Abstract
Purpose:
To characterize the impact of neovascular glaucoma (NVG) and its underlying etiology on mortality and determine the association between NVG and cause of death.
Design:
Retrospective cohort study.
Participants:
A total of 210 subjects with NVG were included in the descriptive study. Of the 210 subjects, 132 with a definite date of NVG onset were included in the survival analysis.
Methods:
Patients with NVG seen at our tertiary care center from 2015–2016 were identified. Demographic information, underlying etiology, and medical conditions were documented. The Centers for Disease Control and Prevention National Death Index (CDC NDI) was used to designate if subjects were alive on December 31, 2021, and to determine the primary cause of death. Kaplan-Meier and Cox proportional hazard regression analyses were performed on subjects with a definitive date of NVG onset.
Results:
Of the 210 subjects, 32% were deceased by December 31, 2021. The commonest diagnosis was proliferative diabetic retinopathy (PDR) (57%), followed by central retinal vein occlusion (CRVO) (29%). Of 132 subjects included in the Kaplan-Meier survival analysis, the mean survival time was 154 ± 15 months (95% CI: 125–183). Cox regression analysis demonstrated that older age (HR 1.03, 95% CI: 1.004–1.06, P = 0.02) and central retinal artery occlusion (CRAO) (HR: 4.63, 95% CI: 1.7–12.44, P = 0.002) were associated with poorer survival.
Conclusions:
Increasing age and CRAO-related NVG (compared to CRVO-unrelated) were important predictors of increased mortality. Our results highlight the need for aggressive management of underlying systemic co-morbid conditions contributing to early mortality in patients with NVG.
Keywords: Mortality, neovascular glaucoma, survival
Neovascular glaucoma (NVG) is a secondary glaucoma caused by abnormal anterior segment blood vessel growth, most commonly due to retinal ischemia/hypoxia, but can also be caused by inflammatory conditions and tumors.[1,2,3] The resultant fibrovascular membrane obstructs the outflow of the aqueous humor, eventually closing the iridocorneal angle and raising intraocular pressure. Panretinal photocoagulation (PRP), intravitreal anti-VEGF injections, and intraocular pressure (IOP) lowering agents and glaucoma surgery are the mainstays of treatment for NVG.[2,4]
Neovascular glaucoma is associated with systemic and ocular disease that causes retinal ischemia. The commonest causes of NVG include proliferative diabetic retinopathy (PDR) and central retinal vein occlusion (CRVO), which account for 66% of all cases,[4,5,6] followed by ocular ischemic syndrome (OIS) (13%).[4] There is a varying prevalence of NVG due to central retinal artery occlusion (CRAO); however, one study found that five out of 33 patients (15.2%) diagnosed with CRAO developed neovascular glaucoma.[7]
These disease processes are associated with increased mortality. A study at an integrated health care center in Southern California found that subjects with diabetic retinopathy were 1.61 times more likely to die in a 5-year follow-up compared to subjects without retinopathy.[8] Compared to patients with no retinopathy, those with diabetic retinopathy had a higher risk for cardiovascular events (heart attack, stroke, heart failure), with risk increasing with the severity of retinopathy.[9] A diagnosis of retinopathy in diabetic subjects increases the risk of each outcome in the disease course compared to diabetics without retinopathy.[9] CRVO is also associated with increased mortality in some studies; out of 439 patients at a secondary referral center in Denmark, 30% of subjects with CRVO died within 5 years compared to 24% in the control population.[10]
While existing literature is focused on the etiology, pathogenesis, and treatment methods for NVG, there is no data about the long-term survival and causes of mortality in patients with NVG. NVG is often an end-organ damage marker of severe cardiovascular disease and may portend a poor prognosis for life. This is important both for clinical decision-making and patient counseling.
In this study, we aim to characterize the differences in mortality among several etiologies of NVG and identify if an association exists between NVG diagnosis and cause of death. We hypothesize that NVG is associated with increased overall mortality compared to the general population. Due to the significant cardiovascular and cerebrovascular associations with PDR and CRAO,[9,11] we also hypothesize that NVG associated with PDR and CRAO will have a higher mortality risk than NVG associated with CRVO or other eye-specific conditions.
Methods
We performed a retrospective chart review at Emory Healthcare in Atlanta, Georgia. The study was approved by the Institutional Review Board (IRB) of Emory University. Informed consent was not required for this study per our IRB regulations because of the retrospective nature of the study and minimal risk to subjects. Our study had two components: a descriptive retrospective study was performed on 210 subjects with NVG. Survival analysis was performed on a subset of these patients (n = 132) with a definite date of NVG onset.
Subjects
We included subjects seen at the Emory Eye Center (EEC) in 2015 and 2016 with a diagnosis of neovascular glaucoma (International Classification of Diseases [ICD] codes H40.5, H40.89, and H40.9). Patients presenting in 2015–2016 were included so as to get at least a 5-year survival data on all subjects.
Onset of NVG diagnosis was defined by new onset of pain, documented presence of neovascularization of the anterior chamber or iris, and IOP > 21 mm Hg, or by documentation in the electronic medical record (EMR) of NVG onset established at an outside facility.
From the list of subjects generated using the above ICD codes, subjects were excluded from the descriptive analysis if: etiology of NVG could not be determined, vital status data was missing from the Centers for Disease Control and Prevention (CDC) National Death Index (NDI)[12] search, or NVG was due to extrinsic diseases (such as ocular oncologic pathologies). Subjects were excluded from the Kaplan-Meier analysis if an exact date of NVG onset could not be determined.
Data collected from the EMR is specified as follows: age at diagnosis, gender, race (White, Black/African American, Other [Asian, American Indian/Alaskan Native, Not reported]), affected eye, NVG etiology (PDR, CRVO, CRAO, OIS, other), visual acuity and IOP at NVG diagnosis, presence of diabetes mellitus (DM) or hypertension (HTN), vital status on December 31, 2021, and date and cause of death for deceased subjects. NVG severity was determined based on presenting IOP (mild: less than 20 mm Hg, moderate: 20–40 mm Hg, severe: >40 mm Hg).
Diagnoses in the “other” category included Axenfeld-Rieger syndrome, Coats disease, inflammatory conditions (uveitis, radiation retinopathy, acute retinal necrosis), sickle cell retinopathy, and retinal detachment. Cases of NVG due to ocular oncologic pathologies were not included due to the pathogenesis of NVG development (compressive disease versus intrinsic disease).
Vital status (alive versus deceased) on December 31, 2021, was then determined. Subjects were identified as alive if they were seen at Emory Healthcare from June 1, 2021–December 31, 2021. To confirm vital status, the CDC NDI Plus service was utilized for subjects whose vital status was unknown.[10] The NDI is a comprehensive database that contains all death records from 1979 to latest (2020 for data collection in 2021) for all 50 states. The NDI also provides an Early Release Program, which allows for access to preliminary death record information for the newest year (2021 in this study) before finalization. At a minimum, the subject first and last name and date of birth were included in the submitted record to qualify as a sufficient search. Outputs from the database for deceased subjects included date of death, state in which death occurred, corresponding death certificate number, and cause(s) of death. The NDI Retrieval Report was used to assess for true matches (matches assigned a Status Code of “1”) and date of death. The Cause of Death Report was used to obtain ICD codes for the cause of death. Cause of death was classified as cardiac, cerebrovascular, respiratory, or other. The ICD code was used to codify each cause of death. Cardiac causes included myocardial infarction, atherosclerotic or ischemic heart disease, heart failure, and shock. Cerebrovascular causes included ischemic or hemorrhagic stroke or anoxic brain injury. Respiratory causes included respiratory failure, lung malignancy, and pneumonia. Other causes were those that did not align with the previous categories.
Statistical analysis
Statistical analysis was performed using Statistical Package for the Social Sciences (SPSS) software (IBM SPSS software platform, USA). Descriptive statistics were used to summarize data. A Chi-square test of proportions was used to determine the significance of differences between groups. For the survival analysis, the interval between the date of NVG diagnosis to either the study endpoint of December 31, 2021, or the date of death was used to determine survival time in months. A Kaplan-Meier analysis was performed on subject records that had a definitive date of diagnosis of NVG. The analysis was also stratified based on ocular diagnosis to compare survival curves between all diagnoses, PDR, and CRVO. Cox proportional hazard regression analysis was performed to assess risk factors predicting mortality in NVG. P values < 0.05 were considered statistically significant.
Results
The initial EMR search generated 163 subjects with the specified ICD codes in 2015 and 431 subjects in 2016. After each subject’s EMR was reviewed for an accurate diagnosis of NVG and duplicates were removed, we included 220 NVG subjects (98 subjects in 2015 and 122 subjects in 2016). An additional 10 subjects were excluded for the following reasons: six with melanoma (five choroidal melanoma/mass, one metastatic melanoma), two subjects with no underlying ocular diagnosis establishing cause of NVG, and two subjects with an unknown vital status after CDC output analysis.
Out of all subjects, 60 (27%) were labeled alive on December 31, 2021 (based on EMR data that showed that they came to appointments in 2022), and data for 160 (73%) subjects was sent to the CDC to determine vital status. Based on the CDC NDI output, 68 (31%) subjects were labeled as deceased.
Descriptive analysis
Demographic information, ocular diagnoses, and outcomes for the 210 subjects who qualified for the study are summarized in Table 1 by ocular diagnosis and in Table S1 by race.
Table 1.
Demographic Characteristics of NVG Subjects in 2015-2016 by Ocular Diagnosis and Outcomes
| Ocular Diagnosis | Vital Status on 12/31/21 | Average Age on 1/1/2016 (years) (mean±SD) | Gender | Race | Average age at death (years) (mean±SD) | Cause of Death | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CRVO (n=60) | Alive (n=42) | 68±12 | Male 23 (55%) Female 19 (45%) |
Black 16 (38%) White 9 (21%) Other 17 (40%) |
86±8 | Cardiac 3 (17%) Cerebrovascular 1 (6%) Respiratory 6 (33%) Other 8 (44%) |
||||||
| Deceased (n=18) | 84±9 | Male 9 (50%) Female 9 (50%) |
Black 5 (28%) White 10 (56%) Other 3 (17%) |
|||||||||
| PDR (n=119) | Alive (n=81) | 58±12 | Male 43 (53%) Female 38 (47%) |
Black 47 (58%) White 17 (21%) Other 17 (21%) |
65±13 | Cardiac 15 (39%) Cerebrovascular 6 (16%) Respiratory 4 (11%) Other 13 (34%) |
||||||
| Deceased (n=38) | 63±14 | Male 22 (58%) Female 16 (42%) |
Black 20 (53%) White 14 (37%) Other 4 (11%) |
|||||||||
| CRAO (n=9) | Alive (n=2) | 81±3 | Female 2 (100%) | Black 1 (50%) Other 1 (50%) |
71±11 | Cardiac 2 (29%) Respiratory 2 (29%) Other 3 (42%) |
||||||
| Deceased (n=7) | 69±9 | Male 3 (43%) Female 4 (57%) |
Black 5 (71%) Other 2 (29%) |
|||||||||
| OIS (n=9) | Alive (n=8) | 77±8 | Male 5 (63%) Female 3 (38%) |
Black 3 (38%) White 2 (25%) Other 3 (38%) |
70 | Other 1 (100%) | ||||||
| Deceased (n=1) | 70 | Male 1 (100%) | White 1 (100%) | |||||||||
| Other (n=13) | Alive (n=9) | 47±22 | Male 5 (56%) Female 4 (44%) |
Black 4 (44%) White 4 (44%) Other 1 (11%) |
64±31 | Cardiac 1 (25%) Other 3 (75%) |
||||||
| Deceased (n=4) | 62±27 | Male 2 (50%) Female 2 (50%) |
Black 1 (25%) White 2 (50%) Other 1 (25%) |
|||||||||
| Total (n=210) | Alive (n=142) | 57±3 | Male 76 (54%) Female 66 (46%) |
Black 71 (50%) White 32 (23%) Other 39 (27%) |
71±16 | Cardiac 21 (31%) Cerebrovascular 7 (10%) Respiratory 12 (18%) Other 28 (41%) |
||||||
| Deceased (n=68) | 64±15 | Male 37 (54%) Female 31 (46%) |
Black 31 (46%) White 27 (40%) Other 10 (15%) |
Table S1.
Outcomes and Ocular Diagnosis by race
| Race | Average Age All Subjects on 1/1/2016 (years) (mean±SD) | Vital Status | Average age on 1/1/2016 (years) (mean±SD) | Ocular Diagnosis | Cause of Death | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Black (n=102) | 63±14 | Alive (n=71) | 60±14 | PDR n=47 (66%) CRVO n=16 (23%) Other n=4 (6%) OIS n=3 (4%) CRAO n=1 (1%) |
Cardiac n=12 (39%) Cerebrovascular n=5 (16%) Respiratory n=3 10%) Other n=11 (35%) |
|||||
| Deceased (n=31) | 69±14 | PDR n=20 (65%) CRAO n=5 (16%) CRVO n=5 (16%) Other n=1 (3%) |
||||||||
| White (n=59) | 64±16 | Alive (n=32) | 62±13 | PDR n=17 (53%) CRVO n=9 (28%) Other n=4 (13%) OIS n=2 (6%) |
Respiratory n=6 (22%) Cardiac n=6 (22%) Cerebrovascular n=1 (4%) Other n=14 (52%) |
|||||
| Deceased (n=27) | 68±20 | PDR n=14 (52%) CRVO n=10 (37%) Other n=2 (7%) OIS n=1 (4%) |
||||||||
| Other (n=49) | 66±16 | Alive (n=39) | 64±17 | CRVO n=17 (44%) PDR n=17 (44%) OIS n=3 (8%) CRAO n=1 (3%) Other n=1 (3%) |
Cardiac n=3 (30%) Respiratory n=3 (30%) Cerebrovascular n=1 (10%) Other n=3 (30%) |
|||||
| Deceased (n=10) | 76±8 | PDR n=4 (40%) CRVO n=3 (30%) CRAO n=2 (20%) Other n=1 (10%) |
CRVO: central retinal vein occlusion, PDR: proliferative diabetic retinopathy, CRAO: central retinal artery occlusion, OIS: ocular ischemic syndrome, NVG: neovascular glaucoma
To summarize the tables, the commonest underlying ocular diagnosis in our cohort was PDR (n = 119, 57%), followed by CRVO (n = 60, 29%). Our cohort had 102 (49%) Black subjects and 59 (28%) White subjects.
Subjects with PDR, whether alive (n = 81, age 58 ± 12) or deceased (n = 38, age 63 ± 14), were significantly younger (t-test, P < 0.001) compared to CRVO subjects in both alive (n = 42, age 68 ± 12) or deceased (n = 18, age 84 ± 9) groups. Subjects who were deceased were significantly more likely to have more severe NVG (Chi-square, P = 0.021).
There was no significant difference in age between races (ANOVA, P = 0.419) [Table S1]. Although a greater proportion of Black subjects had PDR (n = 47, 66%) compared to White (n = 17, 53%) subjects, this difference was not significant (Chi-square, P = 0.100).
Cause of mortality in NVG
The cause of mortality, as determined by the CDC data, is summarized by ocular diagnosis in Table 1 and by race in Table S1. There was no significant difference in causes of death by ocular diagnosis and by race.
However, there were several trends observed. Mortality in CRVO subjects was from respiratory causes (n = 6, 33%) rather than cardiac (n = 3, 17%) or cerebrovascular causes (n = 1, 6%), whereas in PDR, mortality was predominantly from cardiac causes (n = 15, 39%), followed by cerebrovascular (n = 6, 16%) and respiratory (n = 4, 11%).
Survival analysis
Out of the 210 subjects, 132 subjects with a definitive date of NVG onset were included in the Kaplan-Meier survival analysis.
Mean survival times for each etiology of NVG are listed in Table 2. A Log rank (Mantel-Cox) analysis found a significant difference between mean survival times (P = 0.016). Subjects with PDR had the longest survival, whereas subjects with CRAO had the shortest survival. Fig. 1 depicts the survival curves and number at risk for CRVO, PDR, and CRAO. Censored subjects in these groups represent varying follow-up times among subjects.
Table 2.
Kaplan-Meier Survival Analysis
| Diagnosis | Average age at diagnosis (years) (mean±SD) | Mean survival time (months) (95% CI) | Median survival time (months) (95% CI) | 5-year all cause mortality | 10-year all cause mortality | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| CRVO (n=37) | 71±13 | 125±12 (102 – 148) | 152±34 (85 – 219) | 24% | 38% | |||||
| PDR (n=73) | 57±12 | 165±19 (130 – 202) | 188±29 (131 – 245) | 28% | 34% | |||||
| CRAO (n=9) | 70±8 | 57±16 (26 – 88) | 46±19 (8 – 84) | 67% | 83% | |||||
| OIS (n=6) | 74±10 | 117±16 (86 – 148) | - | |||||||
| Other (n=7) | 46±28 | 135±18 (99 – 171) | - | |||||||
| Overall (n=132) | 62±15 | 154±15 (125 – 183) | 156±16 (124 – 188) | 29% | 42% |
CRVO=central retinal vein occlusion; PDR=proliferative diabetic retinopathy; CRAO=central retinal artery occlusion; OIS=ocular ischemic syndrome. *Median survival time not available for ocular diagnosis with >50% survival by study endpoint
Figure 1.
Kaplan-Meier survival analysis (in months) for CRVO (blue), CRAO (pink), and PDR (green). Time is measured from the date of NVG diagnosis to either the date of death or the study end date (December 31, 2021). The number of subjects at risk, reported in 50-month intervals starting at 0 months and ending at 250 months, for each group, is shown in the table immediately below the survival curves
Predictors of mortality in NVG
To determine the predictors of mortality in NVG, a Cox proportional hazard regression analysis was performed with age at diagnosis, race, ocular diagnosis, NVG severity, and visual acuity as the covariates. A univariate analysis demonstrated age at diagnosis (P = 0.008), ocular diagnosis (P = 0.033), and NVG severity (P = 0.036) to be significant. Compared to CRVO, CRAO was the only ocular diagnosis associated with reduced survival (HR: 3.362, CI: 1.345–8.402, P = 0.009). Compared to mild NVG severity, only severe NVG was associated with reduced survival (HR: 0.355, CI: 0.131–0.961, P = 0.042). A multivariate analysis with these covariates was only significant for age at diagnosis (P = 0.062). The outcome of the regression analysis is summarized in Table 3. Race, gender, and visual acuity were not found to be significant predictors of mortality in NVG.
Table 3.
Cox Proportional Hazards Regression Analysis
| Covariate | Hazard Ratio | 95% CI | P | |||
|---|---|---|---|---|---|---|
| Age at Diagnosis | 1.030 | 1.008 – 1.053 | 0.008 | |||
| Gender | 1.092 | 0.615 – 1.939 | 0.763 | |||
| Race | 0.181 | |||||
| Black White Other |
1.000 1.765 1.108 |
Reference 0.939 – 3.315 0.496 – 2.475 |
0.078 0.803 |
|||
| Ocular Diagnosis | 0.033 | |||||
| CRVO PDR CRAO OIS Other |
1.000 0.892 3.362 0.481 0.814 |
Reference 0.458 – 1.736 1.345 – 8.402 0.063 – 3.665 0.185 – 3.588 |
0.736 0.009 0.480 0.786 |
|||
| NVG Severity | ||||||
| Mild (<20) Moderate (20–40) Severe (>40) |
1.000 0.700 0.355 |
Reference 0.266 – 1.841 0.131 – 0.961 |
0.036 0.470 0.042 |
|||
| Visual Acuity | ||||||
| 20/200 or better 20/200 – HM HM – LP NLP |
1.000 1.435 1.621 1.198 |
Reference 0.610 – 3.377 0.717 – 3.664 0.245 – 5.847 |
0.704 0.408 0.245 0.823 |
CRVO: central retinal vein occlusion, PDR: proliferative diabetic retinopathy, CRAO: central retinal artery occlusion, OIS: ocular ischemic syndrome, NVG: neovascular glaucoma
Power calculations: This was a retrospective study. Power calculations were done post hoc.[13] Using the median survival time for PDR and CRVO and an alpha ratio of 0.05, our study was powered at 0.62. Using the hazard ratios for race [Table 3] with an alpha value of 0.05, our study was powered to 0.9.
Discussion
We found a 5-year all-cause mortality of 29% in patients diagnosed with NVG. The average age at death was 71 years, whereas the average life expectancy is 76.4 years. Although we did not have a control population to directly compare outcomes, this difference is what we would expect given the systemic diseases associated with NVG. The demographic data for our study cohort reflects the population distribution of the city of Atlanta and its surrounding regions. We found that a diagnosis of NVG was more prevalent in men and Black subjects. Surprisingly, race was not found to be a significant predictor of mortality. According to mortality statistics in the United States, Black males have a higher death rate compared to the general population (1,380.2 per 100,000 vs 879.7 per 100,000).[14]
PDR was the commonest cause of NVG (54%), followed by CRVO (29%) in our cohort. These diagnoses compromised a large majority of the causes of NVG most likely because of the population we sampled. Subjects with PDR were diagnosed with NVG at a much younger age (57 ± 12 years) compared to other subgroups such as CRVO (71 ± 13 years) and CRAO (70 ± 10 years). Hence, despite a longer median survival time [Table 3] of nearly 14 years compared to 10 years in CRVO, subjects with PDR, on an average, died at an earlier age than CRVO or other NVG diagnoses. The average age of PDR subjects in our population was 59 ± 15 years which was similar to a recent study.[15] Cardiovascular or cerebrovascular causes were the commonest causes of mortality in subjects with PDR, whereas respiratory causes were more common in subjects with CRVO. In the general population, heart disease accounts for 20.1% of deaths, whereas cerebrovascular diseases account for 4.7%. We believe that respiratory causes being more common in CRVO may suggest that it is not as involved in widespread systemic disease such as is seen in diabetes. Our study also found older age and ocular diagnosis (specifically CRAO) to be predictors of mortality.
Poor survival outcomes have previously been described in patients with NVG.[16,17,18] In a recent study of veterans in Minneapolis, Minnesota, with NVG who underwent tube shunt implant and/or cyclodestructive procedure, the survival rate was 62% compared to 80% in an age- and gender-matched control population. The mean survival after the procedure was 4 years.[8] However, these studies focused on prognosis after glaucoma drainage device implantation rather than including all subjects with NVG as our study does. In a cohort of 114 subjects in New Zealand with NVG that needed a Molteno implant, the mean survival was 6.5 years, as compared to 14 years for the age-matched population. Patients with PDR were diagnosed to have NVG at an average age of 65 years and had a mean life expectancy of 5.1 years.[16] Other studies after Krupin valve[17] and Molteno implants[18] have found high mortality rates, but these too suffer from the fact that subjects who did not need surgery were not included, and prognosis was defined as mortality from the time of surgery, not the onset of disease.
Diabetic retinopathy has consistently been identified as a risk factor for poor long-term outcomes in diabetic patients. In a study looking at mortality in diabetic patients with severe tractional retinal detachments (TRDs) requiring vitrectomy, the mean survival was 2.7 years, with a Kaplan-Meier survival curve analysis revealing a 10-year mortality rate of 48.7%.[19] Neovascular glaucoma is also an end-stage complication of diabetes. Our study included all subjects, not just the ones that needed surgery, and our Kaplan-Meier survival analysis has a lower 5-year all-cause mortality of 28% and a 10-year all-cause mortality of 34% for subjects with NVG from PDR [Table 3].
The same study of long-term mortality in patients with TRD found that 86.4% of patients experienced a cardiac complication (such as myocardial infarction, congestive heart failure, or cardiomyopathy), and 21.7% of patients experienced a stroke.[19] This aligns with our evaluation of patients with PDR, where a majority (39%) died from a cardiovascular cause.
We found that a diagnosis of CRAO to be an important predictor of mortality in NVG, with survival analysis demonstrating a startling 5-year and 10-year mortality rate of 67% and 83%, respectively, and a median survival of 4.75 years. Our mortality numbers for CRAO with NVG appear to be much higher than the mortality rates for CRAO reported in the literature. A study in Poland found an all-cause mortality in patients with CRAO of 25.5%, with a median survival time of 35.9 months.[20] Another study in Korea found that between the years 2002 and 2018, out of 15,684 patients with CRAO, approximately 21% were deceased by the end of the 14-year study period.[11] Not all CRAOs lead to NVG; the rates quoted in literature vary from 10–18%,[7,21] and our CRAO numbers are small (n = 9), which may explain the discrepancy.
To the best of our knowledge, this is the most comprehensive study of mortality in subjects with NVG. We elucidate the demographic and mortality statistics in NVG stratified by underlying cause, age, gender, and race. The survival data is critical for planning prospective or interventional studies in patients with NVG. Importantly, our study highlights the necessity to aggressively manage underlying systemic co-morbid conditions such as DM, hypertension, and other cardiovascular risk factors, which contribute to early mortality in patients with a diagnosis of NVG.
We acknowledge the limitations of this study. Due to the retrospective nature, we cannot control for all possible confounders that might influence the mortality statistics in this cohort. One important confounder is COVID-19, as the timing of our mortality assessment coincided with the pandemic. Some deaths may have been directly attributable to the infection, but distinguishing these from deaths due to other causes is difficult, particularly since many patients who succumbed to COVID-19 also had significant comorbidities. Additionally, since our study was retrospective, power calculations were performed post hoc and yielded a power of approximately 60% (at an alpha level of 0.05), which is relatively low. Future studies with higher sample sizes are warranted. We did not distinguish between type 1 and type 2 DM; however, we have included information about the age demographics of our diabetic population above. This study was conducted at a tertiary care referral center in an urban setting. Although we service a large population, our cohort likely under-represents subjects who are from rural areas, are uninsured, are unable to travel, or have difficulty getting access to health care. The first visit documenting a diagnosis of NVG may not be the true onset; however, NVG does not remain asymptomatic for a long period of time. Additionally, many of our subjects were diagnosed and treated for NVG years before the start of recruitment. Hence, we have not included this in our descriptive analysis. For the survival analysis, only subjects with a definitive date of NVG diagnosis were included. Finally, our study was designed to evaluate mortality, not ophthalmic outcomes of NVG, so we did not document procedures that the patients underwent.
Conclusion
To summarize, our cohort of NVG in a tertiary care center in Atlanta, GA, USA had a 5-year all-cause mortality of 29% and a 10-year all-cause mortality of 42% with a mean survival rate of 13 years. We found increasing age and CRAO-related NVG to be important predictors of increased mortality.
Conflicts of interest:
There are no conflicts of interest.
Funding Statement
Research to Prevent Blindness, New York, NY, unrestricted grant; Emory Eye Center Departmental Grant, Atlanta, GA.
References
- 1.Shazly TA, Latina MA. Neovascular glaucoma: Etiology, diagnosis and prognosis. Semin Ophthalmol. 2009;24:113–21. doi: 10.1080/08820530902800801. [DOI] [PubMed] [Google Scholar]
- 2.Rodrigues GB, Abe RY, Zangalli C, Sodre SL, Donini FA, Costa DC, et al. Neovascular glaucoma: A review. Int J Retina Vitreous. 2016;2:26. doi: 10.1186/s40942-016-0051-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Barac IR, Pop MD, Gheorghe AI, Taban C. Neovascular secondary glaucoma, etiology and pathogenesis. Romanian J Ophthalmol. 2015;59:24–8. [PMC free article] [PubMed] [Google Scholar]
- 4.Havens SJ, Gulati V. Neovascular glaucoma. Dev Ophthalmol. 2016;55:196–204. doi: 10.1159/000431196. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Artells de Jorge N, Tabernero SS, Millas SC, Galicia CM, Gálvez MI, Leal LM, et al. High cardiovascular mortality in patients with diabetic neovascular glaucoma. Invest Ophthalmol Vis Sci. 2018;59:1901. [Google Scholar]
- 6.Senthil S, Dada T, Das T, Kaushik S, Puthuran GV, Philip R, et al. Neovascular glaucoma-A review. Indian J Ophthalmol. 2021;69:525–34. doi: 10.4103/ijo.IJO_1591_20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Rudkin AK, Lee AW, Chen CS. Ocular neovascularization following central retinal artery occlusion: prevalence and timing of onset. Eur J Ophthalmol. 2010;20:1042–6. doi: 10.1177/112067211002000603. [DOI] [PubMed] [Google Scholar]
- 8.Zhou Y, Coleman S, Boysen J, Pansegrau ML, Wright MM, Carpel EF, et al. Survival in patients with neovascular glaucoma following tube shunt implant or cyclodestructive procedure. J Curr Glaucoma Pract. 2022;16:74–8. doi: 10.5005/jp-journals-10078-1357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Modjtahedi BS, Wu J, Luong TQ, Gandhi NK, Fong DS, Chen W. Severity of diabetic retinopathy and the risk of future cerebrovascular disease, cardiovascular disease, and all-cause mortality. Ophthalmology. 2021;128:1169–79. doi: 10.1016/j.ophtha.2020.12.019. [DOI] [PubMed] [Google Scholar]
- 10.Bertelsen M, Linneberg A, Christoffersen N, Vorum H, Gade E, Larsen M. Mortality in patients with central retinal vein occlusion. Ophthalmology. 2014;121:637–42. doi: 10.1016/j.ophtha.2013.07.025. [DOI] [PubMed] [Google Scholar]
- 11.Hwang DDJ, Lee KE, Kim Y, Kim MS, Rim TH, Kim M, et al. Incidence of Retinal Artery Occlusion and Related Mortality in Korea, 2005 to 2018. JAMA Netw Open. 2023;6:e233068. doi: 10.1001/jamanetworkopen.2023.3068. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Data Access-National Death Index. Available from: https://www.cdc.gov/nchs/ndi/index.htm . [Last accessed on 2023 Jun 16]
- 13.One arm survival power/sample size calculator. https://stattools.crab.org/Calculators/oneNonParametricSurvival.htm . [Last accessed on 2025 Jun 04] [Google Scholar]
- 14.Products - Data Briefs - Number 456 - December 2022. doi: 10.15620/cdc: 122516. [Google Scholar]
- 15.Frimpong AA, Chang TL, Weiss JM, Assanah-Lewis BG, Lu MC, Nwanyanwu KH. Proliferative diabetic retinopathy disproportionately impacts distressed communities near a northeastern academic center. Ophthalmol Sci. 2025;5:100872. doi: 10.1016/j.xops.2025.100872. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Blanc JP, Molteno ACB, Fuller JR, Bevin TH, Herbison P. Life expectancy of patients with neovascular glaucoma drained by Molteno implants. Clin Experiment Ophthalmol. 2004;32:360–3. doi: 10.1111/j.1442-9071.2004.00837.x. [DOI] [PubMed] [Google Scholar]
- 17.Krupin T, Kaufman P, Mandell AI, Terry SA, Ritch R, Podos SM, et al. Long-Term Results of Valve Implants in Filtering Surgery for Eyes with Neovascular Glaucoma. Am J Ophthalmol. 1983;95:775–82. doi: 10.1016/0002-9394(83)90064-8. [DOI] [PubMed] [Google Scholar]
- 18.Mermoud A, Salmon JF, Alexander P, Straker C, Murray AD. Molteno tube implantation for neovascular glaucoma. Long-term results and factors influencing the outcome. Ophthalmology. 1993;100:897–902. [PubMed] [Google Scholar]
- 19.Shukla SY, Hariprasad AS, Hariprasad SM. Long-term mortality in diabetic patients with tractional retinal detachments. Ophthalmol Retina. 2017;1:8–11. doi: 10.1016/j.oret.2016.09.002. [DOI] [PubMed] [Google Scholar]
- 20.Roskal-Wałek J, Wałek P, Biskup M, Odrobina D, Mackiewicz J, Głuszek S, et al. Central and branch retinal artery occlusion—Do they harbor the same risk of further ischemic events? J Clin Med. 2021;10:3093. doi: 10.3390/jcm10143093. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Jung YH, Ahn SJ, Hong JH, Park KH, Han MK, Jung C, et al. Incidence and clinical features of neovascularization of the Iris following acute central retinal artery occlusion. Korean J Ophthalmol. 2016;30:352–9. doi: 10.3341/kjo.2016.30.5.352. [DOI] [PMC free article] [PubMed] [Google Scholar]