Abstract
Background
Humeral head resurfacing (HHR) has emerged as an alternative treatment for glenohumeral osteoarthritis. We investigated the outcomes of HHR using validated patient-reported outcome (PRO) measures.
Methods
A retrospective review was performed on 213 patients who underwent HHR. A PRO follow-up was performed by administering a questionnaire including the American Shoulder and Elbow Society (ASES) score, Brophy activity survey, short form of the Disabilities of the Arm, Shoulder and Hand (quickDASH) survey, and general shoulder function. PRO scores were stratified by comorbidities and complications.
Results
Survey responses were received from 106 patients (51%), with a mean follow-up of 5.6 ± 1.8 years (range: 9 months to 6.1 years). Preoperative comorbidities were associated with significantly higher quickDASH scores. Postoperative complications were associated with significantly higher rates of current pain, higher visual analog scale scores, night pain, lower subjective shoulder values, and lower ASES pain and total scores. No differences in patient satisfaction were identified between the cohorts with and without preoperative comorbidities and between the cohorts with and without postoperative complications.
Conclusion
In our cross-sectional analysis of mid- to long-term outcomes following HHR, preoperative comorbidities, or postoperative complications had no impact on patient-perceived postoperative satisfaction or most PROs. HHR is clinically viable in a wide variety of patients. Future work is necessary to compare the efficacy of HHR compared with more traditional total shoulder arthroplasty and stemmed hemiarthroplasty regarding long-term outcomes and appropriate indications.
Keywords
Introduction
Osteoarthritis of the glenohumeral joint has been traditionally treated with either total shoulder arthroplasty (TSA) or stemmed hemiarthroplasty. As these methods utilize stemmed prostheses to replace the humeral head, they are associated with a more significant risk of transfusion 1 and iatrogenic fracture, 2 as well as more complicated revision. 3 TSA is commonly complicated by glenoid component loosening, 4 while hemiarthroplasties display evidence of progressive glenoid wear.5,6 Humeral head resurfacing (HHR) is an arthroplasty technique that avoids an anatomic neck cut and preserves bone stock.7–9
Although benefits of HHR may include patient-specific reconstruction and preserved bone stock, clinical outcomes following HHR have been poorly described. 10 Available literature is limited by sample size,11,12 and mid- to long-term patient-reported outcomes (PROs) have not been extensively reported. 13 Prior studies suggest that HHR may reduce the risk of humeral fracture while maintaining normal joint anatomy14,15 and reducing eccentric glenoid wear. 16 HHR may represent an attractive option for younger patients and those with isolated humeral head arthritis because of its theoretical lower risk profile. The goal of our study was to evaluate the mid- to long-term PROs following HHR. We hypothesized that patients with preoperative comorbidities and postoperative complications would have significantly lower PRO scores and lower rates of satisfaction following surgery.
Materials and Methods
Study Design
In an institutional review board-approved study, a cross-sectional survey analysis of the mid- to long-term outcomes of HHR at a single large academic institution was performed. All HHR procedures performed at our institution by 4 different sports medicine and shoulder fellowship trained orthopedic surgeons with 5 to 15 years of experience were identified from January 1, 2000 to December 31, 2014. Implants used were Biomet Copeland resurfacing humeral head and Tornier Aequalis HemiCAP resurfacing humeral head. PROs were paired with a retrospective chart analysis performed in 2015 to record demographic, surgical, and complication data.
Participants/Study Subjects
All patients older than 18 years who had an HHR procedure performed from January 1, 2000 to December 31, 2014 for primary (degenerative) osteoarthritis, posttraumatic osteoarthritis, rheumatoid arthritis, or avascular necrosis of the humeral head were included. Exclusion criteria were as follows: age under 18 years, lost to follow-up earlier than 6 months postoperatively, a history of septic arthritis in the operative shoulder, a history of drug-seeking behavior or chronic pain syndrome prior to surgery, or a history of brachial plexus dysfunction or spinal disease requiring conservative or surgical intervention.
Variables and Outcome Measures
All nondeceased patients were mailed an invitation letter informing them about the PRO study, along with a consent form. Recruitment involved a maximum of 3 mailings and 3 follow-up telephone calls per subject. Those willing to participate had the option of completing a PRO questionnaire online via Research Electronic Data Capture (REDCap) or in paper form. Paper responses were copied into the REDCap system upon receipt. The questionnaire was comprised of metrics previously validated in the shoulder arthroplasty population, including the American Shoulder and Elbow Society (ASES) score, 17 the Brophy activity survey, 18 and the short form of the Disabilities of the Arm, Shoulder and Hand (quickDASH) survey. 19 The questionnaire asked respondents for demographic information (occupation, education, disability status) and included questions about their postoperative satisfaction with a symptom-specific and a composite (0%–100%) subjective shoulder value (SSV; Figure 1). For the purpose of this study, the primary determinant of overall procedure “success” was defined as an SSV ≥80%, with a concurrent response of satisfied with current symptoms, which was either “somewhat satisfied” or “very satisfied” to the question “If you had to spend the rest of your life with the symptoms you have now, how would you feel about it?” PRO scores were compared between subjects with or without preoperative comorbidities (body mass index ≥40, diabetes mellitus, smoker, prior rotator cuff dysfunction, such as prior rotator cuff tear, partial rotator cuff tear, and impingement symptoms, prior nonarthroplasty surgery, or autoimmune/rheumatologic disease) as well as between subjects with or without postoperative complications. Complications were defined as postoperative pain for any reason, extended narcotic needs, infections, wound complications, a return to the operating room, and the need for revision arthroplasty.
Questionnaires included survey questions about postoperative satisfaction and current symptom complaints.
Accounting for All Study Subjects
A total of 213 patients met our inclusion/exclusion criteria. Of these, 6 were deceased. Survey responses were received from 106 patients (112 shoulders) for a response rate of 51.2% (Figure 2). Mean follow-up was 5.6 ± 1.8 years, with a range of 9 months to 6.1 years. When compared with the non-PRO responders, the 106 subjects of the PRO study cohort were not significantly different in terms of demographics, preoperative comorbidities, complications, and surgical characteristics (Table 1).
Two hundred fifteen patients underwent HHR between 2000 and 2014, with 213 eligible for the study based on inclusion and exclusion criteria. After recruitment, 106 patients were included in the PRO analysis. HHR, humeral head resurfacing; PRO, patient-reported outcome.
Demographics, Preoperative Comorbidities, and Shoulder Characteristics of Subjects Who Did Not Response and Responded to the PRO Survey Are Shown.
Abbreviations: BMI, body mass index; CAD, coronary artery disease; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; PRO, patient-reported outcome.
Statistical Analysis—Study Size
Statistical analysis was performed by the investigators using GraphPad Prism 7.0 (LaJolla, CA). Normality in all cases was confirmed using the Kolmogorov–Smirnov test. Categorical variables were compared with Fisher’s exact test, while Student’s t test was used for continuous variables. Significance was defined in all cases as P < .05.
Results
Preoperative comorbidities were present in 56 of 106 subjects and were associated with significantly higher quickDASH scores (48.0 ± 8.1 vs 42.9 ± 11.9, P = .040; Table 2). However, preoperative comorbidities were not associated with significant differences in rates of current pain, narcotic use, ASES scores, or Brophy scores. Postoperative complications occurred in 30 of 106 subjects. Heavy laborers were more likely to have a postoperative complication. Impingement-related pain was the most prevalent source of complication, occurring in 26 (24.5%) patients. Four patients returned to the operating room: 2 for adhesive capsulitis, 1 for hematoma evacuation, and 1 for an acute supraspinatus tear 9 months postoperatively. No patient required a revision arthroplasty at the time of final follow-up. Postoperative complications were associated with significantly higher rates of current pain (86.2% vs 50.0%, P < .001), higher visual analog scale (VAS) pain ratings (3.6 ± 2.6 vs 1.9 ± 2.4, P = .002), higher rates of night pain (63.6% vs 38.6%, P = .02), ASES total scores (44.7 ± 25.9 vs 55.0 ± 22.3, P = .04), and SSV (65.8 ± 26.7% vs 75.9 ± 21.7%, P = .04; Table 2). Postoperative complications were not associated with significant differences in the rate of narcotic use, ASES function scores, quickDASH scores, or Brophy scores.
Patient-Reported Outcome Scores for Subjects With and Without Preoperative Comorbidities and Postoperative Complications Are Shown.
Abbreviations: ASES, American Shoulder and Elbow Society; DASH, Disabilities of the Arm, Shoulder and Hand; VAS, visual analog scale. P < .05 are bolded.
Preoperative comorbidities were not associated with significant differences in overall patient-reported satisfaction with current symptoms (73.8% vs 69.5%, P = .687) or rates of procedure “success,” defined as a self-reported SSV above 80% and a satisfaction with current symptoms as “somewhat satisfied” or “very satisfied” (67.9% vs 64.0%, P = .678; Table 2). The majority of patients felt that their symptoms improved after surgery, whether they had preoperative comorbidities (78.7% vs 74.6%, P = .668) or postoperative complications (72.7% vs 78.2%, P = .630). Postoperative complications were also not associated with differences in satisfaction with current symptoms (66.7% vs 73.6%, P = .500) or procedure success (56.7% vs 69.7%, P = .256). Of note, although the mean SSV for each groups were lower than 80, the majority of the patients (70 of the 106) reported an SSV of 80 or higher. The average value was skewed by the few patients with very low SSVs—13 of the 106 patients had an SSV less than the standard deviation below the average SSV of the entire cohort.
Discussion
Our study with a mean follow-up of 5.6 years found that neither preoperative comorbidities nor postoperative complications impact overall patient-perceived satisfaction or most other PRO measures following HHR. These findings suggest that HHR may be a viable alternative to more extensive arthroplasty procedures for a broad range of patients. HHR is an alternative to stemmed hemiarthroplasty for the treatment of glenohumeral and isolated humeral head arthritis. 3 There is a paucity of prior research evaluating long-term outcomes following HHR, and the existing evidence appears to be inconsistent. HHR has been associated with high revision rates 20 and less accurate humeral head positioning when compared with hemiarthroplasty. 11 Additional prior work recommends HHR only for concentric osteoarthritis without glenoid erosion due to high failure rates from glenoid wear. 21 Although HHR has been thought to preserve bone stock, permitting a later revision if needed, a loss of bone volume was noted below the HHR implant, 22 and revisions after HHR had lower Western Ontario Osteoarthritis of the Shoulder Index scores. 23 It has been proposed that HHR results in poor postoperative range of motion in those with preexisting rotator cuff injury, 24 although various studies have shown that pain relief, 20 functional improvement, patient satisfaction, 25 and improved range of motion were noted postoperatively.20,26 A major limitation of current HHR literature is small sample sizes. The largest recent study evaluated intraoperative and postoperative complications in patients undergoing all forms of shoulder arthroplasty (n = 12 559) and found that patients undergoing HHR (n = 2329) had fewer complications than those undergoing TSA, and that the superior approach resulted in fewer complications than the deltopectoral approach. 27 Two other large registries have been used to look at outcomes of HHR and focus primarily on outcomes from revision surgery (n = 1210), 23 as well as patient outcome comparison among resurfacing prosthesis (n = 195), hemiprothesis (n = 408), total prosthesis (n = 86), and reverse total prosthesis (n = 170). 28 Although there are several recent studies with longer follow-up, ours is the largest, aside from the 3 previously mentioned, to examine PROs following HHR (Table 3).
Sample Size and Follow-up Time Are Shown for Recently Published HHR Studies.
Abbreviations: HHR, humeral head resurfacing; NA, not applicable.
In our cohort, patients who experienced a postoperative complication did not report a significant difference in PROs with the exception of pain and had significantly higher rates of current pain, night pain, and higher VAS pain severity. This is also reflected in worse ASES pain and ASES total scores. However, despite a high 50% prevalence of some level of pain following HHR in our entire cohort, ASES scores are lower than most recent studies (Table 4).14,25,29 Heavy laborers were also more likely to have a postoperative complication. This may be useful information to consider when evaluating the clinical utility of HHR in this population, especially since HHR is often considered in younger, more active patients.
Overall Patient-Reported Outcome Scores for All Subjects Are Shown, Along With Available Values From Prior Studies.
Abbreviations: ASES, American Shoulder and Elbow Society; DASH, Disabilities of the Arm, Shoulder and Hand; VAS, visual analog scale.
Although prior work has mainly focused on objective clinical outcomes following HHR, such as range of motion, radiographic measurements (fracture, component loosening, osteolysis), complications,11,13 and osseous integration, 30 our primary definition of procedure success was long-term patient-reported satisfaction. Patient satisfaction and whether they would have the procedure again was not significantly different for patients with preoperative comorbidities or postoperative complications, or even those who required reoperation. In addition, no difference was observed between “successful” and “unsuccessful” procedures. On the whole, 71.7% of patients were satisfied with their current symptoms, while the “success” criteria were met in 66.0% of cases. Our findings suggest that while complications result in higher rates of pain postoperatively, this pain is mild and does not significantly impact shoulder function or patient satisfaction. Patient satisfaction following TSA has been reported in several recent studies, 4 but few HHR studies include satisfaction as an outcome measure. Levy et al. 7 reported that 81.6% of patients felt better or much better from preoperatively, and Alizadehkhaiyat et al. 8 reported that 85% were satisfied with their results, which are consistent with our result of 76.7% (Table 4). Sweet et al. 25 and Pritchett 31 reported 90% and 95% patient satisfaction with the procedure, respectively, compared to our 71.7% satisfaction with current symptoms, although definitions of satisfaction are inconsistent across studies. Although objective data are certainly important, patient-reported satisfaction may be a more clinically relevant outcome measure. This is further supported by our study, as satisfaction and most subjective outcome measures were not influenced by rates of complication. A recent study by Verstraelen et al. demonstrated a similar concept, showing that despite poor radiographic outcomes at a mean of 7.2 years after HHR, clinical outcomes were good. 32
Our study has several limitations. First, patients were given the option of completing the PRO survey electronically or in paper form. Although this allowed for a better response rate and potentially a more representative sample, ambiguous responses on the paper survey had to be interpreted and converted into the electronic form. To provide consistency in interpretation, a single author performed this task for all paper surveys. Second, the design of our study is inherently subject to response bias, especially with 51.2% survey response rate, which can be considered low. However, this response rate is quite comparable to other survey-based studies. A systematic review of 219 studies on mail surveys published in medical journals reports an average response rate of around 60%. 33 In addition, comparison of the respondent subset with the nonresponders showed no significant differences in demographics or patient health or shoulder characteristics, indicating that the sample of patients in this study is representative of the entire population of patients undergoing HHR at our institution over the past 15 years (Table 1). However, the low follow-up still remains a weakness, as patients who needed revision or conversion of their arthroplasty may have been more likely to seek care elsewhere and thus higher in the lost to follow-up group. Third, our study did not include an analysis of objective data, such as pre- and postoperative range of motion and pre- and postoperative radiographic measures, as we were limited by inconsistent reporting in the chart and the lack of consistent advanced imaging postoperatively. Finally, a thorough review of the electronic medical record was used to identify preoperative comorbidities and postoperative complications. However, electronic medical records are not always accurate and/or thorough and preoperative comorbidities may have been missed. In addition, complications may have been missed if a patient was lost to follow-up. With more complete collection of complications, it would have been valuable to further stratify patients to different magnitudes of complications and compare the outcome measures.
Conclusion
In conclusion, in our cross-sectional analysis of mid- to long-term outcomes following HHR, preoperative comorbidities or postoperative complications had no impact on patient-perceived postoperative satisfaction or most PROs. HHR is clinically viable in a wide variety of patients and can be considered in patients particularly with unipolar humeral disease or minimal glenoid disease, and younger patients with a desire to preserve bone stock. Future work is necessary to compare the efficacy of HHR compared with more traditional TSA and stemmed hemiarthroplasty regarding long-term outcomes and appropriate indications.Acknowledgments
The authors thank Dr Gregory V Gasbarro for assistance with data collection as well as Dr Caiyan Zhang for assistance with statistical analysis.
Footnotes
Ethical Statement
All studies were performed in accordance with the ethical standards in the 1964 Declaration of Helsinki.
Ethical Review
This study was the product of the University of Pittsburgh IRB PRO14090251—“Evaluation of Clinical Outcomes and Complications of HemiCAP Humeral Head Resurfacing.”
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: The senior author (Albert Lin) is a consultant for Arthrex and Wright/Tornier.Each author certifies that he or she has no commercial associations (eg, consultancies, stock ownership, equity interest, patent/licensing arrangements) that might post a conflict of interest in connection with the submitted article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by an Albert Ferguson Orthopaedic Research Grant (FY 2015) administered by the Department of Orthopaedic Surgery at the University of Pittsburgh.