Introduction

Myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD) is a relatively new addition to the category of central nervous system (CNS) inflammatory demyelinating diseases [1, 2]. CNS inflammatory demyelinating conditions, including multiple sclerosis (MS) and neuromyelitis optica spectrum disorders (NMOSD), are differentiated based on severity, clinical phenotype, imaging, laboratory, and pathological findings [2] (Table 1). While patients with MS, MOGAD and NMOSD may present with similar clinical manifestations, such as optic neuritis and myelitis, those with MOGAD lack a clear sex predilection, and more commonly experience a monophasic course [2,3,4,5]. MOGAD also has the greatest predilection in children, representing 20–30% of inflammatory CNS syndromes in this population as compared to approximately 5% in adults. The current estimated range of incidence in the pediatric population is 3.1 per 1 million, as compared to 1.6 and 2.39 per 1 million among adults [3]. Notably, these numbers, along with the estimated worldwide prevalence of 20 million [5], are likely to increase with growing recognition of the disease and improved availability of serological testing. Unlike NMOSD, cases of MOGAD are not strongly associated with other systemic autoimmune disorders or circulating autoantibodies [3]. Yet, disease manifestations may occur after prodromal infections, particularly those caused by viral pathogens, such as influenza, Epstein–Barr virus, herpes simplex virus, and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), to name a few [3]. Occasionally, patients with MOGAD have an overlap syndrome with anti-NMDA receptor encephalitis, characterized by clinical features of demyelination associated with encephalopathy, seizures, dyskinesias, or psychosis [5]. As the clinical spectrum of MOGAD continues to expand, so too does our appreciation for diagnostic and management challenges associated with this enigmatic condition. Key areas of ongoing research include determining the specificity and pathogenicity of MOG autoantibodies, identifying immunopathologic targets for future therapies, discovering and validating biomarkers that detect disease activity, and deciphering which patients with MOGAD require long-term immunotherapy.

Table 1 Disease Characteristics that Distinguish MOGAD, MS, and NMOSD [2,3,4,5,6, 11,12,13,14,15,16,17]
Full size table

MOGAD: the evolving clinical spectrum

Our growing appreciation of the full clinical spectrum of MOGAD will likely mirror the NMOSD experience. Once considered to be a severe form of MS targeting the optic nerves and spinal cord, NMOSD has since been recognized as a distinct autoimmune astrocytopathy with pathognomonic clinical features [2, 6]. Similarly, MOGAD has now been identified as a separate entity from both MS and NMOSD. Recently, diagnostic criteria proposed by an international panel of experts highlight optic neuritis, myelitis, acute disseminated encephalomyelitis (ADEM), cerebral mono-focal or multifocal deficits, brainstem or cerebellar syndromes, and cerebral cortical encephalitis (often with seizures) as cardinal features of MOGAD (Tables 2 and 3) [5]. Unlike MS, neurological deterioration does not typically progress in the absence of relapses [5]. In real-world settings, there will be challenges in diagnosing MOGAD despite having clearly proposed criteria for the disease, because, as mentioned, many clinical manifestations of MOGAD overlap with other CNS inflammatory syndromes, including but not limited to MS and NMOSD. Diagnostic confusion may also arise from the interpretation of radiological and serological findings since these features may all be caused by other etiologies and are not specific to MOGAD. For this reason, it will be important to not overweigh any single finding in isolation. As a rule of thumb, clinical features of MOGAD, particularly acute optic neuritis, may closely resemble those of NMOSD with severe vision loss (often bilateral) at its nadir. Patients with MOGAD, however, often show greater therapeutic response to high-dose corticosteroid treatment and may also demonstrate significant spontaneous improvement [7,8,9,10]. Importantly, the phenotypic expression of MOGAD may vary with age, such that ADEM-like lesions are more likely to affect children, whereas optic neuritis and isolated myelitis tend to be more common among adults [5]. In addition, relapse risk in MOGAD patients is often higher in adults than children [3].

Table 2 Clinical Features of MOGAD [2,3,4,5, 14, 15, 18,19,20,21,22,23,24,25,26,27,28,29,30,31]
Full size table
Table 3 Proposed MOGAD diagnostic criteria by an international panel of experts. This table is adapted from Banwell et al., 2023 [5]
Full size table

MOG autoantibodies: specificity, pathogenicity, and interpreting titers

Seropositivity status for MOG-IgG autoantibodies is key to applying the recently proposed criteria for MOGAD (Table 3) [5]. It is also imperative to exclude alternate diagnoses, just as we do for patients with suspected MS or NMOSD. Without rigorous clinical characterization and thoughtful deliberation, individuals with true MOGAD run the risk of being under-treated, whereas those without MOGAD might be inappropriately exposed to long-term immunotherapies or incorrect therapies (Case vignettes here contrasting two scenarios of overlooked MOGAD [Case 1] and false positive [Case 2]). As highlighted by the recently proposed MOGAD criteria [5], there are challenges to relying on a serum MOG-IgG antibody with imperfect specificity and uncertain pathogenicity because causation can be confused with association. Notably, cell-based assays (CBA) are used to establish a MOGAD diagnosis, with live CBA techniques performing slightly better than fixed-based assays. Importantly, enzyme-linked immunosorbent assay (ELISA) testing should not be used to make a diagnosis, as findings are less specific and often discordant with CBA results [32]. While access to CBA techniques has improved diagnostic accuracy, the positive predictive value (the probability that a patient with a positive serum MOG-IgG result harbors MOGAD) still varies with the prevalence of MOGAD in any given patient population. Investigators at the Mayo Clinic have reported a positive predictive value (PPV) of MOG-IgG testing by live CBA of 72% using a cut-off of 1:20 despite having an overall specificity of 98% [33]. Their study found there was a positive correlation between serum MOG-IgG titer cut-off value and PPV, such that the PPV was 100% for 1:1000, 82% for 1:100, 51% for antibody titers 1:20–1:40 [33]. Notably, a ≥ 1:40 titer “cut-off,” (n = 65) yielded a PPV of 93.8%, but lowered sensitivity [34]. Subsequently, Manzano and colleagues also performed an institutional cohort study to determine the PPV of serum MOG-IgG for clinically defined MOGAD, using the same live MOG-IgG CBA [34]. Among 1,877 patients tested, 78 (4.2%) patients tested positive for MOG-IgG with titer ≥ 1:20; of those who were seropositive, 67 had a validated MOGAD diagnosis, yielding a PPV of 85.9%. Undoubtedly, low positive titers need to be interpreted with caution. Yet, it is worth mentioning that excluding patients with MOG IgG titers of 1:20 without proper clinical analysis may lead to a significant number of missed diagnoses of MOGAD [33,34,35], further reiterating the need for thoughtful clinical judgment. Ideally, patients should be tested for MOG-IgG sero-status at their incident event, and when possible, before administration of corticosteroids, immunoglobulins, or apheresis [5]. When there are concerns regarding a false positive MOG-IgG result, repeat serum testing should be performed after 3 months or at the time of relapse [5]. Finally, like aquaporin 4 (AQP4)-IgG, MOG autoantibodies are induced in the peripheral circulation, though the triggering cause for their production remains unclear [2, 4]. Serological testing with CBA techniques is more sensitive for detecting MOG autoantibodies than CSF sampling. Occasionally, patients with MOGAD may have negative serum MOG-IgG results, but show seropositivity with CSF testing [5, 36, 37]. For this reason, it is reasonable to consider CSF testing for MOG-IgG to support the diagnosis of MOGAD when the serum results are negative, but the clinical and MRI findings are otherwise highly suggestive of the diagnosis [36,37,38,39,40].

When considering specificity, MOG autoantibodies are uncommon in patients with MS (0.4%) [4]. However, because of the high prevalence of MS compared to MOGAD, many false positive results may occur if all patients with possible MS are tested for MOG-IgG, therefore discretion is required [33]. Co-existing seropositivity for MOG-IgG is also uncommon in cases of AQP4-IgG-positive NMOSD; when this occurs, AQP4 IgG titers are often higher, and the patient usually follows a NMOSD phenotype [5, 41]. Yet the high prevalence of MOG autoantibodies among patients with AQP4-IgG-seronegative NMOSD has created some diagnostic confusion, such that MOGAD patients have been erroneously labeled as seronegative NMOSD. With the newly proposed MOGAD criteria (Table 3), patients with presenting features of NMOSD who are negative for AQP4-IgG should be tested for MOG-IgG, as part of their diagnostic evaluation for MOGAD [5]. In fact, with more accessible use of reliable CBA for MOG-IgG, rigorous classification of clinical phenotypes and improved understanding of pathognomonic radiological findings, many patients previously diagnosed with autoimmune demyelinating diseases including seronegative NMOSD, atypical MS, chronic relapsing inflammatory optic neuropathy (CRION), and ADEM may be re-classified as MOGAD.

Going forward, it will be important to consider whether clinical manifestations in patients with suspected MOGAD reflect pathogenic effects of MOG autoantibodies, as opposed to a scenario in which MOG-IgG seropositivity more likely represents a “bystander” phenomenon [42]. The pathogenicity of MOG autoantibodies is the subject of ongoing debate. In vivo and in vitro studies suggest that MOG autoantibodies may cause primary demyelination in the CNS with loss of microtubule cytoskeleton in oligodendrocytes and altered expression of proteins [3, 4]. Evidence of the pathogenicity of MOG autoantibodies has also been inferred from models of experimental autoimmune encephalitis (EAE) [3, 4]. Specifically, serum from individuals with MOGAD administered in rat models of EAE has been associated with worse clinical disease and increased evidence of demyelination and axonal loss [3, 4]. In other experimental work, serum derived from pediatric patients with MOGAD has coincided with disruption to F-actin and β-tubulin networks in immortalized oligodendrocyte cells [3]. These data support potential pathologic roles for MOG antibodies in the clinical expression of MOGAD, but further research is needed.

Established and emerging biomarkers for the diagnosis and monitoring of MOGAD [3, 4, 33]

The clinical features of MS, MOGAD, and NMOSD significantly overlap. Therefore, biomarkers will be needed to distinguish these diagnoses accurately and reliably, especially in acute care settings where antibody results are often delayed (Table 4). Future research will also need to explore the impact of patient-related factors (biological sex, racial background, and pregnancy status), co-morbidities (infections), and treatment effects (long-term immunotherapies) on relapse risk and severity. In the path toward patient-centered “precision-based medicine”, it would be beneficial to identify biomarkers that not only facilitate diagnosis and predict relapses, but also prognosticate recovery for those with MOGAD.

Table 4 Established and Emerging Biomarkers to help Diagnose MOGAD, Monitor Disease Activity, and Prognosticate Recovery [2,3,4,5, 9, 12, 14, 43,44,45,46,47,48,49,50,51]
Full size table

Management of MOGAD patients

Acute treatment

The Optic Neuritis Treatment Trial (ONTT) has informed our understanding regarding the relative risks and benefits of high-dose corticosteroid therapy for patients presenting with acute optic neuritis [15]. This seminal study reported that administering IVMP (1 g/day for 3 days, followed by 1 mg/kg for 11 days) to patients presenting with acute ON expedited visual recovery, but it did not improve long-term visual outcomes [15]. Yet, the lessons of the ONTT may not be extendable to patients with MOGAD, due to study design (bilateral ON was a basis of exclusion). Notably, an examination of serum samples obtained from 177 out of the 448 ONTT participants revealed that only 3 (1.7%) were positive for MOG-IgG antibody [8]. Thus, the conclusions drawn from the ONTT regarding the role of corticosteroids in optic neuritis cannot be extrapolated to predict long-term visual outcomes in MOGAD patients. MOGAD attacks are often very steroid responsive and therefore most experts recommend high-dose corticosteroids as first-line treatment for patients presenting with acute MOGAD attacks [54,55,56,57]. Like the established NMOSD treatment approach, there is observational evidence suggesting that early management of MOGAD-related ON with IVMP is crucial for optimal visual recovery, as it has been shown that early steroid treatment is associated with better visual outcomes and less thinning of the peri-papillary retinal nerve fiber layer [9, 57,58,59]. This signifies a shift in the treatment paradigm of ONTT for ON, as it is felt that in NMOSD and MOGAD, time is of the essence to preserve vision [60]. However, it is important to note that there are cases of MOGAD ON with spontaneous improvement without steroid treatment [9, 10].

Despite the lack of a universally accepted high-dose steroid treatment regimen, the use of IVMP at a dosage of 1 g/day for 5 days is the most prevalent strategy [61, 62]. While no study has compared the efficacy of IVMP to oral corticosteroid in treating acute MOGAD attacks, Morrow et al. have reported that bioequivalent doses of oral corticosteroids are noninferior to IVMP in treating acute ON [63], and therefore 1250 mg of oral prednisone is sometimes offered as an alternative to IVMP. Further research is needed to establish the role of oral corticosteroids in the acute management of MOGAD manifestations. Patients suffering from a recent MOGAD attack are commonly considered to be at risk for relapse upon cessation or reduction of corticosteroid therapy. Therefore, considering multiple observational studies, to mitigate relapse risk, it is recommended that corticosteroids are tapered gradually over a period of 1–2 months [57, 64,65,66,67].

Most MOGAD attacks are responsive to corticosteroids, yet a subset of MOGAD patients do not respond well to IVMP, necessitating further interventions. Both plasmapheresis (PLEX) and immuno-adsorption (IA) have been proposed as therapeutic options for those failing IVMP treatment. PLEX and IA have been successfully utilized to treat immune-mediated neurological diseases, including acute NMOSD attacks [68,69,70,71,72,73,74,75]. An international multicenter retrospective study of 92 MOGAD ON attacks treated with PLEX demonstrated significant improvement in all attacks except for one eye that was treated over 6 months after the initial attack [76]. While preliminary reports are encouraging, prospective studies are required to confirm the efficacy of PLEX for acute MOGAD attacks [35]. Intravenous immunoglobulin (IVIG) is another therapeutic option that has been employed in other demyelinating conditions, such as NMOSD. While no published studies have shown IVIG efficacy in treating acute MOGAD [69, 77], a multicenter retrospective study looking at IVIG for the treatment of acute MOGAD suggests it may be associated with improved outcomes (unpublished data). Further research is required to uncover the utility and role of PLEX, IA, and IVIG in the acute management of MOGAD patients.

Maintenance therapy

The best maintenance therapy for MOGAD patients remains a topic of ongoing debate, particularly considering that approximately 50% of MOGAD patients exhibit monophasic disease, and may not require long-term immunosuppression [9, 14]. Unfortunately, it is difficult to predict if a newly diagnosed patient is destined to develop a relapsing course. Longitudinal MOG-IgG seropositivity was found to be associated with relapsing MOGAD, but many of the patients who remain seropositive do not relapse. Roughly 25% of MOGAD patients will become seronegative and are more likely to have monophasic disease [78,79,80,81]. Furthermore, findings by Cobo-Calvo et al., suggest that adults are at a higher risk of relapse and may experience worse clinical outcomes than children [80]. Long-term immunotherapy is typically initiated after a second MOGAD attack [61]. However, the type of maintenance treatment for recurrent MOGAD remains disputed, and has been extrapolated from NMOSD studies [60]. This is not ideal due to the well-described differences in pathophysiology, demographics, severity, and prognosis between the two disease entities. Common maintenance therapies employed may include oral steroids, azathioprine (AZA), mycophenolate mofetil (MMF), and B cell-targeting biologics, such as rituximab (RTX), and tocilizumab, all of which have been employed in NMOSD, in addition to maintenance IVIG, which is not a common treatment for NMOSD [60, 62].

Multiple studies have shown that maintenance therapy with oral corticosteroids has been associated with a reduction in relapse rate [64, 82, 83]. Bridging immunosuppressive agents with oral steroids is linked with better therapeutic outcomes [56, 64]. However, the potential benefits of chronic corticosteroids use must be weighed against the health risks associated with their long-term use.

Broad spectrum immunosuppressants, such as AZA and MMF, are well tolerated and have been proposed to be effective maintenance therapies for MOGAD patients [84,85,86]. Chen and colleagues reported that MOGAD patients maintained on AZA had a significant reduction in the annualized relapse rate (ARR) from 1.6 (range: 0–9.7) to 0.2 (range: 0–3.2) [84]. A modest reduction in relapse rates was noted for those treated with MMF [84]. Additionally, a prospective observational cohort study provided Class IV evidence that MMF therapy reduces relapse in MOGAD patients [87]. When initiating AZA and MMF, it is recommended to bridge these patients with prednisolone for the first few months until a decrease in lymphocyte counts is observed [56, 62]. While observational studies have shown that AZA and MMF are associated with a reduction in relapses, it is important to note that relapses can still occur and there is no randomized clinical trial data in MOGAD patients.

The efficacy of RTX in reducing relapses and EDSS scores in NMOSD patients has been well-established through RCTs [88, 89]. However, RTX’s efficacy in MOGAD patients remains ambiguous as there are no randomized clinical trials evaluating its effectiveness in this patient population. Meta-analyses have suggested that RTX reduces relapses and EDSS scores in MOGAD, though to a lesser extent than in NMOSD [90, 91]. Additionally, retrospective studies have also reported varying degrees of ARR reduction for MOGAD patients treated with RTX [61, 84, 92]. In addition, MOGAD relapses can occur despite depletion of memory B cells [93].

More recently, IVIG has emerged as a promising maintenance therapy for MOGAD patients [55]. A multicenter retrospective cohort study conducted by Chen et al. reported that maintenance IVIG therapy was associated with a reduced relapse rate in adult MOGAD patients, especially when administered at a dose of 1–2 g/kg every 4 weeks [84, 94]. Other retrospective studies have suggested maintenance IVIG therapy is effective in pediatric MOGAD patients [55]. Looking forward, IL-6 inhibitors may play a role in MOGAD maintenance therapy with growing evidence suggesting that tocilizumab might be effective for highly relapsing MOGAD patients, and satralizumab is being investigated in a multicenter randomized clinical trials [95, 96]. Additionally, rozanolixizumab, an anti-neonatal Fc receptor inhibitor, is also a promising therapy that is currently in randomized clinical trials for the treatment of relapsing MOGAD patients [14]. As evidences for therapeutic approaches continue to accumulate, randomized clinical trials with sufficient follow-up are necessary to uncover the most efficacious and well-tolerated maintenance therapy for patients with recurrent MOGAD.

Prognosis

The duration of immunosuppressive treatment for MOGAD remains uncertain, as the field currently lacks robust biomarkers to inform therapy. Specifically, it is challenging to identify patients who will remain monophasic after their first attack. Several indicators of disease activity have been proposed, including attack severity and frequency, MOG-IgG sero-status and titers, patient age, relapse-free interval, spinal cord involvement, and resultant neurological disability [61, 97,98,99]. The identification of prognostic factors indicating the likelihood of relapse and disability in a patient would catalyze a personalized treatment approach for MOGAD patients.

Adult MOGAD patients tend to have more recurrent episodes and poorer functional recovery compared to pediatric patients [82, 97, 100, 101]. MOG-IgG sero-status, particularly in adult patients, is proposed to be a predictor of disease course, with ‘seroreversion’ (from positive to negative) being indicative of a low relapse risk [67, 79, 82]. Yet, the utility of MOG-IgG titers for treatment planning remains debatable [102]. Nevertheless, recurrent disease has been associated with elevated initial MOG-IgG titers, while transiently low titers have been shown to be associated with a monophasic course [80, 82, 103].

Long-term outcomes and disability are quite challenging to predict as studies with long follow-up are sparse for patients with MOGAD. While it is believed that, on average, even frequently relapsing MOGAD results in less disability than NMOSD, inter-patient variability has been documented [16, 104, 105]. Indeed, the clinical heterogeneity of MOGAD is an increasingly recognized phenomenon [97]. Sufficiently powered prospective studies with extended follow-up would enable a more comprehensive understanding of the natural history of MOGAD, identify predictors of its course and establish guidelines for its treatment.

figure a
figure b