Retina••

Prevalence of homocysteinemia and effect of vitamin supplementation in retinal vein occlusion

Abstract

Objective The objective of this study was to determine the prevalence of homocysteinemia in patients with retinal vein occlusion (RVO). We investigated the association of B complex vitamin (BCV) and multivitamin (MVI) supplementation on homocysteine levels in RVO patients.

Methods and analysis This cross-sectional study occurred at the Retina Center of New Jersey. We investigated 312 patients diagnosed with RVO between 2011 and 2019. Homocysteine levels were measured on diagnosis of RVO and 4–8 weeks after, following recommended daily supplementation with BCV, MVI or combination MVI+BCV.

Results The median patient age was 71.00, IQR (61.00, 78.25), with 164 (52.6%) being women. Prevalence of homocysteinemia (>13 µmol/L) was 150 of 312 (48.1%), with a median baseline homocysteine level of 12.80 (10.17, 15.90) µmol/L. The follow-up cohort, 105 patients, demonstrated significant reduction in median homocysteine levels from 14.50 (12.30, 17.90) to 10.70 (9.30, 13.50) µmol/L following vitamin supplementation. Multivariate analysis found having baseline homocysteinemia was significantly associated with at least 25% reduction in homocysteine following vitamin supplementation.

Conclusions In this study, 48.1% of patients with RVO had elevated homocysteine (>13 µmol/L). Supplementation with BCV or MVI+BCV was associated with a significant reduction in homocysteine from 14.50 (12.30, 17.90) to 10.70 (9.30, 13.50) µmol/L.

What is already known on this topic

  • Elevated homocysteine levels may lead to a hypercoagulable state, with some studies reporting this to be an independent risk factor for arterial and venous thrombotic diseases, such as stroke and coronary artery disease. Data from previous studies investigating the relationship between elevated homocysteine levels and incidence of retinal vein occlusion (RVO) have been mixed.

What this study adds

  • Comparing our results to the findings from the National Health and Nutrition Examination Survey dataset suggest a higher prevalence of homocysteinemia in patients with RVO than similarly aged adults in the USA general population. B vitamin supplementation was associated with a significant decrease in homocysteinemia in this subpopulation; while the effects of vitamin supplementation on lowering homocysteine levels have been reported in the general population,25 our study identified a research gap in the present literature, as we could not find any studies demonstrating homocysteine levels in patients with RVO at baseline and following supplementation with B complex vitamin and multivitamin.

How this study might affect research, practice or policy

  • Patients who present with RVO may warrant broader medical examination for diseases related to underlying hypercoagulable states including the testing of baseline homocysteine levels, which are not currently covered by Medicare, and further research is needed to determine if homocysteine reduction with vitamin supplementation would be effective in reducing the incidence or improving the disease progression in patients with RVO.

Introduction

Retinal vein occlusions (RVOs) are severe causes of visual impairment that occur more frequently in the elderly population.1 2 RVOs affect vessels of different sizes and can be segmented into six main categories, including ischaemic central retinal vein occlusion (CRVO), non-ischaemic CRVO, ischaemic branch retinal vein occlusion (BRVO), non-ischaemic BRVO, ischaemic Hemi-CRVO and non-ischaemic Hemi-CRVO.3 The aetiology underlying RVOs is not well understood. Previous studies have demonstrated that risk factors associated with RVOs include age, hypertension, elevated homocysteine levels, open angle glaucoma, obesity, cardiovascular disease, diabetes mellitus, dyslipidaemia and antiphospholipid antibodies.4–9 It has been postulated that patients with clotting cascade mutations involving factor V Leiden and prothrombin, in addition to patients with deficiencies in physiological coagulation inhibitors such as plasminogen activator inhibitor, may also be at higher risk for RVO.7 10 11

Elevated homocysteine levels may lead to a hypercoagulable state, with some studies reporting this to be an independent risk factor for arterial and venous thrombotic diseases, such as stroke and coronary artery disease.12–14 Data from previous studies investigating the relationship between elevated homocysteine levels and incidence of RVO have been mixed, with some suggesting a positive association,10 15–17 or no association.18 Similarly, mixed findings have also been published regarding the relationship between C677T and A1298C methylenetetrahydrofolate reductase (MTHFR) genotype polymorphisms and RVO.15–19

Homocysteine is an amino acid derived from methionine whose metabolism is facilitated by three enzymes: methionine synthetase, MTHFR and cystathionine synthetase.20 Folate, vitamin B12, vitamin B6 and flavin adenine dinucleotide are coenzymes necessary for these reactions.20 It has been reported that elevated levels of homocysteine may lead to vascular disease via a combination of vascular endothelial dysfunction, vascular smooth muscle proliferation and coagulation abnormalities, leading to venous thrombosis.21 Predisposing factors of elevated homocysteine levels include diseases such as hypothyroidism, renal failure, proliferative diseases, and type II diabetes mellitus, drugs such as metformin, estrogens, methotrexate, levodopa, and alcohol, vitamin deficiencies including folate, B12, B6 and polymorphisms of the MTHFR gene.22 It has been reported that a homozygous C677T polymorphism may result in a 70% decrease in MTHFR enzyme function,23 and a compound heterozygous A1298C and C677T polymorphism may result in a 50%–60% decrease in MTHFR enzyme function.24

Supplementation with the coenzymes necessary for homocysteine metabolism, specifically vitamins B12, B6 and B9, may increase the efficiency of the viable homocysteine metabolising pathways enough to reduce homocysteine levels to a normal level. It has been reported that folic acid supplementation (0.5–5 mg/day) can reduce total homocysteine levels by 25%, and B12 supplementation (0.5 mg/day) can reduce total homocysteine levels by an additional 7%, with a combination therapy reducing homocysteine levels up to 33%.25 26 Hence, the objective of this study was to determine the prevalence of elevated homocysteine levels in patients with RVO and to investigate the association of supplementation with B complex vitamins (BCVs), multivitamins (MVIs) or a combination of both (MVI+BCV) on homocysteine levels in these patients.

Materials and methods

Study design

This was a cross-sectional study in which patients with RVO were tested for elevated homocysteine levels and followed for 4–8 weeks after recommendation for vitamin supplementation.

Treatment and vitamin supplementation

On diagnosis of RVO, all patients were treated to the standard of care. Patients presenting with macular oedema were treated with intravitreal injections of bevacizumab, ranibizumab or aflibercept. Patents without macular oedema or decrease in vision were monitored and not treated with intravitreal injection.

Following standard of care treatment, patients were recommended to supplement their diets with a BCV containing B12, B6 and B9 and/or an MVI. The BCV formulations recommended were either FABB, TL Guard and Folplex, all containing 25 mg B6 as pyridoxine HCl, 2.2 mg B9 as folic acid and 1 mg B12 as cyanocobalamin. The MVI formulation recommended was Centrum Silver.

Homocysteine levels were recorded at baseline and at 4–8 weeks after initiation of vitamin supplementation for patients who returned to the clinic and reported following a supplementation regimen. All vitamins were advised to be taken daily for the entirety of this period. These patients were divided into three cohorts depending on which regimen they reported following: MVI only, BCV only or a combination therapy of MVI+BCV.

Diagnostic criteria

RVOs were diagnosed via ophthalmoscopic fundus examination and fluorescein angiography. Ophthalmoscopic fundus examination assessed for dilated or tortuous veins, disc swelling and hyperaemia, cystoid macular oedema, retinal haemorrhages, microaneurysms, cotton–wool spots, and retinal and iris neovascularisation. Fluorescein angiography evaluated increased venous transit time, areas of capillary closure and venous filling defects.

Fasting homocysteine levels were measured, and patients were considered to have elevated homocysteine (homocysteinemia) if their total fasting plasma homocysteine level was>13 µmol/L. MTHFR polymorphisms were assessed via PCR and restriction enzyme analysis.

Study population selection

We identified 475 consecutive patient records with a diagnosis of BRVO, CRVO or Hemi-CRVO who presented to the Retina Center of New Jersey from October 2011 to June 2019. We excluded all patients who did not have baseline data for homocysteine levels and patients who had pre-existing RVO before the defined study period (N=104). We then excluded all duplicate patient records (N=59), yielding a total study population of 312 unique patients for our baseline characteristic analyses. For our follow-up analyses on the association of vitamin supplementation on homocysteine levels in RVO patients, we identified 109 patients who were previously untreated with MVI or BCV who had baseline and follow-up homocysteine levels after 4–8 weeks of reporting either MVI only, BCV only or combination MVI+BCV supplementation. Since there were only four patients who reported vitamin supplementation with MVI only, we excluded these patients from our follow-up cohort, yielding a total of 105 patients. A flowchart of our study population selection process is displayed in figure 1.

Figure 1
Request permissionrequest permission icon
Figure 1

Study population. Flowchart of study population selection. BCV, B complex vitamin; MVI, multivitamin; RVO, retinal vein occlusion.

Statistical analysis

We first characterised the demographics, RVO type, comorbidities, MTHFR genotype and complete blood counts (CBCs) for our primary case series with baseline homocysteine levels, stratified by age. Specifically, we characterised patients age≥60 in one cohort, and patients age<60 in the other cohort, as well as the total study population. We then characterised the demographics, prevalence of RVO type and reduction in homocysteine levels for our follow-up series with homocysteine levels after vitamin supplementation, stratified by the regimen followed (MVI only, BCV only or combination MVI+BCV). Since there were only four patients who reported supplementation with MVI only, we solely characterised patients who were on BCV only or combination MVI+BCV therapy.

We created paired boxplots to compare the median homocysteine levels before and after vitamin supplementation, stratified by the regimen followed (BCV only or combination MVI+BCV). We then created boxplots to compare the calculated medians of differences in homocysteine levels after vitamin supplementation. We used multiple logistic regression to analyse if age≥60 versus age<60 was associated with the primary outcome of this study, which was at least 25% reduction in homocysteine levels after vitamin supplementation. We chose 25% as a cut-off based on the results of a large meta-analysis finding that folic acid supplementation reduced homocysteine concentration by 25%.26 We created three models: an unadjusted model, a model adjusted for sex and a final model additionally adjusted for homocysteinemia, BRVO versus CRVO/Hemi-CRVO and any MTHFR polymorphism. We created a forest plot to visualise the ORs for the final model.

We used R V.4.0.1 for our statistical analyses.27–30 We reported categorical variables as proportions (%) and continuous variables as medians (IQR). We used the χ2 test with Yates’ continuity correction or Fisher’s exact test as appropriate to compare categorical variables. We used the Mann-Whitney U/Wilcoxon rank-sum tests to compare continuous variables and our boxplots of the medians of differences in homocysteine levels by regimen of vitamin supplementation followed. The statistical test used for our paired boxplots comparing median homocysteine levels before and after vitamin supplementation was the Wilcoxon signed-rank test. We considered p<0.05 to be statistically significant.

Patient and public involvement

Patients were not directly involved with the research process. This study was conducted as a retrospective cross-sectional study. Research questions were developed through physician experience with managing RVO and influenced by the previous literature. Patients, their priorities, experiences and preferences played no role in the development of research questions. Patient’s data was deidentified with all analysis blinded. Patients were not involved with the design and conduct of the study, choice of outcome measures or recruitment. Patients were not directly contacted in regard to this study and therefore were not asked about the burden or time involved regarding participation. Patients were not involved in planning for the dissemination of the study results; however, patients will be shown aggregated results of the study. In the interest of increasing patient and public involvement, we will elucidate from our population how to best involve them in the future.

Results

Baseline characteristics

Table 1 displays the baseline characteristics of the entire case series, stratified by two cohorts: age≥60 and age<60. Of the 312 patients with RVO and baseline homocysteine levels in this study, the median age was 71.00 (61.00, 78.25), with 241 (77.2%) being age≥60 and 164 (52.6%) being women. There were 141 patients with BRVO, 83 of these being ischaemic; 136 patients with CRVO, 68 of these being ischaemic; and 35 patients with Hemi-CRVO, 21 of these being ischaemic. The median baseline homocysteine level for the entire case series was 12.80 (10.17, 15.90) µmol/L, with 150 (48.1%) patients found to have homocysteinemia. Of these 312 patients with baseline homocysteine levels, 258 were not taking BCV or MVI supplementation at the time of RVO diagnosis and had a median baseline homocysteine level of 13.50 (11.00, 16.90) µmol/L. The other 54 patients who were taking BCV or MVI at the time of RVO diagnosis had a median baseline homocysteine level of 9.90 (7.90, 11.80) µmol/L. A variety of MTHFR genotypes were identified in this study, with 246 (81.5%) patients having at least one MTHFR polymorphism. All CBCs had median values within normal range.

Table 1
|
Baseline characteristics

Notably, homocysteinemia was more prevalent in patients age≥60 (53.9% vs 28.2%, p<0.001), with median baseline homocysteine levels being higher as well (13.40 (10.70, 16.50) vs 11.60 (8.10, 13.35), p<0.001). Hypertension was more prevalent in patients age≥60 (87.0% vs 72.5%, p=0.007). Having at least one MTHFR polymorphism was more prevalent in patients age≥60 (84.0% vs 72.3%, p=0.05). There were no significant differences between patients age≥60 and <60 regarding sex, race/ethnicity, hypercholesterolaemia, diabetes and individual MTHFR polymorphisms. Similarly, we did not identify associations for most CBCs, except for median RBC and platelet levels.

Association of vitamin supplementation on homocysteine levels

Of the 105 patients in our follow-up cohort, 56 reported using combination MVI+BCV and 49 reported using BCV only. There were no significant differences in demographics between patients who reported using combination MVI+BCV and BCV only (table 2). Within each regimen followed, there was a statistically significant decrease in median homocysteine levels following supplementation with BCV only (−3.40 (–5.20, –2.10) µmol/L) and combination MVI+BCV (−3.30 (−6.12, –1.80) µmol/L, figure 2A). Supplementation with BCV only reduced median homocysteine levels from 14.40 (12.30, 17.90) to 11.10 (9.30, 13.70) µmol/L (p<0.001, figure 2A), while combination MVI+BCV supplementation reduced median homocysteine levels from 14.60 (12.23, 18.90) to 10.35 (9.40, 13.20) µmol/L (p<0.001, figure 2A). When comparing the medians of differences in homocysteine reduction for each regimen followed, there was no statistically significant difference (p=0.992, figure 2B).

Figure 2
Request permissionrequest permission icon
Figure 2

Boxplots demonstrating association of vitamin supplementation on homocysteine levels. (A) Supplementation with B complex vitamin (BCV) only and combination multivitamin (MVI)+BCV were both associated with significantly reduced homocysteine levels in patients with retinal vein occlusion after 4–8 weeks of follow-up. (B) There was no statistically significant difference in the medians of differences following either supplementation regimen.

Table 2
|
Effect of vitamin supplementation in follow-up patients

Analysis of the entire follow-up cohort revealed the median homocysteine level before vitamin supplementation was 14.50 (12.30, 17.90) and 10.70 (9.30, 13.50) µmol/L after vitamin supplementation, a significant reduction (table 2). Of these patients, 69 (65.7%) met the criteria for homocysteinemia before vitamin supplementation, while 29 (27.6%) met the criteria after vitamin supplementation, and hence vitamin supplementation was associated with a significant decrease in homocysteinemia within this subpopulation (p<0.001).

Multiple logistic regression

In our three models, the ORs for a homocysteine level reduction of at least 25% were not statistically significant when comparing patients age≥60 vs age<60. In our final adjusted model, this OR was 1.69 (95% CI: 0.57 to 5.26). All covariate ORs in our final adjusted model are displayed in a forest plot in figure 3. We did not find any statistically significant ORs for most of the other covariates (age, sex and any MTHFR polymorphism) except for homocysteinemia, which had an adjusted OR of 4.26 (95% CI: 1.74 to 11.14).

Figure 3
Request permissionrequest permission icon
Figure 3

Association of vitamin supplementation on homocysteine reduction of at least 25%. Multiple logistic regression adjusting for age≥60, homocysteinemia, branch retinal vein occlusion (BRVO), any methylenetetrahydrofolate reductase (MTHFR) polymorphism and sex. Baseline homocysteinemia was the only covariate associated with a homocysteine reduction of at least 25% following vitamin supplementation. **This was the only covariate studied associated with a statically significant homocysteine reduction of at least 25% following vitamin supplemenation.

Discussion

The prevalence of homocysteinemia in this population of 312 patients with RVO was 150 (48.1%), with a median baseline homocysteine level of 12.80 (10.17, 15.90) µmol/L. In the general US population, data from the National Health and Nutrition Examination Survey (NHANES) found a prevalence of homocysteinemia in 18.5% (95% CI: 16.3% to 21.0%) of adults≥60 between 2003 and 2006.31 In our study, for 241 adults≥60, the prevalence of homocysteinemia was 130 (53.9%). In the NHANES dataset, for adults from ages 40–59, the prevalence of homocysteinemia was 7.0% (95% CI: 5.6% to 8.0%), while in our study the prevalence of homocysteinemia was 20 of 71 (28.2%) in similarly aged adults<60.31 Homocysteine levels increase with age due to a variety of factors, including decreased renal function and nutrition, and so this trend was expected in both our results and in NHANES.32 Importantly, the NHANES dataset defined elevated homocysteine as>13 µmol/L,31 the same cut-off used in our study. Comparing our results to the findings from the NHANES dataset suggest a higher prevalence of homocysteinemia in patients with RVO than similarly aged adults in the US general population. Considering that elevated homocysteine levels are associated with arterial and venous thrombotic diseases, such as stroke and coronary artery disease, patients who present with RVO may warrant broader medical examination for related diseases.10–12 Additionally, the relatively high prevalence of homocysteinemia in RVO patients provides a strong rationale for testing baseline homocysteine levels in this patient population, a lab not currently covered by Medicare.33

Regarding the effect of vitamin supplementation on homocysteine levels in RVO patients, there was a statistically significant decrease in median homocysteine levels following recommendation of BCV only or combination MVI+BCV supplementation (figure 2A). Overall, in the follow-up cohort of 105 patients, 69 patients (65.7%) met the criteria for homocysteinemia before vitamin supplementation, while 29 (27.6%) met the criteria after vitamin supplementation, a significant reduction. BCV only supplementation was associated with a median homocysteine level reduction of −3.40 (−5.20, –2.10) µmol/L, decreasing median levels from 14.40 (12.30, 17.90) to 11.10 (9.30, 13.70) µmol/L. Combination MVI+BCV supplementation was associated with a median homocysteine level reduction of −3.30 (−6.12, –1.80) µmol/L, decreasing median levels from 14.60 (12.23, 18.90) to 10.35 (9.40, 13.20) µmol/L. When comparing the medians of differences in homocysteine reduction for each regimen followed, there was no statistically significant difference (figure 2B). Additionally, multiple logistic regression found that having baseline homocysteinemia was associated with an outcome of at least 25% reduction in homocysteine levels following vitamin supplementation (figure 3). As homocysteinemia potentially requires additional enzyme activity to metabolise, supplemental coenzymes in the form of B6, B9 and B12 may have helped catalyse these reactions, and resulted in a significant reduction in homocysteine.20 In patients with RVO, there is likely benefit in having the baseline B6, B9 and B12 levels determined to assess for obvious deficiencies in the homocysteine metabolism pathway.

Additionally, in our multiple logistic regression analysis, age, sex, type of RVO and having at least one MTHFR polymorphism did not have any statistical effect on the association of vitamin supplementation with lowering homocysteine levels by at least 25%. Taken together, our study suggests that supplementing with either BCV only or combination MVI+BCV is sufficient in lowering homocysteine levels in patients with RVO, regardless of baseline demographics or genotyping. Based on the increased prevalence of homocysteinemia in older patients, screening this population for elevated homocysteine and potentially treating them with vitamin supplementation may be warranted. Importantly, no conclusion can be made from this study regarding whether lowering homocysteine levels would have any effect on RVO incidence, prevention or progression.

Strengths

While the effects of vitamin supplementation on lowering homocysteine levels have been reported in the general population,25 26 our study identified a research gap in the present literature, as we could not find any studies demonstrating homocysteine levels in patients with RVO at baseline and following supplementation with BCV and MVI. Despite this study only taking place in New Jersey, the patient population consisted of a wide demographic that closely mirrored the NHANES dataset; specifically, our patients were 27.1% Hispanic, 15.5% black and 52.6% white, while the NHANES population was 20.2% Mexican Americans, 20.8% non-Hispanic black and 51.9% non-Hispanic white,31 which helps increase the generalisability of our study to the US population.

Limitations

Being a case series, our study lacked matched controls, limiting the conclusions we could make regarding our findings and the general population without RVO. As a single-centre study, our patient population was limited to a narrow geographic location, which limited generalisability. We implemented an observational study design due to the logistic constraints of monitoring subjects’ daily vitamin intake, which limited our level of evidence. Despite our study finding an association with vitamin supplementation and reduced homocysteine levels in patients with RVO, it does not demonstrate an association between homocysteine reduction therapy and RVO incidence, prevention or progression, and further research is needed in this area.

Conclusion

In this study, 150 of 312 (48.1%) patients with RVO met the criteria for diagnosis of homocysteinemia at baseline. The median homocysteine level of the baseline cohort of 312 patients was 12.80 (10.17, 15.90) µmol/L. For the follow-up cohort of 105 patients previously not supplemented with BCV or MVI who had homocysteine levels before and after vitamin supplementation, the median homocysteine level before supplementation was 14.50 (12.30, 17.90) and 10.70 (9.30, 13.50) µmol/L after. Overall, 69 (65.7%) patients met the criteria for homocysteinemia before vitamin supplementation, while 29 (27.6%) met the criteria after vitamin supplementation, and hence vitamin supplementation was associated with a significant decrease in homocysteinemia in this subpopulation (p<0.001). Further research is needed to determine if homocysteine reduction with vitamin supplementation would be effective in reducing the incidence or improving the disease progression in patients with RVO.

Overview

Article metrics
Altmetricopen-url
Dimensionsopen-url