Weight Reduction as an Adjunctive Management Strategy for Diabetic Retinopathy

Abstract

As rates of global obesity and diabetes increase, diabetic retinopathy continues to grow as a frequent cause of visual impairment. Despite tremendous recent strides in therapy, a significant fraction of patients remain poorly responsive to modern interventions. Adjunctive therapy in such settings could be widely beneficial. A growing body of evidence suggests that weight reduction strategies for obesity-related diabetes have the potential to serve as important supplements to modern ophthalmic care for preservation of vision.

Introduction

Diabetic retinopathy (DR), among the most frequent causes of vision loss in working-age adults, affects about one-third of all patients with diabetes mellitus (DM). Incidence rates of DR and type 2 DM are rising, likely stemming from rising global rates of over-nutrition and obesity.¹

Although DR was historically associated with relatively poor visual prognosis, a major development in therapy came about with retinal photocoagulation for diabetic macular edema (DME) and proliferative diabetic retinopathy (PDR).^2,3 An even greater transformation in DR therapy came with the advent of ophthalmic pharmaceuticals that antagonize vascular endothelial growth factor (VEGF), the primary molecular culprit in clinically-manifested DR. Many such agents, including bevacizumab, ranibizumab and aflibercept, not only prevent vision loss among patients with DR, but often provide visual improvement. Despite such efficacy, these therapies – and other contemporary treatments such as ophthalmic corticosteroid therapy – are ineffective in about a third of patients, highlighting remaining needs for improvement of DR care.⁴

In this regard, management of DR requires management of the underlying systemic illness. The benefits of glucose, lipid, and blood pressure control are well-established. ⁵ In our practice, we also advocate the use of fenofibrate given its efficacy in preventing DME and PDR.^6,7 In type 2 DM, strategies for weight loss can also be powerful, as even modest reductions in obesity can have tremendous beneficial metabolic effects.⁸ The effects of these approaches on DR are incompletely studied, falling far short of the evidence needed to make broad recommendations for patient care – but they deserve attention and further study. Here, I review a case study from my practice along with selected preclinical studies and small clinical trials that offer some insight into the effects of weight reduction strategies on DR. Moreover, I also briefly review the evidence for the effects of glucagon-like peptide 1 receptor agonists (GLP1-RAs) on DR, since these drugs have received wide acclaim for their beneficial effects on type 2 DM control and obesity reduction while gaining notoriety for risk of promoting vision loss.

Case Example

A 56-year-old Caucasian man with a 15-year history of type 2 DM was referred to me for severe vision loss in both eyes. He was taking metformin and glyburide, achieving moderate glycemic control on that regimen with a glycated hemoglobin index (HbA1c) of 7.8%. Like many patients with type 2 DM, he had elevated blood pressure and dyslipidemia, and was being treated for both conditions.

His initial eyesight was 20/100 in the right eye and 20/150 in the left eye. Retinal examination, including the use of optical coherence tomography (OCT), indicated thick, center-involving DME in the setting of severe non-proliferative DR in both eyes. A notable and worrisome finding was the presence of thick subfoveal hard exudates especially in the left eye, as shown on his photos and initial OCT scan (Figures 1A, B, G, H). This finding is typically associated with poor visual outcomes, as shown in the Early Treatment for Diabetic Retinopathy Study (ETDRS).⁹

Figure 1.

Impact of an intermittent fasting (IF) diet on treatment-resistant diabetic macular edema. Retinal photographs from the right (A, C, E) and left (B, D, F) eyes of a 56-year-man with type 2 DM and vision loss. Baseline images (A, B) show center-involving macular edema, delineated by the presence of yellow, hard exudates and scattered intraretinal hemorrhages. After beginning IF, his disease severity lessened progressively with reduced hemorrhages and exudates at six months-post diet (C, D) and nearly resolved lesions by one year-post diet (E, F). His central macular thickness, shown by optical coherence tomography (G – right eye, H – left eye), was poorly responsive to ocular pharmacotherapy (I, J, K, L), but gradually improved three months (M, N) and six months after IF (O, P). By one year after IF, he had near-normal retinal thicknesses in both eyes and vastly improved visual acuity (Q, R).

After considering several options, we agreed to start with intravitreous injections of 0.3 mg ranibizumab to both eyes. Despite injections spaced four weeks apart, his DME was not reduced and his hard exudates worsened. More importantly, his vision remained unimproved. We added fenofibrate 160 mg daily, as the ACCORD and FIELD trials showed an approximately 30% risk reduction in need for DME therapy using this drug.^6,7 We also switched his treatment to 2 mg aflibercept monthly, since this drug has good visual outcomes in patients with poor baseline vision.¹⁰ Yet, his vision and retinal exam only nominally improved (Figures 1 I, J). We then switched to corticosteroid injections – in the form of bolus triamcinolone (4 mg) and fluocinolone implant – but these therapies also were not successful, and his vision deficits remained (Figure 1 K, L). He preferred not to undergo vitrectomy, though this could have been helpful, and instead opted to defer all further treatment.

This gentleman returned several months later, at which point I expected to see fully blossomed bouquets of macular edema involving both eyes. Instead, he had – for the first time in years of therapy – a significant reduction in macular thickness (Figure 1 M, N). Noticing that he appeared a lot thinner, I inquired about his general health. He had begun an intermittent fasting regimen on his own accord, resulting in a ~30% loss of body weight and lowering of body mass index (BMI) from 32 to 27 over a span of only several weeks. More importantly, he found the regimen much easier to adhere to compared to other dietary modifications he attempted in the past. His HbA1c improved to 6.5%, and he managed to come off of the sulfonylurea. Furthermore, he noticed improvement in vision. Indeed, on exam he improved to 20/50 in the right eye and 20/100 in the left eye. Remarkably, his hard exudates were also reduced in size (Figures 1 C, D). Given his recovery without ophthalmic treatment, we deferred further therapy, and monitored his progress every three months. From the moment he began intermittent fasting – a regimen that he continued to maintain – his macular thicknesses continued to decline and his vision gradually kept improving. A year later, he was seeing 20/25 in the right eye and 20/40 in the left eye. He has a residual hard exudate in the left eye, much smaller than on the initial examination, and minimal residual exudate in the right eye (Figures 1 E, F, Q, R). He had received no ophthalmic intervention since beginning intermittent fasting, is no longer receiving any pharmacologic therapy for type 2 DM, and has a HbA1c of 5.7%.

Of note, the initial gain in eyesight that he achieved with anti-VEGF therapy (20/70, improved from 20/100 in the right eye) would likely have been considered a treatment success (i.e., 15-letter improvement or more) in trials such as RISE/RIDE, VISTA/VIVID and DRCR Protocol I.¹¹ Yet, visual acuity of 20/70 in the better eye is far less than ideal for many individuals, including this particular patient who was actively employed as a salesperson and hoped to resume independent driving. Moreover, diabetes-induced deficits of many other aspects of vision that may not be routinely measured in clinical setting or in studies – such as contrast sensitivity, color discrimination, and dark adaptation – often have even poorer responses to anti-VEGF treatment. This gentleman’s case illustrates the remaining limitations of currently-available ophthalmic pharmaceuticals for the treatment of DME and highlights the need for adjunctive interventions. While his experience is an isolated example, I have cared for several other patients with similar success stories using intermittent fasting, restrictive caloric intake, bariatric surgery, or rigorous exercise regimens in the management of type 2 DM who improved their eyesight without ophthalmic intervention. Combined, these anecdotes offer a glimpse into the potential for nutritional and lifestyle-associated adjustments in treating diabetes, including those that ophthalmologists could endorse. Time and again, surveys among patients with diabetes show that the fear of losing vision provides a powerful motive to improve their metabolic control, putting the ophthalmologist in a key position to participate in the holistic care of people suffering from diabetes while simultaneously helping them preserve a cherished form of sensation.

Body Mass and Diabetic Retinopathy

The strongest predictors of DR risk include metrics of plasma glucose concentration and duration of exposure to diabetes.¹² Less consistent are associations between obesity indicators and risk of DR, with some studies showing positive relationships, others showing inverse relationships, and several failing to find any association at all.

One confounding factor accounting for these inconsistencies are varying definitions of obesity. The most common definition of obesity relies on the BMI. According to the World Health Association, a BMI of >30 kg/m² is considered obese, but this cutoff value may not be appropriate for persons from different parts of the world.¹³ In certain individuals, for example those from India, China, and Japan, metabolic disease manifests at lower BMI levels. In such groups, better assessments of obesity could include waist circumference (WC) and the waist-to-hip ratio (WHR), since these measurements better reflect abdominal or visceral obesity – a form of excess body weight that is linked to disease. High WC, for example, is one of the five diagnostic criteria for the metabolic syndrome, and is a good predictor of cardiovascular risk among South Asians.¹⁴ Few studies have carefully studied the association of WHR or WC as independent risk factors for DM, but those that have suggest a positive risk relationship.

Second, obesity may have different effects on different types of DM. At least one study suggests that type 2 DM could be classified into three subdivisions.¹⁵ Two of these subdivisions are associated with high BMI, but only one has elevated insulin resistance. A third type involves normal BMI and mild progressive insulin resistance with advancing age. The relative risk of DR between these various subtypes remains incompletely described and inclusion of all such patients into obesity-DR risk studies could mask important effects of obesity in a particular subtype.

Third, few studies have systemically assessed the effects of weight reduction as an intervention on DR risk. Nonetheless, the benefits of weight reduction for cardiovascular health in DM are high and therefore should be advocated for general health. Effects may vary depending on the form of weight reduction used, but here we will review some of the available evidence for these various methods.

Caloric Restriction

Intermittent Fasting

Weight reduction by intermittent fasting has becoming an increasingly popular eating program but was originally described as early as the 1930s.¹⁶ Multiple regimens exist, although a commonly-used one involves time-restricted feeding intervals in which the majority of daily calories are consumed within a several hour period each day (e.g., 16-hour fast, 8-hour feed; or 20-hour fast, 4-hour feed patterns).¹⁷ No cap is placed on the amount of calories consumed during feeding periods, nor are there restrictions on the types of allowable foods in the diet – features that many people find appealing compared to other diets.

Intermittent fasting schemes have demonstrated benefits on cardiometabolic health in clinic settings and on longevity in animal studies.¹⁸ However, the evidence for a specific beneficial effect on DR is limited. Lifestyle changes in general are very likely to prevent the incidence of DR in at-risk groups. In one study that randomized patients with pre-diabetes to a multi-pronged four-year lifestyle intervention regime consisting of 5% weight reduction, caloric restriction, and exercise, the five-year incidence of retinal microaneurysms was approximately 40% lower than in a control group that was not subjected to such intervention.¹⁹ However, no clinical trial to date has specifically tested effects of intermittent fasting regimens on DR incidence or severity.

Preclinical evidence shows that intermittent fasting robustly preserves vision in experimental diabetes and this protective effect may occur through both retina-intrinsic and non-intrinsic mechanisms. In a common mouse model of type 2 DM associated with hyperphagia and obesity, forced intermittent fasting prolonged overall lifespan, reduced retinal capillary damage and lessened leukocytic infiltration into the retina compared to control animals on a standard feeding regimen.²⁰ An independent study using the same mouse model, showed that intermittent fasting increased the local neuroretinal activity of a protective molecule known as Sirtuin1 – belonging to a class of molecules that have long been implicated in mediating the protective effects of caloric restriction on mortality.²¹

Carbohydrate Reduction

Other popular modern dietary regimens include the Paleolithic diet and ketogenic diets. The Paleolithic diet is based on the premise that feeding patterns among human ancestors prior to modern agricultural practices are more concordant with our inherited metabolic programs. It stresses the avoidance of grains, dairy, legumes, refined sugars and starches, as well as highly-processed foods. Ketogenic diets (similar to the Atkins diet that gained popularity in the 1980s) emphasize reduction in daily carbohydrate intake, with many plans aiming for 7% or less daily caloric intake from carbohydrate. Essentially, Paleolithic and ketogenic diets are carbohydrate-restrictive and both promote metabolic switching of fuel from liver-derived glucose to adipose-derived ketone bodies, thereby fostering reduction of adiposity. Although sparse, some compelling data exist to support the use of these regimens for metabolic control in DM, with effects on weight loss, improvement in insulin resistance, lowering of blood pressure, and improved serum lipid profiles. There are no available data on their efficacy in DR, although in a case report, a switch to ketogenic diet was associated with spontaneous regression of proliferative DR with traction retinal detachment – a finding that may have some merit since this stage of disease is rarely reversible without intensive ophthalmic intervention.²²

Bariatric Surgery

Surgery for weight loss, encompassing many procedures such as gastric bypass and gastric band placement, have immense beneficial effects on glycemic control in patients with type 2 DM. Their reported effects on DR are highly mixed, with some studies reporting worsening of DR after surgery.²³ One reason for this observation is the early-worsening phenotype due to rapid correction of glucose following surgery, similar to the phenotype seen in the Diabetes Control and Complications Trial (DCCT) where rapid diabetes control using intensive insulin therapy was associated with transient increases in DR severity. In studies with one year or less follow-up, significant proportions of patients had worsening DR severity.^24,25 Considering studies with >5 years follow up only, far fewer instances of DR worsening are seen.^26,27 Therefore, early worsening of DR severity after sudden correction of glycemia with bariatric surgery could eventually normalize without loss of vision, as was observed in DCCT. A major caveat is that all available studies on bariatric surgery and DR risk are vastly underpowered to make any definitive conclusions.

Exercise

Our bodies evolved to accommodate high levels of physical activity to provide for many basic needs of life, including finding food, procreating, and fleeing from or fighting against threats. As the conveniences of modern society have reduced the need for us to move in order to achieve these goals, our resulting sedentary lifestyles are poorly matched to suit our inherited metabolic programs. Indeed, these lifestyles are a major contributor to common modern diseases, including DM. Exercise – perhaps by restoring the balance between our fuel availability and expenditure – has numerous benefits across multiple axes of health, in patients with or without DM. Potential mechanisms include improved insulin sensitivity, reduced inflammation, and weight loss itself.²⁸

Clinical studies examining the relationship of exercise on DR have produced inconsistent results. In WESDR, DR risk was lower in physically-active women with early-onset DM (<30 years old at diagnosis) compared to those who were sedentary.²⁹ In several cross-sectional and smaller trials, a protective relationship between exercise and DR severity was recapitulated in people with type 2 DM.^30–32 However, both the Pittsburgh study of Physical Activity and Diabetes Complications and the whole-cohort analysis of the WESDR (including those with older onset DM) failed to show any association between exercise and DR risk.^33,34 Finally, although some authors historically advocated for avoidance of exercise in patients with advanced stage DR, such as proliferative disease, due to potential risks of vitreous or retinal hemorrhage with vigorous activity, the real risks of these events are negligible.^35,36

At least one potential confounder with many of these studies is that change in body weight was not taken into consideration, either because of the nature of the study design (cross-sectional) or because it was not chosen as an outcome measure. Therefore, it remains unclear whether weight loss from exercise itself has an independent impact on DR risk.

Glucagon-Like Peptide-1 Receptor Agonists

Glucagon-Like Peptide-1 receptor agonists (GLP1-RAs) are a novel class of antihyperglycemic agents used in the management of type 2 DM. Given their impressive effects on glycemic control, they have emerged as a valuable second-line tool in the care of patients with refractory DM. An added benefit to these drugs is their tendency to promote weight loss. The effect was confirmed in a highly-touted clinical trial, with recent approval of one agent for obesity management.³⁷

Despite such enthusiasm, fears remain concerning the safety of these agents – specifically in regard to vision. On such GLP1-RA – semaglutide – was associated with worsening DR, with a 76% increase in severe DR complications compared to placebo.³⁸ These results led to black box warnings, causing worry among patients with DR and their non-ophthalmologist care providers. However, in a subsequent study, escalating doses were not associated with increased DR adverse events.³⁹ An important limitation is that many GLP1-RA studies recorded DR outcomes based on adverse event reporting rather than dedicated retinal imaging. Reassuringly, a randomized trial that primarily aimed to examine the safety of liraglutide on retinal health failed to reproduce any significant adverse effects in terms of DR.⁴⁰

Based on the combined data, the evidence supports the use of GLP1-RAs for the treatment of type 2 DM with low concerns for detrimental effects on vision. On the contrary, it remains possible that the dual effects of these drugs on glycemic control and weight reduction could impact long-term retinal health positively. Preclinical evidence shows that GLP1 receptors are located within the neural retina and engage neuroprotective programs when activated. Although clinical evidence for an independent and bona fide protective effect on the retina is lacking, I have been an ardent supporter of the use of GLP1-RAs among my own type 2 DM patients with manifest DR who have poor control with metformin alone.

Conclusions

Modern therapy for DR has evolved greatly over the past 20 years but remains inadequately effective for a significant proportion of patients. As outlined in the case above, control of the underlying diabetes can be far more effective for DR than aggressive ophthalmic therapy. While glycemic control can be difficult to achieve in many patients, weight loss strategies could offer powerful tools for metabolic control in a subset of them. Although it is likely that these strategies – including intermittent fasting, carbohydrate reduction, bariatric surgery, exercise, or GLP1-RA treatment – have benefits on ocular health due to improvement in glycemic control, they may have independent effects on the retina – a concept that deserves further attention and investigation.

Funding Statement

This work was supported by an unrestricted grant to the John F. Hardesty MD Department of Ophthalmology and Visual Sciences from Research to Prevent Blindness.