OBJECTIVE
To determine the predictive value of the Charlson Comorbidity Index (CCI) for outcomes in patients with macular edema secondary to retinal vein occlusion (RVO).
DESIGN
Retrospective clinical cohort study.
SUBJECTS
Patients seen between 2013 and 2023 at the Cole Eye Institute, Cleveland Clinic, were included. All patients were >18, diagnosed with RVO (International Classification of Diseases (ICD)-9 and 10 codes), had a complete CCI score, and had at least 1 year of ophthalmic follow-up data after their first intravitreal injection (baseline). Patients with ocular surgery, trauma, or panretinal photocoagulation were excluded.
METHODS
Age-adjusted CCI scores were calculated for each patient from chart review. For patients with bilateral RVO, one eye was selected randomly. Patients were stratified into tertiles by CCI distribution: tertile 1 (CCI 0-5; mean 3.4), tertile 2 (age-CCI 4.1-6; mean 4.9), and tertile 3 (CCI ≥ 8; mean 9.6). Multivariable linear regression was performed to determine the predictive value of CCI and other covariates on visual and anatomical outcomes.
MAIN OUTCOME MEASURES
Best-corrected visual acuity (BCVA) and central subfield thickness (CST) at 1-year follow-up.
RESULTS
A total of 972 patients met all criteria, with an average age-adjusted CCI score of 6.2. Each one-point increase in CCI predicted 0.38 fewer letters in BCVA at follow-up ( P <.001). Baseline BCVA was a significant predictor of follow-up BCVA in all tertiles ( P <.001). In the third tertile, each one-point increase in CCI was associated with a 0.72 letter reduction in follow-up BCVA ( P <.001). For CST, baseline CST was strongly predictive of final CST ( P <.001), while CCI was only significant in the first tertile, where each point increase in CCI predicted a 13.7 µm increase in CST ( P =.02). In RVO subtype interaction models, the age-adjusted CCI × CRVO interaction was not statistically significant for either 1-year BCVA ( P =.269) or 1-year CST ( P =.695).
CONCLUSIONS
Higher CCI scores are significantly associated with worse visual outcomes in patients with RVO, particularly in the combined population and most comorbid patients (tertile 3). CCI was significantly associated with higher (thicker) CST only among the least comorbid patients (tertile 1).
INTRODUCTION
R etinal vein occlusion (RVO) is the second most common retinal vascular disease after diabetic retinopathy and a significant cause of visual morbidity, particularly among older adults. , RVO is an obstruction of the retinal venous system that elevates venous pressure, causes retinal hemorrhage and macular edema, and can threaten vision. ,,, It includes central RVO (CRVO) and branch RVO (BRVO) subtypes, both of which can result in macular edema, hemorrhage, and neovascular complications. , First-line management of RVO-related macular edema is intravitreal anti-Vascular Endothelial Growth Factor (VEGF) therapy; corticosteroid implants are options in selected cases such as pseudophakia or inadequate anti-VEGF response. ,,, Systemic conditions such as hypertension, diabetes, and cardiovascular disease are established risk factors for the development of RVO and its complications. These conditions may exacerbate retinal ischemia and inflammation, worsen macular edema and neovascularization, and ultimately compromise visual outcomes. , Systemic comorbidities also promote endothelial dysfunction, hypercoagulability, and pro-inflammatory/VEGF-driven signaling, mechanisms that can dampen anatomic and visual responses to anti-VEGF therapy. ,,
While individual comorbidities have been linked to worse ocular outcomes in other retinal diseases, the cumulative burden of systemic illness and its impact on RVO have not been well quantified. , The Charlson Comorbidity Index (CCI) is a validated tool that assigns weighted scores to 19 medical conditions and incorporates age to quantify overall comorbidity burden ( Table 1 ). , It has been widely used across medical specialties to predict mortality, hospital readmission, and surgical outcomes. , In ophthalmology, CCI has been incorporated into electronic health record-based risk stratification approaches, demonstrating its feasibility as a summary measure of systemic comorbidity burden in ophthalmic outcomes research. However, to date, no studies have examined whether CCI is associated with treatment outcomes in patients with RVO.
TABLE 1
Age-CCI Scoring.
| Item | Weight for Each Condition |
|---|---|
| Myocardial infarction | 1 |
| Congestive heart failure | 1 |
| Peripheral vascular disease | 1 |
| Cerebrovascular disease | 1 |
| Dementia | 1 |
| Chronic pulmonary disease | 1 |
| Connective tissue disease | 1 |
| Ulcer disease | 1 |
| Mild liver disease | 1 |
| Diabetes | 1 |
| Hemiplegia | 2 |
| Moderate or severe renal disease | 2 |
| Diabetes with end-organ damage | 2 |
| Any tumor without metastasis | 2 |
| Leukemia | 2 |
| Lymphoma | 2 |
| Moderate or severe liver disease | 3 |
| Metastatic solid tumor | 6 |
| HIV/AIDS | 6 |
| Age 50-60 | 1 |
| Age 60-70 | 2 |
| Age 70-80 | 3 |
| Age 80+ | 4 |
Age-adjusted Charlson Comorbidity Index (Age-CCI) scoring schema. This table lists the weights assigned to each of the 19 Charlson comorbidities and to age bands; a patient’s Age-CCI equals the sum of the condition weights plus the age weight. Age-CCI was used for tertile stratification (T1-T3) and as a predictor in regression models. Age contributes 0 points for patients younger than 50 years, then increases by decade beginning at age 50.
This study seeks to evaluate the association between age-adjusted CCI and visual (best corrected visual acuity, BCVA) and anatomical (central subfield thickness, CST) outcomes in patients with RVO undergoing anti-VEGF therapy over a 1-year period. Given the limited application of the CCI in ophthalmology and the absence of prior investigations within the context of RVO, this study aims to address a notable gap in the literature. Hypothetically, higher CCI scores will correlate with poorer outcomes, and this relationship may be influenced by demographic and socioeconomic variables. If further validated, incorporating CCI at baseline could support risk stratification, closer follow-up intervals, and earlier multidisciplinary coordination for patients at higher risk of suboptimal outcomes.
METHODS
STUDY PATIENTS
This retrospective clinical cohort study was conducted at the Cole Eye Institute, Cleveland Clinic, with Institutional Review Board approval. Data were abstracted from the electronic medical record and handled in compliance with the Health Insurance Portability and Accountability Act. All procedures adhered to the tenets of the Declaration of Helsinki, with appropriate safeguards for privacy and data security. Inclusion criteria were patients aged 18 or older at the time of first anti-VEGF injection, diagnosed with RVO between 2013 and 2023 based on ICD-9/10 codes, with complete data for CCI calculation, and at least 1 year of ophthalmic follow-up after the initial intravitreal injection. Patients were excluded if they had ocular surgery, trauma, or panretinal photocoagulation during or prior to the study period. For patients with bilateral RVO, one eye was randomly selected to preserve independence of observations and avoid bias from inter-eye correlation in models that were not designed for clustered eye-level data.
CCI scores were computed using the age-adjusted 19-item Charlson algorithm. Patients younger than 50 years received 0 age points in the age-adjusted CCI, consistent with the standard Charlson scoring system. Patients were stratified into tertiles based on CCI scores, an approach previously used in other specialties to evaluate associations between comorbidity burden and clinical outcomes. Age-adjusted CCI was computed using the validated 19-item Charlson algorithm with age weighting ( Table 1 ). , Patients were stratified into tertiles by baseline age-adjusted CCI distribution: tertile 1 (age-CCI 0-4; n = 324; mean 2.4), tertile 2 (age-CCI 4.1-6; n = 324; mean 4.9), and tertile 3 (age-CCI ≥ 6.1; n = 324; mean 9.0). In addition to tertile analyses, all patients were analyzed together as the combined population (CP). Outcome measures were BCVA (ETDRS letters) and CST; secondary variables included race, insurance, and smoking status categorized from standardized intake fields (never, former, light smoker, smokes some days, smokes daily) without reclassification. ,
UNIVARIATE ANALYSIS
Descriptive statistics, paired t tests, and ANOVA were conducted in Microsoft Excel. t Tests compared baseline and the visit closest to 1 year after baseline (≈12 months) within each tertile and the CP, while ANOVA assessed mean differences between tertiles. A significance threshold of α = 0.05 was used. The “1-year” outcome visit was defined as the encounter closest to 365 days after the baseline injection, within a 9-15 month window; if multiple visits occurred within that interval, the visit nearest to day 365 was selected. BCVA and CST analyses used all available observations; therefore, the number of included eyes differed between analyses depending on the available data at a given timepoint.
MULTIVARIATE ANALYSIS
Multivariable linear regression was performed in R Studio (version 2024.04.2+764) using the readxl, dplyr, and ggplot2 packages. Models predicted 1-year BCVA and CST using age-adjusted CCI, baseline BCVA or CST, race, sex (defined at birth), smoking status, insurance type, and total intravitreal anti-VEGF injections during the 1-year follow-up period. Treatment was delivered in a real-world, nonprotocolized manner at the discretion of the treating retina specialist. Anti-VEGF agent selection, visit interval, and decisions regarding continued treatment, extension, switching, or addition of corticosteroid therapy reflected routine clinical care rather than a single standardized protocol. Race was categorized as white, black, or non-black minority. Insurance status was grouped into private, government (Medicare/Medicaid), or none. Two-sided P values are reported for model coefficients, with α = 0.05 considered statistically significant; overall model fit was evaluated using F -statistics and R 2. To evaluate whether RVO subtype (central RVO [CRVO] or branch RVO [BRVO]) modified the relationship between systemic comorbidity burden and outcomes, additional multivariable interaction models were performed in the combined population with an age-adjusted CCI × RVO subtype term for 1-year BCVA and 1-year CST. These models adjusted for the same covariates as the primary multivariable models, including baseline outcome value, sex, smoking status, race, insurance class, and total procedures. The x-intercept (Appendix 1) of the fitted regression line was computed as − β ₀/ β ₁.
RESULTS
DEMOGRAPHICS
A total of 972 patients met inclusion criteria; the overall mean age was 77 years. By tertile, mean ages were 70 (T1), 69 (T2), and 82 (T3). White patients comprised 79.1% of the cohort, Black 15.3%, and non-White minority 5.6%. Men comprised 44.8% (435/972) of the combined population ( Table 2 ). Prevalence of diabetes, renal disease, congestive heart failure, and COPD increased monotonically from T1 to T3, consistent with higher age-CCI reflecting greater systemic burden.
TABLE 2
Patient Demographics.
| Cohort | All Patients | Tertile 1 | Tertile 2 | Tertile 3 |
|---|---|---|---|---|
| Patients | 972 | 324 | 324 | 324 |
| Mean CCI score | 5.4 ± 3.2 | 2.4 ± 1.1 | 4.9 ± 0.7 | 9.0 ± 2.7 |
| Average age | 77 | 70 | 69 | 82 |
| 20-29 years old | 2 | 2 | 0 | 0 |
| 30-39 years old | 3 | 3 | 0 | 0 |
| 40-49 years old | 14 | 13 | 1 | 0 |
| 50-59 years old | 44 | 34 | 7 | 3 |
| 60-69 years old | 155 | 87 | 41 | 27 |
| 70-79 years old | 306 | 118 | 105 | 83 |
| 80-89 years old | 294 | 52 | 107 | 135 |
| 90-99 years old | 154 | 15 | 63 | 76 |
| Male | 435 | 144 | 143 | 148 |
| Female | 537 | 180 | 181 | 176 |
| White | 769 | 260 | 254 | 255 |
| Black | 149 | 44 | 52 | 53 |
| Non-Black minority | 54 | 20 | 18 | 16 |
| Never smoker | 489 | 176 | 167 | 146 |
| Secondhand smoke | 0 | 0 | 0 | 0 |
| Former smoker | 419 | 119 | 140 | 160 |
| Light smoker | 3 | 1 | 0 | 2 |
| Smokes some days | 13 | 6 | 4 | 3 |
| Smokes everyday | 48 | 22 | 13 | 13 |
| Private insurance plan | 300 | 156 | 76 | 68 |
| Government insurance plan | 610 | 142 | 228 | 240 |
| No listed insurance plan | 62 | 26 | 20 | 16 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
