Collagen and PRP in Partial Thickness Rotator Cuff Injuries. Friends or Only Indifferent Neighbours? Randomized Controlled Trial

INTRODUCTION: Partial thickness rotator cuff injuries (PTRCI) is the sum of degenerative, overload and microtrauma processes, external supply of collagen and platelet-rich plasma (PRP) could potentially counteract deterioration of degenerative tendinopathy. AIM: Comparison of the effectiveness: collagen with PRP, PRP alone, collagen alone in the treatment of PTRCI. METHODS: Ninety patients with PTRCI treated with US–guided injections into the shoulder bursa every consecutive week: Group A - collagen with PRP (n=30), B - collagen alone (n=30), C - PRP alone (n=30). Primary outcomes: numeric rating scale (NRS), QuickDash and EQ-5D-5L questionnaires in control points: IA (initial assessment), T1, T2, T3 – after 6, 12 and 24 weeks, respectively. Secondary outcomes: number of patients with loss of RC continuity, number of regenerated RC between IA and T3. RESULTS: No statistical difference between groups in primary outcomes, tendency for further improvement in A and C group (opposite to group B) between T2 and T3. The RC discontinuity (n = 3, one case in each group) and RC regeneration (n = 65; 73%, 67% and 77%, in group A, B and C, respectively). CONCLUSIONS: Combined therapy of collagen and PRP in PTRCI is not more effective than monotherapies with collagen or PRP.


Introduction
Rotator cuff injuries (RCI) ranks third in the population prevalence among musculoskeletal system pathologies (16%) after lumbar spine pain (25%) and knee pain (19%). Depending on the source, the prevalence of RCI is from 5 to 39%. It increases very clearly with age and in patients over 60 years of age reaches over 30%, with a great majority being described as rotator cuff tendinopathy (RCT), mostly in the form of partial thickness RCI (PTRCI) as an emanation of the degenerative process [1].
Over 85% of the dry mass of the rotator cuff tendons is type I collagen. Damage, disorganization of collagen bers and a negative metabolic balance of collagen underlie the macroscopic lesions visible in ultrasound (US) or resonance imaging (RI) [2].
Most often, the PTRCI concerns the supraspinatus tendon (SSP), which is the crucial factor in centering of the humeral head in joint glenoid during the act of upper limb elevation. In traumatic cases also subscapularis (SSC) and infraspinatus (ISP) tendons are affected often with long head biceps (LHB) instability. There are several reasons for the degenerative process leading initially to edema, microperforations and then full thickness tendon lesions: 1) natural age-related weakening of blood supply near the SSP insertion, 2) concomitant degenerative spurs of acromioclavicular joint or acromion shape as a direct cause of the subacromial impingement, 3) disturbed muscle timing between RC and deltoid usually associated with cervical spondylosis (scapular dyskinesis), 4) shoulder joint multidirectional instability as a result of capsule-ligamentous elements laxity and disturbed contact of joint surfaces with RC posterior impingement. The consequences of RCI are further destabilization of the shoulder, scapular dyskinesis, upper and anterior migration of humeral head followed by subacromial bursitis. The clinical picture of PTRCI includes shoulder pain radiating to the deltoid area and even to the elbow both at rest and at strain, weakening of muscular strength, impaired limb function and disability of self-service. Due to the risk of surgical treatment, reduced strength of RC tendons affected by the degenerative process and a signi cant risk of injury recurrence, conservative treatment of PTRCI is the rst choice, especially for inactive patients over the age of 60 years. It assumes the alleviation of in ammatory symptoms (physical therapy, general and local pharmacotherapy -most often steroid injections), attempts of regenerative treatment as: Platelet Rich Plasma (PRP), collagen injections, autologous conditioned serum (ACS) and rehabilitation (muscular centering of the humeral head, developing compensatory movement patterns).
Reports on PRP show its positive effects, both alleviating symptoms and slowing down the process of further degeneration of the tendon, demonstrating its advantage over steroid administration or prolotherapy [3,4,5].
Another form of therapy aimed at suppressing the negative balance of collagen metabolism is to supply collagen in the form of an injection in the vicinity of the injured tissue (into the tendon itself or into the subacromial bursa). The premises for this type of injection are reports showing a reduction in pain after collagen injections compared to steroid injections and a signi cant acceleration of the proliferation and migration of tenocytes cultured in an exogenous collagen environment in vitro [6,7]. The same is true for synergic effects of collagen and PRP con rmed in multiple studies utilized tendon-like cell models, where • no consent Three groups of patients, each containing 30 participants, were enrolled in the study. Patients meeting the inclusion criteria were allocated randomly according to the computer-generated randomization list (block randomization; block size = 6). No changes of allocation and no changes in the methodology of the study took place throughout the study.
All data were collected at SMC Clinic.During the Initial Assessment (IA), patients were asked to evaluate intensity of the pain (Numeric Rating Scale, NRS, range from 0 (no pain) to 10 (extreme pain) and to complete widely used, validated questionnaires:QuickDash (0-50) and theEQ-5D-5L (descriptive part and EQ-VAS 0-100).US-examination of the shoulder was performed with the usage of Alpinion E-CUBE 12 device, linear transducer L3-12H (3-12 MHz).
SSP tendon width (cross-section in mm) was measured in the internal rotation position of the arm. We distinguished following ultrasound patterns of PTRCI: bursa-sided (BS), joint -sided (JS), intra-tendon (IT) and oblique or focal (OF). The measurement in BS and JS types was performed in the narrowest point (follow-up measure estimates tendency for increase of the RC width as a sign of regeneration). In IT or OF type of injury the measurement was performed at the thickest point of RC (follow-up measure estimates tendency for reduction of in ammatory and oedematous overgrowth of the RC as a sign of regeneration).
Each group was treated by three US-guided injections into the subacromial bursa using the in-plane technique. Injections were performed every consecutive week by the same physician (P.G.). Group Acollagen (3 vials of Collagen MD Shoulder -total 6 cc) simultaneously with PRP GLOFINN (10 cc whole blood, double centrifugate, leukocyte rich PRP, volume of PRP -2 cc); Group B -collagen alone (3 vials of Collagen MD Shoulder); Group C -PRP GLOFINN alone.
All patients were allowed to continue a rehabilitation protocol with preservation of safe, pain-free range of motion, postural exercises, scapular stabilization exercises. Prohibited were any exercises with resistance which would compromise the healing process of RC.
Secondary outcomes included percentage of patients in each group where the RC continuity was preserved with desired evolution of RC cross-section width and percentage of patients who had US signs of RC regeneration. Secondary outcomes were assessed at IA and T3.
The power of the test was set at 0.8 and the signi cance level at 0.05, assuming that the effect size was f = 0.35. This allowed us to establish that the research sample for the three compared groups should not be smaller than 90 subjects (each group with 30 participants).
Descriptive statistics for demographic data, ANOVA test to proof initial comparability of the groups and to check possible signi cant differences between groups according to age, NRS, QuickDash and EQ-5D-5L questionnaire VAS were performed using IBM SPSS version 25.
In the analysis of the collected research material, the one-way ANOVA test was used, which allowed us to check whether one independent variable (factor) affects the results of the dependent variable. The test results allowed us to determine whether the mean scores of the scales for individual control points differ statistically signi cantly between the groups. In order to determine between which groups there is a statistically signi cant difference, Tukey's post-hoc test was used. A calculation of the difference in value between the baseline IA and the T3 point for every single patient was also performed, and then the mean values of this difference was taken to compare primary outcomes in the groups.

Results
One hundred one patient were screened for eligibility. Ninety patients meeting inclusion criteria were randomized. One person from group A did not nish the therapy (for reason other than therapy intolerance) and had no T1, T2 and T3 observations. Two persons quit the study after T1 (group C/JS and group A/IT). One person left the follow-up appointment after T2 due to lack of improvement and asked for a change of therapy (group C/OF). Two patients faced total RC tear before the end of the observation -between T2 and T3 (group C/JS and A/IT) and one patient was found to have had complete RC injury at the T3 visit. Finally, eight people did not obtain the T3 control; one person did not complete the therapy, three people dropped out of further control after T1 (two patients experienced a complete tear of the RC between the visit T2 and T3) -group C/JS and group A/IT, one person from group C/OF asked to change therapy after T2 control), and one person from group A/JS refused control T3.
Recruitment and follow-up process presents Fig. 1.
The treatment has been accomplished by 89 patients (99%). All check-up visits were passed by 91.1% of patients. In 82 patients who completed therapy and obtained T3 control, three SSP total injuries were observed (3.6%). There was no participant cross-over.   Figure 2 presents the mean NRS evolution in speci c groups. A reduction in pain intensity is seen mostly in the rst 6 weeks of follow-up but no signi cant statistical differences between groups were noticed in ANOVA. There is a slight tendency in A and C group for further improvement beyond T2. Figure 3 presents the mean QuickDash evolution in speci c group where similar pattern of mean values reduction is observed also without signi cant statistical differences between groups in ANOVA test for QuickDash main questionnaire. Figure 4 presents the mean EQ-5D-5L VAS evolution in speci c groups. No statistically signi cant differences were found. The dynamics of changes during six weeks after last injection is similarly more intense.

Page 8/19
The ANOVA test for the EQ-5D-5L Index showed a statistically signi cant difference in baseline values (IA) between the groups, while no statistically signi cant differences were observed in the control points.
In order to check between which groups there are statistically signi cant differences, the post hoc test was used. The Games-Howell test was chosen due to the failure to meet the assumption of homogeneity of variance in the analyzed groups. The analysis showed differences (p < 0.05) between groups A (0.892) and C (0.816). It is worth noting that between groups A (0.892) and B (0.820) there is also a similar difference between the mean EQ-5D-5L indices, but the analysis showed no statistical signi cance (p = 0.063) [Fig. 5].
A calculation of the differences between the baseline IA and T3 values for every single patient was also performed. A mean value of this differences for each studied group were calculated. Figure 6 presents differences in the mean initial and nal values for primary outcomes in the groups. Group B shows the highest differences in all scales, although the ANOVA test did not show statistical signi cance between the groups.
Lost continuity of RC between IA and T3 was found in three cases (one in each group) and the number of cases with RC regeneration con rmed in ultrasound was: A-22, B-20, C-23.
Mean increase of RC width in BS and JS type of injury for speci c groups was: A -0.7 mm, B -0.2 mm, C -1.3 mm. There is statistically signi cant difference for B and C (p < 0.05) in ANOVA test. [Fig. 7].
Mean reduction of width for IT and OF type of injury for speci c groups was: A -0.7 mm, B -0.9 mm, C -0.3 mm. No statistically signi cant difference between groups in ANOVA test were found [ Fig.8]. Fig.8.
Mean reduction of width for IT and OF type of injury for speci c groups.
No signi cant harms, complication or unintended effects of the treatment were reported.

Discussion
Conservative treatment of PTRCI with injections of collagen and PRP as monotherapy or combined therapy showed no signi cant difference in e cacy.
The strength of our study is based on the rst ever performed test in vivo whether really exist the potential synergy between PRP and collagen delivered into subacromial bursa in terms of tendon regeneration by randomized control trial (the authors did not nd a similar study in the literature). In addition to the wellvalidated subjective assessment questionnaires, an ultrasound examination with six-month observation was used, which seems to be long enough to observe changes in echogenicity and possible change in tendon thickness.
The weakness of the study is certainly a small group of participants, imperfections in the methodology of RC thickness measurement in ultrasound as an operator dependent. The bias is mostly connected with di culties to obtain the same cross-section point of reference for precise test-retest measurement.
Another bias of the study which may modify the results is the wide margin of tolerance according to rehabilitation protocol which was implemented for the participants before or in the course of the study beyond of our control, as well a sport or working activities exerted by many of them against recommendations. There were also no restrictions on taking painkillers when needed during observations period.
Degenerative rotator cuff tendinopathy appearing as PTRCI is a condition challenging to treat, mainly because of the poor regenerative potential of the tendons corelated with aging. It has been described many other factors contributing to treatment failure like: overload in the rehabilitation process, drugs (i.e. quinolones), alcohol intake, smoking, corticosteroids [9].
For over two decades there have been a growing interest in biologically active substances like growth factors, stem cells or autologous conditioned serum [10].
There are many publications about PRP's potential enhancement of healing potential after surgical repairing of RCI and decreasing ratio of re-tear. However, the data are con icting [11,12,13,14].
In vitro culture experiments clearly con rm the anabolic effect of PRP on the healing of RC lesions through cell proliferation and synthesis of collagen I [15].
However in vivo, especially without RC repair there is much confusion and the conclusions are con icting. A protocol of PRP usage, similar to our study, brought by Freitag et al. (2014) in a case report of 60 years old patient treated with three doses of PRP for PTRCI in weekly intervals but administered not into bursa but into the partial supraspinatus tear using a lateral approach. He followed-up the outcomes throughout 52 weeks. The NRS, patient percentage perceived improvement (PPPI) and a handheld isometric dynamometer assessment of RC strength was recorded in follow-up intervals occurring at 6th, 17th, 25th and 52nd weeks revealing the best PPPI up to 90% in 17th week and slightly worse at 52nd week (70%) [ 20 -saline). Injection therapy was followed by 6 weeks rehabilitation program. They found no signi cant differences in improving quality of life, pain, disability, and shoulder range of motion than placebo in patients with PTRCI who were treated with an exercise program [19].
However, all the above-cited studies did not have subsequent imaging control to objectify tendon regeneration.
It seems to be reasonable to raise a question of insu cient dose of single shot PRP (especially if a low volume whole blood set was used), as a possible reason of unsatis ed results. Similar questions have been raised with respect to the type of PRP that may be optimal for promoting regeneration that was con rmed in laboratory comparative studies between low and high leucocyte PRP [20,21]. Our study showed clearly that the healing potential of RC no matter how weak still exists and can be activated or augmented by external delivery of biologic active substrates although without clear difference between monotherapy or combined therapy. However many questions about the optimal PRP composition, collagen dose, administration sequence (mixture or sequential administration) and injection location depending on the type of RC injury (intraarticular or intrabursal) remain open. The most interesting seem to be unknown connection between structural integrity of RC and clinical outcomes.

Conclusions
Combined therapy of collagen and PRP in PTRCI is not more effective than separate therapies.

Declarations
Funding