Risk Factors Associated With Retinal Vein Occlusion
Risk Factors Associated With Retinal Vein Occlusion
In this study, the characteristics of a series of patients with a first episode of RVO and the impact of potential associated risk factors were described. Subsequently, we compared the observed prevalence for these factors in the RVO population with three different general populations in the case of arterial risk factors, and with those for one cohort of young patients with DVT/PE or with those found in the literature in the case of venous factors. Among the arterial risk factors, hypertension and overall BP levels were the main associated factors, whereas only hyperhocysteinaemia and aPLAs were relevant among the venous risk factors for thrombosis.
Vascular diseases, especially coronary heart disease and stroke, are the leading causes of morbidity and mortality in industrialised countries. Although retinal vascular disease does not compromise lifespan, it can have a deleterious effect on the quality of life of those affected and, therefore, it is important to determine the main associated risk factors in order to prevent the disease. Both arterial and venous risk factors are thought to be involved in RVO, based on the fact that the retinal veins and arteries share a common adventitious.
In this study, there were more affected males than females (58% vs. 42%), with a mean age of 59 years. This indicates that it is not a disease only of the elderly people, but that it also affects active middle-aged subjects and, therefore, can have a significant impact on productivity. The CRVO or BRVO can, in the short term, lead to macular oedema with or without a decrease in perfusion, and in the long term, to neovascularisation, retinal deterioration, neovascular glaucoma or vitreous haemorrhaging, all of which are causes of vision loss.
The BRVO is three to five times more frequent than is CVRO, and it normally occurs as a result of compression at the crossroads junctions with an adjacent artery. A similar ratio was observed in this study and has been described in previous studies. The venous stasis caused by the increase in pressure or stiffness of the adjacent artery seems to be one of the main predisposing factors for RVO. This hypothesis easily explains why the central retinal vein is less affected, and this is related to its large calibre and location. This study did not have any bilateral occlusion that supports localised stasis problems.
The presence of ocular hypertension as a cause of RVO is discarded, since only a few patients (11) had high ocular pressure (≥ 21 mmHg) at the moment of the event. Even in those 15 previously diagnosed with ocular hypertension and under treatment, only five (33.3%) had high ocular pressure. From our data, it is possible to speculate that ocular hypertension could be particularly relevant in those subjects with very high ocular pressure levels in one eye, as has been shown by the fact that the affected eye was the unilateral one.
The presence of inherited thrombophilia in our series was very rare and similar to that observed in the general population. In fact, none of the patients suffered from protein C, S or antithrombin III deficiency, only seven had Factor V Leiden, and three the G20210A mutation of the prothrombin gene. These percentages are significantly lower than the ones observed in our cohort of young patients with DVT-PE or than the estimations for populations with DVT or PE in the literature. These results seem to be the same even for the young patients with RVO, although this subsample was quite small and still the differences in age as compared with the DVT group were highly significant. Therefore, our results confirm the findings of previous studies in which inherited thrombophilia had little influence on RVO, so its determination would not be justified a priori.
Hypercoagulability, as a result of immune-complexes against the membrane phospholipids or plasmatic proteins (LA, aPLAs) can also lead to RVO. This can be related to the inhibition of the production of prostacyclin by the vascular endothelium, the increase in thromboxane production by the platelet, the increase in the release of tissue factor and the presence of vascular endothelium adhesion molecules that interfere with the fibrinolytic process. In the present study, there were 13 patients with prevalence significantly higher than the one observed in general population and similar to that observed in the cohort of DVT/PE or the estimated one for thrombosis populations. Carbone et al. found that patients with aPLAs and retinal thrombosis had a greater frequency of antinuclear antibodies, low levels of factor 4 of the complement (C4) and the presence of circulating immune-complexes, something that could be the first manifestation of APS or a Systemic Lupus Erythematosus with severe repercussions in other target organs. Homocysteine, which can be considered both an arterial, as well as a venous risk factor, was elevated in 16 patients. Again, this prevalence is significantly higher than that reported for the general population, slightly higher than the estimated one for thrombosis populations and similar to the one observed in our group with DVT/PE. Hyperhomocysteinaemia produces a blockade of the union of the tissue plasminogen activator to the endothelial cell and, therefore, can block the release of thrombomodulin by the endothelial cell, activate coagulation factors V and VII and inhibit protein C. These findings, observed in our cohort, were confirmed by the meta-analysis performed by Janssen and colleagues in patients with RVO, including 11 case–control studies. The odds ratio was 8.9 (95% Cl 5.7–13.7) for hyperhomocysteinaemia and 3.9 (95% Cl 2.3–6.7) for anticardiolipin antibodies.
Hyperhomocysteinaemia has also been associated with BP in the literature, and therefore the effect of these two factors could be closely related. In this sense, we did not find differences in homocystein levels or for the prevalence of hyperhomocysteinaemia in the RVO group for those with or without hypertension (15.2 vs. 13.9 μmol, p-value 0.34 and 14.8 vs. 14.6 μmol, p-value 0.90, for normotensives and hypertensives before and after reclassification of the hypertension status; 19.6 vs. 12.7%, p-value 0.3 and 15.7 vs. 16.4%, p-value 0.9 for the prevalence of hyperhomocysteinaemia for normotensives and hypertensives before and after reclassification of the hypertension status).
Another important issue is that the D-dimer, which is of great value in the diagnosis of an acute episode of DVT/PE or in ischaemic stroke, is only elevated in a minority of patients with RVO (only 17 subjects had D-dimer ≥ 250 mg/dl). This parameter has been studied very little in relation to retinal vascular pathology, and therefore a more comprehensive evaluation would be necessary.
As an arterial risk factor, initially only 47 patients (47%) were diagnosed with hypertension. After reclassification, taking into account BP levels over 140/90, a significant increase in this prevalence was observed (from 47 to 80%). Although we cannot assume a correct diagnosis for all of them, surely among them are many with undiagnosed hypertension. This is important because RVO could be the first clinical manifestation of an essential hypertension in many patients. Moreover, the high systolic and diastolic BP levels seen in both groups, the supposed normotensives and the hypertensives, when compared with those in studies of the general population, indicate a lack of BP control. This is confirmed by the meta-analysis performed by O'Mahoney and collaborators, which included 2916 subjects with RVO. In this meta-analysis, the prevalence of hypertension was 63%, significantly higher than the prevalence observed in the comparative population (36.2%), with an OR for hypertension of 3.5 (95% CI 2.5–5.1). Taking into account our findings and the results of the meta-analysis, hypertension seems to be a key factor in the development of RVO. Probably, the increase in arterial stiffness as a result of high BP levels facilitates the venous stasis because the artery lies next to the vein in the retina.
For type 2 diabetes, the results were not as evident as for those with hypertension. We only observed a trend towards a higher prevalence, especially after reclassification, based on fasting glucose levels ≥ 126 mg/dl. Fasting glucose levels were also significantly higher in the RVO group when compared with two of the populations (Hortega and Segovia), and these differences were more evident in subjects without type 2 diabetes than in those with it. This may indicate that some kind of carbohydrate metabolism abnormality might also play a role in the development of RVO. In the O'Mahoney meta-analysis, of 2877 patients with RVO, 14.6% were diabetic as compared with 11.1% in the general population, which is in agreement with our results. In the O'Mahoney study, however fasting glucose was not evaluated, so some new type 2 diabetes diagnoses could have gone undetected.
The lipid abnormalities results were quite inconsistent. There could be a trend towards high LDL cholesterol in the RVO group, but only when compared with one of the populations; the same trend was observed for HDL cholesterol. These confusing results were also observed in the multicentre study of O'Mahoney et al. and could be related to the different criteria used in defining dyslipidaemia or the different thresholds for normal lipid levels. In this study, however, there was a noticeable twofold increase in the incidence of dyslipidaemia in patients with RVO as compared with the general population.
Finally, in 48 patients with RVO who underwent a carotid intima-media thickness evaluation, this study noted a high prevalence of arteriosclerotic lesions. This result reinforces the hypothesis of arterial risk factors as triggers of RVO. An association among ischaemic stroke, retinal thrombosis and carotid stenosis has been recently described.
Among the strengths of this study is the use of several populations for comparing the arterial factors. This minimises the possibility that the observed findings were by chance. Moreover, the comparison populations were matched by age, sex and BMI to reduce the impact of said variables on the prevalence of the main cardiovascular risk factors. In the case of the venous factors, we took advantage of one large series of young patients with DVT/PE who are perfectly characterised regarding the risk factors for venous thrombosis. Additionally, the observed prevalence of venous risk factors in the literature for both populations with DVT/PE and population-based studies was reviewed.
Among the main weaknesses is that which derives from the differences in the methodology of the different studies used for comparison. This is especially important in the case of dyslipidaemia, not only because of the different definitions of dyslipidaemia but also because of the lack of information. As examples, in the Segovia population, lipids data were missing in a large number of patients, and in the Pizarra one, we did not have a register of those subjects who were already dyslipidaemics or under hypolipaemic treatment. Taking into account the characteristics of the Pizarra population, which has a very high prevalence of obesity and type 2 diabetes, one higher than that reported for other Spanish populations, we can assume that the low lipid levels observed in this population were probably related to a high use of statins or other hypolipaemic treatments.
In conclusion, arterial factors in special hypertension and high BP levels are the main contributors to the development of RVO. Another important point is that RVO may be the first manifestation of an undiagnosed hypertension. Therefore, the early detection of hypertensive patients and good control of BP levels can help prevent the development of RVO. Moreover, the diagnosis of RVO in the context of hypertension raises the possibility of other target organ damage (TOD) such as carotid arteries, the heart and the kidney, given the large number of individuals with carotid plaques. If this is the case, these patients should be treated more aggressively in order to avoid the progression of TOD. Among the venous factors, only hyperhomocysteinaemia and the presence of aPLAs seem to play a role, so it would not be necessary to assess the inherited thrombophilia in those patients with new onset RVO.
Discussion
In this study, the characteristics of a series of patients with a first episode of RVO and the impact of potential associated risk factors were described. Subsequently, we compared the observed prevalence for these factors in the RVO population with three different general populations in the case of arterial risk factors, and with those for one cohort of young patients with DVT/PE or with those found in the literature in the case of venous factors. Among the arterial risk factors, hypertension and overall BP levels were the main associated factors, whereas only hyperhocysteinaemia and aPLAs were relevant among the venous risk factors for thrombosis.
Vascular diseases, especially coronary heart disease and stroke, are the leading causes of morbidity and mortality in industrialised countries. Although retinal vascular disease does not compromise lifespan, it can have a deleterious effect on the quality of life of those affected and, therefore, it is important to determine the main associated risk factors in order to prevent the disease. Both arterial and venous risk factors are thought to be involved in RVO, based on the fact that the retinal veins and arteries share a common adventitious.
In this study, there were more affected males than females (58% vs. 42%), with a mean age of 59 years. This indicates that it is not a disease only of the elderly people, but that it also affects active middle-aged subjects and, therefore, can have a significant impact on productivity. The CRVO or BRVO can, in the short term, lead to macular oedema with or without a decrease in perfusion, and in the long term, to neovascularisation, retinal deterioration, neovascular glaucoma or vitreous haemorrhaging, all of which are causes of vision loss.
The BRVO is three to five times more frequent than is CVRO, and it normally occurs as a result of compression at the crossroads junctions with an adjacent artery. A similar ratio was observed in this study and has been described in previous studies. The venous stasis caused by the increase in pressure or stiffness of the adjacent artery seems to be one of the main predisposing factors for RVO. This hypothesis easily explains why the central retinal vein is less affected, and this is related to its large calibre and location. This study did not have any bilateral occlusion that supports localised stasis problems.
The presence of ocular hypertension as a cause of RVO is discarded, since only a few patients (11) had high ocular pressure (≥ 21 mmHg) at the moment of the event. Even in those 15 previously diagnosed with ocular hypertension and under treatment, only five (33.3%) had high ocular pressure. From our data, it is possible to speculate that ocular hypertension could be particularly relevant in those subjects with very high ocular pressure levels in one eye, as has been shown by the fact that the affected eye was the unilateral one.
The presence of inherited thrombophilia in our series was very rare and similar to that observed in the general population. In fact, none of the patients suffered from protein C, S or antithrombin III deficiency, only seven had Factor V Leiden, and three the G20210A mutation of the prothrombin gene. These percentages are significantly lower than the ones observed in our cohort of young patients with DVT-PE or than the estimations for populations with DVT or PE in the literature. These results seem to be the same even for the young patients with RVO, although this subsample was quite small and still the differences in age as compared with the DVT group were highly significant. Therefore, our results confirm the findings of previous studies in which inherited thrombophilia had little influence on RVO, so its determination would not be justified a priori.
Hypercoagulability, as a result of immune-complexes against the membrane phospholipids or plasmatic proteins (LA, aPLAs) can also lead to RVO. This can be related to the inhibition of the production of prostacyclin by the vascular endothelium, the increase in thromboxane production by the platelet, the increase in the release of tissue factor and the presence of vascular endothelium adhesion molecules that interfere with the fibrinolytic process. In the present study, there were 13 patients with prevalence significantly higher than the one observed in general population and similar to that observed in the cohort of DVT/PE or the estimated one for thrombosis populations. Carbone et al. found that patients with aPLAs and retinal thrombosis had a greater frequency of antinuclear antibodies, low levels of factor 4 of the complement (C4) and the presence of circulating immune-complexes, something that could be the first manifestation of APS or a Systemic Lupus Erythematosus with severe repercussions in other target organs. Homocysteine, which can be considered both an arterial, as well as a venous risk factor, was elevated in 16 patients. Again, this prevalence is significantly higher than that reported for the general population, slightly higher than the estimated one for thrombosis populations and similar to the one observed in our group with DVT/PE. Hyperhomocysteinaemia produces a blockade of the union of the tissue plasminogen activator to the endothelial cell and, therefore, can block the release of thrombomodulin by the endothelial cell, activate coagulation factors V and VII and inhibit protein C. These findings, observed in our cohort, were confirmed by the meta-analysis performed by Janssen and colleagues in patients with RVO, including 11 case–control studies. The odds ratio was 8.9 (95% Cl 5.7–13.7) for hyperhomocysteinaemia and 3.9 (95% Cl 2.3–6.7) for anticardiolipin antibodies.
Hyperhomocysteinaemia has also been associated with BP in the literature, and therefore the effect of these two factors could be closely related. In this sense, we did not find differences in homocystein levels or for the prevalence of hyperhomocysteinaemia in the RVO group for those with or without hypertension (15.2 vs. 13.9 μmol, p-value 0.34 and 14.8 vs. 14.6 μmol, p-value 0.90, for normotensives and hypertensives before and after reclassification of the hypertension status; 19.6 vs. 12.7%, p-value 0.3 and 15.7 vs. 16.4%, p-value 0.9 for the prevalence of hyperhomocysteinaemia for normotensives and hypertensives before and after reclassification of the hypertension status).
Another important issue is that the D-dimer, which is of great value in the diagnosis of an acute episode of DVT/PE or in ischaemic stroke, is only elevated in a minority of patients with RVO (only 17 subjects had D-dimer ≥ 250 mg/dl). This parameter has been studied very little in relation to retinal vascular pathology, and therefore a more comprehensive evaluation would be necessary.
As an arterial risk factor, initially only 47 patients (47%) were diagnosed with hypertension. After reclassification, taking into account BP levels over 140/90, a significant increase in this prevalence was observed (from 47 to 80%). Although we cannot assume a correct diagnosis for all of them, surely among them are many with undiagnosed hypertension. This is important because RVO could be the first clinical manifestation of an essential hypertension in many patients. Moreover, the high systolic and diastolic BP levels seen in both groups, the supposed normotensives and the hypertensives, when compared with those in studies of the general population, indicate a lack of BP control. This is confirmed by the meta-analysis performed by O'Mahoney and collaborators, which included 2916 subjects with RVO. In this meta-analysis, the prevalence of hypertension was 63%, significantly higher than the prevalence observed in the comparative population (36.2%), with an OR for hypertension of 3.5 (95% CI 2.5–5.1). Taking into account our findings and the results of the meta-analysis, hypertension seems to be a key factor in the development of RVO. Probably, the increase in arterial stiffness as a result of high BP levels facilitates the venous stasis because the artery lies next to the vein in the retina.
For type 2 diabetes, the results were not as evident as for those with hypertension. We only observed a trend towards a higher prevalence, especially after reclassification, based on fasting glucose levels ≥ 126 mg/dl. Fasting glucose levels were also significantly higher in the RVO group when compared with two of the populations (Hortega and Segovia), and these differences were more evident in subjects without type 2 diabetes than in those with it. This may indicate that some kind of carbohydrate metabolism abnormality might also play a role in the development of RVO. In the O'Mahoney meta-analysis, of 2877 patients with RVO, 14.6% were diabetic as compared with 11.1% in the general population, which is in agreement with our results. In the O'Mahoney study, however fasting glucose was not evaluated, so some new type 2 diabetes diagnoses could have gone undetected.
The lipid abnormalities results were quite inconsistent. There could be a trend towards high LDL cholesterol in the RVO group, but only when compared with one of the populations; the same trend was observed for HDL cholesterol. These confusing results were also observed in the multicentre study of O'Mahoney et al. and could be related to the different criteria used in defining dyslipidaemia or the different thresholds for normal lipid levels. In this study, however, there was a noticeable twofold increase in the incidence of dyslipidaemia in patients with RVO as compared with the general population.
Finally, in 48 patients with RVO who underwent a carotid intima-media thickness evaluation, this study noted a high prevalence of arteriosclerotic lesions. This result reinforces the hypothesis of arterial risk factors as triggers of RVO. An association among ischaemic stroke, retinal thrombosis and carotid stenosis has been recently described.
Among the strengths of this study is the use of several populations for comparing the arterial factors. This minimises the possibility that the observed findings were by chance. Moreover, the comparison populations were matched by age, sex and BMI to reduce the impact of said variables on the prevalence of the main cardiovascular risk factors. In the case of the venous factors, we took advantage of one large series of young patients with DVT/PE who are perfectly characterised regarding the risk factors for venous thrombosis. Additionally, the observed prevalence of venous risk factors in the literature for both populations with DVT/PE and population-based studies was reviewed.
Among the main weaknesses is that which derives from the differences in the methodology of the different studies used for comparison. This is especially important in the case of dyslipidaemia, not only because of the different definitions of dyslipidaemia but also because of the lack of information. As examples, in the Segovia population, lipids data were missing in a large number of patients, and in the Pizarra one, we did not have a register of those subjects who were already dyslipidaemics or under hypolipaemic treatment. Taking into account the characteristics of the Pizarra population, which has a very high prevalence of obesity and type 2 diabetes, one higher than that reported for other Spanish populations, we can assume that the low lipid levels observed in this population were probably related to a high use of statins or other hypolipaemic treatments.
In conclusion, arterial factors in special hypertension and high BP levels are the main contributors to the development of RVO. Another important point is that RVO may be the first manifestation of an undiagnosed hypertension. Therefore, the early detection of hypertensive patients and good control of BP levels can help prevent the development of RVO. Moreover, the diagnosis of RVO in the context of hypertension raises the possibility of other target organ damage (TOD) such as carotid arteries, the heart and the kidney, given the large number of individuals with carotid plaques. If this is the case, these patients should be treated more aggressively in order to avoid the progression of TOD. Among the venous factors, only hyperhomocysteinaemia and the presence of aPLAs seem to play a role, so it would not be necessary to assess the inherited thrombophilia in those patients with new onset RVO.
Source...