Impact of insulin treatment in diabetic macular edema therapy in type 2 diabetes

Objective

To evaluate the impact of insulin therapy on the outcomes of diabetic macular edema (DME) treatment with vascular endothelial growth factor (VEGF) inhibitors in type 2 diabetics.

Methods

A retrospective consecutive case series of 95 type 2 diabetic patients with DME treated with anti-VEGF therapy. Two cohorts were examined—patients taking only oral anti-diabetic agents and patients on insulin therapy. The main outcome measures were change in visual acuity (VA) and change in central subfield macular thickness (CST) measured by spectral-domain optical coherence tomography (SD-OCT). Additional variables analyzed included HbA1c, creatinine, blood pressure and body mass index and their correlation with clinical findings.

Results

Both groups had a statistical significant improvement in VA (insulin therapy group: 20/61 to 20/49, p=0.003; oral anti-diabetic agents group: 20/76 to 20/56, p=0.005). There was no difference between groups at initial or 12 month examination (p=0.239 and p=0.489, respectively). From an anatomic standpoint, CST also improved significantly in both groups [454.7 μm to 354.9 μm (p<0.001) in the oral anti-diabetic agents group and 471.5 μm to 368.4 μm (p<0.001) in the insulin therapy group]. Again, there was no significant difference between groups at initial or 12 month follow-up examination (p= 0.586 and p=0.591, respectively). Mean HBA1c levels remained relatively stable during the follow-up in both groups.

Conclusion

Anti-VEGF therapy is a useful treatment for DME. This study suggests that chronic insulin therapy, compared to oral anti-diabetic agents, does not modify the anatomic or functional effectiveness of DME treatment.

INTRODUCTION

Diabetes mellitus is a major public health problem affecting approximately 285 million of people worldwide in 2010.¹ The problem is only increasing with data from the Framingham Heart Study also indicating that the incidence of type 2 diabetes has doubled over the last 30 years.² Diabetic retinopathy (DR) and diabetic macular edema (DME) are among the most significant and disabling chronic complications of diabetes mellitus. DME is an important cause of severe vision loss in type 2 diabetes. Hyperpermeability of retinal blood vessels and subsequent formation of edema and hard exudates are the key clinical features. As established in several clinical trials,³–5 strict metabolic control still remains the standard care for prevention of DR, which in many cases is only achieved by intense insulin therapy. Regarding DME treatment, anti-vascular endothelial growth factors (anti-VEGF) intravitreal injections with or without laser photocoagulation has become the gold-standard for reducing macular edema and improving visual acuity.⁶, 7

Previous studies have demonstrated that insulin treatment of diabetes may result in increased retinal vascular permeability and then induce diabetic retinopathy progression and visual impairment.⁸–11 These possible alterations are similar to those changes seen in diabetic retinopathy with increased VEGF load and may have an impact on the responsiveness of an eye to anti-VEGF therapy. The aim of the present study was, therefore, to assess whether chronic insulin therapy impacts DME response to anti-VEGF therapy.

METHODS

Study population

After Cleveland Clinic Institutional Review Board approval was obtained, a retrospective consecutive case series was performed in patients with type 2 diabetes and DME treated with intravitreal anti-VEGF injections from January 2010 to January 2013 at the Cole Eye Institute. The tenets of the Declaration of Helsinki were followed. Patients were included in the study if they met the following inclusion criteria: type 2 diabetes, center-involving DME as defined below, treatment with any intravitreal anti-VEGF therapy, baseline visual acuity less than or equal to 20/25, and SDOCT evaluation at all follow up time points (Cirrus^{, Carl Zeiss Meditec). Exclusion criteria included: type 1 diabetes (defined as juvenile-onset diabetes due loss of insulin production), intraocular surgery within 3 months of initiation of anti-VEGF therapy, intravitreal steroid injection within 3 months of initiation of anti-VEGF therapy, presence of significant media opacity that would limit vision recovery (e.g., significant cataract, vitreous hemorrhage, corneal scar), presence of co-existing macular disease (e.g., age-related macular degeneration, vascular occlusive disease), vitreomacular traction as determined by SD-OCT, macular ischemia if noted by the treating physician based on fluorescein angiography, previous vitreoretinal surgery (e.g., vitrectomy), and less than 1 year follow-up from initial injection. Subjects were divided in two groups: those taking oral anti-diabetic agents and those on insulin therapy. Patients from both groups were at baseline matched according to HbA1c levels to avoid glycemic control interference. If both eyes had treated DME, one eye of each patient was randomly selected for study inclusion.}

Eye-specific clinical variables were collected and analyzed including Snellen visual acuity (VA), intraocular pressure, ophthalmic examination features, and post-injection complications. Snellen VA was converted to LogMar notation for statistical analysis.

Assessment of metabolic parameters

Metabolic parameters were assessed through review of the comprehensive medical records including blood pressure (mm of Hg), body mass index (BMI), and levels of HbA1c, serum glucose, and creatinine. All values aforementioned were assessed both at baseline and at final follow-up. In addition, confounding medical diagnoses, medications, gender, age, duration of DM (years) were also recorded. Glomerular filtration rate was calculated based on appropriate parameters.

Assessment and treatment of DME

Center-involving DME was considered to be present based on the presence of foveal intraretinal fluid SD-OCT in association with the clinical diagnosis of diabetic retinopathy and concurrent appropriate fluorescein angiographic findings (e.g., leaking microaneurysms; if performed). At each visit, SD-OCT (Cirrus^{, Carl Zeiss Meditec) data was collected to assess the central subfield macular thickness (CST).}

All patients underwent intravitreal anti-VEGF therapy for DME with bevacizumab (1.25 mg). Patients were treated with a pro re nata (PRN) protocol with intravitreal bevacizumab given for center-involving DME. Follow-up periods were typically every 4–8 weeks. However, given the retrospective nature of this study, the specific treatment regimen was ultimately at the treating physicians’ discretion. Panretinal photocoagulation was performed in accordance with ETDRS guidelines for high-risk proliferative diabetic retinopathy and focal/grid laser photocoagulation was allowed during the treatment period at the treating physician’s discretion.¹², 13 The main outcomes assessed were mean changes in VA and CST in order to assess the impact of insulin therapy on DME treatment.

Statistical Analysis

Data processing and analysis were performed with SPSS version 20.0 software (SPSS, Inc, Chicago, IL). Patients were separated into two cohorts to assess the influence of diabetes therapy. In addition, these groups were also separated into two sub-cohorts according to HbA1c levels to assess the impact of glycemic control: serum HbA1c values ≤ 7.0% (acceptable control) and serum HbA1c values > 7.0% (poor control). The normal distribution of the variables was verified with Kolmogorov-Smirnov test. Comparisons between groups and primary outcome variables were performed using paired t-test for continuous variables. Categorical variables were expressed in proportions and analyzed by Pearson’s χ ^{test. For all statistical tests p<0.05 was considered statistically significant after Bonferroni correction to adjust for multiple comparisons. Data were expressed as mean ± SD.}

RESULTS

Demographics and Systemic Metabolic Parameters

A total of 95 eyes of 95 patients with type 2 DM were included in the study. Fifty-seven (60%) were male and 38 (40%) were female with a mean age of 65.8 years (range from 38 to 88 years). Of all patients, 49 (51.7%) were on insulin therapy (31 on insulin therapy only, 11 on insulin therapy plus metformin, and 7 on insulin therapy plus sitagliptin/metformin) and 46 (48.3%) were on oral anti-diabetic agents only. In the insulin group, the length of time since initiating the therapy was 6.4 ± 6.7 years (1 to 26 years). At baseline examination, the mean HbA1c of all enrolled patients was 7.2% (range from 5.2 to 13.5), the mean creatinine was 1.4 mg/dL (range from 0.4 to 5.9 mg/dL), the mean BMI was 30.9 Kg/m ^{(range from 20.8 to 60.8 Kg/m) and the mean of systolic blood pressure (SBP) was 136.1 mm of Hg (range from 106 to 203 mm of Hg) and of diastolic blood pressure (DBP) was 74.5 mm of Hg (range from 50 to 100 mm Hg). Eight-four (88.4%) were taking anti-hypertensive medications. Mean duration of diabetes was 14.3 years. The mean number of intravitreal injections was 5.9 ± 2.7 and the mean follow-up was 13.1 months.}

At baseline, the mean HbA1c level was 7.2% ± 1.62 in the oral anti-diabetic agents group and 7.3% ± 0.96 in the insulin therapy group (p = 0.774). The mean duration of DM was 12.4 ± 8.6 years and 15.9 ± 8.3 years in the oral anti-diabetic agents group and insulin therapy group, respectively. The mean creatinine value was 1.1 ± 0.72 mg/dl in the oral anti-diabetic agents group and was 1.63 ± 1.29 mg/dl in the insulin therapy group. There were no statistically significant differences between the two groups at baseline. Epidemiologic and baseline clinical findings of patients divided in oral anti-diabetics medication group and insulin therapy group are summarized and compared in Table 1.

At final follow-up examination, most metabolic parameters were stable. The mean HbA1c was 6.83 (range from 5.5 to 8.4), mean creatinine was 1.6 mg/dL (range from 0.6 to 11.4 mg/dL), mean BMI was 30.1 Kg/m ^{(range from 19 to 54.8 Kg/m), mean SBP was 134.7 mm of Hg (range 103–182 mm of Hg), and mean DBP was 72.0 mm of Hg (range 40–98 mm of Hg). At final follow up, the HbA1c was 6.7% in the oral anti-diabetic agents group and 6.9% in the insulin therapy group (p=0.328). All systemic parameters did not reach a statistically significant difference between groups either at the baseline or final examination. The number of anti-VEGF intravitreal injections was the same in both groups (5.9 ± 3.4 in the insulin therapy group and 5.9 ± 2.6 in the oral anti-diabetic agents group).}

Anatomic and Functional Response to Anti-VEGF Therapy

Regarding functional response, the mean initial BCVA was 20/61 which improved to 20/49 (p=0.003) at 12 months follow-up in the oral anti-diabetic agents group. In the insulin therapy group, the initial BCVA also improved from 20/76 to 20/56 (p=0.005), there was no significant difference between both groups at initial or 12 months examination (p=0.239 and p=0.489, respectively) (Table 2). Similar to the functional analysis, both groups had significant anatomic improvement in CST reduction following anti-VEGF therapy (454.7 μm to 354.9 μm, p<0.001, in the oral anti-diabetic agents group; and 471.5 μm to 368.4 μm p<0.001, in the insulin therapy group). There was no difference between the two groups at baseline or at 12-months follow-up in CST (Table 2).

When glycemic control was considered, patients with HbA1c levels ≤ 7% had a statistically significant improvement in CST in both insulin therapy and oral anti-diabetic agents groups (Table 3). On the other hand, in patients with poor glycemic control (HbA1c levels > 7%), CST only improved significantly in the insulin therapy group (Table 3). Considering visual acuity, there was a trend towards improvement in the well-controlled insulin therapy group and a statistically significant improvement in the poorly controlled insulin therapy group and the well-controlled oral anti-diabetic agents group. Similar to the CST findings, there was no improvement in VA in the poorly controlled oral anti-diabetic agents group (Table 3).

DISCUSSION

DME affects approximately 7% of all people with diabetes.¹, 14 Although the disruption of blood-retinal barrier (BRB), that isolates the retina from the bloodstream, is an important feature of DR, the basic physiological defect that causes retinal vascular leakage is still unknown.⁸ The most well known risk factor for DME is chronic hyperglycemia. It is believed that sustained hyperglycemia damages the BRB directly, thereby leading to increased fluid extravasation and DME.¹⁵ Specifically, hyperglycemia leads to an accumulation of free radicals and advanced glycemic end-products (AGE’s). Additionally the upregulation of VEGF, prostaglandins, and other cytokines alter the structure and function of the BRB.¹⁶ Other systemic associated risk factors include hypertension, nephropathy, anemia, sleep apnea, and pregnancy.¹⁵

Many patients with type 2 DM often require insulin therapy, as a result of the progressive loss of pancreatic beta cell function. However insulin use may also increase the risk of DR and DME in certain patients.¹⁷ Poulaki et al¹⁰ reported that acute, intensive insulin treatment in diabetic rats causes BRB breakdown via VEGF expression. Henricsson et al¹⁷ described a 100% increased risk of DME with insulin treatment compared to oral medication in more than three hundred insulin-treated patients. In addition, Zapata et al¹¹ found that people with type 2 diabetes receiving insulin therapy had increased macular thickness compared to control. Possible mechanisms of action include upregulation of VEFG expression and the vasoactive effects of insulin itself and sudden improvement in glycemic control that further compromises an already-damage BRB¹⁰, 11, 15 Moreover, a recent study have shown that insulin can disrupt tight cell junction in the retinal pigment epithelium cells that regulates the outer BRB.⁸

To our knowledge, no previous study has examined the role of chronic insulin therapy on the responsiveness of the DME treatment with anti-VEGF intravitreal injections. In our study, patients taking oral anti-diabetic medications had a similar improvement in both VA and anatomic parameters compared to patients on chronic insulin therapy. Indeed, the satisfactory outcome of both groups are in agreement with previous clinical trials that have demonstrated a rapid improvement in both visual acuity and CST in patients treated with anti-VEGF therapy for DME.⁶, 7 Both groups of patients had similar levels of HbA1c at the initial and final follow-up removing glycemic control as a variable. Moreover, regardless the modality of DM treatment (oral anti-diabetic agents or insulin therapy), patients with better glycemic control tended to achieve more satisfactory outcomes after anti-VEGF therapy at one-year follow-up.

The present study reveals adequate effectiveness of anti-VEGF therapy in DME treatment regardless of insulin use. A previous study from our group has demonstrated that patients with initial lower HbA1c levels and better glycemic control during the DME treatment achieved a better final outcome measure by both VA and CST.¹⁸ Interestingly, in this study when glycemic control was examined, there appeared to be a particularly significant impact in the oral anti-diabetic agent group on the anatomic and functional response. It is important to recognize that none of our patients were being acutely started on insulin. Therefore, our patients did not have an abrupt decrease of the glycemic level, considered a significant factor for worsening of DR and DME.⁹

It is also significant to note that this research has a number of limitations, some inherent to the retrospective nature of the study. Although DME therapy utilizes an OCT-based treatment regimen with follow-up every 4–8 weeks, this was not mandated and variability may exist, such as possibility of under-treatment in some patients. Patients were treated with a PRN treatment protocol; however, given the retrospective nature of the study, variability may exist from physician to physician of decisions to treat. However, the number of treatments between the two groups was virtually identical. Despite the lack of a strict standard treatment protocol our study involves two groups of individuals with equivalent epidemiologic and metabolic parameters and treated with similar number of anti-VEGF injection. In addition, an ideal situation would include only treatment-naïve DME patients. Another limitations consist on limited information of DME duration and visual acuity assessment without ETDRS standardization. However, we do believe the study brings some value information on the pursuit of understanding the factors involved on the responsiveness of the DME treatment with anti-VEGF therapy.

Based on our findings, anti-VEGF therapy for DME is effective and provides similar results in patients taking oral anti-diabetic agents compared to patients on chronic insulin therapy. There is a particularly robust response to anti-VEGF therapy in both groups when glycemic control is optimized, highlighting the critical importance of communication between the treating physicians managing the DME and the systemic diabetes control. Improving our understanding of the factors that contribute to anti-VEGF response for DME therapy may help to enhance the DME treatment outcomes particularly through the concerted coordinated effort of the treating retina specialist and the endocrinologist or primary care physician.

Acknowledgments