Is Parkinson Disease a Risk Factor After Total Elbow Arthroplasty? Midterm Survivorship in a Matched Cohort

Abstract

Purpose

Parkinson disease (PD) has been associated with elevated rates of periprosthetic fracture, periprosthetic joint infection, and reduced implant survivorship following total hip and knee arthroplasty; however, it remains unclear whether these risks extend to total elbow arthroplasty (TEA). The purpose of this study was to evaluate whether PD is associated with worse midterm survivorship and higher complication risk after TEA compared with matched controls.

Methods

A retrospective, matched cohort study was performed using the PearlDiver administrative claims database. Adult patients undergoing primary TEA were identified, and PD was defined by diagnostic claims recorded on or before the index procedure. Patients with PD were matched 1:4 to non-PD controls by age and sex. Primary end points were aseptic loosening, periprosthetic joint infection, and periprosthetic fracture. Secondary outcomes included revision TEA, operative fixation of periprosthetic fracture, and a composite adverse outcome defined as the first occurrence of any complication or reoperation. Patients were followed until first event, end of enrollment, or 5 years. Kaplan-Meier analysis was used to estimate event-free survivorship, with log-rank testing and Cox proportional hazards regression used for comparison.

Results

After matching, 675 patients were included: 135 with PD and 540 controls. Overall adverse events were uncommon, and survivorship remained high across all end points. There was no significant difference in 5-year composite adverse event-free survivorship between PD and controls (84.4% [95% CI: 77.9−91.4] vs 83.4% [95% CI: 79.9−87.0]; log-rank P = .69). Loosening-free survivorship was also similar between groups at 5 years (96.4% [95% CI: 92.9−100] in PD vs 95.7% [95% CI: 93.7−97.8] in controls; P = .68). No significant differences were observed in survivorship for periprosthetic joint infection or periprosthetic fracture. The study had approximately 80% power to detect a hazard ratio of about 2.0 for adverse events; smaller differences may have gone undetected.

Conclusions

Parkinson disease was not associated with worse 5-year survivorship or higher risk of implant-threatening complications after TEA. These findings suggest that PD diagnosis alone may not confer increased midterm risk following TEA.

Type of study/level of evidence

Prognostic III, retrospective matched cohort study.

Parkinson disease (PD) affects motor control, balance, and autonomic function in ways that can complicate perioperative care and postoperative recovery.¹ Parkinson disease also creates unique medication vulnerabilities because delays or interruptions in dopaminergic therapy during hospitalization can worsen rigidity, tremor, and delirium and may contribute to serious syndromes, such as parkinsonism-hyperpyrexia.² In the context of hand surgery, PD is considered a high-risk comorbid condition because it has been associated with higher complication rates (particularly neurocognitive and medical events), prolonged hospital stays, increased readmissions rates after major elective procedures, and greater hospital charges.3, 4, 5

Upper-extremity outcomes in patients with PD raise additional concerns because tremor, rigidity, and impaired balance can magnify mechanical stress and fall risk during recovery.⁶ Reported adverse outcomes include loss of reduction, nonunion/malunion, and overall higher complication and revision rates, with instability/dislocation as a prominent failure pattern.^6,7 Shoulder arthroplasty has been shown to improve pain in patients with PD; however, functional gains are less predictable and complications remain common, which complicates preoperative counseling and procedural selection in this population.⁸

Total elbow arthroplasty (TEA) presents a separate challenge because it carries one of the highest complication and revision burdens among major joint arthroplasties.⁹ The most consequential failure modes after TEA include periprosthetic joint infection (PJI), aseptic loosening, and periprosthetic fracture (PPF), events that often trigger complex revision surgery and can permanently limit function.⁹ When TEA infection occurs, eradication can be difficult across revision strategies, and outcomes vary substantially by approach, emphasizing the clinical importance of preventing infection and identifying risk factors for failure.¹⁰ Even with staged revision protocols, infected TEA can carry a high downstream re-revision burden because of recurrent infection, loosening, or fracture, highlighting how a single infectious event can threaten long-term implant survivorship.¹¹ Although the effect of PD has been investigated in other areas of arthroplasty including shoulder, hip, and knee, the TEA-specific PD literature is limited and largely consists of small reports.^8,12, 13, 14 Given this evidence gap, surgeons often must extrapolate from lower-extremity arthroplasty and shoulder arthroplasty data when counseling patients with PD about TEA risk, which may not reflect elbow-specific failure patterns.⁸

To address this gap, we used a large administrative claims database to evaluate midterm TEA survivorship in patients with and without PD. Our primary objective was to compare 5-year complication-free survivorship between cohorts with and without PD after TEA, focusing on aseptic loosening, PJI, and PPF as clinically meaningful end points. We hypothesized that PD would be associated with lower midterm TEA survivorship and higher complication risk compared with matched controls, given PD’s documented perioperative vulnerability and adverse orthopedic outcomes in other settings.

Materials and Methods

We conducted a retrospective cohort study using the PearlDiver administrative claims database. PearlDiver contains deidentified, Health Insurance Portability and Accountability Act–compliant inpatient and outpatient medical and procedural claims data with longitudinal follow-up.

Cohort definitions and index dates

We identified adults who underwent TEA using procedural claims data (see Table S1, available online on the Journal’s website at https://www.jhsgo.org, for the full code list). We defined the index date as the date of the first recorded TEA for each patient. Within the TEA population, we identified patients with PD using diagnostic claims recorded before or on the index TEA date (Table S1). We assigned patients to the PD cohort if they met PD criteria and to the control cohort if they had no PD diagnosis recorded at any time in their available claims history.

We recorded baseline demographic characteristics, including age and sex, at the time of the index TEA. We also captured select comorbid conditions relevant to elbow arthroplasty indications and inflammatory arthropathy (Table S2, available online on the Journal’s website at https://www.jhsgo.org).

Matching strategy

We matched patients with PD to patients without PD in a 1:4 ratio based on age and sex. After matching, we summarized cohort characteristics to confirm demographic balance between groups (Table).

Table.

Breakdown of the Study Cohort by Age and Sex

Characteristic	Patients With Parkinson Disease (n = 135)		Controls (n = 540)
	No.	%	No.	%
Age, y
55–59	7	5.2	28	5.2
60–64	23	17.0	52	9.6
65–69	27	20.0	108	20.0
70–74	34	25.2	136	25.2
75–79	43	31.9	172	31.9
80+	5	3.7	20	3.7
Other	6	4.4	24	4.4
Sex
Female	103	76.3	412	76.3
Male	32	23.7	128	23.7

Outcome definition and follow-up

The primary end points were survivorship-threatening complications after TEA, including aseptic loosening, PJI, and PPF (Table S1). We also evaluated clinically meaningful reinterventions, including revision elbow arthroplasty and operative fixation for distal humerus/PPF patterns (Table S1). We defined a composite adverse outcome as the first occurrence of any of these complications or reoperations during follow-up.

We followed each patient from the index TEA until the first occurrence of the outcome of interest, the end of available follow-up, or 5 years, whichever occurred first. Patients who did not experience the outcome during follow-up were censored at their last available claim or at 5 years.

Statistical analysis

We estimated 5-year event-free survivorship using Kaplan-Meier analysis for the matched PD and control cohorts. We generated Kaplan-Meier curves for the composite adverse outcome and separately for loosening, PJI, PPF, and distal operative fixation. We compared survivorship between cohorts using log-rank tests. We then performed Cox proportional hazards regression with PD status as the independent variable and reported hazard ratios (HRs) with 95% confidence intervals (CIs). All tests were two-sided, and P < .05 was considered statistically significant. All analyses were formed using the built-in statistical software within the PearlDiver platform.

Data availability statement

The data that support the findings of this study were obtained from the PearlDiver administrative claims database. These data are available from PearlDiver Technologies through licensed access and are not publicly available to the authors for direct distribution.

Results

After matching, the final study sample included 675 patients who underwent TEA, including 135 patients with PD and 540 matched controls. In the overall matched cohort, complications after TEA were uncommon, and survivorship remained high across all outcomes. There were no significant differences in overall adverse event-free survival rates between patients with PD and matched controls (χ² = 0.162, P = .69) (Fig. 1). At 5 years, adverse event-free survival rates were similar between groups (for patients with PD, HR = 84.4%; 95% CI: 77.9–91.4; for controls, HR = 83.4%; 95% CI: 79.9–87.0). On the basis of the observed number of adverse events over 5 years in the matched cohort (total events = 91), the study had 80% power (two-sided α = 0.05) to detect an HR of approximately 2.1 (or 0.48 in the opposite direction) for adverse events between groups. Smaller effect sizes may not be detectable with the available sample.

Figure 1.

Adverse event-free survival rates for patients with PD and matched controls.

Aseptic loosening was rare in both cohorts. There were no significant differences in loosening-free survivorship between patients with PD and controls (χ² = 0.169, P = .68) (Fig. 2). Five-year loosening-free survival rates remained high in both groups (for patients with PD, HR = 96.4%; 95% CI: 92.9–100; for controls, HR = 95.7%; 95% CI: 93.7–97.8).

Figure 2.

Probability of survival without aseptic loosening in patients with PD and matched controls.

Periprosthetic joint infection-free survivorship also did not differ significantly between cohorts (χ² = 1.01, P = .32) (Fig. 3). At 5 years, infection-free survival rates remained high in both groups, with a small, nonsignificant trend toward higher survivorship among patients with PD (for patients with PD, HR = 92.2%; 95% CI: 87.3–97.3; for controls, HR = 89.3%; 95% CI: 86.4–92.4).

Figure 3.

Probability of survival without PJI in patients with PD and matched controls.

Periprosthetic fractures were infrequent, and fracture-free survivorship did not differ between patients with PD and controls (χ² = 1.46, P = .23) (Fig. 4). At 5 years, fracture-free survival rates remained excellent in both groups (for patients with PD, HR = 98.2%; 95% CI: 95.6–100; for controls, HR = 96.2%; 95% CI: 94.3–98.1).

Figure 4.

Probability of survival without PPFs in patients with PD and matched controls.

Discussion

Parkinson disease poses distinct challenges in orthopedic surgical care. Despite the overlapping risk profiles of PD and TEA, surgeons lack clear TEA-specific data to guide counseling and risk stratification for patients with PD.⁹

In the present large, matched cohort study, PD was not associated with worse midterm TEA survivorship. Adverse event-free survival rates were similar between patients with PD patients and matched controls through 5 years, and we observed no significant differences in aseptic loosening, PJI, or PPF. Kaplan-Meier curves showed substantial overlap across all end points, and 5-year survivorship estimates remained high in both cohorts. Together, these findings suggest that PD alone may not meaningfully increase the risk of implant-threatening mechanical or infectious complications after TEA within a 5-year horizon.

A neutral association between PD and TEA survivorship is biologically and clinically plausible when considered alongside the broader arthroplasty literature. Large database studies and pooled analyses have consistently shown that PD increases perioperative vulnerability, often driven by medical and neurocognitive complications and greater resource utilization, yet have not uniformly demonstrated higher rates of deep periprosthetic infection or revision across joints.⁵ For example, after total hip arthroplasty, patients with PD are at higher risk of delirium, altered mental status, urinary tract infection, and other medical events that prolong hospitalization and complicate recovery, but these complications do not necessarily directly affect loosening, infection, or fracture of the implant.⁴ This distinction matters when interpreting our end points: TEA survivorship outcomes reflect the presence or absence of relatively discrete mechanical and infectious failures, whereas many PD-related perioperative hazards may primarily affect short-term physiology, mobility, and discharge disposition rather than the integrity of the implant-bone interface.¹ It is also possible that modern perioperative workflows that prioritize medication reconciliation and timely dopaminergic dosing reduce preventable PD-related decompensation, further attenuating differences in implant-threatening complications between groups.² In addition, because many TEAs are performed in outpatient or short-stay settings, some PD-related perioperative morbidities may be less relevant to this procedure or may not fully be captured in administrative claims data.

Nevertheless, the existing lower-extremity arthroplasty literature demonstrates that PD can significantly elevate the risk of specific implant-threatening complications. After total hip arthroplasty (THA), PPF is a predominant concern, with meta-analytic data showing a nearly eightfold increase in risk for PD patients (OR = 7.93), and the Swedish Hip Arthroplasty Register has confirmed significantly higher revision risk in PD patients at 9 years (5.44% vs 1.75%).^15,16 Similarly, after TKA, PD has been associated with a 7.48-fold increased odds of PPF and significantly reduced implant survivorship (88.8% vs 94.3% at 13 years), with a 2.7 times higher risk of implant failure.^17,18 Revision rates in patients with PD have been reported at 23.6% at a mean 5.3-year follow-up, with matched pair analyses confirming higher rates of early revision surgery.19, 20, 21 Notably, the pattern of complications differs by joint, with fracture and dislocation predominating after THA and infection appearing more common after TKA.¹⁹ In contrast to these findings from the hip and knee arthroplasty literature, the present study did not demonstrate an association between PD and worse midterm TEA survivorship for any of the analogous end points. This discrepancy may reflect fundamental differences in the biomechanical demands, failure mechanisms, and loading environments across joints. Total elbow arthroplasty operates in a non-weight–bearing context, which may attenuate the contributions of PD-related fall risk and impaired gait that drive PPF in the lower extremity. Additionally, the relatively lower overall volume and complication burden of TEA may limit the statistical power to detect modest differences, particularly given the rarity of individual complications such as loosening and fracture in the elbow.

Our findings also highlight that PD-associated arthroplasty risk may be joint and mechanism specific and that TEA may not behave like other upper-extremity arthroplasties. In the shoulder arthroplasty literature, PD is often linked to higher complication and revision rates, and instability/dislocation appears to be a prominent driver of failure.⁷ Evidence from systematic reviews suggests that although shoulder arthroplasty can improve pain in patients with PD, complication and reoperation rates remain substantial, and functional improvement is less predictable.⁸ Total elbow arthroplasty has a different spectrum of survivorship-threatening complications, and larger TEA series typically emphasize aseptic loosening, infection, and PPF as the dominant implant failure modes.⁹ This difference in failure mechanisms may partially explain why we did not observe a PD signal for the end points studied. At the same time, TEA complications that are clinically important but not captured in our survivorship framework—such as heterotopic ossification, postoperative stiffness, and limited range of motion—may still occur disproportionately in patients with PD and deserve separate study.¹²

These results have practical implications for patient counseling and perioperative planning. For patients with PD who meet standard indications for TEA, clinicians should not assume substantially worse midterm implant survivorship solely because of PD diagnosis. This may be especially relevant in shared decision-making when patients and surgeons weigh TEA against nonsurgical management or alternative procedures in complex elbow pathology. However, PD remains a meaningful perioperative risk condition, and outcomes likely depend on optimizing medical management, maintaining timely dopaminergic therapy, and proactively mitigating aspiration and delirium risk.^1,2 Given the morbidity rates associated with TEA infection and the complexity of revision pathways when infection occurs, attention to perioperative optimization and infection prevention remains essential for all TEA patients, including those with PD.¹⁰

Our study has several strengths. We used a large administrative claims database with longitudinal follow-up, which allowed evaluation of relatively uncommon TEA complications within a matched cohort. We applied a consistent time-to-event framework to assess midterm survivorship and compared outcomes between demographically balanced cohorts with and without PD. These features allowed us to evaluate a clinically relevant question in a setting in which a single-center series would be underpowered and potentially biased by institutional practice patterns.

Several limitations warrant consideration. As with all claims-based studies, our analysis is subject to misclassification of diagnoses, procedures, and outcomes, and we could not verify clinical details, such as radiographic loosening, culture results, or surgical technique. We could not account for PD severity, disease duration, symptom control, frailty, functional status, or specific perioperative medication regimens, all of which may influence complication risk and recovery. Total elbow arthroplasty-specific factors, such as surgical indications (trauma vs inflammatory arthritis vs degenerative disease), implant design and constraint, fixation method, surgeon volume, and postoperative rehabilitation protocols, are not reliably captured in claims data and may confound the relationship between PD and outcomes. In addition, our end points focused on implant-threatened complications and reoperations; we could not evaluate patient-reported outcomes, elbow range of motion, stiffness/heterotopic ossification, or satisfaction, which may be particularly important domains in patients with PD even when implant survivorship remains intact.^8,12 Finally, although adverse events were not rare in this cohort, TEA remains a low-volume procedure, and the overall number of events may still limit power to detect modest differences between-group. As a result, our null findings argue against large differences in midterm survivorship, but we cannot exclude smaller increases in risk or later divergence beyond 5 years.

Future work should evaluate longer-term TEA survivorship and incorporate outcomes that better reflect functional recovery in PD, including range of motion, stiffness, heterotopic ossification, and patient-reported function.¹² Stratifying by TEA indication and incorporating proxies for PD severity and frailty may help identify subgroups at higher risk and clarify whether PD influences specific failure mechanisms over longer follow-up.

In conclusion, in this matched claims-based analysis, PD was not associated with worse 5-year TEA survivorship or higher risks of aseptic loosening, PJI, or PPF. These data support shared decision-making and emphasize that perioperative optimization, rather than PD diagnosis alone, should guide counseling and risk mitigation for patients undergoing elbow arthroplasty.

Conflicts of Interest

No benefits in any form have been received or will be received related directly to this article.