【medical-news】眼睛：是系统健康的心灵之窗吗？
发布于 2009-02-26 · 浏览 1126 · IP 北京北京
这个帖子发布于 16 年零 32 天前，其中的信息可能已发生改变或有所发展。
The eye: window to the soul or a mirror of systemic health?
It has been said that there is no systemic disease which does not have a sign identifiable on ophthalmic examination. For many doctors, retinopathy—microaneurysms, haemorrhages, "hard" lipid exudates, microinfarcts of the retinal nerve fibre layer (cotton wool spots)—is synonymous with diabetes. A clinically distinct form of retinopathy is also recognised in hypertension, although traditionally, abnormalities of arteriovenous crossing architecture are the sine qua non feature of hypertensive retinopathy, with haemorrhagic and exudative changes only following later. However, population-based research has shown that retinopathy is a relatively common finding in older people without diabetes.1
In this edition of the journal, Liew and colleagues report increased coronary heart disease (CHD) mortality associated with retinopathy (see article on page 391).2 Retinopathy was present in 29% (57/199) of those with diabetes and 10% (268/2768) of those without diabetes. The presence of retinopathy in people without diabetes increased the risk of CHD death by the same magnitude as did a previous diagnosis of diabetes without retinopathy. Similarly, the presence of retinopathy in those with diabetes increased the rate of CHD mortality by an additional similar amount over and above that associated with diabetes itself. Adjustment for body mass index, fasting plasma glucose, oral hypoglycaemic drugs and serum creatinine level did not change the results appreciably. Associations were similar in men and women.
In view of previous publications, the findings of retinopathy in people without diabetes, and that retinopathy is a biomarker for vascular causes of mortality are not surprising. Glycated haemoglobin concentration has a monotonic relationship with cardiovascular and all-cause mortality. This relationship has apparently no threshold effect. Importantly, the measurable increase in risk of death occurs across physiological levels of HbA1c.3 A 1% increase in absolute concentrations of glycated haemoglobin is associated with about 10–20% increase in cardiovascular disease risk.4 These observations themselves raise questions about the traditional medical approach of allocating dichotomised diagnosis of disease versus no disease, and lend more weight to the use of continuous variable risk profiles. This concept is intrinsic to the principles of preventive medicine, where making small changes to lifestyle or healthcare activity to lower risk factor levels even modestly in the large majority of people who are at moderately increased risk makes a measureable reduction in the population burden of disease.5
In a previous issue of Heart, the same groups of authors reported findings of a morphometric study of retinal vascular calibre, suggesting that retinal vascular calibre predicted CHD death independently of traditional cardiovascular risk factors in men and in women aged 49–75 years, but not in older women.6 They suggested that their findings might help to explain the observation that the relationship between recognised adverse and protective factors in cardiovascular disease appears exaggerated in women, when compared with men.7 Thus, in addition to the potential for risk prediction, these morphometric assessments may also provide further insights into the aetiology and pathophysiology of cardiovascular disease.
These population-based studies are transforming our conceptual approach to retinal imaging. Clinical interest has hitherto largely focused on pathology in the retina as a manifestation of known systemic disease, such as diabetes or hypertension. However, it now appears that early changes in the retina and vasculature may be seen before established clinical disease, but reflecting pathophysiological processes which influence the risk of such later diseases. The estimated magnitudes of the independent contribution of these retinal changes to coronary heart disease risk prediction are substantial and comparable to established risk factors. This is not surprising since some of the vascular imaging is a direct indication of processes presumably occurring systemically. These findings, of course, need to be replicated in other populations and if they are, the potential for retinal imaging to contribute to the prediction of cardiovascular disease risk is huge.
Prospective measurement of the retinal arteriole:venule ratio (AVR) helps to predict incident hypertension,8 the incidence and progression of diabetes9 and the need for amputations in diabetes.10 These associations persist after adjustment for other risk factors such as age, hypertension, diabetes and smoking. Thus, it appears that measurement of the AVR is an important addition to risk profiling in major causes of morbidity and mortality. However, this technique is "technology dependent", requiring very high-quality, high-resolution digital retinal photographs taken through pharmacologically dilated pupils. Age-related cataract can significantly impair photographic quality, and therefore viability of AVR measurement. Specialised computer programs and experienced technicians are needed to grade the images—retinal vessel calibre cannot be reliably estimated in clinical practice using direct ophthalmoscopy.
A further limitation of retinal vascular morphometry results from interindividual variation in magnification effects of the optical system of the eye. Although algorithms exist for correcting these effects, accuracy depends on more detailed biometric characterisation of the individual eye, and ultimately, therefore, is more time and equipment dependent.11 Use of the arteriole:venule calibre ratio is an attractive, pragmatic way to overcome the problem of variation in ocular magnification. Nonetheless, the increasing body of evidence that points to retinal microangiopathy being a significant predictor of all-cause mortality is an important advance in our ability to assess risks to our patients. In an ideal world, such an assessment would be performed by a trained healthcare professional using a direct ophthalmoscope or other ophthalmic devices. Currently, a weakness in this conceptual model is the finding that 50% of retinopathy lesions are missed by ophthalmoscopy, compared with a reference standard of high-quality retinal photography.12
It seems likely that advances in the performance and availability of imaging technology will offer solutions to the apparent obstacles in translating these discoveries into routine clinical practice. Clearly, ophthalmic imaging offers unique opportunities to examine neural, vascular and connective tissue components of the central nervous system. Scanning laser ophthalmoscopy (SLO) to measure optic nerve head topography has been in widespread clinical use for diagnosis and monitoring of glaucoma, a chronic progressive optic neuropathy, for a decade.13 SLO imaging performs at least as well as human observers grading photographic images of the optic nerve.14 A competing technology, scanning laser polarimetry (SLP), allows quantitative assessment of the thickness of the retinal nerve fibre layer, and is also in widespread use in the management of glaucoma.15 Optical coherence tomography (OCT) was initially shown to be a method of in vitro imaging of the retina and coronary arteries,16 but rapidly became adapted for in vivo ophthalmic imaging, and is now widely used in diagnosis and management of retinal and optic nerve pathology, and as a tool for examination of the cornea and anterior segment of the eye. Commercial SLO and SLP used for glaucoma management incorporate semiautomated morphometry algorithms which make partial correction for ocular magnification (OCT does not). The measurements are compared against a normative database giving provisional diagnostic classifications.
Indeed, ophthalmic imaging has now developed to such a degree that the fine structure of retinal cells is now discernable in vivo.17 Obstacles remain to be overcome, including problems of movement artefact, difficulties with image registration in time series studies and the performance of both hardware and software in the acquisition, storage and automated grading of large numbers of high resolution. However, the inexorable logarithmic growth of computing power, following the predictions of Moore’s Law, suggests that these latter concerns will not remain a constraint for much longer. Indeed it seems likely that ophthalmic imaging has much more to offer in the study and clinical management of major systemic diseases.
Heart 2009;95:348-349
It has been said that there is no systemic disease which does not have a sign identifiable on ophthalmic examination. For many doctors, retinopathy—microaneurysms, haemorrhages, "hard" lipid exudates, microinfarcts of the retinal nerve fibre layer (cotton wool spots)—is synonymous with diabetes. A clinically distinct form of retinopathy is also recognised in hypertension, although traditionally, abnormalities of arteriovenous crossing architecture are the sine qua non feature of hypertensive retinopathy, with haemorrhagic and exudative changes only following later. However, population-based research has shown that retinopathy is a relatively common finding in older people without diabetes.1
In this edition of the journal, Liew and colleagues report increased coronary heart disease (CHD) mortality associated with retinopathy (see article on page 391).2 Retinopathy was present in 29% (57/199) of those with diabetes and 10% (268/2768) of those without diabetes. The presence of retinopathy in people without diabetes increased the risk of CHD death by the same magnitude as did a previous diagnosis of diabetes without retinopathy. Similarly, the presence of retinopathy in those with diabetes increased the rate of CHD mortality by an additional similar amount over and above that associated with diabetes itself. Adjustment for body mass index, fasting plasma glucose, oral hypoglycaemic drugs and serum creatinine level did not change the results appreciably. Associations were similar in men and women.
In view of previous publications, the findings of retinopathy in people without diabetes, and that retinopathy is a biomarker for vascular causes of mortality are not surprising. Glycated haemoglobin concentration has a monotonic relationship with cardiovascular and all-cause mortality. This relationship has apparently no threshold effect. Importantly, the measurable increase in risk of death occurs across physiological levels of HbA1c.3 A 1% increase in absolute concentrations of glycated haemoglobin is associated with about 10–20% increase in cardiovascular disease risk.4 These observations themselves raise questions about the traditional medical approach of allocating dichotomised diagnosis of disease versus no disease, and lend more weight to the use of continuous variable risk profiles. This concept is intrinsic to the principles of preventive medicine, where making small changes to lifestyle or healthcare activity to lower risk factor levels even modestly in the large majority of people who are at moderately increased risk makes a measureable reduction in the population burden of disease.5
In a previous issue of Heart, the same groups of authors reported findings of a morphometric study of retinal vascular calibre, suggesting that retinal vascular calibre predicted CHD death independently of traditional cardiovascular risk factors in men and in women aged 49–75 years, but not in older women.6 They suggested that their findings might help to explain the observation that the relationship between recognised adverse and protective factors in cardiovascular disease appears exaggerated in women, when compared with men.7 Thus, in addition to the potential for risk prediction, these morphometric assessments may also provide further insights into the aetiology and pathophysiology of cardiovascular disease.
These population-based studies are transforming our conceptual approach to retinal imaging. Clinical interest has hitherto largely focused on pathology in the retina as a manifestation of known systemic disease, such as diabetes or hypertension. However, it now appears that early changes in the retina and vasculature may be seen before established clinical disease, but reflecting pathophysiological processes which influence the risk of such later diseases. The estimated magnitudes of the independent contribution of these retinal changes to coronary heart disease risk prediction are substantial and comparable to established risk factors. This is not surprising since some of the vascular imaging is a direct indication of processes presumably occurring systemically. These findings, of course, need to be replicated in other populations and if they are, the potential for retinal imaging to contribute to the prediction of cardiovascular disease risk is huge.
Prospective measurement of the retinal arteriole:venule ratio (AVR) helps to predict incident hypertension,8 the incidence and progression of diabetes9 and the need for amputations in diabetes.10 These associations persist after adjustment for other risk factors such as age, hypertension, diabetes and smoking. Thus, it appears that measurement of the AVR is an important addition to risk profiling in major causes of morbidity and mortality. However, this technique is "technology dependent", requiring very high-quality, high-resolution digital retinal photographs taken through pharmacologically dilated pupils. Age-related cataract can significantly impair photographic quality, and therefore viability of AVR measurement. Specialised computer programs and experienced technicians are needed to grade the images—retinal vessel calibre cannot be reliably estimated in clinical practice using direct ophthalmoscopy.
A further limitation of retinal vascular morphometry results from interindividual variation in magnification effects of the optical system of the eye. Although algorithms exist for correcting these effects, accuracy depends on more detailed biometric characterisation of the individual eye, and ultimately, therefore, is more time and equipment dependent.11 Use of the arteriole:venule calibre ratio is an attractive, pragmatic way to overcome the problem of variation in ocular magnification. Nonetheless, the increasing body of evidence that points to retinal microangiopathy being a significant predictor of all-cause mortality is an important advance in our ability to assess risks to our patients. In an ideal world, such an assessment would be performed by a trained healthcare professional using a direct ophthalmoscope or other ophthalmic devices. Currently, a weakness in this conceptual model is the finding that 50% of retinopathy lesions are missed by ophthalmoscopy, compared with a reference standard of high-quality retinal photography.12
It seems likely that advances in the performance and availability of imaging technology will offer solutions to the apparent obstacles in translating these discoveries into routine clinical practice. Clearly, ophthalmic imaging offers unique opportunities to examine neural, vascular and connective tissue components of the central nervous system. Scanning laser ophthalmoscopy (SLO) to measure optic nerve head topography has been in widespread clinical use for diagnosis and monitoring of glaucoma, a chronic progressive optic neuropathy, for a decade.13 SLO imaging performs at least as well as human observers grading photographic images of the optic nerve.14 A competing technology, scanning laser polarimetry (SLP), allows quantitative assessment of the thickness of the retinal nerve fibre layer, and is also in widespread use in the management of glaucoma.15 Optical coherence tomography (OCT) was initially shown to be a method of in vitro imaging of the retina and coronary arteries,16 but rapidly became adapted for in vivo ophthalmic imaging, and is now widely used in diagnosis and management of retinal and optic nerve pathology, and as a tool for examination of the cornea and anterior segment of the eye. Commercial SLO and SLP used for glaucoma management incorporate semiautomated morphometry algorithms which make partial correction for ocular magnification (OCT does not). The measurements are compared against a normative database giving provisional diagnostic classifications.
Indeed, ophthalmic imaging has now developed to such a degree that the fine structure of retinal cells is now discernable in vivo.17 Obstacles remain to be overcome, including problems of movement artefact, difficulties with image registration in time series studies and the performance of both hardware and software in the acquisition, storage and automated grading of large numbers of high resolution. However, the inexorable logarithmic growth of computing power, following the predictions of Moore’s Law, suggests that these latter concerns will not remain a constraint for much longer. Indeed it seems likely that ophthalmic imaging has much more to offer in the study and clinical management of major systemic diseases.
Heart 2009;95:348-349
最后编辑于 2022-10-09 · 浏览 1126
3 收藏点赞
