Prediction of diabetic retinopathy from clinical variables and color vision data

Predictions about the onset of retinopathy in 295 diabetic patients, all originally having no evidence of retinopathy, have been made in a longitudinal study over 7 yr. Out of many color vision tests and clinical variables, the best individual predictor was a measure of yellow-blue discrimination, using an anomaloscope. The other predictors of significance were the degree of blood glucose control and the duration of diabetes. Although the predictions from a linear logistic model were significant in classifying the diabetic subjects into those whose fundus will remain normal and those in whom it will develop retinopathy, the number of misclassifications was substantial. An examination of the goodness of fit between the data and the model suggested a criterion value (P) of around P = 0.3 for the probability that a patient develops retinopathy. At this value, the probability of being normal for an individual classed as normal was 0.82, and the probability of developing retinopathy for an individual classed as having retinopathy was 0.54.

Alpha 2-macroglobulin and proliferative retinopathy in type 1 diabetes

Serum alpha 2-macroglobulin (alpha 2m) and total glycosylated haemoglobin (HbA1) concentrations were measured in 110 insulin dependent Type 1 diabetics with minimal or no fundoscopic retinopathy, referred to as non-retinopaths, and in 52 proliferative retinopaths. Proteinuria was recorded in 8 (7%) non-retinopaths and 29 (56%) retinopaths and was accompanied by elevated alpha 2m concentrations in both groups of diabetics but only significantly so in the non-retinopaths. Diabetics without proteinuria showed a significant correlation between alpha 2m concentration and duration of diabetes, HbA1 and age (being higher at extremes of age). Alpha 2m concentrations were significantly higher in retinopaths than in non-retinopaths without proteinuria when allowance was made for the influence of age and duration of diabetes on alpha 2m. This difference may be attributed to the higher HbA, levels found in retinopaths than in non-retinopaths and was no longer evident when account was taken of the prevailing HbA1 concentration in individual patients.

HLA antigens and other risk factors in the development of retinopathy in type 1 diabetes

Factors possibly influencing the development of diabetic retinopathy were studied in 112 randomly selected type 1 diabetics having no or minimal retinopathy (group A) and in 82 type 1 diabetics with known severe diabetic retinopathy. The latter comprised those with severe background retinopathy (group B, n = 17) and those having proliferative retinopathy without (group C, n = 38) and with group D, n = 27) diabetic nephropathy. Nonretinopaths (group A) were of similar sex ratio, body weight, and age at diagnosis of diabetes but had been diabetic longer (p less than 0.001) and were thus older (p less than 0.001) than retinopaths (groups B-D). The distribution of HLA antigens of the A, B, and C loci was similar in nonretinopaths and retinopaths with the exception that HLA B7 showed a reduced (p less than 0.05) prevalence in the retinopaths (6% versus 17%) and was singularly underrepresented in group D, where no patients had this antigen. Mean postprandial plasma glucose and HbA1 concentrations were higher (p less than 0.01 and p less than 0.001) and cigarette smoking was more prevalent (p less than 0.01) in the retinopathy groups B-D than in group A. Systolic and diastolic blood pressures were similar in groups A-C, with higher (p less than 0.001) values only in group D. There was no association between insulin antibody binding in the serum or measurable plasma C-peptide immunoreactivity and retinopathy status. The risk of development of diabetic retinopathy in type 1 diabetes may be related to HLA-associated genetic factors and to cigarette smoking.

Relationship of glycosylated haemoglobin to C-peptide secretory status and antibody binding of insulin in insulin-dependent diabetes

Diabetic control, assessed by measuring the concentration in venous blood of total glycosylated haemoglobin (HbA1), endogenous insulin secretion, as estimated by the C-peptide response (delta C-P) to intravenous glucagon, and serum beef insulin antibody binding were measured in 50 juvenile onset insulin dependent diabetics (IDDM) receiving a single daily injection of soluble and protamine zinc insulin. The delta C-P correlated inversely with duration of diabetes (tau = -0.27, p less than 0.01) and daily insulin requirement (tau = -0.22, p less than 0.05) in the 50 IDDM studied of whom 28 exhibited a measurable delta C-P. In C-peptide nonresponders, but not in the C-peptide responders, and inverse regression (t = 2.19, p less than 0.05) was observed between beef insulin antibody and HbA1. In the 25 IDDM having the lowest insulin antibody binding, and inverse correlation (tau = 0.36, p less than 0.02) was observed between delta CP and HbA1, which was not found (tau = 0.05) in the remaining 25 IDDM who had the highest insulin antibody binding. These findings suggest that, in the absence of endogenous insulin secretion, diabetic control in IDDM receiving a single daily injection of conventional beef insulin is better in patients with high beef insulin antibody binding. Conversely, in patients with low beef insulin antibody binding, diabetic control appears to be better in those with persisting endogenous insulin secretion.

Diabetic control in patients treated with once or twice-daily insulin injections, including a comparison of conventional beef and highly purified pork insulins

Twenty patients were changed from a single daily injection of beef insulin (a mixture of soluble and protamine zinc insulin) to two daily injections (mixtures of soluble and isophane insulins). This was associated with a reduction, one month later, in the concentration of glycosylated haemoglobin (HbA1) and in the degree of late evening glycosuria. A reduction was shown 6 months later in antibody binding of beef and pork insulin by serum. Subsequent conversion to a twice daily regime of highly purified pork insulin was not associated with further improvement in diabetic control, but was associated after 1 month with a reduction in daily insulin dose, and after 5 months with a further reduction in antibody binding of beef and pork insulin by serum. Patients failing to show a C-peptide response to intravenous glucagon had a fall in HbA1 after conversion from a once to a twice daily insulin regime, which correlated inversely with insulin antibody binding estimated at the beginning of the study.

Contraception for the insulin-dependent diabetic woman: the view from one clinic

Experience in a large diabetic clinic has confirmed the suspicion that insulin-dependent diabetic women are at considerably increased risk of thromboembolic disease if they take combined estrogen/progestogen oral contraceptive preparations. The most obvious alternative, an intrauterine device, is associated with an unexpectedly high failure rate, probably because of an unusual metabolic interaction with the diabetic endometrium. In a small group of diabetic women the progestogen-only pill was found to be a successful form of contraception not associated with any side effects except for menstrual irregularities. For most diabetic women the choice of contraceptive should therefore be between a progestogen-only pill and a mechanical method. Female sterilization and injectable progesterone each have their place in particular circumstances. Careful counseling of each patient is essential to ensure the best choice of contraceptive and correct application of the chosen method.

Driving and insulin-dependent diabetes

A survey of 250 patients with insulin-dependent diabetes (IDD) holding a full motor vehicle driving licence revealed that 107 (42.8%) had not declared IDD on their application for a driving licence. 70 of these (28% of entire group) claimed that they were unaware of the statutory requirements. There was no difference in the declaration-rate between men and women. 159 patients (66%) declared IDD for their motor insurance. 86 patients (34.4%) had had severe or frequent hypoglycaemia in the preceding six months, during which they had been driving regularly. 34 patients (13.6%) admitted involvement in a driving accident since commencing treatment with insulin, and 13 of these patients were aware that hypoglycaemia had been an important causal factor. The prevalence of diabetic retinopathy and cataracts was considerable, but few patients (2.4%) had severe impairment of vision in both eyes.

Fatal coma in diabetes

Analysis of causes of death in a population of 3,113 diabetics was carried out for a period of eight years and those patients dying of some form of diabetic coma identified. Of 1,274 deaths, only 22 (1.73%) were primarily due to coma; 7 hypoglycaemia, 8 ketoacidosis, 3 hyperosmolar coma and 4 lactic acidosis. Three of the ketoacidosis patients may have died from other causes. Most deaths occurred in patients with long-standing diabetes. In the hypoglycaemic group all were on insulin and several had been difficult to control for many years. Infection was an important precipitating factor for ketoacidosis and hyperosmolar coma. Phenformin was the cause of all cases of fatal lactic acidosis. It is reassuring that death from coma is a comparatively rare event in known treated diabetic patients.

Glucose and insulin responses to oral glucose in overt non-insulin-dependent diabetics with and without the islet cell antibody

Plasma insulin, glucose, and insulin/glucose responses to a 50-g oral glucose tolerance test (OGTT) were compared in 14 islet cell antibody (ICAb) positive non-insulin-dependent diabetics (NIDDM), 14 matched ICAb negative NIDDM, and 14 ICAb negative nondiabetic controls. Both groups of NIDDM exhibited marked carbohydrate intolerance with insulinopenia. Despite having significantly higher plasma glucose concentrations during the study, the ICAb positive NIDDM had significantly lower insulin levels, and thus lower insulin/glucose ratios, than the ICAb negative NIDDM both in the fasting state and in response to the OGTT. Similarily, ICAb positive NIDDM had higher integrated glucose responses (delta G), lower integrated insulin responses (delta I), and lower delta I/delta G values than ICAb negative NIDDM. Three ICAb negative and seven ICAb positive NIDDM subsequently required insulin treatment. These findings show that ICAb positive NIDDM suffer from a greater disturbance of B-cell function than do matched ICAb negative NIDDM.

Serum alpha 2-macroglobulin levels in diabetes

Serum alpha 2-macroglobulin levels have been determined in diabetic patients by quantitative radial immunodiffusion and compared with those observed in age- and sex-matched controls. In addition, the results in diabetics have been analysed with respect to such variables as the age and sex of the patient, the duration of disease, treatment, control, and the occurrence of retinopathy or nephropathy. The alpha 2-macroglobulin levels in diabetic patients were found to be significantly higher than in age- and sex-matched controls, thus confirming previous observations. However, these differences were most apparent in the more extreme age groups. Multiple regression analysis also revealed that the only variables contributing significantly to the regression apart from age and sex were control and retinopathy.

Seasonal onset of insulin dependent diabetes in relation to sex and age at onset

All newly diagnosed insulin dependent diabetics presenting consecutively to the Diabetic Department of the Royal Infirmary from the City of Edinburgh or its environs, between the years 1964 and 1977, were analysed for sex (297 males, 205 females), age at diagnosis (range 10--75 years), month of diagnosis, duration of symptoms and month of symptomatic onset. Males aged 10--19 years showed a significant seasonal variation in diagnosis (p less than 0.025) with an increase in autumn and winter months which was not seen in females aged 10--19 years nor in patients of either sex aged more than 19 years at diagnosis. The duration of diabetic symptoms increased with increasing age at diagnosis in both males and females and was consistently greater in females than in males for each age group at diagnosis. When the month of symptomatic onset was considered in relation to sex and age at diagnosis, having excluded patients with duration of symptoms of more than three months, only males aged 10--19 years at diagnosis showed a significant seasonal variation in incidence (p less than 0.005).

Diabetic mortality in Edinburgh

A prospective study of mortality in 3,113 diabetics was carried out in Edinburgh over a period of eight years; 1,272 patients (41 %) died. Death rates for females equalled those for males and, in relation to the general population, there was a considerable excess mortality which was greater for females. Statistical analysis indicated that the important mortality risk-factors are age, duration of diabetes of greater than ten years and treatment. The risk of oral therapy or insulin were approximately equally greater than that of diet therapy and probably reflected severity of disease. Using international coding for diagnosis, 27 % of deaths were classified as directly due to diabetes and 49 % to vascular disease. Reclassifying the terminal cause of death left only 26 patients (2 %) recorded with diabetes as the direct cause of death. Three hundred and thirty five males (66 %) and 561 females (73 %) died of vascular disease. There was a predominance of myocardial infarction in males and cerebrovascular disease in females. These percentages were a little lower when post-mortem information was available. These results provide additional evidence that diabetes reduces life expectancy by inducing premature vascular disease and that the effect is greater in women than in men.

Glycosylated haemoglobin concentrations in newly diagnosed diabetics before and during treatment

Concentrations of total glycosylated haemoglobins (Hb A1) were measured in 40 diabetics at diagnosis and at monthly intervals after treatment with chlorpropamide, insulin, or diet alone was begun. The mean Hb A1 concentration at presentation in 16 patients treated with chlorpropamide was significantly higher than that in 12 patients treated with insulin, and the duration of glycaemic symptoms was much longer in the chlorpropamide-treated group. In contrast, the mean plasma glucose concentration was similar in both groups. The mean concentrations of Hb A1 and plasma glucose at diagnosis in the 12 patients treated by diet alone were lower than those in the other two groups, and most of these patients were free of symptoms. Treatment quickly relieved symptoms and lowered plasma glucose in all patients. The Hb A1 concentration fell significantly with treatment such that after two months there was no significant difference between the three groups, although results remained above the normal range. These findings support the theory that the Hb A1 concentration reflects the blood glucose control over the previous one to two months and suggest that the duration of hyperglycaemia may be important in determining the Hb A1 concentration as well as the absolute blood glucose concentration.

The value of islet cell antibody in predicting secondary failure of oral hypoglycaemic agent therapy in Diabetes mellitus

The sera of 160 diabetics who were well controlled by oral hypoglycaemic agents (OHA) for at least three months after diagnosis were tested for pancreatic islet cell antibodies (ICAb) either at diagnosis or within two years after diagnosis. 129 were non-obese at diagnosis and of these ICAb was detected in the sera in 20 (16%). 31 were obese at diagnosis and of these ICAb was detected in the sera in three (10%). All of the 160 diabetics were insulin independent at the time of testing the serum for ICAb. The presence of ICAb was associated with a high probability of becoming insulin dependent, calculated from actuarial statistics. 86% of ICAb positive patients initially controlled on OHA may be expected to be insulin dependent at five years from diagnosis as compared to 18% of ICAb negative patients. Obesity at diagnosis did not significantly affect the probability of becoming insulin dependent. ICAb positive diabetes controlled by OHA can be regarded as a less severe form of insulin-dependent (Type I) diabetes.