Study design

The current study is a prospective cross-sectional observational study including patients with COPD entering PR. The medical ethical committee of the Maastricht University Medical Centre, Maastricht, the Netherlands concluded that while all tests, except a structured pain interview, were done as part of the clinical routine, the study did not fall under the Medical Research Involving Human Subjects Act and therefore did not require medical ethical approval. The local Institutional Review Board of CIRO approved the study.

Participants

Eligible participants were patients with a diagnosis of COPD according to the Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease (GOLD) [1] who were referred for PR to CIRO, a centre of expertise for chronic organ failure in Horn, the Netherlands [13]. Participants were excluded if they were unable to understand Dutch or had lung cancer as primary diagnosis. All participating patients gave written informed consent. Participants were recruited between September 2014 and November 2014.

Measurements

As part of the regular three-day assessment performed before the PR programme, the following parameters were recorded: demographics, past medical history (Charlson comorbidity index) [14], smoking history, body mass index (BMI) and use of medication. Severity of dyspnoea was assessed by the Modified Medical Research Council Dyspnoea Scale (mMRC) [15]. FEV₁ and Forced Vital Capacity (FVC) were measured using post-bronchodilator spirometry. Severity of airflow limitation was defined according to GOLD (1). Whole-body plethysmography was performed to measure Functional Residual Capacity (FRC), Residual Volume (RV), and Total Lung Capacity (TLC). Static hyperinflation was defined as transcendence above 120 % of the expected values from one or more of these three parameters. Single breath diffusion capacity for Carbon monoxide (TLCO) and Transfer Coefficient (KCO) were measured [16]. Patients also performed two six-minute walk tests and the test with the longest six-minute walk distance (6MWD) was used for further analysis [17]. Disease-specific health status was assessed using the COPD Assessment Test (CAT) (score ranging from 0 (best) to 40 points (worst)) [18]. Symptoms of anxiety and depression were recorded using the Hospital Anxiety and Depression Scale (HADS) [19]. The HADS consists of an anxiety subscale (HADS-A) and depression subscale (HADS-D), each ranging from 0 (best) to 21 points (worst).

In addition to the tests performed in the clinical routine, pain was assessed using a validated Dutch multidimensional, structured pain interview [20, 21]. This instrument includes items of the McGill Pain Questionnaire and the Brief Pain Inventory (BPI) [22, 23]. Prior to assessment of pain, participants were asked: ‘Are you generally bothered by pain?’ If they responded ‘yes’, pain interviews were completed. If they responded ‘no’, participants did not complete the pain interview. Participants were asked to mark the location of their pain using the body outline diagram of the BPI. Interview questions included the severity and character of pain and pain treatment. Participants were also asked to rate their pain on a numeric rating scale (NRS) ranging from 0 (no pain) to 10 (worst pain), at the time of answering the questionnaire, during the previous week, at its worst and at its least. One assessor (Y.L.P.) completed all pain interviews. Interviews took place on day one or two of the three-day assessment and lasted about 10–15 min.

Statistics

Data were analysed using SPSS for Mac version 21.0. All continuous data were tested for normality and are shown as mean (SD). Frequencies were used to describe the prevalence of pain generally and the prevalence of pain in each location. Demographic and clinical characteristics were compared between participants with and without thoracic pain. Chi square tests were used to compare categorical data between these two groups. Independent sample T-tests or Mann–Whitney U tests were used for continuous variables, according to the variable distribution. The correlations between the severity of thoracic pain and FEV₁, FRC, RV, and TLC as percentages of their predicted values, was determined using Pearson correlation coefficients and are shown in scatter plots. Because of multiple comparisons, a p-value ≤0.01 was considered statistically significant.

Patient characteristics

In total, 72 COPD patients were asked to participate in the study and 69 provided informed consent (response rate 95.8 %). Two patients had a FEV₁/FVC >0.7 and were therefore excluded. The final study population consisted of 67 participants.

Pain prevalence

Pain was reported by 55 participants (82.1 %). Thoracic pain was reported by 36 participants (53.7 %) (Fig. 1). Of these, 27 (75.0 %) only reported thoracic pain. Further common locations for pain were: neck , shoulder blades and lower back (Fig 1). A combination of thoracic pain and pain between shoulder blades was reported by 1 participant (1.5 %), while a combination of thoracic pain and lower back pain was reported by 7 participants (10.4 %). Participants described their thoracic pain mostly as pressing pain (51.4 %).

Characteristics of participants with or without thoracic pain

Relationship between thoracic pain and static hyperinflation

There were no differences in static lung volumes between the two groups. Association between severity of thoracic pain and lung function parameters is shown in Fig. 2. None of the correlations shown in either figure was found to be significant.

Pain prevalence

More than half of the participants entering PR reported thoracic pain. In addition neck, shoulder blade and lower back pain were also reported. The prevalence of thoracic pain was higher in the current study than was previously shown in outpatients with COPD attending a PR program, despite a comparable degree of airflow limitation [10] Borge et al. [24] also showed a high prevalence of pain in general, but a lower prevalence of thoracic pain than in the present study. This lower prevalence of thoracic pain may be explained by the fact that Borge et al. [24] included hospital outpatients, while the current study included patients referred for PR.

Characteristics of participants with or without thoracic pain

This study confirmed our hypothesis that thoracic pain in COPD patients is related to impaired disease-specific health status. Moreover, differences in CAT scores between patients with and without thoracic pain were more than twice the minimum clinical importance difference [18, 25, 26]. The qualitative study from Lohe et al. [27] also showed a significant impact of pain on quality of life among patients with COPD. In fact, pain caused some patients to be homebound or even bedbound.

We could not confirm our hypothesis that static hyperinflation is related with thoracic pain in COPD. In fact, this study showed no significant differences in static hyperinflation between patients with and without thoracic pain, and no correlations between severity of thoracic pain and static hyperinflation. Thoracic pain was also not explained by differences in severity of airflow limitation or diffusion capacity. Some authors suggested that comorbidities may be responsible for pain in COPD [7]. We did not find a relationship between the Charlson comorbidity score and the presence of thoracic pain. Moreover, the literature concerning the role of comorbidities in the presence of pain in COPD remains conflicting [7]. Therefore, mechanisms underlying thoracic pain in patients with COPD remain unclear.

In the present study, prevalence of thoracic pain was higher in younger participants. Habraken et al. [28] previously found in a qualitative study that older patients are less likely to report their symptoms and tend to adapt to their situations or attribute their symptoms to getting older.

The present study showed a trend towards female participants being more likely to report thoracic pain. Higher symptom burden among female participants with COPD has been reported in earlier studies. Indeed, female patients with COPD reported more severe dyspnoea, experienced more physical and mental limitations and had lower health status scores [29, 30]. Lamprecht et al. [31]. stated that female patients, even with the same degree of lung function impairment, more often report symptoms such as dyspnoea and coughing.

We did not find statistically significant differences in scores for anxiety and depression between patients with and without thoracic pain. This may be due to the limited sample size of our study. Indeed, participants with thoracic pain reported clinically relevant [32] higher scores HADS anxiety and HADS depression scores than participants without thoracic pain. Lohne et al. [27] previously described the relationship between psychological symptoms and thoracic pain as the ‘vicious COPD circle’, stating that incomprehensible pain could lead to sleep disorders, anxiety and depression, which then results in worsening of upper body pain.

Earlier studies showed a positive relationship between pain and dyspnoea severity [33, 34] Bentsen et al. [34] hypothesized earlier that airflow limitation increases respiratory muscle work resulting in fatigue, exhaustion and pain. In the current study we did not find a difference in mMRC scores between the two groups. This may be caused by the fact that the mMRC is a situational measure, which is limited responsive to change [35]. Future studies should consider including a NRS or Visual Analogue Scale (VAS) to measure the intensity of breathlessness. A recent study including 266,000 Medicare Managed Care recipients showed that dyspnea and pain commonly occur together [36]. In fact, participants with dyspnea had a considerably higher prevalence of pain, including thoracic pain. However, Clark et al. [36] found no evidence that underlying diseases, such as COPD were completely responsible for the co-occurrence of breathlessness and pain. Moreover, the authors hypothesized that physical deconditioning might be the explanation for the relationship between breathlessness and pain. Both symptoms could be the results of the same underlying process [36].

Treatment of pain in patients with COPD

Evidence now shows that pain, including thoracic pain, is highly prevalent among patients with COPD, and it has a major impact on disease-specific health status. As a consequence, systematic assessment and treatment of thoracic pain should be part of standard clinical care for patients with COPD to minimize symptom distress and optimize disease-specific health status. Nevertheless, only 39 % of the patients in our study were using pain medication, while 82 % of the patients reported pain. Moreover, worst reported pain scores were about 7 points, reflecting severe pain [37]. Under treatment of pain has been reported before among patients with COPD [5]. A possible explanation for pain under treatment is that patients with COPD may not actively express a wish for help for their symptoms, because they adapt to their situation or believe that there are no possibilities to improve their situation [28]. A second explanation for under treatment of pain in COPD might be the lack of knowledge concerning causal mechanisms of pain in COPD and therefore lack of causal treatment. In the light of the above described hypothesis of Clark et al. [36], it might be reasonable to assume that PR, which has been shown to improve physical functioning and decreases dyspnea severity [13], could also reduce pain in patients with COPD. If future studies could confirm this hypothesis, PR might be a non-pharmacological intervention to treat pain in patients with COPD.

Methodological considerations

This is the first study that specifically investigates the relationship between thoracic pain and multiple clinical characteristics in a well-characterized sample of COPD patients. Nevertheless, several limitations need to be considered in interpreting the results of the study. First, the present study only included a limited number of participants, which may result in underpowerment. Second, the present study did not include a control group without COPD. Therefore, it was not possible to verify whether thoracic pain symptoms are more common in patients with COPD than in non-COPD controls. Future studies should include a non-COPD control group to explore whether and to what extent thoracic pain is related with COPD. Third, only patients referred for PR were included which limits the generalizability of the results. Finally, this study was cross-sectional, whereas longitudinal daily recording of thoracic pain may be more reliable.

Ethics approval and consent to participate

The medical ethical committee of the Maastricht University Medical Centre, Maastricht, the Netherlands concluded that while all tests, except a structured pain interview, were done as part of the clinical routine, the study did not fall under the Medical Research Involving Human Subjects Act and therefore did not require medical ethical approval. The local Institutional Review Board of CIRO approved the study. All participating patients gave written informed consent.