Reoperation Rates and Short-Term Complications Following Endoscopic vs. Open Carpal Tunnel Release: A Longitudinal Analysis

Abstract

Background:

Carpal tunnel syndrome (CTS), affecting approximately 8% of the population, is treated with open (oCTR) or endoscopic (eCTR) carpal tunnel release. Previous literature compares outcomes within 1 to 2 years; this study evaluated >5-year reoperation rates and short-term complications using a large electronic health record database.

Methods:

A retrospective analysis using data from the TriNetX Research Network (2007-2024) identified patients with unilateral CTS who underwent either oCTR or eCTR within 1 year of diagnosis, using Current Procedural Terminology (CPT) and International Classification of Diseases (ICD) codes. Propensity score matching and multiple logistic regression calculated adjusted risk and odds ratios (ORs) with 95% confidence intervals (95% CIs) to assess reoperation rates at 2, between 2 and 5, >5 years after operation and 90-day postoperative complications (wound dehiscence, surgical site infection [SSI]).

Results:

Within 2 years of CTR, reoperation rate was higher for eCTR than that for oCTR (relative risk [RR] = 1.15, 95% CI = 1.09-1.22; OR = 1.36, 95% CI = 1.21-1.53). Beyond 5 years, the revision rate of the two approaches was similar (RR = 0.85, 95% CI = 0.74-1.01; OR = 0.76, 95% CI = 0.58-1.00). The number needed to treat to prevent one reoperation within 2 years was 67, and beyond 5 years, it was 473. Within 90 days of surgery, eCTR was associated with decreased wound dehiscence (RR = 0.67, 95% CI = 0.53-0.85; OR = 0.50, 95% CI = 0.36-0.71) and SSI (RR = 0.77, 95% CI = 0.65-0.91; OR = 0.63, 95% CI = 0.48-0.81).

Conclusion:

This study demonstrates the clinical insignificance of the difference in early CTR revision rate between approaches and that eCTR necessitates a similar reoperation rate at long term, supporting eCTR to remain an appropriate intervention for CTR.

Introduction

Carpal tunnel syndrome (CTS) is one of the most common entrapment neuropathies, reportedly affecting up to 8% of the general population in the United States and leading to significant pain, numbness, and functional impairment.^1,2 Surgical intervention, specifically carpal tunnel release (CTR), is a widely accepted treatment for relieving symptoms and improving hand function in patients with moderate to severe CTS who do not respond to conservative management. ³ The two primary surgical techniques for CTR include the traditional open carpal tunnel release (oCTR) and the endoscopic carpal tunnel release (eCTR). Both approaches aim to decompress the median nerve by releasing the transverse carpal ligament, but they differ in terms of surgical exposure, visualization, and postoperative recovery.^3,4

According to the American Academy of Orthopedic Surgeons Clinical Practice Guidelines for surgical CTS management, the efficacy of both oCTR and eCTR is supported by strong evidence. Regardless, an abundance of literature has previously compared many different aspects of these two techniques, including but not limited to complication rate, revision rate, recovery time, cost of care, and procedure length. Employing eCTR can facilitate a quicker recovery time and return to work,^{4 -6} whereas oCTR incurs a lower all-cause revision rate within 1 to 2 years while also reporting lower costs of care and procedure length.^{7 -12} Existing literature has failed to establish one clear superior option, and therefore, both approaches are commonly employed.

Importantly, it is evident that existing literature comparing the long-term outcomes of oCTR to those of eCTR is extremely limited, with most studies focused on perioperative and short-term postoperative outcomes, typically within 1 to 2 years of the procedure.^{4,7 -12} There is an especially notable paucity of research that evaluates and compares the necessity for revision of oCTR and eCTR beyond the 5-year postoperative period.

Given the growing preference for minimally invasive procedures, ¹³ understanding the long-term effectiveness and safety of eCTR compared to oCTR is crucial for informed surgical decision-making. The primary aim of this study is to address the gap in long-term comparative data by evaluating the reoperation rates of oCTR and eCTR beyond the 5-year postoperative period, while also assessing the short-term complication rates, such as surgical site infection (SSI) and wound dehiscence, using a large electronic health record database.

Methods

Data Source

The TriNetX Research Network database (Cambridge, Massachusetts) is a global health-collaborative clinical research platform collecting real-time electronic medical data from more than 400 million patients, 170 health care organizations, and 19 countries. ¹⁴ The organizations participating in the TriNetX network, including academic medical centers, community hospitals, referral center, and other types of healthcare centers, provide health care data in deidentified, pseudo-anonymized, or limited data set formats, adhering to local privacy regulations. These organizations authorize the use of their data for research purposes on the TriNetX platform. In exchange for their data, contributing organizations incur no financial costs and receive access to data query tools, analytics, visualization capabilities, and the necessary hardware for software execution. TriNetX is certified to the ISO 27001:2013 standard and maintains an Information Security Management System to ensure the protection of the health care data and to meet the requirements of the Health Insurance Portability and Accountability Act (HIPAA) Security Rule. In addition, the deidentification process complies with the HIPAA Privacy Rule standards, as verified by a qualified expert, fulfilling the requirements of Section 164.514(b)(1). For this study, the data reviewed were from a secondary analysis of existing data. Collection did not involve an intervention or interaction with human subjects. Because this database is free from all personal health information, this study is exempt from institutional review board approval and informed consent.

Data Preparation

All data used in this study were obtained from the TriNetX Research Network database, which provided access to electronic medical records (diagnoses, procedures, medications, laboratory values, genomic information). All data accessed were collected between 2007 and 2024. However, for this study, all index CTR surgeries were performed before 2019, ensuring sufficient follow-up time to evaluate both short-term (within 2 years) and long-term outcomes (beyond 5 years). Reoperation data were collected up to 2024 to capture revisions occurring during these predefined time frames. Inclusion criteria, exclusion criteria, and grouping for analyses were based on classifications using International Classification of Diseases, Nineth and Tenth Revision, Clinical Modification (ICD-9-CM and ICD-10-CM) codes, Current Procedural Terminology (CPT) codes (appendix), and individual procedure characteristics data collected by and available through TriNetX.

The initial query identified patients with unilateral diagnoses of CTS, followed by CTR within 1 year of diagnosis. Two distinct groups were created: patients with right-handed CTS who underwent right-handed CTR (excluding left-handed and bilateral CTS) and patients with left-handed CTS who underwent left-handed CTR (excluding right-handed and bilateral CTS). These criteria ensured consistency in the dataset and accounted for the correlation between the duration of CTS symptoms, severity, and the likelihood of symptom resolution following CTR. ¹⁵ The initial cohort included 79 496 patients with a history of CTR. After duplicates and individuals with missing data or incomplete demographic information were removed, a final dataset of 43 865 patients was established, comprising 32 512 who underwent oCTR (CPT 64721) and 11 353 who underwent eCTR (CPT 29848).

Matching

Data were then prepared for propensity matching. The covariates chosen for propensity matching included sex, race, ethnicity, marital status, age, body mass index (BMI), disorders involving immune mechanisms, rheumatoid disease, gout, amyloidosis, hypothyroidism, diabetes mellitus, and date of CTR. Patients that did not have data available for all demographic and diagnostic factors were excluded.

1:1 nearest neighbor propensity score matching was then performed, matching the two groups on the previously listed covariates. All matching and analyses were performed in R. ¹⁶ The MatchIt package performed the 1:1 nearest neighbor propensity score matching without replacement, where the propensity score was estimated using a logistic regression of the CTR approach on the covariates. ¹⁷ The standardized mean differences (SMDs) for all matched variables were measured before and after matching. A successful matching threshold was defined by SMD < 0.1. Matching paired 1 oCTR patient to every eCTR patient, and all nonmatched individuals were excluded. The postmatch population included 11 353 oCTR patients and 11 353 eCTR patients. All absolute SMDs and overall mean absolute SMD of covariates were less than 0.1, indicating a successful match.

Comparisons were then made between groups for necessity of CTR revision within 2 years postoperatively and beyond 5 years postoperatively. CTR revision was indicated for persistent, recurrent, or new carpal tunnel symptoms. Incidence of SSI and wound dehiscence within 90 days of CTR was also compared.

Statistical Analysis

Statistical analyses compared the risk and odds of adverse postoperative outcomes, including SSI, wound dehiscence, and reoperation, between oCTR and eCTR procedures. Adjusted statistical tests were performed to further ensure no residual effect of covariates on statistical outcomes.

Adjusted risk ratio and odds ratios were calculated using three multiple logistic regressions of the type of CTR on our three postoperative outcomes. Covariates in the regressions included sex, race, ethnicity, marital status, age, BMI, disorders involving immune mechanisms, rheumatoid disease, gout, amyloidosis, hypothyroidism, diabetes mellitus, and date of CTR. Logistic regression was performed using the native R linear regression function with the weights assigned by the propensity match. The risk ratio and odds ratio with 95% confidence interval were estimated from the multiple logistic regression using the average comparisons function of the “marginaleffects” package in R. ¹⁸ The comparison “lnratioavg” was used to estimate risk ratios and the comparison “lnoravg” to estimate odds ratios within the average comparisons function.

Results

In this study population, 14 079 patients (62%) were female, and 8627 (38%) were male. The study population was predominantly non-Hispanic (94.5%) and white (84%), with other races including black, Asian, American Indian, Native Hawaiian, and others. The mean age of the study population was 60.2 years. After exclusion, grouping, and matching before analysis, the two groups were similar regarding demographic characteristics, comorbidities, and dates of procedures (Figure 1).

Figure 1.

Absolute standardized mean difference between groups before and after matching.

Note. Absolute SMD represent the absolute difference in mean outcome between groups divided by the standard deviation of that outcome among all participants. Any covariate with a postmatch SMD of <0.1 is considered well matched. All covariates had a postmatch SMD of <0.1 indicating a successful match; SMD = standardized mean differences; BMI = body mass index.

Within 2 years after CTR, 501 of 11 353 (4.41%) patients who underwent oCTR and 670 of 11 353 (5.90%) patients who underwent eCTR required surgical revision (Figure 2). These percentages represent reoperations occurring specifically within the first 2 years. For revisions occurring between 2 and 5 years postoperatively, there was no significant difference between groups (RR = 1.07, 95% CI = 0.93-1.23; OR = 1.15, 95% CI = 0.87-1.50), with reoperation rates of 0.92% for oCTR and 1.15% for eCTR during this interval (Figure 3). After 5 years, reoperation rates were further reduced, with 104 of 11 353 (0.92%) patients in the oCTR group and 80 of 11 353 (0.70%) patients in the eCTR group requiring revision (Figure 4). These percentages represent separate intervals of observation and are not cumulative, as fewer reoperations naturally occur over time.

Figure 2.

Incidence of reoperation within 2 years after CTR.

Note. This bar graph represents the total incidence of reoperation within 2 years after CTR between the endoscopic and open group. There were 670 reoperations in the endoscopic group (5.90%) and 501 in the open group (4.41%). The difference between the incidence of reoperation between these groups was significant (RR = 1.15, 95% CI = 1.09-1.22 and OR = 1.36, 95% CI = 1.21-1.53); CTR = carpal tunnel release; RR = relative risk; OR = odds ratio; CI = confidence interval.

Figure 3.

Incidence of reoperation within 2 to 5 years after CTR.

Note. This bar graph represents the total incidence of reoperation between 2 and 5 years after CTR between the endoscopic and open group. 0.92% for oCTR operations required reoperation, and 1.15% of eCTR operations required reoperation during this interval. The difference between the incidence of reoperation between these groups was not significant (RR = 1.07, 95% CI = 0.93-1.23; OR = 1.15, 95% CI = 0.87-1.50); CTR = carpal tunnel release; RR = relative risk; OR = odds ratio; CI = confidence interval.

Figure 4.

Incidence of reoperation more than 5 years after CTR.

Note. This bar graph represents the total incidence of reoperation >5 years after CTR between the endoscopic and open group. There were 80 reoperations in the endoscopic group (0.70%) and 104 in the open group (0.92%). The difference between the incidence of reoperation between these groups was not significant (RR = 0.85, 95% CI = 0.74-1.01 and OR = 0.76, 95% CI = 0.58-1.00); CTR = carpal tunnel release; RR = relative risk; OR = odds ratio; CI = confidence interval.

Within 2 years after operation, 67 oCTR procedures are needed instead of eCTR to prevent one reoperation (the number needed to treat [NNT]). For reoperations occurring between 2 and 5 years postoperatively, the NNT increases to 1032 oCTR procedures. Beyond 5 years, the NNT decreases to 473 oCTR procedures to prevent 1 reoperation. The mean time to reoperation for all patients was 504.4 days for eCTR and 783.9 days for oCTR. For reoperations within 2 years, the mean time to revision was shorter for both groups, with eCTR averaging 117.7 days and oCTR averaging 150.7 days. For reoperations occurring beyond 5 years, the mean time to revision was 2775.6 days for eCTR and 3439.5 days for oCTR. It is important to note that these averages may be influenced by a small number of patients, with revisions occurring more than 10 years after their initial surgery, which could pull the averages higher.

Multiple logistic regression also demonstrated that relative to oCTR, eCTR was associated with decreased incidence of both wound dehiscence (RR = 0.67, 95% CI = 0.53-0.85 and OR = 0.50, 95% CI = 0.36-0.71) (Figure 5) and SSI (RR = 0.77, 95% CI = 0.65-0.91 and OR = 0.63, 95% CI = 0.48-0.81; Figure 6) within 90 days after operation.

Figure 5.

Incidence of dehiscence within 90 days of CTR.

Note. This bar graph represents the total incidence of dehiscence <90 days after CTR between the endoscopic and open group. There were 45 instances of dehiscence in the endoscopic group (0.40%) and 89 in the open group (0.78%). The difference between the incidence of dehiscence between these groups was significant (RR = 0.67, 95% CI = 0.53-0.85 and OR = 0.50, 95% CI = 0.36-0.71); CTR = carpal tunnel release; RR = relative risk; OR = odds ratio; CI = confidence interval.

Figure 6.

Incidence of infection within 90 days of CTR.

Note. This bar graph represents the total incidence of infection <90 days after CTR between the endoscopic and open group. There were 84 instances of infection in the endoscopic group (0.74%) and 134 in the open group (1.18%). The difference between the incidence of infection between these groups was significant (RR = 0.77, 95% CI = 0.65-0.91 and OR = 0.63, 95% CI = 0.48-0.81); CTR = carpal tunnel release; RR = relative risk; OR = odds ratio; CI = confidence interval.

Several covariates were significantly associated with adverse outcomes. Male patients were more likely to experience wound dehiscence after CTR (RR = 1.41, 95% CI = 1.21-1.64; OR = 1.87, 95% CI = 1.37-2.55). Race was significantly associated with reoperation rates (χ² = 20.778, P < .001), although the direction and magnitude of differences could not be determined. Ethnicity was also significantly associated with reoperation, with non-Hispanic individuals experiencing higher reoperation rates (6.98%) than Hispanic individuals (4.20%) (χ² = 13.957, P < .001). Younger age was associated with higher rates of reoperation, wound dehiscence, and SSI, although the differences in mean ages were minimal (e.g., reoperation mean age = 57.23 vs. no reoperation = 60.44). Similarly, BMI was positively associated with all three adverse outcomes, with reoperation patients averaging a BMI of 32.02 compared to 30.70 for those without reoperation.

Visualization of CTR reoperation by group shows that early in the postoperative period, the eCTR approach may require more frequent revisions. However, this difference diminishes over time as the interval since the operation increases (Figure 7).

Figure 7.

(a) Percent of operations requiring reoperation by month (<5 years). (b) Percent of operations requiring reoperation by year (2-10 years).

Note. (a) This line graph represents the percentage of procedures for which a reoperation was performed every month between 0 and 5 years after CTR. (b) This line graph represents the percentage of procedures for which a reoperation was performed every year between 2 and 10 years after CTR; CTR = carpal tunnel release.

Discussion

This study leverages a large, diverse database of over 43 000 patients, offering one of the most comprehensive evaluations of long-term outcomes for oCTR and eCTR. Using advanced methods like propensity score matching and adjusted logistic regression, it ensures valid comparisons while minimizing confounding. As the first study to assess reoperation rates beyond 5 years alongside short-term complications such as wound dehiscence and SSI, it provides actionable insights for surgical decision-making and advances understanding of both immediate and long-term outcomes.

Existing literature has extensively explored outcomes, revision rates, and complications for oCTR and eCTR, but most studies lack the extended follow-up provided by this study. Ferrin et al, ¹⁹ using a national cohort, reported a higher hazard of revision for eCTR than for oCTR (hazard ratio 1.56) while emphasizing the low absolute revision rates (1.59% over 10 years). This study’s longer follow-up aligns with our findings, particularly regarding the durability of eCTR over time. In contrast, Carroll et al ²⁰ focused on 1-year revision rates and identified eCTR as a risk factor for revision, particularly in patients with diabetes and tobacco use. While valuable for understanding short-term outcomes, studies like Carroll et al are limited by their shorter follow-up periods and narrower scopes, leaving critical gaps in understanding long-term surgical durability and safety.

Other studies, such as those by Haglin et al ²¹ and Williamson et al, ¹² emphasized the growing popularity and higher costs of eCTR, noting its steeper learning curve and increased adoption over time. These studies predominantly analyzed procedural trends, costs, and utilization without delving into detailed clinical outcomes. In addition, works like Westenberg et al ⁸ and Wessel et al ⁹ identified patient-specific risk factors for revision but focused on smaller cohorts. In contrast, our study’s large sample size, multicenter database, and inclusion of both short- and long-term outcomes position it uniquely within this body of literature. By addressing short-term complications such as wound dehiscence and SSI alongside long-term revision rates, this study provides a comprehensive perspective that directly informs clinical decision-making and patient care. These contributions underscore its value in bridging critical evidence gaps and advancing the understanding of eCTR and oCTR outcomes.

Additionally, this study demonstrates that within the 90-day postoperative period, eCTR is associated with decreased rates of infection and wound dehiscence. In another large database study, Kishan et al ²² demonstrate findings that reinforce these findings; in a cohort of patients with type 2 diabetes mellitus, the eCTR approach was associated with lower rates of median nerve injury, wound dehiscence, and SSI at 6 weeks after operation. However, in a similar large database study, Khalid et al ²³ reported no difference in rates of nerve injury but also demonstrated that oCTR was associated with a higher risk of developing wound complications and SSI. Based on this evidence, it is likely eCTR is associated with lower rates of postoperative complications related to surgical site dehiscence and infection.

The clinical implications of these findings suggest that postoperative outcomes may benefit from tailoring to the specific risks associated with eCTR and oCTR. While eCTR presents a slightly higher risk of revision in the short-term postoperative period, the long-term outcomes, particularly beyond 5 years, demonstrate no significant difference in the need for reoperation between the two techniques. This indicates that both methods are equally viable in terms of long-term efficacy. Furthermore, the reduced rates of infection and wound dehiscence associated with eCTR, particularly within the first 90 days, highlight its potential advantages for patients at higher risk of wound complications, such as those with diabetes, ²⁴ where it has been illustrated that with each 1.0% point increase in HbA1c, the daily wound-area healing rate decreased by 0.028 cm²/day. ²⁵ While our findings do not provide direct evidence to warrant changes in postoperative care protocols, they emphasize the importance of careful patient selection and risk assessment when choosing between surgical techniques. These insights can help guide surgeons and care teams in optimizing patient outcomes by considering individual risk profiles and procedural benefits.

Limitations

The authors acknowledge several limitations in this study. First, as a retrospective large database study, the analysis relied on the availability of variables, completeness of data, and accuracy of coding and entry. Information on the severity of CTS and the degree of functional impairment was unavailable, limiting the ability to account for how baseline disease severity may have influenced the likelihood of reoperation or outcomes after primary or revision surgery.

In addition, the study lacked data on the specific causes of reoperation. While patients requiring CTR revision were categorized into persistent, recurrent, and new symptom groups, detailed insights into the underlying reasons for revision surgery, such as incomplete symptom relief, recurrence, immediate worsening of symptoms, or complications from the initial surgery, were not available. ²⁶ This limitation may obscure differences in outcomes between these subgroups and limits the ability to draw nuanced conclusions about the true epidemiology of revisions.

Due to ICD-9 and ICD-10 limitations, postoperative complications such as hematoma formation or nerve injury, which could influence both short- and long-term outcomes, were not captured in the dataset. The inability to assess symptom relief after both primary and revision surgeries further restricts the ability to evaluate the overall effectiveness of the interventions. In addition, information on specific surgical details—such as whether revisions were performed using the same or a different technique, or whether the same surgeon performed both the initial and revision procedures—was unavailable.

Finally, while this study used a large real-world dataset, it lacked patient-reported outcomes, such as symptom relief or functional improvement, which would provide a more comprehensive understanding of surgical success. These limitations highlight the need for future studies with more granular clinical data to better elucidate the factors influencing reoperation and postoperative outcomes.

Conclusion

In conclusion, this study provides the first comprehensive comparison of long-term outcomes between eCTR and oCTR. While eCTR may carry a higher risk of revision in the early postoperative period, the NNT values associated with this comparison reflect the potential clinical insignificance of this difference. Importantly, the approaches show similar long-term efficacy. Lasly, eCTR was associated with lower rates of infection and wound dehiscence, particularly in the first 90 days after surgery. These findings should inform surgical decision-making and postoperative care, with a focus on patient-specific risk factors to optimize outcomes.

Appendix

ICD-10 Diagnosis Codes and Current Procedural Terminology Used for Exclusion Criteria and Grouping.

Open Carpal Tunnel Release (ICD-10-PCS-01N50ZZ and CPT-64721)

Endoscopic Carpal Tunnel Release (ICD-10-PCS- 01N54Z and CPT-29848)

Surgical Site Infection (ICD-10-CM-T81.4)

Wound Dehiscence (ICD-10-CM-T81.31)