Independent risk factors for in-hospital outcome of myasthenic crisis: a prospective cohort study

Abstract

Background:

Objectives:

To explore the risk factors for in-hospital outcomes in patients with MC.

Methods:

Using a national neuromuscular center-based cohort of MG with prospective follow-ups from the crisis to the post-crisis phase, we finally included 90 MC episodes from 76 independent patients who received a standard regimen of rescue therapies.

Results:

The mean admission age was 52.89 ± 15.72 years. With a female predominance of 63.16% (48/76) and a high proportion of thymoma-associated MG (TMG) of 63.16% (48/76), the overall in-hospital mortality was 2.63% (2/76) and the average duration for mechanical ventilation (MV) use was 17.09 ± 13.36 days (0–53 days). In contrast to the patients with anti-acetylcholine receptor (AChR) antibodies, muscle-specific tyrosine kinase (MuSK)-associated MC exhibited a shorter MV support (5.20 ± 5.07 versus 17.40 ± 13.24 days, p = 0.023), length of intensive care units (ICU) stay (6.00 ± 4.64 versus 19.16 ± 17.54 days, p = 0.046), and hospital stay (16.00 ± 4.12 versus 34.43 ± 20.48 days, p = 0.011). Thymoma [odds ratio (OR): 0.200, 95% confidence interval (CI): 0.058–0.687, p = 0.011], partial pressure of carbon dioxide (PCO₂) in blood gas before MV (OR: 1.238, 95% CI: 1.015–1.510, p = 0.035), and pneumonia (OR: 0.204, 95% CI: 0.049–0.841, p = 0.028) were identified as independent risk factors for prolonged MV use. TMG patients with thymoma burden exhibited a notable longer MV use (22.08 ± 17.54 versus 8.88 ± 6.79 days, p = 0.001), a prolonged hospital stay (40.40 ± 26.13 versus 23.67 ± 13.83 days, p = 0.009) compared with non-TMG. Even with complete thymoma resection (R0), TMG exhibited an unfavorable outcome versus non-TMG.

Conclusion:

With timely rescue therapies and prospective follow-ups, the in-hospital outcome of MCs was substantially improved. Thymoma, PCO₂ in blood gas before MV, and pneumonia were identified as independent risk factors for prolonged MV use.

Plain Language Summary

Risk factors for in-hospital outcome of myasthenic crisis

Myasthenic crisis (MC) is a life-threatening condition for myasthenia gravis (MG). Therapeutic plasma exchange (TPE) and intravenous immunoglobulin (IVIg) efficaciously treat patients with MC. However, not every MC responds well to rescue therapies, and the determinants for outcome with the evidence from prospective cohorts are still lacking. Using a national neuromuscular center-based cohort of MG with prospective follow-ups from the crisis to the post-crisis phase, we were able to include 90 MC episodes from 76 independent patients who received a standard regimen of rescue therapies. The mean admission age was 52.89±15.72 years. With a female predominance and a high proportion of thymoma-associated MG. The overall in-hospital mortality was 2.63% (2/76) and the average duration for MV use was 17.09±13.36 days (0-53 days). In contrast to the patients with anti-AChR antibodies, MuSK-associated MC exhibited a shorter MV support, length of ICU stay and hospital stay. Thymoma, PCO2 in blood gas before MV, and pneumonia were identified as independent risk factors for prolonged MV use. TMG patients with thymoma burden exhibited a notable longer MV use, a prolonged hospital stay compared with non-TMG. Even with complete thymoma resection (R0), TMG exhibited an unfavorable outcome versus non-TMG. With timely rescue therapies and prospective follow-ups, the in-hospital outcome of MCs was substantially improved. Thymoma, PCO2 in blood gas before MV, and pneumonia.

Keywords

clinical outcomes intravenous immunoglobulin myasthenic crisis therapeutic plasma exchange thymoma

Introduction

Myasthenia gravis (MG) is an autoimmune disorder caused by dysfunction in the neuromuscular junction (NMJ) transmission, resulting in muscle weakness and fatigue.¹ The most prevalent autoantibody associated with MG is the anti-acetylcholine receptor (AChR) antibody, which can be detected in 80–87% of patients. Other antibodies targeting muscle-specific tyrosine kinase (MuSK) and low-density lipoprotein receptor-related protein 4 were positive in 5–8% and 1–5% of MG patients, respectively.^1–4 Based on the concurrence or history with thymoma, generalized MG (gMG) patients with positive anti-AChR antibodies can be further classified as thymoma-associated MG (TMG) and non-TMG, the latter of which included early-onset gMG (EOMG), and late-onset gMG (LOMG). The remaining gMG patients were mainly anti-MuSK antibody positive, or double negative.⁵

Myasthenic crisis (MC) is a potentially life-threatening complication that requires mechanical ventilation (MV) support in MG due to respiratory muscle weakness. Approximately 15–20% of MG patients experience MC during their lifetime. Despite the continuous efforts in medical technologies and the implementation of standardized treatment, MC remains a highly challenging clinical issue due to its high mortality rate and economic burden.^6–8 The hospital mortality rates of MC ranged from 5% to 22%.^9,10 A multicenter retrospective cohort study highlighted a refractory course during the crisis that was associated with anti-MuSK antibodies with a longer need for MV, intensive care units (ICU) stay, and hospital stay.¹¹ For patients with anti-AChR antibodies, LOMG was more prone to recurrent crises in contrast to EOMG.^12,13 In comparison to the AChR-MCs, seronegative MG patients with MC were younger, had a higher rate of thymus hyperplasia, and were more likely to be female. The duration between the diagnosis of MG and the occurrence of MC was longer in seronegative MC.¹⁴

As rescue therapies for patients with crisis, therapeutic plasma exchange (TPE) and intravenous immunoglobulin (IVIg) have been demonstrated equally efficacious and safe.^15,16 Even with these fast-acting therapies, the disease burden of MC was huge mainly due to the long-term immune therapies and ICU stay.¹⁷ During the pandemic, COVID-19 infections or vaccinations were associated with the rapid development of MC and unfavorable outcomes.^18,19 Previous studies have explored the risk factors for the development of MC or MG exacerbation and analyzed the clinical outcome in a multicenter retrospective cohort.^9,20 However, there is currently scarce prospective cohort study following the MG patients from the crisis. It remained unknown what would entail the poor response to therapies in MC and inform the advanced interventions. In addition, it is not known whether the advances in the preferential use of TPE and B cell-depleting therapies would benefit MG patients with crisis or not, in particular, MuSK-antibody-associated MCs, which were previously characterized as more refractory.

In the era with the emergence of several new biological agents with higher selectivity and target specificity, more options would be provided as sequential or maintenance therapies for refractory MG and patients with MC.²¹ Early identification of those who would exhibit unsatisfactory responses to the rescue therapies or long-term conventional immunosuppressants would aid in the timely initiation of targeted therapies.

Therefore, using a prospective cohort that was established in 2020 to follow MG patients from the crisis to the post-crisis state, we analyzed the clinical features and laboratory variables to explore the potential factors affecting the in-hospital outcome after standard rescue therapies.

Methods

Participants and study design

All patients with MC (classified as Myasthenia Gravis Foundation of America Class V, MGFA V) were collected from five neurological intensive care units (NICUs) or neurologically associated interdisciplinary ICUs in North, West, and General campus of Huashan Hospital Fudan University, Shanghai. The diagnosis of MG was established clinically in all patients and confirmed by two or three of the following tests: clinical response to neostigmine test, seropositivity for anti-AChR/anti-MuSK antibodies, or significant decremental response on repetitive nerve stimulation. Other MG mimicking diseases including Lambert–Eaton myasthenic syndrome, peripheral neuropathy, myopathies, and motor neuron diseases were excluded. The definition for MC was as follows: (1) respiratory failure; (2) blood gas analysis revealed that PO₂ <8 kPa, SO₂ < 90%, and/or PCO₂ > 6.65 kPa.

To explore the risk factors affecting the in-hospital outcome, we then only included MCs with rescue therapies with either TPE or IVIg during the crisis. Respiratory failure due to severe congestive heart failure, or adult respiratory distress syndrome were excluded. The MCs with only prednisone administration, or intolerance to complete the consecutive TPE or IVIg were excluded. Standardized prescriptions were used for all patients: five sections of plasma exchange with 1.0 volume for each session every other day or IVIg 0.4 g/kg/day for five consecutive days. We used antecubital veins or central venous catheters for TPE therapy as vascular access. Acid citrate dextrose was used for anticoagulation and 4–5% albumin and frozen plasma were used as replacement fluid. One plasma volume (PV) was calculated according to the following formula: PV = (1 − hematocrit) × 70 × body weight (kg).

Outcome measures and data collection

The primary outcome was the length of MV as measured by days including invasive and non-invasive MV. Late withdrawal was defined as the duration of MV exceeding 15 days and early withdrawal referred to those who were successfully weaned within 15 days.⁹ Secondary outcomes include the length of ICU stay and the length of hospital stay. We collected the blood gas analysis within 3 days before the MV use. Laboratory tests that were performed within 3 days before and after MV use were retrospectively reviewed. The maximum value in blood-glucose, and neutrophil/lymphocyte rate during the crisis was selected for subsequent analysis if there had been multiple tests. For serum potassium, calcium, hemoglobin, and albumin, the minimum value collected during the crisis was selected for further analysis.

The MG-associated clinical scores were prospectively assessed within 3 days after the use of MV. Baseline MGFA-quantitative MG test and MG manual muscle test scores were evaluated by experienced MG specialists. Each participant was evaluated by the MG-activities of daily living scale and MG quality of life 15 questionnaire. Demographic features, medication, and comorbidities were obtained by reviewing medical records. MGFA class at onset, thymoma history, thymectomy, therapies before the crisis, and treatments before the crisis were recorded. Comorbidities included diabetes mellitus type 2, malignancies, cardiovascular diseases, hyper- or hypothyroidism, liver dysfunction, renal dysfunction, chronic hepatitis, blood diseases, pemphigus, and chronic pulmonary diseases also were retrospectively analyzed. Pneumonia during the hospital was recorded. TMG patients who had complete resection of thymoma as the ‘R0-thymectomized’ group, whereas those who still had unresectable or recurrent thymomas as the ‘With thymoma burden’ group.

Statistical analysis

IBM SPSS version 25.0 (SPSS Inc., Chicago, IL, USA) and GraphPad Prism 8.0 (GraphPad Software, LLC, San Diego, CA, USA) were used for statistical analysis. Continuous variables were presented as the mean ± standard deviation. Categorical variables were expressed as frequencies and proportions. Missing values are replaced by sequence averages. Analysis of variance testing or the Kruskal–Wallis test was performed for the analysis of groups depending on the distribution of variables. Fisher’s exact test was used for categorical variables comparison. Considering the multiple comparisons, statistical significance was corrected using the Bonferroni correction. Factors associated with prolonged use of MV were analyzed using the total cohort with univariate and multivariate logistic regressions. Variables with a p-value of ⩽0.2 in the odds ratio (OR) defined by univariate analysis were further included in the multivariate analysis. Statistical significance was set at a p < 0.05.

Results

Baseline clinical features and outcomes

The prospective cohort was registered (NCT04837625) and collected from January 2020 through July 2023 with 112 MC episodes from 96 MG patients. Among them, three patients were excluded because their medical records were incomplete. Of the remaining 93 patients, we excluded 19 MC from 17 MG patients because of the prolonged interval from the initiation of MV to admission for over 3 days. Finally, we enrolled 90 MCs from 76 independent MG patients who had received standard rescue therapies at crisis (Figure 1).

Figure 1.

The workflow of the study design. A total of 90 MC episodes from 76 independent MG patients were enrolled in the final analysis.

Baseline demographic information is provided in Table 1. The mean age at admission to the hospital was 52.89 ± 15.72 years old. Of all 76 patients, 48 (63.16%) were female. Among the patients who experienced the first crisis (n = 64), the average duration from onset to this crisis was 34.71 ± 54.50 months. Among them, 61% (39/64) had the first crisis within 2 years after the MG diagnosis. For all 90 MC events, the interval from the MG onset to this crisis was 50.05 ± 69.40 months (range: 0.3–264). Regarding the antibody status, 92.11% (70/76) were anti-AChR antibodies, and 5.26% (4/76) were anti-MuSK antibodies, 2.63% (2/76) were seronegative. Of all anti-AChR antibody-associated MG patients, the proportion of the TMG was 68.57% (48/70). EOMG and LOMG both accounted for 15.71% (11/70). Two cases associated with thymoma (2.6%, 2/76) were seronegative for both anti-AChR and anti-MuSK antibodies.

Table 1.

Clinical characteristics and outcomes for all included MC.

Clinical Variables	Mean ± SD(range) or No.(%)
Myasthenic crisis episodes (n)	90
Number of independent MG patients	76
Age at admission, years, M ± SD	52.89 ± 15.72
Female, No. (%)	60 (66.67, of crisis)
	48 (63.16, of patients)
Weight, kg	59.29 ± 11.78
Height, m	1.64 ± 0.07
BMI, kg/m²	21.89 ± 3.63
Age at onset, years
Mean ± SD	48.70 ± 15.87
Range	7–84
Age at first crisis, years
Mean ± SD	51.86 ± 15.73
Range	18–88
First crisis, No. (%)	64 (71.11)
Multiple crisis, No. (%)	26 (28.89)
Disease duration, months
Mean ± SD	50.05 ± 69.40
Range	0.3–264
Classification by onset, No. (%)	(Of patients)
Ocular-onset	0
TMG	48 (63.16)
Early-onset	11 (14.47)
Late-onset	11 (14.47)
MuSK-MG	4 (5.26)
Seronegative-MG	2 (2.63)
Thymoma history, No. (%)	50 (65.79, of patients)
Thymectomy, No. (%)	37 (48.68, of patients)
Pathology: thymoma	36 (97.30)
Pathology: thymus hyperplasia	1 (2.70)
MG associated therapies before crisis, No (%)
Pyridostigmine	75 (83.33)
Glucocorticoids	74 (82.22)
Tacrolimus	18 (20.00)
Mycophenolate mofetil	3 (3.33)
Azathioprine	3 (3.33)
CTX	1 (1.11)
Non	16 (17.78)
Comorbidities, No. (%)	41 (53.95, of patients)
MGFA at onset, No. (%)
I	28 (31.11)
II	52 (57.78)
III	7 (7.78)
IV	3 (3.33)
Rescue therapies, No. (%)
TPE	18 (20.00)
IVIg	53 (58.89)
TPE + IVIg	19 (21.11)
Length of MV, days
Mean ± SD	17.09 ± 13.36
Range	0–53
Length of ICU stay, days
Mean ± SD	18.83 ± 17.38
Range	0–78
Length of hospital stay, days
Mean ± SD	34.02 ± 20.79
Range	9–85
Status at discharge, No. (%)
Independent	81 (90.00)
Discharged with MV	3 (3.33)
In a care facility	4 (4.44)
Dead	2 (2.22)
	2 (2.63, of patients)

BMI, body mass index; CTX, cyclophosphamide; ICU, intensive care units; TPE, therapeutic plasma exchange; IVIg, intravenous immunoglobulin; MC, myasthenic crisis; MG, myasthenia gravis; MGFA, Myasthenia Gravis Foundation of America; MuSK-MG, muscle-specific tyrosine kinase MG; MV, mechanical ventilation; TMG, thymoma-associated MG.

Pyridostigmine was used in 83.33% (75/90) of cases prior to the crisis. Immunotherapies before admission included glucocorticoids (74/90, 82.22%), tacrolimus (18/90, 20%), mycophenolate mofetil (3/90, 3.33%), azathioprine (3/90, 3.33%), and cyclophosphamide (1/90, 1.11%). The longest time of glucocorticoids and tacrolimus used were 216 and 242 months, respectively. A total of 53.95% (41/76) of participants had one or more chronic diseases including hypertension, diabetes, cardiac disease (congestive heart failure, coronary artery disease, and arrhythmia), tuberculosis, cancer, liver disease, kidney disease, thyroid disease.

In our study, the overall in-hospital mortality was 2.22% (2/90), both of which were associated with unresected thymoma and multiorgan failure. For all alive MG patients who used MV during the crisis (n = 81), 25% of patients were able to wean by hospitalization on Day 8 and 50% by Day 14. On Day 25.5, 75% of patients successfully weaned. When discharged from the hospital, 90% of MG patients did not require any assistance in walking, while three patients still relied on MV and four patients required further rehabilitation (Figure 2).

Figure 2.

Kaplan–Meier survival curve depicting the proportion of patients with MV dependence over time in MC.

Outcome difference between AChR-associated MC and MuSK-associated MC

There were no differences in baseline information between patients in AChR-associated MC (n = 83) versus MuSK-associated MC (n = 5). Females accounted for the majority in both groups (66.27% in the AChR-MC group and 80% in the MuSK-MC group). The age and BMI at admission, age, and MGFA at onset, and disease duration were equal in AChR-MC and MuSK-MC groups.

Regarding the clinical outcome, MuSK-MC had a favorable outcome in early ventilatory weaning in contrast to that in AChR-MC (p = 0.024). Other outcome measures were consistent with the evidence from the length of MV (5.20 ± 5.07 versus 17.40 ± 13.24, p = 0.023), the length of ICU stay (6.00 ± 4.64 versus 19.16 ± 17.54, p = 0.046), and the length of hospital stay (16.00 ± 4.12 versus 34.43 ± 20.48, p = 0.011) (Table 2).

Table 2.

In-hospital outcome of MC in AChR-associated MG versus MuSK-associated MG.

Clinical variables	AChR-MC (n = 83)	MuSK-MC (n = 5)	p
Sex, No. (%)			0.885
Male	28 (33.73)	1 (20)
Female	55 (66.27)	4 (80)
BMI			0.119
Mean ± SD	22.04 ± 3.65	19.39 ± 3.34
Range	15.06–35.05	16.94–24.80
Age at admission, year			0.200
Mean ± SD	53.42 ± 15.67	43.60 ± 18.72
Range	18–88	25–66
Age at onset, year			0.291
Mean ± SD	49.05 ± 15.91	41.20 ± 18.21
Range	7–84	24–65
MGFA at onset, No. (%)			0.366
I	27 (32.53)	0
II	46 (55.42)	5 (100)
III	7 (8.43)	0
IV	3 (3.61)	0
Disease duration,^a month			0.822
Mean ± SD	52.60 ± 71.48	25.60 ± 26.94
Range	0.3–264	3–57
Early or late-withdraw,^a No. (%)			0.024*
Early	46 (55.42)	5 (100)
Late	37 (44.58)	0
Rescue therapies, No. (%)			0.015*
PE	14 (16.87)	4 (80.00)
IVIg	50 (60.24)	1 (20.00)
PE + IVIg	19 (22.89)	0
Length of MV, days			0.023*
Mean ± SD	17.40 ± 13.24	5.20 ± 5.07
Range	0–53	0–13
Length of ICU stay, days			0.046*
Mean ± SD	19.16 ± 17.54	6.00 ± 4.64
Range	0–78	0–13
Length of hospital stay, days			0.011*
Mean ± SD	34.43 ± 20.48	16.00 ± 4.12
Range	9–85	13–21

Disease duration is defined as the time from the onset to admission for MC. Early withdrawal is defined as length MV use of 15 days or less, whereas late withdrawal is defined as the MV use time for more than 15 days.

AChR-MC, acetylcholine receptors-myasthenic crisis; BMI, body mass index; ICU, intensive care units; IVIg, intravenous immunoglobulin; MGFA, Myasthenia Gravis Foundation of America; MuSK-MC, muscle-specific tyrosine kinase MC; MV, mechanical ventilation; SD, standard deviation.*p<0.05.

Factors associated with prolonged use of MV

According to the days for ventilation, we then divided all 90 MC events into two groups: the early withdrawal group (n = 51) and the late withdrawal group (n = 39). Next, we performed univariate analysis including baseline data (sex, BMI, age at admission, age at onset), MGFA at onset, thymoma history, therapeutic regimen before MC, comorbidities, and laboratory testing results. The variables with p < 0.2 were subjected to multivariate analysis. Univariate [OR: 0.180, 95% confidence interval (CI): 0.060–0.541, p = 0.002] and multivariate (OR: 0.200, 95% CI: 0.058–0.687, p = 0.011) analyses showed that the thymoma history was associated with an increased risk of prolonged use of MV. PCO₂ before MV used (OR: 1.238, 95% CI: 1.015–1.510, p = 0.035) and pneumonia during hospitalization (OR: 0.204, 95% CI: 0.049–0.841, p = 0.028) was another two independent risk factor for prolonged MV use. Age at onset (OR: 1.030, 95% CI: 1.001–1.059, p = 0.042), blood glucose (OR: 1.194, 95% CI: 1.029–1.387, p = 0.020), serum calcium (OR: 0.020, 95% CI: 0.001–0.388, p = 0.010), and albumin (OR: 0.910, 95% CI: 0.839–0.987, p = 0.022) also associated with late withdrawal of MV, but all of them were insignificant in the multivariate analysis (Table 3).

Table 3.

Clinical and laboratory variables associated with a prolonged use of MV during the crisisa.

Clinical Variables	Univariate analysis		Multivariate analysis
	OR and 95% CI	p Value	OR and 95% CI	p Value
Male	1.500 (0.621–3.626)	0.368
BMI	1.017 (0.906–1.141)	0.778
Age at onset	1.030 (1.001–1.059)	0.042*
Type of crisis
First crisis	Reference
Multiple crisis	0.619 (0.244–1.571)	0.312
MGFA at onset
I	Reference
II	0.667 (0.054–8.240)	0.752
III	0.389 (0.033–4.585)	0.453
IV	0.500 (0.028–8.952)	0.638
Thymoma history
No	Reference
Yes	0.180 (0.060–0.541)	0.002*	0.200 (0.058–0.687)	0.011*
Immunotherapies before admission
None	Reference
One	0.533 (0.126–2.265)	0.394
Two	1.000 (0.370–2.699)	1.000
Comorbidities	0.682 (0.290–1.604)	0.380
Baseline MMT	0.989 (0.961–1.018)	0.463
Baseline MGFA-QMG	0.990 (0.922–1.064)	0.794
Baseline ADL	1.003 (0.913–1.101)	0.955
Baseline QoL-15	1.031 (0.977–1.088)	0.260
PCO₂ before MV	1.183 (1.022–1.368)	0.024*	1.238 (1.015–1.510)	0.035*
PO₂ before MV	0.989 (0.939–1.040)	0.658
Lactic acid before MV	1.282 (0.813–2.023)	0.285
Serum bicarbonate before MV	0.958 (0.863–1.064)	0.425
Blood glucose	1.194 (1.029–1.387)	0.020*
Hemoglobin	0.978 (0.954–1.003)	0.088
Serum potassium	1.148 (0.436–3.023)	0.780
Serum calcium	0.020 (0.001–0.388)	0.010*
Albumin	0.910 (0.839–0.987)	0.022*
Neutrophil/lymphocyte	1.025 (0.999–1.053)	0.061
Pneumonia during hospitalization	0.129 (0.035–0.476)	0.002*	0.204 (0.049–0.841)	0.028*

Enrolled participants (n = 90). Blood gas analysis was the last time before MV. Blood glucose and neutrophil/lymphocyte were the highest values for 3 days prior to and 3 days after MV. Serum calcium, serum potassium, hemoglobin, and albumin were the lowest value for 3 days prior to and 3 days after MV.

ADL, activities of daily living; BMI, body mass index; CI, confidence interval; MGFA-QMG, Myasthenia Gravis Foundation of America-quantitative MG test; MMT, MG manual muscle test; MV, mechanical ventilation; OR, odds ratio; PCO₂, partial pressure of carbon dioxide; QoL-15, quality of life 15.

p<0.05.

Thymoma concurrence was associated with a higher disease burden in MC

Since thymoma was associated with an unfavorable outcome for MC, we then asked if there were any differences in this impact between the TMG with thymoma burden and those who had R0 thymectomy. To avoid the potential effects of thymectomy, we excluded the post-thymectomy crisis (n = 7). Then we classified all crisis events (n = 83) into three groups including the non-thymoma associated group (n = 29), R0 thymectomized group (n = 34), and those with thymoma burden (n = 20). We defined patients with no thymoma as ‘non-thymoma-associated’ (Table 4).

Table 4.

Subgroup analysis in MC stratified by the thymomaa.

Clinical variables	No thymoma history	With thymoma history		p Value
	Non-thymoma associated (n = 29)	R0 thymectomized (n = 34)	With thymoma burden (n = 20)	p Value
Sex, No. (%)				0.135
Male	10 (34.48)	14 (41.18)	3 (15.00)
Female	19 (65.52)	20 (58.82)	17 (85.00)
Age at admission, years				0.164
Mean ± SD	49.31 ± 18.50	51.09 ± 13.32	58.15 ± 15.96
Range	18–77	24–75	27–88
Age at onset, years				0.032*
Mean ± SD	42.97 ± 17.81	47.59 ± 13.21	55.15 ± 16.09
Range	7–77	24–72	27–84
BMI				0.037*
Mean ± SD	21.90 ± 4.41	20.92 ± 2.74	23.59 ± 3.71
Range	15.06–35.05	15.63–26.12	16.00–28.60
Disease duration, months				0.036*
Mean ± SD	76.53 ± 86.36	40.93 ± 44.79	36.79 ± 76.67
Range	0.5–264	0.5–171	0.3–259
Classification by onset, No. (%)				0.895
Early-onset	15 (51.72)	17 (50.00)	9 (45.00)
Late-onset	14 (48.28)	17 (50.00)	11 (55.00)
MGFA at onset, No. (%)				0.657
I	7 (24.14)	12 (35.29)	6 (30.00)
II	19 (65.52)	18 (52.94)	11 (55.00)
III	3 (10.34)	3 (8.82)	1 (5.00)
IV	0	1 (2.94)	2 (10.00)
Comorbidities, No. (%)	13 (44.83)	22 (64.71)	13 (65.00)	0.213
Immunotherapies before admission, No. (%)				0.488
No	7 (24.14)	3 (8.82)	2 (10.00)
One drug	15 (51.72)	24 (70.59)	12 (60.00)
Two drugs	7 (24.14)	7 (20.59)	6 (30.00)
PCO₂ before MV, kPa				0.232
Mean ± SD	8.35 ± 2.95	7.81 ± 2.46	9.36 ± 3.76
Range	4.95–15.07	4.49–14.28	4.46–17.25
Pneumonia during hospitalization, No. (%)	18 (62.06)	28 (82.35)	15 (75.00)	0.189
Length of MV, days				<0.001*
Mean ± SD	8.88 ± 6.79	18.93 ± 11.25	22.08 ± 17.54
Range	0–52	0–44	0–53
Length of ICU stay, days				0.055
Mean ± SD	13.53 ± 15.69	22.32 ± 17.52	18.20 ± 18.84
Range	0–78	0–68	0–61
Length of hospital stay, days				0.008*
Mean ± SD	23.67 ± 13.83	37.65 ± 20.83	40.40 ± 26.13
Range	9–82.50	11–82	12–85

BMI, body mass index; ICU, intensive care units; MC, myasthenic crisis; MGFA, Myasthenia Gravis Foundation of America; MV, mechanical ventilation; PCO₂, partial pressure of carbon dioxide; SD, standard deviation.

p<0.05.aThe post-thymectomy crisis (n=7) were excluded.

There is no significant difference in sex, age at admission, MGFA at onset, comorbidities, immunotherapies before admission, and the incidence of pneumonia among the three groups. The patients with thymoma burden exhibited elder onset age (p = 0.032), higher BMI (p = 0.037), and shorter disease duration from the onset to the crisis (p = 0.036) at admission. In addition, TMG patients with thymoma burden had a significantly longer length of MV (p < 0.001) and a longer length of hospital stay (p = 0.008).

Further pairwise comparisons also demonstrated a longer length of MV in the group with thymoma burden (22.08 ± 17.54 versus 8.88 ± 6.79, p < 0.001), as well as in hospital stay (40.40 ± 26.13 versus 23.67 ± 13.83, p = 0.017) compared with the non-thymoma-associated group. R0 thymectomized group also had a longer length of MV (18.93 ± 11.25 versus 8.88 ± 6.79, p = 0.004) and a longer hospital stay (37.65 ± 20.83 versus 23.67 ± 13.83, p = 0.023). However, there are no significant differences in length of MV, length of ICU stay, and length of hospital stay between R0 thymectomized group and the patients with thymoma burden (Figure 3).

Figure 3.

Baseline features and in-hospital outcome with significant differences among MC stratified by the thymoma. Baseline features included age at onset (a), BMI (b), and Disease duration (c). In-hospital outcome variables included length of MV (d), Length of ICU stay (e), and Length of hospital stay (f). Subgroups included non-thymoma associated group (A), R0 thymectomized group (B), and with thymoma burden group (C).

Discussion

In this study, we analyzed a cohort with 90 MC events in 76 individual patients with standard rescue therapies. The prospective nature allowed better clinical decision-making based on the MG subtypes and disease status. Although it is a single-center analysis, it utilizes the resources of the National Centre for Neurological Disorders and represents the MG patients referred from different neuromuscular diagnostic centers in China.

The in-hospital mortality rate of MC decreased dramatically from 42% in the 1960s to 5–12%.^9,22 Recently, a national population-based registry study for MG in China reported that the admission mortality was 6.5%,⁸ while the mortality rate of MC in the USA is approximately 4.5%.²³ In comparison to the previous retrospective studies of MC, our cohort consisted of MG patients with younger age, female predominance, and a high proportion associated with thymoma. TMG was reported to have more severe myasthenic symptoms, a worse prognosis, and higher mortality.²⁴ However, the mortality rate in our cohort was as low as 2.63% (2/76). This progress might be attributed to the rapid stratification and active intervention in MC patients. As our cohort consisted entirely of patients treated with rescue therapies, the efficacy of rescue therapies in treating crisis was again validated in this real-world cohort.²⁰

MuSK-antibody was reported to be an independent risk factor for disease exacerbation in MG patients.²⁰ MuSK-antibody was associated with increased days of MV use (43.0 ± 53.1 days), length of ICU stay (45.3 ± 49.5 days), and length of hospital stay (55.9 ± 47.6 days) for patients with MC, which were mainly collected between 2006 and 2015.¹¹ Unlike the AChR antibody-mediated immunopathology, MuSK-MG is consistent with a Th2 response, and it is largely governed by the IgG4 subclass that mediates pathology by physically blocking NMJ protein–protein interactions. They have a very limited capacity to induce complement and cell activation.^25,26 Further, MuSK-associated MG might not respond well to IVIg probably because IVIg has limited capacity for neutralizing IgG4 or significantly enhancing IgG4’s catabolism through FcRn. Another reason is that IgG4 cannot activate the complement and form immune complexes with antigens. Additionally, there is no recruitment of immune cells via Fc receptors, and no antigenic destruction through phagocytosis or antibody-dependent cellular cytotoxicity in IgG4-mediated autoimmune neuropathy. These factors hinder the effectiveness of IVIg treatment in MuSK-associated MG.^27,28 It has been reported that patients with MuSK MG tend to respond better to TPE than to IVIg with evidence showing a shorter duration of MV and a higher rate of improvement,^11,29,30 which is in line with our previous study in the MuSK-MG cohort.³¹ In this prospective cohort recruited from 2020, we used TPE as the rescue therapy and Rituximab as off-label therapy for the maintenance therapies for MG patients once the anti-MuSK antibody was identified. Owing to these recent advances in therapies, the proportion of MuSK-MC in the whole MC cohort was relatively low (5.6%, 5/90) in comparison to the German multicenter study (7.6%, 19/250). Further, our MuSK-MC patients had a shorter duration of MV use, ICU stay, and hospital stay compared with the MuSK-MC in the previous retrospective multicenter cohort.

Pathological abnormalities in the thymus were identified in more than 80% of AChR-MG patients.³² Thymoma has been identified in 10–30% of MG patients with positive anti-AChR antibodies, while up to 50% of patients with thymoma develop MG.² Previous reports have indicated that thymoma was a risk factor for disease exacerbation and MC recrudescence.^20,33 Multiple crises occurred in 28.89% (26/90) of the overall cohort, among which 76.9% (20/26) were associated with thymoma. In TMG, autoreactive T cells are released from the thymoma due to defective negative selection, which activates B cells to produce AChR-Ab in the periphery.^34,35 The over-expansion in the cortical region of thymomas leads to insufficient regulatory T-cell generation for maintaining the peripheral autoimmune process.³⁶ These are the main causes of the association between thymoma and MG exacerbation.

Thymectomy is a crucial treatment for MG patients with thymoma which leads to superior clinical improvement and remission rates.^2,37 It is even considered the only treatment that offers the possibility of achieving complete remission.³⁸ The therapeutic effect was observed not only in patients with thymoma,³⁹ but also in patients with non-thymomatous MG.^40,41 However, there has been no report on whether surgery has an impact on patients experiencing a relapse crisis after thymectomy (not post-thymectomy MC). A total of 34 crisis cases (40.96%, 34/83) from 27 patients (38.57%, 27/70) that experienced crisis after complete surgical remission in our study. We referred to this group as the ‘R0 thymectomized’ group. These patients had a similar fundamental state, but had completely different outcomes compared to patients in the ‘non-thymoma associated’ group. These patients who underwent thymectomy have a longer duration of MV (p = 0.004) and hospital stay (p = 0.023) compared to patients in the non-thymomatous associated group.

Our study observed that TMG patients would still have MG exacerbation or MC, even after the thymectomy. It may be explained by the following reasons. First, a large number of thymoma-derived, potentially autoreactive T cells gradually replace the patient’s more tolerant T-cell repertoire in the peripheral circulation.⁴² Second, MG-associated B-cell clones mature in the thymus and later enter the blood circulation, which was detected 12 months after thymectomy, and their persistence correlated with less favorable changes in clinical symptoms after thymectomy.⁴³ In addition, the spleen, bone marrow, and lymph nodes may serve as suitable niches or producers for autoreactive cells, thus perpetuating them in MG patients after thymectomy.⁴⁴ It has also been reported by earlier research that lymphocytes in blood and bone marrow can synthesize AChR antibodies in vitro.⁴⁵ All these reasons, together with the loss of thymic immunomodulatory roles after thymectomy,⁴⁴ lead to the impact of the thymoma on the peripheral immune system.

Prolonged duration of MV increased the risk of ventilator-associated pneumonia and vice versa.^9,46,47 In MG with extended thymectomy, postoperative pneumonia prolonged the time of MV use and hindered the weaning process after rescue therapies.⁴⁸ Thus the early prevention and intervention for pneumonia according to our previous identification of MG-relevant microbial spectrum for pneumonia is critical to improve the in-hospital outcome for MC.⁴⁹

The results in our prospective study are consistent with the previous retrospective study, in which pneumonia was identified as a risk factor for prolonged use of MV in MC.⁹ Another retrospective study has highlighted the importance of thymoma on the MG exacerbation, which was mainly focused on the development of MC, but not the clinical outcome of MC. Our study revealed two additional risk factors, thymoma, and PCO₂ in blood gas, which might help in the early stratification of patients with worse prognoses and intervention for a better in-hospital outcome of MC.²⁰

The therapeutic goal of MG is to achieve complete remission or alleviate symptoms to a mild extent. For patients with MC, some were very refractory to repeated rescue therapies and required a long ICU stay. Emerging biological agents targeting FcRn and complements exhibit favorable safety and tolerability in generalized or refractory MG. Future application of these biologics as add-on therapies to facilitate the weaning process and improve the clinical outcome is anticipated.^50,51

Our study has limitations. First, it has a selection bias toward those severe cases since these patients were referred to the National Centre for Neurological Disorders in China. This was also reflected by a large proportion of TMG patients which were associated with worse prognosis. Second, a small sample size from one center is another main limitation. We only recruited those patients with MC who had the same regimen of rescue therapies to avoid the confounders from treatment. A future multicenter cohort study with a large sample size is required to validate the risk factors for in-hospital outcomes. Although it is prospective, the immunotherapies prior to the crisis and the comorbidities were heterogeneous which may bring confounders to this study.

Conclusion

In this prospective cohort study for MC with rescue therapies, we achieved a low in-hospital mortality rate of 2.63%. Thymoma, PCO₂ in blood gas before MV, and pneumonia were identified as independent risk factors for prolonged MV use. TMG patients had a worse in-hospital outcome. These findings highlighted the importance of timely rescue therapies, pneumonia prevention, and early risk stratification based on the clinical phenotype in patients with MC.

Footnotes

Author note

For the Pan-Yangtze River Delta Alliance for Neuromuscular Disorders Co-investigator – Li Zeng: Department of Neurology, Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China, Sichuan, China; Jie Yang: Department of Neurology, Wuhan No. 1 Hospital, Wuhan, China; Heng Li: Department of Neurology, Central Hospital Affiliated to Shandong First Medical University, Jinan, China; Zhangyu Zou: Department of Neurology, Fujian Medical University Union Hospital, Fuzhou, China; Song Yang: Department of Neurology, Changzhou First Hospital, Changzhou, China; Yali Zhang: Department of Neurology, Inner Mongolia Medical University, Inner Mongolia, China; Qing Ke: Department of Neurology, the First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China.

Declarations

ORCID iD

Sushan Luo

References

Gilhus

Skeie

Romi

, et al. Myasthenia gravis – autoantibody characteristics and their implications for therapy. Nat Rev Neurol 2016; 12: 259–268.

Mane-Damas

Molenaar

Ulrichts

, et al. Novel treatment strategies for acetylcholine receptor antibody-positive myasthenia gravis and related disorders. Autoimmun Rev 2022; 21: 103104.

Pechlivanidou

Ninou

Karagiorgou

, et al. Autoimmunity to neuronal nicotinic acetylcholine receptors. Pharmacol Res 2023; 192: 106790.

Vakrakou

Karachaliou

Chroni

, et al. Immunotherapies in MuSK-positive myasthenia gravis; an IgG4 antibody-mediated disease. Front Immunol 2023; 14: 1212757.

Huijbers

Marx

Plomp

, et al. Advances in the understanding of disease mechanisms of autoimmune neuromuscular junction disorders. Lancet Neurol 2022; 21: 163–175.

Mevius

Jores

Biskup

, et al. Epidemiology and treatment of myasthenia gravis: a retrospective study using a large insurance claims dataset in Germany. Neuromuscul Disord 2023; 33: 324–333.

Sobieszczuk

Napiorkowski

Szczudlik

, et al. Myasthenia gravis – treatment and severity in nationwide cohort. Acta Neurol Scand 2022; 145: 471–478.

Chen

Tian

Zhang

, et al. Incidence, mortality, and economic burden of myasthenia gravis in China: a nationwide population-based study. Lancet Reg Health West Pac 2020; 5: 100063.

Neumann

Angstwurm

Mergenthaler

, et al. Myasthenic crisis demanding mechanical ventilation: a multicenter analysis of 250 cases. Neurology 2020; 94: e299–e313.

10.

Hsu

Chen

Huang

, et al. Hemogram parameters can predict in-hospital mortality of patients with myasthenic crisis. BMC Neurol 2021; 21: 388.

11.

Konig

Stetefeld

Dohmen

, et al. MuSK-antibodies are associated with worse outcome in myasthenic crisis requiring mechanical ventilation. J Neurol 2021; 268: 4824–4833.

12.

Zivkovic

Clemens

Lacomis

Characteristics of late-onset myasthenia gravis. J Neurol 2012; 259: 2167–2171.

13.

Chaudhuri

Behan

PO.

Myasthenic crisis. QJM 2009; 102: 97–107.

14.

Mergenthaler

Stetefeld

Dohmen

, et al. Seronegative myasthenic crisis: a multicenter analysis. J Neurol 2022; 269: 3904–3911.

15.

Barth

Nabavi

, et al. Comparison of IVIg and PLEX in patients with myasthenia gravis. Neurology 2011; 76: 2017–2023.

16.

Ebadi

Barth

Bril

Safety of plasma exchange therapy in patients with myasthenia gravis. Muscle Nerve 2013; 47: 510–514.

17.

Phillips

Abreu

Goyal

, et al. Real-world healthcare resource utilization and cost burden assessment for adults with generalized myasthenia gravis in the United States. Front Neurol 2021; 12: 809999.

18.

Rodrigues

de Freitas

Lima

, et al. Myasthenia gravis exacerbation and myasthenic crisis associated with COVID-19: case series and literature review. Neurol Sci 2022; 43: 2271–2276.

19.

Sonigra

Sarna

Vaghela

, et al. An interesting case of fatal myasthenic crisis probably induced by the COVID-19 vaccine. Cureus 2022; 14: e23251.

20.

Nelke

Stascheit

Eckert

, et al. Independent risk factors for myasthenic crisis and disease exacerbation in a retrospective cohort of myasthenia gravis patients. J Neuroinflammation 2022; 19: 89.

21.

Menon

Bril

Pharmacotherapy of generalized myasthenia gravis with special emphasis on newer biologicals. Drugs 2022; 82: 865–887.

22.

Cohen

Younger

Aspects of the natural history of myasthenia gravis: crisis and death. Ann N Y Acad Sci 1981; 377: 670–677.

23.

Claytor

Cho

Myasthenic crisis. Muscle Nerve 2023; 68: 8–19.

24.

Alvarez-Velasco

Gutierrez-Gutierrez

Trujillo

, et al. Clinical characteristics and outcomes of thymoma-associated myasthenia gravis. Eur J Neurol 2021; 28: 2083–2091.

25.

Gomez

Van Den Broeck

Vrolix

, et al. Antibody effector mechanisms in myasthenia gravis – pathogenesis at the neuromuscular junction. Autoimmunity 2010; 43: 353–370.

26.

Masi

O’Connor

KC.

Novel pathophysiological insights in autoimmune myasthenia gravis. Curr Opin Neurol 2022; 35: 586–596.

27.

Dalakas

MC.

IgG4-mediated neurologic autoimmunities: understanding the pathogenicity of IgG4, ineffectiveness of IVIg, and long-lasting benefits of anti-B cell therapies. Neurol Neuroimmunol Neuroinflamm 2022; 9: e1116.

28.

Dalakas

MC.

Autoimmune neurological disorders with IgG4 antibodies: a distinct disease spectrum with unique IgG4 functions responding to anti-B cell therapies. Neurotherapeutics 2022; 19: 741–752.

29.

Binks

Vincent

Palace

Myasthenia gravis: a clinical-immunological update. J Neurol 2016; 263: 826–834.

30.

Pasnoor

Wolfe

Nations

, et al. Clinical findings in MuSK-antibody positive myasthenia gravis: a U.S. experience. Muscle Nerve 2010; 41: 370–374.

31.

Zhou

Yan

, et al. Short-term effect of low-dose rituximab on myasthenia gravis with muscle-specific tyrosine kinase antibody. Muscle Nerve 2021; 63: 824–830.

32.

Cavalcante

Le Panse

Berrih-Aknin

, et al. The thymus in myasthenia gravis: site of ‘innate autoimmunity’? Muscle Nerve 2011; 44: 467–484.

33.

Ozyurt

Nazli

Tutkavul

, et al. Occurrence and severity of myasthenic crisis in an unselected Turkish cohort of patients with myasthenia gravis. Front Neurol 2023; 14: 1201451.

34.

Marx

Willcox

Leite

, et al. Thymoma and paraneoplastic myasthenia gravis. Autoimmunity 2010; 43: 413–427.

35.

Thomas

Mayer

Gungor

, et al. Myasthenic crisis: clinical features, mortality, complications, and risk factors for prolonged intubation. Neurology 1997; 48: 1253–1260.

36.

Marx

Pfister

Schalke

, et al. The different roles of the thymus in the pathogenesis of the various myasthenia gravis subtypes. Autoimmun Rev 2013; 12: 875–884.

37.

Bachmann

Burkhardt

Schreiter

, et al. Thymectomy is more effective than conservative treatment for myasthenia gravis regarding outcome and clinical improvement. Surgery 2009; 145: 392–398.

38.

Godoy

Mello

Masotti

, et al. The myasthenic patient in crisis: an update of the management in Neurointensive Care Unit. Arq Neuropsiquiatr 2013; 71: 627–639.

39.

Blalock

Mason

Morgan

, et al. Myasthenia gravis and tumors of the thymic region: report of a case in which the tumor was removed. Ann Surg 1939; 110: 544–561.

40.

Wolfe

Kaminski

Aban

, et al. Randomized trial of thymectomy in myasthenia gravis. N Engl J Med 2016; 375: 511–522.

41.

Wolfe

Kaminski

Aban

, et al. Long-term effect of thymectomy plus prednisone versus prednisone alone in patients with non-thymomatous myasthenia gravis: 2-year extension of the MGTX randomised trial. Lancet Neurol 2019; 18: 259–268.

42.

Buckley

Douek

Newsom-Davis

, et al. Mature, long-lived CD4+ and CD8+ T cells are generated by the thymoma in myasthenia gravis. Ann Neurol 2001; 50: 64–72.

43.

Jiang

Hoehn

Lee

, et al. Thymus-derived B cell clones persist in the circulation after thymectomy in myasthenia gravis. Proc Natl Acad Sci USA 2020; 117: 30649–30660.

44.

Chen

Yang

Poor responses and adverse outcomes of myasthenia gravis after thymectomy: predicting factors and immunological implications. J Autoimmun 2022; 132: 102895.

45.

Lisak

Laramore

Zweiman

, et al. In vitro synthesis of antibodies to acetylcholine receptor by peripheral blood mononuclear cells of patients with myasthenia gravis. Neurology 1983; 33: 604–608.

46.

Papazian

Klompas

Luyt

CE.

Ventilator-associated pneumonia in adults: a narrative review. Intensive Care Med 2020; 46: 888–906.

47.

Rabinstein

Mueller-Kronast

Risk of extubation failure in patients with myasthenic crisis. Neurocrit Care 2005; 3: 213–215.

48.

Huang

Zhang

, et al. Risk factors for postoperative myasthenic crisis after thymectomy in patients with myasthenia gravis. J Surg Res 2021; 262: 1–5.

49.

Jin

Jiao

, et al. Pneumonia in myasthenia gravis: microbial etiology and clinical management. Front Cell Infect Microbiol 2022; 12: 1016728.

50.

Dalakas

MC.

Advances in the therapeutic algorithm for myasthenia gravis. Nat Rev Neurol 2023; 19: 393–394.

51.

Chen

Qiu

Gao

, et al. Myasthenia gravis: molecular mechanisms and promising therapeutic strategies. Biochem Pharmacol 2023; 218: 115872.