The most important finding of the present study is that postoperative stiffness at 3 months after isolated ARCR in patients without preoperative stiffness is a common complication with an overall incidence of 14.2% (39/274). Additionally, DM and the timing of rehabilitation were independent risk factors for early postoperative stiffness following ARCR.
Postoperative shoulder stiffness following ARCR may affect functional outcomes and a patient's satisfaction for the procedure. Because of the clinical importance of postoperative shoulder stiffness, an improved understanding of its incidence and a consistent definition is crucial for orthopedic surgeons. In a review of the literature, the incidence of postoperative stiffness following ARCR varies from 2.3 to 28.5% [6–10]. Reasons for this wide range were the potential subjectivity of the criteria, and inconsistent timing of measurements. Several authors defined shoulder stiffness as passive forward flexion of less than 100˚ and external rotation less than 30˚ [6, 9, 23]. Of the authors using this definition, Brislin et al. [23] reported that 23 of 263 patients (8.7%) had shoulder stiffness 3 months after ARCR, and Parsons et al. [9] noted that 10 of 43 patients (23.3%) experienced shoulder stiffness 6 to 8 weeks after ARCR. While Tan et al. [6] reported that 32 of 290 patients (11%) experienced postoperative shoulder stiffness at their 3 months follow-up visit, 25/32 (78.1%) had resolution of stiffness by 9–12 months. Kim et al. [24] defined shoulder stiffness as forward flexion of less than 140˚ or external rotation with the arm in 90˚ abduction of less than 40˚. They noted that 74 of 209 patients (35.4%) experienced postoperative stiffness within 6 weeks after ARCR. Chung et al. [12] set the criteria of shoulder stiffness for forward elevation at less than 120˚, external rotation with the arm at the side at less than 30˚, or internal rotation at the back as lower than the third lumbar vertebral level, as previously described by Oh et al. [12, 20]. According to these criteria, they reported that postoperative stiffness was observed in 18.8% (54/288) of patients at 3 months after ARCR, 2.8% (8/288) at 6 months, and 6.6% (19/288) at the final follow-up (mean 13.5 months) [12]. However, these previous studies included patients with preoperative stiffness or additional procedures that can affect postoperative stiffness. In the present study, 39 of 274 patients (14.2%) who underwent ARCR developed postoperative stiffness at 3 months after surgery. Unlike previous studies, our study included patients with ARCR only and excluded patients with preoperative stiffness, open or mini-open rotator cuff repair, and any who underwent additional procedure. Stiffness was defined as forward flexion of less than 120˚ and external rotation with the arm at the side of less than 30˚ in our study. These criteria were selected because they are easy to examine in the outpatient clinic. Several studies have included a motion of internal rotation in the definition of stiffness [3, 20]. However, we excluded an internal rotation from the diagnostic criteria of postoperative stiffness because we thought that a hand-behind-the-back ROM might not accurately assess active and passive internal rotation of the shoulder [25].
Although the etiology of stiffness after ARCR might be multifactorial and not completely understood, various risk factors (e.g., female gender, younger in age, DM, preoperative stiffness, hypothyroidism, systemic lupus erythematosus) are reported to be associated with postoperative stiffness [8, 13, 21, 26]. In particular, DM—which is a risk factor for frozen shoulder—has been widely studied for stiffness after ARCR [11–13, 21, 27]. Although some articles had not found an association between DM and postoperative stiffness [11, 12], several studies have reported that DM could be a potent risk factor for postoperative stiffness [13, 21]. Blonna et al. [13] reported that the overall incidence of postoperative stiffness was 29% (19/65) in patients who underwent ARCR or arthroscopic subacromial decompression. In their study, of the 12 patients who had DM or pre-diabetes conditions, 5 (42%) developed postoperative stiffness (relative risk = 5.7, p = 0.03). Burrus et al. [21] analyzed 232 of 19,229 patients (1.2%) who underwent lysis of adhesions or manipulation under anesthesia after isolated ARCR using the PearlDiver Patients Records Database. They reported that type-1 DM was a significant risk factor (odds ratio = 2.7, p < 0.0001). In the present study, it was also noted that DM was independently associated with postoperative stiffness at 3 months; 20.5% (8/39) of DM patients were in the stiffness group compared to 8.9% (21/235) in the non-stiffness group.
The present study revealed that rehabilitation was significantly associated with postoperative stiffness, and previous investigations also have shown that rehabilitation was closely related to postoperative stiffness [22]. Parsons et al. [9] retrospectively evaluated 43 patients who underwent full-time sling immobilization without formal therapy for 6 weeks after ARCR. They concluded that slower rehabilitation does not result in increased long-term stiffness, but 23% (10/43) patients were determined to have postoperative shoulder stiffness at 6 to 8 weeks after surgery. Koo et al. [8] performed primary ARCR in 152 patients and patients with risk factors identified in the previous study (i.e., calcific tendonitis, adhesive capsulitis, partial articular surface tendon avulsion type rotator cuff tear, concomitant labral repair, single-tendon cuff repair) were enrolled in a modified rehabilitation protocol that added early overhead closed-chain passive motion exercises. They reported that no patients experienced postoperative stiffness at a mean of 8-month follow-up compared to a control group (7.8%). However, Galatz et al. [28] found that early motion, even passive motion, may result in devastating consequences. This group reported a high percentage (94.4%) of recurrent defects in patients with early passive rehabilitation after ARCR.
Another potential variable is GHJ synovitis or SAS bursitis [6, 29–31]. Tan et al. [6] reported that the GHJ synovitis score was independently associated with postoperative shoulder stiffness at 3 months after ARCR. Tauro [7] analyzed 72 patients with rotator cuff tears and concomitant preoperative shoulder stiffness who underwent ARCR and also found that bursitis and articular synovitis were more advanced in the group with severe stiffness. However, scores of GHJ synovitis and SAS bursitis were not significantly different between both groups in the present study.
This study has several limitations. First, it is a retrospective analysis, however, our data were collected prospectively by a single research coordinator. The second limitation is the short follow-up. Patients were followed up for 3 months. But, this decision was made since our primary goal was to analyze the incidence and risk factors of early postoperative stiffness. Third, patients in this study were not routinely assessed with postoperative imaging for the presence of cuff re-tear. Nevertheless, the strength of this study is that patients treated with isolated ARCR were included, and patients who underwent additional procedures or preoperative stiffness were excluded. Furthermore, the investigation was performed in a homogenous group of patients who underwent ARCR by a single surgeon, with a relatively large number of cases (274 patients) with various potential risk factors.