Reliability and Validity of Cough Peak Flow Measurements in Myasthenia Gravis

Decreased cough strength in myasthenia gravis (MG) leads to aspiration and increases the risk of MG crisis. The aim of this study was to clarify the reliability and validity of cough peak flow (CPF) measurements in MG. A total of 26 patients with MG who underwent CPF measurements using the peak flow meter by themselves were included. MG symptoms were evaluated by pulmonary function tests and clinical MG assessment scales before and after immune-treatments. The relationship between CPF and pulmonary function tests and MG comprehensive were assessed. The cut-off value of CPF for aspiration risk was determined and the area under the curve (AUC) was calculated. The intraclass correlation coefficient was more than 0.95 for pre-and post-treatment. Positive correlations were found between CPF and almost all spirometric values as well as between the differences of pre-and post-treatment in CPF and quantitative myasthenia gravis score. The CPF for identifying the aspiration risk was used to calculate the CPF cut-off value of 205 L/min with a sensitivity of 0.77, specificity of 0.90, and AUC of 0.85. The CPF, a convenient measure by patients themselves, has a high reliability in patients with MG, and is a useful biomarker reflecting MG symptoms.


Introduction
Myasthenia gravis (MG) is an autoimmune disease in which antibodies bind to acetylcholine receptors or to functionally related molecules in the postsynaptic membrane at the neuromuscular junction [ 1 ].Clinical features of MG range from ocular symptoms, such as ptosis and diplopia, to bulbar signs, weakness of the extremities, facial muscle, and respiratory muscles [ 2 ].Among the symptoms of MG, respiratory muscle weakness causes acute worsening of respiratory failure and increases the risk of MG crisis requiring intubation or noninvasive ventilation [ 3 ].In patients with MG, the evaluation of cough strength is crucial in clinical settings as it is linked to the risk of infections such as aspiration pneumonia, which accounts for 38% of MG crises [ 4 ].Additionally, decreased cough strength also contributes to the risk of aspiration pneumonia [ 5 ].
Recently, cough peak flow (CPF) has gained attention as an objective measure reflecting cough strength.Many reports described below have emphasized the importance of measuring CPF using a peak flow meter in general practice [ 6 ].Peak flow meters have been reported to be useful in the establishment of asthma diagnosis and have been widely adopted for monitoring patients with asthma [ 7 ].Peak flow meters are applied to the CPF as a measure of coughing force.CPF has been mainly studied in patients with neuromuscular diseases and to predict the success for extubating patients in the critical care setting [ 6 ].Moreover, CPF has been utilized in patients with Duchenne muscular dystrophy [ 8 ] and amyotrophic lateral sclerosis [ 9 ].However, the relationship between clinical symptoms and CPF in patients with MG has not been thoroughly investigated.There are spirometric values available as a method for evaluating respiratory function in MG patients, but these tests have only been conducted in medical facilities, and there is a paucity of self-assessment measures for patients to evaluate their respiratory function and MG symptoms in their daily lives.Therefore, the worsening of MG symptoms often relies on the patient's self-awareness.
The purpose of this study was to determine the reliability of CPF measurements in MG patients and to establish the validity of CPF measurements by investigating their relationship with respiratory function and MG symptoms.If the reliability and validity of CPF measurements in assessing patients with MG are established, it is conceivable that CPF measurements could serve as a surrogate indicator for assessing MG symptoms of respiratory and bulbar function through a more convenient method.Furthermore, this study holds clinical significance as CPF measurement provides a convenient assessment tool that can be performed by the patients themselves at the bedside or at home.This allows for the possibility of timely treatment in case of acute worsening of MG symptoms.

Study design
A retrospective cohort study of patients with MG who were evaluated at Chiba University Hospital between January 2015 and December 2020 was conducted.Electronic Medical Records were reviewed for clinical and demographic details.

Study population
This study included inpatients with MG aged over 20 years old who underwent immunotherapy at the Chiba University Hospital and whose consent for CPF measurement was obtained.The diagnosis of MG was supported by positive serology for binding acetylcholine receptor autoantibodies, muscle-specific tyrosine kinase (MuSK) antibodies and also a positive edrophonium test, abnormal repetitive nerve stimulation, or abnormal singlefiber electromyography.Exclusion criteria were patients who were unable to follow the instructed actions due to impaired consciousness, dementia, or psychiatric symptoms, and patients with poorly controlled subjective symptoms such as dyspnea and pain.
Our study was approved by the institutional review board at Chiba University Graduate School of Medicine (approval no.4089) and conducted in accordance with the amended Declaration of Helsinki.All participants provided informed consent before participation.

Procedures and outcomes
To measure CPF, a face mask was attached to a peak flow meter (Assess Peak Flow Meter HS710012, Philips Respironics, Murrysville, PA, USA) [ 10 ].Subjects were instructed to take a full inspiration and voluntary cough with maximum strength into the device.CPF measurement was conducted in the sitting position [ 11 ] with the measuring device attached to the face mask to prevent air leakage ( Fig. 1 ).A total of three maneuvers were performed for each patient.Before the measurements, an orientation and pre-practice of the measurement method was given by the examiner to ensure that the patients were able to perform the measurements themselves, followed by the actual measurements after at least one hour rest.Three time measurements were performed each at before breakfast, lunch, and dinner, for a total of 9 measurements per day conducted by the patients themselves.CPF measurements were conducted before meals in order to minimize the effects of acetylcholinesterase inhibitors taken by patients.Between each CPF measurement within a session, patients took a rest interval of 30 s.

Statistical analysis
The rationale for setting the sample size design based on a two-sided one-way analysis of variance was intraclass correlation coefficient (ICC) = 0.97 (assumed from a previous study [ 14 ]), three measurements (morning, noon, and evening) and a confidence interval range of 0.05 (lower limit = 0.938, upper limit = 0. 987).From this, the required number of cases was 19, and 23 was set to take into account 20% of the excluded cases.
Quantitative variables were described as mean and standard deviation or standard error, and qualitative variables as number and percentage.Differences between pre-and post-treatment groups were analyzed using the paired t -test.
The intra-rater reliability of CPF measurements was assessed by conducting CPF measurements three times before each meal (breakfast, lunch, and dinner) resulting in a total of nine measurements daily at pre-and post-treatment.The ICC was used for analysis.The effects of MG fatigability by measuring three times in one session and the effects of daily fluctuations by measuring three sessions a day were confirmed using repeated measures analysis of variance, respectively.
The validity of CPF measurements was evaluated from both criterion-related validity and convergent validity perspectives.Criterion validity testing compares measurements against an objective gold standard to assess whether indicators measure what they intend to, and can provide accurate evidence to inform programs [ 15 ].Criterion-related validity was assessed by examining the relationship between CPF values and various indicators obtained from pulmonary function tests and comprehensive MG evaluations using Pearson's correlation coefficient.Convergent validity is the demonstration of substantial and significant correlation between different instruments designed to assess a common construct [ 16 ].Convergent validity was evaluated by examining the relationship between CPF values and the item of swallowing function of QMG score using Pearson's correlation coefficient.In addition, the CPF of the group with aspiration and without aspiration were compared using the Wilcoxon rank-sum test, with aspiration defined as 1 or more in the swallowing items of the QMG score.This definition of aspiration using the swallowing item of QMG score because of a correlation between QMG bulbar scores and the assessment of aspiration using videofluoroscopic examination [ 17 ].Moreover, the aspiration risk prediction was assessed by receiver operating characteristic curve analysis, with CPF as the independent variable and the presence of aspiration as the dependent variable.The area under the curve (AUC) was calculated for the CPF receiver operating characteristic curve.All analyses were performed using SAS version 26.0 (IBM, Armonk, NY, USA), and statistical significance was defined as a p-value of < 0.05.

Patients characteristics
The study comprised 26 inpatients with MG who required immunotherapy.Demographic and clinical characteristics of the study patients are displayed in Table 1 .There were no patients with a history of asthma.

Reliability of the CPF
Comparison between pre-and post-treatment groups based on voluntary cough, spirometric values, and MG comprehensive assessments showed that all parameters were significantly different between pre-and post-treatment ( Table 2 ).There were no significant differences in the number of CPF measurements and time of day before and after treatment, and the ICC was more than 0.95 for all items ( Table 3 ).

Criterion-related validity of CPF
The relationship between CPF and spirometric values showed positive correlations in almost all spirometric values on pre-and post-treatment ( Table 4 ).There was no correlation between CPF and the MG comprehensive assessments; however, the difference in CPF or QMG scores between pre-treatment and post-treatment was denoted with , and CPF and QMG scores were found to be positively correlated ( r = 0.51, p-value = 0.009).

Convergent validity of CPF
When examining the relationships between CPF and each of the QMG items, the swallowing and percent predicted FVC items showed correlation at pre-treatment ( Table 4 ).The QMG swallowing item defined aspiration as more than 1, with 16 patients with aspiration and 10 patients without aspiration the CPF were significantly lower in patients with aspiration than in patients without aspiration [aspiration = 196.7 (110 -308) L/min, without aspiration = 264.4(190 -345.6)L/min, median (range), p-value = 0.004].Moreover, the receiver operating characteristic curve was used to calculate the CPF cut-off value of 205 L/min with a sensitivity of 0.77 (95% confidence interval = 0.53-0.91),specificity of 0.90 (95% confidence interval = 0.60-0.98),and AUC of 0.85 (95% confidence interval = 0.70-1.00)( Fig. 2 ).

Discussion
This study validated the reliability and validity of CPF measurements in assessing voluntary cough in patients with MG.Based on the demographic data ( Table 2 ), there were significant differences in both pulmonary function tests and the MG comprehensive assessments between pre-treatment and posttreatment.This indicates that immunotherapy treatment improved MG symptoms and respiratory function.The reliability of CPF in patients with MG was exceptionally high.The results were also positive in terms of criterion-related validity, where CPF was associated with respiratory function and MG symptoms, and convergent validity, where CPF was associated with aspiration.
The most significant result of this study was the high intrarater reliability obtained when patients themselves performed CPF measurements at both pre-treatment and post-treatment, with an ICC of 0.95 or higher.This suggests measuring CPF with the   peak flow meter can effectively induce their appropriate selfmanagement and maximum cough strength even at the bedside.
The characteristic symptoms of MG include easy fatiguability and daily fluctuations in weakness, and it is important to consider whether these symptoms of MG affect the measurement of CPF.Specifically, there were no significant differences in the three CPF measurements within a session as well as in CPF measurements between the sessions before breakfast, lunch, and dinner.These findings suggest that the effects of fatiguability and daily fluctuations were minimal in the measurement of CPF in patients with MG.The reasons for this are twofold.Firstly, the measurement protocol took into account the patient's susceptibility to fatigue, as it consisted of three measurements with 30-second rests at each measurement interval.This ensured a more accurate assessment of their CPF.Secondly, regarding daily fluctuations, the measurements were conducted during the patient's hospital stay when they were in a relatively low activity state.This might minimize the impact of daily variations on the CPF measurement.
For criterion-related validity, CPF measurements in this study were associated with pulmonary function tests and MG comprehensive assessment which are both the gold standard for MG assessment [ 19 ].Almost all spirometric values showed a correlation with CPF at pre-treatment and post-treatment.Previous studies have also reported that CPF is associated with FVC and FEV 1 in Duchenne muscular dystrophy [ 18 ].This indicates that CPF measurements can reflect respiratory function in MG patients as well as in Duchenne muscular dystrophy patients.
In relation to the MG comprehensive assessment, there was no statistically significant correlation between QMG and CPF.However, a relationship between the degree of improvement in QMG and CPF measurements after treatment (i.e.QMG and CPF) was identified.This might suggest that CPF measurement was a biomarker reflecting the variability of MG symptoms.
In demonstrating the convergent validity of CPF measurements in MG, we focused on the association with aspiration in view of the risk of MG crisis.Results showed a correlation between CPF and the item of swallowing function in QMG at the time of pre-treatment with worsening MG symptoms.This association between CPF and swallowing could be attributed to the role of the coughing mechanism.Coughing consists of three phases: inspiratory, compressive, and expiratory phases, with glottis closure affecting the compressive phase [ 5 ].This closure of the glottis is also an important movement for smooth swallowing [ 20 ], and thus may be associated with coughing and swallowing function.
In the present study, the risk of aspiration in CPF values was calculated to have a CPF cut-off value of 205 L/min with a sensitivity of 0.77, specificity of 0.90, and AUC of 0.85.In a previous study in amyotrophic lateral sclerosis (ALS) patients [ 21 ], the cut-off value was 238 L/min, which is comparable to our calculated cut-off value, and with similar discrimination accuracy (sensitivity of 0.73, specificity 0.78, and AUC 0.78).The definition of unsafe swallowers in the study of ALS was the penetrationaspiration scale of 3 or more, although the score of 6 or more typically corresponds to aspiration [ 22 ].In contrast, the present study defined aspiration risk as QMG swallowing of 1 or more, which was different from the study of ALS; therefore, it is difficult to totally compare their results.However, the results of the present study may reflect a higher risk of aspiration due to the lower CPF cut-off value and higher specificity of the CPF compared to the ALS study.These findings support the convergent validity of the CPF measurements in MG.Moreover, QMG swallowing is an actual water test, which involves the risk of aspiration, whereas CPF is a voluntary cough test, which safely measures aspiration risk without the risk of aspiration; thus, the self-monitoring of MG symptoms is considered a strong point of CPF measurements.Therefore, we suggest that CPF measurement could be a useful and safe tool for detecting aspiration risk in MG patients.
Finally, while various devices have been used for CPF measurements in previous studies, recent research has focused on objective data, utilizing devices such as spirometers [ 23 ], pneumotachographs [ 24 ], and digital peak flow meters [ 25 ].However, the peak flow meter used in the present study offers the advantages of being inexpensive and user-friendly.Moreover, the findings of this study indicate that even when performed by the patients themselves, the measurements yielded highly reliable results, suggesting its effectiveness as a self-monitoring tool for MG symptoms.While this study was conducted on inpatients, we are contemplating the future possibility of utilizing home monitoring as a tool to promptly detect acute exacerbations of MG.
There are three main limitations in this study.Firstly, this study was a retrospective cohort study at a single institution.The nature of the study may have led to potential biases such as selection bias and measurement bias.Further multicenter prospective cohort studies of patients with MG will provide more conclusive information regarding the longitudinal changes in CPF in patients with MG.Secondly, there was a bias toward autoantibodies of MG in this study.Autoantibodies of the patients were mainly against the acetylcholine receptor antibodies, and there were only 4 patients with MuSK.In particular, MuSK antibodies may affect CPF measurements because the main symptoms of MuSK antibodies are bulbar and respiratory symptoms [ 26 ].Therefore, further subgroup analysis by autoantibody will be required in the future to identify the effect on CPF measurements.Finally, the effect of intra-day variability on CPF may have been minimal, as the patients were less active during the inpatient CPF measurements than in their home life.However, as symptoms may fluctuate in home life due to high activity levels, an outpatient cohort will be required in the future.

Conclusion
In conclusion, this study revealed that the measurement of CPF in MG patients is a highly reliable method that can be measured by the patients themselves.In addition, CPF was found to be a criterion-relevant validity measure, with results reflecting the respiratory function of MG patients and MG symptoms.Furthermore, the results of an increased risk of aspiration at CPF < 205 L/min supported the convergent validity of CPF in MG.Therefore, we encourage CPF measurement in MG patients as a tool for self-monitoring of MG symptoms and detecting aspiration risk.

Fig. 1 .
Fig. 1.Cough peak flow measurement.The patients were positioned sitting, with the measuring device tightly pressed against their faces to prevent air leakage.

Fig. 2 .
Fig.2.The sensitivity and specificity of the risk of aspiration in cough peak flow values at pre-treatment.The cut-off value of 205 L/min showed a high discrimination power with sensitivity of 0.75 and specificity of 0.90.

Table 1
Demographic and clinical characteristics of the study patients.
MG, myasthenia gravis; MGFA, Myasthenia Gravis Foundation of America; AChR, acetylcholine receptor; MuSK, muscle-specific tyrosine kinase.Values are presented as mean (range) and as number (percentage) of patients.

Table 2
Comparison between pre-and post-treatment groups based on voluntary cough, spirometric values, and MG comprehensive assessments.

Table 3
Reliability of measurements through comparison between the three cough peak flow measurements within a session and between the three session times in pre-and post-treatment.

Table 4
Correlation between the cough peak flow values and spirometric values and MG comprehensive assessments of the pre-and post-treatment groups.