early period up until 31 December 2014, and a late period, from 1 January 2015, after the 2014 Japanese clinical practice guidelines for MG came into force.13 For the NHI and LSEHS, the selection period lasted from 1 October 2014 to 31 December 2019.

The index date was defined as the date of the first delivery of immunotherapies at, or within 90 days following, the first documented International Classification of Diseases 10th edition (ICD-10) diagnosis code for MG (G70.0) during the selection period. The immunotherapies of interest in this study were OCS, IVMP, PLEX, IVIg, CNIs, methotrexate, azathioprine, mycophenolate mofetil or eculizumab.

Eligible patients with MG were identified from the ICD-10 code G70.0 associated with any reimbursement claim. The inclusion criteria were documentation of a confirmed diagnosis of MG during the selection period, together with a documented claim for an immunotherapy within 90 days of the first MG diagnosis, a documented MG-related serological test (AChR antibodies, MuSK antibodies) during the baseline period of 180 days before the index date or at the index date, age ≥16 years at the index date and presence of the patient in the database throughout the 180-day baseline period. The exclusion criteria were a claim for immunotherapies covering a period of >90 days any time until up to 90 days before the index date, a claim for any MG treatment (medication, radiation therapy for thymoma or thymectomy) any time until 90 days before the index date, or the absence of a visit with an MG diagnosis claim in the 6 months after the index date.

Patients were followed up from the index date until the end of the study (31 December 2021), end of insurance enrolment or until they died, whichever occurred first. The total study period thus lasted from 1 January 2005 to 31 December 2021. The study design is illustrated in online supplemental figure 1.

Supplemental material

Data sources

Data were retrieved from three Japanese health insurance claims databases, which are summarised below. Further information can be found in the Inventory of Japanese databases for clinical and pharmacoepidemiology research.19

JMDC Claims Database

The JMDC is a large claims database covering social health insurance of salaried workers and their dependents aged <75 years provided by JMDC. It contains inpatient, outpatient and pharmacy claims representing around 19 million cumulative beneficiaries in Japan since 2005. The database includes longitudinal, anonymised data on disease class, identified from ICD-10 disease codes and medication prescription, classified by anatomical therapeutic chemical (ATC) class and medical procedures. The database also contains information from annual health check-ups for certain beneficiaries, which includes height and weight.

NHI database

The NHI database is part of the DeSC database, a new administrative claims database developed by DeSC Healthcare. It contains information on around 15% of all the beneficiaries of NHI, which covers individuals <75 years old who are unemployed, self-employed or retired, and their dependents since 2014. The database includes similar information to the JMDC.

LSEHS database

This database is also part of the DeSC database and contains similar information to the NHI. The LSEHS contains data specifically on individuals aged ≥75 years covered by the Advanced Elderly Medical Service System for elderly people provided by the Japanese government. Data on around 17% of all LSEHS beneficiaries in Japan are available.

Study variables

Demographic variables documented were age, gender and employment status. All deliveries of MG treatments, including immunotherapies (as listed above) and AChEIs, were documented. These were identified from the relevant ATC code in reimbursement claims. The Charlson Comorbidity Index (CCI)20 was calculated from ICD-10 codes during the baseline period. Obesity at baseline was defined by a body mass index ≥30 kg/m² documented during the year preceding the index date or up to 90 days after the index date. The size of the facility was documented and assigned to one of seven bed size categories (unknown, 0–19 beds, 20–99 beds, 100–199 beds, 200–299 beds, 300–499 beds and 500+ beds). No data imputation was performed. For all study variables (including patient characteristics), missing data were presented as a category ‘missing’ or ‘unknown’.

Exposure to MG treatments

The initial MG treatment was defined as any immunotherapy (as defined above) prescribed at, or within 90 days following, the index date. Prescription of OCS was calculated as the estimated daily dose from the number of tablets prescribed and the period covered by the prescription and converted into prednisolone dose equivalents according to the equivalence table proposed by Asare.21 The daily dose was then classified into the four following categories: 0 mg; >0 mg to 5 mg; >5 mg to 20 mg and >20 mg.

Outcomes

The primary outcome variable was achievement of an OCS dose regimen of ≤5 mg/day (OCS ≤5 mg/day) for ≥90 consecutive days and without any gap ≥60 days between two consecutive claims. Discontinuation of OCS for ≥60 days or definitive discontinuation without resumption of OCS (OCS 0 mg/day) was considered to indicate achievement of the ≤5 mg/day target. The date each patient achieved OCS ≤5 mg/day was documented. The proportion of patients achieving the endpoint at each time point was calculated, and the time to achieve OCS ≤5 mg/day estimated using Kaplan-Meier survival analysis. These outcomes were determined separately for patients in each of the three cohorts (JMDC, NHI and LSEHS cohorts) and separately for patients in the JMDC early period and late period cohorts (with an index date before and after 2014 when the Japanese guidelines were published).13

Results

Study population

Over the 174-month selection period, 3553 patients with a confirmed diagnosis of MG were identified in the JMDC, of whom 516 met inclusion criteria and 314 were eligible. For 63 (20.1%) of these patients, the index date was before 31 December 2014 (early period). For the remaining 251, the index date was in 2015 or later (late period). In the NHI, 2787 patients with a confirmed diagnosis of MG were identified during the 63-month selection period, of whom 372 patients met inclusion criteria and 239 were eligible. In the LSEHS, 3201 patients were identified, of whom 397 met inclusion criteria and 258 were eligible. A patient flow diagram is presented in figure 1.

Figure 1

Patient flow diagram. JMDC, Japan Medical Data Center; LSEHS, Late-Stage Elderly Healthcare Service; MG, myasthenia gravis; NHI, National Health Insurance.

The characteristics of the study population are presented by cohort in table 1. The proportion of women was 46% in the overall JMDC and NHI cohorts and 59% in the LSEHS cohort. The mean age was lowest in the JMDC cohort (49 years), followed by the NHI cohort (61 years), and was highest in the LSEHS cohort (80 years). The median CCI was 1.0 in the JMDC and NHI cohorts, and higher (2.0) in the LSEHS cohort. The consultation on the index date when the first immunotherapy was prescribed typically took place in large hospitals (with >500 beds in around two-thirds of all cases).

Table 1

Patient characteristics at baseline in the JMDC, NHI and LSEHS study cohorts

Initial treatments of MG

During the first 90 days following the index date, the most frequently delivered treatments were OCS. In the JMDC, 264 patients (84.1%) were delivered OCS (all doses combined) during this period (table 2). The proportion of patients receiving OCS decreased slightly from 90.5% in the early period cohort to 82.5% in the late period cohort. The proportion of patients receiving OCS was 83.3% in the NHI cohort and 78.7% in the LSEHS cohort (table 2). During the initial 90-day period, AChEIs were delivered to 60.5% of patients in the JMDC cohort, with no major differences in prescription rates across cohorts or between early and late period cohorts. In addition, CNIs were delivered to 38.5% of patients in the JMDC cohort with an increase from 27.0% during the early period to 41.4% during the late period (table 2). Delivery of CNIs was also higher in older patients in the LSEHS cohort (44.2%) (table 2). Only a minority of patients received intravenous treatments during the initial treatment period (for the JMDC cohort overall, 16.9% received IVMP, 10.2% received IVIg and 7.3% underwent PLEX). No major differences in the proportion of patients receiving intravenous treatments were observed between the early and late period cohorts, except for IVIg, for which use increased during the late period (table 2). No major differences in rates of delivery of these intravenous treatments were observed between the three cohorts. No patients received eculizumab.

Table 2

Treatments of myasthenia gravis in the JMDC, NHI and LSEHS study cohorts

Maintenance treatments of MG

The patterns of oral treatments (OCS, AChEIs and CNIs) delivered during the first year, as well as the proportion of patients receiving them, were very similar to those observed during the 90-day initial treatment phase (table 2). However, the proportions of patients who had received intravenous treatments during the first year had increased compared with the initial treatment period. Use of IVMP increased by around 50% (eg, from 16.9% to 24.2% in the JMDC cohort), with smaller increases observed for IVIg or PLEX (table 2). No major between-time-period differences in the JMDC cohort or major between-cohort differences were observed for these intravenous treatments during the first year (table 2).

Time to achieve the OCS dose target of ≤5 mg/day

In the overall JMDC cohort, the proportion of patients achieving OCS ≤5 mg/day was 48.5% (95% CI: 42.7% to 54.7%) at 12 months and 65.5% (95% CI: 59.8% to 71.2%) at 24 months, with a median time to target of 13.0 (95% CI: 9.8 to 16.1) months. The median time to achieve OCS ≤5 mg/day was significantly shorter (p=0.042; log-rank test) in the late period cohort (11.0 months) than it was in the early period cohort (17.9 months). Time to achieve OCS ≤5 mg/day in the different cohorts is presented in the form of Kaplan-Meier survival curves in figure 2. Although time to achieve OCS ≤5 mg/day was very similar between the JMCD and NHI cohorts, median time to achieve this target in the LSEHS cohort was shorter (6.5 months) with a significant difference in time between the three cohorts (p<0.001; log-rank test).

Figure 2

Time to achieve OCS ≤5 mg/day. Data are presented with their 95% CIs in square brackets. JMDC, Japan Medical Data Center; LSEHS, Late-Stage Elderly Healthcare Service; NHI, National Health Insurance; OCS, oral corticosteroids.

OCS dose regimen over time by class

Shifts to a lower dose of OCS occurred through the follow-up period, as well as OCS discontinuation. Some increases in OCS dose were also observed, but these were much rarer than dose reductions. Patterns in the different cohorts evaluated are illustrated as Sankey plots in figure 3. In patients initially treated with OCS in the JMDC early period cohort, the proportion of patients receiving OCS ≤5 mg/day rose from 35% at week 24 to 42% at week 48, 40% at week 72 and 49% at week 104. Concomitantly, the proportion of patients receiving OCS >20 mg/day fell from 26% at week 4 to 12% at week 24 and 4% at week 48, remaining low thereafter. In patients in the JMDC late period cohort, more patients received OCS ≤5 mg/day in the first 4 weeks immediately following the index date (23% vs 9% in the JMDC early period cohort), and fewer patients received OCS >20 mg/day (19% vs 26%, respectively). In the NHI cohort, the proportion of patients receiving OCS ≤5 mg/day rose from 36% at week 24 to 48% at week 104, and the proportion receiving OCS >20 mg/day fell from 22% at week 4 to 9% at week 24 and remained ≤6% thereafter. The LSEHS cohort included the most patients using OCS ≤5 mg/day: 44% of patients by week 4, and 56% by week 104. In parallel, the proportion of patients in this cohort receiving OCS >20 mg/day was low: 14% for week 4, falling to ≤5% by week 12 and subsequently to ≤3% for the remainder of the study. In all cohorts, 16%–19% of patients receiving OCS as initial treatment were no longer receiving them at weeks 100 and 104 (figure 3).

Figure 3

Oral corticosteroid (OCS) dose class over the first 2 years of follow-up. ■: No OCS; ■: OCS >0–5 mg; ■: OCS>5–20 mg; ■: OCS>20 mg. Percentages are calculated with respect to the total number of patients in the cohort. JMDC, Japan Medical Data Center; LSEHS, Late-Stage Elderly Healthcare Service; NHI, National Health Insurance.

Discussion

10 years after the publication of Japanese guidelines for MG,13 prescriptions for OCS in general, and for moderate-to-high dose regimens (>5 mg/day) in particular, remain high in the real-world setting in Japan. In the JMDC and NHI cohorts, >80% of patients received OCS during the initial 90-day treatment phase, and >50% were still receiving an OCS 2 years later, at a dose regimen of >5 mg/day for over half of these. Nonetheless, faster tapering was observed in patients starting treatment after the Japanese guidelines were published, compared with patients starting treatment in the early period, with a median time to achieve OCS ≤5 mg/day falling from 18 to 11 months. Moreover, the proportion of patients who reached this target by 12 months rose from 39% to 51%. In addition, in the late-period cohort, OCS doses declined progressively over follow-up. However, in the early-period cohort, some fluctuations were observed, which may reflect increases in OCS dose following disease recurrence. These findings may indicate gradual appropriation of the Japanese guidelines by healthcare providers.

In the LSEHS cohort, composed exclusively of people aged >75 years, the proportion of patients who received OCS as an initial treatment was similar to that observed in the JMDC and NHI cohorts. However, 44% of the LSEHS cohort with initial OCS prescriptions at Week 4 received ≤5 mg/day (compared with 23% in the JMDC cohort and 28% in the NHI cohort). In addition, dose tapering in the LSEHS cohort was more rapid (median time-to-target: 6.5 months), and a higher proportion achieved the treatment target after 1 year (62.6%), than in the other two cohorts. The more cautious prescription of OCS in elderly patients with MG, and closer adherence to the Japanese guidelines, may reflect awareness and perceptions of healthcare providers of the risks of these medications in the elderly, for example, relating to osteoporotic fracture or depression.7 9 Furthermore, patients with late-onset MG are more likely to achieve MM-5 mg when receiving immunotherapies than patients with early-onset MG.16 22

The strengths of this study include the relatively large sample size included for a rare disease. To our knowledge, this is one of the largest health insurance database studies of patients with MG to have been performed in Japan. The three source cohorts contained around 9500 individuals with an MG diagnosis, which, assuming no overlap between the cohorts, corresponds to around one-third of the estimated 29 000 people with MG in Japan.2 Of these, 811 who received a first documented diagnosis of MG and were treated during the selection period were considered incident cases. Our findings were also consistent across the three cohorts. A major limitation is that the databases contain no information on clinical variables such as disease severity or QoL, so it was not possible to determine whether the patients were adequately controlled by their maintenance therapy or not, or whether they in fact achieved minimal manifestations. In addition, we only included patients with MG starting immunotherapy within 90 days of diagnosis, which may restrict inclusion to patients with more severe disease, and thus limit the extent to which the findings may be generalised. Furthermore, OCS prescription was documented from medication delivery, with no information available on whether it was actually taken. Adherence to OCS in other indications such as severe asthma or lupus is poor, possibly due to the limited acceptability of side effects.23–25 Since medication adherence in MG is also poor (39%–48%),26–28 disease control may be suboptimal, and healthcare providers might not encourage their patients to reduce the OCS dose. Moreover, in such claims databases, it is difficult to confirm whether a treatment such as IVIg qualifies as an EFT in the sense of the specific dosing regimen recommended in the Japanese guidelines, and to what extent it is used, due to limited information on the dose regimen. In addition, if individual beneficiaries in the JMDC leave employment because of their MG, they were considered censored in the survival analysis. Finally, the quality of coding of rare diseases in insurance claims databases may be suboptimal.

Initial treatment with OCS for MG is widespread in Japan, and the majority of patients receive these medications for long periods. According to a survey conducted in Japan in 2000, 64% of patients with MG received OCS.29 Studies in other countries have revealed much lower levels of OCS use. For example, in claims database studies from Germany and France, the proportion of patients receiving OCS in the year following diagnosis was 39.4% and 58.3%, respectively.30 31 In both countries, the platform immunosuppressant treatment was azathioprine. A cross-sectional study conducted in the JAMG-R has evaluated factors associated with achieving MM-5 status for over 6 months.32 The probability of achieving this treatment target was higher in patients who received low-dose OCS (maximum dose ≤20 mg/day) in the first 6 months compared with those receiving higher-dose OCS (maximum dose: 50–60 mg/day). The probability of achieving the target was also higher in those receiving a CNI or intravenous treatment. However, such treatments were used considerably more extensively in the JAMG-R (39% were prescribed an EFT) than in the present claims database study (around 10% received an intravenous treatment within 90 days).18 This may suggest that standards of care in everyday clinical practice in Japan may differ from those in specialist centres.18

The question thus arises as to why patients continue to receive long-term treatment with higher doses of OCS in Japan, being prescribed so often with no clear understanding of its drawbacks,17 33 suggesting a lack of awareness of the risks associated with long-term OCS use or of the availability of alternative treatments.34 This was one of the issues motivating the publication of Japanese guidelines in 2015, and since that time, additional treatment options for MG have become available.13 Japanese studies using the MG-QOL-15 have highlighted the dose-dependence of the impact of OCS on QoL in patients with MG. While lower doses of OCS (OCS ≤5 mg/day) generally correlate with better QoL, it was demonstrated that, even among patients achieving minimal manifestations status, those requiring higher-dose OCS reported significantly worse QoL.35 This specific finding underscores that minimising OCS dose is associated with better patient-reported QoL, even when good disease control (MM status or better) is achieved.8 10 Furthermore, a higher dose of OCS may represent a major risk factor for depression in MG.9 The Japanese guidelines also recommend the implementation of EFT strategies in order to achieve MM-5 status as soon as possible. The high proportion of patients using high doses of OCS for long periods, together with the limited use of EFT during the first 3 months (<30% of patients), indicates that the guidelines are still not being followed for many patients with MG. Awareness campaigns for healthcare providers about the risks associated with long-term higher-dose OCS in patients with MG may be useful to move the management of MG towards the optimal standard of care. Recommended EFT requires hospitalisation, and PLEX is more invasive than oral or intravenous injection therapy. For these reasons, it is possible that patients avoid EFT, and healthcare providers need to inform their patients about the potential benefits and risks.36 37 In addition, if centres lack the resources to provide EFT, it may be appropriate for them to refer patients to specialised centres where they can benefit from the recommended therapy.

One of the important unanswered questions raised by our findings is whether the changes in treatment patterns, or the type of treatment regimen, were associated with a change in disease activity. Insurance claims studies such as ours cannot answer questions, since clinical state is not documented in the databases. Nonetheless, recent data from the JAMG-R (in which 39% of patients received EFT) reported that 50% of patients achieved the target of MM-5 status.18 One possible interpretation of our findings would be that many patients with MG remained on higher-dose OCS for long periods because their disease could not be controlled otherwise. This interpretation would be important to verify. The slower tapering rates and prolonged dependency on OCS could also reflect a conservative approach to treatment adjustment, possibly due to concerns about exacerbating the symptoms of MG. Several factors might contribute to the difference in care between what is observed in our study and that recommended in current Japanese guidelines.14 For example, healthcare providers, especially those not specialising in neuroimmunology, may not be familiar with the Japanese guidelines. Additionally, it is possible that resources to provide EFT may be limited, leading healthcare providers to rely more heavily on OCS. Concerns about side effects and the complexity of managing certain immunosuppressant treatments or EFT could also deter their use. Finally, the findings may illustrate the need for better-tolerated and effective molecular-targeted drugs to help achieve MM-5 status and provide more treatment choices for patients with MG.38 39

In conclusion, continued widespread use of OCS in the treatment of MG in Japan, despite publication of the Japanese guidelines aimed at reducing exposure to OCS, highlights several key issues for the management of MG. First, persistent use of higher-dose OCS suggests that the recommendations to limit this strategy, as put forward in the Japanese guidelines, which are either unfamiliar to healthcare providers or not being followed. Ensuring awareness of the Japanese guidelines in healthcare providers should be a key target in the immediate future in order to address this shortfall. This could involve targeted educational programmes and workshops that not only explain the rationale behind the guidelines, but also provide practical training on managing MG with EFT. Second, since patients also have access to the Japanese guidelines, it is essential to educate them. By sharing information through patient support groups and online communities, patients can recognise the importance of guideline-based treatment and make their own choices. Third, increasing the availability and accessibility of alternatives to OCS could help reduce reliance on OCS. Finally, ongoing research into more effective and less burdensome treatment modalities such as molecular-targeted drugs will be essential to decreasing the burden of higher-dose OCS and to improving patient QoL.

Data availability statement

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

The study complied with all relevant international and national legislation on medical research and data privacy. In particular, it complied with the Declaration of Helsinki (Fortaleza Revision, 2013) and with the Japanese Act on the Protection of Personal Information (Act No. 57, 2003 and subsequent revisions). Furthermore, the study (RI2022016) was approved by the Ethics Review Committee for database studies of the Research Institute of Healthcare Data Science, Tokyo, Japan, on 20 April 2023. The requirement for ethics approval and informed consent is regulated by the Japanese Pharmaceuticals and Medical Devices Agency guidelines for conducting pharmacoepidemiological research using medical databases. In accordance with these guidelines, collecting patient consent to participate is not relevant to this study since the data came from an administrative healthcare insurance database in which all patient data are anonymised and the individual patients whose data are used cannot be identified.