Does This Patient Have Myasthenia Gravis?

A 45-year-old man has a 2-month history of fluctuating double vision, a “droopy” right eye that improves with rest, and a complaint that “food gets stuck halfway down.” Your examination confirms severe right eyelid ptosis that dramatically improves with rest. His right eye adduction and upgaze are markedly impaired. The left eye demonstrates complete horizontal ophthalmoplegia. The limb muscle strength and reflexes are normal. You wonder whether there is an accurate and clinically useful bedside test to help confirm the diagnosis of myasthenia gravis.

A 69-year-old man has a 2-month history of intermittent spells of double vision, generalized weakness that worsens toward the evening, and unspecified dizziness. Although he has normal strength and reflexes and no ophthalmoplegia, he does report fluctuating diplopia during the examination. As in Case 1, you must decide whether to perform additional bedside tests, obtain electrodiagnostic or acetylcholine antibody testing, or pursue a broader diagnostic evaluation of the various causes of dizzy spells and fatigue.

Myasthenia gravis is an autoimmune disease associated with circulating acetylcholine receptor antibodies, modification of the synaptic cleft, and destruction of the postsynaptic neuromuscular membrane. The clinical hallmark of the disease is fatigable weakness. The clinical severity ranges from mild, purely ocular, forms to severe generalized weakness and respiratory failure. Myasthenia gravis is a rare disease; its prevalence in the United States is reported at 14.2/100 000. Prevalence rates have been increasing steadily over the past decades, likely due to decreased mortality, longer survival, and higher rates of diagnosis.^{1 - 3} Men older than 50 years have the highest incidence in the population, with the peak around age 70 years. Women have 2 incidence peaks: one around ages 20 to 40 years and one around age 70 years.^{4 - 5}

Clinicians must be alert to the symptoms and signs of myasthenia gravis, as it is an eminently treatable disease and the earlier treatment is started, the better the clinical response.^{6 - 8} Only 54% to 69% of patients with myasthenia gravis are diagnosed within 1 year of onset, and the mean time to diagnosis is more than 1 year.^{3 ,9 - 12} Untreated patients are at risk for deterioration and “crisis,” which occurs when weakness becomes severe enough to require mechanical ventilation.^{13 - 14} Left untreated, reversible and fatigable weakness may become fixed. An erroneous diagnosis of myasthenia gravis may expose patients to unnecessary diagnostic procedures and treatments.

The acetylcholine receptor antibody test is the most specific diagnostic test for myasthenia gravis. This test has reasonable sensitivity in generalized myasthenia gravis (80%-96%), but up to 50% of patients with purely ocular myasthenia are seronegative.^{15 - 19} Single-fiber electromyography, performed by highly trained experts at specialized centers, is highly sensitive for disorders of the neuromuscular junction but is not specific for myasthenia gravis.

The purpose of this review was to determine the value of clinical symptoms and signs, as well as the results of simple provocative clinical tests, in deciding whether myasthenia gravis should be considered as a diagnosis and in enabling the physician to determine if further confirmatory testing (including the highly specific and sensitive antibody test) is warranted.

In the normal neuromuscular junction acetylcholine is released into the synaptic cleft, diffuses to the postsynaptic membrane, binds to ligand-sensitive ion channels (nicotinic acetylcholine receptors), and causes an excitatory postsynaptic end-plate potential. If the threshold depolarization is achieved, an action potential will spread along the muscle fiber membrane, causing muscle contraction. Acetylcholine is cleared from the synaptic cleft by presynaptic reuptake and by the metabolic action of acetylcholinesterase (Figure 1).

The failure of transmission at many neuromuscular junctions in myasthenia results in diminished end-plate potentials that are insufficient to generate action potentials in a number of muscle fibers.²⁰ This results in fatigable weakness of striated muscles, which is the basis for the clinical diagnosis. Sustained or repetitive muscle contraction causes fatigue and weakness of myasthenic muscles. Cooling a weak muscle improves neuromuscular transmission.²¹ Rest and acetylcholinesterase inhibitors transiently increase acetylcholine levels in the synaptic cleft. The change in strength following these manipulations can be assessed during the clinical examination.

Patients with myasthenia gravis complain of weakness in specific muscles. Up to 65% of patients initially have ocular symptoms of double vision and drooping of the eyelids. Less than one fourth of patients present with bulbar weakness (ie, in lower cranial nerve–innervated oropharyngeal muscles) and report slurred or nasal speech, alterations of the voice (eg, softness, breathiness, hoarseness), and difficulty chewing or swallowing. Limb weakness is a rare initial complaint (14%-27%) and should be differentiated from nonspecific generalized fatigue. Patients may report shortness of breath. The symptoms of myasthenia are typically better upon awakening or after rest and become progressively worse with prolonged use of the affected muscles or later in the day.^{3 ,22 - 24}

Reduced muscle power by manual testing in specific muscles that worsens with repetition and improves with rest is the characteristic examination finding in myasthenia. Most muscles with voluntary activation have a large variability of strength even under normal conditions due to effort. Evaluating extremity strength greatly depends on the experience of the examiner. Ptosis and extraocular muscle deficits are relatively free of a voluntary component and provide a more objective measure.

Fatigable and rapidly fluctuating asymmetric ptosis is a hallmark of myasthenia gravis. The rapid fluctuation results from improvement during even very short periods of rest, such as blinking. Besides fast variability in the degree of ptosis, it may altogether shift quickly from one eye to the other, known as “shifting ptosis.”²² Ptosis should be evaluated with the patient sitting comfortably, the head held in primary position without tilting. The patient fixates on a distant object (eg, a spot on the wall) and is asked to refrain from blinking and to relax the forehead muscles. Frontalis contraction, a mostly involuntary compensatory mechanism, is a common and characteristic sign in myasthenic patients with ptosis. Relaxing the forehead muscles may be difficult for some patients. The examiner measures palpebral fissure width at eye level during forward gaze and again during prolonged upward or lateral gaze for 30 seconds.^{22 ,25} The more ptotic eyelid should be used for additional provocative tests, such as the ice pack, rest, and sleep tests.

The ice pack test is performed by placing a latex glove finger filled with crushed ice over the more ptotic eyelid for 2 minutes. During the rest test the patient places a glove filled with cotton (a placebo) over the more ptotic eyelid while holding the eyes closed for 2 minutes. During the sleep test the patient is left in a quiet dark room with the eyes closed for 30 minutes. Complete or almost complete resolution of ptosis or at least a 2-mm increase in palpebral fissure width constitutes a positive response to these maneuvers. It is very important to evaluate the improvement immediately following the test, as the lids may quickly begin to droop again.

The curtain sign (also known as “enhanced ptosis” or “paradoxical ptosis”) is usually observed in patients with some initial ptosis. The patient looks straight ahead and refrains from blinking. The examiner holds one eye open, which results in the other lid starting to droop more (like a curtain falling). The lid twitch sign occurs when the patient opens the eyes after gentle closure or follows the examiner’s finger down and then back up to eye level. The lids overshoot or “twitch” for a fraction of a second before settling into position and starting to droop.²⁶

Asymmetrical weakness of extraocular muscles is commonly observed in myasthenia when sustained lateral gaze or upgaze worsens or induces double vision. The cover-uncover test may be performed to bring out subtle extraocular weakness. As the patient fixates on an object in the distance, the examiner covers one eye while observing for deviation of the uncovered eye during lateral and then upward gazing. With extraocular weakness, the uncovered eye will drift. The examination is completed by repeating the procedure for the opposite eye. Quiver eye movements are very fast, small-twitch, “lightning-” or “jerk”-like movements of the eyes upon changing direction of gaze. They are said to occur even in the setting of profound ophthalmoplegia.²⁷

Although patients rarely complain of facial weakness, it is often found on examination. Severe facial weakness results in a characteristic transverse smile. Orbicularis oculi weakness is demonstrated as the examiner tries to separate the eyelids against forced eye closure. Orbicularis oculi fatigue may be observed on gentle eye closure. After complete initial apposition of the lid margins, they separate within seconds and the white of the sclera starts to show (positive “peek sign”) (Figure 2).²⁸ The iris should not be visible due to the eyeballs being rolled up (Bell phenomenon). The iris may be visible if the patient is not trying to close the eyes voluntarily (in the case of a conversion reaction and functional weakness) or in case of severe ophthalmoplegia.

Tongue and pharyngeal weakness will result in the patient’s speech becoming slurred or nasal, especially with prolonged speaking. Other commonly weak muscles include neck flexors, deltoids, hip flexors, finger/wrist extensors, and foot dorsiflexors. The muscles should be repeatedly tested against manual resistance, with a brief rest between repetitions. Having the patient hold the head above the pillow in the supine position or hold the arms outstretched in abduction at the shoulder for 1 minute are ways to test for fatigability of neck flexors and deltoids, respectively. Involvement is often asymmetrical. The remainder of the neurologic examination results, including those for deep tendon reflexes and sensory examination, must be normal.

Edrophonium chloride is a fast- and short-acting acetylcholinesterase inhibitor that may be administered in the office setting to diagnose myasthenia gravis ( Article ). Its effect usually occurs within 30 seconds and lasts less than 5 minutes. Most myasthenic muscles respond to the test dose of 2 mg, but many will require more. Adverse effects are rare and usually mild (excess salivation, sweating, abdominal cramps, or fecal incontinence). Serious adverse effects, such as bradycardia, asystole, and bronchoconstriction, occur infrequently (<0.2%) but warrant that the patient be receiving cardiac monitoring during the test and that a bag-mask be available should the patient need ventilatory assistance.^{30 - 31} Reactive airway disease or cardiac bradyarrhythmias are relative contraindications. Using a 3-way stopcock setup may be feasible in a patient already equipped with a peripheral intravenous line (in an intensive care unit, for example). One concern with such a setup is the possibility of an accidental mix-up of the syringes, with resultant injection errors—the syringes should always be labeled clearly. Due to the short action of the drug, the examiner must be able to quickly assess for improvement. Evaluating extraocular muscle abnormalities or changes in manual muscle testing require skill and time; therefore, most experts recommend performing the edrophonium test only when the patient has easily observed baseline weakness in specific muscles.³² One author feels that only a clearly ptotic eyelid or visibly abnormal extraocular muscles provide “objective” end points.²⁷ Unequivocal improvement in ptosis or extraocular muscles constitutes a positive response. The administering physician (especially one with less experience) should consider blinding the edrophonium administration to avoid expectation bias.

Establish reliable peripheral intravenous access

Prepare a syringe with 2 mg of atropine (available in 0.4-mg/mL or 1-mg/mL ampules) to have handy

Prepare 1 mL (10 mg) of edrophonium in a tuberculin syringe (edrophonium is available in a 10-mg/mL solution in a 1-mL ampule [10 mg] or in a 10-mL vial [total of 100 mg])

Inject 2 mg (0.2 mL) slowly over 15 seconds while observing for an objective improvement in target muscles

Improvement should occur within 30 seconds and disappear in 5 minutes; if no response and no significant adverse effects, administer the remaining edrophonium (8 mg [0.8 mL]) for a total dose of 10 mg

Atropine should be injected (0.5 or 1 mg) in case of clinically significant bradycardia, respiratory distress, or syncope*

*Routine administration of atropine simultaneously with edrophonium for the purpose of diagnostic testing for myasthenia gravis is not recommended. Bartley and Bullock²⁹ recommend using a 3-way stopcock with the edrophonium-containing syringe attached to the direct port and the atropine-containing syringe attached to the side port, so that atropine may be quickly injected in case of severe adverse effects.

Neostigmine bromide is an anticholinesterase agent used to treat myasthenia gravis. Parenteral preparations are available in vials containing 0.25-, 0.5-, and 1-mg/mL doses. The recommended dose for the diagnosis of myasthenia is 0.02 mg/kg given intramuscularly. A standard dose of 1 or 1.5 mg may be used. The response should be evaluated 30 minutes after injection, at the time of peak effect. The half-life after intramuscular administration is 50 to 90 minutes. Adverse effects, precautions, and need for good intravenous access (to administer atropine in case of an adverse event) are the same as for edrophonium.

Pyridostigmine bromide is an analog of neostigmine with a slightly longer duration of action and fewer adverse effects. It is the most commonly used anticholinesterase agent for the symptomatic treatment of myasthenia gravis. It has been used for diagnosis in patients in whom edrophonium or neostigmine is relatively contraindicated though it is not generally used for diagnostic purposes.³³ It is available for injection in 2-mL vials containing 5 mg/mL. A 2-mg intramuscular or intravenous dose is equivalent to 60 mg orally. Precautions should be exercised just as with edrophonium and neostigmine.

English-language articles in the MEDLINE database from January 1966 through January 2005 were searched using the terms myasthenia gravis, diagnosis, and test. One of the authors (K.S.) identified potential articles by screening the retrieved titles and abstracts (when available) and searching through the bibliographies of the retrieved articles. Two authors (K.S., R.S.B.) independently reviewed the retrieved articles. An article was included where agreement existed that the study had met our inclusion criteria.

Eligible studies evaluated a particular symptom or sign in patients with myasthenia gravis and in controls. Studies requiring sophisticated equipment or subspecialty trained physicians (otolaryngology, ophthalmology, etc) were excluded. Studies based on small numbers of patients were not excluded, as most series are comparatively small in the literature. Out of 640 total articles, the search identified 33 potential articles. Of these, 15 met inclusion criteria and form the basis of this review.^{28 ,33 - 46} Quality of evidence in each study was classified based on a published classification scheme for levels of evidence developed for The Rational Clinical Examination series (Table 1). In addition, studies that were carried out prospectively and/or on a consecutive series of patients were ranked higher (Table 2). Only 2 studies included an independent blinded comparison of signs and symptoms to a criterion standard.^{34 ,36}

Sensitivity was defined as the proportion of patients with myasthenia gravis who had the particular symptom or sign; specificity, as the proportion of nonmyasthenic patients without the sign or symptom. The positive likelihood ratio (LR) was defined as the likelihood of a positive test result (or presence of a sign or symptom) in a myasthenic patient compared with the likelihood of a positive test result in a nonmyasthenic patient, that is, the increase in odds that the patient has myasthenia gravis when the test result is positive (or when the sign or symptom is present). Positive LR is expressed as sensitivity/(1 − specificity). The negative LR is the likelihood of a negative test result (or absence of a sign or symptom) in a myasthenic patient compared with the likelihood of a negative test result (or absence of a sign or symptom) in a nonmyasthenic patient, that is, the decrease in odds that the patient has myasthenia gravis when the test result is negative (or when the sign or symptom is absent). Negative LR is expressed as (1 − sensitivity)/specificity. Summary LRs were derived using random-effects measures that provide conservative confidence intervals (CIs) around the estimates.^{48 - 50}

Fifteen studies reported findings on patients both with and without myasthenia gravis^{28 ,33 - 46} (Table 2). Seven studies evaluated the ice test, including 3 that also evaluated the response to anticholinesterase agents and 1 that also evaluated the rest test. Four additional studies reported on the response to anticholinesterase agents and 1 additional study on the sleep test. The remaining 3 articles included 1 study reporting on 2 symptoms and 2 studies evaluating 1 sign each. The results across studies for the ice test and anticholinesterase tests were homogeneous; we report random-effects summary LRs for these signs (Table 3).

Only 1 eligible study was identified and it evaluated 2 symptoms.⁴⁶ The history was taken from patients via a questionnaire. The presence of food remaining in the mouth after swallowing increases the likelihood of myasthenia gravis, but the wide CI indicates that the finding is not reliable. Speech becoming unintelligible during prolonged speaking has an LR of 4.5 (95% CI, 1.2-17.0). Neither normal swallowing nor normal speech rules out myasthenia gravis (LR, 0.70; 95% CI, 0.58-0.84 and LR, 0.61; 95% CI, 0.46-0.80, respectively).

Two eligible studies were identified and reported on 1 sign each.^{28 ,45} The presence or absence of quiver eye movements increased the likelihood of myasthenia gravis, but the broad CIs around the LR indicate that the examiner may not rely on the finding. The presence of the peek sign might be more useful (LR, 30.0; 95% CI, 3.2-278.0) but also has broad CIs.

Seven studies investigated the ice test and all had similar findings.^{34 - 40} The overall prevalence (prior probability) of myasthenia gravis in these studies was 49% (92 of 189 patients total). All but 1 of these studies were carried out prospectively. The LR for a positive ice test result suggests that the finding is useful (summary LR, 24.0; 95% CI, 8.5-67.0). A negative ice test result lessens the likelihood of myasthenia gravis (summary LR, 0.16; 95% CI, 0.09-0.27).

Two studies evaluated the precision (ie, interobserver variation) of the ice test. Kubis et al³⁴ used the signed-rank test to evaluate interobserver variability and found no significant difference between observers (P = .79). Ertas et al³⁶ reported complete agreement among their observers. Neither of the studies evaluated the intraobserver variation.

Seven studies reported the results of anticholinesterase tests, and all had similar findings.^{33 ,35 - 36 ,40 ,42 - 44} Five of these studies evaluated the edrophonium test; 1 study included response to pyridostigmine, and another included response to neostigmine as an alternative. All but 1 of these studies were prospective and 3 were carried out on consecutive patients. One hundred fifty-six of 320 patients (49%) had myasthenia gravis. The likelihood of myasthenia gravis increases for a positive test result (summary LR, 15.0; 95% CI, 7.5-31.0), whereas the lack of improvement makes myasthenia gravis much less likely (summary LR, 0.11; 95% CI, 0.06-0.21).

Two studies evaluated the sleep or rest test on 93 patients, including 52 (56%) with myasthenia gravis.^{34 ,41} An abnormal rest test result increases the likelihood of myasthenia but the wide CI indicates uncertainty about the true significance. A positive sleep test result may be more useful (LR, 53.0; 95% CI, 3.43-832.0). Both the rest and sleep test make the probability of myasthenia unlikely when the result is normal (LR, 0.52; 95% CI, 0.29-0.95 and LR, 0.01; 95% CI, 0.00-0.16, respectively).

Fluctuating weakness (ie, reduced muscle power) that worsens with exertion and improves with rest or with application of ice or cold is never normal. It is important to differentiate fluctuating weakness from patients’ reports of “weakness,” which most often refers to fatigue or exertion. True fluctuating weakness, as demonstrated by manual muscle testing, is the cardinal feature of myasthenia gravis. Other neuromuscular conditions (including amyotrophic lateral sclerosis and periodic paralyses) may be associated with fluctuating weakness; however, the fluctuation in myasthenia is more dramatic and occurs much more rapidly. Ptosis or diplopia may be present in a number of conditions (congenital exotropia or esotropia, strabismus, congenital ptosis, cranial nerve palsies, myopathies, progressive external ophthalmoplegia, brainstem lesions, and neurodegenerative disorders such as progressive supranuclear palsy), but the constant degree of involvement and associated neurologic findings (pupillary abnormalities, nystagmus, vertigo, sensory involvement) commonly exclude myasthenia gravis as a diagnosis. One must bear in mind that even the initially fluctuating weakness of myasthenia gravis will become fixed over time if severe enough. The hypomimia (“masked facies”) of parkinsonism may be mistaken for facial weakness, but on examination no true weakness is found and associated features of parkinsonism are evident. It may be a challenge to differentiate true fatigable weakness due to myasthenia gravis from conversion reactions. In the latter conditions one may often find that various elements of the examination are inconsistent with pathophysiologic conditions. Conversion reactions commonly produce “give-away” weakness, in which an initial full resistance suddenly gives away under the hand of the examiner, as opposed to true weakness that gradually worsens or is present from the start. Ptosis produced by conversion reactions is commonly symmetrical and bilateral. Since it occurs with contraction of the orbicularis oculi, one can observe that the lower lid elevates. It may completely disappear with diverting the patient’s attention. Eye closure weakness due to poor effort results in the iris showing between the eyelids.

Fluctuating diplopia and ptosis are highly characteristic of myasthenia gravis. The presence of a positive rest test result may increase the likelihood of myasthenia. The physician must carefully question the patient regarding his complaint of “food getting stuck halfway down.” If it is food remaining in the mouth after swallowing, it may also increase the probability of myasthenia. The available evidence-based data, however, do not allow the examiner to rely on these findings to confirm the diagnosis. These positive test results should prompt the clinician to confirm the diagnosis with the acetylcholine receptor antibody test and to refer this patient to a specialist (neurologist or neuro-ophthalmologist).

The presentation of an elderly patient complaining of fluctuating double vision and weakness worsening toward the end of the day raises the possibility of myasthenia gravis. The lack of quiver eye movements, peek sign, or history of “unintelligible speech after prolonged speaking” or “food in the mouth after swallowing” does not significantly reduce the likelihood of myasthenia based on the studies we reviewed. This patient does not have any objective ptosis or visible diplopia, so provocative tests cannot be performed. A search should be undertaken for causes of nonspecific dizziness and generalized fatigue. If, however, he continues to complain of fluctuating double vision, he should be referred for specialist evaluation to rule out myasthenia despite normal physical examination findings.

The presence of certain historical features (speech becoming unintelligible after prolonged periods) or signs (peek sign) may be useful in confirming the diagnosis of myasthenia gravis, though their absence does not rule it out. The ice test, the sleep test, and response to anticholinesterase agents (especially the edrophonium test) are useful in confirming the diagnosis and reduce the likelihood when results are negative. A positive test result should prompt proceeding with acetylcholine receptor antibody testing and specialist referral for electrophysiologic tests and should help confirm the diagnosis in patients who have negative results for the acetylcholine receptor antibody panel.

This review has several limitations, and the results should be interpreted with caution. The results may not be generalizable for a number of reasons. Myasthenia gravis is a rare disorder, and the number of studies evaluating its symptoms and signs are few. The studies included in this review examined only a few symptoms and signs in a selected group of patients with a confirmed diagnosis of myasthenia gravis. Due to possible verification bias in this selected population of patients with myasthenia (in whom confirmation of the diagnosis is more likely with clear-cut cases), it is expected that in the general population these tests have a lower sensitivity but even higher specificity. Because of the uncertainty regarding sensitivity, patients with persistent symptoms but normal physical examination findings should be referred to specialists for diagnosis. The specificity and sensitivity of the described tests may also greatly depend on the skill and experience of the examiner. Future studies are needed that evaluate not only intraobserver variability but agreement between experts and nonexperts. There are other signs that may be more useful than those tested historically and that await scientific study. This review underscores the need for more studies to evaluate symptoms and signs predictive of myasthenia to improve physicians’ ability to recognize and evaluate patients at the time of presentation.