Insulin antibodies (IAs) affect blood glucose control in patients receiving insulin therapy.

To investigate the relationship between different hypoglycemic treatments and IAs in patients with type 2 diabetes mellitus (T2DM).

This cross-sectional, retrospective study included 1863 patients with T2DM who were receiving exogenous insulin therapy. All patients received stable antidiabetic therapy in the last 3 months and IA levels were measured using an iodine-125 array.

A total of 1863 patients were enrolled. There were 902 (48.4%) patients who had positive IAs (IA level > 5%), with a mean IA level of 11.06% (10.39%-11.72%). IA levels were positively correlated with high fasting blood glucose (odds ratio = 1.069, P < 0.001). The proportion of positive IAs was lowest in patients using glargine only (31.9%) and highest in patients using human insulin only (70.3%), P < 0.001. The IA levels in patients using sulfonylureas/glinides (8.3%), metformin (9.6%), and dipeptidyl peptidase-4 inhibitors (8.2%) were all lower than in patients without these drugs (all P < 0.05).

Nearly half of patients on insulin therapy have positive IA antibodies, and IA antibody levels are associated with blood glucose control. Insulin glargine and a combination of oral glucose-lowering drugs were correlated with lower IA levels.

Autoimmune glycaemic dysregulation and hyperinsulinaemic hypoglycaemia mediated by insulin autoantibodies is an increasingly recognised but controversial phenomenon described in both exogenous insulin naïve (insulin autoimmune syndrome) and exposed (exogenous insulin antibody syndrome) individuals. There has been a significant proliferation of case reports, clinical studies and reviews in the medical literature in recent years which have collectively highlighted the discrepancy between experts in the field with regard to the nomenclature, definition, proposed pathophysiology, as well as the clinical and biochemical diagnostic criteria associated with the condition. The essential characteristics of the condition are glycaemic dysregulation manifesting as episodes of hyperglycaemia and unpredictable hyperinsulinaemic hypoglycaemia associated with high titres of endogenous antibodies to insulin. Although the hypoglycaemia is often life-threatening and initiation of targeted therapies critical, the diagnosis is often delayed and attributable to various factors including: the fact that existence of the condition is not universally accepted; the need to exclude surreptitious causes of hypoglycaemia; the diverse and often complex nature of the glycaemic dysregulation; and the challenge of diagnostic confirmation. Once confirmed, the available therapeutic options are expansive and the reported responses to these therapies have been variable. This review will focus on our evolving understanding, and the associated diagnostic challenges - both clinical and laboratory - of this complex condition.

Insulin autoimmune syndrome (IAS), characterized by glycemic dysregulation and life-threatening hypoglycemia, can occur in patients with type 1 diabetes (T1D). Diagnostic confirmation is complex but important in order to ensure timely initiation of definitive therapy.

We aimed to quantitate the degree of immunoglobulin-insulin complex (IIC) formation and its effects on glycemic control in a patient with T1D and IAS compared with T1D and non-T1D controls and before and after therapeutic plasma exchange (TPE).

The prospective descriptive study was conducted between June 2015 and December 2015 in a quaternary children's hospital in Brisbane, Australia. Percent Free "Immunoreactive" Insulin (%FII) as assessed by polyethylene glycol precipitation studies and its relationship to plasma glucose and serum insulin concentration.

Samples from the patient with T1D and IAS demonstrated lower mean %FII compared to T1D (23.8 ± 2.0 vs 52.0 ± 6.7; P < .0001) and non-T1D (23.8 ± 2.0 vs 102.9 ± 2.7; P < .0001) controls. This was associated with loss of glycemic predictability and frequent severe hypoglycemia. TPE increased %FII (23.8 ± 2.0 before TPE vs 83.6 ± 2.5 after TPE, P < .0001) and reestablished plasma glucose responsiveness to exogenous insulin.

IAS should be considered in T1D patients with unexplained glycemic instability and hypoglycemia. The laboratory plays an integral diagnostic role.

TPE is an effective method for removing IICs and normalizing insulin-mediated glucose responses.

To study the long-term development (104 weeks) of insulin antibodies during treatment with insulin detemir (IDet) and insulin aspart (IAsp) in children with type 1 diabetes aged 2-16 years.

A 52-week, two-arm, randomized trial comparing IDet and neutral protamine Hagedorn insulin, both in combination with IAsp, was followed by a one-arm, 52-week extension trial of the IDet + IAsp arm. The present analysis was conducted in children who completed the randomized trial and entered into the extension trial.

Of the 177 children randomized to IDet treatment, 146 entered the extension trial. IDet-IAsp cross-reacting antibodies peaked within the first 39 weeks of treatment before gradually declining. A similar pattern was seen for IDet-specific and IAsp-specific antibodies. At end of trial (EOT), no correlation was observed between the level of IDet-specific or IAsp-specific antibodies or IDet-IAsp cross-reacting antibodies and either glycated hemoglobin (HbA1c) or basal insulin dose. Mean HbA1c was stable during the treatment period, with a slight increase over time from 8.41% (68.4 mmol/mol) at baseline to 8.74% (72 mmol/mol) at EOT. Mean IDet dose increased from 0.43 U/kg at baseline to 0.66 U/kg at EOT. Mean IAsp dose increased from 0.46 U/kg to 0.51 U/kg at EOT.

Although treatment with IDet and IAsp is associated with development of specific and cross-reacting antibodies, no correlation between insulin antibodies and basal insulin dose or HbA1c was found.

Novo Nordisk A/S. ClinicalTrials.gov identifiers: NCT00435019 and NCT00623194.

This study investigated the prevalence of islet autoantibodies in children and adults with T1DM according to their age and the duration of disease.

We measured the levels of islet autoantibodies, including antiglutamic acid decarboxylase antibody (anti-GAD Ab), and combined these with anthropometric measurements and laboratory tests of 137 patients newly diagnosed with T1DM during the last 20 years. The subjects were subdivided into four groups according to their age at the onset of the disease. We then compared the prevalence of islet autoantibodies in the different age groups with the duration of disease.

Among the 137 patients, 68.9% tested positive for islet autoantibodies (71.4% within 1 year; 67.7% after 1 year of the disease onset). Within 1 year of the onset of the disease, 66.3% of the patients were positive for the anti-GAD Ab, and 35.6% were positive for IAAs. The prevalence of islet autoantibodies was significantly higher in the prepubertal groups than in the postpubertal groups (80.0% vs. 58.3%). The rate of positive islet autoantibodies changed with the duration of disease, and it differed according to the type of autoantibody and the age of the patient.

The rates of positive islet autoantibodies were significantly higher in younger than in older patients at the time of the diagnosis of the disease. The positive rates were significantly changed 1 year after the onset of the disease in the preschool and the children groups. So these findings suggest that we need to diagnose type 1B diabetes distinguished T2DM in aldolescent group, carefully.

To investigate the influence of two insulin administration modalities, continuous subcutaneous insulin infusion (CSII) and multiple daily insulin injections (MDI) therapy with insulin analogues, on the development of insulin antibodies (IAs) in patients with type 1 diabetes mellitus and to assess the impact of IAs on glucose control and hypoglycaemia. 96 patients with type 1 diabetes mellitus treated with CSII (n = 48) or MDI (n = 48) were included in the study. Age, duration of diabetes, A1c, preprandial and postprandial blood glucose and hypoglycaemic events were compared between IA positive and negative patients. IA levels were higher in the CSII group (% 24.6 ± 14.2) than the MDI group (% 13.2 ± 9.9). Duration of diabetes and age were not associated with IA positiveness. While A1c, preprandial blood glucose and the frequency of hypoglycaemic events were similar in two groups, postprandial blood glucose was lower in IA positive group (P = 0.03). Patients with type 1 diabetes mellitus treated with CSII with insulin analogues had higher IA levels when compared to MDI therapy. However, the development of IAs did not impair the glycaemic control.

The purification of animal insulin preparations and the use of human recombinant insulin have markedly reduced the incidence, but not completely suppressed, the development of anti-insulin antibodies (IAs). Advances in technologies concerning the mode of delivery of insulin, i.e. continuous subcutaneous insulin infusion (CSII), continuous peritoneal insulin infusion (CPII) and more recently inhaled insulin administration, appear to significantly increase circulating levels of immunoglobulin G (IgG) anti-IAs in diabetic patients. However, the increase is usually moderate and mostly transient as compared to previous observations with poorly purified animal insulin preparations. The clinical impact of these circulating anti-IAs remains unclear. Nevertheless, several studies have suggested that antibodies could retard insulin action, leading to a worsening of postprandial hyperglycaemia and/or serve as a carrier, thus leading to unexpected hypoglycaemia. CPII may be associated with more marked and sustained increase in IAs levels, possibly related to the use of an unstable insulin and the formation of immunogenic aggregates of insulin. The possible clinical consequences of these high levels of IAs remain to be evaluated because a low-glucose morning syndrome or severe insulin resistance with ketone bodies production have been reported in some cases. In conclusion, even if CSII and CPII may promote the development of circulating IAs, this increase does not lead to immunological insulin resistance, compared to that previously described with animal non-purified insulin preparations, and seems to have only marginal influence on blood glucose control or complications in most diabetic patients.

A 46-year-old nonatopic woman who had been suffering from type 2 diabetes for 17 years was hospitalized at the Endocrinology Department of Selcuk University due to very high glucose levels after recovery from acute hepatitis A infection. She had never used insulin before. After first subcutaneous dose of human regular insulin, severe local allergic reaction developed. Desensitization to insulin was tried. One day later, ketoacidosis developed. Human regular insulin was again subcutaneously injected to the patient. Severe anaphylactic reaction occurred, and in spite of all the medical attempts to save the patient, she died.

Obese, insulin-resistant patients have been shown to have metabolic inflexibility. The goal of this study was to examine the effect of insulin administration on energy metabolism in lean, type 1 diabetic (DM1) patients. Eleven DM1 patients without vascular complications and 11 healthy controls (C) were examined. We performed a 2-step hyperinsulinemic euglycemic clamp (240 minutes; period 1: 1 mU. kg(-1). min(-1) and period 2: 10 mU. kg(-1). min(-1)) combined with indirect calorimetry during basal period B (B, -45 to 0 minutes), period 1, and period 2 of the clamp. The metabolic clearance rates of glucose (MCR) were lower in DM1 compared with C in period 1 (12.54 +/- 3.38 v 17.41 +/- 6.18 mL. kg(-1). min(-1); P <.02), as well as in period 2 (21.63 +/- 6.47 v 26.61 +/- 4.45 mL. kg(-1). min(-1); P <.05). Basal respiratory quotient (RQ) was lower in DM1 compared with C (0.72 +/- 0.04 v 0.75 +/- 0.04; P <.03). Insulin administration was accompanied by an increase in RQ in both groups, which was lower in DM1 compared with C (period 1: +0.09 +/- 0.04 v +0.11 +/- 0.07; P <.001; period 2: +0.13 +/- 0.04 v +0.16 +/- 0.04; P <.001). Glucose oxidation did not differ between the groups in period B; however, it was lower in DM1 compared with C in periods 1 (1.17 +/- 0.67 v 3.28 +/- 1.11 mg. kg(-1). min(-1); P <.003); and 2 (2.10 +/- 0.64 v 3.28 +/- 0.93 mg. kg(-1). min(-1); P <.009). Lipid oxidation was higher in DM1 in all periods compared with C; period B (3.28 +/- 0.77 v 1.16 +/- 0.55 mg. kg(-1). min(-1); P <.001), period 1 (1.10 +/- 0.41 v 0.67 +/- 0.54 mg. kg(-1). min(-1); P <.05), and period 2 (0.99 +/- 0.29 v 0.52 +/- 0.58 mg. kg(-1). min(-1); P <.01). The groups did not differ in protein oxidation. In conclusion, DM1 patients with secondary insulin resistance (IR) are characterized by metabolic inflexibility manifesting itself by smaller increases in RQ and glucose oxidation after insulin administration during the euglycemic clamp.

Nonviral gene transfer was investigated as a potential treatment of growth hormone deficiency (GHD) using hypophysectomized mice as a model. After a single hydrodynamic administration of naked plasmid DNA containing the human growth hormone (hGH) gene controlled by an ubiquitin promoter, sustained elevation of circulating hGH was observed the entire observation period (68 days), with a concomitant normalization of circulating insulin-like growth factor I (IGF-I) and IGF-binding protein-3. Furthermore, longitudinal growth was corrected in terms of normalization of tibia length, tail length, and body weight gain. Liver, spleen, and lung weights were normalized, whereas heart weight was normalized partly. hGH mRNA was expressed exclusively in liver tissue. In conclusion, we showed that nonviral hGH gene transfer normalizes longitudinal growth in hypophysectomized mice, indicating that this method potentially could be relevant as a new therapeutic tool in the clinical handling of GHD.

To compare steroids and their associations in men with type 1 diabetes and healthy control subjects.

We studied 52 adult men with type 1 diabetes without microvascular complications, compared with 53 control subjects matched for age and BMI. Steroids and their binding globulins were assessed in a single venous blood sample and a 24-h urine sample.

In adult men with type 1 diabetes, total testosterone did not differ from healthy control subjects, but sex hormone-binding globulin (SHBG) (42 [14-83] vs. 26 [9-117] nmol/l, P < 0.001), cortisol-binding globulin (CBG; 0.87 +/- 0.17 vs. 0.73 +/- 0.10 nmol/l, P < 0.001), and cortisol levels (0.46 +/- 0.16 vs. 0.39 +/- 0.14 nmol/l, P < 0.01) were higher. The free testosterone index was lower (60 [17-139] vs. 82 [24-200], P < 0.001), and the calculated free testosterone was slightly lower (497 [115] vs. 542 [130], P < 0.064), but the pituitary-gonadal axis was not obviously affected in type 1 diabetes. The calculated free serum cortisol was not different, and 24-h urinary free cortisol excretion was lower in type 1 diabetes (121 [42-365] vs. 161 [55-284] nmol/24 h, P < 0.009). Testosterone was mainly associated with SHBG. Estimated portal insulin was a contributor to SHBG in control subjects but not in type 1 diabetes. Cortisol was associated with CBG. HbA(1c) contributed to CBG in men with diabetes but not in control subjects, whereas estimated portal insulin did not contribute.

Adult men with fairly controlled type 1 diabetes without complications who are treated with subcutaneous insulin have a tendency to hypogonadism, as reflected by lower free testosterone levels in the presence of similar total testosterone levels and higher SHBG levels.

To assess the antigenicity of the insulin Hoechst 21PH (Hoe21PH) using continuous subcutaneous insulin infusion (CSII) and to compare the antigenicity of this insulin when administered intraperitoneally or subcutaneously. RESEARCH DESIGN AND METHODS; Peritoneal administration of Hoe21PH (Hoechst-Roussel, Somerville, NJ) insulin using implantable devices (continuous peritoneal insulin infusion [CPII]) increases anti-insulin antibody (AIA) levels in type 1 diabetic patients. Intraperitoneal administration, addition of a stabilizer (polyethylene polypropylene glycol), or insulin modifications due to storage in the pump may be involved in this antigenicity. In this nonrandomized study, 24 type 1 diabetic patients were treated with either CSII (n = 11, group 1) or CPII (n = 13, group 2). AIA levels were measured by radioimmunoassay (RIA) or enzyme-linked immunosorbent assay (ELISA) before starting patients on Hoe21PH and again after 3 and 6 months.

Patients were comparable in the two groups. AIA levels (RIA) remained stable (24.3 +/- 8.5% [month 0] to 24.9 +/- 8.5.5% [month 6]) in group 1 and increased (21.8 +/- 6.7% [month 0] to 41.8 +/- 6.9% [month 6]) in group 2 (P = 0.005, Wilcoxon's rank-sum test). Using ELISA, AIA remained stable in the patients in group 1 (n = 9; 3.8 +/- 0.8 units/ml [month 0] and 4.1 +/- 1.0 units/ml [month 6]) and tended to increase in the patients in group 2 (n = 12; 4.1 +/- 0.7 units/ml [month 0] to 17.5 +/- 4.6 units/ml [month 6]) (P = 0.07). Comparison of the evolution of AIA formation between the two groups, using RIA at months 0, 3, and 6 showed a significant difference (analysis of variance, P = 0.009).

No increase in AIA levels was demonstrated when Hoe21PH insulin was administered subcutaneously as assessed by two different assays. CPII is proven to be more antigenic than CSII, and this is not related to a specific antigenicity of Hoe21PH insulin. The intraperitoneal route of administration or insulin modifications due to insulin storage in implantable devices might explain this antigenicity.

Systemic lupus erythematosus (SLE) is characterized by the finding of ample serum autoantibodies. The role and the origin of many of these antibodies are still obscure. The aim of this work was to study the occurrence of anti-insulin antibodies (AIA) in SLE, and to postulate, based on AIA determination, on the mechanisms involved in the production of some autoantibodies in SLE. IgG and lgM AIA, anti-DNA antibodies (ADA) and anti-tetanus toxoid antibodies (ATA) were determined using ELISA in sera and B-lymphocytes culture media of 24 SLE patients, 10 healthy controls and 19 insulin-dependent diabetes mellitus (IDDM) patients. B- and T-lymphocytes were isolated using Ficoll gradient, depleted of T-cells using cyclosporin A, EBV infected and grown in medium. The frequencies of IgM-AIA and IgG-ADA were higher in SLE patients than in healthy controls (P < 0.02 and P < 0.05, respectively). The rate of IgM-AIA in SLE and IDDM was comparable, while IgG-AIA was significantly less common in SLE than in IDDM (P < 0.05). The prevalence of ATA in SLE patients and healthy controls was similar. These findings increase the spectrum of the humoral autoimmune response in SLE and suggest that part of it (natural autoantibodies) is independent of antigen driven response.

The aim of this study was to measure the pharmacokinetics and pharmacodynamics of subcutaneously injected 40 IU/ml porcine lente insulin preparation (Caninsulin, Intervet BV, The Netherlands) in diabetic cats. The pharmacological properties of the insulin in poorly controlled or untreated cats were compared with those after several weeks of treatment, to determine if improved diabetic stability altered the pharmacology of this insulin. In addition, the pharmacological properties of intravenously injected 100 IU/ml regular porcine insulin (Actrapid MC, NovoNordisk, Denmark) were measured. Serial plasma samples were collected after subcutaneous injection of porcine lente insulin from 25 diabetic cats in the first week of admission to a 12-month diabetic treatment trial. Samples were also collected after 4 or 8 weeks of treatment, in those cats which had not achieved diabetic remission by this time. At this time, serial plasma samples were also collected from these cats after intravenous injection of porcine regular insulin. Plasma samples were assayed for glucose, anti-insulin antibodies were extracted using a PEG technique, and samples were assayed for insulin using an RIA kit with low sensitivity for endogenous feline insulin, but high sensitivity for exogenous porcine insulin in feline plasma. Caninsulin injected subcutaneously in diabetic cats led to a peak insulin concentration in plasma after 1.7+/-0.1 h, and a nadir of blood glucose after 4.1+/-0.3 h. Insulin and glucose concentrations returned to baseline within 12 h. There was no significant change in the onset or duration of Caninsulin action between the first week of treatment and 5 or 9 weeks of treatment. Actrapid MC injected intravenously had a peak insulin at 0.36+/-0.03 h, and a nadir of blood glucose at 1.9+/-0.3 h. Insulin and glucose returned to baseline within 6 h. It was concluded that Caninsulin injected subcutaneously has suitable pharmacological properties for the twice-daily treatment of diabetes mellitus in cats. In addition, Actrapid MC insulin injected intravenously has suitable pharmacological properties for injection every 4-6 h in diabetic cats.

The majority of young diabetics in India prefer to use low-cost bovine insulin for economic reasons. Therefore, the question of insulin antibody response to bovine insulin and its functional significance is still relevant in the Indian context. We assessed insulin antibody response in 52 young diabetics (type 1, n=25, malnutrition modulated form of diabetes, n=19 and fibrocalculous pancreatopathy (FCP) n=8) on bovine insulin therapy (mean duration 3.0+/-2.1 years) using an internationally standardised in-house radioligand assay. The functional significance of insulin antibody was assessed by calculating their affinity constant, maximum binding capacity and total insulin binding power by Scatchard analysis (type 1, n=14, malnutrition modulated form of diabetes, n=11). All the patients treated with bovine insulin showed high titers of insulin antibodies with S.D. score ranging from 5.1 to 42.0. No significant difference was observed in the mean S.D. score of insulin antibodies in the three diabetic groups. The mean daily insulin dose, maximum insulin binding capacity and total insulin binding power were significantly higher in type 1 when compared to the malnutrition modulated form of diabetes (36+/-8 vs. 26+/-11 IU/day, P<0.05; 9. 7+/-7.8 vs. 4.0+/-3.9 nmol/l, P=0.03 and 59+/-29 vs. 29+/-43, P=0.01, respectively). Insulin antibodies S.D. score and its affinity did not show significant relationship with daily insulin dose and glycemic control (HbAl) at admission. Only 24+/-7% variations in daily insulin requirement were accounted for by total insulin binding power. There was a significant inverse relationship between insulin antibody S.D. score and duration of insulin therapy (r=-0. 4172, P<0.0004). To conclude, insulin antibody response following bovine insulin therapy is not different among type 1, malnutrition modulated form of diabetes and FCP diabetes. The insulin antibody response to bovine insulin therapy does not contribute significantly to increase in daily insulin requirement in bovine insulin treated insulin requiring young diabetics.

Many diabetic patients taking multiple subcutaneous insulin injections cannot adjust their dosage appropriately to maintain blood glucose within a normal range. It is hard to predict how dosage changes and physiological fluctuations affect insulin levels and subsequently glucose control. To examine these issues, we have developed a model representing the link between dosage and blood insulin levels. Our model adequately predicts insulin concentrations for individual patients and could be incorporated into an overall glucose-insulin representation. More importantly, parameter and sensitivity analysis results highlight insulin kinetic features that are difficult to isolate in a clinical setting and that may significantly influence glucose dynamics. For example, large interpatient variation, measured quantitatively by model parameters, emphasizes the need for individualized design of insulin regimens. Intrapatient variations are also large in some patients. Improved control for these patients may only be possible through more frequent sampling and control action. The sensitivity coefficient for absorption suggests a significant overlapping injection effect that is not considered in present patient management strategies.

In the horse, adenomata of the pairs intermedia of the pituitary gland have been associated with the distinct clinical entity of Cushing's disease which arises largely as a result of excessive secretion of adrenocorticotropin (ACTH) or other proopiomelanocortin (POMC) peptides. Pars intermedia peptide secretion is under dopaminergic control and compounds such as pergolide or bromocriptine, which are dopamine agonists, can palliate the clinical signs. A variety of endocrinological abnormalities, relevant to both pathogenesis and diagnosis, may be demonstrated in equine Cushing's disease, including hyperadrenocorticism, peripheral insulin resistance and excessive POMC-peptide secretion from the pituitary gland. Preliminary studies on carbohydrate metabolism suggest that quantification of insulin activity may be a useful prognostic index in cases of equine Cushing's disease, and that insulin therapy of secondary diabetes mellitus may be indicated in some cases.

Although it has often been stated that insulin antibodies cause insulin resistance, this concept is still controversial. The effect of insulin antibody GP30, commonly used in insulin radioimmunoassay, on insulin action was investigated in Wistar rats in vivo by the euglycemic glucose clamp technique. As a preliminary experiment, the equilibrium time required for insulin antibody to bind with endogenous insulin was examined. One hundred microliters/kg insulin antibody took 60 min or more to attain equilibrium, but 10 microliters/kg insulin antibody almost immediately equilibrated with endogenous insulin. During a 60-min glucose clamp study, 2 mU/kg/min porcine insulin was infused with 100 microliters/kg insulin antibody. At steady state, during the last 20-min period, the mean glucose infusion rate was 2.10 +/- 0.85 mg/kg/min (n = 5, mean +/- SD), significantly lower than the 5.77 +/- 1.61 mg/kg/min of the control, indicating insulin resistance before equilibrium was reached. However, the glucose infusion rates during the clamp with 10 microliters/kg insulin antibody and 100 microliters/kg insulin antibody infused 75 min before the insulin were 6.10 +/- 1.44 and 7.12 +/- 1.19 mg/kg/min, respectively, no different from the control. In these instances, free insulin levels measured by radioimmunoassay using the polyethyleneglycol method were 43.8 +/- 20.4 and 15.4 +/- 6.1 microU/ml, respectively, lower than the control (77.0 +/- 16.1 microM/ml).(ABSTRACT TRUNCATED AT 250 WORDS)

Humoral immune factors related to type 1 diabetes have been investigated in children with coeliac disease. Anti-insulin (IAAb), immunoglobulin (alpha IgAb), islet cell (ICA) and glucagon autoantibodies were examined in 15 children with coeliac disease at diagnosis (group 1), in 15 children with coeliac disease following a gluten-free diet (group 2) and in 30 control patients (groups 3 and 4). IAAb were present in 27% of group 1 and in 20% of group 2 patients and alpha IgAb were significantly increased in group 1 and 2 patients; two patients in group 2 were positive for ICA; none of the coeliac disease patients were positive for anti-glucagon antibodies. The levels of anti-gliadin antibodies in group 1 were positively correlated with those of alpha IgAb. Coeliac disease-related HLA antigens were not correlated with antibody presence. The presence of diabetes-related humoral immune factors in coeliac disease raises the question as to whether or not they are predictive of subclinical pancreatic damage or whether they are simply indicators of a more general autoimmune diathesis.

Continuous insulin infusions are a valuable way of managing highly selected patients, although patients and healthcare practitioners must be aware of the limits and the increased risks involved with this type of technology. Maximum benefit from the CSII technology is achieved when the patient is part of a complete healthcare team accessible on a daily basis to respond to the changing nature of the underlying diabetes. Intranasal and pulmonary delivery of insulin, in contrast, represent a minor technology that will potentially add convenience to some diabetic management plans and possibly provide a new treatment approach for noninsulin-dependent diabetic patients.

Insulin allergy and antibody-mediated resistance may complicate therapy with animal insulins. We describe a 53-year-old man manifesting both resistance and persistent systemic allergy despite treatment with recombinant human insulin. Insulin resistance and symptoms of allergy appeared in this patient several months after initiating therapy with mixed beef-pork insulin, as is often the case. Symptoms initially improved, but persisted, and then worsened again, despite continuous human insulin therapy. Total insulin-binding capacity by Scatchard analysis, high plasma insulin-binding capacity, and specific anti-insulin antibody levels were consistent with an immunologic form of insulin resistance. Glucocorticoid therapy was required both to reduce allergic findings and to restore glycemic control. Although recently available human insulins may be less immunogenic than animal forms, immune responses to exogenous human insulin still may pose significant clinical problems.

In order to investigate the role of insulin as a potential target autoantigen of cellular immunity in the prediabetic period, proliferative responses of T lymphocytes to human insulin were studied in nine islet-cell antibody (ICA) + first-degree relatives of patients with Type I diabetes (individuals at high risk for the development of Type I diabetes, or the 'prediabetic' group, which was never treated with insulin) and in 12 control individuals. Insulin autoantibodies were present in 6/9 (67%) of the prediabetic subjects and none of the controls. Peripheral blood lymphocytes were collected on Ficoll and incubated with human insulin, control antigens, or media alone for 5-6-day and 9-10-day incubation periods. Cells were pulsed with 3H-thymidine, harvested, and analysed in a scintillation counter. Results are expressed as stimulation index (SI = cpm with antigen/cpm without antigen), with a SI greater than or equal to 1.5 considered a positive response. Eight of nine (89%) prediabetic individuals responded positively to insulin after a 9-10-day incubation period, in contrast to four of 12 (33%) control subjects, P less than 0.05. The mean proliferative response to insulin after 9-10 days' incubation was 2.1 +/- 0.4 and 1.2 +/- 0.1, for the prediabetic and control groups, respectively. The proliferative response to insulin was not directly correlated with levels of insulin autoantibodies (r = -0.05, NS). These data suggest that most individuals at high risk for the development of Type I diabetes display a cellular immune response to insulin, and a subset of these individuals does not display a concomitant humoral immune response to insulin based on the presence or absence of insulin autoantibodies.

Subcutaneous insulin absorption is a complex process, whose quantitative aspects have important clinical implications. In this review we briefly discuss the rationale of modelling techniques before introducing some of the more common types of models (empirical vs mechanistic, simple vs complex, compartmental) found in the biological literature. The various approaches are compared regarding their suitability to model subcutaneous absorption of insulin. Methods are described (monitoring residual depot activity or the appearance of insulin in the systemic circulation) which allow the determination of model parameters from experimental data. The degree to which current model predictions describe the available experimental data is discussed. Since the absorption of insulin involves a number of poorly understood events it would be difficult, at this time, to construct a complex model which completely describes all aspects of the absorption process. Although the simpler techniques (such as the use of a one-pool model) provide only an approximate description of subcutaneous kinetics they are likely to remain useful tools in routine investigation.

Glucose intolerance is a nearly universal finding in patients with chronic renal failure and in animal models of uremia. The glucose intolerance results from impaired insulin-mediated glucose disposal by muscle, adipose, and liver tissue. Insulin binding by these tissues is not reduced. Rather, several defects exist in the postreceptor cascade of insulin action. Although impaired insulin-mediated glucose uptake and metabolism occur, the primary defect and causative agent are not established. The purpose of the present article is to review recent literature on the potential mechanisms underlying the insulin resistance of chronic renal failure.

Human sera were tested for insulin antibodies by fluid and solid phase assays. Radioimmune titres determined with 125-I Tyr A14 insulin were not correlated with those obtained using insulin coated microplates and enzyme linked immunodetection (n = 60). Several reasons for this lack of correlation were found. Iodine substitution on the A14 residue of insulin may significantly alter the avidity of some insulin antibodies for their ligand; hence, disclosing a heretofore unsuspected pitfall for antibody determination by radioimmunoassay. Specificity for bovine insulin was easily demonstrable in fluid phase by comparing the binding of monoiodinated bovine, porcine and human insulin. By contrast, in solid phase assay, titres obtained with microplates coated with bovine or human insulin were almost equal, regardless of the serum specificity for bovine insulin. This lack of specificity of the solid phase assay is not due to denaturation or unavailability of the bovine specific epitope because: bovine specificity could be demonstrated by competitive assay, after preincubation of the serum with insulin of the different species; and, coating with crosslinked insulin dimers or oligomers instead of monomers did not unmask bovine specificity. It is concluded that radioimmune methods are best suited to study specificity but may be biased by the presence of the radioiodine label whereas solid phase assay detects low avidity antibodies with great efficiency but is less appropriate to study specificity.

In order to define the detection limit of a radioimmunoassay for insulin-antibody a correction was made for binding in the presence of an excess of unlabelled insulin and assay precision was calculated. One hundred forty control sera were assessed; all were islet cell antibody negative. For each sample, binding of 125-I human insulin was determined both with and without excess unlabelled insulin, subtraction of the latter acting as a correction. The distribution of uncorrected binding was skewed while corrected binding was normally distributed, (mean (SD) = 0.149 (0.298%)) Precision, defined as the mean of the standard deviations of replicates, was 0.263%. Detection limits calculated from the estimate of precision (0.263%) or from the standard deviation of the corrected binding (0.298%) were similar. Two hundred thirty sera from insulin-treated patients were studied. Precision was plotted as a 'precision profile' and the detection limit calculated from the precision for binding of less than 1% [0.261%]; 88% of the sera were positive [cut-off 1.3%, p less than 0.01]. We conclude that corrected binding is normally distributed in antibody-negative sera and that an estimate of assay precision can be used to define the detection limit of the assay.

Clearance of immune complexes made of antiinsulin antibodies and 123I-insulin was studied with scintillation scanning in anesthetized rats. Complexes made with purified guinea pig antiinsulin IgG2 (cytophilic isotype) were rapidly cleared by the liver whereas those made with IgG1 remained in the plasma, as did 123I-labeled IgG1 or IgG2 of control animals. Hepatic clearance of insulin-antiinsulin IgG complexes was not inhibited by either an excess of insulin or decomplementation, thereby ruling out interaction with insulin and C3b receptors. Insulin and guinea pig antiinsulin serum or its purified IgG isotypes formed large aggregates exceeding 5 IgG. Antiinsulin antibodies of diabetics, mostly IgG1 and IgG3 (cytophilic isotypes), formed complexes that either remained in plasma (small aggregates) or were cleared by the liver (large aggregates). In conclusion, clearance of insulin-antiinsulin IgG complexes is probably mediated by Fc gamma receptors on macrophages and requires cytophilic subclass composition and formation of large IgG aggregates.

A combined pharmacokinetic/pharmacodynamic model was proposed to describe the pharmacokinetics of intravenously administered regular insulin (0.55 units/kg) in alloxan-induced diabetic dogs. Serum insulin concentrations were described by either a one- or two-compartment open model, in which a hypothetical effect compartment was linked to the central pharmacokinetic compartment, or in which the effect compartment was linked to the peripheral compartment. Response, as measured by percent change in glucose concentration from adjusted basal plasma concentrations, was modeled using the sigmoidal Emax effect model, a linear effects model, a log-linear effects model, and a gamma-linear effects model, using the insulin pharmacokinetic parameters to describe the amount in the hypothetical effect compartment. The results indicated that insulin pharmacokinetics are usually described by a two-compartment open model. Response to insulin was predicted more accurately in half of the dogs using the gamma-linear effects model in which the effect compartment was linked to the central compartment. In the other half of the dogs the best model was the sigmoidal Emax model in which the effect compartment was linked to the central pharmacokinetic compartment. The parameters in the latter model were correlated with each other and the confidence limits of the parameter estimates were larger than the parameters of the gamma-linear effects model. These models should be further investigated, but may offer an alternative method for distinguishing rapid insulin metabolism from insulin resistance.

Plasma and urinary C-peptide determinations in the discrimination between insulin-requiring and non-insulin-requiring diabetes were elevated in 61 adult diabetics. Specimens for C-peptide determinations were taken on two consecutive days: on the first day plasma C-peptide concentrations were determined before and 6 min after intravenous glucagon administration. On the second day 2- and 4-h urinary C-peptide excretion was measured after an individual breakfast. Results of urinary C-peptide analyses were expressed as molar concentration and also as molar quantity excreted (without any corrections and related to creatinine excretion). Glucagon-stimulated plasma C-peptide turned out to be a reliable criterion for the detection of insulin requirement. Sixty-nine per cent of diabetics included in this study were classifiable by basal plasma C-peptide concentrations. Two-hour postprandial urinary C-peptide/creatinine quotient turned out to be slightly less sensitive (89%) than the glucagon test (94%) and of equal specificity (96%). Glucagon-stimulated plasma C-peptide and postprandial urinary C-peptide excretion correlated significantly among insulin-requiring diabetics (r = 0.73), but not among non-insulin-requiring diabetics (r = 0.23). We regard determination of stimulated plasma C-peptide as a primary investigation for the direct assessment of endogenous insulin secretory reserves for clinical management decisions. Determination of postprandial urinary C-peptide is applicable in selected situations for non-invasive assessment of insulin secretion.

Sera containing insulin antibodies from 20 insulin-treated diabetic patients, sera containing insulin autoantibodies from 20 insulin-naive non-diabetic patients, and from 10 normal controls, were tested at 1:20 dilution in three different radioimmunoassays (RIA) and an enzyme linked immunosorbent assay (ELISA), using a highly purified human insulin ligand. The RIA using insulin radiolabelled at multiple sites detected insulin antibodies in 17/20 and insulin autoantibodies in 13/20 sera. The same RIA using A-14-monoiodinated insulin was sensitive to antibodies and autoantibodies in all the sera. The same RIA using sera after insulin extraction detected only 13/20 diabetic sera and 9/20 autoimmune sera as positive, owing to a substantial rise in non-specific binding of the control sera. ELISA was sensitive to insulin antibodies and autoantibodies in every case. When binding curves for ELISA and the most sensitive RIA were compared using serial dilutions of four insulin antibody containing sera and four insulin autoantibody containing sera, antibody titres varied from 1.1 to 3.8 times higher in ELISA, and autoantibody titres from 10.6 to 28.6 times higher in ELISA. These studies indicate that ELISA is more sensitive than RIA to insulin antibodies, and in particular to insulin autoantibodies.

To determine the influence of insulin antibodies (and their equilibrium kinetic properties) on the pharmacokinetics of insulin, we examined the relationship between insulin antibody binding and the initial rate of increase, time to peak, and return to baseline of therapeutic doses of insulin injected subcutaneously (0.15 U/kg) and the half-life, distribution space, and metabolic clearance rate of intravenously infused insulin (2 mU/kg/min) in insulin-treated patients with diabetes mellitus. Compared to age-weight-matched nondiabetic subjects, the diabetic subjects had reduced initial rates of increase (0.33 +/- 0.2 v 0.44 +/- 0.03 microU/mL/min, P less than 0.05), delayed time to peak (130 +/- 12 v 86 +/- 8 min, p less than 0.02), and prolonged return to baseline (485 +/- 37 v 313 +/- 13 min, P less than 0.01) of plasma free insulin levels after subcutaneous injection of insulin, and a prolonged half-life (19.8 +/- 5.8 v 4.3 +/- 0.3 min, P less than 0.02), increased distribution space (904 +/- 284 v 109 +/- 10 mL/kg, P less than 0.001), and augmented metabolic clearance rate (28.5 +/- 1.8 v 17.3 +/- 0.7 mL/kg/min, P less than 0.001) after intravenously infused insulin. All of these abnormal parameters were significantly correlated with binding of insulin to insulin antibodies at tracer insulin concentrations (Bo) and with the high affinity of insulin antibody binding sites as determined by Scatchard analysis. However, patients with 125I insulin antibody binding (Bo) less than 10 percent had normal or near normal plasma free insulin pharmacokinetics.(ABSTRACT TRUNCATED AT 250 WORDS)

A double-blind comparative trial of the effectiveness and antigenicity of semisynthetic human insulin (Novo) and highly purified (Monocomponent) porcine insulin was performed over a 12 month period in 20 diabetic subjects newly treated with insulin. Human insulin was shown to be indistinguishable from porcine insulin of comparable purity with respect to plasma glucose and glycosylated hemoglobin levels and insulin dose requirements. Human insulin was no less antigenic than porcine insulin; significant IgG-associated insulin binding activity was detected in six of the ten patients in the human insulin treated group and four of the ten patients in the porcine insulin treated group. In all patients, the binding activity was of low capacity. No adverse reactions to human insulin were noted. It is concluded that semisynthetic human insulin, like purified porcine insulin, is safe and effective. Although there may be advantages for human insulin in the setting of insulin allergy, this study does not indicate that human insulin has advantages over purified porcine insulin with respect to elicitation of antibodies of the IgG class.

Exposure of insulin solutions to elevated temperatures for prolonged periods of time will inevitably lead to chemical modifications of the hormone. Contact with different materials in dosing devices, other design-related factors and motion appear to be chemically more detrimental than storage in glass vials at the same temperature. An in vitro test, designed to mimic the in vivo situation, consisted of delivery of insulin at 37 degrees C while the device was constantly moved on a shaking apparatus. Insulin quality was assessed using high performance liquid chromatography. A polyethylenepolypropylene glycol-stabilized neutral human insulin solution (HOE 21 PH) was used. A single insulin derivative is the major modification product which, after passage of the complete infusion system, amounts to up to 10%. The biological potency of the derivative is indistinguishable from native insulin. Delivery of acidic insulin under implant conditions, leads to extensive and multiple insulin derivatization, even though the biological potency remains 95% after 4 weeks.

Insulin autoantibodies (IAA) were studied in newly diagnosed insulin-dependent diabetics before the start of insulin treatment and in unaffected identical twins of insulin-dependent diabetics. In 15 of the 40 (38%) diabetics and 27 of the 58 (47%) twins IAA levels exceeded those of 100 controls. Frequency of IAA in unaffected twins was not related to duration of diabetes in their affected twin. In 11 unaffected twins, IAA levels differed in two samples taken 1-12 years apart; IAA were detected at least once in all twins and in one on both occasions. IAA in the twins were not related to the presence of islet-cell antibodies or to HLA-DR 3 or 4. As the unaffected twins of longstanding diabetics are unlikely to develop diabetes, these observations suggest that IAA do not always presage diabetes and are probably not a consequence of the disease; they may reflect an inherited autoimmune tendency to diabetes.

A rise in blood glucose concentration at the end of the night, and consequent morning hyperglycaemia, are well recognized events in some diabetic patients. In 94 patients on twice daily insulin injections we have examined the prevalence and extent of morning hyperglycaemia, and its relation to control, insulin therapy, and insulin antibody levels. Blood glucose reached the highest level of the day before or after breakfast in 83% of patients, and in 50% this value was 2 mmol/l greater than any other time of day. Patients with higher fasting concentrations did not have worse blood glucose control over the rest of the day. No correlation was found between fasting blood glucose concentrations and the evening dose of intermediate acting insulin or the level of insulin antibodies. No consistent change in fasting blood glucose concentrations occurred with changes in antibody levels in patients switched between pork and beef insulin. Morning hyperglycaemia was as common with both insulin species. Pre- and post-breakfast hyperglycaemia is common and significant in insulin-treated diabetic patients. It is not directly related to diabetic control at other times of the day, and is independent of insulin species and insulin antibody levels.

A method is described for the measurement of T-cell cytotoxicity to purified preparations of insulin, their high molecular weight fractions and C-peptide derivatives. The technique assesses cellular cytotoxicity to membrane associated insulin in an autologous system. Peripheral blood lymphocytes from 19 type 1 diabetic patients were stimulated in vitro for 6 days with the pharmaceutical preparation(s) used therapeutically. At the end of this period, they were set up in a 4-h cytotoxicity assay against 51Cr labelled PHA blast transformed lymphocytes in the presence of therapeutic insulin, monocomponent insulin, B-component or C-peptide. The results presented here demonstrate cytotoxicity to insulin B-component in diabetic patients who have received pork insulin for 2-6 years. The possible reasons for this finding are discussed.

The disappearance rate of insulin antibodies was studied after cessation of insulin treatment which had been given for 3 months to 6 years in 42 Type 2 (non-insulin-dependent) diabetic patients. Insulin antibodies were measured before and 15 days after interruption of insulin treatment, and every 30 days until the disappearance of insulin antibodies. The mean +/- SD value of insulin binding in the entire group before the interruption of insulin treatment was 32 +/- 14%. There was no relationship between the antibody level at that time and the duration of insulin treatment. However, the insulin antibody level was significantly higher in 17 diabetic patients on an insulin dose of greater than 20 U/day (p less than 0.02) than in 25 on an insulin dose of less than 20 U/day (39 +/- 13% versus 28 +/- 12%). A positive correlation was found between initial insulin binding and the time required for it to fall below 10% (r = 0.74). Antibodies were absent 60 days after discontinuing insulin treatment in eight of ten subjects presenting with initial binding of less than 20%. In contrast, in only two of 12 patients with an initial binding of greater than 40% were insulin antibodies detectable 150 days after discontinuation of insulin therapy. Disappearance of insulin antibodies sometimes took up to 1 year and occasionally even more than 2 years.

In 25 insulin-dependent diabetics, 14 managed by conventional insulin injection treatment (CIT) and 11 treated by continuous subcutaneous insulin infusion (CSII), there was a highly significant correlation between urinary insulin excretion rate (IER) per 1.73 m2 and mean serum free insulin concentration (r = 0.73, p less than 0.001), measured over a 24 h period. Urinary IER and mean daily serum free insulin levels were significantly higher in diabetics than in non-diabetics. CSII-treated patients had significantly lower mean 24 h plasma glucose levels than CIT-treated patients despite similar values of urinary IER and mean daily serum free insulin in the two groups. Urinary IER may be a useful indicator of average insulinaemia in large scale studies, avoiding the problems of multiple blood sampling and immunoassay in the presence of anti-insulin antibodies.

Bovine insulin was labelled with carrier-free Na 123I and the species labelled on the A14 tyrosyl residue was purified by reverse phase high pressure liquid chromatography. After sterilization by filtration through a 0.22 micrometer millipore filter, the labelled hormones (123I-insulin) was injected intravenously into normal volunteers and into two insulin-immunized insulin-dependent diabetic patients. Patient 1 was treated with 74 U insulin/day, whereas, despite daily insulin doses greater than 100 U, patient 2 remained hyperglycaemic and continued to lose weight. Plasma insulin binding capacity was 10 U/l in patient 1 and greater than 20 U/l in patient 2. In both diabetic patients, heart activity (i.e. blood pool) decreased more slowly than in control patients, an observation consistent with the reduction of plasma insulin clearance rate in immunized patients. Kidney activity was decreased in patient 1 and undetectable in patient 2, suggesting that most of the circulating 123I-insulin was bound to antibodies and not filtrable. Finally, in both patients, the profile of liver radioactivity was markedly altered. Maximum liver activity was delayed and the liver image persisted for greater than 45 min. According to analysis of the time activity profile in various organs, we consider that specific antibodies act predominantly either as 'carrier-proteins' retarding the action of soluble insulin as in patient 1 or as insulin 'scavengers', the immune complexes being cleared by the reticulo-endothelial system as in patient 2.

Sera of patients with type 1 and type 2 diabetes were examined for IgA- and insulin-containing immune complexes (IgA-ICs, ICs-insulin) using a solid-phase anti-C3 enzyme immunoassay. IgM-ICs and IgG-ICs were also investigated. IgA-ICs were detected in 6 of 26 type 1 diabetics, in 9 of 25 insulin-treated type 2 diabetics, and in 8 of 34 type 2 diabetics on oral hypoglycemic agents, but only in 2 sex- and age-matched controls. ICs-insulin was detected in 9 of 25 type 1 diabetics and in 1 of 19 insulin-treated type 2 diabetics, irrespective of the time of insulin treatment. ICs-insulin did not appear to be related to the presence of microangiopathy. IgA-ICs were found to be associated with the presence of microangiopathy, suggesting that they may play a role in the pathogenesis of the late diabetic complications.

Two methods of synthesis of human insulin have been developed to the stage of commercial production. One entails synthesis of human insulin by bacteria ("biosynthetic" human insulin) and one entails the conversion of pork insulin into human insulin by an amino acid substitution ("semisynthetic" human insulin). Both forms of human insulin have been shown to be safe, and to have similar efficacy and pharmacokinetics to purified pork insulin. Human insulin given by subcutaneous injection has been shown to elicit antibody formation in man, although the extent of this may be slightly less than in the case of purified pork insulin, and is certainly less than in the case of beef insulin and beef-pork combinations. The clinical significance of these differences in immunogenicity is doubtful. However there are rare defined situations, such as allergy to pork insulin and antibody-mediated insulin resistance where human insulin has been shown to have advantages over purified pork insulin. Bacterial synthesis of insulin provides an assured supply for the world's future needs, and it is conceivable that further technological refinement may make bacterial synthesis cheaper than extraction and purification of animal insulins. However, current evidence provides no basis for recommending human rather than purified animal insulin for the routine management of insulin dependent diabetes.

Insulin withdrawal studies were performed in 12 Type 1 (insulin-dependent) C-peptide negative diabetic patients with low to moderate insulin antibody levels, to assess the biological availability of antibody-bound insulin and its clinical significance. There was a highly significant correlation between the extent to which the free insulin concentration was maintained during the period of insulin withdrawal and both the level of insulin-binding by serum and the total insulin concentration at the start of the study. During insulin withdrawal, the patients who best maintained their circulating free insulin levels showed the smallest increases in blood glucose and 3-hydroxybutyrate concentrations. We conclude that antibody-bound insulin is available for physiological action, and that in those individuals with moderate antibody concentrations it is capable, in the fasting state, of maintaining free insulin levels. In these circumstances insulin antibodies are behaving as simple carrier proteins.

24 diabetic patients stabilised on conventional bovine insulins and possessing insulin antibodies underwent a study of the immunological and clinical consequences of changes in both purity and species of their insulin preparations. After a 2-month run-in period patients were treated for 3 consecutive 4-month periods with (a) purified bovine insulin (20-40 ppm proinsulin), (b) highly purified porcine insulin, and (c) semisynthetic human insulin, without elective dose changes. Mean insulin antibody levels changed little on purified bovine insulin (22.2 leads to 23.4 micrograms/l) but fell on highly purified porcine (23.4 leads to 12.9 micrograms/l) and remained much the same on Semi-Synthetic human insulin. In contrast, C-peptide antibodies fell significantly and continuously throughout the study. The slower rate of fall in C-peptide antibody levels is likely to be due to the prolonged half-life of circulating exogenous proinsulin in the presence of insulin antibody. Although insulin dose remained constant the incidence of hypoglycaemic episodes did not increase and glycosylated haemoglobin levels rose significantly when patients were on porcine insulin. The deterioration in diabetic control may have been due to greater temporal mismatch between insulin needs and insulin availability with pork or human insulin than with beef insulins, and to reduced insulin antibody levels. The use of purer insulins which more closely resemble the human form can cause a significant reduction in levels of insulin and C-peptide antibodies. Although these changes may have other benefits they do not necessarily produce better diabetic control. Subcutaneous insulin regimens need to be tailored to the individual patient and, indirectly, to his antibody status.

In the past 10 years the techniques of gel filtration and ion exchange chromatography have made available insulins of markedly enhanced purity. These highly purified insulins have made immunological insulin resistance a rarity, and result in absent or clinically insignificant levels of insulin antibodies in insulin-treated diabetics. Insulin allergy has not been reported with highly purified insulins alone, and is rare even when the patient has previously received recrystallised insulin. Generalised allergic reactions to insulin and insulin resistance are associated with the enhanced immunological reaction to intermittent insulin therapy. The use of highly purified insulins for short courses of treatment is therefore mandatory, particularly in patients with infections. Injection-site lipoatrophy, a relatively common occurrence with the older insulins, disappears on changing to highly purified preparations. Following a change to highly purified insulins, insulin dose requirements will fall gradually with insulin antibody levels. When switching from conventional beef to highly purified pork insulins, a more immediate change in dose requirements may occur so that prospective reductions in insulin dose are indicated. It is still uncertain whether moderate levels of insulin antibodies are associated with any difference in metabolic control. This is partly a reflection of difficulties in measuring diabetic control, and partly a lack of properly designed studies. Current insulins of both older and newer types give plasma insulin profiles that are far from physiological. Insulin antibodies cross the placenta and may contribute to increased fetal insulin secretion and neonatal hypoglycaemia. Pre-pregnant patients should be changed to the newer preparations. Highly purified insulins cost little more than conventional insulins in the free market, and should be used in all newly diagnosed insulin-requiring diabetics. More recently, human insulin has become available through both DNA recombinant technology and amino acid substitution techniques. It has proved to have identical characteristics to pork insulin both in vitro and in normal subjects. Clinical trials with human insulin are at present in progress.