Lower-extremity total joint arthroplasty in shoulder arthroplasty patients: does the order of the lower-extremity total joint arthroplasty matter?

J Shoulder Elbow Surg (2019) -, 1–7

www.elsevier.com/locate/ymse

Lower-extremity total joint arthroplasty in shoulder arthroplasty patients: does the order of the lower-extremity total joint arthroplasty matter? Heather A. Prentice, PhDa,*, Priscilla H. Chan, MSa, Mark T. Dillon, MDb, Nithin C. Reddy, MDc, Ronald A. Navarro, MDd, Elizabeth W. Paxton, PhDa a

Surgical Outcomes and Analysis, Kaiser Permanente, San Diego, CA, USA Department of Orthopaedic Surgery, The Permanente Medical Group, Sacramento, CA, USA c Department of Orthopaedic Surgery, Southern California Permanente Medical Group, San Diego, CA, USA d Department of Orthopaedic Surgery, Southern California Permanente Medical Group, Harbor City, CA, USA b

Background: As total joint arthroplasty (TJA) utilization increases, arthroplasties of multiple joints in a patient are more common. An understanding of the success of shoulder arthroplasty patients also requiring a lower-extremity (hip or knee) TJA is lacking. We evaluated the following questions: (1) Is there a difference in the revision risk following shoulder arthroplasty in patients who also undergo a lower-extremity TJA compared with those who do not? (2) Does the revision risk differ depending on the sequence of the procedures? Methods: Patients who underwent elective primary shoulder arthroplasty from 2009 through 2015 were identified using Kaiser Permanente’s shoulder arthroplasty registry. Patients with a lower-extremity TJA were identified using the institution’s total joint replacement registry. Revision related to the index shoulder was modeled via Cox regression stratified by procedure type and adjusted for confounders. Results: Of the 4751 shoulder arthroplasties identified, 1285 (27.0%) underwent a prior hip and/or knee arthroplasty and 483 (10.2%) underwent a hip and/or knee arthroplasty following the index shoulder arthroplasty. No difference was found in all-cause shoulder revision risk with lower-extremity TJA before (hazard ratio, 1.38; 95% confidence interval, 0.97-1.96) or after (hazard ratio, 1.30; 95% confidence interval, 0.82-2.06) the index shoulder arthroplasty compared with patients who underwent a shoulder arthroplasty only. Conclusion: In our study sample, we did not observe shoulder revision surgery risk to be different in patients who also underwent a lower-extremity TJA, regardless of the sequence of the 2 procedures. Future prospective studies should investigate whether the timing of the lower-extremity TJA in relation to the shoulder procedure impacts the latter’s success.

Institutional review board approval from Kaiser Permanente was obtained prior to the beginning of the study.

*Reprint requests: Heather A. Prentice, PhD, Surgical Outcomes and Analysis, Kaiser Permanente, 8954 Rio San Diego Dr, Ste 406, San Diego, CA 92108, USA. E-mail address: [email protected] (H.A. Prentice).

1058-2746/$ - see front matter Ó 2019 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved. https://doi.org/10.1016/j.jse.2019.07.002

2

H.A. Prentice et al. Level of evidence: Level III; Retrospective Cohort Design; Treatment Study Ó 2019 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved. Keywords: Shoulder arthroplasty; total hip arthroplasty; total knee arthroplasty; revision; registry; multiple arthroplasties

The volume of shoulder arthroplasties in the United States is growing exponentially20; Kim et al13 reported a 2.5-fold increase in shoulder arthroplasty procedures from 1998-2008 and Dillon et al5 reported a 2.7-fold increase in shoulder arthroplasty from 2005-2013. Despite this growth, the volume of shoulder arthroplasties performed annually is far outpaced by that of lower-extremity total joint arthroplasty (TJA) (ie, total hip arthroplasty and total knee arthroplasty). With an estimated prevalence of 7 million persons, TJA is one of the most common elective procedures in the United States,18 and the number of procedures performed is also expected to increase through 2030.16 This growth in arthroplasty is attributed to several factors, including a growing elderly population, a higher rate of diagnosis, an increased demand for more mobility, advances in implant design and surgical techniques, an increasing number of fellowship-trained surgeons, and increasing surgeon case volume.5,13,16,18 As the number of arthroplasties performed continues to increase, so too will the number of patients with arthroplasties in more than 1 joint. Evidence-based recommendations on the management of shoulder arthroplasty patients also requiring a lowerextremity TJA are lacking: Whether the lower-extremity TJA should be performed before or after the shoulder arthroplasty is currently not known. One hypothesis is that the shoulder procedure should be performed first so that the patient is then able to use ambulatory aids following the hip or knee procedure. However, there is a paucity of literature on the potential impact this additional procedure may have on the success of the shoulder procedure. Therefore, we sought to evaluate the following questions: (1) Is there a difference in the risk of revision following shoulder arthroplasty in patients who also undergo a lower-extremity TJA compared with those who do not? (2) Does the revision risk differ depending on the sequence of the lower-extremity TJA in relation to the shoulder arthroplasty procedure?

Materials and methods Study design and study sample Kaiser Permanente’s shoulder arthroplasty registry (SAR) and total joint replacement registry (TJRR) were used for this retrospective cohort study.4,22 This integrated health care system serves over 12.3 million members in 8 geographical

regions of the United States.10 Health plan membership has previously been shown to be largely socioeconomically and demographically representative of the geographical area it covers.11,14 Using the SAR, we identified all primary elective shoulder arthroplasty procedures performed from 2009 through 2015 in the 2 largest regions (Northern California and Southern California). These cases were performed by 225 surgeons in 37 medical centers in the state of California. Shoulder arthroplasty procedures included total shoulder arthroplasty (TSA), reverse total should arthroplasty (RTSA), hemiarthroplasty, and humeral head resurfacing arthroplasty. Some patients could have undergone a shoulder arthroplasty prior to the time of registry establishment (2009), and inclusion of these patients in the study sample would lead to inflation of the risk set and potentially bias the estimates. To minimize this bias, only patients younger than 75 years by 2009 (ie, born in 1934 or later) were included, as these younger individuals were less likely to have received an arthroplasty prior to 2009 compared with older patients.

Data collection The TJRR and SAR were established in 2001 and 2009, respectively. All arthroplasty procedures performed within our organization are captured by both registries, with 100% coverage since 2009. A detailed summary of data-collection procedures, coverage, and participation rates for both registries has been provided previously.3,4,22,23 In brief, both registries collect patient, procedure, implant, surgeon, and hospital information using electronic intraoperative forms that are completed at the point of care by the operating surgeon. Information is then supplemented using data from the electronic health record, administrative claims data, membership data, and mortality records. Outcomes, including revisions, are prospectively monitored using electronic screening algorithms and validated by trained clinical content experts using the electronic health record.

Exposure of interest Lower-extremity TJA, as well as the order of the procedures, was the exposure of interest. This was evaluated through 2 pathways: (1) The impact of a hip or knee procedure before the shoulder arthroplasty was modeled as a binary variable (yes or no). (2) The impact of a hip or knee procedure after the shoulder arthroplasty (an intermediate event) was modeled as a timedependent covariate, Xp(t) for primary and Xr(t) for revision, capturing hip or knee events, whichever came first. Hence, a patient is considered to have no intermediate lower-extremity arthroplasty, X(t) ¼ 0, until an additional arthroplasty occurs. For example, if a patient underwent a shoulder arthroplasty and then a total knee arthroplasty 100 days apart, this was modeled as

Multiple arthroplasties in shoulder patient Xp(0 < t < 100) ¼ 0 and Xp(t  100) ¼ 1. X(t) was treated as a categorical variable. Exposure status was identified through operative dates using the TJRR and SAR. Both registries were linked using unique patient identifiers assigned to patients through our health care system.

3 were not modeled. Analyses were performed using R (version 3.1.2; R Foundation for Statistical Computing, Vienna, Austria), and a ¼ .05 was the statistical significance threshold used for this study.

Results Outcomes of interest The primary outcome of interest was all-cause revision of the shoulder arthroplasty. Secondary outcomes included aseptic revision and septic revision. Revision was defined as any procedure related to the primary shoulder arthroplasty procedure in which at least 1 component was removed or replaced. Aseptic revision was defined as a revision for any reason other than an infection, and septic revision was defined as a revision for infection. Revisions were prospectively monitored by the SAR from the operative date through the date of final registry surveillance (March 31, 2016) and validated through manual chart review by trained clinical content experts.

Covariates Variables evaluated as potential confounders were related to the shoulder arthroplasty procedure and included age (continuous per 5-year increment), body mass index (continuous per 5year increment), sex (female or male), American Society of Anesthesiologists classification (1-2 or 3-5), and surgical indication (osteoarthritis, rotator cuff arthropathy, or other).

Statistical analysis Means, standard deviations (SDs), medians, and interquartile ranges were used to describe continuous variables, and frequencies and proportions were used to describe categorical variables. Revision outcomes were modeled as time-to-event outcomes using mixedeffects Cox models. The sequence of the lower-extremity TJA, that is, prior to shoulder arthroplasty vs. after shoulder arthroplasty, was evaluated using time-dependent categorical variables. For each outcome of interest, patients were censored on the date of leaving the health care plan membership, date of death, or study end date, whichever came first. Analysis was stratified according to 3 mutually exclusive groups based on indication and procedure type: hemiarthroplasty or humeral head resurfacing arthroplasty, RTSA, and TSA. These 3 groups were analyzed using a separate baseline hazard function because of the potentially unequal risks inherited from different procedures and the indication leading to the choice of procedure. Separate baseline hazard functions were fitted for each stratum. All models were fitted with a Gaussian frailty term at the surgeon level to control for stable surgeon effects and adjusted for the aforementioned confounders. To account for missing values in confounders, fully conditional specification multiple imputations using the Markov chain Monte Carlo estimation method were performed to create 50 versions of the analytical data set.29 Each data set was separately analyzed using the same model, and the results were combined using the Rubin rules.25 The imputation model included all variables and the cumulative baseline hazard of revision. Assumptions of proportional hazards in exposure variables were checked by plotting the Schoenfeld residuals against the vector of unique failure times. Numbers of events of less than 5

The final study group consisted of 4751 shoulder arthroplasties in 4253 patients. Among the 4751 shoulder arthroplasties, the mean age was 66.7 years (SD, 8.4 years), the mean body mass index was 30.6 kg/m2 (SD, 6.2 kg/m2), 51% were performed in female patients, 44.9% were performed in patients with an American Society of Anesthesiologists classification of 3 or higher, 69.4% were performed in patients with osteoarthritis as the indication for the shoulder arthroplasty procedure, and 61.5% were performed in patients who underwent a TSA (Table I). At the time of the index shoulder arthroplasty, 1285 shoulder arthroplasties (27.0%) had at least 1 prior lower-extremity TJA; 483 (10.2%) had a lower-extremity TJA (primary or revision) after the shoulder arthroplasty procedure, before any shoulder revision occurred. Among the 1285 shoulder arthroplasties with a prior lower-extremity TJA, the median time from TJA to shoulder arthroplasty was 2.8 years (range, 0.2-13.5 years). The median time from shoulder arthroplasty to subsequent TJA for the entire cohort was 1.5 years (range, 0.2-6.3 years). We observed 41 shoulder revisions (3.2%) in the prior lower-extremity TJA group and 102 revisions (2.9%) in the group with no history. We did not observe a difference in all-cause revision risk when comparing shoulder arthroplasty with and without a prior lower-extremity TJA (hazard ratio [HR], 1.38; 95% confidence interval [CI], 0.97-1.96; P ¼ .072) (Table II). Similarly, no difference was found in the risk of aseptic revision (HR, 1.39; 95% CI, 0.93-2.06; P ¼ .105) or septic revision (HR, 0.85; 95% CI, 0.36-1.99; P ¼ .709). There were 11 shoulder revisions (2.3%) in the group undergoing lower-extremity TJA after shoulder arthroplasty and 129 revisions (3.2%) in the shoulder arthroplasty–only group. We failed to observe a difference in all-cause revision (HR, 1.30; 95% CI, 0.82-2.06; P ¼ .264) or aseptic revision (HR, 1.79; 95% CI, 0.91-3.55; P ¼ .093) in patients who experienced a lower-extremity TJA following the shoulder arthroplasty procedure compared with patients with no subsequent lower-extremity TJA (Table II). Septic revisions could not be evaluated because of the low event frequency. In evaluating the order regarding the lower-extremity TJA, there was no difference in all-cause revision risk (HR, 1.03; 95% CI, 0.45-2.33; P ¼ .949) or aseptic revision risk (HR, 1.44; 95% CI, 0.12-3.38; P ¼ .391) when we compared TJA prior to vs. after shoulder arthroplasty. The primary reasons for shoulder revision in patients undergoing a shoulder arthroplasty only and those undergoing a lower-extremity TJA prior to the shoulder

4

H.A. Prentice et al. Table I Characteristics of 4751 primary elective shoulder arthroplasties (2009-2015) Characteristic

Data

Age, mean (SD), yr BMI, mean (SD),* kg/m2 Female sex, n (%) ASA classification  3,* n (%) Surgical indication,y n (%) Osteoarthritis Rotator cuff arthropathy Other Procedure, n (%) Hemi or HHR RTSA TSA

66.7 30.6 2425 1994

(8.4) (6.2) (51.0) (44.9)

3299 (69.4) 984 (20.7) 624 (13.1) 751 (15.8) 1079 (22.7) 2921 (61.5)

SD, standard deviation; BMI, body mass index; ASA, American Society of Anesthesiologists; Hemi, hemiarthroplasty; HHR, humeral head replacement; RTSA, reverse total shoulder arthroplasty; TSA, total shoulder arthroplasty. * Data were missing for BMI in 649 cases (13.7%) and ASA classification in 310 (6.5%). y Patients may have had multiple indications.

arthroplasty procedure were similar, including infection (24.8% and 20.0%, respectively), dislocation (20.0% and 20.0%, respectively), glenoid component loosening (16.8% and 22.9%, respectively), and rotator cuff tear (16.0% and 22.9%, respectively) (Table III). In contrast, the top reasons for revision for those undergoing a lower-extremity TJA following the shoulder arthroplasty procedure were glenoid wear after hemiarthroplasty (45.5%), humeral component loosening (18.2%), and rotator cuff tear (18.2%).

Discussion The number of patients requiring multiple arthroplasties in different joints is increasing. For patients requiring both shoulder and lower-extremity TJA procedures, it is unknown whether the sequence of procedures impacts the success of the shoulder replacement. In our registry-based cohort study, we did not observe a difference in the risk of revision following shoulder arthroplasty in patients who also underwent a TJA compared with patients undergoing a shoulder arthroplasty procedure only. This held true when evaluating shoulder arthroplasty patients who underwent a lower-extremity TJA prior to or after the shoulder procedure. Several studies have investigated the incidence of a subsequent arthroplasty following the initial procedure in lower-extremity TJA patients, finding contralateral arthroplasties in the same joint to be the most common subsequent procedure.7,17,19,24,26-28 The average time interval previously reported between arthroplasties ranged from 3 to 4.7 years.1,7,28 Most of these prior studies focused on patients requiring more than 1 lower-extremity TJA. A recent study by Lamplot et al,17 including hip, knee, and shoulder arthroplasties, reported that 27.3% of patients went on to undergo an arthroplasty in a subsequent joint, with 23.6% in the joint contralateral to the index procedure and 3.7% in a joint at a different anatomic site. In our cohort of shoulder arthroplasty patients, 27% had a history of lower-extremity TJA with a median time of 2.8 years from the preceding procedure to the shoulder arthroplasty. Moreover, 10.2% of patients underwent a lower-extremity TJA during follow-up after the shoulder arthroplasty procedure, similar to the 13.1% previously observed.17

Table II Adjusted revision risk following SA for patients who also underwent lower-extremity TJA compared with patients without lower-extremity TJA* SA outcome

Crude rate, n (%)

Adjusted estimate

TJA and SA

SA only

HRy (95% CI)

P value

41 (3.2) 35 (2.7) 6 (0.5)

102 (2.9) 80 (2.3) 28 (0.8)

1.38 (0.97-1.96) 1.39 (0.93-2.06) 0.85 (0.36-1.99)

.072 .105 .709

11 (2.3) 10 (2.1) 1 (0.2)

129 (3.2) 115 (2.4) 31 (0.8)

1.30 (0.82-2.06) 1.79 (0.91-3.55) d

.264 .093 d

z

Lower-extremity TJA prior to SA All-cause revision Aseptic revision Septic revision Lower-extremity TJA after SAx All-cause revision Aseptic revision Septic revisionk

SA, shoulder arthroplasty; TJA, total joint arthroplasty; HR, hazard ratio; CI, confidence interval. * Lower-extremity TJA procedures included both hips and knees and both primary and revision cases. y HRs were adjusted for age, body mass index, sex, American Society of Anesthesiologists classification, surgical indication, and procedure. z Estimates were interpreted as the risk of the SA outcome for patients with a history of lower-extremity TJA prior to the SA procedure compared with patients with no lower-extremity TJA history. x Estimates were interpreted as the risk of the SA outcome for patients who underwent a lower-extremity TJA following the SA procedure compared with patients who did not undergo a subsequent lower-extremity TJA. k Septic revision showed an event rate of less than 5 and therefore was not modeled.

Multiple arthroplasties in shoulder patient Table III Reason

5

Reasons for revision SA, stratified by patients with and without TJA n (%)*

Arthrofibrosis Dislocation Glenoid component loosening Glenoid fracture Glenoid wear after hemiarthroplasty Hematoma or seroma Hemiarthroplasty cuff dysfunction Humeral component loosening Implant malposition Infection Liner wear Other Periprosthetic fracture Rotator cuff tear Scapular notching

SA only (n ¼ 125)

TJA prior to SA only (n ¼ 35)

TJA after SA only (n ¼ 11)

1 25 21 2 18 1 10 8 10 31 1 6 4 20 0

2 7 8 0 3 0 3 0 2 7 0 2 2 8 0

0 1 1 0 5 0 1 2 1 0 0 0 0 2 0

(0.8) (20.0)y (16.8)y (1.6) (14.4) (0.8) (8.0) (6.4) (8.0) (24.8)y (0.8) (4.8) (3.2) (16.0) (0.0)

(5.7) (20.0)y (22.9)y (0.0) (8.6) (0.0) (8.6) (0.0) (5.7) (20.0)y (0.0) (5.7) (5.7) (22.9)y (0.0)

(0.0) (9.1) (9.1) (0.0) (45.5)y (0.0) (9.1) (18.2)y (9.1) (0.0) (0.0) (0.0) (0.0) (18.2)y (0.0)

SA, shoulder arthroplasty; TJA, total joint arthroplasty. * Some patients had more than 1 reason for revision. y Top reasons.

Although there are no evidence-based recommendations on the sequence of procedures for patients requiring multiple arthroplasties, one philosophy is to perform the shoulder procedure first so that the patient can then use ambulatory aids following the hip or knee procedure. However, it is unknown whether any increased forces acting on the shoulder from these ambulatory aids, even if for a short period, would predispose the patient to implant loosening or rotator cuff pathology necessitating revision. One study found that shoulder loads for patients using crutches following total hip replacement vary depending on the crutch setup.6 Forces from actions such as transfers across patients’ shoulders can be greater than 360% of normal.15 Data from studies looking at outcomes following shoulder arthroplasty in wheelchair-dependent patients could help guide the orthopedic surgeon in this decision-making process, given these patients’ increased reliance on the upper extremities for mobilization and transfers.8,9,12,21 In 1 study of 6 wheelchair-dependent patients who underwent shoulder arthroplasty, all had radiographic changes consistent with rotator cuff tearing at final follow-up.8 In another study, wheelchair-dependent patients undergoing RTSA were protected from transfers for 12 weeks to allow for integration of the baseplate, with the hope of mitigating the risk of implant loosening and dislocation.12 After 12 weeks, though, these increased forces could still be felt by the prosthesis during patient transfers. Although, in theory, there should be osseous ingrowth of the components, it is ultimately unknown how such forces may impact the longevity of the shoulder prosthesis over time. Whereas we did not observe

a difference in the revision risk for shoulder patients also requiring a lower-extremity TJA, we did observe some differences in the reasons for revision. The primary reasons for revision in patients who underwent a lower-extremity TJA before the shoulder procedure were similar to those in patients who underwent a shoulder arthroplasty only (infection, dislocation, glenoid component loosening, and rotator cuff tear), although frequencies did vary. In contrast, patients who underwent a lower-extremity TJA after the shoulder procedure appeared to have shoulder revisions that related more to accelerated wear of the implant, with almost half of those revised (45.5%) undergoing a revision for glenoid wear after a hemiarthroplasty. To our knowledge, this is the first study to investigate septic revision surgery for shoulder arthroplasty patients with and without a prior lower-extremity hip or knee arthroplasty, and we failed to observe a difference in risk. Other studies have instead compared patients with and without a history of periprosthetic joint infection and the risk of another periprosthetic joint infection following a subsequent arthroplasty.1,2 There are a number of limitations to this study. First, our exposure of interest was undergoing a lower-extremity TJA; our evaluation did not include history of replacement in joints other than the knee or hip. Furthermore, although we observed a wide range in the timing of when lower-extremity TJA procedures were performed in relation to the shoulder procedure (minimum, 0.2 years; maximum, 13.5 years), we were not able to evaluate whether differences in the revision risk exist within specific time intervals (eg, TJA performed within 6 months of the shoulder procedure

6

H.A. Prentice et al.

vs. at >6 months) because of low event rates within the different periods. Second, it is possible that a shoulder arthroplasty patient underwent a lower-extremity TJA at a different institution or underwent one at our institution prior to initiation of our TJRR and therefore was misclassified as undergoing a shoulder arthroplasty only, potentially creating bias. However, we expect this to be minimal as patients have little incentive to seek care outside of the health care plan and the TJRR started tracking all hip and knee procedures in 2001, well before the study start date. Furthermore, this type of misclassification would have made the shoulder arthroplasty–only patients more similar to the patients with a lower-extremity TJA, biasing our effect estimates toward the null value of 1. We also attempted to minimize the potential for this bias by excluding patients older than 75 years by the year 2009 when full coverage of all arthroplasty procedures performed within our integrated health care system was implemented by the registry. In addition, we were not able to evaluate other outcomes following shoulder arthroplasty, such as dislocations or rotator cuff tears not requiring revision surgery, as these are not captured by our SAR. Finally, our study was not able to include patient-reported outcomes and socioeconomic data. Perceptions regarding past surgical procedures, as well as financial constraints, may factor into patients’ willingness to undergo another procedure and outcomes.24 Future studies evaluating multiple arthroplasties may want to account for these factors. The study’s strengths are related to the data sources for our cohort study and the sophisticated statistical analysis performed. We included a large and representative sample of shoulder arthroplasty patients from our integrated health care system. Both the TJRR and SAR prospectively collect information on total hip arthroplasty, total knee arthroplasty, and shoulder arthroplasty procedures performed within Kaiser Permanente, and outcomes were validated through chart review. The statistical model used for testing our research hypotheses allowed us to evaluate the effect of multiple arthroplasties, with granular detail of the sequence and timing between procedures. It also differentiated between a primary or revision procedure in the lower extremity and stratified by procedure type for the shoulder.

Conclusion Evidence regarding the best sequence of procedures for patients requiring multiple arthroplasties is lacking. We did not observe an association with shoulder revision risk in shoulder arthroplasty patients who also underwent a lower-extremity TJA, regardless of the sequence of procedures. These findings may serve to help practicing orthopedic surgeons when counseling shoulder patients who have undergone or may require arthroplasties in other joints. Future study evaluating whether

the timing of the lower-extremity TJA in relation to the shoulder impacts the success of the shoulder procedure is needed.

Acknowledgments The authors acknowledge the Kaiser Permanente orthopaedic surgeons who contribute to the Kaiser Permanente Shoulder Arthroplasty Registry and the Kaiser Permanente Total Joint Replacement Registry, as well as the Surgical Outcomes and Analysis Department staff, which coordinates registry operations. They also acknowledge Brian H. Fasig, PhD, and William E. Burfeind, BS, for their ongoing support of the registry databases and quality-control management.

Disclaimer The authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.

References 1. Aibinder WR, Schoch BS, Cofield RH, Sanchez-Sotelo J, Sperling JW. Safety and efficacy of shoulder arthroplasty following lower extremity periprosthetic joint infection. J Shoulder Elbow Surg 2017;26:79-84. https://doi.org/10.1016/j.jse.2016.05.024 2. Bedair H, Goyal N, Dietz MJ, Urish K, Hansen V, Manrique J, et al. A history of treated periprosthetic joint infection increases the risk of subsequent different site infection. Clin Orthop Relat Res 2015;473: 2300-4. https://doi.org/10.1007/s11999-015-4174-4 3. Chen C, Garrido T, Chock D, Okawa G, Liang L. The Kaiser Permanente electronic health record: transforming and streamlining modalities of care. Health Aff (Millwood) 2009;28:323-33. https://doi. org/10.1377/hlthaff.28.2.323 4. Dillon MT, Ake CF, Burke MF, Singh A, Yian EH, Paxton EW, et al. The Kaiser Permanente shoulder arthroplasty registry: results from 6,336 primary shoulder arthroplasties. Acta Orthop 2015;86:286-92. https://doi.org/10.3109/17453674.2015.1024565 5. Dillon MT, Chan PH, Inacio MC, Singh A, Yian EH, Navarro RA. Yearly trends in elective shoulder arthroplasty, 2005 through 2013. Arthritis Care Res (Hoboken) 2017;69:1574-81. https://doi.org/10. 1002/acr.23167 6. Freddolini M, Esposito F, Latella L, Marcucci M, Corvi A. Shoulder joint loadings in post total hip replacement surgery patients during assisted walking: the influence of the crutch setup. J Biomech 2018; 72:46-52. https://doi.org/10.1016/j.jbiomech.2018.02.021 7. Gillam MH, Ryan P, Salter A, Graves SE. Multi-state models and arthroplasty histories after unilateral total hip arthroplasties: introducing the summary notation for arthroplasty histories. Acta Orthop 2012;83:220-6. https://doi.org/10.3109/17453674.2012.684140

Multiple arthroplasties in shoulder patient 8. Hattrup SJ, Cofield RH. Shoulder arthroplasty in the paraplegic patient. J Shoulder Elbow Surg 2010;19:434-8. https://doi.org/10.1016/j. jse.2009.07.003 9. Jordan RW, Sloan R, Saithna A. Should we avoid shoulder surgery in wheelchair users? A systematic review of outcomes and complications. Orthop Traumatol Surg Res 2018;104:839-46. https://doi.org/10. 1016/j.otsr.2018.03.011 10. Kaiser Permanente. Fast facts about Kaiser Permanente. https:// share.kaiserpermanente.org/article/fast-facts-about-kaiser-permanente/; 2018. Accessed March 1, 2018 11. Karter AJ, Ferrara A, Liu JY, Moffet HH, Ackerson LM, Selby JV. Ethnic disparities in diabetic complications in an insured population. JAMA 2002;287:2519-27. https://doi.org/10.1001/jama.287.19.2519 12. Kemp AL, King JJ, Farmer KW, Wright TW. Reverse total shoulder arthroplasty in wheelchair-dependent patients. J Shoulder Elbow Surg 2016;25:1138-45. https://doi.org/10.1016/j.jse.2015.11.006 13. Kim SH, Wise BL, Zhang Y, Szabo RM. Increasing incidence of shoulder arthroplasty in the United States. J Bone Joint Surg Am 2011; 93:2249-54. https://doi.org/10.2106/jbjs.j.01994 14. Koebnick C, Langer-Gould AM, Gould MK, Chao CR, Iyer RL, Smith N, et al. Sociodemographic characteristics of members of a large, integrated health care system: comparison with US Census Bureau data. Perm J 2012;16:37-41. https://doi.org/10.7812/TPP/ 12-031 15. Kulig K, Rao SS, Mulroy SJ, Newsam CJ, Gronley JK, Bontrager EL, et al. Shoulder joint kinetics during the push phase of wheelchair propulsion. Clin Orthop Relat Res 1998:132-43. 16. Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 2007;89:780-5. https://doi.org/10.2106/ jbjs.f.00222 17. Lamplot JD, Bansal A, Nguyen JT, Brophy RH. Risk of subsequent joint arthroplasty in contralateral or different joint after index shoulder, hip, or knee arthroplasty: association with index joint, demographics, and patient-specific factors. J Bone Joint Surg Am 2018; 100:1750-6. https://doi.org/10.2106/jbjs.17.00948 18. Maradit Kremers H, Larson DR, Crowson CS, Kremers WK, Washington RE, Steiner CA, et al. Prevalence of total hip and knee replacement in the United States. J Bone Joint Surg Am 2015;97: 1386-97. https://doi.org/10.2106/jbjs.n.01141

7 19. McMahon M, Block JA. The risk of contralateral total knee arthroplasty after knee replacement for osteoarthritis. J Rheumatol 2003;30: 1822-4. 20. Padegimas EM, Maltenfort M, Lazarus MD, Ramsey ML, Williams GR, Namdari S. Future patient demand for shoulder arthroplasty by younger patients: national projections. Clin Orthop Relat Res 2015;473:1860-7. https://doi.org/10.1007/s11999-0154231-z 21. Patel RM, Gelber JD, Schickendantz MS. The weight-bearing shoulder. J Am Acad Orthop Surg 2018;26:3-13. https://doi.org/10.5435/ jaaos-d-15-00598 22. Paxton EW, Inacio MC, Khatod M, Yue EJ, Namba RS. Kaiser Permanente national total joint replacement registry: aligning operations with information technology. Clin Orthop Relat Res 2010;468:264663. https://doi.org/10.1007/s11999-010-1463-9 23. Paxton EW, Kiley ML, Love R, Barber TC, Funahashi TT, Inacio MC. Kaiser Permanente implant registries benefit patient safety, quality improvement, cost-effectiveness. Jt Comm J Qual Patient Saf 2013;39: 246-52. https://doi.org/10.1016/S1553-7250(13)39033-3 24. Ravi B, Croxford R, Hawker G. Exclusion of patients with sequential primary total joint arthroplasties from arthroplasty outcome studies biases outcome estimates: a retrospective cohort study. Osteoarthritis Cartilage 2013;21:1841-8. https://doi.org/10.1016/j.joca.2013.08.020 25. Rubin DF. Multiple imputation for nonresponse surveys. New York: John Wiley & Sons; 1987. 26. Sanders TL, Maradit Kremers H, Schleck CD, Larson DR, Berry DJ. Subsequent total joint arthroplasty after primary total knee or hip arthroplasty: a 40-year population-based study. J Bone Joint Surg Am 2017;99:396-401. https://doi.org/10.2106/jbjs.16.00499 27. Sayeed SA, Trousdale RT, Barnes SA, Kaufman KR, Pagnano MW. Joint arthroplasty within 10 years after primary Charnley total hip arthroplasty. Am J Orthop (Belle Mead NJ) 2009;38:E141-3. 28. Shao Y, Zhang C, Charron KD, Macdonald SJ, McCalden RW, Bourne RB. The fate of the remaining knee(s) or hip(s) in osteoarthritic patients undergoing a primary TKA or THA. J Arthroplasty 2013;28:1842-5. https://doi.org/10.1016/j.arth.2012.10.008 29. van Buuren S. Multiple imputation of discrete and continuous data by fully conditional specification. Stat Methods Med Res 2007;16:21942. https://doi.org/10.1177/0962280206074463