Racial differences in reoperation rates following dacryocystorhinostomy

Abstract

Background

Dacryocystorhinostomy is a commonly performed procedure for nasolacrimal duct obstruction with generally favorable outcomes. However, variability in long-term success has been reported, and the role of racial and ethnic factors remains incompletely understood. This study aimed to assess differences in long-term reoperation risk following dacryocystorhinostomy across racial and ethnic groups.

Methods

A retrospective multicenter cohort study was conducted using deidentified electronic health records from a global research network (2005–2025). Adult patients undergoing primary dacryocystorhinostomy were included and categorized as White, Asian, Black or African American, or Other. Patients with congenital lacrimal anomalies or malignancy were excluded. Propensity score matching was performed in a 1:1 ratio between each non-White group and White patients, adjusting for demographic characteristics, socioeconomic factors, comorbidities, prior procedures, laboratory parameters, and visit type. Time-to-event analyses were performed using Kaplan–Meier estimation, log-rank tests, and Cox proportional hazards models. The primary outcome was time to reoperation, defined as repeat dacryocystorhinostomy occurring more than 90 days after the index procedure.

Results

After matching, 1,820 Asian, 744 Black or African American, and 629 patients in the Other group were compared with equal numbers of White patients. Reoperation occurred in 14.0% of Asian patients versus 10.4% of White patients (hazard ratio 1.32, 95% confidence interval 1.09–1.59; P = 0.004). In the Other group, reoperation rates were 11.9% versus 9.9% (hazard ratio 1.58, 95% confidence interval 1.13–2.22; P = 0.007). Among Black or African American patients, reoperation occurred in 11.2% versus 8.6% of White patients, with no statistically significant difference (hazard ratio 1.27, 95% confidence interval 0.92–1.76; P = 0.153).

Conclusions

Differences in long-term reoperation risk following dacryocystorhinostomy were observed across racial and ethnic groups, with higher risk in Asian and Other groups compared with matched White patients. These findings are hypothesis-generating and should be interpreted in light of the absence of data on surgical technique, operative variables and socioeconomic factors. Prospective studies capturing these variables are needed to clarify underlying mechanisms and support equitable clinical practice in lacrimal surgery.

Trial registration

Not applicable.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12886-026-05139-3.

Introduction

Dacryocystorhinostomy (DCR) is a cornerstone procedure for the management of nasolacrimal duct obstruction, offering high overall success rates in restoring lacrimal drainage [1–4]. Despite its effectiveness, long-term outcomes are not uniform and are influenced by a range of factors, including patient demographics, systemic comorbidities, and anatomical variability [5–7]. Notably, the potential role of race and ethnicity in shaping DCR outcomes remains insufficiently characterized, representing an important gap in the current literature [3, 8–11].

In clinical practice, variability in DCR outcomes is often attributed to anatomical and patient-related factors; however, whether such variability differs systematically across racial and ethnic groups remains unclear [12–14]. While racial and ethnic categories are increasingly examined in health outcomes research, they represent complex constructs that may reflect a combination of anatomical variation, healthcare access, and broader systemic influences rather than intrinsic biological differences [15, 16]. Clarifying their potential association with surgical outcomes is therefore essential for a more comprehensive understanding of DCR success and failure [17–19].

Previous studies have suggested that population-level anatomical variation might contribute to surgical challenges. For example, radiological comparisons of the lacrimal sac fossa between Black African and Caucasian individuals have demonstrated differences in maxillary bone thickness and nasal mucosa configuration that could affect the size of the osteotomy and long-term patency after DCR [20, 21]. Similar population-level differences in the morphology of the frontal process of the maxilla and intranasal structures have also been described in Asian cohorts [22]. These findings suggest that anatomical variability, together with structural and systemic influences, may contribute to disparities in long-term DCR outcomes across diverse populations.

In this study, we used the TriNetX Global Research Network to evaluate whether racial and ethnic background is associated with long-term risk of reoperation following DCR. Our objective was to characterize disparities in surgical outcomes after adjusting for demographic and clinical covariates and to support efforts toward equitable, data-informed care in oculoplastic and lacrimal surgery.

Methods

This retrospective cohort study used data from the TriNetX Global Collaborative Network, which aggregates de-identified electronic health records from approximately 150 healthcare organizations worldwide. Contributing organizations span 21 countries across North America, Europe, the Middle East, Africa, and the Asia-Pacific region (Supplementary Table 1). Available data include demographic characteristics, diagnoses, procedures, medications, and laboratory results. The study adhered to the Declaration of Helsinki and STROBE reporting guidelines. All data were fully de-identified in compliance with the Health Insurance Portability and Accountability Act (HIPAA) and the General Data Protection Regulation (GDPR). The Hadassah Medical Center Institutional Review Board reviewed the study protocol and determined that it did not require formal ethics approval. TriNetX is ISO 27001:2013 certified and operates secure federated data architecture. Analyses were performed on a single date in May 2025 using TriNetX LIVE™ analytical tools to ensure consistency in cohort definition.

Cohorts were defined using diagnostic and procedural coding systems, including the International Classification of Diseases, Tenth Revision (ICD-10), Current Procedural Terminology (CPT), ICD-10 Procedure Coding System (ICD-10-PCS), the Systematized Nomenclature of Medicine (SNOMED), and TriNetX visit records. Adult patients (≥ 18 years) with diagnostic records from May 1, 2005, onward were eligible. Four race-based cohorts were created: White, Black or African American, Asian, and Other.

Racial and ethnic classification in TriNetX follows the Health Level Seven (HL7) Centers for Disease Control and Prevention (CDC) Version 1 standard. This information may be self-reported or provider-assigned and may reflect one or more racial identities. The ‘Other’ category encompasses individuals who identified with multiple racial identities simultaneously, those whose recorded race did not correspond to a standard HL7 classification (e.g., individuals identifying as Middle Eastern, North African, or Native Hawaiian/Pacific Islander), and patients for whom the EHR documented a non-standard or free-text racial descriptor. Patients with missing or undocumented racial or ethnic data were excluded from the analysis to ensure accurate cohort classification.

Within each cohort, patients who underwent DCR were identified using the procedure codes listed in the Supplementary Table 2. The index event was defined as the first recorded DCR procedure for each patient during the study period. Patients were excluded if they had congenital malformations of the eyelid, lacrimal apparatus, or orbit (ICD-10 code Q10.x) or malignant neoplasm of the lacrimal gland or duct (ICD-10 code C69.5).

The primary outcome was recurrent DCR, identified using the same procedural codes. To minimize misclassification of early planned procedures or revisions, follow-up began 90 days after the index DCR and patients were followed for up to 10 years. In the primary analysis, each non-White cohort was compared to the White cohort.

Baseline characteristics were assessed from clinical encounters within the two years preceding the index event. Covariate imbalance was defined as a standardized mean difference (SMD) ≥ 0.1. Propensity score matching (PSM) was performed using a 1:1 greedy nearest-neighbor algorithm without replacement, applying a caliper width of 0.1 standard deviations of the logit of the propensity score. This caliper width was chosen to prioritize covariate balance; a wider caliper (e.g., 0.2) retains more patients but risks including poorly matched pairs with residual confounding. The sample size reduction in Cohort 2, while substantial, reflects the degree of baseline differences between racial groups and the stringency required to achieve adequate balance. Covariate balance after matching was assessed using standardized mean differences, with values < 0.1 considered indicative of adequate balance. Matching covariates included demographic variables, socioeconomic indicators, systemic comorbidities, probing of the nasolacrimal duct, laboratory parameters, and visit type. Covariates were selected based on their clinical relevance to DCR outcomes and their known or potential association with both race and reoperation risk. Comorbidities such as chronic sinusitis and skull or facial bone fractures were included because mucosal inflammation and prior anatomical disruption are recognized determinants of long-term DCR patency; their selection was clinically motivated, not driven by automated platform defaults. Time-to-event analyses were conducted using Kaplan-Meier estimation, log-rank testing, and Cox proportional hazards modeling. The proportional hazards assumption was evaluated using Schoenfeld residuals.

Patients with missing race or ethnicity data were excluded from the analysis. For other variables, analyses were conducted using available data as recorded in the database, and propensity score matching was performed based on complete cases.

Results

Before matching, the Asian (N = 1,830), Black or African American (N = 758), and Other race (N = 634) cohorts differed from the White cohort (N = 9,018) across multiple baseline characteristics. In the Asian cohort, female sex was recorded in 74.3% vs. 70.3% in White patients; diabetes in 10.2% vs. 8.8%; chronic sinusitis in 5.7% vs. 10.2%; and chronic kidney disease in 2.6% vs. 3.7%. Probing of the nasolacrimal duct was documented in 8.7% vs. 12.5%, and ambulatory visits in 86.1% vs. 87.8%. After matching, each cohort included 1,820 patients; age at index was 60.1 ± 14.7 vs. 60.1 ± 14.8 years, and female sex was 74.2% vs. 74.8%. All other variables were balanced (Supplementary Table 3).

In the Black or African American cohort, female sex was recorded in 62.9% vs. 70.3%; diabetes in 19.0% vs. 8.8%; chronic sinusitis in 14.7% vs. 10.2%; and chronic kidney disease in 6.6% vs. 3.7%. Probing was recorded in 11.6% vs. 12.5%, and emergency visits in 24.0% vs. 13.2%. After matching, each cohort included 744 patients; age at index was 53.5 ± 15.5 vs. 52.5 ± 16.5 years, and female sex was 63.2% vs. 60.1%. All other variables were balanced (Supplementary Table 4).

In the Other race cohort, female sex was recorded in 65.3% vs. 70.3%; diabetes in 11.5% vs. 8.8%; chronic sinusitis in 13.9% vs. 10.2%; and overweight status in 5.2% vs. 7.9%. Probing was recorded in 17.7% vs. 12.5%, and emergency visits in 13.1% vs. 13.2%. After matching, each cohort included 629 patients; age at index was 56.7 ± 16.5 vs. 55.8 ± 17.0 years, and female sex was 65.8% vs. 65.6%. All other variables were balanced (Supplementary Table 5).

In the comparison between Asian and White patients (n = 1,820 per group), the mean follow-up time was 1,598 ± 1,354 days in the Asian cohort and 1,515 ± 1,326 days in the White cohort. Recurrent DCR occurred in 255 Asian patients (14.0%) and 190 White patients (10.4%). The risk ratio was 1.34 (95% CI, 1.13–1.60), and the risk difference was 3.57% (95% CI, 1.45%-5.70%; p = 0.001). The hazard ratio for recurrence was 1.32 (95% CI, 1.09–1.59), with a log-rank p value of 0.004. Cumulative incidence curves based on Kaplan-Meier estimates are shown in Fig. 1A.

Fig. 1.

Kaplan–Meier curves for time to reoperation after dacryocystorhinostomy by racial and ethnic group. Cumulative incidence of reoperation (expressed as 1 − survival probability) following dacryocystorhinostomy (DCR), stratified by race, over 10 years. Shaded areas represent 95% confidence intervals. (A) Asian vs. White patients (Log-rank p = 0.004). (B) Black or African American vs. White patients (Log-rank p = 0.153). (C) Other race vs. White patients (Log-rank p = 0.007)

In the comparison between Black or African American and White patients (n = 744 per group), the mean follow-up time was 1,581 ± 1,263 days in the Black cohort and 1,536 ± 1,314 days in the White cohort. Recurrent DCR occurred in 83 Black patients (11.2%) and 64 White patients (8.6%). The risk ratio was 1.30 (95% CI, 0.95–1.77), and the risk difference was 2.55% (95% CI, -0.48% to 5.58%; p = 0.0988). The hazard ratio for recurrence was 1.27 (95% CI, 0.92–1.76), with a log-rank p value of 0.153. Cumulative incidence curves based on Kaplan-Meier estimates are shown in Fig. 1B.

Following propensity score matching, reoperation outcomes were compared across all three non-White groups and their respective White comparators. Mean follow-up was 1,598 ± 1,354, 1,581 ± 1,263, and 1,041 ± 1,136 days in the Asian, Black or African American, and Other cohorts, and 1,515 ± 1,326, 1,536 ± 1,314, and 1,511 ± 1,323 days in their respective matched White cohorts, respectively.

A summary of reoperation outcomes across all matched comparisons is presented in Table 1. Cumulative incidence curves based on Kaplan-Meier estimates are shown in Fig. 1C.

Table 1.

Reoperation outcomes following dacryocystorhinostomy across propensity score-matched racial and ethnic groups

Comparison	n (per group)	Reoperations -Non-White n (%)	Reoperations - White n (%)	Risk Ratio (95% CI)	Risk Difference % (95% CI)	HR (95% CI)	Log-rank p
Asian vs. White	1,820	255 (14.0%)	190 (10.4%)	1.34 (1.13–1.60)	3.57 (1.45–5.70)	1.32 (1.09–1.59)	0.004
Black or African American vs. White	744	83 (11.2%)	64 (8.6%)	1.30 (0.95–1.77)	2.55 (− 0.48 to 5.58)	1.27 (0.92–1.76)	0.153
Other vs. White	629	75 (11.9%)	62 (9.9%)	1.21 (0.88–1.66)	2.07 (− 1.37 to 5.51)	1.58 (1.13–2.22)	0.007

HR = hazard ratio; CI = confidence interval. All comparisons are based on 1:1 propensity score-matched cohorts. Bold HR values indicate statistically significant results (p < 0.05)

Discussion

In this large, multicenter cohort, we identified clinically relevant differences in long-term reoperation risk following DCR across racial and ethnic groups. Asian and Other patients demonstrated higher reoperation rates compared with matched White patients, whereas no statistically significant difference was observed among Black or African American patients. These findings extend existing knowledge on DCR outcomes by suggesting that variability in surgical success may not be fully explained by conventional clinical factors alone, and may instead reflect a broader interplay of anatomical, healthcare, and systemic influences.

Several mechanisms may contribute to these observed differences. From a surgical perspective, population-level anatomical variations such as a thicker frontal process of the maxilla, narrower intranasal space, or differences in mucosal configuration may increase the technical complexity of osteotomy creation and marsupialization. Given that inadequate osteotomy size and suboptimal mucosal apposition are well-recognized causes of DCR failure, even subtle anatomical constraints may translate into reduced long-term patency. In addition, variation in wound healing responses, including a tendency toward more pronounced fibrosis in certain populations, may further increase the risk of postoperative restenosis [21–23].

These findings should be interpreted within a broader clinical and systemic context. Racial and ethnic categories do not represent discrete biological entities but may instead capture differences in anatomical characteristics, access to care, healthcare utilization, and social determinants of health [15, 16, 24–26]. As such, the observed variation in outcomes is unlikely to reflect intrinsic biological differences alone, but rather a multifactorial interplay of patient, surgeon and system-level factors [18, 19, 27]. Recognizing potential disparities thus requires consideration of how preoperative assessment, availability of subspecialty care, follow-up practices, and social determinants of health may influence the likelihood of achieving and maintaining DCR patency.

Several limitations should be acknowledged. Racial and ethnic data in TriNetX may be self-reported or provider-assigned and may not fully capture heterogeneous or multiracial identities, raising the possibility of misclassification. Patients without documented racial or ethnic information were excluded, which could introduce selection bias. The “Other” category represents a broad and diverse population, limiting the specificity of conclusions for this group.

Additionally, key surgical variables such as operative approach, stent use, osteotomy size, surgeon experience, and postoperative management were not available in the dataset, and unmeasured confounding from these factors is possible. In contemporary literature, external and endonasal dacryocystorhinostomy have shown broadly comparable success in some series and meta-analyses, although outcomes vary by technique, case selection, and surgical setting [3, 28]. Both approaches also involve osteotomy creation in the region of the maxillary bone and lacrimal sac fossa, but they differ in surgical access, visualization, and instrumentation [29]. Because the TriNetX database does not record which operative approach was used, we could not determine whether differences in the distribution of surgical techniques contributed to the observed racial differences in reoperation rates. Future studies with procedure-level surgical detail are needed to address this limitation more directly. Administrative coding may not always precisely reflect clinical practice. Finally, follow-up duration varied among cohorts, which may influence observed event rates. The stringent caliper matching in the Black or African American cohort resulted in a substantially reduced matched sample (744 pairs from an original 758 patients), which may limit the generalizability of findings for this group; the matched pairs represent those most similar to White patients on measured covariates and may not reflect the broader population of Black patients undergoing DCR.

Despite these limitations, the large sample size, multicenter design, and rigorous propensity score matching strengthen the reliability of the results. Importantly, identifying disparities in reoperation risk should not be interpreted as evidence of innate biological differences between racial groups; rather, these findings highlight the need to investigate modifiable factors that contribute to unequal outcomes.

In conclusion, Asian and Other racial groups demonstrated a higher risk of long-term reoperation following DCR compared with White patients. These findings suggest that surgical outcomes in DCR may be influenced by factors extending beyond standard clinical parameters, including anatomical variability and healthcare-related influences. However, given the limitations of administrative database research, particularly the absence of procedure-level surgical detail, causal interpretation is not warranted. These findings should be regarded as hypothesis-generating, identifying a disparity that warrants further prospective investigation with detailed surgical and socioeconomic data. Such studies would be needed to determine whether the observed differences are attributable to modifiable factors and to guide more equitable clinical practice in lacrimal surgery.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

DCR

Dacryocystorhinostomy

EHR

Electronic health records

PSM

Propensity score matching

SMD

Standardized mean difference

CI

Confidence interval

HR

Hazard ratio

RR

Risk ratio

RD

Risk difference

ICD-10

International Classification of Diseases, Tenth Revision

ICD-10-PCS

International Classification of Diseases, Tenth Revision, Procedure Coding System

CPT

Current Procedural Terminology

SNOMED

Systematized Nomenclature of Medicine

HL7

Health Level Seven

CDC

Centers for Disease Control and Prevention

HIPAA

Health Insurance Portability and Accountability Act

GDPR

General Data Protection Regulation

Author contributions

N.E. and A.G. contributed equally to study conception and design. I.N. and T.S. performed data extraction from the TriNetX platform and conducted the primary data analysis. N.E., A.G., T.S., and I.N. drafted the initial manuscript and prepared the final manuscript. O.Z., O.B and D.W. contributed to data interpretation and critical revision of the manuscript. O.B, O.Z and D.W. contributed to methodology development and validation. Z.G. contributed to study conception, supervised the study, and contributed to study design, data interpretation and validation. All authors reviewed, revised, and approved the final manuscript.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Data availability

The data that support the findings of this study are available from TriNetX, LLC but third-party restrictions apply to the availability of these data. The data were used under license for this study with restrictions that do not allow for the data to be redistributed or made publicly available. However, for accredited researchers, the TriNetX data is available for licensing at TriNetX, LLC. Data access may require a data sharing agreement and may incur data access fees.

Declarations

Ethical approval

The study adhered to the Declaration of Helsinki and STROBE reporting guidelines. All data were fully de-identified in compliance with the Health Insurance Portability and Accountability Act (HIPAA) and the General Data Protection Regulation (GDPR). The Hadassah Medical Center Institutional Review Board reviewed the study protocol and determined that it did not require formal ethics approval.

Consent to participate

Informed consent was waived because this retrospective study used fully deidentified data from the TriNetX Global Research Network, in accordance with applicable regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.