Arthroscopic rotator cuff repair (ARCR) is a minimally invasive surgery that is used to treat rotator cuff tears with minimal postoperative pain and rapid functional recovery. Although ARCR has favorable postoperative functional outcomes1, dissatisfaction due to residual pain and disability due to postoperative stiffness has been previously reported2.
Postoperative stiffness, which can cause pain and restrict daily activities, is one of the most frequent complications of rotator cuff repair, reportedly occurring in 18.6% of patients at 3 months postoperatively3. It can cause pain and restrict daily activities3. Adequate rehabilitation is essential to restore function after ARCR; however, patients with painful stiffness may find it difficult to follow the recommended rehabilitation.
Treatment with intra-articular steroid injections has been attempted to manage postoperative painful stiffness after ARCR4. However, intra-articular steroid injection after ARCR may lead to steroid-related complications, including infection and healing failure of the repaired tendon5,6.
The accuracy of intra-articular injections is another possible issue. The accuracy of intra-articular injection without image guidance is reportedly <50%7,8. Although image guidance may improve the accuracy of intra-articular injection9, it requires additional instruments, such as ultrasound or fluoroscopy.
The subacromial space is located superficial to the glenohumeral joint and anatomical structures adjacent to the subacromial space, including the acromion and deltoid, can be easily palpated. Therefore, subacromial steroid injection (SAI) requires less technical precision than intra-articular injection of the glenohumeral joint. Furthermore, overexpression of inflammatory factors has been observed in the joint capsule as well as the subacromial space in patients with frozen shoulder10. Oh et al.11 reported that SAIs are more efficacious than intra-articular injections for relieving pain and improving the range of motion (ROM) in patients with primary frozen shoulder.
Most ARCR procedures are conducted in the subacromial space; therefore, we considered that inflammation in the subacromial space due to operative procedures during ARCR may contribute to postoperative pain and stiffness. We hypothesized that an SAI at 3 months after ARCR may improve early postoperative functional outcomes, including pain and ROM in patients with painful postoperative stiffness, without increasing the occurrence of steroid injection-related complications. Therefore, we evaluated the effectiveness and safety of SAIs in patients with persistent stiffness 3 months after ARCR.
We retrospectively reviewed the medical records of 777 consecutive patients who underwent ARCR at the senior author’s institution between January 2012 and May 2014. All ARCR procedures were performed by a single surgeon. This study was approved by the Institutional Review Board of Seoul National University Bundang Hospital (No. B-2206-765-101) and performed in accordance with the principles of the Declaration of Helsinki.
The data used in this study were fully anonymized and the requirement for informed consent was waived due to the retrospective study design. Before administering the SAI, patients were individually informed about the potential risks, including local complications such as infection, hemorrhage, healing failure, and possibility of reoperation, as well as systemic complications such as temporary uncontrolled glycemic control and aggravation of diabetes mellitus or chronic kidney disease5,6,12-14. The SAI was administered only after obtaining informed consent from each patient. The influence of confounders was minimized by excluding patients with partially repaired tendons (n=48), isolated infraspinatus tear (n=1), isolated subscapularis tear (n=12), concomitant subscapularis tear (n=207), revision ARCR (n=12), history of pyogenic arthritis (n=3), fracture of the ipsilateral shoulder (n=3), two or more steroid injections within 3 months after ARCR (n=16), and follow-up of <1 year (n=75). We only included patients without re-tear who were evaluated by ultrasonography at 3 months after ARCR; therefore, patients with re-tear within 3 months after ARCR (n=28) or who were not evaluated using ultrasonography (n=72) were also excluded. A total of 300 patients were finally enrolled in this study.
A shoulder fellowship-trained orthopedic surgeon (JKK) collected data on demographics (age at operation, sex); intraoperatively- measured tear size; functional outcomes including ROM, visual analog scale for pain (pVAS), and Constant score; and healing failure at postoperative radiologic follow-up through a retrospective review of medical records. Functional outcomes were independently measured by a clinical researcher who was not involved in this study, while healing failure was determined based on official reports from a musculoskeletal radiologist with over 15 years of experience, who was also not involved in this study.
Functional outcomes were assessed by evaluating the preoperative functional status as the baseline on the day before surgery. ROM was measured at 3, 6, and 12 months after ARCR and at every annual follow-up visit; the pVAS and Constant scores were evaluated at 6 and 12 months after ARCR and at every annual follow-up visit. Forward flexion and external rotation were measured using a goniometer in the neutral position and with the arm at the side, respectively, and internal rotation was assessed based on the highest level of the spine that the patient could touch with the ipsilateral extended thumb. Postoperative stiffness was diagnosed if any one of the following criteria for passive ROM at 3 months post-ARCR were met: forward flexion <120°, external rotation <30°, or internal rotation to a level lower than L315. To assess the presence of stiffness, passive ROM measured at 3 months post-ARCR was utilized, whereas active ROM measured at 6 months after ARCR was used to evaluate functional recovery.
The integrity of the repaired tendon was assessed using ultrasonography at 3 and 6 months after ARCR and using magnetic resonance imaging at 12 months after ARCR. Subsequently, ultrasonography was performed at each annual follow-up. Healing failure was defined as Sugaya classification type IV or V16.
Among the 300 patients enrolled in the study, 148 who received an SAI due to limited ROM at 3 months after ARCR were classified as the SAI group, and the remaining 152 patients who did not receive an SAI were classified as the control group.
The SAI consisted of 1 mL of triamcinolone (40 mg) and 7 mL of ropivacaine (7.5 mg/mL; 52.5 mg) and was administered to patients who presented with postoperative stiffness without re- tear at 3 months after ARCR. Prior to receiving the SAI, the integrity of the repaired tendon was reconfirmed by a shoulder fellowship-trained orthopedic surgeon using ultrasonography. SAI was performed with the patients in a sitting position and the ipsilateral shoulder was internally rotated at the back. Injection into the rotator cuff or deltoid was avoided by using an 18-gauge needle, which was inserted into the subacromial space above the repaired tendon and the entire injection process was monitored using ultrasonography.
All operative procedures were performed in the lateral decubitus position using the Spider Limb Positioning System (Smith & Nephew) under general anesthesia. The tear size was measured using a calibrated probe during ARCR and categorized according to the Cofield classification as: small, <1 cm; medium, 1–3 cm; large, 3–5 cm; and massive, >5 cm or ≥2 completely torn tendons17.
After surgery, immobilization was maintained with an abduction brace for 4–6 weeks according to the tear size as follows: small, 4 weeks; medium, 5 weeks; and large-to-massive, 6 weeks18. Active assisted ROM exercises were allowed after the immobilization period. Muscle strengthening exercises started approximately 3 months after ARCR and sports activities were usually permitted 6 months after ARCR. All rehabilitation procedures were supervised by the Department of Rehabilitation Seoul National University Bundang Hospital.
All statistical analyses were performed using IBM SPSS version 22.0 (IBM Corp.). The distribution of continuous variables was evaluated using the Shapiro-Wilk normality test. Continuous variables were analyzed using the independent t-test or Mann-Whitney U-test, according to the normality. For categorical variables, the chi-square test or Fisher exact test was used to determine differences. All statistical tests were two-sided, and the significance level was set at 0.05.
The mean follow-up period was 18.1±4.7 months (range, 12.1–37.2 months) and did not differ significantly between the two groups (p=0.731) (Table 1). Demographic factors, including age at operation, sex, and tear size, also did not differ significantly between the SAI and control groups (all p>0.05) (Table 1). None of the patients experienced early local or systemic steroid injection-associated complications requiring additional intervention, such as infection, hemorrhage, uncontrolled glycemic control, or acute aggravation of diabetes or chronic kidney disease.
Table 1 . Comparison of demographic characteristics between the steroid injection and control groups
| Characteristic | Steroid injection group | Control group | p-value |
|---|---|---|---|
| No. of patients | 148 | 152 | |
| Age at operation (yr) | 64.4±7.8 | 61.8±8.6 | 0.062 |
| Sex, male:female | 46:102 | 60:92 | 0.133 |
| Tear size, small:medium:large* | 28:74:46 | 36:71:45 | 0.604 |
| Follow-up (mo) | 18.0±4.3 | 18.2±4.4 | 0.731 |
Values are presented as number only or mean±standard deviation.
*Tear size was categorized into small (<1 cm), medium (1−3 cm), and large (3−5 cm).
At 3 months after ARCR, ROM was significantly lower in the SAI group than in the control group (all p<0.001) (Table 2). However, functional outcomes, including ROM, pVAS, and Constant scores, were not significantly different at 6 months after surgery and the final follow-up (all p>0.05) (Table 2).
Table 2 . Comparison of functional outcomes between the steroid injection and the control groups
| Variable | Steroid injection group | Control group | p-value |
|---|---|---|---|
| Postoperative 3 mo | |||
| Forward flexion (°) | 132.3±11.3 | 149.4±10.9 | <0.001* |
| External rotation (°) | 29.2±11.8 | 43.3±10.8 | <0.001* |
| Internal rotation (VL) | L3.7±2.1 | T11.9±1.8 | <0.001* |
| Postoperative 6 mo | |||
| Forward flexion (°) | 162.9±10.3 | 165.5±10.9 | 0.072 |
| External rotation (°) | 66.2±10.5 | 70.6±9.2 | 0.064 |
| Internal rotation (VL) | T9.6±2.3 | T9.2±2.5 | 0.171 |
| Pain (VAS) | 1.3±1.0 | 1.0±0.9 | 0.083 |
| Constant score | 90.3±9.1 | 90.1±9.9 | 0.752 |
| Final follow-up | |||
| No. of patients at each postoperative time point, 1 yr:2 yr:3 yr | 134:13:1 | 139:13:0 | 0.918 |
| Forward flexion (°) | 168.8±11.0 | 169.4±11.1 | 0.754 |
| External rotation (°) | 77.7±10.5 | 79.5±8.3 | 0.271 |
| Internal rotation (VL) | T8.4 ± 2.1 | T8.4 ± 1.9 | 0.442 |
| Pain (VAS) | 0.8 ± 0.9 | 0.7 ± 0.8 | 0.523 |
| Constant score | 93.4 ± 6.2 | 93.4 ± 6.5 | 0.964 |
Values are presented as mean±standard deviation or number only.
VL: vertebral level, L: lumbar vertebrae, T: thoracic vertebrae, VAS: visual analog scale.
*Statistically significant.
In this study, the administration of a single SAI 3 months after ARCR was found to be an effective method for reducing pain and facilitating rehabilitation of patients with painful postoperative stiffness after ARCR. No steroid-related adverse effects were observed during the follow-up period, and the functional and radiological outcomes were not significantly different between the SAI and control groups.
These findings are in line with those of previous studies that reported the effectiveness of SAIs for pain reduction and ROM recovery after surgery19,20. However, these studies selected patients for steroid injection based on the pVAS alone, without considering the ROM19,20. Although pain may affect the ROM, when considering the definition of adhesive capsulitis, which is defined as the gradual development of global limitation of active and passive shoulder ROM without specific radiographic findings other than osteopenia21, the limitation of ROM appears to be more important than pain. Therefore, we selected patients for steroid administration depending on the presence of postoperative stiffness rather than postoperative pain to evaluate the effect of SAI on postoperative stiffness19,20.
In this study, no steroid-related complications, including infection, were observed in the SAI group and the healing failure rate was not significantly different between the two groups. Steroids have potent anti-inflammatory and immunosuppressive properties. Therefore, they may effectively reduce pain originating from inflammatory reactions by reducing the production of prostaglandins or leukotrienes and inhibiting phospholipase A222. Steroids may have an adverse impact on the healing process of repaired tendons after rotator cuff repair, as the anti-inflammatory effects of the steroid in the early postoperative period may inhibit healing after ARCR23. Clinically, we were unable to precisely estimate the duration of the inflammatory process after ARCR; however, several previous studies, have shown that healing failure occurs more frequently in the early postoperative period within 3 months after ARCR24,25. Therefore, steroids were only administered to patients without a re-tear, which was confirmed by ultrasonography at 3 months after ARCR, to decrease healing failure because of the suppression of the inflammatory process in the early postoperative period. Our study results indicate that the administration of an SAI 3 months after ARCR did not increase the risk of steroid-related complications.
In this study, limited ROM at 3 months after ARCR was effectively managed by the administration of an SAI without increasing the occurrence of re-tear. A previous study demonstrated that early postoperative stiffness after ARCR is advantageous for cuff healing26, and a recent animal experimental study showed that serum from patients with early postoperative stiffness after rotator cuff repair could induce fibroblast activation and induce capsular fibrosis27. Therefore, fibrosis induced by fibroblasts after ARCR may promote healing but also contribute to early postoperative stiffness28.
Hettrich et al.29 reported that intra-articular steroid injection could decrease the presence and amount of fibrosis in patients with primary adhesive capsulitis. In the present study, steroids were administered via SAI, however, Oh et al.10 showed that the efficacy of intra-articular injection was not superior to that of SAI in patients with adhesive capsulitis. Furthermore, most operative procedures during ARCR are performed in the subacromial space, and not in the glenohumeral joint. Although we did not perform histological analyses, we presumed that an SAI may have effects similar to those of an intra-articular injection; consequently, it may resolve the adhesions in the subacromial space originating from fibrosis after ARCR.
This study has some limitations. First, we were unable to exclude the inherent bias associated with retrospective studies. However, the demographic factors and tear characteristics, which could act as confounders, did not significantly differ between the two groups. Furthermore, we included a large number of patients who were operated upon by the same surgeon. The relatively short follow-up period may be another limitation. However, per several previous studies, most structural failures after ARCR occurred within 3 months after surgery24,25. Moreover, Oh et al.30 reported that the evaluation of healing at 6 months after surgery may be relevant for predicting structural failure after ARCR. Therefore, we considered that the follow-up duration of this study may be sufficient to evaluate the efficacy and safety of SAI at 3 months after ARCR in patients with postoperative stiffness after ARCR. Lastly, in this study, the recovery of ROM was evaluated exclusively using active ROM, from 6 months postoperatively. Although both passive and active ROM were routinely measured at 3 months after surgery, passive ROM was not assessed beyond 6 months unless patients exhibited shoulder stiffness. Due to the retrospective nature of this study, we were unable to separately analyze the effect of SAI on passive ROM. However, since passive ROM is generally greater than active ROM, using passive ROM at 3 months to diagnose shoulder stiffness and active ROM to assess functional recovery is unlikely to have overestimated the effect of SAI. In conclusion, in patients with postoperative stiffness 3 months after ARCR, a single SAI may relieve pain and improve functional outcomes without increasing the risk of healing failure.
Table 3 . Healing failure rates at final follow-up according to whether steroid injection was given and tear size
| Tear size | Steroid injection group | Control group | p-value |
|---|---|---|---|
| Overall | 22/148 (14.9) | 20/152 (13.2) | 0.671 |
| Small, <1 cm | 1/28 (3.6) | 1/36 (2.8) | 0.862 |
| Medium, 1−3 cm | 6/74 (8.1) | 5/71 (7.0) | 0.814 |
| Large, 3−5 cm | 15/46 (32.6) | 14/45 (31.1) | 0.883 |
Values are expressed as number (%).
No potential conflict of interest relevant to this article was reported.
Conceptualization: JHO. Data curation, Formal analysis, Investigation: AW, YKM, JHY, JKK. Methodology, Project administration, Resources, Supervision: HJJ, JHO. Writing–original draft: all authors. Writing–review & editing: HJJ. YKM, JHO.
