Article Text

Download PDFPDF

When the patient fails to respond to treatment: myasthenia gravis
  1. D Hilton-Jones
  1. Clinical Director, Myasthenia Centre, West Wing, John Radcliffe Hospital, Oxford, OX3 9DU, UK
    ; david.hilton-jones{at}


    Myasthenia gravis is one of the most satisfying neurological disorders to treat. There are few other conditions in which therapeutic intervention can take a patient from being bed-bound and ventilated to normality. Most patients present with less severe symptoms, but even mild extraocular muscle weakness can be profoundly disabling. The standard therapeutic approach is successful for most patients, which can make the non-specialist neurologist somewhat blasé about its management. However, panic can set in when the standard approach fails. Failure is often the result of incorrect diagnosis, or inappropriate use of first-line treatments. This article outlines the main reasons for failure and gives advice on alternative therapeutic strategies.

    Statistics from

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


    The management of autoimmune acquired myasthenia gravis has been reviewed in this journal recently1 and algorithms summarising the treatment of ocular and generalised myasthenia gravis are reproduced in figures 1 and 2. Most patients respond extremely well to such an approach. But what to do when this approach fails, or appears to fail?

    Figure 1 Management of ocular myasthenia gravis. (Reproduced from Pract Neurol 2005;5:18–27.)
    Figure 2 Management of generalised myasthenia gravis. AChR, anti-acetylcholine receptor antibody. (Reproduced from Pract Neurol 2005;5:18–27.)

    The principle causes of “failure” are listed in the box. The first part of this article will address the approach to such failures, while the second part will discuss what to do when the standard approach to treatment really has failed.

    Causes of failure, or apparent failure, of myasthenia gravis to respond to conventional treatment

    • Inappropriate expectations

    • Wrong diagnosis

    • Misinterpretation of persisting symptomatology on treatment

    • Inappropriate use of pyridostigmine

    • Incorrect use of prednisolone

    • Inappropriate dose and expectations of azathioprine (or other immunosuppressants)

    • Compliance


    Despite the occasional Lazarus-like response to anticholinesterase drugs, it must be remembered that it typically takes months for myasthenia to be brought under adequate control with immunosuppression (or thymectomy in those who respond), and the clinician may have unreasonable expectations, which are passed on to the patient. Also, there may be a disparity between the patient’s and the clinician’s view of success—even if there is a dramatic improvement in the range of eye movements or of limb strength, persisting diplopia and fatigability may still leave the patient functionally very disabled. I typically advise patients at the outset of treatment that they “are starting a marathon not a sprint”. Prednisolone works faster than the commonly used second-line immunosuppressant drugs, and although most patients will show a good response within a couple of months of attaining the maximum dose, in some the response is delayed for as much as 9–12 months. Failure to appreciate this may lead to inappropriate changes in medication, as discussed below. Of note is that myasthenia associated with muscle specific kinase (MuSK) antibodies may be more resistant to treatment than that associated with acetylcholine receptor (AChR) antibodies.


    This merits a separate paper, but the most common errors are:

    • Misinterpretation of neurophysiology. Attaching undue significance to a decremental response of less than 10%, or failing to appreciate that jitter is also seen in denervating disorders and in mitochondrial myopathies.

    • Congenital myasthenic syndromes. Despite the term congenital, these can present in adolescence and adult life (for example, rapsyn, slow channel syndrome, and Dok-7 syndromes) and are frequently initially misdiagnosed (following neurophysiology and negative antibody tests) as seronegative myasthenia. They of course do not respond to immunosuppression but some respond to pyridostigmine (but not slow channel syndrome and Dok-7 syndromes).

    • Lambert-Eaton myasthenic syndrome. Another cause of misdiagnosis, particularly if the neurophysiologist is inexperienced and fails to look for post-tetanic potentiation.

    • Laboratory error. Although rare, this can happen, and I have seen a number of patients diagnosed incorrectly on the basis of a positive AChR antibody assay. These patients typically have a “low-positive” titre, but another error can be sample confusion and a true high-titre on the sample, but it is the wrong sample.

    • False-positive Tensilon (edrophonium) test. This is possibly the commonest error seen in a tertiary referral centre. A diagnosis based on Tensilon testing alone, without antibody or neurophysiological support, should always be viewed with considerable caution. It is most commonly misinterpreted in those with fatigue when objective assessment is difficult. Its use should be restricted largely to patients with external ophthalmoplegia, when a subjective response can readily be assessed, and performed by somebody with experience of the test. A true objective response may be seen in conditions other than myasthenia and indeed anticholinesterase drugs have been used in a range of neurological disorders for symptom relief.

    • Mimics of ocular myasthenia. Mitochondrial chronic progressive external ophthalmoplegia (CPEO) can easily be mistaken for myasthenia (and the Tensilon test may be “positive”). The severity of limitation of eye movements is often gross in CPEO, whereas diplopia tends to be infrequent. In oculopharyngeal muscular dystrophy the ptosis, which is always more marked than limitation of eye movements, can be asymmetric as is typical of myasthenia. Thyroid ophthalmopathy must always be considered as a cause of diplopia. When the signs are purely ocular, and the antibody and neurophysiological studies negative, the patient needs brain MRI to exclude a structural lesion (fig 3).

    • Mimics of generalised myasthenia gravis. There are many, including most neuromuscular disorders. The patient’s complaint of “fatigue” may be misinterpreted in the absence of demonstrable fatigue. Patients with chronic fatigue syndrome often seek referral to a myasthenia clinic, having read about the condition on the web—a non-specific response to Tensilon may cause confusion.

    Figure 3 MR scans showing periaqueductal lesion (presumed glioma) causing obstructive hydrocephalus in a patient who had been treated for one year, without success, for “seronegative” ocular myasthenia (unilateral ptosis and diploplia). The diploplia resolved following shunting but the ptosis, which showed mild fatigability, was unchanged. (Courtesy of Dr P Anslow; reproduced from Pract Neurol 2002;2:173–7.)


    Non-specific complaints of “fatigue”, tiredness, lethargy and sleepiness may all be considered by the patient to be residual symptoms of their myasthenia, and misinterpreted by the clinician who then gives inappropriate advice with respect to treatment. High doses of prednisolone may cause feelings of fatigue, and muscle discomfort, which of course will respond to dose reduction not increase. Sustained high-dose steroid therapy may cause steroid myopathy with the weakness being interpreted as recurrence/exacerbation of the myasthenia leading to an inappropriate increase in dose. In my experience even sustained high doses of prednisolone on alternate days rarely cause steroid myopathy, whereas it appears more common with daily regimes. When tapering prednisolone and pyridostigmine it is very common, particularly just after each dose reduction, for the patient to complain of fatigue due to the loss of the non-specific stimulatory effect of these drugs—patients frequently interpret this as recurrence of their disease and unless suitably counselled may inappropriately increase the dose, or be reluctant to make further reductions in dose.


    The maximum daily dose should not exceed about 360 mg. Higher doses are very unlikely to produce useful additional benefit but will lead to more adverse effects, including the possibility of “cholinergic crisis”, which can be interpreted wrongly as worsening of the myasthenia with disastrous consequences. Once immunosuppression has brought the myasthenia into remission there is no need, or indeed logic, for continuing anticholinesterases, but frequently they are continued inappropriately. Adverse effects, such as muscle twitching and cramps, may be wrongly interpreted as indicating continuing activity of the myasthenia.


    The main causes of treatment failure are either not giving a large enough dose of prednisolone at the outset, or reducing the dose too quickly. The target dose for either ocular (0.75 mg/kg body weight on alternate days, maximum 60 mg) or generalised (1.5 mg/kg body weight on alternate days, maximum 100 mg) myasthenia gravis is readily calculated and should be adhered to:

    • It is better to get the disease under control as quickly as possible, and then find the minimum maintenance dose, than to try to get away with lower doses that may fail to achieve remission.

    • In general, the dose should not be reduced from the target dose until the patient has entered remission, and as noted above that can take many months. Premature dose reduction will inevitably lead to poorer disease control.

    • Inappropriate continuation of a high dose, once remission has been achieved, carries the risk of inducing steroid myopathy as well as causing other adverse effects which may be misinterpreted as ongoing myasthenia disease activity (see above).


    Azathioprine is my preferred immunosuppressant/steroid-sparing drug because of its proven benefit in a randomised placebo-controlled trial and its safety record, including use in pregnancy. Its use is somewhat complicated by individual polymorphisms that affect the rate of metabolism. Some centres make use of the TPMT-assay to determine optimum dose but arguably more important is to monitor closely haematological and biochemical parameters. The target dose is based on 2.5 mg/kg bodyweight/day but some patients require up to 3 mg/kg bodyweight/day. Macrocytosis and lymphopenia are normal responses to the drug, and indeed if they have not developed then a further dose increase is appropriate if the myasthenia is not adequately controlled; not infrequently I have seen the dose being reduced inappropriately because these are perceived to be undesirable adverse effects. Azathioprine is very slow to act and should be considered to have little or no effect for at least 12 months, with maximum benefit within 24 months. A fair trial should therefore last at least 18 months before considering it ineffective and trying another drug. All other immunosuppressants almost certainly work more quickly although there are no trial data. Even so, a minimum trial should probably be at least six months.


    It is not surprising that this can be an issue given the undoubted adverse effects of steroids and their associated bad publicity. Indeed, virtually all patients will develop some adverse effects, and cosmetic ones may be a particular problem in the young female population. Absence of adverse effects with failure of therapy strongly suggests non-compliance.



    Azathioprine cannot be judged to have failed until the patient has received it for at least 18 months at the optimum dose (see above). But what next if azathioprine fails, or the patient is intolerant of it? There are essentially no trial data to help with this decision but much anecdotal experience, which suggests that all similar drugs work for some people. Although guidelines are available,2 the choice is often based on personal experience and concerns about specific adverse effects. For each, the minimum duration of therapy before declaring lack of efficacy should be at least six months. The following are discussed in alphabetical order:

    • Ciclosporin. The typical dose is 2.5 mg/kg bodyweight/day, at which dose serum level estimations are not required (a retrospective review gave support for its use, but there have been no formal trials3). However, even at this relatively low dose renal impairment and hypertension can be problematic. Hirsutes is an unacceptable adverse effect for some, and tremor fairly common. Doses of up to 7 mg/kg bodyweight/day can be used, with serum level monitoring. There remain concerns about the risk of developing cancer (especially skin cancers, as with other immunosuppressants) and facilitating cancer progression.

    • Cyclophosphamide. Pulsed intravenous therapy (500 mg/m2 body surface area monthly) has been used in patients resistant to other forms of immunosuppression.4 High-dose cyclophosphamide (50 mg/kg bodyweight daily for four days intravenously) has been used to “reboot” the immune system by inducing partial or complete immunoablation5 but this must still be regarded as experimental.

    • Methotrexate. Cannot be used by either males or females planning reproduction, but in others I think it is an underused treatment. There are no useful published trial data. An advantage is that general practitioners are experienced in its prescription and monitoring. It is usually given at a dose of 5–15 mg once weekly.

    • Monoclonal antibodies. In vogue and expensive! The anti-TNFα drugs infliximab and etanercept have been reported in very small studies to have an effect in myasthenia gravis, although in at least one case exacerbation of symptoms was noted.6, 7 The anti-CD20 lymphocyte antibody rituximab looks rather more promising and has been reported to be of benefit in MuSK myasthenia, which is often more treatment resistant than anti-AChR antibody myasthenia.8 Our own limited experience to date has been encouraging.

    • Mycophenolate mofetil. Has also been very much in vogue in the last few years. As with other immunosuppressants, there are many anecdotal reports of its (sometimes dramatic) efficacy. A recent double-blind study (as yet published only in abstract) failed to show benefit over placebo but the trial design has been criticised. Compared with azathioprine it probably has fewer hepatic and haematological adverse effects. The dosage regime is simple (1 g twice daily) and apart from those patients who develop gastrointestinal upset it is well tolerated. It merits further study and I, as have other specialists in the area, have had some success (as well as failures) with it.

    • Tacrolimus. Yet another vogue drug with vocal anecdotal support but inadequate formal trial data.9 It is from the same class of immunosuppressants as ciclosporin.

    Intravenous immunoglobulin and plasma exchange

    Neither of these, except in very rare circumstances, should be considered as long-term therapies because of cost and safety issues, as well as their inconvenience. Neither will give benefit for more than two months. However, they may have a useful supportive role. At the outset of management, a good clinical response to either offers evidence that one is dealing with an antibody-mediated disease, if that was ever in doubt, and that subsequent conventional immunosuppression is likely to be effective, even if it takes time to become so. In other words, it may encourage the patient and physician to persevere. While waiting to gain the benefits of immunosuppression, intravenous immunoglobulin (IVIg) or plasma exchange may aid in symptom management. Trials suggest that they are generally equally efficacious, and which is used will depend on local resources, but some patients may respond better to one than the other (with plasma exchange tending to be more effective).


    With time, and the development of effective immunosuppressant drug regimes, the role of thymectomy has become less certain—hence the current international thymectomy trial. If a patient with generalised myasthenia has declined early thymectomy (fig 2), then the question of surgery may be revisited if other treatments appear to be failing. On current evidence, even in the face of therapeutic failure, I would not consider thymectomy in patients who are not AChR-positive or who are over the age of 45 years (except that older AChR-positive patients might be considered for the international thymectomy trial; = 1), or who have restricted ocular myasthenia.

    Ocular myasthenia

    A few patients are relatively resistant and may need to follow the regime advised for generalised myasthenia. Even so, it is not that uncommon for patients to be left with residual diplopia or ptosis despite optimal treatment. The possibility of surgical intervention should be considered (fig 1).


    The “standard” approaches to the management of ocular myasthenia gravis (pyridostigmine and prednisolone) and generalised myasthenia gravis (pyridostigmine, prednisolone, azathioprine, thymectomy) are successful in most patients. Failure can often be attributed to wrong diagnosis or not following treatment to the letter, but a few patients prove to be treatment resistant. The Editor commissioned this paper to answer the question of what to do with such patients. He is renowned for wanting both an evidence-based approach, and a didactic statement of what the “expert” actually does in practice—in this case the two are incompatible! If the standard approach, rigorously applied, has failed then in practice I would use an alternative immunosuppressant drug, supplemented in the short term with IVIg or plasma exchange if symptomatic relief is required. But which immunosuppressant? The choice is generally dictated by the patient following extensive discussion of the (known) benefits and hazards of each. Specific additional factors may tilt the balance one way or another:

    • In a patient with hypertension or renal disease I would avoid ciclosporin—mycophenolate mofetil may suit the younger patient, or methotrexate the patient who does not want children.

    • In Britain, azathioprine is considered safe in pregnancy and ciclosporin probably safe (see British National Formulary). Other immunosuppressant drugs are, according to their manufacturers, contraindicated, but that is generally because of lack of data and fear of litigation rather than their being evidence of problems. Inevitably some women, dependent on immunosuppression, will become pregnant on these other drugs and some evidence will slowly emerge.

    • In the older patient I have a liking for methotrexate. I find it effective, generally well-tolerated, and easy to use in liaison with the patient’s general practitioner.

    If two immunosuppressant drugs (from azathioprine, ciclosporin, mycophenolate mofetil and methotrexate) have failed, because of intolerance or lack of efficacy, then on current evidence and experience I would be inclined towards tacrolimus or rituximab, rather than cyclophosphamide.

    As important as the precise therapy is the experience of the clinician. In general, patients with “difficult” myasthenia should ideally be managed in specialist centres. This will also increase the likelihood of there being formal trials of the various therapies now available.


    Other content recommended for you