Safety and Effectiveness of Intra-articular Corticosteroid Injection Following Arthroscopic Shoulder Surgery


 Background

To evaluate the safety and effectiveness of intra-articular corticosteroid injection following arthroscopic shoulder surgery.
Method

PubMed, Cochrane Library, EMBASE, Web of Science and CNKI were retrieved from the inception of electronic databases to June 2020. All analyses were performed using Stata/SE 15.1 version (StataCorp).
Result

6 papers were included in this meta-analysis. There was no significant difference in tear rate (OR = 0.713, 95% CI: 0.450 to 1.129, P = 0.149), constant score (MD = 6.212, 95% CI: 2.552 to 9.971, P = 0.001), rating scale of the American shoulder and elbow surgeons (ASES) score (MD=-0.116, 95% CI: -1.769 to 1.546) and the university of California at Los Angeles shoulder rating (UCLA) score (MD = -1.461, 95% CI: -3.221 to 0.299). The infection rate of patients who received corticosteroid injection within 1 month after operation was significantly higher (P < 0.05), but there was no significant change in the infection rate of patients who received corticosteroid injection within 2–4 months after operation.
Conclusion

The use of corticosteroids after shoulder arthroscopy will not increase the rate of postoperative tears, but the injection of corticosteroid within 1 month will increase the postoperative infection rate.


Introduction
Adhesive shoulder arthritis, calci ed supraspinatus tendonitis, shoulder synovitis and shoulder osteoarthritis could cause irreversible shoulder joint damage with the main clinical manifestations of limited mobility and pain [1][2][3][4]. Although conservative treatment and open surgery have been the main treatment methods for these diseases, with the popularization of arthroscopy technology, it has become one of the main treatment option [5,6]. The main application advantages of shoulder arthroscopy are less injury of shoulder joint, su cient removal of focus, low postoperative pain and fast recovery [7]. It is estimated that more than 270,000 patients underwent rotator cuff repair in 2006 in the United States and more than 250,000 patients completed rotator cuff repair under shoulder arthroscopy [8]. However, despite the replacement of surgical instruments and the increasingly sophisticated technique of the operators, patients often have severe postoperative pain and shoulder stiffness after arthroscopic rotator cuff repair, which is closely related to postoperative rotator cuff tear and poor functional recovery [9,10].
Local corticosteroid blocking therapy has become one of the vital treatments for in ammatory shoulder pain because of reducing synovial in ammation, accelerating early functional recovery after surgery and is widely accepted by orthopedists [11]. Nicholas [12]compared the effect of two groups with highgrade, partial-thickness rotator cuff tears based on whether they received a subacromial corticosteroid injections or not and showed that the injection subgroup experienced a signi cant increase in rating scale of the American shoulder and elbow surgeons (ASES) and Constant score. However, there are still concerns about whether corticosteroid use will increase the incidence of infection, early tendon rupture and delayed tendon healing after arthroscopy [13,14]. It is worth noting that 50,478 shoulder arthroscopy patients, of whom 4115 received injections in the 6-month preoperative period was accompanied with a signi cant increase in both the overall infection rate and severe infection rate in patients who received corticosteroid injections within 2 weeks.
Meanwhile,some researchers believe that timing and dose of injections prior to arthroscopic rotator cuff repair impacts the risk of infection [15]. Therefore, there is still great controversy regarding the safety and clinical e cacy of intra-articular corticosteroid injection after shoulder arthroscopy. The purpose of this research is to explore the safety and clinical e cacy of intra-articular corticosteroid injection after shoulder arthroscopy, with a view to providing clinical guidance.

Literature and search strategy
The retrieved object is the research literature on the analysis of safety or effectiveness of corticosteroid injections in patients with arthroscopic shoulder surgery published publicly in the electronic databases including PubMed, Cochrane Library, EMBASE, and Web of Science from the inception of electronic databases to June 2020. We retrieved the following keywords in combination with Boolean logic: "arthroscopy" "shoulder Injuries" "rotator cuff injuries" "shoulder impingement syndrome" "corticosteroids" "steroids" "triamcinolone" "betamethasone" "dexamethasone". Beyond that, the research of the appraisal reference list was manually checked to determine other potential quali cation trials. The process iterates until no more articles could be determined. The metaanalysis was based on acknowledged PRISMA guideline (the prioritized reported items for systematic review and meta-analysis) [16].

Inclusion And Exclusion Criteria
The articles will be incorporated into the present meta-analysis if the literatures meet the following principles in accordance with PICOS. Population: patients with de nite diagnosis of shoulder disease such as adhesive shoulder arthritis, calci ed supraspinatus tendonitis or rotator cuff tear; Intervention: patients receiving arthroscopic shoulder surgery; Comparison: the safety and effectiveness of patients with or without corticosteroid injection; Outcome measures: one or more adequate data of the outcomes could be conducted statistical analysis; Study design: an o cial published RCT or RCS. Exclusion criteria:(I) Non-English written literature (II) abstracts, letters, editorials, expert opinions, case reports, review and basic research including animal and cell experiments (III) non-comparative study (IV) inadequate raw data.

Data Extraction And Outcome Measures
Two of the reviewers (Wen-chen Lu, Ting-jiang Wang) respectively extracted data from the included studies. The following essential information was captured: rst author names, publication year, samples size, study design, outcomes and other relevant data. The extracted data [median, range, mean difference, 95% con dence interval (CI) and standard deviation (SD)] is input into the designed standardized table. When there are differences of opinion, another authority author has the nal decision. The outcome measurements were tear rate, postoperative infection, constant score, ASES score and the university of California at Los Angeles shoulder rating (UCLA) score.

Quality Assessment And Statistical Analysis
Newcastle-Ottawa Scale (NOS) and Risk Bias in Non-randomized intervention studies (ROBINS-I) were respectively conducted to evaluate the methodological quality of the included RCTs and RCS. The literature quality evaluation was conducted separately by two reviewers (Wen-chen Lu, Zhi-hong Tang). Consensus was reached through consultation for divergence. The GRADEpro software was used for evidence quality assessment. We use the Stata 15.1 version (Stata Corporation, College Station, Texas, USA) for statistical analyses. When I 2 50%, the data was considered as obvious heterogeneity. We conduct a metaanalysis using random-effect model according to Cochrane Handbook for Systematic Reviews of Interventions (version 5.1.0). Otherwise, xed-effect model was performed. For continuous outcomes (constant score, ASES score and UCLA score), Weighted Mean Difference (WMD) were expressed for the evaluation.
For discontinuous various outcomes (tear rate, infection), Odds Ratio (OR) were applied for the assessment.

Search results
A total of 895 studies were identi ed as potentially relevant literature reports. There were no additional studies identi ed through other sources. We got 570 articles when the duplicate was removed. By scanning the title and abstract, 500 studies were excluded according to the eligibility criteria. Another 64 articles were further excluded by reading the full text. Ultimately, 6 articles [17][18][19][20][21][22] were eligible for data extraction and meta-analysis. The searching process is shown in Fig. 1.

Characteristics Of Included Studies
The characteristics of the 6 included studies are summarized in Table 1. 7682 individuals were incorporated into our trial of whom 4184 patients with corticosteroid injection and 3498 patients with corticosteroid injection. Among them, 5 articles were RCS, and 1 article was a RCT. 4 studies from Korea, 1 study from the USA and 1 study from France.

Study Quality And Risk Of Bias
The quality of the included studies was evaluated by the scale of ROBINS-I. The details were presented in Table 2.

Methodological Quality Assessment
The total qualities of the evidence were low for the tear rate, postoperative infection, constant score, ASES score and UCLA score (  Fig. 2).
A total of three papers mentioned postoperative complications and two of them claimed that their study had no complications. Only a review of one large database including 3946 patients reported the relevant infection data. Due to the lack of su cient data, a descriptive analysis of postoperative infection rates will be used. KEW [19]compared the incidence of postoperation infection of patients undergoing shoulder arthroscopy operation between two groups. Finally, they found that the infection rate of patients who received corticosteroid injection within 1 month after operation was signi cantly higher (P < 0.05), but there was no signi cant change in the infection rate of patients who received corticosteroid injection within 2-4 months after operation (P > 0.05).

Discussion
Oral nonsteroidal anti-in ammatory analgesic drugs, intra-articular corticosteroid injection, physical therapy and joint capsule release are used to treat shoulder stiffness and pain after repair surgery for rotator cuff injuries [23,24]. It was found that compared with hyaluronic acid, broblast proliferation was lower and the rate of healing failure was higher when the corticosteroids were injected at the surgical site in the studies in animal repair models of rotator cuff injury [25,26]. In addition, human tissue cytology studies have shown that corticosteroid treatment is associated with higher cell apoptosis at the surgical site [27]. Although some basic studies have shown that corticosteroids may have a negative effect on postoperative tissue healing, some clinical studies have shown that early intraarticular injection of corticosteroids after arthroscopic rotator cuff repair does not increase the tear rate [21,28]. Our research demonstrated that there was no signi cant difference in tear rate, constant score, ASES score and UCLA score. However, corticosteroids injection within 1 month after arthroscopic shoulder surgery will signi cantly increase the early infection rate.
The results of this meta-analysis showed that the corticosteroids injection group had no signi cant difference in tear rate compared with the control group, which is inconsistent with the research of BAVEREL [21]. We consider that different characteristics of the included patients may be one of the potential reasons. only severely injured patients with rotator cuff full-thickness tearing were included in BAVEREL's research [21]. In this study, patients with partial or fullthickness rotator cuff tears and a mixture of the two were included. The differences of the study subjects may lead to inconsistent conclusions. Secondly, betamethasone is the intervention of the former study, while triamcinolone acetonide is the main intervention of this study. To some extent, the difference of corticosteroid types will affect the conclusion.
In this research, 2 included articles mentioned that no postoperative adverse reactions were found after corticosteroids administration. The most likely reason was that the sample size was relatively small. KEW [19]compared the incidence of infection of 3946 patients undergoing shoulder arthroscopy operation between two groups. Finally, they found that the infection rate of patients who received corticosteroid injection within 1 month after operation was signi cantly higher, but there was no signi cant change in the infection rate of patients who received corticosteroid injection within 2-4 months after operation and this result is in line with previous studies [29,30]. Postoperative infections are caused when pathogens enter the body during surgery or corticosteroid injection. The rst month after the operation is considered as the in ammatory response period. Due to the immunosuppressive effect of corticosteroid, the ability of bacteria resistance declines causing infection.
Previous studies have reported that long-term use of corticosteroid will increase the incidence of complications within 30 days after arthroscopic shoulder surgery [29][30][31]. MARTIN [31]reported the incidence of corticosteroid relevant complications such as reoperation (0.31%), super cial infection of the surgical site (0.16%), deep infection (0.01%), deep vein thrombosis or thrombophlebitis (0.09%), peripheral nerve injury (0.01%), pulmonary embolism (0.06%) in a study involving 9410 patients. HEYER [29] found that in addition to the long-term use of corticosteroid, patients older than 65 years old, male patients, ASA rating greater than level 2, history of chronic obstructive pulmonary disease, hypertension and operation time more than 90 minutes were all related to the increased incidence of adverse reactions.
The shortcomings of the research: First, only one of the included literatures is an RCT, which may be one of the sources of heterogeneity. Second, all the literatures are short-term observation studies with a follow-up time of no more than 3 years. This conclusion should not be applied to long-term observation results.
The use of corticosteroids after shoulder arthroscopy will not increase the rate of postoperative tears, but the injection of corticosteroid within 1 month will increase the postoperative infection rate.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
Not applicable.

Competing interests
The authors declare that they have no competing interests.

Funding
Not applicable.

Authors' contributions
Wen-chen Lu conceived of the design of the study. Ting-jiang Wang performed and collected the data and contributed to the design of the study. Zhi-hong Tang prepared and revised the manuscript. All authors read and approved the nal content of the manuscript.   Figure 1 Flowchart of the study selection process     PRISMAChecklist.doc