Does perioperative high-dose prednisolone have clinical benefits for generalized myasthenia gravis?

Myasthenia gravis (MG) is an autoimmune disorder that results from dysfunction of the post-synaptic acetylcholine receptor (AchR), which may be due to reduced receptor density, increased turnover, or antibodies (AchR-Ab). Immunosuppressant therapies have been shown to suppress the destructive activity of AchR-Ab at neuromuscular junctions. Anticholinesterase inhibitors are used as a first-line medical therapy, and corticosteroids are used for patients with moderate to severe symptoms. However, anticholinesterase medications, such as pyridostigmine, do not prevent the AchR-Ab-mediated destruction of neuromuscular junctions. Furthermore, long-term use of anticholinesterases can reduce AchR responses due to the destruction of AchR, leading to increased anticholinesterase doses [1] that can suppress cardiac function [2]. Extended thymectomy is a well-accepted surgical treatment for selected MG patients. However, AchR-Ab is produced not only in the thymus, but also in external thymic tissues, including peripheral lymph nodes and bone marrow. Given these characteristics, we proposed that immunosuppressive therapy combined with extended thymectomy would be of benefit in generalized MG. In 1978, we began to offer high-dose steroid therapy combined with extended thymectomy to all generalized MG patients, and reported initial favorable results [3]. In the present study, we wished to clarify the clinical benefits of perioperative administration of high-dose prednisolone combined with extended thymectomy by examining the long-term outcomes of a consecutive series of 116 patients with generalized MG. We conducted a retrospective review of 116 patients diagnosed with generalized myasthenia gravis (MG) that received alternate-day oral administration of high-dose prednisolone and had undergone transsternal extended thymectomy between November 1978 and July 1997. Collected data included detailed patient characteristics, medical history, disease status, administered steroid dose, operative procedures, postoperative complications, pathologic diagnosis, medication history, and detailed follow-up data. These data were collected at the time of hospital discharge and from outpatient charts on a prospective basis. A diagnosis of generalized MG was made when there was clinical evidence of muscle weakness and abnormal fatigability, and where the response to edrophonium or intramuscular neostigmine was positive. Characteristic responses to low frequency repetitive stimulation and anti-acetylcholine receptor antibody (anti-AChR) assays were supportive of the diagnoses. A detailed investigation was made of the main demographic variables for each patient included in the study, and clinical status at symptom onset, maximal worsening, and postoperative observations were assessed according to the Myasthenia Gravis Foundation of America (MGFA) clinical classification scheme (Table 1) [4]. MGFA clinical classification based on the report by Jaretzki et al. [4] To reduce the risk of initial worsening of MG symptoms due to steroid administration, alternate-day oral administration of prednisolone (PSL) was initiated at 10–20 mg/day, and then gradually increased by 5–10 mg/week up to 80–100 mg/day [5]. When patients achieved a stable condition after the maintenance of high-dose PSL for 2–4 weeks, transsternal extended thymectomy was performed. Although pyridostigmine was used initially, it was reduced and discontinued after stable PSL use, if possible. Other immunosuppressive agents were used on a case-by-case basis. Plasmapheresis and high-dose intravenous immunoglobulin were introduced in cases of severe generalized weakness with/without bulbar symptoms. Histamine H2 blocker, Vitamin D, and potassium were routinely administered to all patients for prophylaxis of adverse steroid effects. Transsternal extended thymectomy with full vertical sternotomy was performed with the patients in the supine position according to the report by Masaoka et al. [6]. Two hundred milligrams of intravenous hydrocortisone sodium succinate was administered to each patient in the morning of the operative day, and then intraoperatively and postoperatively every 8 h to give steroid coverage. At the end of general anesthesia, an end-tracheal tube was extubated if PaO2 was greater than 300 Torr and PaCO2 less than 60 Torr under spontaneous ventilation at FiO2 of 1.0. Upon resumption of oral nutrition (usually at the evening of postoperative day 1), oral PSL was continued the next morning. If bulbar symptoms appeared or respiratory condition deteriorated, tracheal re-intubation and mechanical ventilation were introduced. Post-thymectomy myasthenic crisis was defined as total postoperative mechanical ventilation support for more than 48 h with no cardiopulmonary complications or cholinergic crisis [7]. After discharge from the hospital, patients visited the Neurology outpatient clinic regularly every month, unless medications were discontinued. From 2 to 4 months postoperatively, reduction of PSL was started at the rate of 5 mg/month or less according to patient symptoms. Patients were evaluated for disease status and medication requirement at every visit. Remission of MG symptoms was defined when a patient had no symptoms or signs of MG for the latest 1 year before the time point of evaluation. Transition of PSL dose was also recorded. Three representative side effects of steroid were recorded; compressive vertebral fracture with osteoporosis, cataract and steroid-induced diabetes diagnosed by a specialist in each field. Data were analyzed using the SAS software package Version 8.02 (Statistical Analysis System, Cary, NC, USA). Results were expressed as mean ± SD, or as median and range. Complete stable remission (CSR), pharmacologic remission (PR) and improvement (Imp) were set as analysis endpoints according to the report by Jaretzki et al. [4] (Table 1). Postoperative MGFA classification of each patient was determined at the month of operation every year under various interval length of patient follow-up. CSR was defined as no symptoms or signs of MG for at least 1 year with no therapy for MG during that time. PR was the same criteria as for CSR except for the use of some form of therapy. Imp was defined as a substantial decrease in pretreatment clinical manifestations or a sustained substantial reduction in MG medication. Recurrence (Rec) was defined as patients with CSR that subsequently developed clinical findings or restarted medication for MG. The crude CSR rate was calculated as: the number of patients with CSR/number of patients evaluated. Life table analysis was performed by the Kaplan–Meier method for CSR, PR + CSR, Imp and Rec [8]. Life table analysis of recurrence started at zero time, i.e. the time at which CSR could be assessed. Life table analysis with the log-rank test was used to investigate treatment effects with respect to MGFA classification before treatment or existence of thymoma. Cox's proportional hazard model was conducted to identify important factors for CSR and Rec. Patient characteristics are summarized in Table 2. Pretreatment distribution according to MGFA Classification was 79 patients (68.1%) with class II disease, 17 (14.7%) with class III, 9 (7.8%) with class IV, and 11 (9.5%) with class V. Following alternate-day administration of high-dose PSL, 62 patients (53.4%) showed improved MG symptoms before surgery. Although 53 patients with class II and one patient with class IV showed no improvement in MGFA class, no patients deteriorated. All patients received alternate-day administration of PSL that escalated up to 100 or 80 mg with an average duration of 114 days preoperatively. Of the 66 patients that initially received pyridostigmine, 42 (63.6%) discontinued treatment preoperatively. Patient characteristics While 44 patients (37.9%) experienced an initial exacerbation of MG, this problem was resolved with pyridostigmine or continuous administration of PSL in all cases. Postoperatively, seven patients (6.0%) needed blood transfusion and two patients underwent reoperation for continuous bleeding. Three patients (2.6%) experienced pneumonia. However, no patient was dead related to operation. Six patients (5.2%) experienced a post-thymectomy myasthenic crisis, and tracheotomy was required for three patients for prolonged mechanical ventilation. The average duration of mechanical ventilation was 202 ± 232 min. Discontinuation of PSL took at least 20 months following surgery due to the gradual dose reduction. Only 35 patients (30.1%) could discontinue PSL 2 years after surgery. There were eight patients (6.9%) with compressive vertebral fracture due to osteoporosis, 16 (13.8%) with cataract, and 6 (5.2%) with steroid-induced diabetes. All patients were followed up for more than 3 years. Sixty-six patients reached CSR during the study period, with crude cumulative CSR and PR + CSR rates of 44.8 and 62.7%, respectively. From life table analysis, 63.4, 83.5 and 95.5% of patients were in CSR, PR + CSR and Imp, respectively (Fig. 1), with 41.8, 52.8 and 63.4% of patients in CSR at 3, 5 and 10 years, respectively. Stratification according to pretreatment MGFA classification showed statistically significant differences with respect to CSR (p = 0.0039, Fig. 2), PR + CSR (p = 0.0125) and Imp (p = 0.0001). Existence of thymoma did not influence CSR (51.5 and 64.3% in patients without thymoma and 57.3 and 57.3% in patients with thymoma at 5 and 10 years, respectively). PSL dose transitions were similar between the pretreatment MGFA classes (Fig. 3). Multivariate analysis revealed that age and pretreatment MGFA classification tended to be independent predictors of CSR, but was not statistically significant (Table 3). Kaplan–Meier curves for time to complete stable remission (CSR), pharmacologic remission (PR) + CSR, and improvement (Imp). CSR and PR were defined when a patient had no symptoms or signs of MG for the latest 1 year before the time point of evaluation. Life table analysis showed that 63.4, 83.5 and 95.5% of patients were in CSR, PR + CSR and Imp, respectively. Kaplan–Meier curves for complete stable remission (CSR) according to pretreatment MGFA classification. CSR was defined when a patient had no symptoms or signs of MG for the latest 1 year before the time point of evaluation. Transition of prednisolone (PSL) dose according to pretreatment Myasthenia Gravis Foundation of America (MGFA) classification. Multivariate analysis of variables associated with complete stable remission Twenty-six patients with CSR experienced a recurrence of MG during the follow-up period (26/66, 39.4%). Of these, eight patients returned to CSR and eight defined as PR, with MG symptoms persisting in 10 patients. Recurrence rate after CSR by life table analysis was 36.8 and 46.0% at 3 and 5 years, respectively. Patients with thymoma exhibited a significantly higher rate of recurrence than those without thymoma (p = 0.001, Fig. 4). Twenty-six (22.4%) of the 116 MG patients had thymoma preoperatively. According to Masaoka's classification, 11 patients had stage I disease, 11 stage II, 3 stage III, and 1 stage IVa. Fifteen patients with stages II to IVa received prophylactic irradiation (mean 38.3 ± 8.0 gy, range 16–60 gy) to the whole mediastinum where tumor and thymus existed. No radiation was performed to the lung. Only one patient with stage IVa had a recurrence of thymoma. Probability of recurrence rate after complete stable remission with respect to the existence of thymoma. Life table analysis of recurrence started at zero time, i.e. the time at which CSR could be assessed. Myasthenia gravis is an autoimmune neuromuscular disorder with a chronic clinical course that requires long-term multiple-modality therapy. Despite the absence of randomized control trials, immunosuppressive therapy combined with thymectomy appears to exhibit superior outcomes compared to immunosuppressive therapy alone. Crude complete remission rates of thymectomy upon medical treatment range from 27 to 46% [9]. In this study, the crude cumulative CSR rate was 44.8% and life table analysis revealed probabilities of CSR at 3, 5 and 10 years of 41.8, 52.8 and 63.4%, respectively. Due to differences in complete remission criteria, it is difficult to compare our results with previous reports. Nonetheless, as our CSR criteria tended to be equivalent to or stricter than those used in previously published studies, our present results can still be regarded as consistent with previous reports [6,8–10]. In our study, preoperative steroid use improved MG symptoms, and only 5.2% of patients experienced a myasthenic crisis after extended thymectomy. To reduce the risk of initial drug-related worsening of MG and myasthenic crisis, alternate-day administration and gradual increases in PSL dose were employed. However, it still took an average of almost 4 months from the start of PSL to surgery (114 ± 131 days) to stabilize MG symptoms. Postoperatively, average PSL doses were 40–60 mg on alternate days during the 1st year and 10–20 mg on alternate days after the 2nd year. Steroids are associated with a variety of significant adverse effects. Indeed, a 50% incidence of cushionoid symptoms after steroid therapy has been reported in MG patients [11]. Therefore, long-term steroid use remains problematic and controversial [12,13]. As it can be difficult to determine the specific adverse effects of a steroid, we investigated the three representative side effects, namely compressive vertebral fracture with osteoporosis, cataract and steroid-induced diabetes. Wakata et al. [14] reported that the frequency of osteoporosis in MG patients that received alternate-day high-dose prednisolone was 11.5%, which was lower than the presumptive osteoporosis rate of the general population in Japan (22.6%). Rivner declared that diet and exercise could be used to decrease the probability of weight gain and diabetes, that calcium and Vitamin D supplements and hormone replacement could be used to minimize steroid-induced bone loss, and that hyperglycemia, hypertension, infection and ulcers could be treated with other medications [13]. Although it cannot be said that the incidence of steroid side effects in our study was acceptable, it should be possible to minimize the side effects of steroid use through the routine use of prophylactic medications, careful management, and appropriate reduction of PSL. However, as increased age and high pretreatment MGFA classification grade were possible risk factors for uncontrollable MG, as confirmed by multivariate analysis, and given the potentially serious side effects of steroids, the appropriate indications for high-dose steroid therapy should be clarified and other immunosuppressant medications considered. In our study, the incidence of myasthenic crisis after transsternal thymectomy was only 5.6%. This was consistent with or lower than observed in previous reports (6–34%); for instance, Kas et al. [15] and Watanabe et al. [7] reported rates of 33.6 and 11.5%, respectively. Both these reports applied the same myasthenic crisis criteria as our study. The reason for the low incidence of myasthenic crisis in the present study may be that MG symptoms were minimized prior to surgery, such that 63.6% of patients (42/66) had ceased anticholinesterase drugs before thymectomy. This is supported by the finding that anticholinesterase drug cessation reduces the risk of myasthenic crisis and cholinergic crisis [15,16]. Anticholinesterase can induce respiratory insufficiency due to bronchial spasms and increased airway secretions, and induce cholinergic crisis due to increased postoperative sensitivity for anticholinesterase drugs [17]. Results of the multivariate analysis demonstrated that age and pretreatment MGFA classification tended to be independent predictors of CSR, but not statistically significant, which was consistent with previous reports [18,19]. Many studies have shown that younger patients and lower MG stage show better responses to thymectomy with high remission rates. It has also been reported that steroid use after thymectomy is a predictor of poor outcome [18], as steroids are used in cases with poor responses to thymectomy. Our results indicated that cases with good responses to thymectomy could reduce PSL treatment levels. Although few articles have examined MG recurrence after temporary remission, it is certainly an important issue in the management of MG patients. While the presence of thymoma did not influence the outcome of MG treatment in our study, consistent with a report by Olanow et al. [17], multivariate analysis demonstrated that thymoma was a risk factor for recurrence of MG after CSR. Few previous reports have identified thymoma as a risk factor for MG recurrence after remission. Wakata et al. [20] reported that the periods between the onset of MG, thymectomy, and initiation of PSL were longer in relapsed cases with ocular symptoms than in non-relapsed cases after remission. Scherpbier and Oosterhuis [21] reported that the speed of dose reduction was not responsible for relapse in most MG patients, and that purely ocular cases and patients that underwent thymectomy without thymoma had higher chances to remain in remission after the discontinuance of prednisolone. Although the mechanisms of how thymoma induces MG recurrence remain unclear, several possibilities exist. For instance, thymoma is more common among older patients so that MG may be relatively more severe in patients with thymoma [22], or patients with thymoma may have an increased tendency for associated autoimmune disorders and extrathymic neoplasms [23]. It is also possible that residual thymoma or thymus after thymectomy may influence the outcome of MG treatment [9]. However, only one patient with stage IVa thymoma exhibited tumor recurrence, and transsternal extended thymectomy is less likely to leave thymic remnants than simple or transcervical thymectomy [9]. In conclusion, alternate-day administration of high-dose prednisolone may reduce the risk of post-thymectomy myasthenic crisis. Preoperative existence of thymoma was a risk factor of MG recurrence after CSR.