Abstract
Purpose
Emerging evidence suggests that there may be a link between upper-extremity surgeries and carpal tunnel syndrome (CTS), but whether shoulder arthroplasty is associated with a greater risk of CTS remains unknown. This study investigates the 1-year incidence of ipsilateral CTS following shoulder arthroplasty using the incidence of contralateral CTS as an internal control.
Methods
This retrospective cohort study used Medicare claims from 2016 to 2022 to evaluate adults ≥18 years old undergoing shoulder arthroplasty. The incidence rate of ipsilateral CTS and contralateral CTS up to 1-year following surgery was estimated for the entire cohort and by subgroups based on the indication for the shoulder arthroplasty. The incidence rate ratio comparing ipsilateral to contralateral incidence of CTS was estimated before and after adjusting for covariates. Among those that developed ipsilateral CTS, the rate of operative CTS and injection to treat CTS was estimated up to 1-year following the incident ipsilateral CTS event.
Results
There were 4,261 adults identified. For the entire cohort, the ipsilateral incidence of CTS was 46% higher than the contralateral side. A higher CTS incidence on the ipsilateral versus contralateral side was consistent across the subgroups before and after adjusting for covariates. Of 136 beneficiaries who developed ipsilateral CTS, 44.9% received operative carpal tunnel release or injection to treat CTS within 1-year of their CTS diagnosis.
Conclusions
This analysis found a significantly increased incidence of ipsilateral CTS within 1 year of shoulder arthroplasty, with consistent findings for both degenerative and traumatic indications.
Type of study/level of evidence
Differential diagnosis/symptom prevalence study II.
Key words: Carpal tunnel syndrome, Reverse shoulder arthroplasty, Shoulder arthroplasty, Total shoulder arthroplasty
Carpal tunnel syndrome (CTS) is the most common entrapment neuropathy worldwide, accounting for up to 90% of nerve compression disorders, with a prevalence of 5% in the Western world.1 Carpal tunnel syndrome impairs quality of life, reduces work productivity, and increases health care costs, emphasizing the need for timely diagnosis and management.1, 2, 3 Although systemic conditions such as diabetes and pregnancy are known risk factors, the impact of surgical interventions proximal to the wrist remain less explored.1 There is increased recognition that the risk for developing CTS extends beyond just periods of localized mechanical compression; in fact, underlying subclinical neuropathies may also increase median nerve susceptibility to injury.1 In this context, the perioperative and postoperative changes associated with shoulder surgeries could lead to proximal neuropathic stressors that predispose patients to CTS.
Among these interventions is shoulder arthroplasty, a procedure performed with increasing prevalence over the past two decades.4 In a small prospective cohort study, Lädermann et al5 reported a higher incidence of acute, transient postoperative peripheral neurological lesions in patients undergoing reverse shoulder arthroplasties (RSAs) versus anatomic total shoulder arthroplasty (TSA), suggesting possible procedure-specific neurological risks; however, this study did not investigate long-term nerve injury. It also found subclinical worsening of prior preoperative compressive peripheral neuropathy (CPN) in 3 of 19 RSA patients, but no onset of new postoperative CPN lesions. Yian et al6 evaluated the incidence of new postoperative CPNs using retrospective cohorts comparing patients undergoing shoulder arthroplasty with nonsurgical patients with shoulder arthritis. They did not report a significant increase in CPNs; however, they did discuss potential perioperative risk factors for CPNs specific to shoulder arthroplasty and acknowledged those inherent to the nonsurgical arthritis control group.6 Given the growing prevalence of shoulder arthroplasty, understanding its relationship with CTS is clinically relevant to hand surgeons.
Despite the potential link between shoulder arthroplasty and CTS, the relationship remains poorly characterized. This study aimed to investigate the 1-year incidence of ipsilateral CTS following shoulder arthroplasty. The secondary objective was to characterize the rate of CTS treatment among those who developed CTS within 1-year after surgery. We hypothesize that patients have an increased incidence of developing CTS after shoulder arthroplasty. This research seeks to inform postoperative expectations and improve patient counseling about shoulder arthroplasty.
Materials and Methods
Data source
This retrospective cohort study used administrative claims data from January 1, 2016, to December 31, 2022, from a 20% random sample of the Medicare fee-for-service database Part A (hospital insurance) and B (medical insurance, covering outpatient services). For research purposes, surgery and medical conditions were identified by searching for Current Procedural Terminology (CPT) codes and International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10) codes (Table E1, available online on the Journal’s website at www.jhandsurg.org). Because data were deidentified prior to administering to researchers, the university’s institutional review board approved this study as nonregulated.
Cohort selection and follow-up time
We identified Medicare beneficiaries aged ≥18 years who underwent either a TSA or RSA using CPT code 23472 linked to traumatic, post-traumatic, degenerative, or other diagnoses requiring shoulder arthroplasty (Table E1). The start date of follow-up was the date of shoulder arthroplasty occurring between January 1, 2017, and December 31, 2020. Beneficiaries were included if they had evidence of laterality on the same day as the surgery, had continuous enrollment in Part A and B during the 1-year baseline period and for ≥1 day of follow-up, and had evidence of unilateral surgery. The requirement of ≥1 day of follow-up was selected to mitigate sample selection bias as the analytic techniques account for right censoring because of loss to follow-up and mortality. Unspecified or bilateral diagnoses were excluded to ensure accurate laterality.
Outcomes
Carpal tunnel syndrome was identified using ICD-10 codes G56.01 and G56.02. Evidence of CTS on either side (ipsilateral or contralateral) during the 1-year baseline period was also identified and used for adjustment. For those that had diagnostic evidence of CTS in both the baseline and follow-up periods, we required a gap of at least 6 months between CTS dates to ensure that the follow-up CTS date was an incident event, as opposed to follow-up clinical care of CTS that originated in the baseline period. To do this, we first calculated the difference in days from the first follow-up CTS date minus the most recent baseline date of CTS closest to day 0. If the gap was 6 months or greater, the individuals were considered to have both baseline and follow-up evidence of distinct CTS events. If the gap was <6 months, the follow-up variables were flagged as having no evidence and other follow-up information was used to identify the right censoring event (eg, end of 1-year follow-up, mortality, and lost to follow-up).
The primary outcome is the incidence of ipsilateral CTS. The contralateral incidence of CTS was used as an internal control that accounts for the cohort’s baseline risk of CTS. Thus, the extent of incident ipsilateral CTS above the risk of contralateral CTS was assessed to better examine the association between shoulder arthroplasty and CTS.
Among those that developed CTS within 1-year of their shoulder arthroplasty, the first surgical date of operative carpal tunnel release (CPT 29848 and 64721) or injection (CPT 20526) to treat CTS were identified up to 1-year following the incident CTS date.
Age, sex, race, US region of residence, the original reason for Medicare entitlement, and dual eligibility with Medicaid were retrieved from the database. Baseline evidence of CTS on any side was also included, as well as diagnostic evidence of the following 8 comorbidities during the 1-year baseline period: obesity, diabetes mellitus, hypertension, chronic obstructive pulmonary disease, chronic kidney disease, congestive heart failure, rheumatoid arthritis, and hypothyroidism (Table E1). A multimorbidity variable was constructed that summed the presence of the 8 comorbidities as this may better capture overall medical complexity, with scores ranging from 0 (0 comorbidities) to 8 (all 8 comorbidities).
Statistical analysis
Baseline descriptive characteristics and follow-up information were summarized for the entire shoulder arthroplasty cohort and by subgroups based on the indication for the surgery.
For the primary objective, a crude and adjusted ratio of ratios approach was used to evaluate the effect of shoulder arthroplasty on incidence of ipsilateral CTS while accounting for the incidence of contralateral CTS in the entire cohort and by subgroups. First, the crude incidence rate (IR with 95% confidence intervals [CIs]) of CTS per 100 person-years was estimated for the ipsilateral and contralateral sides as the number of events divided by person-time for the entire cohort and subgroups. Individuals were followed from day 0 to the CTS date, mortality, lost to follow-up, or end of the 1-year follow-up period, whichever came first. Second, the crude ratio of ratios, as the IR ratio (IRR), was estimated by comparing the IR of ipsilateral CTS to the IR of contralateral CTS. Third, the adjusted IRR was estimated using Cox proportional hazards regression, where the cumulative incidence, as 1 minus the baseline hazard, of ipsilateral CTS incidence and contralateral CTS incidence was estimated in separate models after adjusting for covariates. Because of limited sample sizes for some of the subgroups, two sets of models were developed to account for confounding. Model 1 adjusted for age (as continuous) and baseline evidence of CTS on any side. Model 2 adjusted for the model 1 covariates plus sex, race (as White vs non-White because of limited sample size), baseline multimorbidity, baseline diabetes mellitus, and dual eligibility with Medicaid, and for the entire cohort, the indication for shoulder arthroplasty (degenerative, post-traumatic, traumatic, and other) was additionally adjusted for. The proportional hazards assumption was tested for each covariate in the models based on the weighted Schoenfeld residuals.
For the secondary objective, the cohort was restricted to those that developed ipsilateral CTS within 1-year of their shoulder arthroplasty. For this restricted cohort, the start date of follow-up, or day 0, was revised to be the date of incident CTS. The IR (with 95% CI) of operative carpal tunnel syndrome and injection to treat CTS per 100 person-years was estimated.
Estimates that had <11 cases were not reported or were suppressed to comply with the Data Use Agreement for patient de-identification purposes. Analyses were performed using Statistical Analysis Software version 9.4 and P < .05 (two-tailed) was considered statistically significant.
Results
There were 4,261 beneficiaries with an eligible shoulder arthroplasty. A flow chart to derive the analytic cohort is presented in Figure 1. Given the small sample size for those with a post-traumatic indication for shoulder arthroplasty (n = 51), degenerative and post-traumatic indications were grouped together because both conditions may share overlapping clinical features and treatment approaches and combining them allowed for sufficient statistical power to perform meaningful analyses without compromising the integrity of the classification. Thus, most (n = 3,576, 83.9% of entire cohort) were because of degenerative including post-traumatic indications, followed by traumatic indications (n = 422, 9.9% of entire cohort) and other indications (n = 263, 6.2% of entire cohort). Baseline descriptive characteristics of the entire cohort and subgroups based on the indication for surgery are presented in Table 1.
Figure 1.
Study flow. Flow chart of inclusion and exclusion criteria using Medicare claims database.
Table 1.
Baseline Descriptive Characteristics of the Entire Cohort of Medicare Beneficiaries Who Underwent Shoulder Arthroplasty and by Subgroup Based on the Indication for Shoulder Arthroplasty
| Description | Entire Cohort (n = 4,261) | Degenerative (Including Post-traumatic) Indication Cohort (n = 3,576) | Trauma Indication Cohort (n = 422) | Other Indication Cohort (n = 263) |
|---|---|---|---|---|
| Indication for surgery, % (n) | ||||
| Trauma | 9.9 (422) | - | - | - |
| Degenerative (including post-trauma) | 83.9 (3,576) | - | - | - |
| Other | 6.2 (263) | - | - | - |
| Age, median (IQR) | 72.9 (69.0–77.6) | 72.7 (68.9–77.2) | 74.9 (69.9–80.7) | 73.4 (69.2–78.1) |
| Sex, % (n) | ||||
| Male | 40.2 (1,714) | 42.7 (1,528) | 17.3 (73) | 43.0 (113) |
| Female | 59.8 (2,547) | 57.3 (2,048) | 82.7 (349) | 57.0 (150) |
| Race, % (n) | ||||
| Asian | ∗ | ∗ | ∗ | ∗ |
| Black | 3.8 (161) | 4.0 (143) | ∗ | 5.7 (15) |
| Hispanic | 0.7 (30) | 0.7 (24) | ∗ | ∗ |
| North American Native | ∗ | ∗ | ∗ | 0 (0) |
| White | 92.2 (3,930) | 92.1 (3,294) | 95.0 (401) | 89.4 (235) |
| Other/unknown | 2.6 (110) | 2.5 (90) | ∗ | ∗ |
| US region of residence, % (n) | ||||
| Northeast | 14.1 (602) | 14.4 (514) | 14.2 (60) | 10.7 (28) |
| Midwest | 22.6 (963) | 22.9 (817) | 20.6 (87) | 22.4 (59) |
| South | 42.6 (1,813) | 41.2 (1,472) | 50.0 (211) | 49.4 (130) |
| West | 20.6 (878) | 21.5 (769) | 14.9 (63) | 17.5 (46) |
| Original reason for Medicare entitlement, % (n) | ||||
| Old age and survivor’s insurance | 79.8 (3,399) | 79.5 (2,843) | 83.9 (354) | 76.8 (202) |
| DIB | 20.0 (853) | 20.3 (725) | 16.1 (68) | 22.8 (60) |
| ESRD | ∗ | ∗ | ∗ | ∗ |
| Both DIB and ESRD | ∗ | ∗ | ∗ | ∗ |
| Dual eligibility with Medicaid, % (n) | 9.3 (398) | 9.2 (328) | 10.9 (46) | 9.1 (24) |
| Shoulder arthroplasty side, % (n) | ||||
| Right side only | 57.9 (2,469) | 57.4 (2,051) | 56.9 (240) | 67.7 (178) |
| Left side only | 42.1 (1,792) | 42.7 (1,525) | 43.1 (182) | 32.3 (85) |
| Baseline occurrence of CTS on any side, % (n) | 4.3 (181) | 4.5 (161) | ∗ | 4.6 (12) |
| Ipsilateral side, % (n) | 2.9 (125) | 3.1 (110) | ∗ | ∗ |
| Contralateral side, % (n) | 2.5 (105) | 2.7 (96) | ∗ | ∗ |
| Baseline comorbidities, % (n) | ||||
| Obesity | 34.9 (1,486) | 35.5 (1,269) | 30.1 (127) | 34.2 (90) |
| Diabetes mellitus | 32.1 (1,366) | 31.5 (1,128) | 34.8 (147) | 34.6 (91) |
| Hypertension | 82.6 (3,521) | 82.3 (2,944) | 85.1 (359) | 82.9 (218) |
| Chronic obstructive pulmonary disease | 16.1 (685) | 15.9 (569) | 13.3 (56) | 22.8 (60) |
| Chronic kidney disease | 17.2 (733) | 16.8 (600) | 21.3 (90) | 16.4 (43) |
| Congestive heart failure | 11.1 (474) | 10.3 (367) | 15.4 (65) | 16.0 (42) |
| Rheumatoid arthritis | 7.7 (326) | 7.8 (278) | 5.5 (23) | 9.5 (25) |
| Hypothyroidism | 29.6 (1,261) | 29.7 (1,061) | 31.3 (132) | 25.9 (68) |
| Multimorbidity of baseline comorbidities | ||||
| Median (IQR) | 2 (1–3) | 2 (1–3) | 2 (1–3) | 2 (1–3) |
| 0 comorbidities, % (n) | 7.8 (333) | 7.9 (281) | 7.8 (33) | 7.2 (19) |
| 1 comorbidity, % (n) | 23.3 (993) | 23.6 (844) | 22.8 (96) | 20.2 (53) |
| 2 comorbidities, % (n) | 28.1 (1,196) | 28.3 (1,013) | 26.8 (113) | 26.6 (70) |
| 3 comorbidities, % (n) | 21.7 (926) | 21.5 (769) | 20.6 (87) | 26.6 (70) |
| 4 comorbidities, % (n) | 11.8 (501) | 11.6 (414) | 13.3 (56) | 11.8 (31) |
| ≥5 comorbidities, % (n) | 7.3 (312) | 7.1 (255) | 8.8 (37) | 7.6 (20) |
DIB, disability insurance benefits; ESRD, end-stage renal disease; IQR, interquartile range.
N ≤ 11 cases in at least 1 cell, and data are suppressed due as part of the Data Use Agreement for patient deidentification purposes.
Primary objective
The IRs of CTS on the ipsilateral and contralateral sides are presented numerically in Table 2 and visually in Figure 2. For the entire cohort, the ipsilateral incidence of CTS (IR, 3.28; 95% CI 2.73–3.83) was 46% higher than the contralateral incidence of CTS (IR, 2.25; 95% CI, 1.80–2.71) (IRR, 1.46; 95% CI, 1.12–1.89; P = .005). A higher CTS incidence on the ipsilateral versus contralateral side was consistent across the subgroups (IRR range, 1.33–3.30) and was significantly higher for the degenerative (including post-traumatic) and traumatic indication cohorts (P = .046 and .037, respectively) but not for the other indication cohort (IRR, 2.71; 95% CI, 0.72–10.23, P = .141). This was likely because of the few outcome events in the stratified analysis.
Table 2.
Censor Reason and IR of CTS on the Ipsilateral and Contralateral Side of the Shoulder Arthroplasty, and IR ratio Comparing the Ipsilateral to Contralateral Side for the Entire Cohort of Medicare Beneficiaries Who Underwent Shoulder Arthroplasty and by Subgroup Based on the Indication for Shoulder Arthroplasty
| Censor Reason∗ |
Person-Years | IR per 100 Person-Years (95% CI) | IRR (95% CI) | P Value | ||
|---|---|---|---|---|---|---|
| End of Follow-up, % (n) | Event, % (n) | |||||
| Entire cohort (n = 4,261) | ||||||
| Ipsilateral CTS incidence | 95.0 (4,046) | 3.2 (136) | 4,148 | 3.28 (2.73–3.83) | 1.46 (1.12–1.89) | .005 |
| Contralateral CTS incidence | 95.9 (4,087) | 2.2 (94) | 4,172 | 2.25 (1.80–2.71) | Reference | |
| Degenerative (including post-traumatic) indication cohort (n = 3,576) | ||||||
| Ipsilateral CTS incidence | 95.2 (3,405) | 3.2 (115) | 3,491 | 3.29 (2.69–3.90) | 1.33 (1.01–1.75) | .046 |
| Contralateral CTS incidence | 96.0 (3,432) | 2.4 (87) | 3,504 | 2.48 (1.96–3.00) | Reference | |
| Trauma indication cohort (n = 422) | ||||||
| Ipsilateral CTS incidence | 92.7 (391) | 3.1 (13) | 403 | 3.23 (1.47–4.98) | 3.30 (1.07–10.11) | .037 |
| Contralateral CTS incidence | 94.8 (400) | † | 408 | 0.98 (0.02–1.94) | Reference | |
| Other indication cohort (n = 263) | ||||||
| Ipsilateral CTS incidence | 95.1 (250) | ∗ | 254 | ∼3.1‡ | 2.71 (0.72–10.23) | .141 |
| Contralateral CTS incidence | 97.0 (255) | ∗ | 259 | ∼1.2‡ | Reference | |
Lost to follow-up was exceedingly rare for the full cohort and subcohorts (<1%). As lost to follow-up had N ≤ 11 cases, data are suppressed for this variable and the mortality variable under the censor reason as part of the Data Use Agreement for patient deidentification purposes. Mortality as a censor reason accounted for 1.5% to 4.3% of the full cohort and subcohorts.
N ≤ 11 cases and data are suppressed due as part of the Data Use Agreement for patient deidentification purposes.
Exact incidence rates cannot be reported to prevent “back calculation” for patient deidentification identification purposes.
Figure 2.
Cumulative incidence plot of incident CTS up to 1-year following shoulder arthroplasty for the ipsilateral and contralateral side of the surgery for the entire cohort (n = 4,261).
After adjusting for covariates (Table 3), the finding of a higher CTS incidence on the ipsilateral side than on the contralateral side remained. However, the effect was slightly attenuated for the entire cohort (crude IRR, 1.46; model 1 adjusted IRR, 1.43; model 2 adjusted IRR, 1.31; all P < .05), whereas the effect was augmented for the degenerative (including post-traumatic) indication cohort (crude IRR, 1.33; model 1 adjusted IRR, 1.47; model 2 adjusted IRR, 1.46; all P < .05). There were too few outcome events to adjust for the model 2 covariates for the traumatic and other indications cohorts. Adjusting for the model 1 covariates, the finding of a higher CTS incidence on the ipsilateral side than on he contralateral side remained, but the effect was slightly attenuated for the trauma indication cohort (crude IRR, 3.30; model 1 adjusted IRR, 3.07; both P < .05) and other indication cohort (crude IRR, 2.71; model 1 adjusted IRR, 2.59; P = .141 and .160, respectively). There was no evidence that the proportional hazards assumption was violated in any model (all P > .05).
Table 3.
Cumulative Incidence From Cox Regression (Calculated as 1 – Baseline Hazard) of CTS on the Ipsilateral and Contralateral Side of the Shoulder Arthroplasty, and the IRR Comparing Ipsilateral to Contralateral Sides for the Entire Cohort of Medicare Beneficiaries Who Underwent Shoulder Arthroplasty and by Subgroup Based on the Indication for Shoulder Arthroplasty
| Model 1 |
Model 2 |
|||||
|---|---|---|---|---|---|---|
| Cumulative Incidence, % | IRR (95% CI) | P Value | Cumulative Incidence, % | IRR (95% CI) | P Value | |
| Entire cohort (n = 4,261) | ||||||
| Ipsilateral CTS incidence | 2.92 | 1.43 (1.10–1.87) | .007 | 2.60 | 1.31 (1.01–1.70) | .044 |
| Contralateral CTS incidence | 2.04 | Reference | 1.99 | Reference | ||
| Degenerative (including post-traumatic) indication cohort (n = 3,576) | ||||||
| Ipsilateral CTS incidence | 2.94 | 1.47 (1.11–1.94) | .007 | 2.80 | 1.46 (1.11–1.93) | .008 |
| Contralateral CTS incidence | 2.01 | Reference | 1.91 | Reference | ||
| Trauma indication cohort (n = 422) | ||||||
| Ipsilateral CTS incidence | 2.83 | 3.07 (1.01–9.40) | .049 | ∗ | ∗ | ∗ |
| Contralateral CTS incidence | 0.92 | Reference | ∗ | ∗ | ∗ | |
| Other indication cohort (n = 263) | ||||||
| Ipsilateral CTS incidence | 2.82 | 2.59 (0.69–9.77) | .160 | ∗ | ∗ | ∗ |
| Contralateral CTS incidence | 1.09 | Reference | ∗ | ∗ | ∗ | |
Model 1 adjusted for age and baseline occurrence of CTS, and additionally for the entire cohort, the subgroup indication for surgery was also included for adjustment. Model 2 adjusted for age, baseline occurrence of CTS, sex, race (as White vs non-White), baseline multimorbidity, baseline diabetes, and dual eligibility with Medicaid, and additionally for the entire cohort, the subgroup indication for surgery was also included for adjustment.
Analyses were not conducted because of insufficient sample size for adjustment.
Secondary objective
Of the 136 that developed incident CTS on the ipsilateral side of their shoulder arthroplasty, 61 (44.9%) received either operative carpal tunnel release or injection to treat CTS within 1-year of their CTS diagnosis. These outcomes could not be differentiated because the overwhelming majority (>90%) received operative carpal tunnel release, leaving <11 cases for the injection group. Therefore, the data cannot be reported for patient deidentification purposes. The IR and censor information for the combined outcome of either operative carpal tunnel release or injection to treat CTS is presented numerically in Table 4 and visually in Figure 3, which demonstrates that the majority of outcome events occurred within the first 100 days.
Table 4.
Among Those Who Developed Ipsilateral CTS Within 1-Year of Their Shoulder Arthroplasty (n = 136), the Censor Reason and IR of Receiving Either Operative Carpal Tunnel Release (OCTR) or Injection for CTS Following the Date of Incident CTS
| Censor Reason∗ |
Person-Years | IR per 100 Person-Years (95% CI) | ||
|---|---|---|---|---|
| End of Follow-up, % (n) | Event, % (n) | |||
| OCTR or injection for CTS | 53.7 (73) | 44.9 (61) | 83 | 73.3 (54.9–91.7) |
Lost to follow-up and mortality were exceedingly rare (<2%). As lost to follow-up and mortality had N ≤ 11 cases, data are suppressed for these variables as part of the Data Use Agreement for patient deidentification purposes.
Figure 3.
Cumulative incidence plot of receiving either operative carpal tunnel release or injection to treat CTS up to 1-year following incident CTS among those that developed incident ipsilateral CTS within 1-year following shoulder arthroplasty (n = 136).
Discussion
The present study characterized the rate of incident CTS on the ipsilateral upper extremity versus contralateral side following shoulder arthroplasty. Study findings suggest that shoulder arthroplasty is associated with an increased rate of ipsilateral CTS. These findings were consistent across the subgroups based on the indication for the surgery. Of those that developed ipsilateral CTS within 1 year of shoulder arthroplasty, 44.9% received an injection or carpal tunnel release within 1 year of developing CTS, with the majority undergoing surgical decompression. These findings have important implications for multiple stakeholders. For patients, awareness of CTS may inform preoperative expectations and decision making. For providers, heightened vigilance for early CTS symptoms, through screening and postoperative surveillance, may improve timely diagnosis and treatment. For payers, recognition of this association may influence coverage decisions related to postoperative care.
Consistent with the overall cohort, the subgroup analyses revealed that the association with ipsilateral CTS after shoulder arthroplasty was observed across most surgical indications. The effect was most pronounced in the degenerative (including post-traumatic arthritis) subgroup, where this subgroup exhibited a 46% higher rate of ipsilateral CTS than that of the contralateral upper extremity. Although the trauma subgroup also exhibited an elevated rate of ipsilateral versus contralateral CTS, the small sample size limited statistical inferences. Nevertheless, the consistency of findings across various indications, despite underlying differences in patient populations and pathologies, suggests that the association between shoulder arthroplasty and subsequent ipsilateral CTS may not be solely attributable to surgical indications and comorbid conditions alone, but rather to perioperative or postoperative factors common to shoulder arthroplasty more broadly.
Our findings add to the literature. Medvedev et al7 demonstrated an increased CTS incidence following ipsilateral arthroscopic labral and rotator cuff repairs when compared with the nonsurgical arm, raising the possibility of a broader link between proximal upper extremity surgery and distal entrapment neuropathies. Yian et al6 also reported a trend toward increased CTS following shoulder arthroplasty compared with nonsurgical arthritic controls, although without statistical significance. Several mechanisms may explain this association. Medvedev et al7 suggested that perioperative edema, soft tissue swelling, and local hematoma formation may promote endoneurial edema and median nerve compression within the carpal tunnel. Intraoperative factors that may predispose nerves to dysfunction include surgical duration or limb positioining.6,8,9 Postoperative conditions following shoulder arthroplasty, such as prolonged sling immobilization and repetitive rehabilitation exercises, may increase carpal tunnel pressure or exacerbate pre-existing subclinical pathology; however, further investigation is needed.6
It is important to consider how perioperative mechanical nerve injury, such as inadvertent traction or compression during surgery, may contribute to downstream neuropathic changes or heightened susceptibility to compressive neuropathy. In fact, recent electrodiagnostic studies have shown that intraoperative trauma to the brachial plexus, axillary, or suprascapular nerves occurs more commonly than previously appreciated in shoulder arthroplasties.10 The concept of double-crush syndrome provides one explanatory framework for how these proximal intraoperative nerve injuries may contribute to downstream neuropathic changes. Double-crush syndrome posits that a proximal nerve lesion lowers axonal resilience and increases vulnerability to distal entrapment.11 It would follow that perioperative neural stress around the shoulder could act as a “first hit,” predisposing the median nerve to symptomatic CTS. However, evidence for double-crush syndrome remains mixed. Multiple electrophysiological studies as well as a large retrospective study have shown no consistent relationship between cervical radiculopathy and CTS severity, although nerve injuries in the shoulder region appear less studied.12, 13, 14 This mixed evidence indicates that, although our findings are compatible with a double-crush mechanism, causality cannot be assumed.
The study’s results should be interpreted within the context of some assumptions using claims-based databases. We selected the Medicare database because we wanted to investigate older adults, whom we assumed were more likely to receive shoulder arthroplasties. Beneficiaries are also less likely to exit the database. However, our findings may not generalize to the wider population. It is also impossible to know the real clinical managements in claims databases, which is relevant when trying to interpret what happened to the other 55.1% of the patients with CTS who did not undergo surgical decompression or injections. These beneficiaries could have been treated conservatively with a brace, demonstrated transient symptoms that did not require treatment, or still undergoing diagnostic work-up or surgical planning at the time of data collection. Finally, we could not differentiate between RSA versus TSA using available CPT and ICD-10 coding. We acknowledge the possibility that RSA may confer additional risks to neuropathies because of distalization and laterialization of the implant design.15,16 Likewise, we grouped post-traumatic indications for shoulder arthroplasty with degenerative indications to enhance statistical power. We recognize that the former can be associated with local scarring and prone to neurological injuries. It is not clear if the effects in this combined cohort are driven by one or the other subgroup.
This study has other limitations to consider. First, as data are derived from claims, there can be inaccuracies in detecting variables used in this study. Second, similar to all observational studies, our investigation is subject to unmeasured confounding bias or even residual confounding. Our study findings can be confounded by factors not attainable in claims data, such as hand dominance, subtle preoperative symptoms, severity of systemic neuropathy, and ergonomic exposures. Third, there is the possibility of surveillance bias, as postoperative follow-up necessarily involves more detailed evaluation of the operative limb. This can increase the likelihood that symptoms of CTS are diagnosed and treated in the ipsilateral hand, thereby inflating the observed incidence compared to the contralateral, less-monitored limb. Lastly, an important limitation of our study is the use of the contralateral limb as an internal comparator for assessing side-specific incidence rates of CTS after shoulder arthroplasty. As highlighted by the work of Camarillo et al17 and Park et al,18 bilateral limb data from the same individuals may not constitute independent observations because of shared systemic, metabolic, and occupational risk factors, as well as the potential for subclinical or bilateral disease. However, our cohort-level approach differs from traditional side-to-side comparisons, as bilateral procedures were excluded and side-specific incidence rates of CTS were independently modeled. For each limb, statistical models were constructed to estimate adjusted CTS rates, comparing cohort-level ipsilateral and contralateral outcomes rather than directly comparing limbs within individuals. To this end, dependency is still possible but less from a statistical perspective and more from a study design perspective.
In conclusion, shoulder arthroplasty is associated with an increased incidence of CTS after surgery. Recognition of this association can lead to improved preoperative patient counseling.
Conflicts of Interest
No benefits in any form have been received or will be received related directly to this article.
Supplementary Data
References
- 1.Dahlin L.B., Zimmerman M., Calcagni M., Hundepool C.A., van Alfen N., Chung K.C. Carpal tunnel syndrome. Nat Rev Dis Primers. 2024;10(1):37. doi: 10.1038/s41572-024-00521-1. [DOI] [PubMed] [Google Scholar]
- 2.US Bureau of Labor Statistics, US Department of Labor Nonfatal occupational injuries and illnesses requiring days away from work, 2015. Published 2011. https://www.bls.gov/news.release/osh2.nr0.htm
- 3.Hubbard Z.S., Law T.Y., Rosas S., Jernigan S.C., Chim H. Economic benefit of carpal tunnel release in the Medicare patient population. Neurosurg Focus. 2018;44(5):E16. doi: 10.3171/2018.1.FOCUS17802. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Wagner E.R., Farley K.X., Higgins I., Wilson J.M., Daly C.A., Gottschalk M.B. The incidence of shoulder arthroplasty: rise and future projections compared with hip and knee arthroplasty. J Shoulder Elbow Surg. 2020;29(12):2601–2609. doi: 10.1016/j.jse.2020.03.049. [DOI] [PubMed] [Google Scholar]
- 5.Lädermann A., Lübbeke A., Mélis B., Stern R., Christofilopoulos P., Bacle G., Walch G. Prevalence of neurologic lesions after total shoulder arthroplasty. J Bone Joint Surg Am. 2011;93(14):1288–1293. doi: 10.2106/JBJS.J.00369. [DOI] [PubMed] [Google Scholar]
- 6.Yian E.H., Dillon M., Sodl J., Dionysian E., Navarro R., Singh A. Incidence of symptomatic compressive peripheral neuropathy after shoulder replacement. Hand (N Y) 2015;10(2):243–247. doi: 10.1007/s11552-014-9701-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Medvedev G., La P., Tidwell M., Owens B.D. The incidence of carpal tunnel syndrome diagnosis increases after arthroscopic shoulder surgery. J Hand Surg Glob Online. 2023;5(5):624–629. doi: 10.1016/j.jhsg.2023.05.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Byl C., Puttlitz C., Byl N., Lotz J., Topp K. Strain in the median and ulnar nerves during upper-extremity positioning. J Hand Surg Am. 2002;27(6):1032–1040. doi: 10.1053/jhsu.2002.35886. [DOI] [PubMed] [Google Scholar]
- 9.Kam A.W., Lam P.H., Haen P.S.W.A., Tan M., Shamsudin A., Murrell G.A.C. Preventing brachial plexus injury during shoulder surgery: a real-time cadaveric study. J Shoulder Elbow Surg. 2018;27(5):912–922. doi: 10.1016/j.jse.2017.11.030. [DOI] [PubMed] [Google Scholar]
- 10.Florczynski M., Paul R., Leroux T., Baltzer H. Prevention and treatment of nerve injuries in shoulder arthroplasty. J Bone Joint Surg Am. 2021;103(10):935–946. doi: 10.2106/JBJS.20.01716. [DOI] [PubMed] [Google Scholar]
- 11.Ghali M., Ehlen Q.T., Kholodovsky E., Cacciatore J., Parrish J., Jenkins N., Dodds S.D. Double crush syndrome: a review of the literature. Hand (N Y) 2025;20 doi: 10.1177/15589447251352122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kwon H.K., Hwang M., Yoon D.W. Frequency and severity of carpal tunnel syndrome according to level of cervical radiculopathy: double crush syndrome? Clin Neurophysiol. 2006;117(6):1256–1259. doi: 10.1016/j.clinph.2006.02.013. [DOI] [PubMed] [Google Scholar]
- 13.Richardson J.K., Forman G.M., Riley B. An electrophysiological exploration of the double crush hypothesis. Muscle Nerve. 1999;22(1):71–77. doi: 10.1002/(sici)1097-4598(199901)22:1<71::aid-mus11>3.0.co;2-s. [DOI] [PubMed] [Google Scholar]
- 14.Morgan G., Wilbourn A.J. Cervical radiculopathy and coexisting distal entrapment neuropathies: double-crush syndromes? Neurology. 1998;50(1):78–83. doi: 10.1212/wnl.50.1.78. [DOI] [PubMed] [Google Scholar]
- 15.Lenoir H., Dagneaux L., Canovas F., Waitzenegger T., Pham T.T., Chammas M. Nerve stress during reverse total shoulder arthroplasty: a cadaveric study. J Shoulder Elbow Surg. 2017;26(2):323–330. doi: 10.1016/j.jse.2016.07.020. [DOI] [PubMed] [Google Scholar]
- 16.Burnett Z., Werner B.C. Risk factors, management, and prognosis of brachial plexopathy following reverse total shoulder arthroplasty. Orthop Clin North Am. 2022;53(2):215–221. doi: 10.1016/j.ocl.2021.11.007. [DOI] [PubMed] [Google Scholar]
- 17.Camarillo N.D., Jiménez-Silva R., Sheehan F.T. Using bilateral data in controls and patients with bilateral and unilateral pathology requires increased scrutiny. J Biomech. 2024;162 doi: 10.1016/j.jbiomech.2023.111855. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Park M.S., Kim S.J., Chung C.Y., Choi I.H., Lee S.H., Lee K.M. Statistical consideration for bilateral cases in orthopaedic research. J Bone Joint Surg Am. 2010;92(8):1732–1737. doi: 10.2106/JBJS.I.00724. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.