Effect of platelet-rich plasma on the degenerative rotator cuff tendinopathy according to the compositions

Rotator cuff (RC) tendinopathy is a degenerative musculoskeletal disease caused by overuse. The prevalence of RC disease, specifically partial and full thickness RC tendon tears, increases as a function of age starting at 40 years [1] and may be high as 50% by the age of 70 years [2].

Conservative treatments for RC tendinopathy include exercise, medication, and injection therapy. Medication and steroid injection therapy only help to reduce the pain, but exercise is effective as a treatment for regaining function in addition to reduction of pain in RC tendinopathy [3]. However, exercise takes a long time for recovery, and patient compliance is needed to attain maximal effect [4]. Other conservative therapies, such as prolotherapy and ESWT, have been tried to regain function for RC tendinopathy but the clinical evidence is not yet clear [5, 6]. There remains no consensus in conservative therapy except exercise.

Recently, autologous growth factors like platelet-derived growth factor and vascular endothelial growth factor found in platelet-rich plasma (PRP) have been established to play a critical role in cell proliferation, chemotaxis, cell differentiation, and angiogenesis [7]. Several studies have reported favorable clinical outcomes with the use of PRP in the treatment of acute and chronic tendinopathies [8‐12] and rotator cuff tears [13]. In contrast to the positive potential of PRP reported in the basic research literature, clinical outcomes have been reported as better [14‐16], not different [17‐20], or even worse [21] in RC tendinopathy. Therefore, there are controversies about PRP as an established treatment option for RC tendinopathy [22].

In our thought, these studies used different PRP producing methods, which resulted in differences in platelet concentrate and other compositions [23] and diverse clinical outcomes. In addition, these differences in cellular composition can affect regeneration effect benefits. A recent study documented that leukocytes in PRP increased catabolic signaling molecules such as matrix metalloproteinase-9 (MMP-9) and interleukin-1β (IL-1β) [24] and these catabolic proteases can perpetuate inflammation and inhibit tissue healing [25]. Dohan commented that PRP was classified as four main families based on their fibrin architecture and cell content, which affects clinical effects [26]. Therefore, we think that it is important to understand the composition of PRP to enhance tendon healing in RC tendinopathy.

PRP treatment has many advantages in terms of relative safety, easy production, and cost-effectiveness. The purpose of the study was to find the relation of cellular component with clinical efficacy in RC tendinopathy and to find the composition of PRP in treating RC tendinopathy.

A total of 30 patients were recruited from April 2015 to January 2017. Table 1 shows the initial baseline characteristics of the patients. In the PRP group, three patients were lost to follow-up from 6 weeks after the PRP injection, one patient was lost to follow-up from 12 weeks after the procedure, and four patients were lost to follow-up from 24 weeks after the procedure. In the exercise group, five patients were lost to follow-up from 6 weeks after the PRP injection, two patients were lost to follow-up from 12 weeks after the procedure, and no patients were lost to follow-up from 24 weeks after the procedure. Lost data was managed by the worst-case imputation method. Among patients lost to follow-up in the PRP group, one patient paradoxically increased their pain after PRP injection and was diagnosed as having newly developed adhesive capsulitis. Exercise compliance in the exercise group was 90.8 ± 30.6% at 6 weeks, 94.2 ± 37.4% at 12 weeks, and 87.4 ± 39.0% at 24 weeks.

ASES scores in the PRP group changed from 42.8 ± 18.4 to 62.7 ± 19.4 at 6 weeks (P < 0.001, paired t test), 72.4 ± 17.3 at 12 weeks (P < 0.001, paired t test), and 68.0 ± 23.8 at 24 weeks after PRP injection (P = 0.003, paired t test). ASES scores in the exercise group changed from 59.0 ± 13.4 to 65.4 ± 16.4 at 6 weeks (P < 0.012, paired t test), 72.3 ± 11.0 at 12 weeks (P < 0.001, paired t test), and 79.7 ± 14.1 at 24 weeks after exercise (P < 0.001, paired t test). Because baseline ASES scores showed significant difference between the PRP and exercise groups, linear regression analysis was conducted for adjustment of baseline ASES scores. Linear regression analysis showed no significant difference of ASES scores between the groups at 6 weeks (P = 0.582) and at 12 weeks (P = 0.969) but showed a significant difference at 24 weeks (P = 0.050; Fig. 2). Per protocol analysis showed the same results with intention-to-treat analysis with worst-case imputation.

Constant-Murley scores in the PRP group changed from 66.5 ± 17.7 to 76.3 ± 14.9 at 6 weeks (P = 0.008, paired t test), 81.6 ± 15.3 at 12 weeks (P = 0.001, paired t test), and 81.7 ± 17.4 at 24 weeks after PRP injection (P = 0.022, paired t test). Constant-Murley scores in the exercise group changed from 80.9 ± 11.6 to 81.2 ± 16.1 at 6 weeks (P = 0.787, paired t test), 82.7 ± 13.3 at 12 weeks (P = 0.953, paired t test), and 90.2 ± 9.5 at 24 weeks after exercise (P = 0.003, paired t test). Because baseline Constant-Murley scores showed significant difference between the PRP and exercise groups, linear regression analysis was conducted for adjustment of baseline Constant-Murley scores. Linear regression analysis showed no significant difference of Constant-Murley scores between groups at 6 weeks (P = 0.258) and at 12 weeks (P = 0.795) but showed a significant difference at 24 weeks (P = 0.048; Fig. 2). Per protocol analysis showed the same results with intention-to-treat analysis with worst-case imputation.

NRS scores in the PRP group changed from 5.7 ± 2.3 to 3.6 ± 2.6 at 6 weeks (P = 0.004, paired t test), 2.9 ± 2.6 at 12 weeks (P = 0.003, paired t test), and 2.9 ± 2.7 at 24 weeks (P = 0.007, paired t test) after PRP injection. NRS scores in the exercise group changed from 4.8 ± 1.6 to 4.4 ± 1.8 at 6 weeks (P = 0.202, paired t test), 3.3 ± 1.1 at 12 weeks (P < 0.001, paired t test), and 2.3 ± 1.5 at 24 weeks (P < 0.001, paired t test) after exercise. Independent t test showed significant group difference at 6 weeks (P = 0.031) but not at 12 and 24 weeks (P = 0.147 and 0.935). Per protocol analysis showed the same results with intention-to-treat analysis with worst-case imputation.

Thickness of the supraspinatus in the PRP group decreased 6 months after PRP injection while thickness in exercise group slightly increased 24 weeks after exercise.

The distribution of PRP components is presented in Additional file 2: Figure S2. The correlation of PRP components with changes of ASES, Constant-Murley score, and NRS showed that significant correlation was found between IL-1β and change of Constant-Murley score at 12 weeks (P = 0.046), and between TGF-β1 and change of NRS at 12 weeks (P = 0.048; Table 2). ROC curves were drawn to acquire cutoff values of IL-1β and TGF-β1 using meaningful improvement of Constant-Murley score and NRS (Fig. 3) and 5.19 pg/ml in IL-1β and 61.79 μg/ml in TGF-β1 were acquired as cutoff values to predict meaningful improvement.

Patients in the PRP group were divided into two subgroups according to the cutoff values for IL-1β and TGF-β1 and clinical outcomes in these two subgroups were again compared with those in the exercise group. The PRP subgroup above the TGF-β1 cutoff value showed significant difference in ASES and Constant-Murley score compared with the exercise group while the PRP subgroup below cutoff value showed no significant difference in linear regression analysis (Table 3). The PRP subgroup above IL-1β cutoff value showed significant differences in all clinical outcomes compared with the exercise group while the PRP subgroup below the cutoff value showed no significant differences in linear regression analysis (Table 3).

In this study, we found that TGF-β1 and IL-1β among cellular components of PRP were related to clinical efficacy for RC tendinopathy and concentration of IL-1β above 5.19 pg/ml and TGF-β1 above 61.79 μg/ml in PRP had better clinical outcomes for RC tendinopathy than the exercise group.

ASES score improved about from 10 to 20 in Cai’s study [34], from 10 to 30 in Kim’s study [35], and from 20 to 30 in Shams’ study [36] after PRP injection for partial tear of the rotator cuff tendon, which are similar to the results in our study. Constant-Murley score improved about 40 in Holtby’s study [37], from 15 to 25 in Shams’ and von Wehren’s studies [36, 38], which are similar to the results in our study. Although significant improvement in clinical outcomes after PRP injection was found to be similar to prior studies, this improvement was not superior to exercise. Considering that exercise must be continuously followed for several months to acquire desired clinical outcomes, PRP injection has an advantage over exercise in terms of 109pt?>In terms of time course, all clinical outcomes in the PRP group showed gradual improvement until 12 weeks but plateau or slight decrease at 24 weeks while those in the exercise group showed gradual improvement until 24 weeks and the greatest improvement was found at 24 weeks. This can be explained by the fact that PRP injection was done just one time without additional following injections, so the effect of PRP was diminished while exercise was maintained until 24 weeks, so the exercise effect was cumulative. We think that repeated injection of PRP is necessary to enhance the effect of PRP on RC tendinopathy.

The thickness of the supraspinatus was decreased at 24 weeks after PRP injection but was increased at 24 weeks after exercise. Tendon thickness in degenerative tendinopathy was seen to be increased on ultrasound examination [39] and decreased after adequate treatment [40], which can explain our results. Tendon thickness is known to be increased after exercise [41], which also explains our results.

IL-1β and TGF-β1 were correlated with clinical outcomes while other PRP components were not. IL-1β is a major cytokine that induces catabolic action on tendon fibroblasts via the upregulation of inflammatory mediators and plays a role in the tendon’s degenerative changes in tendinopathy [42] or regenerative capacity [43]. These roles might be affected by the concentration of IL-1β and our study showed IL-1β above 5.19 pg/ml had positive effects. TGF-β1 inhibits MMP-9 and MMP-13 expression to increase collagen accumulation [44] and is known to improve tendon strength and tendon healing [45, 46]. Our study demonstrated that TGF-β1 above 61.79 μg/ml was related to functional improvement in RC tendinopathy.

Okamura et al. demonstrated that IL-8 had a positive correlation with shoulder function improvement [47]. In other studies, PDGF, VEGF, and EGF, which are known growth factors in PRP, showed a positive correlation with shoulder function but our study did not show any significant correlation [48, 49]. This might be due to different concentration levels of these growth factors.

Unfortunately, baseline ASES and Constant-Murley scores in the exercise group were higher than those in the PRP group, which might cause biased results. To minimize this bias, baseline adjustment was done through linear regression analysis. However, to validate our results, further studies with larger sample sizes and randomized control designs are needed.

We did not analyze leukocyte concentration in the PRP because our study was focused on growth factors in the PRP. Several studies showed that leukocyte-rich PRP showed harmful effects on tendon healing while leukocyte-poor PRP showed beneficial effects [50‐52]. Hilber et al. [50] reported that leukocyte-reduced PRP stimulates the proliferation of tenocytes and Zhang et al. [51] demonstrated that proliferation of tendon stem cells cultured in leukocyte-rich was significantly decreased and tendon stem cells cultured in leukocyte-poor PRP produced more collagen and formed tendon-like tissue. However, these were in vitro studies and clinical studies did not show any significant difference between leukocyte-rich and leukocyte-poor PRPs [53, 54].

Plasma IL-1β and TGF-β levels vary according to the patient condition. Plasma IL-1β level, proinflammatory marker, is higher in rheumatoid arthritis patients [55] than in normal controls and TGF-β, anti-inflammatory marker, also increases in chronic obstructive pulmonary disease patients [56] or in multiple sclerosis [57]. Basic physical condition of patients like these might affect the IL-1β and TGF-β level in the PRP. However, these changes are less than 10-folds, while changes from the concentration of PRP are more than 1000 times. We think that concentration procedure during PRP preparation would affect the IL-1β and TGF-β level in the PRP more strongly than basic physical condition of patients.

In this study, we found IL-1β above 5.19 pg/ml in 58.3% of patients and TGF-β above 61.79 μg/ml in 61.2% of patients. PRP compositions might be affected by the centrifuge protocols [58] and kits used rather than patient characteristics. We used the GPS® III kit and this kit takes PRP from the buffy coat while other kits including ACP® (Device Technologies, MA, USA) take PRP from plasma [23]. In our study, baseline laboratory results including white blood cell and platelet counts from patients were not related to PRP compositions. Smoking and alcohol history did not have any correlation with the PRP compositions.

In our study, exercise compliance was measured by asking how many times the patient followed exercise. Considering that quality of measures of exercise adherence in musculoskeletal diseases is low [59], exercise compliance measures in our study would be limited. Further study to improve measurement of exercise compliance will be necessary.

We found that TGF-β1 and IL-1β among cellular components of PRP were related to clinical efficacy for RC tendinopathy and concentration of IL-1β above 5.19 pg/ml and TGF-β1 above 61.79 μg/ml in PRP had better clinical outcomes for RC tendinopathy than the exercise group. Despite several limitations, our study can help to find the optimal PRP condition and to enhance the effect of PRP on RC tendinopathy.