Introduction

Background and rationale

Total shoulder arthroplasty is a widely used surgical treatment for glenohumeral arthritis. Glenoid component failure has been identified as the major mechanical cause of failure of total shoulder arthroplasty [1,2,3,4,5,6,7,8,9,10]. A recent review indicates that this is still the case: glenoid loosening accounted for 38% of all total shoulder complications [11]. In an attempt to address the high rate of glenoid failure, new glenoid components are being submitted and cleared by the U.S. Food and Drug Administration’s 510 (k) process each year (Fig. 1). However, in spite of the introduction of these new designs, a recent analysis failed to show evidence of substantially improving outcomes for total shoulders over the last two decades [12]. It is difficult to know whether the new designs of glenoid components are yielding results that exceed the results of components that have been in widespread international usage for many years.

Fig. 1
figure 1

Cumulative number by year of United States Food and Drug Administration 510(k) clearances for glenoid components or total shoulder arthroplasty systems including a glenoid component

Full size image

One of the commonly used glenoid designs is an all-polyethylene component that is fixed to bone with cemented peripheral pegs and an uncemented fluted central peg. The outcomes for patients receiving this glenoid component have been the subject of recent reports by surgeons from different countries [13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31]; however, the patients in these articles have not been analyzed together as a cohort of individuals. We contacted the authors of these publications requesting their most recent data on each patient having a total shoulder using this glenoid design so that we could perform a “by patient” analysis in which the clinical outcome for each patient was weighted individually.

We hypothesized that a great majority of patients with glenohumeral arthritis from 11 independent centers using this design of glenoid component would achieve clinically significant improvement following total shoulder arthroplasty assessed by generally accepted outcome instruments.

Methods

This is a multicenter retrospective observational study approved by our Institutional Review Board (HSD# STUDY00001714).

Study design and participants

A literature search was performed to identify institutions that had clinical outcome data for total shoulder arthroplasties that used one of two very similar designs of glenoid component: a DePuy Anchor Peg (DePuy, Warsaw, IN, USA) (Fig. 2) or a Wright Medical Perform Cortiloc (Wright Medical, Memphis, TN, USA) fluted central peg glenoid (Fig. 3). The corresponding authors were contacted and invited to participate by submitting de-identified data on each patient to this study.

Fig. 2
figure 2

The Anchor Peg Glenoid Component, DePuy Synthes, 325 Paramount Dr., Raynham, MA 02767

Full size image
Fig. 3
figure 3

Wright Medical Perform Cortiloc fluted central peg glenoid, Wright Medical, 1023 Cherry Road, Memphis, TN 38117

Full size image

Importantly, this investigation represented the variety of practice preferences of the participating surgeons using these glenoid components. There was no attempt to standardize patient evaluation and management, except for the design of glenoid component used. Each surgeon applied his own approach to preoperative radiographic evaluation (plain films, MRI or CT scan) and each applied his own indications and technique for total shoulder arthroplasty using this glenoid component design. Each participating surgeon completed a standardized data sheet including age at surgery, sex, year of surgery, diagnosis, preoperative glenoid type [32,33,34], glenoid retroversion, prior surgery, humeral component type, glenoid component, and humeral and glenoid component articular surface diameters of curvature. Each center used one or more of the following validated outcome scores to evaluate the clinical results of the arthroplasty: Simple Shoulder Test (SST), American Shoulder and Elbow Score (ASES), Constant Score (CS), or Penn Score.

Subjects were included in our analysis if they were between ages 18–99, had a shoulder arthroplasty procedure performed for arthritis with an all-polyethylene fluted central peg glenoid component between 2000 and 2016, and had functional outcome scores preoperatively and at 1 or 2 years after surgery.

The clinical significance of the preoperative to postoperative improvement in each clinical score was assessed in two ways. First, the published value for the minimal clinically important difference (MCID) for each outcome scale was obtained from the literature: the Simple Shoulder Test (1.5), the American Shoulder Elbow Surgeons Score (13.6), the Constant Score (5.7), and the Penn Score (11.4) [35,36,37]. Each patient’s scores were examined to determine if the improvement exceeded the MCID. Second, the maximal possible improvement for each patient was determined as the difference between the maximal possible value for the outcome scale and the patient’s preoperative score. The amount of improvement achieved by each patient was then divided by the maximal possible improvement to obtain the percent of maximal possible improvement (%MPI)—clinically significant improvement has been defined as improvement of at least 30% of the maximal possible improvement [12, 27, 38,39,40,41,42].

One- and 2-year outcomes were characterized as the amount of improvement on each scale in comparison to the MCID and as the percentage of maximum possible improvement (%MPI). To enable comparison among the different scales, the 12-point Simple Shoulder Test was rescaled to a 0–100 scale [36]. Associations between risk factors and outcomes were estimated using multiple linear regression of the outcomes on each risk factor, adjusting for the practice site and the baseline value of the outcome score.

All calculations were carried out by an experienced statistician in R version 3.4.3 (Vienna, Austria) [42].

Results

In all, 1270 individual patients from 11 centers met the criteria for inclusion (Table 1). The typical patient had osteoarthritis, was 66 years of age at the time of surgery, and had surgery in 2011. The glenoid types reported by the surgeons were A1 (28.5%), A2 (29.2%), B1 (15.3%), and B2 (26.1%). Preoperative glenoid version averaged 12 ± 9°; 30% had retroversion greater than 15°. Seventy percent of the humeral components were standard stems, 18% short stems, and 12% stemless. The Anchor Peg glenoid was used in 65% and the Cortiloc glenoid in 35%. The average diameter of curvature of the humeral component was 49 ± 4 mm and that of the glenoid was 51 ± 6 mm. Preoperative and postoperative function was assessed by the Simple Shoulder Test in 654, the ASES score in 757, the Constant score in 446, and the PENN score in 16 (note that some patients were assessed by multiple scores).

Table 1 Descriptive statistics for preoperative characteristics
Full size table

Twenty-five shoulders (2%) were reported to have had a second surgical procedure following their initial total shoulder arthroplasty. The types of second surgery were not further detailed in this study of clinical outcomes. With the exception of patient age, no preoperative patient or shoulder characteristics were significantly associated with the need for a second surgery. The odds ratio for revision by age at surgery (per 10 years) was 0.62 (95% CI 0.44–0.90, p = 0.014).

The great majority of these patients improved by more than the minimal clinically important difference and improved by more than 30% of the maximal possible improvement (MPI) (Table 2). The mean ± SD preoperative scores improved from SST 3 ± 2, ASES 37 ± 15, Constant score 36 ± 16, and Penn score 30 ± 19 to SST 10 ± 2, ASES 90 ± 12, Constant 76 ± 13, and Penn 80 ± 24 (p < 0.001 for each). For patients assessed with the SST, over 90% exceeded the MCID of 1.5 and improved by 30% of the MPI. For patients assessed with the ASES score, over 95% exceeded the MCID of 13.6 and improved by over 30% of the MPI. For patients assessed with the Constant score, over 90% exceeded the MCID of 5.7 and improved by over 30% of the MPI. Finally, for patients assessed with the Penn score, over 85% exceeded the MCID of 11.4 and improved by over 30% of the MPI.

Table 2 Descriptive statistics: preoperative, 1-year postoperative, and 2-year postoperative for four different outcome scales
Full size table

For patients with data at both 1 and 2 years after surgery, there was little clinical improvement in outcome scores between year 1 and 2, even though some of the differences were statistically significant (Fig. 4).

Fig. 4
figure 4

Average SST, ASES, and Constant scores before and at 1 and 2 years after surgery for patients receiving a standard all-polyethylene glenoid component with a fluted central peg. Vertical axis indicates the percentage of a perfect score. Data for the Penn Score are not shown because of the small number of patients assessed with this score

Full size image

These outcomes for the standard, non-augmented glenoid component were realized in cases with a range of patient and shoulder characteristics (Table 1), including the 41% of cases that had type B glenoids and the 30% of cases that had more than 15° of glenoid retroversion. The outcomes for shoulders with retroverted or type B glenoids were not inferior to those with neutral version or type A glenoids, respectively (Appendix Tables 4 and 5). The clinical outcomes were not consistently associated with the degree of mismatch between the humeral and glenoid diameters of curvature [43, 44].

A comparison of the percent of maximal possible improvement for the SST, ASES, and Constant scores for ten different sites is shown in Table 3. Each scale at each site showed values for the average percent of maximal possible improvement that easily exceeded the threshold of 30%.

Table 3 % MPI (mean ± SD) by outcome scale for the 10 sites using ASES, Constant, and SST scales
Full size table

Discussion

This study design is unique in that it analyzes the early total shoulder clinical outcomes at the discrete time points of 1 and 2 years for 1270 individual patients from 11 independent international centers using a standard design all-polyethylene glenoid component with a fluted central peg and three cemented peripheral pegs.

As assessed by generally accepted outcome instruments, the average improvement in comfort and function achieved by these individual patients easily exceeded the published values for the minimal clinically important difference and provided well over 30% of the maximum possible improvement.

For patients having results at both the year 1 and year 2 time points, the improvement after the first year was small: 0.2 ± 1.5 for the SST, 0.8 ± 8.9 for the ASES score, and 0.5 ± 8.0 for the Constant score. This observation suggests that most of the clinical improvement after a total shoulder arthroplasty occurs within the first year; therefore, year 1 data may be sufficient for characterizing the early outcomes for total shoulder arthroplasty with different glenoid component designs [36].

An interesting finding in this analysis was that with this basic glenoid component design, type B glenoids and glenoids with more than 15° of retroversion did not have outcomes that were inferior to those with less severe glenoid pathoanatomy. While special glenoid components have been designed for the type B retroverted glenoid [26, 45, 46], our study and other recent publications [27, 39, 47, 48] point out that the role for these special components has yet to be clarified.

The clinical outcomes reported were qualitatively similar among the different scoring systems used by the different centers. The application of published values for the MCID and the use of the percent of maximal possible improvement offer the possibility of comparing outcomes obtained with different outcome instruments.

The results of this study should be considered in light of certain limitations. First, this study focused only on the improvement in standard, validated, widely used, and universally available clinical outcome scales—Simple Shoulder Test, American Shoulder and Elbow score, Constant Score, and Penn Score—each of which has a defined MCID. Range of motion estimates and radiographic interpretations by the surgeons were not included in this analysis. Second, this study focused on the clinical improvement at discrete time points: 1 and 2 years after surgery, rather than mixing outcomes recorded over a range of years. While both 1 and 2-year follow-up data were not available for all patients, the results available at both of the two time points were quite similar [36]. Third, while most of the variables we assessed did not have a consistent, statistically significant effect on the outcome, some of these effects may have become significant if the number of patients was larger. Fourth, there was no attempt on our part to standardize the indications for surgery, the preoperative clinical or radiographic evaluation, the choice of humeral component, the surgical technique, or the postoperative rehabilitation program; instead, the surgeons treated each of their patients according to their personal practice guidelines. The consistency of the outcomes among these varied practices indicates the general utility of this design of glenoid component. Fifth, this study did not investigate component revision and survivorship, which require follow-up of 5 years or longer to obtain meaningful data [9, 22, 49, 50]. Finally, the surgeons whose patients were included in this analysis each had substantial experience in shoulder arthroplasty; the outcomes reported here might not be generalizable to the practices of surgeons with less experience [47, 51,52,53,54,55,56].

This report provides a broad-based and well-characterized data set against which the 1- and 2-year clinical outcomes for different glenoid components can be compared using universally available assessment scales (SST, ASES, Constant, Penn). Such comparisons are becoming increasing necessary and timely: new shoulder arthroplasty systems are being introduced annually without evidence that their clinical outcomes are superior to or even equal to those in current use [12]. Thus, a strong case exists for establishing a reference of 1- and 2-year clinical outcomes for standard glenoid components—such as that presented in this report—against which the early performance of new components can be compared. The results of this study suggest a possible benchmark for shoulders having pre and postoperative SST, ASES, Constant, or Penn scores: at 1 and 2 years after total shoulder arthroplasty, 90% of the shoulders are improved by the minimal clinically important difference and are improved by 30% of the maximal possible improvement.

Conclusion

The strength of this study lies in its demonstration that surgeons in 11 independent practices using their individual approaches to patient selection, preoperative evaluation, surgical technique, and outcome evaluation were able to obtain robust early clinical outcomes in an international group of over 1200 individual patients using a basic all-polyethylene glenoid component. These data provide a basis for comparison with newer designs to determine whether these new designs lead to better early clinical outcomes.