Real-World Prevalence, Incidence and Management of Systemic Lupus Erythematosus in Germany: A Retrospective Claims Data Analysis

On the basis of our real-world data, the prevalence of SLE in the German SHI population is continuously increasing, whereas SLE incidence rates remain stable. In our study, SLE treatment, especially in later SOTs, seemed to provide disease stabilisation to some extent. Organ manifestation in newly diagnosed patients with SLE did not worsen across SOTs, and an increasing proportion of patients were receiving a lower corticosteroid dose in later SOTs than in their previous SOT. However, treatment patterns, such as gaps in SLE therapy and continued corticosteroid usage above the recommended daily dose across SOTs, indicate an inadequate treatment and noncompliance with the current guidelines.

The observed age and sex distribution of the incident cohort was similar to that reported in the National Database of the German Collaborative Arthritis Centres [25, 28]. Data from this study confirm that SLE occurs across all age groups but most commonly affects people in their middle age, in the 45–54 years age group. Women were predominantly affected, regardless of organ manifestations, confirming the reported distribution of SLE between men and women in Germany and from studies worldwide [25, 28, 33, 34]. In our study, a female-to-male ratio of approximately 3.6:1 was noted in the disease incidence cohort, representing a higher incidence of men with SLE than reported in other studies (female-to-male ratio ranging from 4.3:1 to 13.6:1) [6, 35]. Similarly, a 2002 study in the German SHI population reported a lower female-to-male ratio (2:1), compared with other reports [35, 36]. However, the observed variations in disease incidence across studies could also reflect inherent differences in population demographics and socioeconomic factors [5, 35].

One of the key findings of our study is that the overall prevalence of SLE increased almost linearly over the study period, from 40.47 per 100,000 population in 2012 to 59.87 in 2019 (corresponding to approximately 49,800 patients with SLE in Germany in 2019 [32]), while the annual incidence rates remained relatively stable, albeit with some fluctuation. A similar epidemiological stage, referred to as compounding prevalence, is also observed in other chronic inflammatory diseases such as inflammatory bowel disease [37, 38]. An increase in prevalence despite stable incidence could indicate a reduced lupus-related mortality, owing to improvements in the treatment landscape or disease management, including increased physician awareness, as observed in other autoimmune diseases [37, 39]. These findings are in line with recent studies demonstrating a general trend towards improved survival in patients with SLE over time [40, 41]. A general trend for SLE prevalence to increase over time has been noted across many studies worldwide [33, 34, 42]. In contrast to our observations, some studies reported a decline in the incidence of SLE over time. A retrospective UK cohort study noted an annual 1.8% decline (p < 0.001) in SLE incidence from 1999 to 2012 [43]. Another study from Italy also observed a stable decline in SLE incidence from 2013 to 2020 [42]. However, findings vary across regions, indicating that SLE incidence may differ geographically and could vary intra-regionally depending on ethnicity [33, 42, 43]. Therefore, care should be taken when comparing incidence rates from different countries.

Estimated SLE prevalence rates similar to those reported in our study were also reported in a 2021 German study by Schwarting et al., using the German Betriebskrankenkassen (BKK) health insurance fund database between 2009 and 2014, with a trend in increasing prevalence over time (38.61 to 55.80 per 100,000 people) [27]. However, in contrast to our study, Schwarting et al. observed an increase in SLE incidence (6.1 to 8.82 per 100,000 people), which could be partly explained by the different methodologies used [27]. The WIG2 database used in our study contains data from different statutory health insurers, whereas the BKK database used by Schwarting et al. represents only one type of insurance category, resulting in different patient populations.

When stratified by age, the prevalence pattern in our study seems to align with that of the German National Database, with peaks at a relatively young age and again at a later stage in patients’ lives [25, 28]. SLE prevalence observed between 2017 and 2019 in our study corroborates the prevalence noted in a systematic review by Albrecht et al. (56 per 100,000) [24]. On the basis of the German National Database, the average age of SLE onset was 41 years for the period 2019–‍2021, and 40 years for the period 2020–2022; however, data stratified by sex were not reported [25, 28]. Our 2019 snapshot data indicated a‍ later onset of SLE in men than in women (55–64 vs 45–49 years of age). Moreover, disease onset in women occurred later than previously reported in the literature (45–49 vs 20–30 years of age) [3, 44, 45].

Although previous studies are generally consistent with our data on SLE prevalence stratified by sex and age, some differences were noted worldwide [34, 46]. These findings reiterate the importance of ethnicity when interpreting epidemiological data and the limitations of any generalisations.

SLE treatment patterns in the incident cohort indicated that treatment gaps were less common in subgroups with severe organ involvement (SLE + LN-only, SLE + CNS-only, SLE + LN + CNS subgroups), potentially confirming that increasing disease severity in terms of critical organs involved does not allow interruptions in treatment. Throughout the patient journey, treatment gaps were observed on a regular basis, with more than half of the patients in each SOT having no SLE treatment for > 60 days. This pattern was observed for the overall incident cohort, as well as for the subgroups. This could be potentially attributed to nonadherence due to inadequate responses, which was also reflected by the substantial proportion of patients switching their current SOT to another treatment after a treatment gap. Poor adherence to treatment is a common issue in chronic diseases such as SLE and could lead to an increased risk for flares, hospitalisations, morbidity and mortality, as well as poor renal outcomes in patients with renal involvement [47, 48]. In the incident cohort, more than half of the patients (52.4%) in SOT1 started a monotherapy as their first treatment, preferentially with an antimalarial or a corticosteroid. This is consistent with the 2023 EULAR guidelines, which recommend antimalarials for all patients with SLE and the use of corticosteroids for rapid symptom relief [16]. At least 40.0% of patients were receiving medium-to-high doses (7.6 to > 15.0 mg/day) of corticosteroids. Thus, guideline recommendations on using steroid-sparing therapy are not being implemented to a satisfactory level, leading to the use of high corticosteroid doses as observed in our study. We also noted a high proportion of patients with cardiorespiratory/cardiopulmonary diseases, which could be a result of increased risk of cardiovascular diseases in patients with SLE due to atherosclerosis or other organ involvement [49]. It is also plausible that the long-term corticosteroid use observed in our study could have contributed towards organ damage accrual. Persistent use of high-dose corticosteroids has been associated with an increased risk of severe adverse events and organ damage accrual [50]. Nevertheless, caution is required when interpreting the daily dosages of prescribed medications extracted from claims data. On the basis of the German National Database, 46.0% of patients with SLE received corticosteroids in 2022, among whom 78.0% had daily prednisolone dosages ≤ 5.0 mg [28]. Additional factors for worsening organ manifestations could be a lack of efficacy of available therapies in a subset of patients, infrequent use of biologics, or insufficient prescription of drugs by rheumatologists/general physicians. Indeed, we note that the use of biologics was limited in our study: 4.39% of patients in 2019 and 9.7% of patients in SOT3. In comparison, on the basis of the German National Database, belimumab use linearly increased from 7.0% in 2019 to 11.0% in 2021 and 15.0% in 2022 [25, 28]. However, it must be noted that the German National Database was based on data explicitly from rheumatologists; in comparison, the WIG2 database used in our study covered data from a broader set of healthcare professionals, including general physicians who might not prescribe biologics, most likely due to a lack of knowledge or acceptance, as access and reimbursement are granted for the German SHI population. Against this background, it is to be hoped that efforts throughout Europe for improved patient care, such as the European Reference Networks, in addition to the recent approval of anifrolumab in Europe, will foster the use of biologics in SLE in Germany [20, 51, 52].

We acknowledge that our study has some limitations. First, there are potential limitations due to the retrospective nature of patient data collection from an administrative claims database. In particular, patients could have been misclassified as a result of limitations in the inclusion/exclusion criteria defined in the study protocol, given that the definitions can only be considered the best possible approximation to reality. Misclassification of patients may occur because of incorrect capture of clinical conditions and/or treatments (i.e. symptom overlap with other disorders; cutaneous lesions; misdiagnosed as SLE topical treatment; erroneous inclusion of patients who received only pain relief instead of SLE-specific treatment), missing diagnostic codes or coding errors. In this respect, a potential misclassification of SLE in our study is reflected by a group of patients (27.6%) in the incident cohort with no documented treatment, which might result in an overestimation of prevalence and incidence data. Second, medication data were limited to drugs administered in outpatient settings and were collected via medication receipts in the claims database, which does not necessarily equate to correct usage of medication by patients. In addition, patients may have received drugs that might not have been captured in this database, which could suggest an underestimation of the number of patients in the incident cohort. The self-injectable formulation of belimumab was only approved in late 2017, and this could have contributed to an underestimation of the use of biologics in our study [53]. Also, data on daily dose of medication in the study were based on an approximate calculation derived from prescriptions. Third, the relatively small sample size for some of the disease manifestation subgroups and the lack of formal statistical testing of data may hinder interpretation of data. Moreover, organ involvement data could be skewed, owing to the challenges in differentiating between actual SLE organ manifestation and occurrence of comorbidity. Finally, the study might not be generalisable to patient populations in other countries. Despite the limitations, consistency of our findings with previous studies provides confidence in the data presented.