FormalPara Key Summary Points

Why carry out this study?

Relapse prevention and acute attack management are important to prevent disease worsening in neuromyelitis optica spectrum disorder, but existing guidelines do not provide recommendations for assessment of relapse severity

This study sought to delineate how assessments of disease, including classification of relapse severity and disease stability, are made in clinical practice and gain patients’ perceptions of these measures

What was learned from the study?

There is no clear consensus on how relapse severity or disease stability are defined in clinical practice; neurologists focus on clinical assessments when determining relapse severity, whereas patients focus on the impact that the relapse has on daily life and comparison with prior relapses. Similarly, there was a disconnect in the definition of disease stability: the complete absence of relapses was more important for patients than for neurologists

To ensure that patients receive the appropriate therapy in a timely manner, a clear consensus on how to assess relapse severity and disease stability that combines clinical measures with patient-focused assessments is required

Introduction

Neuromyelitis optica spectrum disorder (NMOSD) is a rare and debilitating autoimmune disease of the central nervous system that predominantly affects the optic nerve and spinal cord [1, 2]. The presence of aquaporin-4-immunoglobulin G (AQP4-IgG) autoantibodies in at least two-thirds of patients is a key underlying pathology that distinguishes NMOSD from multiple sclerosis (MS) [3, 4].

Patients with NMOSD present with a wide variety of symptoms and disabilities, including visual impairment due to optic neuritis; motor paralysis, sensory loss, and bladder/bowel dysfunction due to transverse myelitis; and severe nausea, vomiting, and hiccups due to lesions in the area postrema [1, 5,6,7,8]. Patients may also experience fatigue, depression, cognitive dysfunction, spasticity, and pain [9,10,11]. Unsurprisingly, quality of life is also reduced in patients with NMOSD [12].

NMOSD is characterized by unpredictable relapses that are associated with permanent neurological damage and disability [1, 13,14,15,16]. Although the severity of each relapse may vary [17], recovery is usually incomplete, meaning that neurological damage and disability accumulates with each relapse [1, 18, 19]. Therefore, relapse prevention and management of acute attacks are therapeutic priorities in NMOSD [20].

This study aimed to provide global insights into current clinical practice and management of AQP4-IgG-seropositive patients with NMOSD. Published data from this study have demonstrated that misdiagnosis is common, with approximately 25% (228/910) of patients with NMOSD initially being misdiagnosed with conditions such as MS, idiopathic myelitis, optic neuritis, and stroke [21]. Data from the same study also showed that decisions regarding treatment are driven by subjective assessments of disease severity by neurologists [21]. This analysis focuses on how assessments of disease, including classification of relapse severity and disease stability, are made in clinical practice, as well as patient perceptions of these measures.

Methods

Study Design

Full details of the study design, including screening criteria, have been reported elsewhere [21]. Briefly, in this cross-sectional study, neurologists from six countries (the USA, Germany, Italy, Brazil, South Korea, and China) were asked to take part in a 30–60 minute online survey, during which they completed two to four clinical records for their adult patients with AQP4-IgG-seropositive NMOSD. Neurologists were recruited via an International Organization for Standardization-certified medical research panel from a mixture of academic and non-academic hospitals, public, private, solo, and group practices, and received an honorarium for their participation. Quality checks occurred daily during the fieldwork and if quality control measures were not fulfilled, data were removed from the overall sample.

Separate to this survey, 30-minute qualitative telephone interviews were conducted with patients diagnosed with NMOSD to understand their perceptions of disease and management. None of the patients included in the qualitative interviews were the same patients as those included in the clinical record review. Patients were recruited by telephone and email from the same six countries as the participating neurologists via physician referrals, patient panels, and patient groups/associations. Patients received an honorarium for their participation. Data were quality checked by how engaged patients were and whether their responses during the interview matched those given during screening.

Study Population

As part of the online survey, neurologists submitted clinical records for patients who were ≥18 years of age, had a confirmed diagnosis of NMOSD (Supplementary Material) and tested seropositive for AQP4-IgG, and fell into one of three categories: (1) newly diagnosed with NMOSD within the last 2 years, but not in the last 3 months (included to understand disease/patient characteristics in newly diagnosed patients not yet initiated on or receiving their first maintenance therapy); (2) diagnosed with NMOSD for ≥2 years and who had their maintenance therapy changed, meaning a treatment switch or add-on, within the past 2 years (included to understand disease/patient characteristics in patients who had a change to their maintenance therapy); or (3) the most recent patient with NMOSD seen by the neurologist in clinical practice (included to represent a “random” patient, and ensure that a representative sample is included to allow for “sizing” of behaviors or disease/patient characteristics).

Patients participating in the telephone interviews were required to be ≥18 years of age, have a confirmed diagnosis of NMOSD, and be receiving maintenance therapy for their NMOSD. Due to the qualitative nature of the patient interviews, self-reported patient AQP4-IgG status was not always established.

Data Collection and Assessments

The clinical record review was designed to collect information on patient demographics and diagnosis, including the time between the appearance of first symptoms and diagnosis, date of diagnosis, diagnostic criteria used, any previous misdiagnoses, AQP4-IgG status, and the timing of AQP4-IgG testing.

Relapse history and characteristics were also assessed with details collected on the number of relapses, date of most recent relapse and implications (i.e., change in maintenance treatment and reason), clinical signs/symptoms of the relapse, duration of relapse (i.e., time between onset of symptoms and maximum recovery), acute treatment of relapse, duration of hospitalization if applicable, extent of recovery from relapse, change in function/disability due to relapse, level of decline related to relapse assessed by Expanded Disability Status Scale (EDSS) and timed 25-foot walk (T25FW) scores, and perceived severity of relapse and rationale. Detailed information on specific assessments is provided in the Supplementary Material. Relapse severity (i.e., “mild,” “moderate,” or “severe”) and disease stability (i.e., “stable” or “unstable”) were subjectively classified by each individual neurologist in the clinical record review. Differentiation between relapses and pseudo-relapses was not ascertained from the chart review.

The patient interviews were designed to gather information about the patient’s journey from diagnosis to treatment, including perceptions about relapse severity or disease stability. A copy of the online survey used in the clinical record review and the patient interview guide can be found in the Supplementary Material.

Statistical Analysis

Sample sizes were based on the maximum feasible sample of neurologists that could be achieved in each country within the 4-week fieldwork period. Statistical analysis involved cross-tabulations with significance testing. Ordinal regression was performed to hierarchize factors determining relapse severity. The p-values were calculated using SPSS software, version 26 (IBM SPSS Statistics, IBM Corporation, Armonk, New York), and a significance level of p < 0.05 was established.

The chi-squared (χ2) test was used to detect an association between two variables. Pairwise comparisons of proportions were performed using a z-test when the test variable was categorical, and pairwise comparisons of means were performed using a t-test when the variable was scale based. Bonferroni corrections were conducted for the z- and t-tests.

For the qualitative patient interviews, the patient interview audios were used to develop verbatim English transcripts, which then underwent thematic analysis to identify any common patterns.

Ethics Statement

This was a market research study and, as such, it was conducted in accordance with the European Pharmaceutical Market Research Association’s (EphMRA) industry code of conduct. Market research as defined in this code of conduct does not require clinical research ethics committee or independent review board approval. All subjects provided informed consent to participate in the study (in accordance with the EphMRA code of conduct), and research was compliant with all international and national data protection laws. Quote responses provided are vignettes of real answers that have been amended to maintain anonymity of respondents without altering the theme of their statements.

Results

Evaluable Study Population and Patient Characteristics

In total, 389 neurologists completed the online survey between 14 July and 15 August 2020 [21]. Reasons for non-qualification are included in the Supplementary Material. The neurologists provided records for 1185 AQP4-IgG-seropositive patients with NMOSD. This sample consisted of 472 patients (40%) diagnosed with NMOSD ≤ 2 years ago (but not in the last 3 months), and who were either receiving their first maintenance therapy [typically oral corticosteroids and/or immunosuppressive therapies (ISTs) or monoclonal antibodies] [21] or had not yet been initiated on a maintenance therapy; 655 patients (55%) with a change in maintenance therapy (switch or add-on) within the past 2 years; and 58 patients (5%) diagnosed > 2 years ago who had not changed their maintenance therapy in the past 2 years.

A total of 33 qualitative patient telephone interviews were also conducted between 20 October and 27 November 2020 [21]. Twenty-one (64%) patients were newly diagnosed with NMOSD within the last 2 years, of whom six reported recent/upcoming therapy changes. Of the remaining 12 “ongoing” patients (i.e., not newly diagnosed), eight reported recent/upcoming therapy changes. In total, 17 therapy changes (either a treatment switch or an add-on) were reported in 14 patients from the patient interview group.

The geographical distribution of the participating neurologists and patients is presented elsewhere [21].

Classification of Relapse Severity

Neurologist Clinical Record Review

Details of the most recent relapse (or initial attack in newly diagnosed patients) were recorded for all 1185 patients included in the clinical record review. In total, 12% (143/1185) of relapses were classified by neurologists as “severe,” with 44% (520/1185) being classified as “moderate,” and the remaining 44% (522/1185) as “mild.” Geographical variation in relapse classification was noted, with more relapses being classified as “mild” in China, South Korea, or Brazil than in Germany (p < 0.05). Similarly, more relapses were likely to be categorized as “moderate” in Germany versus Brazil (p < 0.05), and as “severe” in Germany and Brazil versus South Korea (p < 0.05; Fig. 1).

Fig. 1
figure 1

Classification of relapse/initial attack severity by country. *Indicates a difference (p < 0.05) versus Germany. **Indicates a difference (p < 0.05) versus Brazil. †Indicates a difference (p < 0.05) versus South Korea. The z-test was used for pairwise comparisons of proportion, and the Bonferroni correction was applied to this test

Full size image

The three clinical factors most likely to be associated with relapse severity were identified and hierarchized as follows: (1) extent and speed of recovery after relapse, (2) extent of spinal cord involvement, and (3) level of residual disability. Residual disability was measured in three ways: the extent of function before relapse and 3 months after relapse (e.g., visual, motor, bowel, bladder, sensory function), EDSS or T25FW scores before relapse and 3 months after relapse, and the ability to perform activities of daily living before relapse and 3 months after relapse (e.g., shopping, cooking, driving, household chores). Optic neuritis and brainstem symptoms did not contribute to assessments of residual disability.

There was an association between relapse severity and the highest degree of recovery reached following a relapse (p < 0.05 for interaction). A higher proportion of patients with a “mild” relapse experienced a full recovery versus those with a “moderate” and those with a “severe” relapse [51% (264/522) versus 13% (69/520) and 6% (9/143), respectively], and this observation was consistent across all six countries (Supplementary Material). Similarly, a higher proportion of patients with a “severe” relapse experienced poor/no recovery following a relapse than those with a “moderate” or “mild” relapse [24% (34/143), 2% (10/520), and 0% (0/522)], and this observation was also consistent across all six countries (Supplementary Material).

Patients with a “severe” relapse also had both a longer duration of relapse and a slower recovery versus patients with a “moderate” (p < 0.05) or “mild” relapse (p < 0.05; Table 1). In addition, patients with a “moderate” relapse had a slower recovery versus patients with a “mild” relapse (p < 0.05).

Table 1 Relapse duration and time to “highest” recovery according to relapse severity in all patients from the clinical record review
Full size table

The presence of acute transverse myelitis at diagnosis predicted relapse severity, with a higher proportion of patients with a “severe” relapse displaying acute transverse myelitis at diagnosis than patients with a “moderate” or “mild” relapse [57% (82/143), 42% (218/520), and 27% (139/522), respectively; p < 0.05 for both comparisons]. Of those patients who had spinal magnetic resonance imaging (MRI) conducted at diagnosis (n = 663), more patients with a “severe” relapse had longitudinal extensive transverse myelitis (LETM; defined as extension over ≥ 3 vertebral segments) than patients with a “moderate” or “mild” relapse [76% (87/115), 61% (189/308), and 49% (117/240), respectively; p < 0.05 for both comparisons], whereas short transverse myelitis was observed at diagnosis in more patients experiencing a “mild” relapse versus those with a “moderate” or “severe” relapse [36% (86/240), 26% (80/308), and 11% (13/115), respectively; p < 0.05 for both comparisons]. Unilateral optic neuritis at diagnosis (rather than bilateral) was also an indicator that the patient was more likely to experience a “mild” relapse compared with a “moderate” or “severe” relapse [49% (254/522), 38% (196/520), and 31% (44/143), respectively; p < 0.05 for both comparisons].

Relapses were more likely to be classified as “mild” if the patient was considered to have “normal” function prior to their relapse and did not exhibit any functional decline after relapse. For instance, more patients who had a “mild” relapse had normal bowel/bladder function, sensory function, and motor function prior to their relapse than those who then had a “moderate” or “severe” relapse (p < 0.05 for both comparisons; Fig. 2i). Correspondingly, a persistent change from baseline in bowel/bladder, and sensory and motor function 3 months after relapse was noted in more patients with a “severe” relapse versus those with a “mild” relapse (p < 0.05). When comparing patients with a “severe” relapse versus those with a “moderate” relapse, relevance was only demonstrated with regards to bowel/bladder dysfunction (p < 0.05) (Fig. 2ii).

Fig. 2
figure 2

Proportion of patients with (i) “normal” function prior to their relapse (n = 570) and (ii) a persistent change from baseline function 3 months after relapse (N = 1185), according to the severity of their relapse as assessed by the neurologist. Data before relapse were only collected for patients’ most recent relapse and not for initial attack (n = 570), whereas data after relapse were collected for all patients (n = 1185). *Visual acuity n numbers were lower than for other measures of function as “do not know” responses were removed: right eye, n = 536 (before) and n = 1118 (after); left eye, n = 537 (before) and n = 1117 (after). †Difference versus a “moderate” or “severe” relapse (p < 0.05). ‡Difference versus a “severe” relapse (p < 0.05). §Difference versus a “mild” relapse (p < 0.05). Difference versus a “mild” or “moderate” relapse (p < 0.05). The z-test was used for pairwise comparisons of proportions, and the Bonferroni correction was applied to this test

Full size image

Worsening of EDSS and T25FW scores post-relapse (which were assessed pre-relapse and 3 months post-relapse, with “worsening” defined as any increase in score or time) also demonstrated an association with relapse severity (p < 0.05 for interaction for both variables). Additionally, the degree of worsening in EDSS and T25FW scores was also correlated with relapse severity (Supplementary Material). The total proportion of patients showing an increase in EDSS score with “mild,” “moderate,” or “severe” relapses was 22% (28/129), 61% (96/157), and 78% (40/51), respectively, whereas the proportion of patients experiencing an increase in T25FW score was 32% (17/53), 65% (46/71), and 85% (11/13), respectively.

“Severe” relapses were also more likely than a “mild” relapse (p < 0.05) to decrease the patient’s ability to perform activities of daily living post-relapse, including household chores, shopping, working, cooking/preparing food, driving, washing/dressing, taking medication, and eating (Fig. 3).

Fig. 3
figure 3

Proportion of patients experiencing a persistent change from baseline in ability to perform activities of daily living 3 months after relapse, according to relapse severity. Data shown for patients in whom ability to perform activities of daily living was recorded before and after relapse. *Indicates a difference versus a “mild” or “moderate” relapse (p < 0.05). †Indicates a difference versus a “mild” relapse (p < 0.05). The z-test was used for pairwise comparisons of proportions, and the Bonferroni correction was applied to this test

Full size image

No clear pattern was observed between the occurrence of a severe relapse and the severity of the preceding relapse. In patients who had at least one relapse prior to their most recent “severe” relapse (n = 66), the previous relapse was classified as being “mild,” “moderate,” or “severe” in 30% (20/66), 44% (29/66), and 26% (17/66) of cases, respectively (p-values for all comparisons did not indicate a difference). In addition, 36% (24/66) of patients had their preceding relapse < 6 months prior to their most recent “severe” relapse, 30% (20/66) had it within the previous 6–12 months, 12% (8/66) within the previous 13–18 months, and 21% (14/66) experienced their preceding relapse > 18 months ago.

Patient Interviews

Seventy-nine percent of patients (26/33) who participated in the telephone interviews had experienced more than one relapse and, of these patients, approximately two-thirds reported a mix of relapse severities:

I had the same vision loss (in the mild versus severe relapse)… In the severe one, where it was constant vomiting. This was more just occasional vomiting… some symptoms were not there at all and the rest were milder.

When assessing their own relapse severity, patients tended to “benchmark” their current relapse against their previous relapses or what they had heard about other patients’ experiences. A relapse was not considered to be “moderate” or “severe” by the patient if they thought that it could have been worse:

I classed my relapse as mild because I have read about what they can be like and it was nowhere near as bad as some people have had.


This relapse was bad, but not as bad compared to others I have experienced; for example, there was no paralysis, the pain was milder, and it did not last as long as more severe relapses I have had.

Additionally, when categorizing relapse severity, patients were found to focus on the type and intensity of the symptoms that they experience, the change in their daily life (during the relapse), and the degree of recovery or residual disability following the relapse. Severe motor weakness or paralysis, blindness, loss of bowel/bladder control, severe pain and severe vomiting are symptoms that patients typically associated with a “severe” relapse, whereas mild vision impairment (e.g., double/blurred vision), numbness, a “prickling” or “tingling” sensation, headaches, and fatigue are typically associated with a “mild” relapse by patients. The perceived level of disruption to daily life is assessed by the need for hospitalization, relapse duration, and the ability to perform activities of daily living during the relapse. After a “mild” relapse, patients can typically continue to work and perform activities of daily living at the same level as before the relapse, whereas after a “severe” relapse, patients may need to make important adjustments to their daily life, e.g., change their job role/hours, need assistance with activities of daily living, and/or require mobility aids such as a wheelchair or cane. The length of the recovery period is also a factor for the patients, with a longer recovery period (such as several months) being indicative of a more “severe” relapse, and a relapse that resolves within days/weeks being categorized as “mild”.

Patient perception of relapse severity did not always align with that of their neurologist, with two patients reporting that, although they were told that tests and MRIs showed their relapse to be “non-severe” or “non-inflammatory,” they still felt they had been severely affected:

[My doctor] said the relapses were mild, but they felt worse to me because I could not feel my arm and I could barely see.


My doctor said the relapse was not severe, it was just non-inflammatory because I was stressed… but he has said that about other relapses that were very severe, so I do not think I agree.

Classification of Disease Stability

Neurologist Clinical Record Review

In the clinical record review, 91% of all patients (1078/1185) were considered to be in a “stable” disease state, with the remaining 9% (107/1185) being classed as being “unstable.” “Stable” patients (n = 1078) compared with “unstable” patients (n = 107) were found to experience fewer relapses in total up to the point of the survey (mean number of total relapses: 2.8 versus 3.4 respectively; p < 0.05) and less-frequent relapses (mean number of relapses experienced in the past 12 months: 0.9 versus 1.1, respectively; p < 0.05) (Fig. 4).

Fig. 4
figure 4

Number of (i) total relapses experienced and (ii) relapses experienced in the past 12 months in “stable” (n = 1078) and “unstable” (n = 107) patients. *Indicates a difference versus “unstable” patients (p < 0.05). **Indicates a difference versus “stable” patients (p < 0.05). The z-test was used for pairwise comparisons of proportions, and the Bonferroni correction was applied to this test

Full size image

Patients who were “unstable” were also more likely to have had a relapse more recently (p < 0.05) and to experience a “severe” relapse than those who were categorized as “stable” (p < 0.05); the mean time since the most recent relapse in “stable” versus “unstable” patients was 18.6 months [standard deviation (SD) 21.6 months; range 1–263 months] versus 13.0 months (SD 10.5 months; range 3–51 months), respectively, and the proportion of patients whose most recent relapse was classed as “severe” in “stable” versus “unstable” patients was 9% (95/1078) versus 24% (26/107), respectively.

More “stable” versus “unstable” patients also showed signs of a full recovery after their most recent relapse: 32% (257/810) versus 16% (13/80), respectively (p < 0.05). Patients who were classed as “unstable” were more likely (p < 0.05) to experience the following types of symptoms with residual disability than “stable” patients during their most recent relapse (p < 0.05): vision impairment, orbital pain, nausea, vomiting, band-like sensation around the trunk, loss of bowel/bladder control, fatigue, painful tonic spasm, muscle spasm, and weakness. There was also a difference (p < 0.05) in the duration of the most recent relapse between “stable” and “unstable” patients (37.7 days versus 51.9 days, respectively).

Patient Interviews

Seventy-nine percent (26/33) of patients who participated in the telephone interviews believed their NMOSD to be currently stable, as they had not seen any recent change in disease activity (i.e., no recent relapse or no relapse since they had started on their most recent maintenance therapy, and/or no new symptoms in the last 6–9 months):

I think I am stable; I have not relapsed again, and I do not have any symptoms because of my NMO.

In contrast, those patients who viewed their disease as being “active” (7/33) had often experienced a recent relapse within the last 6–9 months (5/7), and/or had seen a change in residual symptoms, which could have been the appearance of new symptoms or worsening of existing symptoms.

Discussion

This analysis provides valuable insights into current clinical practice of and disease characteristics in NMOSD. Through the recruitment of neurologists from six different countries (and concomitant patient interviews), we now have a global picture of how relapse severity and disease stability are defined in clinical practice, and what, if any, potential predictors for relapse severity there may be. Additionally, patient interviews provided information on how patients view their disease course and relapses, and help us to understand what differences might exist between patients and physicians.

Existing guidelines do not currently provide recommendations for how relapse severity should be assessed in practice, although it is also worth noting that it has been some time since these were last updated [22,23,24,25]. Published data from this study suggest that decisions about whether to initiate treatment following a relapse and which treatment to administer are highly influenced by the severity of the relapse [21]. A clear consensus on what constitutes a “severe,” “moderate,” or “mild” relapse is therefore important, as it can help to avoid ambiguity over whether to start treatment and help to ensure that patients have access to the appropriate treatments at the appropriate time. Of note, patients with AQP4-IgG-seropositive NMOSD who meet the diagnostic criteria [6] are currently recommended treatment with maintenance ISTs after their initial attack. The use of off-label ISTs and corticosteroids were (and in some regions continue to be) the primary treatment strategy [7, 25, 26]. At the time of writing, three monoclonal antibodies have been approved for the treatment of AQP4-IgG-seropositive NMOSD [26,27,28,29,30,31].

Regional differences also highlight the need for consistency in global definitions of relapse severity, with a greater proportion of relapses being categorized as “mild” in Brazil, China, and South Korea than in Germany, and more relapses being categorized as “severe” in Germany versus Brazil or South Korea. Whether regional differences in relapse categorizations were due to the patient sample assessed here, or due to differing perspectives of severity in different countries, is currently unclear. Interestingly, a recent global study reported that race affected the clinical phenotype of NMOSD, as well as the age of onset and the severity of attacks; however, the overall outcome was found to be most affected by early and effective IST treatment [32].

The finding that the extent of both spinal cord involvement at diagnosis and presence of unilateral optic neuritis may be indicative of whether patients are likely to experience “mild” or “severe” relapses emphasizes the importance of assessing both at disease onset. Revised diagnostic criteria from 2015 for patients with AQP4-IgG-seropositive NMOSD require (in addition to a positive test for AQP4-IgG and the exclusion of alternative diagnoses) the presence of optic neuritis, or acute myelitis with LETM, or area postrema syndrome [6].

It is also important to recognize that the severity of a relapse may need to be put into the context of a patient’s ability prior to the relapse; if a patient already has some level of motor impairment, then a relapse that might be classified as “mild” may be enough to push them to the point of requiring a mobility aid, meaning that although the individual relapse is mild, it may be perceived as “severe” by the patient due to the long-lasting effects and residual disability. This suggests that appropriate and effective treatment following a relapse should be sought after every relapse, and also as a preventative measure, as we do not know which relapse may cause a worsening of, or new symptoms.

This also ties in with the observation that no pattern was observed between the most recent “severe” relapse and the severity of the preceding relapse, with 74% of patients experiencing a “mild” or “moderate” relapse prior to their most recent “severe” relapse. The duration between the preceding and most recent “severe” relapse was also shown to potentially be quite short, with 67% of patients reporting “severe” relapses within 12 months of their prior relapse. This again highlights the importance of a single relapse, regardless of its severity, as even a “mild” or “moderate” relapse may be followed by a “severe” relapse. Additionally, this suggests that a delay in treatment initiation could prove to be harmful to patients, as patients with a “mild” attack at disease presentation could still experience a “severe” relapse next. These findings again reinforce the need for timely optimization of therapeutic strategies for relapse prevention following even a mild relapse, given that a “severe” relapse could occur within the next 12 months.

This analysis also raised areas of potential disconnect between patients’ and neurologists’ perceptions of the disease. For instance, from the telephone interviews, it became apparent that patients tend to “benchmark” their current relapse against their previous relapses and/or other patients’ experiences, whereas neurologists are more likely to assess each relapse in isolation. This is most likely a consequence of the neurologists’ assessment of clinical deficit compared with a prior baseline (impairment), whereas patients will instead assess severity based on the result of their relapse (disability). Assessments that combine clinical measures with the effect on patients’ daily living in the form of patient-reported outcomes are perhaps thus needed in both clinical trials and real-world clinical practice, to help provide a more complete picture of the situation.

The majority of patients from the clinical record review were classed as having “stable” disease, which was typically based on a number of different factors, including fewer relapses, less-frequent relapses, a longer period of time lapsing since the last relapse, and fewer “severe” relapses. However, although patients who are categorized as “stable” may have had fewer relapses overall and less-frequent relapses, this did not mean that they were not experiencing relapses. This is another area of potential disconnect, as the majority of patients from the interviews would only consider their disease to be stable if they had not experienced a relapse in 6–12 months, or had no new relapse since starting their most recent maintenance therapy. The occurrence of relapses, however mild, poses the question of how “stable” the disease actually is, and whether current notions of disease stability need to be challenged.

This study was limited by retrospective data collection and the inability to obtain complete data for all patients as part of the clinical record review. There is a risk of bias associated with physician self-reporting of clinical records, e.g., selection bias, recall bias, and desirability bias, and this may have skewed the population towards including patients with more “favorable” results. Similarly, as physician referral was one method of recruiting patients for the telephone interviews, those referred were more likely to be “satisfied” patients, rather than those who viewed their management less positively. These biases should be taken into consideration when interpreting the findings of this study.

Additionally, this study is limited by the fact that individual relapses were not assessed by both patient and physician, rather the clinical records of one group of patients was compared with interviews from a different group of patients. Furthermore, as there are no well-established definitions of relapse severity, the classifications in this study were subjective and based on the individual physician. However, understanding how severity is defined was one of the core objectives of this study, and has subsequently allowed us the opportunity to understand the real-world classification of relapse severity. In this study, level of residual disability was one of the classifications used to determine relapse severity. Residual disability was primarily measured by the extent of function before relapse and 3 months after the relapse rather than the presence of different symptoms, such as optic neuritis and brainstem symptoms, although these may have affected the patient's level of function.

This study provides valuable insights into the assessment of AQP4-IgG-seropositive NMOSD in current clinical practice worldwide, with data collected from more than 1000 clinical records. There is still no clear consensus on how to measure a relapse or its severity, although this study has highlighted that the degree/speed of recovery following a relapse, extent of spinal cord involvement, and level of residual disability following a relapse are important determinants when classifying relapse severity. The need for clear definitions of relapse severity and disease stability is further highlighted by differences between the views of the neurologist and those of the patient. There is therefore a need to tie together clinical assessments of relapse severity with more patient-focused assessments, such as symptoms experienced and the effects on daily living. Equally important to consider is whether a change is needed in how “disease stability” is viewed. From the patient’s perspective, disease stability is defined as the absence of any relapse and not the absence of severe relapses alone. It should also be noted that the severity of one relapse cannot be used to predict the severity of the next relapse, which underlines the need for appropriate and preventative treatment before a patient experiences a “severe” relapse and a worsening of, or new symptoms.

Conclusions

Currently, existing guidelines do not detail recommendations for assessing relapse severity in clinical practice. Published data from this study highlighted that treatment decisions are highly influenced by relapse severity. Therefore, a clear consensus determining how to assess relapse severity and disease stability is required to ensure that patients receive the appropriate therapy in a timely manner. Our findings also suggested that there are potential areas of disconnect between neurologists and patients on how relapse severity and disease stability in NMOSD are perceived. As such, clinical measures should be combined with patient-reported outcomes in the future to help provide a more patient-inclusive perception of the current disease state.