Summary
Our understanding of OMG and its variable phenotypes continues to evolve. Autoantibody testing increasingly provides valuable diagnostic and prognostic information. Despite these improvements, a lack of quality treatment trials creates significant challenges for evidence-based management guidelines.
Original Article
Ocular myasthenia gravis: updates on an elusive target
Current Opinion in Neurology
Andrew Melson, Collin McClelland, Michael Lee
Abstract
Purpose of review
Ocular myasthenia gravis (OMG) is a complex condition with heterogenous phenotypes and ill-defined diagnostic criteria. Understanding concomitant risk factors and autoimmune serology can help inform prognosis for generalization and guide treatment.
Recent findings
Although antibodies to acetylcholine receptors or muscle-specific kinase likely increase risk of generalization, they are less frequent in OMG. Patients without either antibody tend to have a milder disease process and often have variable antibodies to other end-plate proteins such as LRP4, agrin, or cortactin. The treatment of OMG begins with pyridostigmine and is supplemented by oral prednisone if treatment-resistant or high risk for generalization. Variable oral prednisone regimens have been used with success and further immunosuppression may be best achieved with mycophenolate mofetil and azathioprine. Checkpoint inhibitor-induced myasthenia gravis is increasingly recognized and likely has high rates of mortality associated with myocarditis.
Summary
Our understanding of OMG and its variable phenotypes continues to evolve. Autoantibody testing increasingly provides valuable diagnostic and prognostic information. Despite these improvements, a lack of quality treatment trials creates significant challenges for evidence-based management guidelines.
INTRODUCTION
Myasthenia gravis is an autoimmune disorder characterized by fatigable muscle weakness. Auto-antibodies to proteins in the postsynaptic terminal of the neuromuscular junction (NMJ) impair neuromuscular transmission. Ocular myasthenia gravis (OMG) is a subset of myasthenia gravis where patients experience symptoms of ptosis and/or diplopia related to exclusive involvement of the extraocular muscles (EOMs) and the orbicularis oculi. In contrast, generalized myasthenia gravis (GMG) involves the bulbar, limb and/or respiratory muscles.
The final common pathway in myasthenia gravis is reduction in the concentration or function of the nicotinic acetylcholine receptors (AChR) at the NMJ. Although this most often results from autoantibodies to AChR, numerous other protein targets have been identified in the cellular signaling pathways and clustering of AChR to the muscle endplate. This is of particular relevance in OMG, where the pathophysiology underlying the predilection for isolated ocular involvement is not fully understood, and a relatively large proportion do not have identifiable antibodies to AChR. Although one of the endplate proteins, muscle-specific kinase (MuSK), has pathogenic antibodies that have been well established, recent literature has focused on the potential pathogenicity and prognostic implications of other auto-antibodies. As the array of antibodies to endplate proteins associated with myasthenia gravis continues to expand, we discuss an appropriate approach to serologic work up and summarize recent literature regarding risk factors for disease generalization.
The challenges of defining evidence-based diagnostic and treatment algorithms for OMG extend far beyond the nebulous pathophysiology. The fluctuating, heterogenous, and often masquerading manifestations of OMG have led to ill-defined and inconsistent diagnostic criteria among studies. We will examine the recent and evolving literature on the diagnosis and management of OMG including the increasingly recognized checkpoint inhibitor-induced myasthenia gravis and a review of the challenges facing OMG evidence-based management.
AUTO-ANTIBODIES AND THE ROLE OF SEROLOGIC TESTING
Because AChR antibodies were first described in 1973 [1], auto-immune antibodies play an important role in the diagnosis and pathogenesis of myasthenia gravis. As our understanding of the NMJ has evolved, many clinically relevant proteins integral to its function have been identified: acetylcholine receptor (AChR), muscle-specific kinase (MuSK), low-density lipoprotein receptor-related protein 4 (LRP4), agrin, and cortactin. Specifically, the binding of agrin to the membrane protein LRP4 leads to an activation of MuSK, which phosphorylates cortactin in the process of recruiting AChR to clusters within the motor neuron synapse [2▪]. This clustering facilitates muscle excitability by allowing AChR to adequately potentiate neurotransmission in the NMJ. Thus, these muscle end-plate proteins serve as potential immunologic targets for both pathogenic auto-immune processes and therapeutic interventions.
The frequency of AChR antibody seropositivity in patients with GMG is approximately 80–90% [3]. MuSK antibodies account for the next largest proportion and occur in 5–8% [4]. For unclear reasons, patients with OMG have lower seropositivity for AChR antibodies, with recent reports ranging 40–70% [5▪,6▪,7,8,9▪,10▪]. Historically, the most common MuSK-positive myasthenia gravis phenotype is a woman in the fourth decade of life with predominant cranial and bulbar weakness [4]. The perceived rarity of seropositivity with isolated ocular symptoms deters some providers from ordering MuSK antibodies in the initial work up of OMG. However, screening using cell-based assays (using live cells expressing proteins of interest) have improved sensitivity and call into question whether MuSK-positive OMG may be an underdiagnosed, broader phenotype than previously suspected [4,11▪]. Patients with MuSK antibodies reportedly have more symmetric disease and progress to generalization at higher rates, stressing the importance of improved screening to facilitate early treatment of at-risk populations [5▪]. Rarely, patients can be ‘double seropositive’ with antibodies to both AChR and MuSK, presenting most frequently with nonocular symptoms [12].
Patients negative for both AChR and MuSK are termed double-seronegative myasthenia gravis (dSNMG). In OMG, this constitutes a large population where biomarkers for additional risk stratification would prove valuable. Therefore, many authors have examined antibodies to the membrane protein, LRP4. Although LRP4 antibodies have not been fully validated as pathogenic, animal research and clinical reviews heavily supports this notion [13,14]. Notwithstanding their likely pathogenicity, antibodies to LRP4 have unclear specificity for OMG as they have also been found widely in amyotrophic lateral sclerosis [15–16]. A recent systematic review found that anywhere from 0 to 50% of patients with dSNMG have anti-LRP4 antibodies [2▪]. The authors concluded that prototypical patients tend to be young women with mild, isolated ocular symptoms that are responsive to pyridostigmine and oral corticosteroids. Similarly, 20% of a cohort of patients with dSNMG were reported to have anticortactin antibodies, and this subgroup seemed to have mild disease without bulbar symptoms [17▪▪]. Seropositivity to either LRP4 or MuSK is uncommon in the general population of patients with myasthenia gravis (less than 10% each), making the frequency of double seropositivity very rare and the clinical significance remains unclear. For patients with seronegative disease, more recent work evaluating microRNA, specifically miR-30e-5p as a potential biomarker, shows promising results [11▪,18].
The heterogeneity of myasthenia gravis phenotypes in conjunction with a relatively high proportion of patients with OMG without known pathogenic antibodies may suggest a fundamental difference in the underlying pathophysiology and treatment response. Specifically, a lack of directly pathogenic antibodies in some patients with OMG might explain why a subgroup of these patients experience treatment resistance where robust immunosuppressive therapeutic regimens fail to alleviate the symptoms. Until further high-quality research identifies the phenotypic and prognostic implications of these various end-plate proteins, these authors suggest limiting initial antibody serology to Anti-AChR and Anti-MuSK antibodies. Subsequent testing for anti-LRP4 and anticortactin can be considered in cases of dSNMG on a case-by-case basis.
Although advancements in serology continue to identify a larger proportion of patients with OMG, there remains a large subset who are widely seronegative. For these patients, a skillful clinical assessment is essential to raising diagnostic suspicion. Testing with electromyography, ice-packs, and short-lived acetylcholinesterase inhibitors such as edrophonium or neostigmine remain viable options to support a clinical diagnosis in the absence of positive serology. Unfortunately, these ancillary tests often have variable accuracy, tolerability, and availability, which limit their utility. As such, seronegative OMG remains a clinical diagnosis.
FREQUENCY AND RISK FACTORS FOR GENERALIZATION
The progression from isolated ocular to generalized myasthenia gravis has significant prognostic implications given the potential for life-threatening respiratory muscle involvement and the frequent requirement for immunosuppressive therapy. Classic studies indicated that 50% of patients with myasthenia gravis present with isolated ocular symptoms, and 50–60% of those patients with OMG develop GMG within 2 years of symptom onset [19–22]. Though these studies shaped our initial understanding of the natural history of all individuals with myasthenia gravis, they were not designed to evaluate the predictive value of clinical or serologic features on conversion rates. Additionally, the consideration of which patients ‘convert’ from OMG to GMG requires a clear definition of each clinical entity. Unfortunately, there are no well defined diagnostic criteria for OMG which limits homogeneity of study design and inclusion criteria.
There are many factors that limit the utility and accuracy of literature on conversion rates to GMG. The majority suffer from flaws inherent with retrospective study design and lack well defined, clinically relevant criteria for diagnosis. One of the most common challenges is variable inclusion requirements for duration of isolated ocular symptoms. In the absence of guidelines specifying otherwise, many studies considered all patients who had isolated ocular symptoms at initial visit or up to one month after symptom onset to be OMG and uniformly reported high generalization rates (46–74%) [9▪,10▪,23▪,24▪▪]. However, some studies required subjects to maintain purely ocular involvement for between 3 months and 2 years to be categorized as OMG and had much lower conversion rates (15–21%) [5▪,6▪,11▪]. One study did not specify a minimum timeframe for purely ocular symptomatology and had a generalization rate of 10% [8]. Some contribution to the variability may be geographic [8,10▪,23▪], although no specific pathophysiologic explanation has been offered. In one study with inclusion criteria only requiring isolated ocular symptoms at initial visit, 42% of the patients who generalized did so in the first 2 months [9▪]. Taken on the whole, recent literature would suggest that conversion rates might be lower for patients who have not generalized by 3 months compared to those with less than a month of isolated ocular symptoms.
In an effort to further risk stratify patients, several recent studies have retrospectively analyzed OMG cohorts for specific clinical and immunologic predictors of generalization. Unfortunately, these studies share many of the same retrospective limitations and have identified highly-inconsistent, and at times conflicting, independent risk factors for progression (Table 1). Amongst the variable outcome measures and results, a few common trends emerged. First, patients with AChR antibodies seemed to have a higher likelihood of conversion [5▪,6▪,9▪,10▪,11▪,24▪▪]. Two studies showed a positive correlation between AChR titers [10▪,23▪] and generalization and female gender [5▪,6▪], although other studies did not show gender predilection [8,10▪,23▪]. Similarly, patients with MuSK antibodies generally had very high rates of conversion [5▪,10▪], although one study utilizing cell-based assay testing showed only 10% generalized [11▪]. Furthermore, the presence of a thymoma was associated with increased risk of progression in three of the four studies [8,10▪,23▪]. Age at onset showed mixed results, with at least one study supporting each conflicting position [6▪,9▪,10▪,11▪]. Prospective, controlled studies with uniform data collection and inclusion criteria across multiple centers would help clarify these risk factors to guide patient education and treatment options.
Review of recent publications reporting risk factors for generalization of OMG
OCULAR ADVERSE EVENTS WITH IMMUNE CHECKPOINT INHIBITOR THERAPY
Pembrolizumab and nivolumab are monoclonal antibodies to programed cell death protein 1 (PD-1) called immune checkpoint inhibitors (ICIs). They are predominately used to treat advanced and metastatic carcinomas by activating the immune system to target cancer cells. Despite their efficacy, a growing number of patients have experienced significant immune-related adverse events secondary to induction of autoimmunity. Of particular interest, ICI use appears to induce a spectrum of neuromuscular disorders including myasthenia gravis [25–29], with nearly 50% of cases involving oculomotor symptoms [30▪▪]. Drug-induced myasthenia gravis from ICI tends to have a severe phenotype, often coexisting with myositis and myocarditis. Although it can occur in patients with known myasthenia gravis or de novo, a high proportion of patients have positive AChR antibodies [30▪▪].
Unlike typical myasthenia gravis, ICI induced myasthenia gravis with isolated ocular involvement can rapidly progress to life-threatening respiratory or cardiac involvement [29]. Patients presenting with signs of possible OMG and ICI use should undergo a broad diagnostic work up for coexisting myositis and myocarditis. Despite drug cessation and appropriate treatment with immunosuppression, ICI-induced myasthenia gravis reportedly showed mortality rates above 40% [27], but this may either relate directly to adverse drug effects or underlying disease. Future research into this unique scenario may yield interesting information regarding pathophysiology, unmasked latent versus de novo disease, and potential paraneoplastic features.
CHALLENGES IN MANAGING OCULAR MYASTHENIA GRAVIS
A wide variety of treatment options are available for OMG. Because the disease is limited to the oculomotor system, interventions can either treat the underlying disease or directly address the resulting ptosis and/or diplopia. With ptosis, the primary issue is variable occlusion of the visual axis. As such, the problem is alleviated through any means of elevating the lid above this axis including mechanical or surgical techniques. An open-source, 3D printed, modular ptosis crutch has been developed that mechanically elevates the eyelids. It improves upon the skilled-labor intensive individualized ptosis crutch by adapting to a wide variety of spectacle frames and facial structures while being highly affordable to anyone with access to a 3D printer [31]. Another simple technique involves the utilization of cosmetic eyelid tape to mechanically clear the visual axis. Diplopia, however, occurs with highly variable ophthalmoplegia patterns that lead to incomitant strabismus in OMG. Mechanical interventions for diplopia are limited to monocular occlusion or prismatic correction which often fail to alleviate diplopia symptoms (prisms) or result in complete loss of binocularity and depth perception (occlusion). The variable nature of OMG symptoms makes it difficult to find a single mechanical solution to address all potential ocular misalignments.
The ideal treatment strategy for OMG is to restore normal muscle function, either through pharmacologic enhancement of muscle contractility or suppression of the autoimmune processes causing NMJ dysfunction. First-line treatment for OMG consists of acetylcholinesterase inhibitors with or without oral corticosteroids (Fig. 1). Pyridostigmine dosing starts at 60 mg three times daily and can escalate to 180 mg every 4 h. Although ptosis is often highly responsive to pyridostigmine alone, ophthalmoplegia is frequently more resistant and likely to necessitate immunosuppressive therapy [32▪]. Corticosteroids are the current immunosuppressant of choice and have been shown to be effective [32▪,33,34], although there is wide debate as to timing, initial dosage, and effect on prevention of generalized disease. In the absence of clear guidelines, treatment regimens typically fall into two established paradigms: increase prednisone to 60 mg daily within a week then slowly taper over a few months to a dose below 7.5 mg daily [35] or, start prednisone at a low dose of 10–20 mg and slowly increase until symptoms resolves, then taper back down to a dose below 7.5 mg daily [36]. Moderate doses of oral corticosteroids such as 20–40 mg of prednisone per day for 8 weeks may represent a compromise between the two aforementioned commonly used strategies and have been employed with success in AChR antibody positive OMG [33].
Stepwise treatment algorithm for nonthymomatous ocular myasthenia gravis.
If corticosteroids fail, are poorly tolerated, or are unable to be weaned to an acceptable chronic dose due to residual ophthalmoplegia, a host of other immunosuppressive medications have purported benefit for OMG, with mycophenolate mofetil and azathioprine having superior safety profiles. Mycophenolate mofetil has a typical starting dose of 1000 mg twice daily and requires no enzymatic testing prior to initiation. Azathioprine is started at 1 mg/kg daily and patients should be tested for low thiopurine methyltransferase activity to avoid potential toxicity. While both mycophenolate mofetil and azathioprine can take months to show maximum effect, prednisone can deliver beneficial effects within a few weeks [35] and is often started concurrently.
The evidence for other treatment modalities is primarily extrapolated from GMG studies, since the literature evaluating their use in isolated OMG is limited. This is particularly relevant to advanced immunomodulatory treatments developed primarily for refractory GMG. While medications such as eculizumab, a new complement inhibitor, show promise in clinical trials limited to AChR+ GMG patients [37], their efficacy has not been evaluated in OMG patients who were excluded from participation. Similarly, thymectomy for OMG specifically has not been well studied. Although, the MGTX trial [39] demonstrates that thymectomy plus prednisone confers a significant benefit to nonthymomatous GMG patients, a recent retrospective report of nonthymomatous OMG patients found no difference in prednisone dose or symptom severity after thymectomy [40]. The only meta-analysis of thymectomy for nonthymomatous OMG patients concluded that it may be a viable treatment option but found insufficient evidence to make a conclusive recommendation [41]. Until further high-quality evidence based guidelines are published for OMG, thymectomy is generally reserved for patients with thymoma although regional clinical practices vary.
There are many challenges that limit evidence-based approaches to the diagnosis and treatment of OMG. Between the highly variable clinical presentations, inconsistent response to treatment, and the difficulty in ascertaining the impact of ocular sequelae, meaningful prospective research on OMG is quite limited. While the rare disease designation for myasthenia gravis has accelerated research into controlled trials and immune therapies for GMG [38▪], the inherent risk/benefit discussion for treatment modalities in OMG is drastically different and requires individualized research beyond subgroup analyses. Despite tangible needs identified in several articles [42▪,43,44] and preliminary research showing feasibility [45], there remains an absence of randomized control trials for OMG (Table 2).
Summary of ocular myasthenia gravis specific treatment publications over last 20 years
CONCLUSION
OMG is a heterogenous condition with unclear determinates of progression to generalized disease. Although antibodies to AChR and MuSK likely increase risk of generalization, they are less frequent in OMG. Research on antibodies to an increasing array of end-plate proteins has yet to prove pathogenicity, though they may hold value as biomarkers for severity of disease. Evidence-based clinical decision making remains markedly limited by an absence of randomized, controlled clinical trials for the treatment of OMG. Medical treatment should start with pyridostigmine plus or minus corticosteroids, with the option of adding mycophenolate mofetil or azathioprine if needed.
