Incidence, Prevalence, Hospitalization Rates, and Treatment Patterns in Myasthenia Gravis: A 10-Year Real-World Data Analysis of German Claims Data

Abstract

Background

Myasthenia gravis (MG) is a rare chronic autoimmune disease caused by autoantibodies directed against postsynaptic antigens of the neuromuscular junction. Over the last decades, increasing incidence and prevalence rates have been reported. Epidemiological data on prevalence and incidence in Germany are lacking. Furthermore, the MG treatment landscape is rapidly changing due to the continued approval of novel monoclonal antibodies.

Method

This is a retrospective study assessing incidence, prevalence, and hospitalization rates of MG as well as treatment patterns in Germany over 10 years based on medical claims data covering 6.1 million insured persons.

Results

Between 2011 and 2020, the prevalence rate of MG increased from 15.7 to 28.2 per 100,000 person-years. The age-adjusted incidence rate was 2.8 per 100,000 person-years within the study period (95% confidence interval, 2.43–3.22) and decreased dramatically in 2020, the year of the COVID-19 pandemic. Similarly, the hospitalization rate fluctuated within the study period but reached an overall low of 8.3% in 2020 (mean hospitalization rate 11.5%). Treatment patterns showed that most MG patients are treated with base therapy. However, crisis intervention is necessary for 2–5% of MG patients, and therapeutic monoclonal antibodies, including rituximab and eculizumab, are increasingly used.

Conclusion

This is the first study on MG prevalence and incidence rates in Germany. Data show an increase in prevalence by 1.8-fold over 10 years. Decreasing incidence and hospitalization rates in 2020 hint at the impact of the COVID-19 pandemic. Treatment patterns in MG are changing with the advent of therapeutic monoclonal antibodies in this indication.

Introduction

Myasthenia gravis (MG) is a rare chronic autoimmune disease caused by antibodies directed against postsynaptic antigens of the neuromuscular junction [1, 2]. The disease is characterized by muscle weakness that worsens with muscular activity, which can result in a marked reduction in quality of life and potentially be life-threatening if respiratory muscles are affected [1, 3]. MG meets the criteria of an orphan disease with a prevalence of <5 per 10,000 people. However, data consistently shows a worldwide increase in prevalence and incidence over the last decades [4, 5]. A recent review reported incidence rates (IRs) ranging from 1.0 to 2.9 per 100,000 people and prevalence rates (PRs) between 10 and 35 per 100,000 people globally [6]. To date, no study on IR and PR in Germany has been published.

Treatment strategies for MG can be grouped into base therapy and escalation therapy [1, 7−10]. Depending on disease severity, base therapy includes symptomatic treatment with acetylcholinesterase inhibitors (ACIs) (e.g., pyridostigmine) and immunosuppressive agents, including glucocorticosteroids (CS) (e.g., prednisolone) and steroid-sparing agents such as azathioprine and mycophenolate mofetil (immunosuppressive therapy [IST]). If no adequate response to base therapy is reached, treatment can be escalated (escalation therapy) to monoclonal antibodies (e.g., rituximab, eculizumab) depending on the antibody status. Thymectomy is the surgical option to target the immunological pathway, although class 1 evidence exists only for AchR-ab-positive generalized MG [11]. Treatment of acute MG exacerbations and myasthenic crisis includes intravenous immunoglobulins (IVIgs) and the rapid removal of antibodies through plasma exchange (PLEX)/immunoadsorption (IA), and intensive care unit treatment. During the past years, the therapeutic landscape of MG is evolving rapidly. Three novel treatment agents targeting specific disease pathways (namely, targeted complement inhibition with eculizumab, ravulizumab, and FcRn-inhibition with efgartigimod) have been approved for the treatment of AchR-ab-positive generalized MG [10, 12−15]. The approval of other therapeutic monoclonal antibodies is expected [16]. To observe and monitor the change in MG treatment strategies, it is valuable to analyze the contemporary usage of available authorized drugs within informative data sources.

Medical claims data have been used to estimate the epidemiology and treatment patterns of MG across different countries [17−22]. However, to date, there is no published real-world data analysis of the German care patterns in MG. In Germany, medical claims data are collected by statutory healthcare insurances and have been used for epidemiological research in other indications for over 2 decades [23, 24]. This study aimed to investigate the IR, PR, and hospitalization rates (HR) of MG in Germany and to analyze treatment patterns over the past decade using medical claims data covering 6.1 million insured persons.

Methods

Data Source

An anonymized dataset of German claims data for approximately 6.1 million insured persons for 2010 to 2020 was provided by the Corporation for Efficiency and Quality in Health Insurance (GWQ ServicePlus AG, Gesellschaft für Wirtschaftlichkeit und Qualität bei Krankenkassen). Medical claims data consist of the billing codes that physicians, hospitals, pharmacies, and other healthcare providers submit to insurance companies. These data have the benefit of following a nationwide standardized format using a set of pre-established codes that describe specific diagnoses, procedures, and operations and drug prescriptions. Nearly every encounter a patient has with the medical system leads to the generation of a claim, creating a comprehensive and standardized source of information for each patient. However, the dataset does not include clinical details such as the results of autoantibody testing, electrophysiological testing, or disease patterns/severity.

Case Definition

First, the MG population was defined as all patients with at least one confirmed primary or secondary inpatient diagnosis of MG (ICD-10-GM code: G70.0) or two confirmed outpatient diagnoses in two different quarters. Second, only patients who received MG-related treatment were accepted further to verify the MG population. Patients were under MG-related treatment if they received a base therapy prescription (i.e., ACI, CS, or IST) in at least four different quarters during their individual observation period. Patients were considered newly diagnosed if they were insured for at least 1 year before their first confirmed MG diagnosis. Patients were counted as MG prevalent starting from the first diagnosis, independent of the first MG-related treatment. For the subsequent analyses, the above-defined population was further restricted to patients at least 18 years old (in the respective year of analysis) and with at least 1 year of observed data after their first confirmed G70.0 diagnosis.

Selection of MG-Related Drugs and Treatment

For the analysis of MG-related therapy and the grouping into treatment stages, the guideline of the German Society for Neurology's therapy was followed [7]. Three treatment stages were identified: base therapy, escalation therapy, and exacerbation/crisis intervention. Base therapy was further subdivided into patients that received ACI, CS/IST, or ACI plus CS/IST within a given year. Escalation therapy contained patients receiving rituximab or eculizumab treatment (efgartigimod and ravulizumab were only approved at the end of 2022 and, thus, do not fall within the survey period). Exacerbation/crisis intervention consisted of PLEX/IA, IVIg treatment, and patients undergoing ventilation procedures (intubation, machine-supported breathing, or oxygen supply). Nonpharmacological inpatient treatment was considered if a relevant diagnosis-related group was provided (e.g., related to ventilation) or G70.0 was given as the only main diagnosis during the respective visit. The HR was calculated for the years 2010–2020, including a subanalysis for hospitalizations with a length of stay ≥3 days.

Statistical Analysis

The age- and sex-specific incidence was calculated for 100,000 person-years (PY). PY were calculated: across all persons in the dataset, insurance days were summed up and divided by 365 for a given year. The yearly prevalence of MG was determined per 100,000 persons insured that year. The IR was calculated for the years 2011–2020. The PR was calculated for each year as the number of patients having their first diagnosis in 2011–2020 or before that. IR and PR were calculated in total, for males and females, and 4 age groups starting with patients aged 18–39 years. The mean and 95% confidence interval (CI) for each age-sex group were calculated across the respective observation period. The total mean and confidence interval were calculated across the age groups. Mean IR and PR across all age groups were age adjusted using the German 2011 standard population [25]. Hospitalization and treatment rates were calculated with respect to the total number of prevalent MG patients in the respective year. All data processing and analysis were done on the software R-Studio 2022.02.0 using R, version 4.1.3.

Results

Population Characteristics

The dataset contained 2,741 MG patients ≥18 years (male:female, 0.96:1) over 10 years with at least 1 year of follow-up data. Applying the two-step approach described above, this population was narrowed down to 2,317 patients with a minimum of one inpatient or two outpatient G70.0 diagnoses and, ultimately, to 1,493 patients with confirmed MG treatment. The MG population included 676 females (mean age at diagnosis 56.6 years) and 817 males (63.8 years).

Age- and Sex-Specific Incidence and Prevalence

The mean IR between 2011 and 2020 was 2.1 per 100,000 PY (age adjusted: 2.8), with 2.4 (adj: 3.1) and 1.9 (adj: 2.5) for male and female patients, respectively. In patients aged ≥80 years, the mean IR was 11.6-fold higher (9.3) compared to patients aged 18–39 years (0.8). Male patients had a higher IR than female patients in every age group except for patients aged 18–39 years (0.6 in males vs. 1.1 in females) (Table 1). The yearly IR did not follow a clear trend between 2011 and 2020. A maximum of 2.6 was reached in 2014 and 2017. After 2017, the IR declined dramatically to 1.5 in 2020 (shown in Fig. 1a).

Table 1.

Summary statistics for estimated MG IR per 100,000 patients-years between 2011 and 2020

Age group	Mean annual population			Cases, N			Incidence per 100,000 patient-years (95% CI)
	M	F	total	M	F	total	M	F	total
0–17 years	410,434	384,000	794,434	0	0	0	0	0	0
18–39 years	664,128	636,491	1,300,619	36	68	104	0.6 (0.33–0.79)	1.1 (0.63–1.57)	0.8 (0.53–1.05)
40–59 years	670,569	673,837	1,344,406	125	116	241	1.9 (1.53–2.31)	1.7 (1.24–2.24)	1.8 (1.48–2.14)
60–79 years	305,765	302,123	607,888	277	157	434	9.3 (7.63–10.99)	5.3 (4.34–6.36)	7.3 (6.44–8.24)
80+ years	45,156	60,978	106,134	56	39	95	12.7 (7.29–18.03)	6.9 (3.5–10.24)	9.3 (6.02–12.66)
Total	2,096,052	2,057,429	4,153,481	494	380	874	2.4 (2.08–2.76)	1.9 (1.51–2.23)	2.1 (1.86–2.42)
Adj. Rate							3.1 (2.61–3.64)	2.5 (2–2.92)	2.8 (2.43–3.22)

CI, confidence interval; M, male; F, female; Adj. Rate, age adjusted to standard German population 2011.

Fig. 1.

IR and PR of MG (a) and HR of prevalent MG patients (b).

The mean PR between 2011 and 2020 was 22.2 per 100,000 patients (adj: 29.6). The yearly PR of MG increased 1.8-fold between 2011 and 2020, from 15.7 to 28.2 (shown in Fig. 1a). Male patients had a higher PR than female patients (23.9 vs. 20.5) (adj: 31.5 vs. 26.2). The mean PR between 2011 and 2020 was 19.3-fold higher in patients aged ≥80 years (135.3) compared to patients aged 18–39 years (7.0). Female patients had a higher PR in the age groups ≤60 years compared to male patients. This difference was most pronounced in the age group 18–39 years, with a 2.7-fold higher female PR than males. In contrast, male patients had a 2-fold higher PR above 60 years compared to females (Table 2).

Table 2.

Summary statistics for estimated MG PR per 100,000 patients between 2011 and 2020

Age group	Mean annual population			Cases, N			Prevalence per 100,000 patients (95% CI)
	M	F	total	M	F	total	M	F	total
0–17 years	410,434	384,000	794,434	0	0	0	0	0	0
18–39 years	664,128	636,491	1,300,619	25	64	89	3.8 (3.20–4.48)	10.3 (8.63–11.95)	7.0 (5.89–8.09)
40–59 years	670,569	673,837	1,344,406	116	152	268	17.2 (14.93–19.39)	22.5 (19.40–25.56)	19.8 (17.22–22.42)
60–79 years	305,765	302,123	607,888	270	146	416	88.3 (80.03–96.51)	47.6 (42.74–52.46)	68.1 (61.82–74.32)
80+ years	45,156	60,978	106,134	91	57	148	195.7 (178.63–212.75)	91.6 (83.16–99.96)	135.3 (122.14–148.52)
Total	2,096,052	2,057,429	4,153,481	502	419	921	23.9 (20.57–27.17)	20.5 (17.44–23.58)	22.2 (19.04–25.4)
Adj. Rate							31.5 (28.56–34.44)	26.2 (23.37–29.11)	29.6 (26.69–32.52)

CI, confidence interval; M, male; F, female; Adj. Rate, age adjusted to standard German population 2011.

HR and Treatment Patterns

The mean HR between 2010 and 2020 was 10.7% of all prevalent MG patients. Between 2010 and 2017, the HR of MG patients fluctuated around a mean of 12.3%, with a peak of 12.9% in 2017. After 2017, HR gradually declined, with an overall low of 8.1% in 2020. The HR with length of stay of ≥3 days developed parallel to the absolute HR (shown in Fig. 1b).

Base therapy with ACI, CS, IST monotherapy, or combination therapy was prescribed to 93.1% of prevalent MG patients in 2010 (Table 3). The rate of prescriptions declined over the observation period to 83.5% in 2020. ACI, combined with CS/IST, was the most prescribed base therapy across all study years, with rates between 52.6% (2010) and 43.6% (2020). Similarly, for the second most prescribed base therapy, ACI monotherapy, the rate increased from 28.3% in 2010 to 30.2% in 2012 and then decreased to 24.9% in 2020. CS/IST monotherapy rate ranged between 12.1% in 2010 and 15.0% in 2020.

Table 3.

Treatment of MG patients between 2010 and 2020

Treatment	2010		2011		2012		2013		2014		2015		2016		2017		2018		2019		2020
	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%
HR	61	12.3	63	10.5	60	9.1	76	10.4	99	12.1	92	10.6	102	10.7	136	12.9	120	10.7	119	10.0	99	8.1
HR LOS ≥3	55	11.1	58	9.7	53	8.1	68	9.3	92	11.3	78	9.0	85	8.9	119	11.3	104	9.3	100	8.4	82	6.7
Thymectomy	6	1.2	8	1.3	4	0.6	5	0.7	6	0.7	9	1.0	13	1.4	11	1.0	11	1.0	13	1.1	9	0.7
Base therapy	460	93.1	556	92.8	594	90.3	665	90.8	733	89.8	767	88.1	821	86.1	903	85.6	963	85.7	1,014	85.1	1,021	83.5
ACI	140	28.3	172	28.7	199	30.2	211	28.8	225	27.6	223	25.6	249	26.1	263	24.9	278	24.7	291	24.4	305	24.9
CS/IST	60	12.1	84	14.0	84	12.8	107	14.6	120	14.7	155	17.8	149	15.6	173	16.4	174	15.5	198	16.6	183	15.0
ACI+ CS/IST	260	52.6	300	50.1	311	47.3	347	47.4	388	47.5	388	44.5	423	44.3	467	44.3	511	45.5	525	44.0	533	43.6
Escalation therapy	3	0.6	6	1.0	4	0.6	4	0.5	5	0.6	11	1.3	13	1.4	23	2.2	23	2.0	23	1.9	21	1.7
Rituximab	3	0.6	6	1.0	4	0.6	4	0.5	5	0.6	11	1.3	13	1.4	23	2.2	21	1.9	21	1.8	18	1.5
Eculizumab	0	0.0	0	0.0	0	0.0	0	0.0	0	0.0	0	0.0	0	0.0	0	0.0	2	0.2	2	0.2	3	0.2
Exacerbation/crisis intervention	12	2.4	12	2.0	21	3.2	28	3.8	34	4.2	35	4.0	44	4.6	44	4.2	45	4.0	39	3.3	41	3.4
PLEX/IA	6	1.2	3	0.5	5	0.8	9	1.2	8	1.0	7	0.8	10	1.0	10	0.9	8	0.7	9	0.8	6	0.5
Mg	9	1.8	10	1.7	20	3.0	21	2.9	30	3.7	33	3.8	38	4.0	36	3.4	40	3.6	31	2.6	37	3.0
Ventilation procedure	5	1.0	10	1.7	8	1.2	6	0.8	12	1.5	17	2.0	9	0.9	12	1.1	15	1.3	16	1.3	13	1.1
Total escalation + exacerbation/crisis	15	3.0	17	2.8	23	3.5	32	4.4	38	4.7	41	4.7	55	5.8	61	5.8	64	5.7	54	4.5	57	4.7
Total	494	100.0	599	100.0	658	100.0	732	100.0	816	100.0	871	100.0	954	100.0	1,055	100.0	1,124	100.0	1,192	100.0	1,223	100.0

n, number of patients; % percent of total; LOS, length of stay; ACI, acetylcholinesterase inhibitor; CS, glucocorticosteroids; IST, nonsteroidal anti-inflammatories; PLEX, plasma exchange; IA, immunoadsorption; Ig, immunoglobulin.

Within this dataset, rituximab was first prescribed in 2010 at a rate of 0.6% of prevalent MG patients. The rate of rituximab treatment increased to 1.3% in 2015 and peaked at 2.2% in 2017. Eculizumab was first prescribed to MG patients in this dataset in 2018. The treatment rate of 0.2% remained stable until 2020. The total rate of patients undergoing escalation therapy increased from 0.6% in 2010 to 2.2% in 2017, followed by a slight decline to 1.7% in 2020.

The rate of exacerbation/crisis intervention developed as follows: IVIg treatment rate was 1.8% of prevalent MG patients in 2010 and 3.0% in 2020. Between 2014 and 2016, higher rates of up to 4.0% were observed. PLEX/IA fluctuated between a minimum of 0.5% in 2011 and a maximum of 1.3% in 2013. The PLEX/IA rate was 0.5% in 2020. The rate of necessary ventilation procedures ranged between 0.8% in 2013 and 2.0% in 2015.

Discussion

To the best of our knowledge, this is the first study on the PR, IR, HR, and longitudinal treatment patterns of MG in Germany. The age-adjusted mean PR in our study was 29.6 per 100,000 patients, increasing 1.8-fold from 15.7 in 2011 to 28.2 in 2020. A systematic review by Carr et al. [4] of population-based epidemiological studies (mostly European studies from 1968 to 2000) in MG showed a pooled IR between 0.18 and 2.1, and a recent review reported worldwide PRs between 10 and 35 per 100,000 people [6]. Thus, our findings align with previous studies showing comparable and increasing PRs during the last decade. Our data also shows that the IR for females in early-onset MG is higher than that in male patients, whereas late-onset MG is more common in males. This bimodal distribution was observed in 5 of 14 studies by Carr et al. [4] and in a Polish cohort [18]. The absolute gender ratio, however, showed a male predominance for incidence and prevalence. This finding contrasts with the previously reported female predominance and could be due to the skewed population structure in the dataset. The data used for this study stem from the Corporation for Efficiency and Quality in Health Insurance which is an association of company health insurance funds (Betriebskrankenkassen) which opened to the public in 1996 and might therefore insure a younger, healthier, and male-dominated population since men in Germany are more likely to be employed than women [26].

The age-adjusted mean IR was 2.8 per 100,000 PY between 2011 and 2020, dramatically decreasing in 2020. The latter might mainly be explained by the COVID-19 pandemic, which disrupted the usual healthcare system at the beginning of 2020 in many countries, including Germany. Delayed diagnosis due to the COVID-19 pandemic has been shown for other indications such as cancer [27] and might have also led to delayed diagnosis of MG, especially in milder cases. Similarly, HR decreased to an overall low of 8.1% in 2020. This decrease might also partly be explained by the COVID-19 pandemic. With the first lockdown starting in March 2020, only emergency care was possible for months in Germany. Elective procedures such as thymectomy were canceled during this time. Consistently, the thymectomy rate was low in 2020 compared to previous years (Table 3). Furthermore, myasthenic exacerbations without respiratory insufficiency might have been treated more frequently in an outpatient than a hospital setting during the COVID-19 pandemic. However, declining HRs were already emerging in the years before the Corona pandemic. One explanation might be that in 2010, the German Myasthenia Gravis Society started to certify integrated MG centers. To date, there are 18 centers providing care for almost 4,000 MG patients in Germany. Specialized treatment by MG experts and integrated infusion outpatient clinics might decrease the number of MG hospitalizations. Interestingly, Sipilä et al. [17] reported a higher HR of MG patients of ∼25% in Finland between 2004 and 2014. However, the Finnish study was less stringent on defining an “MG-related hospital visit” which could partly explain the higher rate reported.

Although rituximab is an off-label drug for MG treatment, its usage continuously increased from 0.6% (2011) to 2.2% (2017). In 2020, the rate was 1.5%. Eculizumab was first prescribed in 2018 in our cohort. At the end of 2017, eculizumab was approved in Germany for refractory, AchR-ab-positive, generalized MG, which might explain the still low treatment rate of 0.2% in 2020 [15]. However, Phillips et al. [20] reported similar rates of 1.1% rituximab and 0.4% eculizumab usage across a large US population between 2014 and 2019, though usage of eculizumab in the USA is not restricted to refractory MG.

The rate of PLEX/IA as MG exacerbation/crisis treatment fluctuated between 0.5% and 1.2%, and the rate of IVIg treatment fluctuated between 1.7% (2011) and 4.0% (2016). Sipilä et al. [17] reported an approximated mean of 0.7% of all Finnish MG patients received PLEX/IA in 2006 and 2014. Phillips et al. [20] reported higher rates than this study, 2.8% of all MG patients in their US population were treated with PLEX, and 9.0% of patients underwent treatment with IVIg between 2014 and 2019, suggesting different treatment patterns across countries.

Strengths and Limitations

Despite the large number of observed patients and the long follow-up time, this cohort might not be representative of the German population since data might be skewed toward a younger, healthier, and male-dominated population [28]. Prescription rates might not give an exact picture of how patients are taking their medication in real life, and reasons for hospitalization might, despite best practice in analysis, be other than due to MG. The criterion for a “true” MG case of at least four base therapy prescriptions might miss incident cases receiving a first prescription toward the end of the study period and those who did not receive a second prescription due to side effects or lack of therapeutic benefit. Also, by including pharmacological treatment in the case definition for MG, patients in clinical remission might have been missed. However, we opted for that definition to avoid including patients misclassified in the dataset to prevent overestimation of IR and PR.

The COVID-19 pandemic probably had a major effect on the data from 2020. A follow-up analysis in the upcoming years is required to estimate the long-term effect of the COVID-19 pandemic on IR, PR, and HR and treatment patterns in Germany. Lastly, our data only map diagnoses that are ICD-10 code relevant. These include data such as neoplasms of the thymus. Of the 1,493 identified MG cases, 44 (2.95%) received a diagnosis of a malignant neoplasm of the thymus (ICD-10-GM code: C37) and 90 (4.14%) received a diagnosis of a benign neoplasm of the thymus (ICD-10-GM code: D15.0). However, the dataset does not provide detailed information on MG patients. It therefore does not allow for epidemiological analysis of clinical (i.e., ocular vs. generalized MG) and serological subtypes (i.e., autoantibody status) or correlation of disease patterns and severity with HR and treatment patterns.

Statement of Ethics

Ethical approval was not required, as German law allows for analyzing anonymous data for research purposes without patients' consent.

Conflict of Interest Statement

H.W. is an employee of SmartStep Consulting GmbH and T.V. is a managing partner of SmartStep Consulting. SmartStep Consulting received funding from Argenx Switzerland SA for the conduction of the research and the writing of the publication. S.H. and T.R. both received consultancy fees from argenx. B.D. is employed at the GWQ ServicePlus AG, which was founded and is owned by a group of health insurance companies. None of the funding organizations interacted with the authors on any content of the enclosed research or publication. The authors have disclosed their interest fully, and no further competing interest exists.

Funding Sources

This study was funded by argenx Switzerland SA. The authors had complete autonomy to establish the protocol, conduct the analyses, and interpret the results. This autonomy also includes the full right to publish the results without limitation.

Author Contributions

H.W. and T.V. designed the study. H.W. analyzed the data. S.H., H.W., and T.V. interpreted the results. H.W. and S.H. drafted the manuscript. T.R. and C.N. reviewed the manuscript. B.D. supplied the data and methodological support. All authors participated in revising and finalizing the manuscript and approved it for submission.

Data Availability Statement

Because of the confidential nature of in- and outpatient claims data, permission for public availability of the data is not possible. The permission to access the data is restricted to research and subjects to the consent of the health insurance funds.

Funding Statement

This study was funded by argenx Switzerland SA. The authors had complete autonomy to establish the protocol, conduct the analyses, and interpret the results. This autonomy also includes the full right to publish the results without limitation.