High prevalence of shoulder girdle muscles with myofascial trigger points in patients with shoulder pain

This observational study was embedded in a clinical trial (registered at current controlled trials ISRCTN75722066) addressing a specific treatment of patients with shoulder pain [50]. The Committee of Human Research of the region Nijmegen-Arnhem, the Netherlands, has approved the study protocol [CMO 2007/22].

Study participants were recruited from patients participating in a controlled trial investigating the effectiveness of physical therapy on patients with unilateral, non-traumatic shoulder pain. This study was conducted at a primary care practice for physical therapy, which specializes in the treatment of patients with disorders of the shoulder, the neck, and upper extremities. A power analysis was performed prior to beginning this study, and it was calculated that 104 subjects were needed for the clinical trial.

All patients who contacted the practice for non-specific shoulder complaints from September 2007 until September 2009 were requested to participate in the study. The inclusion criteria were 1) age between 18 and 66 years; 2) unilateral non-traumatic shoulder pain; and 3) duration of symptoms of more than six months. Patients were excluded from the study if they presented with a prior diagnosis of shoulder instability, shoulder fractures, any systemic diseases, or a medical history or examination suggestive for the presence of neurological disease, internal diseases, or psychiatric disorders. All patients signed a written informed consent before participating in the study.

To determine the potential general applicability of this study to primary care shoulder pain patients, we searched for Dutch studies conducted on primary care patients from 1995 until 2009. Eight studies were found and sociodemographic data (age, gender, education level, and duration of shoulder pain) were analyzed and compared to the current study population [2, 5, 51‐55].

At baseline, age, gender, hand dominance, and education level were recorded. For comparison reasons we classified the education level as high education (university and higher vocational school), medium education (middle vocational school and higher or middle general secondary school), and low education (lower vocational school, lower general secondary school, primary school, or no education) [54]. Shoulder-pain related data (duration of shoulder-pain, recurrence rate and location of the complaints) were collected and the study subjects were asked to complete a set of standardized self-report measures, including the Disabilities of the Arm, Shoulder, and Hand outcome measure - Dutch Language Version (DASH-DLV), Visual Analogue Scale for Pain (VAS-P) and the Beck Depression Inventory- Second Version- Dutch Language Version (BDI-II-DLV) [50]. The BDI-II-DLV is used to discriminate between patients with major depression and those with only minor depressive feelings or no depression, which may be a confounding factor. The BDI-II has good predictive value, is widely accepted, and is commonly used in both clinical and experimental research. A BDI-II-DLV score equally or ≥ 21 indicates major depression (specificity 78.4%) [56].

For every study participant, one of the two available observers measured the passive range of motion (PROM) of the shoulder in flexion, internal and external rotation, abduction, and (horizontal or cross-body) adduction with a handheld digital inclinometer (The Saunders Group Inc, Chaska, MN). Range of motion was expressed in degrees and presented as the sum of the value measured for the non-affected shoulder minus the value measured for the affected shoulder. A positive value means that the affected shoulder had impaired range of motion as compared to the non-affected shoulder.

Next, the observer examined each subject for the presence of MTrPs in the shoulder muscles of their affected shoulder according to the guidelines outlined in Simons et al [26]; the non-affected shoulder was examined as a control. Following these guidelines, an MTrP is defined as: a nodule in a taut band that is extremely painful upon compression, and may produce referred pain or sensations. MTrPs were classified as either 'active' when the pain was recognized by the patient as a familiar pain, and 'latent' when the observer found a firm nodule in a taut band, which was painful on compression, but did not produce a recognizable pain. The inter-examiner reliability of trigger point palpation has been established in several studies [35, 57, 58]. All 17 muscles that are known to produce pain in the shoulder or may result in dysfunction of shoulder muscles were systematically examined and the number of muscles with MTrPs in the affected shoulder was counted, regardless of the number of MTrPs per muscle (Table 1). The two observers were physical therapists, each with 30 years of clinical experience in primary care practice. Both observers had attended an extensive, postgraduate course on MTrP diagnosis and therapy and had more than 5 years experience in identifying MTrPs and treating patients with MTrPs prior to the start of the study.

The DASH-DLV is a widely used multidimensional (physical, emotional and social) 30-item self-reporting questionnaire that focuses on physical function, pain and other symptoms. DASH-DLV scores ranges from 0 to 100, with higher scores indicating greater disability. DASH is a reliable and valid questionnaire, with good to excellent intra- and inter-rater reliability, and good correlation with the Shoulder Pain and Disability Index. Because of these advantages, DASH is considered to be one of the best questionnaires available for shoulder symptoms (http://​www.​dash.​iwh.​on.​ca/​) [59, 60].

The VAS-P is a self-report scale consisting of a 100 mm horizontal line anchored by word descriptions on each side [61]. VAS-P can be used to measure pain current pain levels (VAS-P1), the average pain over the last 7 days (VAS-P2), and the most severe pain over the last 7 (VAS-P3)). VAS-P scores ranges from 0 (no pain) to 100 (the worst pain imaginable). The Visual Analogue Scale has properties consistent with a linear scale for patients with mild to moderate pain.

Data was collected and transferred to a worksheet by a research assistant (who was not involved in the physical examination or palpation of MTrP).

Frequency distributions, means, medians, standard deviations, and 95% confidence intervals were calculated for descriptive purposes. The Shapiro-Wilk W test was used to test for normality of the data. Because the number of muscles with MTrPs (active, latent and total) was not normally distributed we used the Spearman's rank-order correlation (ρ) test for all variables. For interpretation of the ρ-values, we used the classification proposed by Feinstein [62]. A correlation coefficient < 0.30 was considered to be indicative of a poor correlation. A correlation coefficient ≥ 0.30 and ≤ 0.70 was considered to be indicative of moderate correlation, and a correlation coefficient ≥ 0.70 was defined as substantial or a good correlation. To test for differences in age, gender distribution, and education level between the current study population and study populations from Dutch shoulder studies (from 1995 until 2009), we used a one sample T-test. The α level for statistical significance was set at 0.05. All analyses were performed using Systat 12 or Sigmastat 3.1 for Windows (Systat Software, Inc. Chicago, IL, USA).

Muscles containing active MTrPs were found in all 72 subjects. The median number of muscles with active MTrPs per subject was 6 (range 2 to 16). Muscles containing latent MTrPs were found in 67 subjects. The median number of muscles with latent MTrPs per subject was 4 (range 0 to 11). Figure 6 shows the frequency distribution of active and latent MTrPs per subject. Neither active MTrPs nor latent MTrPs were normally distributed (Shapiro W = 0.95; p < 0.05; W = 0.96; p < 0.05 respectively).

Active MTrPs were found in the infraspinatus muscle in 56 subjects and in the upper trapezius muscle in 42 subjects. In addition, active MTrPs were highly prevalent in the middle trapezius (n = 31), anterior deltoid (n = 34), middle deltoid (n = 36), posterior deltoid (32), and teres minor (n = 34) muscles.

Latent MTrPs were found in the infraspinatus muscle in 11 subjects and in the upper trapezius in 27 subjects. Latent MTrPs were found in the teres major muscle in 35 subjects and in the anterior deltoid muscle in 27 subjects. Figure 7 presents the distribution of active and latent MTrPs per muscle.

The mean score on the DASH was 30.8 (SD 14.1; 95% CI 27.5 to 34.1). Mean VAS-P scores were follows: the VAS-P score for 'current pain' (VAS-P1) was 30 (SD 23.9; 95% CI 27.0 to 39.9), for 'average pain in the last seven days' (VAS-P2) was 42.1 (SD 17.7; 95% CI 37.4 to 50.0) and for 'for the most severe pain in the last seven days' (VAS-P3) was 56.6 (SD 19.8; 95% CI 51.2 to 61.9). The mean PROM score, calculated as the sum the PROM value measured for the non-affected shoulder minus the PROM value measured for the affected shoulder, was 32.4 degrees (SD 34.8; 95% CI 24.2 to 40.6), where a positive value indicates that the affected shoulder has a impaired range of motion. Both DASH and PROM scores were normally distributed (W = 0.97; p < 0.05 and W = 0.91; p < 0.05 respectively). VAS-P1, VAS-P2, and VAS-P3 scores were also considered to be normally distributed, although the Shapiro-Wilk test did present borderline results for VAS-P2 and VAS-P3.

The number of muscles with active MTrPs only moderately correlated with the DASH-DLV (ρ = 0.30; p < 0.05) and VAS-P1 scores (ρ = 0.33;p < 0.05), and poorly correlated with VAS-P2 (ρ = 0.28; p < 0.05) and the duration of the shoulder pain (ρ = 0.26, p < 0.05). We were unable to detect statistically significant correlations between the number of muscles with MTrPs (either active or latent) and VAS-P3 (ρ = 0.09; p > 0.05) or the PROM (ρ = 0.13; p > 0.05) scores. Table 4 provides an overview of the correlations and Figure 8 shows a scatterplot of DASH scores versus the number of active MTrPs.

All subjects with unilateral, chronic, non-traumatic shoulder pain presented with multiple shoulder muscle MTrPs. In addition, MTrPs were found in all 17 muscles examined. However, the number of shoulder muscles with MTrPs appeared to vary greatly among subjects. In particular, MTrPs were most frequently located in the infraspinatus and upper trapezius muscles, in agreement with results from Skootsky [37] and Simons [26], who found that infraspinatus muscles were frequently associated with myofascial shoulder pain shoulder. There are very few other prevalence studies in the literature, and to the best of our knowledge, this is the first extensive report on the prevalence of MTrPs in patients with chronic, non-traumatic unilateral shoulder pain.

The mean DASH-DLV score measured for the current study population is comparable with the mean baseline scores measured for other study populations for subjects with shoulder and arm pain [63‐65]. According to Beaton[66] subjects (n = 200) with DASH scores < 23.6 are still able to perform all desired daily activities, although they may experience some discomfort. For comparison, in a study population from the US (n = 1706), the mean DASH score was 10.10 (SD 14.68) and in young active and healthy adults the mean DASH score was 1.85 (SD 5.99) [67]. Importantly, the DASH-DLV score primarily reflects the level of dysfunction with less emphasis on pain and other symptoms. While 23 items refer to the ability of the subject to perform activities, only 7 items assesses the severity of symptoms. Subjects with long-standing shoulder complaints may alter the way in which they perform activities by using compensatory movements. In addition, DASH-DLV does not discriminate between activities performed using the affected or non-affected arm, which may influence the magnitude of the disability and therefore the final DASH-DLV score. In support of this, several subjects in our study commented that their DASH score would have been different if the activities in question were related to the affected arm.

The number of muscles containing active MTrPs moderately and positively correlated with DASH-DLV, VAS-P1, VAS-P2 scores, and the duration of the shoulder pain, suggesting that the number of muscles with active MTrPs explained only 10% of the variation of the outcome measures, including pain and disability. In addition, other clinically relevant factors may have contributed to the primary and secondary outcome scores. First, although we did not measure the pain intensity at the MTrP, this may have a significant impact on pain and functioning. Hidalgo et al found that patients with shoulder pain had a larger number of both active and latent MTrPs than healthy subjects. They also found that active MTrPs were associated with greater pain intensity, and that lower Pain Pressure Thresholds (PPT) were reported for active MTrPs compared to latent and patients with shoulder pain displayed lower PPT than healthy subjects [49]. Second, in this study we did not take into consideration the number of MTrPs per muscle, which may have contributed to the moderate correlation observed between the number of muscles with MTrPs and the DASH-DLV and VAS-P scores. The total number of muscles with MTrPs was poorly but positively correlated with the duration of the complaints, indicating that the number of shoulder muscles with MTrPs may increase over time regardless of whether the MTrPs were active or latent. Finally, because one of the characteristics of the DASH-DLV score is, that it does not discriminate between the affected and the non-affected shoulder, one could speculate that patients with chronic shoulder pain may develop strategies to overcome pain and disability caused by their shoulder disorder, for instance by using the non-affected arm, resulting in decreased DASH-DLV and VAS-P scores. All these factors may have a substantial influence on the correlation coefficient. Although the number of shoulder muscles with active MTrPs correlates only moderately with the various outcome measures, this does not imply that MTrPs are clinically unimportant. Future studies of chronic shoulder pain examining the total number of trigger points and their pressure sensitivity in the muscles studied could substantially impact the magnitude of the effect of presence of myofascial trigger points on shoulder pain and disability.

To date, unilateral shoulder pain has mainly been proposed to be due to either the presence of inflammation in the subacromial tendons and bursae, or degenerative rotator cuff ruptures (diagnosed using modern imaging techniques, such as MRI or sonography). Although these pathological structures may cause pain, it is also known that similar abnormalities have been found in asymptomatic shoulders.

Active MTrPs, which are painful spots that produce familiar shoulder pain during contraction, stretching or compressing, these MTrPs may provide an alternative explanation for shoulder pain, which is independent of the presence of subacromial abnormalities. According to Simons, Travell and Simons [26], MTrPs within the infraspinatus muscle (which were most prevalent) cause pain in the anterior and middle deltoid regions which expands into the frontal upper arm, as well as referred pain and referred sensations felt in the wrist and the hand. In addition, internal rotation and cross-body adduction may be impaired, which is often the case in patients with shoulder pain. Both experimentally induced and spontaneous muscle pain lead to an aberrant motor activation pattern that is also present in patients with shoulder pain [68, 69]. Although latent MTrPs are not usually an immediate source of pain, they can elicit referred pain when mechanically stimulated, or during sustained or repeated muscle contraction. In addition, latent MTrPs may disturb normal motor recruitment patterns and movement efficiency. Lucas et al. showed that subjects who received myofascial dry needling, followed by passive muscle stretching to remove latent MTrPs, showed normalized motor activation patterns within 20 to 30 minutes following the treatment [48]. Therefore, it is reasonable to expect that treatment of MTrPs may lead to normalization of motor activation patterns and may facilitate spontaneous recovery of shoulder pain, either without exercising or by making exercise more effective.

Based on the results of this study, we propose that an alternative approach may be indicated for the assessment and management of patients with chronic, non-traumatic shoulder pain. Current treatment regimens consist primarily of pharmacological interventions, including anti-inflammatory medications, or muscle strengthening exercises. If MTrPs are one of the main reasons for shoulder pain (active MTrPs) and altered motor activation patterns (active and latent MTrPs), as several authors have proposed, then anti-inflammatory treatment [26, 48, 70] and muscle strengthening exercises should not be the treatment of first choice. Instead, the treatment should begin with MTrP inactivation. Manual techniques, including manual compression of the MTrP, known as ischemic compression or trigger point release, trigger point dry needling or injection therapy are used to inactivate MTrPs. After MTrP inactivation, muscle stretching and relaxation exercises, heat applications, dynamic exercises to improve range of motion and muscle reconditioning are instructed as appropriate. This therapy is accompanied with a gradual increase in daily activities.

If the above hypothesis is true, treatment of MTrPs could provide an innovative, promising therapy for shoulder pain. This study shows the results of patients' characteristics for a sample of patients with chronic, unilateral non-traumatic shoulder pain, who were recruited for a randomized clinical trial to study the results of MTrPs directed interventions by physical therapists in this group. The results of this study are published[71].

We compared sociodemographic data from the current study population with similar data from several other Dutch shoulder pain research studies. Because none of these studies investigated MTrPs, we made this comparison to see whether there was reason to expect that the high prevalence of MTrPs we observed was unique to our population. In our study population more females were included, and the subjects were significantly younger and more highly educated than subjects from the other Dutch populations, although a specific explanation for these differences is lacking. There is no reason to suspect that educational levels correlate with the number of MTrPs and awareness of educational levels is mainly important for effectiveness studies, because they may impact the patients' motivation and compliance [72, 73]. However, increased age may also be associated with increased number of MTrPs [74]. Because the subjects of the present study were younger, and musculoskeletal complaints tend to increase with age [74], there is no reason to suspect that we overestimated the prevalence of MTrPs in our population. On the other hand, there were more females in our study population, and females may be more prone to musculoskeletal disorders in general [75]. Thus, for this reason there may be a chance that MTrPs were slightly more prevalent in our study population [76‐78]. Despite the above-mentioned differences, we conclude that our subjects are comparable with other patients with chronic shoulder pain and the findings of this study can be generalized to other patients.

One of the limitations of our study is that we only examined patients with unilateral chronic shoulder pain and dysfunction, whereas MTrPs are thought to be responsible for both acute and chronic pain. It is conceivable that patients who developed chronic shoulder pain may have more MTrPs, and persistent MTrPs in the acute phase than patients who recover easily. In future research projects assessment of MTrPs in patients with acute shoulder problems should also be included. The small sample size is another limitation of this study. Before starting this study a power analysis was performed and it was calculated that 104 subjects would be needed for the clinical trial. After two years (one year more than originally planned, 72 subjects were enrolled in the study. For practical reasons, the study was completed with this smaller sample size, which may have influenced some of the results of this study. We used two observers in this study with identical clinical experience and post-graduate training on myofascial trigger point therapy. In addition, both observers found a comparable mean number of active MTrPs. Because there was no statistically significant difference in mean DASH scores obtained by the two observers, we consider both groups to be comparable and the findings obtained by both observers to be similar.