Preventing Visual Loss From Chronic Eye Disease in Primary Care: Title and subTitle BreakScientific Review FREE

Visual disability is common in the United States and can have profound consequences for function and quality of life. Most causes of visual impairment are readily diagnosed and at least 40% of blindness and visual impairment is treatable or preventable.^{1 - 2} Nevertheless, many people living in the United States, especially elderly persons³ and minorities,^{4 - 5} do not receive necessary eye care.⁶ As a consequence many of these individuals develop visual disability or blindness needlessly.² The problem of undiagnosed visual disorders is growing, with the number of blind and visually impaired elderly individuals expected to double in the next 3 decades.¹

Clinicians in primary care settings play a critical role in reducing visual disability by managing systemic disease with ocular consequences and ensuring that patients receive timely specialty eye care. They may be the only health care professionals to recognize the need for an eye examination because of a new-onset visual disability. They also may be uniquely aware of risk factors, such as diabetes mellitus (DM) or a family history of glaucoma. Clinicians can educate patients about their need for eye care services and can advocate for their patients in obtaining access to such care. In the context of this article, we define a clinician as a primary care physician, nurse practitioner, or other mid-level practitioner who provides ongoing general medical care and serves as the primary contact between the patient and the health care system.

This review synthesizes current recommendations for primary care clinicians regarding preventing the most common causes of visual loss in adults and summarizes the evidence that supports these recommendations. The goal is to help clinicians develop an informed strategy for vision care in their clinical practice. This review does not deal with acute complaints, such as pain, redness, and new-onset blurred vision; these topics have been well covered in recent comprehensive reviews for the primary care clinician.^{7 - 9} Readers interested in the evaluation and management of red eye are also referred to standard texts and to an algorithm for triaging acute eye complaints available from the American Academy of Family Physicians (http://familydoctor.org/flowcharts/505.html).

We performed a systematic review of the literature regarding the diagnosis and management of vision impairment in adults and concentrated on those aspects of care that are within the domain of a typical primary care practice. A content expert (S.R.) worked with experts in systematic reviews (P.G.S. and C.H.M.). In general, our procedures followed those recommended by the Cochrane collaboration. With the assistance of a reference librarian, we electronically searched MEDLINE, EMBASE, The Cochrane Database of Systematic Reviews, and HealthSTAR between 1985 and 2003 using key words and free-text terms (eg, vision screening, glaucoma prevention and control) to identify potentially relevant studies. We identified additional citations through reference lists and expert consultation. In addition to these strategies, we sought relevant clinical practice guidelines by searching the Internet, including Web sites for clinical societies and the database maintained by the National Guidelines Clearinghouse (http://www.guidelines.gov). Searches were updated periodically throughout the preparation of the manuscript.

The content expert (S.R.) reviewed all citations with the inclusion criterion being that a study assessed the relationship between a specific health process and health outcomes in humans; studies were included regardless of the magnitude or direction of the reported effect. Those studies with the strongest possible research design for the question of interest (eg, the use of randomized controlled trials [RCTs] for questions of efficacy and prospective cohort studies for questions of risk and prognosis) earned the highest priority. If such studies were rare or absent, we reviewed articles using other study designs. We included prior systematic reviews and meta-analyses when relevant. The clinical practice guidelines yielded additional citations for original literature and the expert opinions summarized in the recommendations.

The major causes of visual disability among US adults are age-related or worsen with advanced age.¹ These causes include refractive error, cataract, diabetic retinopathy, glaucoma, and macular degeneration (Table 1).^{10 - 24} The incidence of blindness and vision impairment increases with age, increasing especially rapidly in adults older than 75 years.¹ A population-based study of US adults identified subjective functional visual impairment in up to 7% of people aged 71 to 74 years, rapidly increasing to 39% of those adults 90 years or older.^{25 - 27} Using measured visual acuity as an outcome, the Beaver Dam Eye Study²⁸ found visual impairment with worse than 20/40 visual acuity among 5% of individuals aged 65 to 74 years and 21% of individuals 75 years or older.

Control Glucose in Patients With DM. There is abundant evidence that patients with DM have improved visual outcomes when serum glucose is well controlled. Two RCTs and 2 decision analyses have addressed the effects of tight glycemic control in patients with type 2 DM, and 1 study has addressed this issue in patients with type 1 DM.

The Diabetes Control and Complications Trial established that tight glycemic control in patients with type 1 DM leads to significant reductions in the risk of early microvascular complications.²⁹ Another RCT, United Kingdom Prospective Diabetes Study 33 (UKPDS 33), extended these findings to patients with type 2 DM.³⁰ This RCT randomly assigned 3876 patients newly diagnosed with type 2 DM (median age, 54 years) to either tight glycemic control or to usual care. During 10 years of follow-up, hemoglobin (Hb)A_1C levels averaged 7.0% in the tight control group compared with 7.9% in the usual care group. The tight control group experienced a statistically significant 12% lower relative risk for any diabetes-related end point, which was mainly due to a reduction in the number of microvascular events (primarily photocoagulation for diabetic retinopathy).

The Kaplan-Meier method curves published with UKPDS 33 show that at least 2 to 3 years of tight glycemic control are needed before its benefits become apparent. However, this level of control was associated with an increase from 0.7% to about 1.5% in the annual incidence of major hypoglycemic episodes. These outcomes are in general agreement with a smaller RCT from Japan.³¹

The first of the decision analyses we identified used a Markov model to estimate the benefits of glycemic control on microvascular complications in type 2 DM.³² The results indicated that most of the benefit of glucose control was achieved by decreasing very elevated levels of HbA_1C to 9% and that comparatively little was achieved by further reducing the HbA_1C level from 9% to 7%.

The second decision analysis, using a different set of assumptions, concluded that tight control of noninsulin-dependent DM was far less cost-effective for patients who developed DM at age 75 years (>$200 000 per quality-adjusted life-year) compared with patients who developed DM before age 50 years ($20 000 per quality-adjusted life-year).³³ This model also concluded that the cost per quality-adjusted life-year increased greatly for HbA_1C values of less than 9%.

Very high levels of blood glucose can have immediate visual consequences. Fluctuations in blood glucose levels can be associated with dynamic shifts in the refractive power of the eye, resulting in blurry vision that is theoretically correctable with new lenses.^{34 - 36} Although a change in glasses prescription might temporarily restore clear vision, this approach can be impractical and expensive because the refractive power of the lens often is not stable during these episodes and may not return to baseline for weeks following normalization of serum glucose.³⁴ Thus, glycemic control remains the mainstay of therapy for these patients.

Control Hypertension.Hypertension in Patients With DM. In patients with type 2 DM, the risk of diabetic retinopathy is strongly associated with higher blood pressure.^{10 ,37} Recent data indicate that tight control of hypertension reduces the risk of diabetic retinopathy, as well as all other major diabetes outcomes. In UKPDS 38, 1148 patients with hypertension and type 2 DM (mean age, 56 years; mean blood pressure, 160/94 mm Hg) were randomly assigned to either tight control of blood pressure (<150/85 mm Hg) or less tight control (<180/105 mm Hg).³⁸ Despite relatively small differences in mean blood pressures (144/82 mm Hg vs 154/87 mm Hg), the outcomes between the 2 groups began to diverge between 2 and 3 years after initiation of therapy, and were pronounced by 5 years. After 9 years of follow-up, the treatment group had a 34% reduced risk of retinopathy worsened by 2 steps (worsening by 2 steps often heralds the need for laser photocoagulation treatment) (99% confidence interval [CI], 11%-50%; P<.001). Patients who were treated had a 47% reduced risk of significant deterioration in visual acuity (loss of 3 lines on a standard eye chart from 20/30 to 20/70) (95% CI, 7%-70%; P = .004).³⁸

The UKPDS 38 further showed that the greater the reduction in blood pressure, the greater the reduction in microvascular complications, such as retinopathy. In fact, for each 10-mm Hg decrease in mean systolic blood pressure, there was a 13% reduction in the risk of microvascular complications, such as diabetic retinopathy (95% CI, 10%-16%; P<.001).³⁷ Within this study, there was no threshold of systolic blood pressure below which this benefit began to wane.

These data suggest that the previously recommended blood pressure target of 140/90 mm Hg among middle-aged patients with type 2 DM may be inappropriately lenient. However, although evidence exists that lower blood pressure is better, the appropriate target for optimal blood pressure in older persons with DM has not been evaluated in randomized trials.

Hypertension in Other Patients. Control of hypertension can be expected to benefit the visual health of patients without DM as well, although this has not been studied directly in RCTs. Prolonged hypertension leads to decreased vision from hypertensive retinopathy^{39 - 40} and has been implicated in central retinal artery and vein occlusions,^{41 - 43} ischemic optic neuropathy,^{39 - 40 ,44 - 45} and macular degeneration.⁴⁶ The optimal threshold for blood pressure control in patients without DM also has not been established.

Control Hyperlipidemia, Especially in Patients With DM.Diabetes Mellitus. Although there is a lack of experimental evidence supporting lipid-lowering interventions aimed at diabetic vision loss, most observational studies suggest an association between elevated serum lipid levels and diabetic retinopathy.^{47 - 52} The Early Treatment Diabetic Retinopathy Study evaluated the relationship between elevated serum lipids and decreased visual acuity in 2709 patients at a baseline examination and after 5 years.⁵³ After controlling for age, baseline HbA_1C, and retinopathy severity, baseline serum cholesterol of more than 240 mg/dL (6.22 mmol/L) increased the risk of significant visual loss by 50% when compared with serum cholesterol of less than 200 mg/dL (5.18 mmol/L) (odds ratio [OR], 1.5; 95% CI, 1.1-2.1; P<.001). These results are in general agreement with the World Health Organization's multicenter study on the determinants of visual impairment in patients with DM.^{54 - 55} Nevertheless, other studies have failed to confirm this association.^{56 - 57} Randomized controlled trials are currently under way to assess the effect of lipid-lowering medications on the incidence and progression of diabetic retinopathy.⁴⁸

Age-Related Macular Degeneration. Observational case-control studies have suggested a link between self-reported dietary intake of certain fats (vegetable, monounsaturated, and polyunsaturated fats and linoleic acid) and increased risk for advanced age-related macular degeneration (AMD).⁵⁸ The Nurses' Health Study and the Health Professionals Follow-up Study showed similar findings for high intakes of fat, including linoleic acid.⁵⁹ Other studies show links between AMD and saturated fat and cholesterol.⁶⁰ However, these findings were not replicated in the Third National Health and Nutrition Examination Survey, which found no such association.⁶¹ One possible explanation for this disparity may be that the National Health and Nutrition Examination Survey defined AMD based on a single photograph through an undilated pupil as opposed to a more comprehensive examination as in the Age-Related Eye Disease Study.⁴² Because no cause and effect have yet been established for these links, expert guidelines have not included recommendations regarding dietary fat intake in patients with AMD.

Other Eye Diseases. Hyperlipidemia and lipid-related atherosclerotic disease have been implicated as risk factors in a variety of ophthalmic diseases, including retinal artery and vein occlusions, ischemic optic neuropathy, cataract, and even dry eye.^{41 ,62 - 68}

Advocate Smoking Cessation.Age-Related Macular Degeneration. Cigarette smoking is one of the few known modifiable risk factors for AMD,^{11 ,46} with a well-documented dose-response curve.⁶⁹ In the Beaver Dam Eye Study, women who smoked at the time of the study had an OR of AMD-related retinal findings (large soft drusen) of 2.20 (95% CI, 1.04-4.66) compared with women who never smoked.⁷⁰ The OR for male smokers was 3.21 (95% CI, 1.09-9.45).⁶⁹

Cataract. Associations between cataract formation and smoking appear in multiple observational studies.^{12 ,71 - 72} Evidence suggests that smoking cessation may reduce the likelihood of cataract formation.^{71 ,73} However, this benefit may not accrue, at least measurably, for many years following cessation.⁷²

Diabetic Retinopathy. Smoking is linked to incidence and progression of diabetic retinopathy in multiple studies.^{10 ,47} The effect may be dose-related.⁷⁴ Smoking cessation is generally encouraged by diabetic retinopathy experts.^{13 ,47 ,75}

Thyroid Eye Disease. Smoking significantly increases the risk for Graves ophthalmopathy^{76 - 80} and worsens the condition among those who already have it.⁷⁸ In 1 case-control study among 450 patients with Graves disease, the odds of developing ophthalmopathy was 7.7 (95% CI, 4.3-13.7) for smokers relative to nonsmokers.⁸¹ There is some evidence that these effects are longstanding and may be irreversible.⁷⁶

Assess the Ocular Effects of Systemic Medications. Many systemic medications are associated with ocular symptoms, complications, or both. Table 2 summarizes the most common culprits, including their signs and symptoms, and strategies to address each.

Consider the Role of Antioxidants and Protection Against UV Light. Patients who have been diagnosed by an eye care professional with specific types of macular degeneration may benefit from taking certain antioxidants. Recently released data from the Age-Related Eye Disease Study indicate that a daily high-dose antioxidant supplement reduces the chance of further vision loss for selected patients with macular degeneration.⁸⁴ In this RCT, 3640 patients who had intermediate macular degeneration in 1 or both eyes, and those with advanced AMD or AMD-related vision loss in 1 eye only, showed a 6% reduction in absolute risk of vision loss over 6.3 years (23% vs 29%; OR, 0.73; 99% CI, 0.54-0.99). Vision loss was defined as at least a doubling of the visual angle, equivalent to a change from 20/20 to 20/40 or worse, or a change from 20/50 to 20/100 or worse. Based on data in this study, we estimated that 11 patients would need to be treated for 5 years for 1 patient to benefit from high-dose antioxidants. No benefit was detected for patients with early AMD in this time frame, although the study was not powered to detect small benefits. The study dosages were vitamin C (500 mg), vitamin E (400 IU), beta carotene (15 mg), zinc (80 mg as zinc oxide), and copper (2 mg as cupric oxide). (Copper should be taken with zinc because high-dose zinc is associated with copper deficiency.⁸⁵ ) No statistically significant serious adverse effects from nutritional supplements were reported. However, the study noted that smokers and ex-smokers should not use beta carotene, because previous studies have suggested an association between beta carotene use and lung cancer in smokers.^{86 - 87} There are limited data to support or refute the use of antioxidants and to prevent or treat visual loss other than that from AMD as described above.

Protection against UV light has long been recommended by eye care professionals.¹⁶ Even low levels of exposure to UV light confer a measurable increase in the risk of certain types of cataract.^{88 - 91} A recent comprehensive literature review concluded that there is sufficient evidence of increased risk of cataract with exposure to UV light to justify public health messages advocating simple measures such as sunglasses to decrease ocular exposure.⁹² However, few data unequivocally support protection against UV light in preventing or treating visual loss other than cataract.

A systematic review performed for the American College of Physicians concluded that both screening for retinopathy and subsequent treatment are clearly beneficial for patients with DM.⁹³ Decision analytic models have concluded that screening for diabetic retinopathy is highly cost-effective and even cost-saving when payments for disability due to blindness are included, although the cost-effectiveness of screening decreases with increasing age.^{94 - 95}

There is widespread agreement among most major authorities that patients with adult-onset DM should undergo dilated retinal evaluation at least every 2 years and that patients with poorly controlled DM need yearly examinations.^{75 ,93 ,96 - 100} Most major authorities recommend annual retinal examinations for all patients with adult-onset DM regardless of how successful their disease has been controlled.^{75 ,93 ,96 - 100} Table 3 summarizes specific recommendations by major authorities.

Recommendations for annual examinations are based in part on the finding that 5% to 10% of patients with no retinopathy will progress to retinopathy within 1 year.^{104 - 106} However, some authorities believe that patients with well-controlled DM may not require retinal examinations as frequently. In the Wisconsin Epidemiologic Study of Diabetic Retinopathy,^{104 - 105} a large prospective study of 1990 predominantly white patients, the subset of patients with type 2 DM with no retinopathy on skilled reading of standard 7-field stereoscopic color fundus photographs, no gross proteinuria, and glycemic control within 2 standard deviations of the nondiabetic population, generally did not require follow-up retinal examinations for 4 years. These data suggest that for selected low-risk patients who have excellent glucose control (as measured by HbA_1C <8.0%), and healthy eye examinations (as determined by gold standard methods), the interval between retinal evaluations may extend to 2 years with little increased morbidity. A recent cost-utility analysis highlights the potential cost-effectiveness of this strategy⁹⁵ ; the National Committee for Quality Assurance and Veterans Affairs acknowledge these findings in their current diabetes quality measures for Health Employer Data and Information Set (HEDIS)¹⁰⁷ and Veterans Affairs performance measures. Nevertheless, many experts express concern that a 2-year target would increase the rate of needless blindness, because gold standard evaluations such as those performed by the Wisconsin Epidemiologic Study of Diabetic Retinopathy may not be representative of the screening techniques available to most patients, and because preventable blindness occurs due to poor compliance even when the goal is annual screening.⁹⁹

Most major guidelines concur that diabetic retinopathy screening should be performed by an eye care professional who is knowledgeable regarding diabetic retinopathy or by stereoscopic fundus photography.^{75 ,93 ,96 - 100} Studies have reported that either ophthalmologists or diabetologists are more sensitive at detecting early diabetic retinopathy than are general internists or family physicians.¹⁰⁸ Although 7-field dilated stereoscopic fundus photography remains the gold standard for detection of diabetic retinopathy, clinical examination by a skilled practitioner is more sensitive in detecting certain features of the disease.^{75 ,109}

There is evidence and professional consensus that a dilated pupil is necessary to ensure optimal examination of the retina.^{13 ,75} In 2 studies, retinal examination through undilated pupils failed to correctly classify the presence and severity of retinopathy 50% of the time compared with the standard 7-field stereo photographs.^{110 - 111} In contrast, examination through a dilated pupil correlated with photographs approximately 80% of the time.

There is increasing evidence that some systems of retinal photoscreening may have comparable sensitivity and specificity to dilated retinal examination in the detection of treatable diabetic eye disease,^{112 - 114} although experts have not reached consensus on this point.^{13 ,75}

We did not identify any published RCTs that assessed the effect of routine visual screening on visual outcomes, although a recent Cochrane Database of Systematic Reviews (updated in 2003) found no convincing evidence for community-based screening of older people by asking questions about subjective visual impairment.¹¹⁵ All 5 studies included in this review were multicomponent assessments, which used self-reported measures of visual impairment based on a limited number of questions, both as screening tools and as outcome measures. There were no qualifying studies that assessed the value of testing visual acuity, performing or referring for an eye examination, using a formally validated quality of life instrument as a screening tool, or combining several of these modalities. Hence, there are no RCTs that have evaluated routine visual screening as it would be performed as part of typical care by a primary care practitioner.

Although there is lack of evidence from RCTs supporting vision screening, nearly every major clinical authority recommends a periodic vision evaluation of asymptomatic adults, particularly in older populations (Table 4).^{102 ,116 - 122} However, there is no consensus as to who should perform the evaluation or its optimal content or timing. Most recommendations reflect consensus of expert opinion rather than evidence-based analysis. Some authorities simply recommend periodic screening tests such as visual acuity, which can be performed by primary care practitioners or other health care clinicians.^{102 ,116 - 117 ,120} Other authorities, recognizing that specialized ophthalmic testing is essential to detect most major causes of visual loss, such as glaucoma, macular degeneration, and diabetic retinopathy, recommend comprehensive examinations by eye care specialists, with interim screening tests as needed.^{118 - 119 ,121 - 122} Depending on the recommendation, primary care clinicians have several options: vision screening older patients themselves during the periodic health examination, referring older patients for annual complete eye examinations, or combining these strategies.

Most guidelines recommend increasingly frequent evaluations with advancing age. The rationale for a shorter interval is that increasing age is associated with a higher incidence of eye disease.² Recommendations for persons with DM, persons with known eye disease, and persons at risk for glaucoma are more stringent than for low-risk patients (Table 1, Table 3, and Table 4).

Visual impairment is common and debilitating, especially in elderly persons. Many patients suffer needlessly from preventable or treatable visual disorders. Primary care physicians have unique opportunities to help prevent visual impairment and blindness through patient education, medical therapy, and specialty referral. They can play a critical role by educating patients about the importance of treatment and prevention of eye diseases, by optimizing systemic treatment for illnesses, such as diabetes and hypertension, and by recognizing the need for specialty referral when there is visual impairment or risk factors for common eye diseases.

The value of routine screening for vision impairments has yet to be proven in clinical trials. Nevertheless, a strong indirect argument can be made favoring periodic vision evaluations, especially in elderly persons. Although some version of periodic vision evaluation is recommended by all major clinical authorities, the appropriate role for vision screening in the primary care setting has not been rigorously assessed. Visual acuity measurement is a component of every major vision screening guideline and can be performed in the primary care setting. However, many causes of vision impairment can be diagnosed only by persons with specialty ophthalmic knowledge and equipment, and most treatments require specialty skills to deliver. Thus, effective collaboration between primary care clinicians and eye care professionals has the greatest potential to improve eye care and enhance patients' vision and quality of life.