Highlight
-
•
Failure occurred in approximately half of the MIGS cases within 2 years.
-
•
MIGS had higher failure rates than filtration surgeries in traumatic glaucoma cases.
-
•
Glaucoma due to other eye disorders was associated with higher failure rates.
-
•
Trabeculectomy had best success in steroid-induced glaucoma but with more hypotony.
-
•
Insufficient IOP reduction was the leading cause of surgical failure overall.
Purpose
To evaluate the outcomes of minimally invasive glaucoma surgery (MIGS), trabeculectomy, and tube shunt surgery among young patients with secondary glaucoma.
Design
Retrospective clinical cohort study
Subjects
A total of 2483 eyes with secondary glaucoma aged 18 to 40 years who underwent MIGS, trabeculectomy, or tube shunt surgery (2013-2024) in the IRIS Registry (Intelligent Research in Sight).
Methods
The 2-year failure rates were calculated using the Kaplan-Meier estimator, with failure defined as having any of the following: (1) insufficient intraocular pressure (IOP) reduction, defined as either IOP > 21 mm Hg or IOP reduction < 20% from preoperative IOP, (2) hypotony, defined as IOP < 5 mm Hg, (3) any additional glaucoma procedures, or (4) development of no light perception.
Main Outcome Measures
The 2-year failure rates per surgery type per glaucoma subtype.
Results
A total of 454 MIGS, 337 trabeculectomies, and 1692 tube shunt surgeries were analyzed. The overall failure rates at 2 years were 50.6% for trabeculectomy, 52.1% for tube shunt surgery, 53.3% for goniotomy, and 49.0% for canaloplasty. The failure rates among traumatic glaucoma were 39.2% for trabeculectomy, 44.2% for tube shunt surgery, 55.6% for goniotomy, and 50.0% for canaloplasty. Glaucoma due to “other eye disorders,” mostly following prior ocular procedures, had the highest failure rates following trabeculectomy and tube shunt surgery (58.5% and 62.2%, respectively). The failure rate for trabeculectomy was the lowest in steroid-induced glaucoma (37.6%), but hypotony was observed in 10.0% of cases. In uveitic glaucoma, the failure rates were 56.0% for trabeculectomy, 45.6% for tube shunt surgery, 55.1% for goniotomy, and 35.9% for canaloplasty.
Conclusions
MIGS accounted for nearly one in 5 of the glaucoma surgeries recorded in young patients with secondary glaucoma, with approximately half of cases remaining failure-free at 2 years. Surgical outcomes varied depending on the subtype of secondary glaucoma, with glaucoma due to other eye disorders being associated with higher failure rates. Trabeculectomy for steroid-induced glaucoma was successful in nearly two-thirds of the cases, but hypotony was more common in these individuals. The comparison between procedure types should be interpreted with caution as the results may be influenced by baseline disease severity.
Introduction
S econdary glaucoma is a heterogeneous group of glaucomas in which increased intraocular pressure (IOP) results from a variety of identifiable causes such as trauma, uveitis, and corticosteroids. Compared to primary glaucoma, secondary glaucoma is significantly less prevalent, accounting for approximately 10% of all glaucoma cases. , Due to the relatively low prevalence of any one subtype of secondary glaucoma, evidence on the surgical outcomes of each type of surgery in each kind of secondary glaucoma is limited.
Previously reported data demonstrate high surgical failure rates. One retrospective review of 356 eyes with uveitis reported failure in 30% of cases within 36 months for both trabeculectomy and tube shunt procedures, highlighting the challenges of managing uveitic glaucoma. Another retrospective study compared success rates between ab externo trabeculotomy and trabeculectomy in 163 eyes with steroid-induced glaucoma, showing higher and similar 5-year success rates of 73.5% and 74.5%, respectively. However, evidence comparing surgical outcomes across different surgical options remain limited, particularly for minimally invasive glaucoma surgeries (MIGS), which became available relatively recently with a handful of published case series in these patients. ,,,
The clinical characteristics and surgical outcomes of glaucoma vary substantially with age. Management patterns differ across age groups because of differences in medication adherence, levels of frailty, and remaining life expectancy. In addition, the success rates of trabeculectomy are lower in young patients than in elderly patients, likely due to the generally more aggressive nature of the disease and the more robust wound healing responses in younger individuals, which can compromise the success of filtration surgery. ,
This study aimed to investigate the surgical outcomes of MIGS, trabeculectomy, and tube shunt surgery in young patients with traumatic glaucoma, uveitic glaucoma, steroid-induced glaucoma, and glaucoma due to “other eye disorders” using a large ophthalmic registry in the United States. To minimize heterogeneity in disease characteristics and management patterns and to better characterize surgical outcomes in a biologically and clinically distinct population, the present study focused on young patients, excluding those aged over 40 years.
METHODS
Study Population
This is a retrospective analysis of the American Academy of Ophthalmology IRIS Registry (Intelligent Research in Sight). The IRIS Registry is a centralized data repository and reporting tool that can be used for research purposes. This does not constitute human subject research because data in the IRIS Registry is de-identified, and the investigator does not have access to study identifiers. Therefore, institutional board review and informed consent are not required. This study adheres to the Declaration of Helsinki.
We identified patients aged 18 to 40 years who had a diagnosis of secondary glaucoma, based on the International Classification of Disease (ICD)-10 codes H40.3-H40.6, between January 1, 2013, and October 31, 2024. Based on ICD-10 codes, secondary glaucoma was classified into 4 groups: traumatic glaucoma (H40.3), uveitic glaucoma (H40.4), steroid-induced glaucoma (H40.6), and glaucoma due to other eye disorders (H40.5). If multiple glaucoma subtypes were recorded for an eye, priority was assigned in the following order: traumatic glaucoma > uveitic glaucoma > steroid-induced glaucoma > glaucoma due to other eye disorders. The first recorded glaucoma surgery (trabeculectomy, tube shunt surgery, and MIGS) for each eye was identified using Current Procedural Terminology (CPT) codes listed in Supplementary Table 1 and was designated as the index surgery. The date of index surgery was designated as the index date. The Ex-PRESS Glaucoma Filtration Device (Alcon Laboratories, Inc., Fort Worth, TX, USA) was included in the group “trabeculectomy,” because both procedures share the same surgical rationale and have demonstrated comparable clinical outcomes. , MIGS was further categorized into goniotomy (CPT 65820), canaloplasty (CPT 66174, 66175), endoscopic cyclophotocoagulation (CPT 66711), iStent and Hydrus microstent (CPT 0191T, 0253T, 0376T), and Cypass (CPT 0474T). Postoperative follow-up was defined as ending on the last recorded IOP or visual acuity (VA) measurement date, up to October 31, 2024. Eyes with less than 1 year of postoperative follow-up were excluded from the study population ( Figure 1 ).
Flowchart of study population selection ICD, International Classification of Disease; MIGS, minimally invasive glaucoma surgery
Demographic data
Patient demographics included age at the index date, sex, race and ethnicity, urban/rural status, household income, and educational attainment. Sex, race and ethnicity in the IRIS Registry are extracted from electronic health records, which may be documented based on self-report or by observation of the practice. We used the following race and ethnicity categories: Asian, Hispanic, non-Hispanic Black, non-Hispanic White, and Other. Urban/rural status was assigned based on each patient’s most recent valid 5-digit ZIP code, using the 2020 U.S. Census Urban Areas Relationship Files (available at https://www.census.gov/geographies/reference-files/time-series/geo/relationship-files.2020.html#uacomp ). ZIP codes located in Urbanized Areas (population ≥ 50,000) or Urban Clusters (2500-49,999) were classified as urban; those not listed were classified as rural. ZIP codes for patients without a valid ZIP code were labeled as “Missing.” Household income and educational attainment were defined using data from the American Community Survey’s 5-year estimates, linked to each patient’s most recent 5-digit ZIP code. Household income was categorized into quintiles based on ZIP code-level estimates. Educational attainment, defined as a high school graduate or higher, was measured by the percentage of individuals aged 25 years and older with a high school diploma or higher in each ZIP code.
Ophthalmic variables
All IOP and VA measurements were extracted for each eye. The VA values were converted to logMAR (logMAR) within the IRIS Registry. We removed the top 0.1% IOP measurements to reduce the influence of outliers. We defined preoperative IOP as the mean of all measurements within 60 days before the index surgery. Preoperative VA was defined as the last measurement within 182 days before the index surgery. We also estimated postoperative IOP and VA at 1 month, 3 months, 6 months, and 1 year by taking the average of measurements between 30 days to 90 days, 91 days to 180 days, 181 days to 365 days, and 366 days to 730 days after the index surgery.
Glaucoma severity was identified using ICD-9 codes (365.70: unspecified, 365.71: mild, 365.72: moderate, 365.73: advanced, 365.74: indeterminate) and the fourth decimal place of ICD-10 codes (0: unspecified, 1: mild, 2: moderate, 3: advanced, 4: indeterminate). Eyes with only unspecified or indeterminate codes, or without any severity codes, were categorized as “Missing.” If multiple severity codes among mild, moderate, and advanced were recorded for an eye, the most severe code was assigned as the eye’s severity. The surgeon type linked to each index surgery was categorized as either a glaucoma subspecialist or a non-glaucoma subspecialist. We considered the index surgery to be combined with cataract extraction if cataract extraction defined by CPT codes (Supplementary Table 1) was recorded within 31 days of the index surgery.
To investigate potential secondary causes of glaucoma, we identified preceding non-glaucoma procedures, including cornea-, retina-, trauma-, and lens-related procedures, among eyes with glaucoma due to other eye disorders using the CPT codes listed in Supplementary Table 2.
Outcome measures
The primary outcomes were the 2-year failure rates of each surgery type by secondary glaucoma subtype. Secondary outcomes included risk factors for failure, incidence of postoperative complications, and postoperative changes in IOP and VA.
Failure criteria
We calculated the 2-year failure rates for each surgery type using the Kaplan-Meier estimator, with failure defined as having any of the following: (1) insufficient IOP reduction, defined as either IOP > 21 mm Hg or IOP reduction < 20% from preoperative IOP, (2) hypotony, defined as IOP < 5 mm Hg, (3) any additional glaucoma procedures, or (4) development of no light perception (LogMAR 4.0). IOPs in criteria (1) and (2) were evaluated over 2 consecutive follow-up visits, at least 90 days apart, and recorded more than 90 days after the index date. Additional glaucoma procedures included trabeculectomy, tube shunt surgery, MIGS, cyclophotocoagulation, and needling or revision defined by CPT codes (Supplementary Table 1). Criterion (3) was applied to records more than 90 days after the index date. We defined these criteria based on the landmark Tube vs Trabeculectomy (TVT) Study and the Primary Tube vs Trabeculectomy (PTVT) Study. , We performed an additional analysis in which the failure criterion of insufficient IOP reduction was defined as IOP > 21 mm Hg, excluding IOP reduction < 20%, to make the results comparable to the Systemic Immunosuppressive Therapy for Eye Diseases (SITE) Cohort Study, which evaluated failure rates of filtration surgeries among eyes with uveitis. Failure rates were compared using log-rank tests between trabeculectomy and tube shunt surgery, as well as between goniotomy and canaloplasty. For eyes meeting multiple failure criteria, the earliest occurring criterion was designated as the reason for failure. If multiple criteria occurred on the same day, all were recorded as the reasons for failure for that eye. To investigate the causes of failure, we calculated the proportion of each reason relative to the total number of surgeries. Eyes that did not meet the failure criteria were censored upon reaching the end of the follow-up period.
Additionally, to understand racial differences in surgical failure patterns, we compared failure rates between eyes of non-Hispanic Black and non-Hispanic White patients for each surgery type and glaucoma subtype using log-rank tests.
Risk factors for failure
Among patients with available preoperative IOP and VA, we imputed missing data on patient sex, race and ethnicity, urban/rural status, household income, educational attainment, and surgeon’s subspecialty, using multiple imputation by the chained equations method with 5 iterations. Multivariable Cox regression models were applied on this imputed cohort to identify predictors of surgical failure. The models incorporated the following variables: surgery type, glaucoma subtype, surgeon’s subspecialty, patient age, sex, race and ethnicity, urban/rural status, household income, educational attainment, preoperative IOP, and preoperative VA. Analyses were conducted using both failure criteria from the TVT/PTVT Study and the SITE Cohort Study. In addition, separate multivariable Cox regression models were applied to each individual component of the TVT/PTVT failure definition.
Postoperative complications
The percentage of eyes with postoperative IOP spikes was calculated for each surgery type and glaucoma subtype, with an IOP spike defined as IOP > 30 mm Hg and ≥ 10 mm Hg more than preoperative IOP within the first postoperative week, based on criteria from previous studies. , The percentage of eyes with postoperative hand motion or worse vision was calculated per surgery type per secondary glaucoma classification, among eyes with preoperative VA better than hand motion. Postoperative hand motion or worse was defined as having a VA worse than 2.3 logMAR , at 2 consecutive visits occurring more than 90 days after the index date. Postoperative complications defined by ICD codes (Supplementary Table 3) were identified within 90 days after surgery, and the percentage of eyes with postoperative complications was calculated by surgery type and glaucoma subtype. Finally, the percentage of eyes that underwent additional IOP-lowering procedures defined by trabeculectomy, tube shunt surgery, MIGS, and cyclophotocoagulation was calculated for each type of surgery and secondary glaucoma subtype.
Sensitivity Analysis
We conducted 2 sensitivity analyses to assess the robustness of our findings. First, to address potential misclassification of uveitic glaucoma as steroid-induced glaucoma, we reclassified eyes initially labeled as steroid-induced glaucoma as uveitic glaucoma if they had preceding ICD codes for uveitis (Supplementary Table 4) and lacked records of prior ocular procedures (identified by CPT codes beginning with 65, 66, or 67). Using this alternative glaucoma classification, we calculated the 2-year failure rates for each procedure type.
Second, we employed failure criteria that required the IOP conditions in criteria 1 and 2 to be observed at 3 consecutive follow-up visits, at least 90 days apart.
Statistical Analysis
P values <.05 were considered statistically significant. All statistical analyses were performed in R version 4.5.0 R (R Foundation for Statistical Computing, Vienna, Austria).
RESULTS
Patient characteristics
A total of 2483 eyes that underwent glaucoma surgeries were included. Of these, 358 (14.4%) eyes had traumatic glaucoma, 949 eyes had uveitic glaucoma (38.2%), 243 eyes had steroid-induced glaucoma (9.8%), and 933 eyes had glaucoma due to “other eye disorders” (37.6%, Table 1 ). Overall, the mean age at the index date was 31.0 ± 6.0 years. Males accounted for 73.2% of the patients with traumatic glaucoma, compared with 52.3% in the overall study population. The percentage of Non-Hispanic Black patients was the highest in uveitic glaucoma (21.3%) and the lowest in steroid-induced glaucoma (8.2%). Non-Hispanic White patients accounted for 58.0% of the patients with steroid-induced glaucoma, compared with 40.6% in the overall population. Among eyes with available ICD-based severity, more than half were categorized as advanced stage, except for steroid-induced glaucoma, in which mild stage was the most common.
Table 1
Characteristics of Included Eyes.
| Glaucoma subtype | Overall | Trauma | Uveitis | Steroid-induced | Other eye disorders | P value |
|---|---|---|---|---|---|---|
| N (%) | 2483 (100) | 358 (14.4) | 949 (38.2) | 243 (9.8) | 933 (37.6) | |
| Procedure type (%) | <.001 | |||||
| MIGS | 454 (18.3) | 56 (15.6) | 167 (17.6) | 89 (36.6) | 142 (15.2) | |
| Trabeculectomy | 337 (13.6) | 91 (25.4) | 116 (12.2) | 30 (12.3) | 100 (10.7) | |
| Tube shunt surgery | 1692 (68.1) | 211 (58.9) | 666 (70.2) | 124 (51.0) | 691 (74.1) | |
| Surgeon’s subspecialty (%) | .02 | |||||
| Glaucoma subspecialist | 1897 (76.4) | 267 (74.6) | 741 (78.1) | 177 (72.8) | 712 (76.3) | |
| Non-glaucoma subspecialist | 484 (19.5) | 78 (21.8) | 158 (16.6) | 54 (22.2) | 194 (20.8) | |
| Missing | 102 (4.1) | 13 (3.6) | 50 (5.3) | 12 (4.9) | 27 (2.9) | |
| Severity (%) | <.001 | |||||
| Mild | 224 (9.0) | 23 (6.4) | 117 (12.3) | 38 (15.6) | 46 (4.9) | |
| Moderate | 342 (13.8) | 55 (15.4) | 151 (15.9) | 27 (11.1) | 109 (11.7) | |
| Advanced | 710 (28.6) | 131 (36.6) | 285 (30.0) | 33 (13.6) | 261 (28.0) | |
| Missing | 1207 (48.6) | 149 (41.6) | 396 (41.7) | 145 (59.7) | 517 (55.4) | |
| Age (years, mean (SD)) | 31.0 (6.0) | 31.0 (5.9) | 30.5 (5.9) | 30.9 (6.1) | 31.5 (6.1) | .005 |
| Sex (%) | <.001 | |||||
| Female | 1124 (45.3) | 88 (24.6) | 480 (50.6) | 129 (53.1) | 427 (45.8) | |
| Male | 1299 (52.3) | 262 (73.2) | 450 (47.4) | 106 (43.6) | 481 (51.6) | |
| Missing | 60 (2.4) | 8 (2.2) | 19 (2.0) | 8 (3.3) | 25 (2.7) | |
| Race and ethnicity (%) | <.001 | |||||
| Asian | 72 (2.9) | 17 (4.7) | 28 (3.0) | 9 (3.7) | 18 (1.9) | |
| Hispanic | 414 (16.7) | 63 (17.6) | 150 (15.8) | 33 (13.6) | 168 (18.0) | |
| Non-Hispanic Black | 447 (18.0) | 54 (15.1) | 202 (21.3) | 20 (8.2) | 171 (18.3) | |
| Non-Hispanic White | 1009 (40.6) | 137 (38.3) | 357 (37.6) | 141 (58.0) | 374 (40.1) | |
| Other | 379 (15.3) | 55 (15.4) | 146 (15.4) | 27 (11.1) | 151 (16.2) | |
| Missing | 162 (6.5) | 32 (8.9) | 66 (7.0) | 13 (5.3) | 51 (5.5) | |
| Urban/rural status (%) | .36 | |||||
| Rural | 257 (10.4) | 27 (7.5) | 94 (9.9) | 25 (10.3) | 111 (11.9) | |
| Urban | 2220 (89.4) | 330 (92.2) | 853 (89.9) | 218 (89.7) | 819 (87.8) | |
| Missing | 6 (0.2) | 1 (0.3) | 2 (0.2) | 0 (0.0) | 3 (0.3) | |
| Household income (%) | <.001 | |||||
| < $34,999 | 85 (3.4) | 8 (2.2) | 35 (3.7) | 5 (2.1) | 37 (4.0) | |
| $35,000- $74,999 | 1178 (47.4) | 160 (44.7) | 433 (45.6) | 101 (41.6) | 484 (51.9) | |
| $75,000- $149,999 | 670 (27.0) | 111 (31.0) | 262 (27.6) | 99 (40.7) | 198 (21.2) | |
| > $150,000 | 27 (1.1) | 1 (0.3) | 15 (1.6) | 0 (0.0) | 11 (1.2) | |
| Missing | 523 (21.1) | 78 (21.8) | 204 (21.5) | 38 (15.6) | 203 (21.8) | |
| Educational attainment (High school graduate or higher, %) | .005 | |||||
| ≤ 60 | 22 (0.9) | 0 (0.0) | 8 (0.8) | 2 (0.8) | 12 (1.3) | |
| 61-70 | 57 (2.3) | 10 (2.8) | 20 (2.1) | 6 (2.5) | 21 (2.3) | |
| 71-80 | 176 (7.1) | 22 (6.1) | 72 (7.6) | 12 (4.9) | 70 (7.5) | |
| 81-90 | 698 (28.1) | 92 (25.7) | 272 (28.7) | 52 (21.4) | 282 (30.2) | |
| 91-100 | 1015 (40.9) | 157 (43.9) | 376 (39.6) | 133 (54.7) | 349 (37.4) | |
| Missing | 515 (20.7) | 77 (21.5) | 201 (21.2) | 38 (15.6) | 199 (21.3) | |
| Follow up period (days, median [IQR]) | 1415 [841, 2220] | 1344 [779, 2244] | 1479 [861, 2228] | 1445 [956, 2346] | 1380 [796, 2105] | .11 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
