The treatment of diabetic retinopathy: a view for the internist

PURPOSE

To review the status of surgical and medical therapy for diabetic retinopathy from the perspective of the non-ophthalmologist.

DATA SOURCES

Relevant English-language articles published from January 1981 to July 1991 were identified through MEDLINE. Other relevant articles were obtained from the authors' personal database.

STUDY SELECTION

For the review of surgical treatment, large randomized, controlled trials were selected. For the review of medical treatment, randomized studies comparing intensive insulin treatment with conventional insulin therapy were selected, as were double-blind, randomized, controlled trials of aldose-reductase inhibitor therapy and antiplatelet therapy in patients with diabetic retinopathy.

DATA EXTRACTION

Emphasis is on findings from large, multicenter, randomized, controlled studies.

DATA SYNTHESIS

Surgery is effective in three clinical situations: Panretinal (scatter) photocoagulation is effective treatment for proliferative retinopathy that is likely to progress to severe visual loss, with such therapy resulting in a 50% to 60% decrease in the main outcome (visual acuity of 5/200 or less; focal photocoagulation decreases the incidence of deterioration of visual acuity by 60% in patients with clinically significant macular edema, but no benefit of photocoagulation has been shown in patients with mild-to-moderate background diabetic retinopathy; and vitrectomy is effective in improving visual acuity only in patients with severe, complicated proliferative retinopathy. Intensive insulin therapy has not been consistently effective in short-term studies with small numbers of subjects. Results of the Diabetes Control and Complications Trial should show whether intensive insulin therapy affects the course of diabetic complications. Aldose-reductase inhibitors have not shown efficacy in changing the course of diabetic retinopathy. Results of trials using antiplatelet agents are controversial.

CONCLUSIONS

Current therapy of diabetic retinopathy is based on detection and surgical treatment of advanced lesions. Medical interventions that effectively halt the progression or prevent the development of diabetic retinopathy are needed.

Lipoprotein (a) levels in diabetes mellitus: relationship to metabolic control

OBJECTIVE

To determine the influence of diabetes control on serum lipoprotein (a) concentrations.

SETTING

Diabetes clinic of a large metropolitan public hospital, with primary- and secondary-care patients.

DESIGN

A cross-sectional study. Comparisons of lipoprotein (a) concentrations were made between a normal control group, a group of diabetic patients with glycated hemoglobin (HbA1c) less than 8.0%, and a group of diabetic patients with HbA1c of 8.0% or higher.

PATIENTS

Ninety-five normal controls and 93 diabetic subjects (49 with insulin-dependent diabetes mellitus and 44 with noninsulin-dependent diabetes mellitus).

RESULTS

Sixty diabetic subjects with HbA1c levels of 8.0% or higher had higher (25 mg/dL) median levels of lipoprotein (a) when compared with either 93 normal controls (8.8 mg/dL) or 33 diabetic patients with HbA1c less than 8.0% (7.5 mg/dL) (P = 0.008 and P = 0.012, respectively). A similar pattern of distribution of lipoprotein (a) levels according to degree of metabolic control was seen in patients with insulin-dependent diabetes mellitus and noninsulin-dependent diabetes mellitus. No difference in the lipoprotein (a) distribution was noted between diabetic men and women. No correlation was observed between lipoprotein (a) levels and total cholesterol, low-density lipoprotein cholesterol, and triglyceride levels.

CONCLUSION

Lipoprotein (a) levels are elevated in poorly controlled diabetic patients. Increased levels of lipoprotein (a) may be a contributing factor to the high risk for atherosclerosis observed in diabetic patients.

Management of hypertension in diabetes

Hypertension should be detected and treated early in diabetic patients. It markedly affects the morbidity and mortality of diabetic individuals as a result of both atherosclerosis and microvascular disease. Antihypertensive treatment is an effective tool in slowing the progression of early and advanced diabetic nephropathy. No prospective studies have addressed the effects of antihypertensive regimens on the incidence of congestive heart failure, stroke, and coronary artery disease in large groups of diabetic patients. Such studies are urgently needed. Special consideration should be given to the effects of antihypertensive drugs on glycemic control and the lipid profile of the diabetic patient. Because hyperinsulinemia (and insulin resistance) have been advocated as hypertensive and atherosclerotic risk factors, the effects of antihypertensive drugs on insulin action and plasma insulin levels may become an important element in the selection an antihypertensive agent. More information, however, is needed in these areas. ACE inhibitors, calcium channel blockers, and alpha-adrenergic blockers probably offer a more favorable metabolic profile as compared with diuretics and beta-blockers. The former agents should be used as initial drugs in most clinical settings.

Treatment of diabetic impotence with a vacuum device: efficacy and effects on psychological status

Twelve patients with erectile impotence related to diabetic neuropathy were treated with a vacuum device, Pos-T-Vac. Efficacy of the device and psychological evaluation (Dyadic Adjustment Scale for marital satisfaction and Hamilton Rating Scale for depression) were performed before and 3 months after treatment. Vacuum therapy was successful in 75% of the patients. Patients with successful impotence treatment and normal baseline marital satisfaction scores showed a modest increase in the scores of marital satisfaction (from 114 +/- 3 points, baseline, to 121 +/- 3 points, posttreatment; p less than 0.05). Vacuum therapy for the treatment of erectile dysfunction due to diabetic autonomic neuropathy appears to be safe and effective.

No correlation between glycemic control and an increase in erythrocyte aldose reductase activity in type I and type II diabetic patients

Aldose reductase, the first enzyme of the polyol pathway, has been related to the pathogenesis of diabetic complications. The regulation of the enzyme in diabetes patients, however, has not yet been clarified. We recently reported that the activity of aldose reductase was increased in erythrocytes of insulin-dependent diabetes mellitus patients but short-term hyperglycemia did not affect the enzyme activity. It is still unclear, however, whether or not the increase in the enzyme activity is caused by long-term hyperglycemia and thus would be seen equally in both type I (insulin-dependent diabetes mellitus) and type 2 (non-insulin-dependent diabetes mellitus) individuals. To further clarify these issues we measured erythrocyte aldose reductase activity in 46 type I patients and 30 type II patients who had variable glucose control and in 16 nondiabetic subjects. We compared the enzyme activity with plasma glucose levels and hemoglobin A1c levels. The results show that erythrocyte aldose reductase activity is increased in both type I and type II patients as compared with nondiabetic subjects (7.1 +/- 0.3 U/L and 6.8 +/- 0.4 U/L erythrocytes versus 5.6 +/- 0.2 U/L erythrocytes, p less than 0.001 and p less than 0.01, respectively), but there were no significant differences between the two groups of diabetic patients. The enzyme activity varied by approximately four times among the diabetic individuals but there was no correlation between the enzyme activity and plasma glucose or hemoglobin A1c levels. We conclude that the increased activity of erythrocyte aldose reductase seen in diabetes is not related to hyperglycemia.

Personality traits as predictors of good diabetic control

To identify personality characteristics that might contribute to overall good control of type I diabetes mellitus, we used a biological correlate of control, glycosylated hemoglobin A1c values, as a means of selecting patients. Patients with evidence of good control (HbA1c less than 7.5%) were compared with patients with evidence of poor control (HbA1c greater than 10.4%). All patients were administered the Personality Research Form E. Need for achievement and a socially desirable response style were associated with good glycemic control. This finding is placed in the context of the complexity of health care behaviors required for adequate self-management.

The effect of the aldose reductase inhibitor, ponalrestat, on the progression of diabetic retinopathy

The objective of this study was to evaluate the effects of ponalrestat, an aldose reductase inhibitor, on the progression of diabetic retinopathy. In this study, 62 patients with diabetes mellitus underwent a double-masked placebo-controlled clinical trial comparing the effect of ponalrestat 600 mg per day with a placebo on the progression of diabetic retinopathy. Both groups were comparable in terms of age, gender distribution, diabetes duration, metabolic control, and presence and severity of diabetic retinopathy. Seven-field stereo fundus photographs were performed at 0 (baseline), 12, and 18 months; 49 patients completed the study (26 in the ponalrestat group and 23 in the placebo group). In both treatment groups, a significant progression of diabetic retinopathy as evaluated by the Early Treatment Diabetic Retinopathy Study classification was observed (Wilcoxon Rank-Sum Test, p less than 0.05). No difference was observed in the progression of retinopathy between the two treatment groups (p = 0.96). The number of microaneurysms increased in the two study groups (from 5.6 +/- 1.2 to 10.5 +/- 1.3 in the placebo group and from 10.3 +/- 1.4 to 12.7 +/- 1.4 in the ponalrestat group); however, the increase was statistically significant only in the placebo group (p less than 0.05). When the increase in the number of microaneurysms was evaluated by change of category of microaneurysm count, no significant difference was observed. We conclude that ponalrestat at a dose of 600 mg per day has no clinically significant effect on the progression of diabetic retinopathy.

Tight glycemic control? Negative

Many poorly controlled patients do not develop proliferative retinopathy and nephropathy, whereas some tightly controlled patients do. In any event, normalization of blood glucose may be impossible and dangerous.

Crucial role of aldose reductase activity and plasma glucose level in sorbitol accumulation in erythrocytes from diabetic patients

Increased sorbitol levels have been demonstrated in tissues of diabetic patients. Although tissue sorbitol levels correlate with plasma glucose levels, a large variability in sorbitol levels has been observed among diabetic patients with similar plasma glucose levels. This variability in tissue sorbitol levels may be due to differences in the activity of aldose reductase, the enzyme that converts glucose to sorbitol. In this study, we isolated aldose reductase from erythrocytes of 31 diabetic patients and 6 nondiabetic control subjects, measured its activity, and compared it to simultaneously measured erythrocyte sorbitol levels. The activity of erythrocyte aldose reductase was increased in diabetic patients compared with control subjects (28.1 +/- 1.4 vs. 22.4 +/- 1.7 nmol.min-1.g-1 Hb, P less than 0.05), but there was an approximately threefold variation in aldose reductase activity among diabetic patients. Erythrocyte aldose reductase activity and fasting plasma glucose levels significantly correlated with the erythrocyte sorbitol level in all individuals (r = 0.48, P less than 0.005 and r = 0.63, P less than 0.005, respectively). The sorbitol level was higher in patients with high aldose reductase activity than in those who had low enzyme activity for any given level of glycemia. The sorbitol production rate calculated from Km and Vmax values showed a better correlation with the erythrocyte sorbitol level (r = 0.80, P less than 0.005), and there was also a good correlation between the erythrocyte sorbitol level and the product of aldose reductase activity by plasma glucose level (r = 0.70, P less than 0.005).(ABSTRACT TRUNCATED AT 250 WORDS)

Relationship between diabetes control and pulmonary function in insulin-dependent diabetes mellitus

PURPOSE

To evaluate the effect of different levels of glycemic control on the pulmonary function of subjects with type I insulin-dependent diabetes mellitus.

PATIENTS AND METHODS

Eighteen subjects with type I insulin-dependent diabetes mellitus with no history or physical findings of respiratory disease. Patients were given insulin therapy with a standard twice-daily insulin injection regimen (standard treatment group) or a subcutaneous insulin infusion device (insulin pump) (intensive treatment group). Glycosylated hemoglobin (HbA1c) levels were determined at quarterly intervals in both groups of patients (standard treatment group, n = 10; intensive treatment group, n = 8). Pulmonary function and diffusing capacity for carbon monoxide (DLCO) were measured after 6 years of continuous follow-up.

RESULTS

The average HbA1c in the standard treatment group was significantly higher than that of the intensive treatment group throughout the 6 years of follow-up (p less than 0.001). The forced vital capacity of the standard treatment group was 85 +/- 3% of predicted as compared with 106 +/- 4% of predicted in the intensive treatment group (p less than 0.001). The DLCO was also significantly diminished in the standard treatment group as compared with that in the intensive treatment group (65 +/- 2% versus 87 +/- 4% of predicted) (p less than 0.001).

CONCLUSION

These data confirm previous reports of abnormal respiratory function in subjects with insulin-dependent diabetes mellitus and suggest that long-term near-normoglycemia may be beneficial in preventing the deterioration of pulmonary function associated with diabetes mellitus.

Transfer of patients with diabetes from semisynthetic human insulin to human insulin prepared by recombinant DNA technology using baker's yeast: a double-blind, randomized study

A multicenter, double-blind, randomized, parallel-group, 12-month study was conducted to determine if patients with diabetes could be transferred safely and effectively from semisynthetic human insulin (ssHI) to human insulin produced by recombinant DNA technology using baker's yeast as the host cell [rDNA HI (BY)]. Sixty-five patients who ranged in age from 19 to 67 years and who had been on a stable dose of NPH or Lente ssHI with or without Regular ssHI participated. Forty-three were transferred randomly to rDNA HI (BY) and 22 continued on ssHI. Evaluation at both six and 12 months of treatment indicated no statistically or clinically significant differences between the ssHI and rDNA HI (BY) groups with regard to the mean changes from baseline in glycosylated hemoglobin, total daily insulin dose, or body weight. In addition, the proportions of patients who had episodes of mild or moderate hypoglycemia, severe hypoglycemia, hyperglycemia, or other clinical experiences related to diabetes or treatment were similar in the ssHI and rDNA HI (BY) groups. No pattern of unexpected retinal changes occurred, nor were there any remarkable changes in human insulin antibody binding or in mean or individual yeast antibody values in either group. These results show that patients with diabetes can be safely and effectively transferred from semisynthetic human insulin to human insulin produced by rDNA technology from baker's yeast with no change expected in the insulin dose.

The use of human insulin derived from baker's yeast by recombinant DNA technology

Human insulin has gradually replaced animal insulin as the therapeutic agent of first choice among insulin-dependent and insulin-requiring patients with diabetes. Like animal insulin, human insulin manufactured by several different methods is available in Regular, NPH, Lente, and 70/30 (NPH/Regular) formulations. The most recently developed method of manufacturing human insulin uses recombinant DNA technology with baker's yeast as the host cell [rDNA HI (BY)], offering potentially limitless supplies of insulin structurally identical to that made by the human pancreas. Clinical studies have demonstrated that the extent of insulin absorption, the glucose-lowering effects, and the clinical effects on glycemic control and on incidence of hypoglycemia with rDNA HI (BY) are similar to those observed when patients are treated with semisynthetic human insulin (ssHI). Dose-for-dose transfer of patients from ssHI to rDNA HI (BY) is therefore appropriate. It is standard practice to recommend that any change in insulin be conducted under medical supervision.

Evaluation of a new automated affinity-chromatographic method for the measurement of glycated hemoglobin

A new automated affinity-chromatographic method for the rapid determination of glycated hemoglobin is described. It has a four min cycle time, does not measure the labile fraction of glycosylated hemoglobin and is not interfered with by abnormal hemoglobins.

Low prevalence of microalbuminuria in normotensive patients with insulin-dependent diabetes mellitus

We studied the prevalence of microalbuminuria (urinary albumin excretion rate [UAER] greater than 20 micrograms/min less than or equal to 200 micrograms/min) as determined in a single, timed, overnight urine collection in 156 normotensive (BP less than 140/90), Albustix negative subjects with type 1 diabetes and its association with arterial blood pressure, the duration of diabetes, levels of glycosylated hemoglobin, body mass index, daily insulin dose and serum cholesterol. Nineteen subjects (12.2%) had a UAER in the microalbuminuric range. The microalbuminuric patients had a significantly longer duration of diabetes, 21 +/- 2 vs 15 +/- 1 years (P less than 0.01), higher diastolic blood pressure, 80 +/- 2 vs 76 +/- 1 mmHg (P less than 0.05) and serum cholesterol concentration, 206 +/- 11 vs 186 +/- 3 mg/dl (P less than 0.05) than did the normoalbuminuric subjects. There were no differences between the normoalbuminuric and microalbuminuric subjects in terms of age, systolic blood pressure, body mass index, daily insulin dose or glycosylated hemoglobin levels. These data indicate that the prevalence of microalbuminuria in type 1 diabetes has probably been overestimated in previous studies due to the inclusion of patients with hypertension. Thus, microalbuminuria, rather than being a predictor of the development of diabetic renal disease, may indicate the presence of diabetic nephropathy with rising blood pressure levels. Further investigation is needed to clarify the relationship between microalbuminuria and coronary risk factors such as serum cholesterol and diastolic blood pressure levels.

Increased erythrocyte aldose reductase activity in type 1 diabetic patients

Elevated erythrocyte sorbitol levels have been demonstrated in diabetic patients. In order to explain the enhanced sorbitol formation, it has been suggested that aldose reductase might be activated by hyperglycaemia. Although aldose reductase activity has been reported to be increased in some tissues of diabetic patients, the effects of varying concentrations of glucose on the enzyme activity in vivo are unknown. To determine whether or not erythrocyte aldose reductase activity is increased in diabetic patients and is affected by glucose levels, we collected blood samples from 10 Type 1 diabetic patients while fasting and 2 h after a standard meal. We measured the activity of erythrocyte aldose reductase after partial purification by column chromatography. The results showed that erythrocyte aldose reductase activity was significantly increased in diabetic patients as compared with non-diabetic subjects (7.3 +/- 0.7 (+/- SE) vs 5.2 +/- 0.3 u l-erythrocytes-1, p less than 0.05). No correlation, however, was observed between fasting plasma glucose levels and the enzyme activity, and acute elevation of the blood glucose level did not affect the enzyme activity.

Immunobiological consequence of regulation of insulin receptor on alloactivated lymphocytes in normal and obese subjects

Acute manipulations of insulin in vivo regulate the display of insulin receptors induced on activated T lymphocytes after presentation of alloantigen. This study explored the immunobiological consequences of regulation of insulin-receptor display by acute manipulations of insulin achieved during the hyperinsulinemic-euglycemic clamp in healthy normal individuals and obese subjects. T lymphocytes were isolated at 0, 1, and 4 h of hyperinsulinemia from seven normal volunteers and seven obese individuals and studied for their capacity to 1) synthesize a complement of insulin receptors on cell membrane, 2) respond to alloantigen in the mixed-lymphocyte culture (an immunologic activity unrelated to manipulations in insulin concentrations in complete medium), and 3) respond to the lymphocyte-mediated cytotoxicity reaction (an immunologic activity known to be modulated by insulin). In the face of a reduction in receptor numbers to 25% of baseline in normal individuals, alloreactivity in the mixed-lymphocyte culture was not affected (95 +/- 9% of time 0 after 4 h of hyperinsulinemia), whereas lymphocyte-mediated cytotoxicity fell from 14 +/- 4 at time 0 to 2 +/- 2% sp Cr release (P less than 0.02). Hyperinsulinemia achieved by the clamp in seven obese subjects did not alter the synthesis of insulin receptors on cell membrane after presentation of alloantigen. In the absence of further reduction of insulin-receptor membrane display, neither the mixed-lymphocyte culture nor lymphocyte-mediated cytotoxicity reaction was affected. It is concluded that those immunologic activities of lymphocytes that can be modulated by insulin are affected by regulation of insulin-receptor display on activated lymphocytes. Therefore, receptor regulation is not effete but carries significant immunologic consequence.(ABSTRACT TRUNCATED AT 250 WORDS)

The effect of aldose reductase inhibition with ponalrestat on the width of the capillary basement membrane in diabetes mellitus

There is evidence to suggest that hyperglycemia is required for the development of the microvascular complications of diabetes. However, the precise mechanism by which hyperglycemia might cause diabetic complications is not completely clear. One possibility is the increased activity of the polyol pathway. Capillary basement membrane thickness is a hallmark histological finding in diabetic microangiopathy. Previous studies in experimental models of diabetes have related the polyol pathway with the thickness of basement membrane in retinal capillaries. To study the effect of aldose reductase inhibition with ponalrestat on the width of the skeletal muscle capillary basement membrane in subjects with diabetes, we measured the capillary basement membrane width in 55 subjects with diabetes in a double masked, placebo controlled randomized trial over a period of 18 months. Twenty-nine patients received ponalrestat (two 300 mg tablets daily) and twenty-six received placebo tablets. The age, sex distribution, type and duration of diabetes were similar in both groups. The glycosylated hemoglobin remained at a constant level throughout the study in both groups. The baseline capillary basement membrane width of the ponalrestat group was 3134 +/- 146 A, it was 3074 +/- 226 A at month 12 and 2548 +/- 182 A at month 18 (P less than 0.001 vs baseline value). The placebo group also had a significant reduction in the width of the capillary basement membrane, from a baseline value of 3026 +/- 147 A to 2818 +/- 144 A at month 12 and 2618 +/- 156 A at month 18 (P less than 0.001 vs baseline value). There was no statistical difference in the capillary basement membrane width between the two groups at any time point.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of restricting dietary protein on the progression of renal failure in patients with insulin-dependent diabetes mellitus

BACKGROUND

Restriction of dietary protein may slow the progression of renal failure in diverse renal diseases, but the extent to which such a diet is beneficial in patients with diabetic nephropathy is uncertain.

METHODS

We studied the effect of reduced intake of protein and phosphorus on the progression of renal disease in 35 patients with insulin-dependent (Type I) diabetes mellitus and clinically evident nephropathy. The low-protein, low-phosphorus diet contained 0.6 g of protein per kilogram of ideal body weight per day, 500 to 1000 mg of phosphorus, and 2000 mg of sodium. The control diet consisted of the patient's prestudy diet with the stipulation that it contain 2000 mg of sodium and at least 1 g of protein per kilogram per day and 1000 mg of phosphorus. Renal function was assessed by measurement of iothalamate and creatinine clearances at intervals of 3 to 6 months, and the patients were followed for a minimum of 12 months (mean, 34.7). The declines in mean glomerular filtration rates were compared between groups by linear-regression analysis of the glomerular filtration rate as a function of time.

RESULTS

The patients who followed the study diet for a mean of 37.1 months had declines in iothalamate clearance of 0.0043 ml per second per month and in creatinine clearance of 0.0055 ml per second per month. The comparable values in the control group were 0.0168 and 0.0135, respectively (P less than 0.05). Blood pressure was well controlled, and the degree of glycemic control was comparable in both groups.

CONCLUSION

Dietary restriction of protein and phosphorus can retard the progression of renal failure in patients with Type I diabetes mellitus who have nephropathy. We believe that wider use of this treatment is indicated.

Blurred vision: an overlooked initial presenting symptom of insulin-dependent diabetes mellitus

Blurred vision, a well-known complication of uncontrolled diabetes mellitus, is occasionally overlooked as the first and only presenting symptom of type I diabetes mellitus. In this report we describe six type I diabetic patients whose first and only symptom of the disease was blurred vision, as documented by increased glycosylated hemoglobin. Institution of therapy was delayed until the classical symptoms of diabetes mellitus appeared.