Disorders affecting the neuromuscular junction (NMJ) are generally pure motor syndromes that usually affect the extraocular muscles but also the proximal limb and bulbar or respiratory function. They can be classified into autoimmune, acquired, toxic, and inherited disorders of the NMJ. The most common NMJ disorders are autoimmune and therefore respond to immunosuppressive therapy. These include myasthenia gravis (MG) and, rarely, the Lambert-Eaton myasthenic syndrome (LEMS). A number of genetically determined disorders of neuromuscular transmission, the congenital myasthenic syndromes, are seen in childhood but may rarely present in adult life. Botulism is a toxin-mediated disorder of the NMJ. All NMJ disorders cause generalized weakness and fatigability with a propensity for oculobulbar involvement. Electrophysiologic studies can detect an impairment of neuromuscular transmission in most of these disorders.¹ Fortunately, most of these disorders are treatable^2–4 (Fig. 74–1).

MG is the best understood autoimmune disease among these.⁵ The immune-mediated nature of MG was suspected as early as 1960 when Simpson speculated that it was an autoimmune disease with antibodies directed against the skeletal muscle acetylcholine receptor (AChR).⁶ A series of breakthroughs in the 1970s confirmed Simpson’s hypothesis. Lindstrom et al developed the animal model experimental autoimmune MG by immunizing rabbits and rats with highly purified AChR from the electric organ of the eel.^7,8 Subsequently, high AChR antibody titers were found in the serum of MG patients.^9,10 Engel et al localized both the IgG antibody and complement to the myasthenia motor endplate.^11,12 This implied that circulating IgG antibody directed against the AChR bound to the postsynaptic membrane and activated the terminal complement sequence (C5b-9), or membrane attack complex (MAC), resulting in lysis of the AChR with subsequent degeneration (Fig. 74–2).

Elevated MAC levels have been demonstrated in the plasma of MG patients.¹³ AChR antibody has also been shown to block neuromuscular transmission and accelerate turnover of AChR cross-linked by IgG.¹⁴ As a result of this process, the postsynaptic membrane becomes simplified with decreased junctional folds and widening of the synaptic cleft¹⁵ (Fig. 74–3).

In addition, the neuromuscular blockade is passively transferred by injecting animals with IgG from MG patients.¹⁶ The same phenomenon occurs when an infant born to a mother with MG exhibits symptoms at birth, so-called neonatal MG.¹⁷

The AChR is a large protein consisting of five subunits, and the antibody response in MG and experimental autoimmune MG is polyclonal. The portion of the protein primarily responsible for inducing antibodies that produce the disease is debated. Although the existence of a main immunogenic region in the alpha subunit has been promoted,¹⁸ other investigators have challenged this evidence.¹⁹ It is possible that if the most pathogenic determinants of the AChR can be identified, a more rational and specific immune therapy can be designed. Multiple mechanisms are known to cause loss of functional AChRs in myasthenia gravis. These include complement-mediated lysis, accelerated internalization and degradation of AChRs, and finally, direct blockade of AChRs by antibodies.^20,21 Among these, complement-mediated lysis is thought to be the most important mode of loss of AChRs.

The process that initiates the immune-mediated NMJ dysfunction is still unknown. The thymus gland may play a role, because 75% of MG patients who undergo thymectomy have thymic pathologic findings, with 15% being tumors of the thymus and the remainder consisting of lymphoid hyperplasia.²² Lymphocytes in the thymus and peripheral blood appear to be sensitized to muscle in MG patients.^23,24 Muscle-like myoid cells are found in the thymus gland, and thymus tissue from MG patients with and without thymoma is enriched in AChR-reactive T cells.^25,26 The close association of lymphocytes and myoid cells in the thymus, along with some stimulus causing the disruption of immune tolerance, may lead to the autoimmune response. There may be a hereditary predisposition to develop MG because there is an increased incidence of certain human leukocyte antigens (HLAs) in various MG populations.²¹ In the overall US case-control cohort,²⁷ genome-wide association study (GWAS) identified associations on signals at CTLA4, HLA-DQA1, and TNFRSF11A. CTLA4 and HLA-DQA1 were replicated for an independent cohort of Italian cases with controls.

MG is a rare disorder with a prevalence of approximately 125 cases per 1 million population.²⁸ Approximately 11% to 24% of all MG patients have disease onset in childhood or adolescence.^29,30 There is a slight female predominance of 3:2, although males predominate in older age groups. The disease can arise at any age, but peaks are observed in the third and sixth decades. About 5% to 10% of MG cases are associated with other autoimmune conditions. Studies have estimated that only between 3.8% and 7.1% of MG patients have a family history of the disease,^31,32 whereas the GWAS study yielded a higher estimate of the heritability associated with MG risk (25.6%) by capturing more of the polygenic variance.²⁷

Natural History and Prognosis

The natural course of MG is not well known. In early series the diagnosis was limited to the more severely affected patients, with mortality rates of 30% to 40%.^33,34 One long-term follow-up study of MG from Netherlands³⁵ with 73 MG patients reported maximum severity of disease during the first 7 years after onset in 87%; 25% achieved complete remission, 18% improved considerably, 16% improved moderately, 16% remained unchanged, and 2 patients had deteriorated. Patients with thymoma have a less favorable course. Since the introduction of anticholinesterases in 1934, the diagnosis has been facilitated even in mild and less prominent cases. Thymectomy was thought to have improved the natural course in early-onset cases without thymoma. Improved intensive care facilities and plasma exchange are of benefit especially to the 20% of severely affected patients with intermittent respiratory insufficiency. The use of improved diagnostic tools such as determination of antibodies to the AChR protein and to muscle-specific tyrosine kinase and recently antibodies to lipoprotein-related protein 4 (LRP4), single-fiber electromyography (EMG), and repetitive nerve stimulation provide for more accurate diagnosis in mild cases and exclude other myasthenic syndromes. Newer therapeutic options also improved the prognosis for these patients. Patients with other associated autoimmune conditions may have a less favorable course.

Symptoms and Physical Findings

MG is characterized by weakness with fatigability of ocular, bulbar, and extremity muscles. The ocular manifestations are ptosis and diplopia, whereas the bulbar manifestations are dysarthria, dysphagia, and dyspnea. Some patients present with jaw fatigue and jaw closure weakness. Proximal limb and axial muscles tend to be weaker than distal extremity muscles. Symptoms of MG tend to worsen with stress, exertion, infections, major surgery, and as the day progresses. Rest may alleviate the symptoms initially. These temporal symptoms, however, may be difficult to elicit in many patients, especially those with severe disease. Myasthenic crisis is characterized by respiratory weakness and an inability to swallow.

Grob et al carefully studied 1976 patients between 1940 and 2000.³⁶ Most patients (85%) present with ocular symptoms. Initial symptoms were ptosis in 32%, diplopia in 14%, ptosis and diplopia in 36%, and blurred vision in 3%. Other patients presented with leg, arm, face, neck, or trunk weakness, bulbar symptoms, and generalized fatigue. Of the patients who presented with purely ocular disease, 56% progressed to generalized disease by 6 months, 78% by the first year; 13% remained purely ocular; 50% had ocular and generalized symptoms; and 13% had ocular and bulbar symptoms. Time to maximum severity also occurred fairly early in the disease course. Maximum or near-maximum severity was reached by 6 months in 37% of patients and by 1 year in 66% of patients. Only 18% of patients reached maximum severity of weakness after 2 years. In our retrospective neuroophthalmological case series of ocular MG, only a small proportion of ocular MG cases progressed to generalized MG, and that likelihood was lessened by immunosuppressive therapy (12 of 59 not treated with prednisone vs. 0 of 38 treated with prednisone).^3,37

Mortality statistics for MG have fallen dramatically over time.³⁸ Between 1940 and 1957, the mortality rate was 31%, whereas between 1966 and 1985, the mortality rate was 7%. The two primary reasons for this reduced mortality rate were the improvement in intensive respiratory care and the introduction of corticosteroids. Death from MG is uncommon in current practice. Ten percent of patients seen between 1940 and 2000 had a thymoma. The overall course of the patients with thymoma was worse than for those with nonthymomatous generalized disease.

A task force of the Myasthenia Gravis Foundation of America (MGFA) developed a classification system^39,40 (Table 74–1). This replaced the Osserman classification system.⁴¹ On examination, it is important to determine whether lid ptosis, upper lid fatigability with sustained upward gaze, any restriction of ocular motility, and diplopia are present in primary position, on horizontal gaze to the right or left, or on vertical gaze.

Testing for weakness in the orbicularis oculi muscle is critical but often overlooked. Many symptomatic MG patients have bilateral weakness of this muscle group. Strength should also be tested in the lower facial muscles (blowing out cheeks against resistance) and in the tongue. Attention to speech patterns may disclose a nasal dysarthria. It is important to check for neck flexion and extension weakness because these muscle groups are frequently involved. Testing of extremity strength should include proximal and distal muscle groups in the arms and legs. Proximal limb muscles tend to be more affected than distal muscles. Rarely, MG patients may demonstrate a predilection for weakness in distal muscle groups, especially finger extensors.⁴²

Several systems have been used for scoring the severity of MG, especially for trials, including the activities of daily living (MG-ADLs)⁴³ and the quantitative MG score.^40,44 The newer composite MG score with ten items has recently been validated in adults.⁴⁵

In most instances, the clinician can be confident about the diagnosis of MG based on abnormalities brought out through the neurologic history and physical examination. However, one or more tests are usually performed to confirm the clinical diagnosis.

Edrophonium Chloride (Enlon) Test

The intravenous (IV) administration of up to 10 mg of edrophonium chloride is a diagnostic test in the evaluation of a potential MG patient. It is not used as frequently now with the advent of antibody testing. The edrophonium test can have a number of pitfalls. The most common mistake is that the physician performing the test does not have an objective parameter to measure before and after edrophonium administration. The most useful parameter is the degree of ptosis in each eye. The best indication of a positive test is a significant increase in the palpebral fissure aperture or the opening of a completely ptotic eye. If no ptosis is present, the edrophonium test may be difficult to interpret even in clear-cut cases of MG. If the patient has a severe restriction of extraocular movement and edrophonium dramatically improves the motility, the test is considered positive. However, subjective diplopia may not resolve unless edrophonium produces orthophoria in the eyes, which is rare. Significant improvement in dysarthria or in swallowing is another indication of a positive edrophonium test. A mild improvement in limb strength or subjective well-being is not sufficient to claim a positive test. In addition, a positive edrophonium test is not specific because transient subjective improvement is reported in other neurologic disorders, such as motor neuron disease and peripheral neuropathy.⁴⁶

The edrophonium test can be performed easily in the outpatient setting and does not need to be done in the hospital setting. In adults, 1 mL (10 mg/mL) of edrophonium is drawn up into a 1-mL tuberculin syringe. Initially, 0.2 mL of the edrophonium is injected directly into a vein (usually the antecubital). If after 30 to 60 seconds the patient experiences no side effects from the drug (e.g., fasciculations, sweating, or nausea), the rest of the 0.8 mL is injected. When injecting into an IV line, line flushes with saline are needed with each aliquot. More serious side effects, such as bronchospasm or light-headedness caused by bradycardia, are quite uncommon, but atropine should be readily available. When either side effects or a positive response are obtained, no further edrophonium need be given. Atropine is kept on hand for significant bradycardia, but in our experience it is rarely, if ever, needed.

Edrophonium used to be supplied by a manufacturer that called the drug Tensilon, later another manufacturer of the drug used the name Enlon. However, even Enlon has been discontinued recently.

If a patient has ptosis, it is critical that the palpebral fissure aperture is measured and its size recorded before and after edrophonium administration. In patients with less objective findings that do not permit easy measurement, edrophonium probably should not be given in the first place. Thus, a placebo injection is rarely necessary.

In infants and younger children who are uncooperative and difficult to monitor over brief time periods, longer-acting neostigmine may be preferred. The intramuscular (IM) dose is 0.15 mg/kg and the IV dose 0.05 mg/kg.⁴⁷ IV use can be hazardous due to severe muscarinic side effects.⁴⁸ A positive response is generally evident by 15 minutes and is most obvious after 30 minutes. It is a good idea to have injectable atropine available in the case of severe side effects whenever using neostigmine.

Electrophysiologic Testing

Repetitive Stimulation

The classic electrophysiologic demonstration of an NMJ transmission defect is the documentation of a decremental response of the compound muscle action potential (CMAP) to repetitive stimulation (RS) of a motor nerve⁴⁹ (Fig. 74–4). This typical pattern is in contrast to the marked increase in the CMAP amplitude seen in Lambert-Eaton myasthenic syndrome with brief exercise or with fast repetitive nerve stimulation (Fig. 74–5). The decrement is due to failure of some muscle fibers to reach threshold and contract when successive volleys of ACh vesicles are released at the NMJ. Failure to reach the endplate threshold (EPP) to achieve muscle contraction is called blocking. The percent decrease in amplitude and area is calculated between the first CMAP produced by a train of stimuli and each successive stimulus. In most laboratories, five responses are obtained at 2 or 3 Hz, and the maximal percent decrement can be measured at the fourth or fifth response. A decrement of greater than 10% is considered a positive RS study.

In some patients, a decremental response can be demonstrated at baseline. However, often a brief period of exercise (usually 1 minute) is required to fatigue the NMJ so that the decrement can be observed. This phenomenon of postexercise exhaustion (PEE) usually occurs at 2 to 4 minutes after exercise. In addition, repair or an improvement in the decrement can sometimes be observed immediately (within seconds) after brief exercise (Fig. 74–6).

RS is typically first recorded in a distal thenar or hypothenar muscle after stimulating the median or ulnar nerve, respectively, for generalized MG. For ocular MG, typically an orbicularis oculi or nasalis response is recorded after stimulating the facial nerve. If no decrement is observed, RS can be performed on a proximal limb muscle (i.e., trapezius, face, deltoid, biceps). An arm board is used to immobilize the hand muscles. False-positive results are more of a problem in the proximal limb muscles because of motion artifact.

Because RS is a reflection of the integrity of NMJ transmission, a decrement is more often observed in clinically weak muscles. Thus, even if a patient has generalized MG, if there is only facial and proximal limb weakness, a decrement in a hand muscle is unlikely. In a pure ocular MG patient, a decrement may not be present in the orbicularis oculi unless that muscle is weak on examination.

A decremental response is more likely present in a proximal muscle than in a distal muscle. In the series by Stalberg and Sanders, a decrement in a distal muscle was reported in 38% of patients, whereas a decrement in proximal muscles occurred in 64%.⁵⁰ Similar findings have been described by other authors.^51,52 In ocular MG, decrements are less common, occurring in 20% to 50% of patients.^50,53 Facial muscle RS should be included when clinical suspicion for anti-MuSK myasthenia exists because facial muscles are much more prominently involved in this group.⁵⁴ RS at faster rates (i.e., 20 or 50 Hz) is performed when there is concern about LEMS.

Single-Fiber EMG

Single-fiber EMG (Figs. 74–7 and 74–8) is a more sensitive measure of neuromuscular transmission than RS and can be considered if other testing is negative and clinical suspicion is high for MG. In MG, the time required for the EPP at the NMJ to reach threshold is extremely variable. Measurement of this variability in the EPP rise time is known as jitter. The jitter value, calculated in microseconds, is the most important piece of data obtained from single-fiber EMG. Everyone, including healthy individuals, has some degree of jitter. MG patients have increased jitter values. In addition, blocking occurs in myasthenics if a muscle fiber’s EPP never reaches threshold and depolarization does not occur. The frequency of blocking, expressed as a percentage is also determined with single-fiber EMG. In healthy individuals, the percentage of blocking is zero.

Single-fiber EMG is undoubtedly the most sensitive test for MG in adults. It is abnormal in 94% of generalized and 80% of ocular MG patients.^51,55 However, single-fiber EMG has several disadvantages. It is a tedious and lengthy study that requires considerable patient cooperation and is poorly tolerated by many people. The need to use nondisposable single-fiber electrodes was also a limitation, but recent normative data have been presented for single-fiber studies with disposable concentric needles.⁵⁶ Stimulated single-fiber EMG can be performed under sedation, requires less patient cooperation, and may be preferred in children, although it is still a lengthy procedure.⁵⁷ An abnormal single-fiber EMG study is not specific for MG because increased jitter commonly occurs as a result of other neuromuscular diseases, including motor neuron disease, peripheral neuropathy, and many myopathies.⁴⁹ Conventional needle EMG has limited diagnostic value in MG, but short-duration, small-amplitude early recruited myopathic units can be seen in MuSK MG patients⁵⁸ and in severe MG cases.

Antibody Testing

Anti-AChR Antibodies

Finding elevated AChR antibody levels in the serum of a suspected MG patient is the most specific diagnostic test. AChR antibody levels are not elevated in all MG patients. The assay is most helpful in adult generalized MG; it is positive in 85% of such patients.^9,28,51,59 Ocular MG patients, however, have a measurable AChR antibody in only 50% of cases.⁶⁰ Seronegativity is more common in pure ocular forms, mild disease, and remission.⁶¹ Because congenital myasthenic syndromes and seronegative autoimmune MG present in early childhood, differentiating these disorders when the family history is negative is often difficult.⁴⁷ Fluctuating weakness or disease severity and good responses to immunotherapy favor an autoimmune basis.⁶²

The most common AChR antibody test is the binding radioimmunoassay using bungarotoxin, measured in nanomoles per liter. The upper limit of normal varies among reference laboratories (usually between 0.03 and 0.5 nmol/L).⁶³ Other assays that block bungarotoxin binding to AChR (blocking assay) or that reduce the density of AChR on cultured human myotubes (modulating antibody assay) are also commercially available.⁶⁴ These additional assays may be useful in patients with suspected MG who test negative with the standard binding assay,⁶⁴ but they do not add significantly to the diagnostic sensitivity. Some laboratories offer all three (binding, blocking, and modulating) antibodies as one serologic test. Recently, low-affinity AChR antibodies against rapsyn-clustered AChRs were seen in 66% of patients who were otherwise seronegative. These were mainly IgG₁ antibodies that can activate complement C3b deposition.⁶⁵

AChR antibody titers correlate poorly with MG severity.⁶⁶ Although the titer often falls as the clinical condition improves, antibody titers in general do not guide therapeutic decisions. Indeed, MG patients in clinical remission may still have elevated titers, but this is not an indication to continue immunosuppressive therapy.

Anti-MuSK Antibodies

Since 2001, IgG from 40% to 70% of seronegative generalized patients has been found to bind to the extracellular domain of muscle-specific receptor tyrosine kinase (MuSK)^67–69 or 7% of all generalized MG cases. It has been hypothesized that anti-MuSK antibodies impede agrin-mediated clustering of AChR and disrupt normal postsynaptic architecture.⁷⁰ Marked female predominance with a mean age of onset in the fourth decade has been typical.^69,71 The earliest reported onset of anti-MuSK MG is 2 years.⁷² Three main patterns of anti-MuSK MG have been observed; one of them is clinically indistinguishable from anti-AChR generalized MG. The other two patterns are severe oculobulbar weakness and prominent neck, shoulder, and respiratory involvement largely sparing ocular musculature. In these two phenotypic variants, limb strength is relatively intact.^54,69 Anti-MuSK antibodies are rarely seen in pure ocular MG.⁷³ MuSK MG is somewhat more refractory to conventional treatment when compared with AChR MG.⁷⁴ Testing for anti-MuSK antibodies should be considered in all suspected MG patients who are AChR antibody negative.

Striated Muscle Antibodies and Other Laboratory Studies

Antibodies to striated muscle in MG patients were discovered before AChR antibodies. These antibodies can be directed against a number of muscle proteins, including myosin, actin, alpha-actinin, titin, and ryanodine (RyR). It is generally believed that if anti–striated muscle antibodies are present in an MG patient, they should raise suspicion for thymoma because they are reported in up to 84% of patients with thymoma.⁷⁵ However, these antibodies may be found in patients without thymoma and in patients with thymoma who do not have MG.^76,77 The absence of anti–striated muscle antibodies also does not rule out thymoma. Anti-titin and anti-RyR antibodies have been observed as markers for more severe disease in MG patients presenting after age 40.⁷⁶ Thyroid function tests are routinely obtained at the time of initial evaluation because thyroid disease often coexists with MG.²¹

Lipoprotein-Related Protein 4

Low-density lipoprotein-related protein 4 (LRP4) is a recently identified antibody. LRP4 interacts with agrin, and this activates MuSK and promotes the clustering of the AChRs and their stabilization at the NMJ. Anti-LRP4 antibodies are found in approximately 9.2% (range 2%–50%) of MG patients who are negative for both anti-AChR and anti-MuSK antibodies.⁷⁸ LRP4 antibody testing has recently become commercially available.

Treatment has undergone major advances during the past 75 to 80 years. The 1930s saw the initial era of acetylcholinesterase inhibitors through the work of Mary Walker.^79,80 In the 1930s and 1940s, Blalock introduced the modern era of thymectomy for MG, although in 1913 a partial thymectomy was reported to improve a patient with both MG and hyperthyroidism. Thymectomy continued to be a frequently performed procedure for MG, although there was a great deal of controversy as to its benefit⁸¹ until the recent thymectomy prospective, randomized, single-blind study was published.⁸² New forms of acetylcholinesterase inhibitors were introduced in the 1950s—the rapid-acting intravenous form edrophonium chloride and oral pyridostigmine.^83–85 Arguably the biggest advance in the1950s was the development of mechanical ventilation in intensive care units (ICUs) so that myasthenics in crisis could be managed and essentially kept alive when their breathing became severely compromised. Corticosteroids and plasmapheresis (PE) were introduced in 1960s and 1970s.

Throughout the next several decades, a number of drugs from the transplant rejection literature were introduced to suppress the immune system. The first of these drugs was azathioprine (AZA),⁸⁶ followed by cyclosporine (CSA),^87,88 and then mycophenolate mofetil (MM).^89–92 In the 1980s and 1990s, intravenous immunoglobulin (IVIg) began being used in MG patients.^93–97

Due to a combination of all these new therapeutic advances, mortality from MG has dropped significantly during the past 50 years. Prior to 1960, the mortality from MG was estimated to be more than 30%—indeed the disease was very grave. Now most experts consider the mortality rate from MG to be less than 5%. We now expect most patients to improve and some to go into remission.

While the costs for older drugs such as prednisone are still relatively modest, as newer immunosuppressive drugs have become available, the cost has risen. This is particularly true for IVIg. PE is another expensive procedure that involves sophisticated equipment and properly trained health care personnel to perform the procedure.

Data from Previously Completed Clinical Trials

While most published studies had been noncontrolled, nonrandomized, and nonblinded, in the past decade there have been a number of controlled, randomized trials. The positive therapeutic trials were for CSA, pulse methylprednisolone, AZA, IVIg, and tacrolimus, and recently trial of prednisone in ocular MG was conducted.⁹⁸ Negative trials were reported with MM and probably the recently completed methotrexate study. The Food and Drug Administration (FDA) recently approved eculizumab based on the recent trials.⁹⁹

Corticosteroids (CS)

While no controlled trial has demonstrated effectiveness of CS, there is general agreement that these are effective in MG and are first-line therapy in MG since the 1970s. In addition, Benatar et al showed that prednisone is effective in ocular MG in a randomized, controlled trial. Prednisone continues to be used as an immunosuppressive agent of first choice in MG. Although CS can potentially suppress the immune system in a variety of ways, including inhibition NF-κB, reduction of cytokines, and immune suppression, an exact explanation for the beneficial response in MG is unknown. CS can be used in a wide range of regimens and routes of administration. In the high-dose approach, administer prednisone 1 to 1.5 mg/kg per day (60–100 mg) for 2 to 4 weeks followed by an abrupt or tapered conversion to an every-other-day (qod) schedule.¹⁰⁰ Another approach is the low/slow method of Seybold and Drachman.¹⁰¹ It starts with 10 mg/day, and the prednisone dose is increased by 10 mg every 5–7 days; then the patient is switched to a qod schedule. A third and in-between approach is the one from the MM trial protocol of using prednisone 20 mg daily. A daily CS schedule is necessary in well-controlled hypertensive patients, in diabetic patients, and in those with MG crises or marked exacerbation. MG improvement is delayed by 2 to 4 weeks after the onset of treatment and in some is delayed by 2 to 3 months, with maximum benefit at approximately 6 months.¹⁰² Patients typically stay on prednisone 60 to 100 mg qod or its equivalent for 2 months. For good responders, it is best to taper slowly by reducing the dose no faster than 5 mg every 2 weeks. When the patient’s dose has been reduced to 20 mg qod, tapering at an even slower rate is advisable. However, it is not uncommon for patients to require low-dose therapy (5–10 mg qod) for many years or indefinitely. Prior thymectomy does not appear to influence the outcome of steroid withdrawal¹⁰³ but facilitates prednisone dose reduction.¹⁰⁴

In severe fulminant MG cases, it is better to start CS on an inpatient basis in the form of IV methylprednisolone, along with PE, for 3 days. In high-risk patients such as those with uncontrolled hypertension, diabetes, osteoporosis, and obesity and in those with baseline severe weakness, additional immunosuppressants are started if the patients have an incomplete response to CS or relapse while on CS. Initiate one of the immunosuppressive drugs such as AZA, MM, CSA, or MTX simultaneously with CS therapy. The main concern when initiating prednisone therapy at high doses is the transient worsening that occurs in one-third to one-half of patients, particularly those with severe MG.^102,105 The mechanism may involve a direct effect of CS to impair NMJ function.¹⁰⁶ In one study, 8.6% of patients who experienced transient worsening required intubation.¹⁰²

Before CS initiation, a tuberculin skin test or a QuantiFERON-TB Gold test should be performed, and a baseline dual-energy x-ray absorptiometry (DEXA) scan and ophthalmologic examination can be done shortly thereafter. Oral calcium 500 to 600 mg two to three times daily with vitamin D 400 IU daily is given to reduce the risk of pathologic fractures. Other side effects that can occur include personality changes and psychiatric side effects. Despite its many potential side effects, prednisone is considered by many to be the most effective oral immunosuppressive agent for the treatment of MG.^107,108

Cyclosporine (CSA)

CSA is a calcineurin inhibitor that reduces interleukin 2 levels and interferon γ and is a good second-line option in MG. CSA inhibits helper cytotoxic T-lymphocytes and allows the expression of suppressor T-lymphocytes. Probably the first placebo-controlled, randomized trials that showed a beneficial effect in MG patients using immunosuppressive therapy were the two studies published by Tindall et al in 1987 and 1993.^87,88 The 1987 study showed that CSA was an efficacious immunosuppressive treatment in patients who have not been on prednisone or other immunosuppressive drugs.⁸⁷ The 1993 study showed that CSA was effective in steroid-dependent MG patients.⁸⁸ The primary endpoint of both studies was a modified version of the quantitative MG (QMG) score. In the 1993 study, Tindall et al were able to show that CSA patients had a mean improved QMG score of 3.5 points, whereas the placebo-treated patients had a mean decreased QMG of 0. In addition, CSA lowered anti-AChR antibody levels, and CS doses could be reduced after CSA was initiated.

Doses for CSA range between 3 and 6 mg/kg per day. The drug is usually given in two divided doses rather than as a single dose to reduce potential nephrotoxicity. The onset of clinical benefit with CSA is 1 to 2 months (faster than AZA but slower than prednisone), with peak effect at 3 to 6 months, after which the patient is tapered to the lowest effective dose by 0.5 to 1 mg/kg every 2 to 3 months down to 2 to 3 mg/kg or even a 100-mg daily dose. Dose adjustments should not exceed more than 1.0 mg/kg per day every 4 weeks. Side effects include hypertension, nephrotoxicity, neoplasia, hirsutism, tremor, gum hyperplasia, paresthesias, headaches, and hepatotoxicity. Close monitoring of blood levels and serum creatine are necessary while on this medication. Certain drugs such as aminoglycoside, vancomycin, Bactrim, amphotericin B, ketoconazole, H2 blockers, tacrolimus, and nonsteroidal anti-inflammatory drugs (NSAIDs) potentiate the nephrotoxicity of CSA, while many others interfere with the blood levels of CSA.

Azathioprine (AZA)

AZA is an antimetabolite purine analog that blocks DNA/RNA synthesis and inhibits T-lymphocyte proliferation. It is a second-line steroid-sparing immunosuppressant in patients who relapse on prednisone or who have been on high-dose prednisone for lengthy periods. AZA is also used at times as a first-line immunosuppressive agent in retrospective case series instead of prednisone.^109–112 Retrospective studies of AZA therapy have demonstrated that 70% to 90% of MG patients improve.^109–113 But there is concern in the only one randomized, controlled trial showed delayed onset of response by about 12 to 18 months.⁸⁶ They authors were able to show that at 18 months, the AZA-treated patients were able to be tapered to a much lower dose of prednisone than the placebo-treated patients. Secondary outcome measures showed that patients receiving AZA had fewer relapses, longer remissions, and fewer side effects with less weight gain.

AZA is administered in divided doses of 2 to 3 mg/kg per day. We start with 50 mg/day for a week before gradually increasing the dose as tolerated to the target dose of 100 to 250 mg. Prior to AZA initiation, it is suggested to test for thiopurine methyltransferase activity because its deficiency predicts an increased risk of leukopenia. AZA is contraindicated in homozygous patients, whereas lower doses may be carefully tried in heterozygous patients.¹¹⁴ A meta-analysis of 54 observational studies and one randomized, controlled trial did not demonstrate sufficient evidence to address the effectiveness of thiopurine methyltransferase activity pretesting.¹¹⁵ In clinical practice, we monitor patients’ blood cell counts weekly at the initiation of AZA and then monthly.

A flulike reversible acute hypersensitivity reaction affects 12% of users in the first 2 weeks of therapy. It is associated with a rash, elevation in liver enzymes, and pancreatitis. Some may tolerate a rechallenge after recovery. Delayed adverse events include myelosuppression, hepatotoxicity, and susceptibility to infection, malignancy, teratogenicity, rash, alopecia, fever, and arthralgia. Complete blood count (CBC) and liver enzyme determinations are done every week for 4 weeks, then monthly for 6 months, and then every 3 months as long as the patient remains on stable the AZA dose. When liver enzymes are markedly elevated (above two times the normal limit), AZA should be stopped for several months until enzymes normalize before the patient may be rechallenged, at times successfully. The dose is adjusted to treatment response and to maintain the white blood cell count (WBC) above 3,500/mm³ and the absolute lymphocyte count below 1,000/μL. Patients taking allopurinol, an inhibitor of the main detoxification pathway, require AZA dose reductions to 25% to 33% of the preceding. Angiotensin-converting enzyme (ACE) inhibitors must be avoided due to the serious risk of severe leukopenia.

Mycophenolate Mofetil (MM)

MM is a well-tolerated immunosuppressive, which blocks inosine monophosphate dehydrogenase, resulting in selective inhibition of B- and T-lymphocyte proliferation by blocking purine synthesis. The most common adult dosing regimen is 1 g by mouth (PO) twice daily (doses of up to 3 g/day have been used). Uncontrolled retrospective observational studies have demonstrated functional improvement in 60% to 70% of patients receiving it.^89,90 Subsequently, Meriggioli et al performed a small randomized and blinded, placebo-controlled study on 14 patients for 5 months on MM versus placebo.⁹² All were also on immunosuppressive therapy. The authors found that patients treated with MM had a QMG score improvement (decrease) of 2.5 points, whereas in the placebo group the QMG score only changed by −0.24. A “trend” toward improvement in the patients receiving MM was suggested, however, the results were not statistically significant. These anecdotal and small placebo-controlled trial experiences with MM have led to two large Phase 3 pivotal studies of this compound in MG.^116,117 Both trials failed to meet the primary endpoint.^118,119 Therefore, based on this information, we consider MM to be a third-line agent in MG. This is controversial, however, and other experts still believe that MM is very effective based on class IV evidence.55,120 The main side effects of MM are diarrhea, vomiting, increased risk of infection, and rarely, leukopenia. CBCs are checked weekly for the first month and then less frequently. Long-term safety for MM is still in question, but there have been reports of primary CNS lymphoma with use of MM in MG.¹²¹

Tacrolimus

Tacrolimus, has been shown to be effective in MG in a randomized, controlled trial,¹²² in addition to other retrospective series. In a trial of 212 patients(of which roughly half were on prednisone or CSA dependent), tacrolimus was given at two divided dosages of 0.1 mg/kg per day, later adjusted for plasma drug concentrations between 7 and 8 mg/mL.¹²³ The mean followup time was nearly 50 months (range 12–79 months); With the addition of tacrolimusprednisone was withdrawn in 95% of patients QMG scores fell significantly, and muscle strength improvement was found. More than 85% of patients achieved complete, stable remission or pharmacologic remission at the end of follow-up, with another 5% reaching minimal-manifestation status. Tacrolimus was well tolerated overall, (hypertension in 1.9%, nephrotoxicity in 2.9%, neurotoxicity such as tremor or paresthesia in 5.9%, and diabetes in 1.4%).

Methotrexate (MTX)

MTX, an antifolate that inhibits lymphocyte proliferation, is a potentially effective third-line steroid-sparing immunosuppressant. There are a few uncontrolled small case series of the use of MTX in MG that led to larger studies.^124,125 One small single-blinded study compared the effect of MTX (17.5 mg weekly) to that of AZA (2.5 mg/kg daily) on the average daily prednisone requirement over a month in 24 patients with generalized MG indicated that MTX has similar efficacy and tolerability to AZA with a cost advantage for MTX.¹²⁶ Another randomized placebo controlled trial of MTX in which the primary endpoint was prednisone dose after 1 year failed to demonstrate.61,62 Statistical significance. However, 7 of 25 patients on placebo dropped out of the study, whereas only 1 of 25 patients on MTX stopped study participation prematurely, and the 3 patients who dropped out due to worsening MG symptoms were in the placebo group. Therefore, there may still be a role for MTX in MG. MTX is usually started at 7.5 mg/week and in 2 weeks increased to 15 mg/week as a single dose or in two divided doses if the patient has gastrointestinal (GI) symptoms. The dose can be increased to 20 mg/week if needed. Co-administer daily folic acid 0.8 to 1 mg/day orally to prevent stomatitis. Besides stomatitis, potential adverse events include alopecia, pneumonitis, teratogenicity, induction of malignancy, susceptibility to infections, and renal insufficiency. For bone marrow suppression and liver toxicity, monitor CBC, differential count, and liver function tests every week in the first 4 weeks, then monthly for 6 months, and then every 3 months thereafter while on a stable dose. Due to the risk of MTX-induced pneumonitis, we avoid the use of MTX in patients with known interstitial lung disease.