Kinetic models for plasma disappearance of insulin in normal subjects

Insulin murder and the case of Colin Norris

Although insulin is an essential medicine and a life-saving drug, it has also been incriminated in many poisoning deaths; accidental, suicidal and some with malicious intent. Overdosing with insulin precipitates a life-threatening state of hypoglycemia and if untreated leads to coma, irreversible brain damage and death. Normally, the pancreatic β-cells secrete equimolar amounts of insulin and C-peptide into the portal venous blood, although under physiological conditions the plasma concentration ratio (insulin/C-peptide) is less than unity, because insulin is more susceptible to hepatic first-pass metabolism. A high ratio of insulin/C-peptide in plasma from a poisoned patient is compelling evidence that pharmaceutical insulin was administered, which does not contain C-peptide. The analysis of insulin and C-peptide was traditionally done by immunoassay methods (RIA and/or ELISA), although high resolution LC-MS/MS is more suitable for forensic purposes and permits the identification of insulin analogues. Use of insulin as a murder weapon is exemplified by the case of Colin Norris, a male nurse found guilty of murdering four elderly patients and the attempted murder of a fifth by injecting them with insulin. However, the prosecution evidence against Norris was mainly circumstantial and hearsay. Toxicological evidence against Norris consisted of a high insulin/C-peptide concentration ratio in plasma from one of the victims. This analysis was done by an immunoassay method at a clinical laboratory and not a forensic laboratory. Analytical procedures, including chain-of-custody routines, are more stringent at forensic laboratories. Since his conviction, some of the medical evidence against Norris has been called into question, especially the prevalence of spontaneous attacks of hypoglycemia in elderly and frail patients with co-morbidities.

Five-compartment model of insulin kinetics and its use to investigate action of chloroquine in NIDDM

We have constructed a five-compartment model of insulin kinetics. The model structure was chosen to reflect insulin distribution in systemic plasma, hepatic plasma, and interstitial fluid and insulin binding to the liver and peripheral receptors, and it included receptor-mediated and non-receptor-mediated insulin degradation. Model parameters were estimated from plasma insulin concentrations measured during hyperinsulinemic euglycemic clamp studies. In the fasting condition, the model-derived mean residence time of endogenously secreted insulin was 71 min, of which 62 min were spent bound to the liver receptor, 6 min bound to the peripheral receptor, 2 min circulating in hepatic or systemic plasma, and 1 min in the interstitial fluid. More than 80% of total insulin was bound to the liver receptor, indicating that the liver is by far the largest insulin reservoir. The model was employed to assess the effect of chloroquine on insulin kinetics in patients with non-insulin-dependent diabetes mellitus (NIDDM). Chloroquine significantly altered parameter vector. However, the mean residence times of insulin in the system and in the periphery were not affected, indicating that the beneficial effect of chloroquine in patients with NIDDM under conditions of euglycemia could not be attributed to changes in insulin kinetics.

Insulin clearance from plasma in type I (insulin-dependent) diabetic patients: influence of glycaemic level

The influence of the actual glycaemic level on the kinetics of insulin disappearance from plasma was studied in 16 type I diabetic patients. Constant intravenous infusion of insulin (1-5 mU.kg-1.min.-1) was used to achieve different levels of steady state plasma free insulin concentrations, while the blood glucose level was clamped at normoglycaemia, 4.1 +/- 0.1 mmol/l (mean +/- S.E.M.), and mild hyperglycaemia, 8.2 +/- 0.1 mmol/l. The experimentally determined data were compared using a previously validated model of first order kinetics for insulin disappearance from plasma in diabetic patients. At the physiological insulin concentration range the median clearance rate of insulin was 23 ml.kg-1.min-1 (range 16-26) at normoglycaemia and 23 ml.kg-1.min-1 (19-35) at hyperglycaemia (P = 0.45). At supraphysiological levels, insulin was cleared at a lower rate at normoglycaemia than at hyperglycaemia. No correlation was observed between the insulin clearance rate and the duration of diabetes or the haemoglobin A1C level (both Spearman's rho = 0.08). In conclusion, the insulin clearance rate from plasma is independent of the actual glycaemic level and hardly influenced by the long-term glycaemic level in type I diabetic patients.

Reduced insulin clearance in normal subjects due to extreme hyperinsulinemia

Insulin clearance was assessed in five normal subjects infused with insulin at a rate of 10 mU.kg-1.min-1 for 12-16 hours, which produced insulin levels of 1500-2000 microU/ml (approximately 10(-8) M). This level approximates the Kd of low affinity insulin binding sites, whereas previous clearance studies had been performed at insulin concentrations of 10(-9) M or less, approximating the Kd of the high affinity insulin receptor. The metabolic clearance rate for insulin during the infusion averaged 214 +/- 29 ml.min-1.m-2, which is half of that reported previously when lower insulin levels were achieved. Upon termination of the insulin infusion, the disappearance of insulin was markedly prolonged with an average "half-life" of 62 minutes. The rapidity with which hyperinsulinemia altered clearance suggested that down-regulation of insulin receptors was probably not the explanation for the reduced clearance. To elucidate the cause for the observed decrease in insulin clearance, five additional subjects were studied. If insulin was infused for 3.0-4.5 hours, the half-life of insulin disappearance was intermediate between that for an insulin bolus dose and that for a 12-16 hour insulin infusion. Administration of an insulin bolus dose at the end of a 12-hour infusion, while the insulin concentration was still approximately 10(-8) M, or 140 min later, when the insulin concentration was 10(-9) M, was followed by rapid disappearance with half-lives of 1.5 and 6-8 minutes, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)

Kinetics of insulin disappearance from plasma in cortisone-treated normal subjects

The effect of glucocorticoid excess on insulin disappearance from plasma was examined in eight normal men during cortisone treatment (50 mg orally twice daily for 4 d) and in the absence of any medication (control) in random order. Constant infusion of insulin (1-5 mU/kg/min) was used to achieve different levels of steady state plasma insulin concentrations; normoglycaemia was preserved by a glucose clamp technique. The experimentally determined data were compared using a previously validated model of saturation kinetics. The amount of glucose required to maintain normoglycaemia during the insulin infusions was significantly less in the cortisone study than in the control study, while the parameter estimates for the kinetics of insulin disappearance from plasma were unaffected by cortisone. Thus, insulin action and insulin kinetics in the steady state are dissociated in normal subjects rendered insulin resistant by short-term cortisone treatment.

Kinetic models for insulin disappearance from plasma in type I diabetic patients

We have tested whether our previous finding in normal subjects that the disappearance of insulin from plasma obeys saturation kinetics alone also applies to type I diabetic patients. In six long-term diabetic patients steady state plasma insulin concentrations resulting from constant insulin infusion at different rates were compared with the predictions of three models for the kinetics of insulin in plasma. The models allowed the existence of non-saturable (first order equation) or saturable (Michaëlis-Menten equation) mechanisms, or both. The minimal acceptable model included saturation kinetics alone in four subjects and first order kinetics alone in two subjects. The clearance of insulin in diabetic patients, calculated from the best fitting model, was 18.0 (median, range 10.0-23.7) ml X kg-1 X min.-1 versus 25.0 (18.6-47.1) ml X kg-1 X min.-1 in six normal subjects (2p = 0.008). Insulin thus disappears from plasma at a lower rate in diabetic patients than in normal subjects at physiological plasma concentrations.

Kinetics of human and porcine insulins in normal and type I diabetic subjects

Human and porcine insulin were infused intravenously at various rates into 4 normal and 6 Type I diabetic subjects, using a double-blind cross-over design and a euglycaemic glucose clamp, to study the relationship between the steady state plasma free insulin concentration and its plasma disappearance rate. By mathematical model validation procedures both human and porcine insulin were found to obey saturation kinetics in normal subjects and first order kinetics in diabetic subjects in the insulin concentration range studied (0-2 nmol/l). No differences in parameters were observed between the two types of insulin in the study groups. The median clearance rate of insulin in normal subjects was 31 ml.kg-1.min-1 at infinitesinal plasma insulin concentrations versus 21 ml.kg-1.min-1 in the diabetic subjects. Thus, at physiological plasma concentrations both human and porcine insulin disappear faster via the saturable mechanism(s) found in normal subjects than via the apparently linear mechanism(s) found in diabetic subjects.