Published ahead of print on April 5, 2007, doi:10.1164/rccm.200612-1841OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200612-1841OC
Predictors of Ineffective Cough during a Chest Infection in Patients with Stable Amyotrophic Lateral Sclerosis1 Respiratory Care Unit and 2 Department of Respiratory Medicine, Hospital Clínico Universitario, Universitat de Valencia, Valencia, Spain Correspondence and requests for reprints should be addressed to Emilio Servera, M.D., Respiratory Medicine Department, Hospital Clínico Universitario, Avda. Blasco Ibáñez 17, E46010 Valencia, Spain. E-mail: emilio.servera{at}uv.es
Rationale: In patients with neuromuscular diseases, a chest infection is associated with a reduction in respiratory muscle function that may result in decreased cough effectiveness. Objectives: To determine if a clinical or functional parameter in patients with amyotrophic lateral sclerosis (ALS) in a stable condition could predict spontaneous cough ineffectiveness during a respiratory tract infection. Methods: Forty consecutive patients with ALS referred to our Respiratory Care Unit were studied during a one-year follow-up. Measurements and Main Results: FEV1, FVC, FEV1/FVC, peak cough flow (PCF), peak velocity time (PVT), maximum inspiratory and expiratory pressures, and bulbar dysfunction evaluation using the Norris scale bulbar subscore (NBS). A total of 26 patients (65%) had spontaneous cough ineffectiveness during a respiratory tract infection. The best variables to predict nonassisted cough during a respiratory tract infection were NBS (p < 0.01) with a cutoff point of 29 (sensitivity, 0.89; specificity, 0.90; positive predicted value, 0.88; negative predictive value, 0.87), PCF (p < 0.001) with a cutoff point of 4.25 L/s (sensitivity, 0.74; specificity, 0.85; positive predictive value, 0.71; negative predictive value, 0.85), and PCF/PVT (p < 0.001) with a cutoff point of 28.88 L/s2 (sensitivity, 0.77; specificity, 0.96; positive predictive value, 0.91; negative predictive value, 0.89). Conclusions: In patients with stable ALS, bulbar dysfunction (NBS < 29), PCF (< 4.25 L/s), and PCF/PVT (< 28.88 L/s2) could predict the risk of ineffective spontaneous cough during a respiratory tract infection.
Key Words: amyotrophic lateral sclerosis • assisted cough • neuromuscular disease • noninvasive ventilation • peak cough flow
Respiratory problems are the principal cause of morbidity and mortality in patients with amyotrophic lateral sclerosis (ALS) (1, 2). Respiratory muscle weakness occurs progressively in these patients as a consequence of the natural course of the disease, and also in an acute form during respiratory tract infections (2). In this potentially life-threatening situation, the force generated by the respiratory muscles suddenly decreases, which can make them ineffective both for maintaining adequate alveolar ventilation and for producing an effective cough. Solving these problems is essential in clinical practice (3). To optimize management of respiratory problems in patients with ALS, an accurate and easily measurable predictor of cough ineffectiveness during a respiratory tract infection, available during periods of medical stability, would be a highly relevant aid in several aspects of clinical practice (2). It would make it possible to recognize the point at which patients should be referred to those responsible for initiating training in different assisted-cough techniques (3). Likewise, patients could be given information on deciding about advance directives and, lastly, it would make it easier for physicians to decide on the appropriate management in acute illness. However, to date, no clinical or functional data provide this predictive information accurately enough to be included in clinical guidelines. It is well known that poor bulbar function interferes significantly with cough effectiveness (4–6), but there is no cutoff point that would make it possible to pinpoint when bulbar impairment would render spontaneous coughing ineffective. Likewise, several functional respiratory measurements are correlated with cough capacity (7, 8), but they have not been studied as potentially useful tools for identifying coughing failure. Only peak cough flow (PCF) measurement, which is largely dependent on the intensity of the cough (9), has been proposed specifically for this purpose (10, 11). The studies that have been performed so far present a highly interesting research line, but they still do not report sufficient reliable data to provide accurate PCF-predictive values. As such, the aim of this study was to determine if a clinical or functional parameter could predict, in patients with medically stable ALS, the failure of spontaneous cough capacity for mucus removal in an acute respiratory tract infection.
Subjects We performed a 1-year prospective study between May 2004 and May 2005 that included 53 consecutive patients with ALS who were referred to the respiratory medicine department at our tertiary hospital. Our respiratory care unit is the reference unit for the management of respiratory problems of neuromuscular patients in the community of Valencia (4,772,000 inhabitants), Spain, and patients with ALS are referred to our unit from the neurology department of our hospital and from the neurology units of other hospitals in the community. The hospital's neurology department sends patients with ALS directly to our unit once the diagnosis of the disease has been confirmed. It may take longer for patients to arrive from other hospitals, but the aim is that they be included in our follow-up program before respiratory problems start. Informed consent was obtained from each subject who took part in the study. The study protocol was approved by the hospital's ethics committee.
Clinical Assessment Bulbar involvement was defined as the presence of dysphagia, dysarthria, drooling, and/or alterations in videofluoroscopy (13), and was assessed in accordance with the Norris scale bulbar subscore (NBS) (14), which ranges from 39 (normal state) to 0 (total dysfunction) and is included in the online supplement. The self-perceived cough capacity was evaluated by asking all subjects one or more of the following questions: "Do you think you would be able to expectorate if you had a chest infection?", "Do you think you can cough up mucus when you have a cold?", or "When you need to cough up mucus, do you think you will be able to do so?" The fact of the matter is that all the patients understood these questions.
Functional Assessment
Follow-up Protocol An acute respiratory tract infection was defined as the presence of fever and one of the following: presence or increase in intrathoracic respiratory secretions, cough, shortness of breath, or abnormal chest sounds. Criteria for hospitalization were respiratory failure, need for continuous, noninvasive, mechanical ventilation, intensive mechanically assisted coughing (more than four sessions per hour), and failure to remove mucus with assisted coughing techniques.
Statistical Analysis
Throughout the follow-up, 40 of the 53 patients with ALS included in the study had an episode of acute respiratory tract infection. The patients' age, sex, and pulmonary function are shown in Table 1. In no case had more than 3 months gone by between the functional assessment made for the study and an acute episode. A total of 29 patients (43.9%) had bulbar dysfunction (NBS, 18.9 ± 6.9).
A total of 26 patients (65%) needed assisted coughing techniques for thoracic mucus removal due to ineffective spontaneous cough during acute respiratory tract infection episodes (Table 1): of these 26, 23 (88.5%) had bulbar dysfunction (NBS, 20.9 ± 9.6). Statistical differences (p < 0.001) were found between ineffective and effective spontaneous coughing groups in all the parameters except for age, FEV1/FVC, PVT, and body mass index (Table 1; Figure 1). Manually assisted coughing was used initially, but when this technique proved ineffective, mechanically assisted coughing was applied. Thus, manually assisted coughing techniques were used for 8 patients and mechanically assisted coughing for 17 (Table 2). No infiltrates or atelectasis were found in chest X-rays of any patients.
Among the patients with bulbar dysfunction and ineffective spontaneous cough, 14 (NBS, 16.5 ± 6.0) also complained of having oropharyngeal secretions, which were managed by suctioning when discomfort made it appropriate. All these patients easily distinguished upper secretions from lower secretions. These supraglottic secretions were already there before the acute episode and had been managed at home in a conventional way (2). None of the patients went to hospital because of such secretions, nor were any hospitalized because of them. Of the patients who required assistance to cough, four needed hospitalization for continuous, noninvasive mechanical ventilation and intensive mechanically assisted coughing (Table 3). In three cases, all with severe bulbar dysfunction (NBS, 9.33 ± 3.05), a tracheostomy was performed because of ineffective noninvasive mechanical ventilation due to failure of mechanically assisted coughing (FVC, 0.8 ± 0.1 L; FVC%, 36.7 ± 3.2; PImax, –24.0 ± 10.6 cm H2O; PEmax, 33.0 ± 10.5 cm H2O; PCF, 1.95 ± 0.80 L/s; PVT, 0.27 ± 0.18 s; PCF/PVT, 10.47 ± 7.23 L/s2). No patients with home management experienced complications.
Regarding subjective, self-perceived cough capacity, when asked in a stable condition, 7 of the subjects included believed that they would be able to expectorate, 19 believed that they would not be able to, and 14 could not answer the question, mainly because they had not had an acute episode during the previous year. During an acute episode, we found that none of the seven subjects who believed they would be able to expectorate at the interview needed assistance. Of the 19 who believed that they would not be able to expectorate, 2 had an effective cough, but 17 did not. Of the 14 who could not give their opinion, 9 needed help and 5 did not. In the statistical data analysis to establish the variables that are predictive of an ineffective cough, only those variables that exhibited a significant association in a previous univariate analysis were included in the multivariable model (Table 4). The forward stepwise logistic regression analysis shows that PCF (p < 0.001; odds ratio [OR], 0.49; 95% confidence interval [CI], 0.31–0.79), PCF/PVT (p < 0.001; OR, 0.90; 95% CI, 0.82–0.98), and NBS (p < 0.01; OR, 0.89; 95% CI, 0.81–0.98) are the variables that most accurately predict the patients who would have an ineffective unassisted cough during an acute chest episode. The patients' subjective evaluation of their cough capacity effectiveness, given at the initiation of the study, did not prove to be a predictive factor of ineffective cough during acute episodes. However, if we exclude those patients who did not know how to answer about their cough effort effectiveness (n = 26), this subjective evaluation becomes a predictive factor (p < 0.001; OR, 0.15; 95% CI, 0.77–1.02).
Regarding the ROC curve analysis (Figure 2), to predict in which patients spontaneous cough would be ineffective during an acute respiratory tract infection, the variables with higher area under the curve (AUC) were NBS (AUC, 0.81; 95% CI, 0.67–0.96), PCF (AUC, 0.82; 95% CI, 0.67–0.97), and PCF/PVT (AUC, 0.87; 95% CI, 0.73–1.02). A cutoff point of 29 for NBS (sensitivity, 0.89; specificity, 0.90; positive predicted value, 0.88; negative predictive value, 0.87), 4.25 L/second for PCF (sensitivity, 0.74; specificity, 0.85; positive predictive value, 0.71; negative predictive value, 0.85) and 28.88 L/second2 for PCF/PVT (sensitivity, 0.77; specificity, 0.96; positive predictive value, 0.91; negative predictive value, 0.89) were the best predictors to identify patients in a stable condition whose unassisted cough would become ineffective during an acute respiratory tract infection.
The findings in this study show that spontaneous cough ineffectiveness during respiratory tract infections of patients with medically stable ALS can be predicted satisfactorily by means of PCF and PCF/PVT measurements, which are the best functional indicators of spontaneous cough effectiveness during an acute respiratory tract infection. Moreover, we have found that those patients with greater bulbar dysfunction are more likely to have an ineffective spontaneous cough during these episodes. Likewise, cough capacity perception itself can be a good predictor of effectiveness, but the fact that many patients in our study were unable to provide an opinion makes this information less useful in clinical practice. Previous studies have reported a reduction in respiratory muscle strength during a respiratory tract infection, both in normal subjects (17) and in neuromuscular patients (18), which, in the latter group, may result in a decrease in cough effectiveness (11). This reduction in respiratory muscle strength can cause an effective cough capacity at baseline to become ineffective during an acute respiratory episode, and change a banal chest cold into a life-threatening situation (1, 2, 19). During an acute respiratory infection, mucus production increases, whereas mucociliary clearance decreases due to inflammation; under these circumstances, coughing is essential to remove respiratory secretions (20). If cough capacity is ineffective, the retained secretions can induce atelectasis and pneumonia, bringing about ventilation/perfusion mismatching, which worsens hypoxemia (21). If this breathing overload is not solved, respiratory failure may make endotracheal intubation necessary to apply mechanical ventilation and suction out respiratory secretions and, thus, maintain life (20). However, the use of appropriate assisted-cough techniques (manual and mechanical) to remove mucus, and the use of noninvasive mechanical ventilation to assist inspiratory muscles, can prevent this outcome (19). In this situation, secretion management is a major factor in the success of noninvasive mechanical ventilation (2, 22). As previous studies have found (23), and as the results of our study show, when assistance is applied correctly, most patients with ineffective spontaneous cough can be managed at home. This means that patients and their caregivers must receive previous training to carry out these techniques (3, 11); however, the quality and efficiency of the medical practice make it necessary to initiate such training, if needed, and to program the intensity according to patients' needs. In fact, for some patients who still have good muscle function and who cough more effectively when unaided, assisted coughing techniques are inappropriate (4). Consequently, a tool that would make it possible to assess the specific needs of each patient would be of great use in clinical practice. The results obtained in this study help to establish, for patients in a stable phase, the point at which the risk of their cough becoming ineffective during an acute bronchial infection makes initiating training of assisted coughing procedures advisable. Measurement of PCF correlates with the intensity of the cough (9), and it is readily available for patients with ALS: it does not require sophisticated equipment, it is easy and quick to determine, and it does not make the patient suffer. Moreover, when values have not yet decreased to below 4.4 L/s, PCF can be measured using a portable device used to measure peak expiratory flow by patients with asthma (24). The findings of this study confirm those based on clinical observations (3, 11): values of baseline PCF below 4.25 L/s in a stable condition suggest an ineffective cough during acute episodes, when the use of assisted coughing for airway secretions removal becomes necessary (11). Therefore, in this group of risk patients, it is mandatory to teach assisted coughing techniques (manual and mechanical) and to check regularly that these maneuvers are effective (23, 25). Cough capacity does not only depend on chest wall muscle performance (26). In patients with neuromuscular diseases, impaired function of the upper airway both decreases the generated pressure into the airway during the expulsive phase and lengthens the time needed for the cough flow to reach its maximal point (PVT) (27). Acceleration of PCF (PCF/PVT) correlates with the explosiveness of the cough (28), and our findings show that PCF/PVT measured when the patients are in a stable state is another good predictor of cough failure during an acute chest infection in patients with ALS. Acceleration values of PCF at baseline lower than 28.88 L/s2 suggest an ineffective unassisted cough during an acute episode. Bulbar involvement is one of the most negative milestones in the course of ALS. It does not only cause important speech and swallowing impairment (the latter is sometimes underestimated), but it also diminishes the effectiveness of spontaneous cough and severely interferes with noninvasive procedures used for management of respiratory problems (2–6, 13, 29). Our results add specific data to the general knowledge, as they show that the NBS, a well-known clinical scale used to measure bulbar impairment (14), offers predictive data regarding failure to cough effectively during acute episodes that are accurate enough to be useful in clinical practice. On the other hand, the value of the difference between the MIC and FVC, which Bach and colleagues (5, 11, 25) proposed as a useful indicator of bulbar dysfunction in neuromuscular patients, did not prove to be an effective predictor during an acute chest episode in our study. The fact that some patients with no, or minimal, respiratory muscle involvement and optimum glottic function (and, therefore, with an effective cough) registered an equally slight difference between MIC and FVC values as those with severe bulbar involvement (and with an ineffective cough) could explain our finding. It is sometimes not possible to carry out either an adequate respiratory function assessment or a quantitative evaluation of bulbar involvement. If patients perceive that their cough is not effective, it is very possible that their appraisal is correct. Consequently, as in the case of patients with impaired PCF, PCF/PVT, or NBS measurements, they should be informed about their situation and asked about their advance planning, with a view to initiating training if this is what they require. In addition, the hospital staff should receive this information during exacerbations, so that they can provide appropriate respiratory secretions management or palliative care (depending on the patient's previous directives) as soon as possible. In conclusion, for patients with ALS in a stable clinical situation, bulbar dysfunction measured by the NBS, PCF, and acceleration of PCF measurements could predict spontaneous cough ineffectiveness during an acute respiratory tract infection. These findings may make it easier to decide when to commence training of patients and caregivers in assisted-coughing techniques in clinical practice.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200612-1841OC on April 5, 2007 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form December 19, 2006; accepted in final form April 5, 2007
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
