Emerging Treatments and Technologies O R I G I N A L
A R T I C L E
Impact of Diabetic Nephropathy on Pharmacodynamic and Pharmacokinetic Properties of Insulin in Type 1 Diabetic Patients KLAUS RAVE1 TIM HEISE1 ANDREAS PFU¨ TZNER2
LUTZ HEINEMANN1 PETER T. SAWICKI3
OBJECTIVE — To quantify pharmacokinetic and pharmacodynamic properties of regular insulin and insulin lispro in type 1 diabetic patients with and without overt diabetic nephropathy. RESEARCH DESIGN AND METHODS — In this double-blind, two-way cross-over, euglycemic (5 mmol/l) glucose clamp study, we investigated the metabolic response to subcutaneous injections of regular insulin and insulin lispro (0.2 U/kg) in 12 type 1 diabetic patients with overt diabetic nephropathy (proteinuria ⬎500 mg/24 h and/or serum creatinine ⬎1.5 mg/dl; NP group) and in a control group of 12 type 1 diabetic patients with normal renal function (DC group). RESULTS — Peak plasma free insulin levels with insulin lispro (359 [NP] vs. 254 pmol/l [DC]) were higher and time to maximal insulin concentrations (85 [NP] vs. 99 min [DC]) shorter than with regular insulin (213 [NP] vs. 144 pmol/l [DC]; 118 [NP] vs. 153 min [DC]) in both patient groups. Overall insulin levels for regular insulin and for insulin lispro were higher in patients with overt diabetic nephropathy compared with control patients. Time to maximal metabolic effect was shorter with insulin lispro than with regular insulin in both patient groups (102 vs. 191 min [NP]; 105 vs. 172 min [DC]). The overall metabolic effect of regular insulin but not of insulin lispro was lower in patients with diabetic nephropathy than in diabetic control patients (967 vs. 1,510 mg/kg, respectively). CONCLUSIONS — Although insulin levels are higher in patients with overt diabetic nephropathy, the metabolic response to regular insulin is reduced. Insulin lispro maintains its characteristic pharmacokinetic and pharmacodynamic properties in patients with overt diabetic nephropathy. Diabetes Care 24:886 – 890, 2001
failure (5,6) and mild to moderate renal impairment (7); however, until now, it has not been studied in patients with diabetic nephropathy. Insulin sensitivity, on the other hand, is attenuated in patients with renal impairment, not only in those with endstage renal failure, but also in patients in earlier stages of renal impairment (7–9). However, to date the metabolic consequences of reduced insulin clearance and diminished insulin sensitivity on pharmacokinetic and pharmacodynamic properties of regular insulin or rapid acting insulin analogues have never been studied in type 1 diabetic patients. In view of the diminished clearance of insulin in patients with altered renal function, the use of a rapid-acting insulin analogue might theoretically be advantageous because of its shorter duration of action, leading to a reduced incidence of hypoglycemic episodes (10,11). Therefore, the aim of this study was to quantify pharmacokinetic and pharmacodynamic summary measures for regular insulin and insulin lispro in type 1 diabetic patients with and without overt diabetic nephropathy. RESEARCH DESIGN AND METHODS
From the 1Profil Institute for Metabolic Research, Neuss; 2Institut fu¨r Klinische Forschung und Entwicklung, Mainz; and 3St. Franziskus Hospital, Cologne, Germany. Address correspondence and reprint requests to Dr. Klaus Rave, Profil Institute for Metabolic Research GmbH, Stresemannallee 6, 41460, Neuss, Germany. E-mail
[email protected]. Received for publication 31 October 2000 and accepted in revised form 18 January 2001. Abbreviations: AUC, area under the curve; GIR, glucose infusion rate; INS, insulin. A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion factors for many substances.
Patients For this study, 12 type 1 diabetic patients fulfilling the following inclusion criteria were recruited: age 20 –50 years; GHb ⬍10%; overt diabetic nephropathy (defined according to the modified Mogensen criteria: proteinuria ⬎500 mg/24 h on repeated measurements and/or serum creatinine ⬎1.5 mg/dl [122 mol/l] with no clinical or laboratory evidence of renal or urinary tract disease other than diabetic glomerulosclerosis) (12); negative insulin-antibody screening; and a negative pregnancy test in women with childbearing potential. The control group consisted of 12 type 1 diabetic patients
886
DIABETES CARE, VOLUME 24, NUMBER 5, MAY 2001
I
nsulin therapy in type 1 diabetic patients with overt diabetic nephropathy is characterized by two features: while suffering from a considerably increased risk of severe hypoglycemic episodes (1,2), they show poorer metabolic control than diabetic patients with normal renal
function (3). The increased hypoglycemic risk may be partially explained by the decline in renal insulin clearance, as 30 – 80% of systemically circulating insulin is removed by the kidneys (4). A reduction of insulin clearance has been described in nondiabetic patients with end-stage renal
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
Rave and Associates
Table 1—Baseline characteristics of type 1 diabetic patients with and without overt diabetic nephropathy
Characteristic
Patients with nephropathy
Patients without nephropathy
P
Sex (female/male, n) Age (years) Weight (kg) BMI (kg/m2) Duration of diabetes (years) HbA1c (%) Serum creatinine (mol/l) GFR (ml/min/[1.73 m2]) Proteinuria (mg/24 h) Daily insulin dosage (U/day)
6/6 41 ⫾ 5 72 ⫾ 16 24.7 ⫾ 4.0 29 ⫾ 9 8.1 ⫾ 0.8 177 ⫾ 81 54 ⫾ 28 1382 ⫾ 864 43 ⫾ 15
3/9 34 ⫾ 10 76 ⫾ 10 24.0 ⫾ 2.5 18 ⫾ 5 7.2 ⫾ 0.8 77 ⫾ 10 90 ⫾ 5 34 ⫾ 12 49 ⫾ 9
⬍0.05 ⬍0.01 0.45 0.64 ⬍0.01 ⬍0.05 ⬍0.01 ⬍0.01 ⬍0.01 0.39
Data are means ⫾ SD or frequencies. GFR, glomerular filtration rate.
without nephropathy; that is, their protein excretion and endogenous creatinine clearance were within the normal range (Table 1). Patients with overt diabetic nephropathy were older and had a longer duration of diabetes than patients without nephropathy (29 vs. 18 years). The mean HbA1c was higher in patients with diabetic nephropathy, whereas the mean daily insulin dosage was comparable for both groups of patients. According to the inclusion criteria, patients with overt diabetic nephropathy showed higher serum creatinine levels and lower glomerular filtration rates. This randomized study was performed in a double-blind fashion (study medication was applied by a person otherwise not involved in the study). The study protocol was approved by the local ethical committee and was carried out according to the guidelines of Good Clinical Practice. After detailed oral and written explanation of the study objectives, all patients gave their written informed consent to participate in the study. After a fasting interval of ⱖ12 h, patients arrived at our institute in the morning (⬃8:00 A.M.) and were connected to a Biostator (Life Science Instruments, Elkhart, IN). The left hand was placed in a box that was warmed to an air temperature of 55°C to allow the sampling of arterialized venous blood. Fasting blood glucose of the patients was lowered during the first 60 min of a 4-h run-in period by means of an insulin infusion to a target level of 5.0 mmol/l (90 mg/dl). After having reached this target level, blood glucose was kept constant DIABETES CARE, VOLUME 24, NUMBER 5, MAY 2001
over the next 180 min by adapting the dosage of a continuous intravenous insulin infusion. At the end of the run-in period (⬃12:00 P.M.), patients received subcutaneous injections of 0.2 U/kg body wt of either regular human insulin (Humulin U 100; Lilly Deutschland, Bad Homburg, Germany) or insulin lispro (Humalog U 100; Lilly Deutschland); patients with nephropathy received 14.3 ⫾ 3.3 U (mean ⫾ SD) and diabetic control subjects received 15.4 ⫾ 2.0 U into a lifted skinfold in the abdominal region. Thereafter, an 8-h euglycemic glucose clamp was established, as described in detail elsewhere (13). Mean blood glucose levels during glucose clamps in patients with overt diabetic nephropathy were 5.0 ⫾ 0.3 mmol/l with regular insulin and 5.1 ⫾ 0.4 mmol/l with insulin lispro. For diabetic control patients, mean blood glucose was 5.0 ⫾ 0.3 mmol/l with regular insulin and 5.2 ⫾ 0.4 mmol/l with insulin lispro during the glucose clamp experiments. All patients were on an intensified insulin therapy before the study and were instructed to take their last short-acting insulin injection in the evening and their long-acting insulin injection at the latest 24 h before the glucose clamp experiments. Insulin pump⫺treated patients were asked to stop the insulin infusion at 6:00 A.M. on the day of the glucose clamp. To exclude ongoing carbohydrate absorption during the time of the glucose clamp, which may occur in patients with severely impaired gastric emptying, a modified 13C-labeled octanoid acid breath test was performed before every
glucose clamp (14). In brief, patients gave a baseline breath sample in an aluminumcoated bag in the evening preceding the glucose clamp. Thereafter, they ingested a test meal including a scrambled egg with the yolk supplemented with 0.1 ml 13 C-octanoid acid (Wagner Analysen Technik, Bremen, Germany) at 8:00 P.M. Additional breath samples were collected the next morning and after completion of the glucose clamp run-in period. All three samples were analyzed for their 13C/12C ratio using an infrared isotope analyzer (Iris Infrared Isotope Analyzer; Wagner Analysen Technik). A 13C/12C ratio ⬎2.5% above baseline would be regarded as a sign of an ongoing carbohydrate absorption attributable to delayed gastric emptying. In those cases, that particular experiment would have to be excluded from further analysis. However, this was never necessary. Blood glucose was measured by a glucose oxidase method (Super G Ambulance, RLT Mo¨hnesee, Germany). HbA1c was measured with high-performance liquid chromatography (Diamat, Munich, Germany; reference range, 4.3– 6.1%). Proteinuria was determined by means of a laser turbidimetric method (Laserphotometer pm4, Lasermed GmbH, Cologne, Germany) in 24-h urine samples. The glomerular filtration rate was assessed by endogenous creatinine clearance in 24-h urine samples, corrected to 1.73 m2 body surface. Plasma concentrations of free regular insulin and free insulin lispro were determined after polyethylene glycol precipitation of antibody bound insulin by means of a commercial radioimmunoassay kit (DPC Coat-A-Count Insulin Radioimmunoassay; DPC, Bad Nauheim, Germany; inter- and intraassay coefficients of variation 15.0 and 6.6%, respectively). This assay has recently been validated for the determination of free human insulin and free insulin lispro with similar sensitivity (15). Insulin-antibodies were determined by means of a radioligand assay (CentAK, Medipan Diagnostika, Selchow, Germany). Statistics A polynomial function of sixth order was fitted to the individual glucose infusion rate (GIR) and free insulin profiles, allowing the calculation of pharmacokinetic and pharmacodynamic summary measures. The area under the curve (AUC) for 887
Diabetic nephropathy and insulin
Figure 1—A: Plasma free insulin levels after subcutaneous injection of 0.2 U/kg regular insulin (䡺) and insulin lispro (F) in type 1 diabetic patients with overt diabetic nephropathy. Thick lines indicate free insulin concentrations; thin lines indicate fitted polynomial function of sixth order. Data represented are means ⫾ SE. B: Plasma free insulin levels after injection of 0.2 U/kg regular insulin (f) and insulin lispro (E) in type 1 diabetic patients without diabetic nephropathy. Thick lines indicate free insulin concentrations; thin lines indicate fitted polynomial function of sixth order. C: Fitted polynomial function of sixth order to glucose infusion rate after subcutaneous injection of 0.2 U/kg regular insulin (heavy black line) and insulin lispro (dotted line) in type 1 diabetic patients with overt diabetic nephropathy. Thin lines indicate glucose infusion rate original registration. D: Fitted polynomial function of sixth order to glucose infusion rate after subcutaneous injection of 0.2 U/kg regular insulin (heavy black line) and insulin lispro (dotted line) in type 1 diabetic patients without diabetic nephropathy. Thin lines indicate glucose infusion rate original registration.
GIR and free insulin was calculated by the trapezoid rule. Wilcoxon’s signed-rank test was used for within-group statistical analyses, and the Mann-Whitney U test was applied for between-group comparisons. Multiple regression analyses were performed by using the pharmacokinetic and pharmacodynamic summary measures as response variables and nephropathy status, diabetes duration, and HbA1c as explanatory factors. The aim of these analyses was to assess the impact of diabetic nephropathy adjusted to diabetes duration and HbA1c on the response variables. Because of the exploratory character of the study, no adjustments for multiple comparisons were performed. With multiple regression analyses, only the overall metabolic activity for regular insulin in patients 888
without diabetic nephropathy was significantly associated with a parameter other than the nephropathy status (i.e., HbA1c). For statistical calculations, StatView 5.0 (SAS Institute, Cary, NC) was used. P ⬍ 0.05 was considered statistically significant. Results are given as means ⫾ SD throughout the text and as means ⫾ SE in the figure. RESULTS — Peak free plasma insulin levels (INSmax) were higher and time to maximal insulin concentration (tINS max) was shorter after subcutaneous administration with insulin lispro than after injection of regular insulin in diabetic patients with (P ⬍ 0.002 for both) and without nephropathy (P ⬍ 0.02 and P ⬍ 0.006, respectively) (Fig. 1A and B; Table 2). Accordingly, insulin AUCs for the first 3 h
after injection were higher with insulin lispro than with regular insulin in both patient groups (P ⬍ 0.003 and P ⬍ 0.005, respectively). When comparing nephropathic patients with the diabetic control subjects, both with regular insulin and with insulin lispro, peak free plasma insulin levels and the insulin AUCs for the intervals 0 –3 h and 0 – 8 h were higher in patients with overt diabetic nephropathy (regular insulin: P ⬍ 0.008 for free plasma insulin, P ⬍ 0.007 for 0 –3 h AUC, and P ⬍ 0.007 for 0 – 8 h AUC; insulin lispro: P ⬍ 0.008 for free plasma insulin, P ⬍ 0.003 for 0 –3 h AUC, and P ⬍ 0.005 for 0 – 8 h AUC). In contrast, time to late half maximal insulin concentration (late tINS 50%) for regular insulin and insulin lispro did not differ DIABETES CARE, VOLUME 24, NUMBER 5, MAY 2001
Rave and Associates
Table 2—Pharmacokinetic and pharmacodynamic summary measures in 12 type 1 diabetic patients with overt diabetic nephropathy and 12 diabetic patients without diabetic nephropathy Patients with nephropathy Regular insulin
Insulin lispro
Patients without nephropathy Regular insulin
Insulin lispro
INSmax (pmol/l) 213 ⫾ 64† 359 ⫾ 99*† 144 ⫾ 86 254 ⫾ 88* tINS max (min) 118 ⫾ 18† 85 ⫾ 14* 153 ⫾ 24 99 ⫾ 16* Early tINS 50% (min) 34 ⫾ 8† 28 ⫾ 6*† 46 ⫾ 18 38 ⫾ 12 Late tINS 50% (min) 270 ⫾ 76 171 ⫾ 24* 283 ⫾ 42 183 ⫾ 15* AUCINS 0–3 h (nmol/l) 28.8 ⫾ 8.9† 43.2 ⫾ 9.2*† 19.0 ⫾ 12.0 31.8 ⫾ 10.9* AUCINS 0–8 h (nmol/l) 51.3 ⫾ 15.1† 61.0 ⫾ 14.9*† 37.3 ⫾ 25.8 41.5 ⫾ 15.4 GIRmax (mg 䡠 kg⫺1 䡠 min⫺1) 3.4 ⫾ 1.5† 5.4 ⫾ 2.6† 5.9 ⫾ 3.1 7.1 ⫾ 3.4 tGIR max (min) 191 ⫾ 70 102 ⫾ 36* 172 ⫾ 38 105 ⫾ 20* Early tGIR 50% (min) 46 ⫾ 23 30 ⫾ 11* 48 ⫾ 24 31 ⫾ 14* Late tGIR 50% (min) 343 ⫾ 85 244 ⫾ 73* 318 ⫾ 46 214 ⫾ 21* AUCGIR 0–3 h (mg/kg) 472 ⫾ 243† 719 ⫾ 339* 757 ⫾ 425 952 ⫾ 428 AUCGIR 0–8 h (mg/kg) 967 ⫾ 364† 1037 ⫾ 337 1510 ⫾ 611 1280 ⫾ 556 AUCGIR 4–8 h (mg/kg) 495 ⫾ 212 318 ⫾ 178* 753 ⫾ 373 328 ⫾ 173* 21.1 ⫾ 10.3† 18.4 ⫾ 6.9† 46.2 ⫾ 24.5 34.2 ⫾ 19.9 AUCGIR 0–8 h/AUCINS 0–8 h *P ⬍ 0.05 for within-group comparison (patients with overt diabetic nephropathy or diabetic patients without nephropathy) between insulin lispro and regular insulin; †P ⬍ 0.05 for between-group comparison for regular insulin or for insulin lispro. INSmax, maximal plasma free insulin concentration; tINSmax, time to maximal free insulin concentration; early/late tINS 50%, time to early/late half-maximal insulin concentration; AUCINS 0 –3 h, AUC for free insulin concentration 0 –3 h; GIRmax, maximal glucose infusion rate; tGIR max, time to GIRmax; early/late tGIR 50%, time to early/late half-maximal glucose infusion rate; AUCGIR 0 –3 h, AUC for glucose infusion rate 0 –3 h.
significantly between the group of patients with and without nephropathy. Comparing the metabolic activity of regular insulin with insulin lispro, the maximal metabolic effect of subcutaneously injected insulin lispro was higher (P ⬍ 0.003) than the effect of regular insulin in patients with overt diabetic nephropathy, and tended to be higher (P ⫽ 0.078) in patients without nephropathy (Fig. 1C and D; Table 2). Time to maximal metabolic effect (tGIR max), time to early half maximal activity (early tGIR 50%), and time to late half maximal activity (late tGIR 50%) were significantly shorter with insulin lispro than with regular insulin in both patient groups (patients with nephropathy: P ⬍ 0.003 for tGIR max, P ⬍ 0.03 for early tGIR 50%, and P ⬍ 0.003 for late tGIR 50%; diabetic control subjects: P ⬍ 0.003 for tGIR max, P ⬍ 0.02 for early tGIR 50%, and P ⬍ 0.003 for late tGIR 50%). The metabolic activity during the first 3 h after injection (AUCGIR 0 –3 h) with insulin lispro was higher than with regular insulin in patients with overt diabetic nephropathy (P ⬍ 0.004). However, the metabolic activity 4 – 8 h after administration of insulin (AUCGIR 4 – 8 h) was higher for regular insulin than for insulin lispro DIABETES CARE, VOLUME 24, NUMBER 5, MAY 2001
in both patient groups (P ⬍ 0.02 for patients with nephropathy vs. ⬍0.002 for diabetic control subjects). When comparing patients with overt diabetic nephropathy with diabetic control subjects, the maximal metabolic effect (GIRmax) for regular insulin was lower (P ⬍ 0.03) and metabolic activity 3 and 8 h after administration of insulin was smaller (P ⬍ 0.05 and P ⬍ 0.05, respectively) in patients with overt diabetic nephropathy than in patients without renal impairment. It is noteworthy that the lower metabolic effect of regular insulin in patients with overt diabetic nephropathy was in contrast to the higher insulin levels observed in these patients. Glucose consumption per plasma insulin level (Table 2) was significantly lower in diabetic patients with overt nephropathy than in control subjects with insulin lispro or regular insulin (P ⬍ 0.008 and P ⬍ 0.03, respectively). CONCLUSIONS — This study demonstrates for the first time that plasma insulin levels—the maximal insulin concentration and insulin levels during 8 h after subcutaneous injection of regular insulin and insulin lispro—were higher
in patients with diabetic nephropathy than in diabetic control patients without nephropathy. This finding suggests a 30 – 40% reduction of insulin clearance in patients with overt diabetic nephropathy compared with diabetic patients without nephropathy. Paradoxically, in spite of higher circulating insulin levels, both the peak metabolic effect and the overall metabolic activity of regular insulin were diminished in patients with overt diabetic nephropathy when compared with diabetic patients without renal impairment. Because of the lower metabolic response to subcutaneous regular insulin, patients with overt diabetic nephropathy would have to inject a higher dosage of regular insulin to achieve a metabolic effect comparable with that of nonnephropathic patients. Based on the metabolic activity of regular insulin observed in this study, it can be calculated that a 50% higher dosage is necessary for patients with nephropathy to achieve the same metabolic effect as diabetic patients with normal renal function. With increasing dosages of regular insulin, however, it has been shown that not only the maximal metabolic response but also the duration of action (i.e., the overall blood glucose⫺lowering effect) increases (16). The prolongation of action with higher insulin dosages, in turn, may increase the risk for late postprandial hypoglycemic episodes. Therefore, the necessity for higher dosages of regular insulin with the resulting prolonged effect may partially explain the substantially higher risk for severe hypoglycemia in patients with diabetic nephropathy (2). In our study, patients with overt diabetic nephropathy had a slightly, but not significantly, lower daily insulin dosage compared with diabetic patients without nephropathy. At first glance, this observation contradicts our finding that diabetic patients with overt diabetic nephropathy need higher dosages of insulin to achieve the same metabolic effect as patients without renal impairment. However, patients with overt nephropathy in this study not only injected less insulin, but at the same time exhibited a metabolic control significantly worse than that of diabetic control patients, thus confirming the reduced response to insulin. In the group of patients with overt diabetic nephropathy, we included subjects exhibiting macroproteinuria and/or elevated serum creatinine. Hence, in pa889
Diabetic nephropathy and insulin
tients classified as having nephropathy, 10 of 12 subjects showed proteinuria ⬎500 mg/24 h and 7 of 12 patients exhibited serum creatinine ⬎1.5 mg/dl (122 mol/l). One may speculate that in nephropathic patients with moderate to severe renal impairment (serum creatinine ⬎1.5 mg/dl), insulin clearance and/or insulin action—that is, pharmacokinetic and pharmacodynamic parameters—may have been different compared with patients with nephropathy but with normal glomerular filtration rate or only mild renal impairment. Therefore, we analyzed the results for these two subgroups of patients in our study separately and interestingly did not find any statistically significant differences between both patient groups. An important finding of this study is that insulin lispro, in comparison to regular insulin, maintained its characteristic pharmacokinetic and pharmacodynamic properties in patients with overt diabetic nephropathy. Given that it has been demonstrated that the duration of action of insulin lispro, in contrast to regular insulin, only moderately increases with higher dosages (16), it seems possible that fastacting insulin analogues like insulin lispro and insulin aspart could be of particular advantage in patients with diabetic nephropathy in reducing the risk of hypoglycemia (17). However, this remains to be shown in adequate clinical studies. The increased risk of patients with overt diabetic nephropathy for severe hypoglycemic episodes cannot be explained simply by a prolonged duration of insulin action attributable to reduced insulin clearance. In fact, the time to late half maximal insulin concentrations for both regular insulin and insulin lispro was not higher in patients with nephropathy. More important, the duration of metabolic activity in terms of time to late tGIR 50% was similar for either insulin formulation in patients with and without overt diabetic nephropathy. Apart from the altered metabolism of insulin and the reduced insulin action, the contribution of the kidneys to gluconeogenesis in the postabsorptive state has to be taken into account in patients with renal impairment (18). Results of two studies in type 1 and type 2 diabetic patients indicate that renal glucose production in the diabetic state per se is increased and accounts for ⬃30% of the overall gluconeogenetic capacity (19, 20). It seems sensible to assume that renal 890
gluconeogenetic capacity decreases with declining renal function, which, consequently, would increase the risk of hypoglycemia in these patients. In summary, we have shown that peak insulin concentrations and overall insulin levels were higher in patients with overt diabetic nephropathy than in diabetic patients with normal renal function. In contrast to the higher insulin levels, the overall metabolic response to regular insulin but not to insulin lispro was lower in patients with diabetic nephropathy as in diabetic control patients. The promising pharmacokinetic and pharmacodynamic results of insulin lispro in patients with overt diabetic nephropathy should promote clinical studies aimed at reducing the hypoglycemic risk and improvement of metabolic control in these patients (11).
9.
10.
11.
12.
13. Acknowledgments — This study was partly supported by a research grant from Lilly Deutschland, Bad Homburg, Germany. We gratefully acknowledge the excellent technical assistance of Uta Eckers, Till Schroer, Andre Feldmann, and Claudia Gottschalk. Our special thanks go to Dr. Ralf Bender for his valuable advice in statistical questions. References 1. Arem R: Hypoglycemia associated with renal failure. Endocrinol Metab Clin North Am 18:103–121, 1989 2. Mu¨hlhauser I, Toth G, Sawicki PT, Berger M: Severe hypoglycemia in type 1 diabetic patients with impaired kidney function. Diabetes Care 14:344 –346, 1991 3. Bending JJ, Pickup JC, Viberti GC, Keen H: Glycaemic control in diabetic nephropathy. BMJ 288:1187–1191, 1984 4. Rabkin R, Ryan MP, Duckworth WC: The renal metabolism of insulin. Diabetologia 27:351–357, 1984 5. Mak RH: Correction of anemia by erythropoietin reverses insulin resistance and hyperinsulinemia in uremia. Am J Physiol 270:F839 –F844, 1996 6. DeFronzo RA, Tobin JD, Rowe JW, Andres R: Glucose intolerance in uremia: quantification of pancreatic beta cell sensitivity to glucose and tissue sensitivity to insulin. J Clin Invest 62:425– 435, 1978 7. Eidemak I, Feldt RB, Kanstrup IL, Nielsen SL, Schmitz O, Strandgaard S: Insulin resistance and hyperinsulinaemia in mild to moderate progressive chronic renal failure and its association with aerobic work capacity. Diabetologia 38:565–572, 1995 8. DeFronzo RA, Alvestrand A, Smith D, Hendler R, Hendler E, Wahren J: Insulin
14.
15.
16. 17.
18.
19.
20.
resistance in uremia. J Clin Invest 67:563– 568, 1981 Schmitz O, Alberti KG, Christensen NJ, Hasling C, Hjollund E, Beck-Nielsen H, Orskov H: Aspects of glucose homeostasis in uremia as assessed by the hyperinsulinemic euglycemic clamp technique. Metabolism 34:465– 473, 1985 Howey DC, Bowsher RR, Brunelle RL, Woodworth JR: [Lys(B28), Pro(B29)]-human insulin: a rapidly absorbed analogue of human insulin. Diabetes 43:396 – 402, 1994 Jehle PM, Aisenpreis U, Bundschu D, Keller F: Vorteile von Insulin Lispro bei terminaler Niereninsuffizienz. Fortschr Med 117:41– 42, 1999 Sawicki PT, Bender R, Berger M, Mu¨hlhauser I: Non-linear effects of blood pressure and glycosylated hemoglobin on progression of diabetic nephropathy. J Intern Med 247:131–138, 2000 Rave K, Heise T, Weyer C, Herrnberger J, Bender R, Hirschberger S, Heinemann L: Intramuscular versus subcutaneous injection of soluble and lispro insulin: comparison of metabolic effects in healthy subjects. Diabet Med 15:747–751, 1998 Pfaffenbach B, Wegener M, Adamek RJ, Wissuwa H, Schaffstein J, Aygen S, Hennemann O: Nicht-invasiver 13COktansaureatemtest zur Messung der Magenentleerung einer festen Testmahlzeit– Korrelation mit der Szintigraphie bei Diabetikern und Reproduzierbarkeit bei gesunden Probanden. Z Gastroenterol 33: 141–145, 1995 Burge MR, Mcleod J, Bowsher R, Schade DS: Validity of Coat-A-Count insulin RIA kit for quantifying total and free Humalog. Clin Chem 42:777, 1996 Heinemann L, Woodworth J: Pharmacokinetics and glucodynamics of insulin lispro. Drugs of Today 34:23–36, 1998 Trajanoski Z, Wach P, Kotanko P, Ott A, Skraba F: Pharmacokinetic model for the absorption of subcutaneously injected soluble insulin and monomeric insulin analogues. Biomed Tech 38:224 –231, 1993 Stumvoll M, Meyer C, Mitrakou A, Nadkarni V, Gerich JE: Renal glucose production and utilization: new aspects in humans. Diabetologia 40:749 –757, 1997 Mitrakou A, Plantanisiotis D, Vlachos L, Mourikis D, Meyer C, Chinatalapudi UY, Gutierrez O, Kreider M, Gerich J: Increased renal glucose production in insuin dependent diabetes (IDDM): contribution to systemic glucose appearance and effect of insulin repletion. Diabetes 45: 33A, 1996 Stumvoll M, Perriello G, Nurjhan N, Bucci A, Welle S, Jansson PA, Dailey G, Bier D, Jenssen T, Gerich J: Glutamine and alanine metabolism in NIDDM. Diabetes 45:863– 868, 1996
DIABETES CARE, VOLUME 24, NUMBER 5, MAY 2001
