Dyspnea Is a Specific Symptom in Parkinson’s Disease

INTRODUCTION

Non-motor symptoms are frequent in Parkinson’s disease (PD). Among them, dyspnea is reported between 11.5% and 40% of the patients [1–3]. Even if other causes are excluded, it seems to be a symptom due to the neurodegenerative process [4]. As dyspnea is disabling for elderly and in PD [2, 5] and perceived as stressful, a better clinical assessment could improve the management of dyspneic PD patients. Furthermore, a controversy still exists regarding the existence and the pathophysiology of respiratory disturbance in PD [6]. Determining the clinical features of dyspnea in PD could help us to better understand the underlying mechanisms in this disease. On one side, shortness of breath could be a somatic expression of anxiety [7]. But on the other side, MRI data and clinical observation tends to classify dyspnea and ventilatory dysfunction as a dysautonomic symptom [8–10]. Four main contributors to ventilatory dysfunction have been identified: an obstructive element, a restrictive component, potential drug effects (i.e., levodopa induced diaphragmatic dyskinesias), and an abnormal central control of ventilation.

Dyspnea is a complex and multidimensional sensation and it needs specific tools to be screened properly. Classically, shortness of breath is mainly assessed through the disability associated with breathlessness by using the mMRC scale (from 0 (no disability) to 4 (complete disability) [11]; however, this scale only questions the patient about the functional impairment due to dyspnea. Since physical autonomy is disabled in neurological diseases like PD (due to akinesia, hypertonia or gait), other questionnaires are needed in order to screen the symptom and not only its consequences when the patient makes effort. Recently, the multidimensional dyspnea profile (MDP) questionnaire has been developed to distinguish sensory and emotional aspects of the dyspnea [12] (more details in Supplementary Material 2). The best sensory quality (which means the quality that most accurately describes the sensation - SQ) and the severity of the emotional response domain may differ between airway, lung or neuromuscular diseases [13, 14].

The aim of our study was to assess the clinical features of dyspnea using the MDP questionnaire. Our hypothesis was that dyspnea in PD is an individualized symptom with two distinct aspects: one strongly linked with emotional response domain and particularly anxiety and another one associated with ventilation control impairment due to dysautonomia.

METHODS

Between October 2016 and March 2017, in a single-center study (movement disorders clinic), consecutive PD patients were asked whether they experienced dyspnea. The inclusion criteria were i) physician-diagnosed PD (according to the Movement Disorders Society Clinical Diagnostic Criteria for PD (2015)) ii) age between 18 and 85, and iii) experience of dyspnea in the last month. The exclusion criteria were i) atypical parkinsonism, ii) cognitive impairment (as evidenced by a Montreal Cognitive Assessment (MoCA) score <24 out of 30 [15]), iii) a pulmonary, cardiac or ENT pathology (affecting the upper airways) that may lead to dyspnea, iv) intercurrent pulmonary, cardiac or ENT infection in the last month, and (v) abnormal cardiac and/or pulmonary results in a clinical examination. An ethical standards committee approved the procedures.

Patients were asked the two following questions: “In the last month, have you experienced breathlessness” and “In the last month, did you feel discomfort while breathing?” Dyspnea was diagnosed if the answer to at least one question was “yes”. Respiratory disability was assessed using the mMRC scale: zero for dyspnea only with strenuous exercise; 1 for dyspnea when hurrying or walking up a slight hill; 2 if the patient walks slower than people of the same age because of dyspnea or has to stop for breath when walking at own; 3 if the patient stops for breath after walking 100 yards (91 m) or after a few minutes; 4 if the patient is too dyspneic to leave house or breathless when dressing [10]. Patients with fluctuations (n = 52) were also asked to determine if the symptom was associated with motor fluctuations (if dyspnea was more predominant during self-reported “on drug” or “off drug” conditions, or “undetermined if the patients was not able to precise this point) and the clinical features of dyspnea were assessed by the MDP. In this study, the patients were instructed to focus on their worst breathing experience in the preceding 4 weeks. First, the patients were asked to rate unpleasantness of breathing sensations (A1 – numerical rating scale - maximum 10). Secondly, they are asked to determine if sensory qualificators apply or not to describe the dyspnea (SQ – 5 different sensations – Table 1) and then, which of the 5 sensations most accurately applied. Thirdly, the patients are asked to rate each of the sensory qualificators that apply (numerical rating scale from 0 to 10 for each item – maximum 50). The immediate perception response score is equal to the sum of A1 plus the numerical rating of each sensory qualificator. Lastly, they are asked to rate each of 5 breathing-related emotions (numerical rating scale from 0 to 10 for each item – maximum 50). The emotional response domain (A2) is the sum of the numerical rating of each emotion.

The neurological examination included the Movement Disorders Society Unified-Parkinson’s Disease Rating Scale (MDS-UPDRS [16]), the Hoehn and Yahr score (H&Y [17]), the Hospital Anxiety and Depression scale (HAD [18]), and the 8-item Parkinson’s Disease Questionnaire (to assess quality of life (PDQ8) [19]).

The quantitative variables were described using the mean and the standard deviation or the median and the interquartile range (IQR). The normality of the distributions was verified with the help of graphs and the Shapiro-Wilk test. The qualitative variables were described using frequencies and percentages. Relationship between variables has been appreciated with the help of Spearman’s correlation coefficients. The Hotelling– Williams procedure tests whether or not two non-independent correlation coefficients are significantly different. The threshold of significance was 0.05. The analyses were performed using SAS software version 9.4 (SAS Institute, Cary NC, USA).

RESULTS

Among the 163 patients of our cohort (flow chart in Supplementary Material 1), sixty patients (39.2% - more details in [3]) reported dyspnea (age = 64.9±8.3 years, sex ratio (M/F) = 31/29, disease duration = 12.0 years (8.0; 16.0) (Table 1). The duration of dyspnea was 4.6±2.4 years with much more occurrence of the dyspnea during wearing-off times (48.3%) than during biphasic and peak-dose dyskinesia (6.7%). The mean degree of disability associated with dyspnea assessed by the mMRC scale was 1.3±0.8.

MDP data are in Table 1. The most frequent SQ described by the patients were: hyperpnoea (35%), physical breathing effort (25%) and air hunger (20%). Hyperpnoea and air hunger were the SQ with the highest intensity (respectively 3.7±2.5 and 3.3±3.4). Regarding the emotional response domain (A2), anxiety had the highest intensity (4.0±3.3).

The intensity of the global SQ was correlated with the MDS-UPDRS I score (Non-motor Experiences of Daily Living - p = 0.002; r = 0.39), the MDS-UPDRS II score (Motor Experiences of Daily Living – p = 0.004; r = 0.36), the MDS-UPDRS III (Motor Examination – p = 0.03; r = 0.29), the Hoehn & Yahr score (global motor severity of the PD – p = 0.03; r = 0.29), the HAD anxiety score (p = 0.02; r = 0.3), the HAD depression score (p = 0.03; r = 0.28) and the quality of life (PDQ8 - p = 0.004; r = 0.37).

The intensity of emotions (A2) was correlated with the MDS-UPDRS I (Non-motor Experiences of Daily Living -p<0.001; r = 0.49), the MDS-UPDRS II (Motor experiences of daily life – p = 0.03; r = 0.28), the HAD anxiety (p < 0.001; r = 0.49), the HAD depression (p = 0.03; r = 0.28) and the quality of life (PDQ8 - p = <0.001; r = 0.42). HAD anxiety score had a stronger statistical correlation with A2 (intensity of emotional response domain) than with the immediate perception response score (0.3 vs 0.49, p = 0.03).

DISCUSSION

In our study, dyspnea seems to be a specific symptom with variable characteristics. Indeed, hyperpnoea was the best sensory quality for 35% of the participants and in a cohort of patients with amyotrophic lateral sclerosis, “air hunger” was the best sensory quality for almost half of the patients [14]. In chronic obstructive pulmonary disease, “air hunger” (27%), followed by “breathing a lot” (19%) were the most chosen qualities [13]. Therefore, “parkinsonian dyspnea” seems to have a specific profile, different to that observed either in lung or in neuromuscular diseases.

With the different sensory qualities, MDP can also help us to better understand the pathophysiology of dyspnea in PD. Hyperpnoea may be the consequence of a muscles command dysfunction due to hypertonia and lack of coordination. Some authors highlighted an abnormal ventilatory control in PD patients [20] and further investigations are needed to determine if dyspnea is associated with autonomic and/or cognitive dysfunction (as attention deficit or hallucinations) in the disease. Regarding the emotional response scores, in our cohort, dyspnea was associated with anxious and depressive feelings. These results are not specific of PD. Yet, fear and frustration are not described with such intensity in lung or neuromuscular diseases [13, 14]. It may reflect the difficulty for PD patients to control properly the respiratory muscles, unlike in asthma for example [21]. Only a study assessing MDP and pulmonary function tests in PD patients could confirm this hypothesis.

Our study suffers from several limitations. Firstly, we did not perform pulmonary function testing in our study. These data might help to understand better the pathophysiological mechanisms underlying dyspnea. By assessing dyspnea and PFT in a prospective cohort, correlations with obstructive/restrictive patterns could be done. Then, we could determine if the respiratory sensations are associated with an objective lung dysfunction and are related with lack of ventilatory control (i.e., dysautonomia) or with anxiety, as part of non-motor fluctuations. Secondly our cohort was recruited from a single center (a tertiary hospital). Nevertheless, the mean age, disease duration and non-motor scores were similar to those published for other cohorts [22]. A multicenter international cohort could avoid this bias. Lastly, we cannot rule out self-reporting bias among the patients. Some participants might have not perceived their dyspnea because of a disease-related physical limitation or might even have mistaken it for a motor impairment. We did not restrict the definition of dyspnea to respiratory disability since we screened it with an open-ended question.

CONFLICT OF INTEREST

GB, TP, FM, KD and CC have no disclosure. DD has received PHRC grants from the French Ministry of Health and research funding from the ARSLA charity, France Parkinson charity and European Commission (H2020). He served on advisory boards, served as a consultant and given lectures for pharmaceutical companies such as Apopharma, Orkyn, Air Liquide, Aguettant, Everpharma, Boston Scientific. LD has received grants from France Parkinson charity and served on advisory boards for Orkyn, Zambon and Abbvie and gave a lecture (honoraria) for UCB. CM has served as CSO InBrain Pharma and Scientific advisor for Apopharma, Orkyn/ Air Liquide and Boston Scientific.