Insulin lispro is a short-acting human

Transcription

1 Emerging Treatments and O R I G I N A L A R T I C L E Technologies Optimized Basal-Bolus Therapy Using a Fixed Mixture of 75% Lispro and 25% NPL Insulin in Type 1 Diabetes Patients No favorable effects on glycemic control, physiological responses to hypoglycemia, well-being, or treatment satisfaction MAUREEN M.J. JANSSEN, MD FRANK J. SNOEK, PHD NATHALIE MASUREL ROEL P.L.M. HOOGMA, MD WALTER L. DEVILLÉ, MD CORRIE POPP-SNIJDERS, PHD ROBERT J. HEINE, PHD OBJECTIVE To investigate the effects of a multiple injection regimen with a mixture of 75% lispro and 25% intermediate-acting insulin (lispro high mixture [HM]) before meals on glycemic control, physiological responses to hypoglycemia, well-being, and treatment satisfaction. RESEARCH DESIGN AND METHODS We studied 35 type 1 diabetes patients. After an 8- to 10-week lead-in period, patients were randomized to HM or human regular insulin therapy for weeks. During the lead-in and treatment periods, and hypoglycemic frequencies were measured, and patients completed the Well-Being Questionnaire and the Diabetes Treatment Satisfaction Questionnaire. In 19 patients, responses to hypoglycemia were tested during stepped euglycemic-hypoglycemic clamps. RESULTS HM treatment improved postprandial glycemia but had no effect on, frequency of hypoglycemia, well-being, or treatment satisfaction. During experimental hypoglycemia, HM therapy was associated with a slightly lower total adrenaline response and a higher autonomic symptom threshold (i.e., the autonomic symptom response occurred at a lower blood glucose level) than human regular insulin therapy. We speculate that this effect resulted from an accumulation of insulin during the night. CONCLUSIONS Multiple injection therapy with HM rather than human regular insulin before meals does not offer advantages regarding glycemic control, frequency of hypoglycemia, well-being, or treatment satisfaction. In addition, this regimen causes an attenuation of the adrenaline and autonomic symptom responses to hypoglycemia. Diabetes Care 23: , 2000 From the Research Institute for Endocrinology, Reproduction and Metabolism and the Department of Epidemiology and Biostatistics (M.M.J.J., F.J.S., N.M., W.L.D., C.P.-S., R.J.H.), Faculty of Medicine, Vrije Universiteit Amsterdam, Amsterdam; and the Department of Internal Medicine (R.P.L.M.H.), Groene Hart Hospital, Gouda, the Netherlands. Address correspondence and reprint requests to Robert J. Heine, PhD, Department of Endocrinology, University Hospital, Vrije Universiteit Amsterdam, P.O. Box 7057, 1007 MB Amsterdam, the Netherlands. Received for publication 28 July 1999 and accepted in revised form 12 January Abbreviations: AUC, area under the curve; CR, counterregulatory; DTSQ, Diabetes Treatment Satisfaction Questionnaire; GH, growth hormone; HBGM, home blood glucose monitoring; HM, lispro high mixture; HPLC, high-performance liquid chromatography; MIT, multiple-injection therapy; NPL, neutral protamine lispro; WBQ, Well-Being Questionnaire. A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances. Insulin lispro is a short-acting human insulin analog that is absorbed more rapidly from the subcutaneous tissue than human regular insulin (1). Several benefits of insulin lispro have been proposed for the treatment of type 1 diabetes. First, insulin lispro can be injected immediately before meals, which may improve patients treatment satisfaction (2). Second, it improves postprandial glycemia (3). Finally, in some studies, insulin lispro therapy has been shown to reduce the frequency of hypoglycemia (3,4). By lowering the frequency of hypoglycemia, insulin lispro may improve counterregulatory (CR) hormone and symptom responses to subsequent hypoglycemia. However, conflicting results have been reported on this subject (5,6). Although multiple-injection therapy (MIT) with lispro improves postprandial glycemia, it does not affect because of a blood glucose increase 4 to 5 h after injection (3). This increase can be prevented by adding NPH insulin to the premeal insulin lispro injections (7). This therapy, referred to as optimized basal-bolus therapy, has been shown to lower (8) but may reduce patient convenience because NPH insulin must be administered by using extra injections or by mixing insulin lispro and NPH insulin immediately before injection. To produce stable fixed mixtures of insulin lispro and intermediateacting insulin, a protamine-based lispro formulation with kinetics similar to NPH insulin was developed: neutral protamine lispro (NPL) insulin (9,10). One of the fixed mixtures, lispro high mixture (HM) insulin, which consists of 75% lispro and 25% NPL insulin, may offer optimized basal insulin replacement while maintaining the convenience of MIT with prefabricated penfills. The aim of the present study was to determine the effects of MIT with HM insulin before meals and NPL insulin at bedtime on glycemic control, frequency of hypoglycemia, and physiological responses DIABETES CARE, VOLUME 23, NUMBER 5, MAY

2 Lispro high mixture therapy Table 1 Baseline characteristics of the HM and human regular insulin randomization groups (entire patient group and subset of patients participating in the euglycemic-hypoglycemic clamp tests) to hypoglycemia. The secondary aim was to study the effects of this regimen on emotional well-being and treatment satisfaction. All patients Clamp patients HM Regular HM Regular insulin insulin insulin insulin n (M/F) 17 (11/6) 18 (11/7) 9 (7/2) 10 (8/2) Age (years) 33.0 ± ± ± ± 5.4 Duration of diabetes (years) 15.7 ± ± ± ± 7.3 (%)* 7.5 ± ± ± ± 0.7 BMI (kg/m 2 ) 24.9 ± ± ± ± 1.8 Total daily insulin dose (U/kg) 0.7 ± ± ± ± 0.3 Data are n or means ± SD. *Reference range %; P 0.05 HM vs. human regular insulin group. RESEARCH DESIGN AND METHODS We studied 35 type 1 diabetes patients in reasonable glycemic control ( 8.3%) by using MIT with human regular insulin before meals and NPH insulin at bedtime. The protocol was approved by the local ethics committee, and patients gave written informed consent before participating. The study had an open randomized parallel design. After an 8- to 10-week leadin period, patients were randomized to human regular and NPH insulin or HM insulin and NPL insulin for weeks. Patients were instructed to inject regular insulin 30 min before and HM immediately before meals. was measured at screening, at randomization, and at the end of the study. During the lead-in and treatment periods, patients kept 6-week home blood glucose monitoring (HBGM) diaries that required 1 7-point profile (before and 2 h after each meal and bedtime) and 6 4-point profiles (on alternating days before or 2 h after each meal and bedtime) a week. If necessary, insulin dosages were adjusted in increments of 2 U every 3 days to attain the following glucose targets: preprandial glucose between 4 and 8 mmol/l and bedtime glucose between 8 and 10 mmol/l. We analyzed the 7-point profiles and the frequency of hypoglycemia (HBGM readings 3.5 mmol/l). At the end of the lead-in and treatment periods, patients completed the Diabetes Treatment Satisfaction Questionnaire (DTSQ) and Well-Being Questionnaire (WBQ) (11,12). The WBQ includes 4 subscales: depression (6 items), anxiety (6 items), energy (4 items) and positive well-being (6 items). Items are scored on a 0 3 Likert scale. The DTSQ measures satisfaction with treatment (6 items). Items are scored on a 0 6 Likert scale. The first 19 patients enrolled in the study participated in euglycemic-hypoglycemic clamps at the end of the lead-in and treatment periods. Hypoglycemia was avoided for 24 h before the experiments as described previously (13). At 7 A.M., 2 catheters were inserted in the nondominant arm: 1 in a large vein for glucose and insulin infusion and 1 in a distal hand vein for arterialized venous blood sampling (13). After 30 min, a continuous insulin infusion (Humulin R; Lilly, Indianapolis, IN) (50 U diluted in 4.5 ml albumin and 45 ml saline) was started at 50 mu kg 1 h 1. Blood glucose was stabilized at 4 mmol/l within 60 min and was kept at that level for another 60 min with a variable glucose infusion (13). Subsequently, blood glucose was lowered to 3.5 mmol/l within 15 min and was clamped at 3.5 mmol/l for 45 min. This was repeated twice to reach plateaus of 3 and 2.5 mmol/l. For measurement of insulin, catecholamines and growth hormone (GH) blood samples were taken at 15-min intervals. At 20-min intervals, patients used a hand-held computer to complete a hypoglycemic symptom questionnaire and 3 cognitive function tests: 4- choice reaction time, subtractions, and a hidden 2 and 7 test (14). Symptoms were ranked on a 0 6 linear analog scale and included autonomic symptoms (palpitations, warmness, tremor, sweating, visual disturbances, irritability, nervousness, anxiety, and restlessness) and neuroglycopenic symptoms (difficulty concentrating, lightheadedness, confusion, difficulty speaking, drowsiness, odd behavior, lack of coordination, and weakness). were measured with high-performance liquid chromatography (HPLC). During the clamps, blood glucose was measured with a YSI analyzer (Yellow Springs, OH). Plasma catecholamines were measured with HPLC (15), and plasma GH (Sorin Biomedica, Saluggia, Italy) and insulin (Medgenix, Fleurs, Belgium) were measured with 2-site immunometric assays. For the insulin assay, plasma samples of patients with insulin antibody of 10% were treated with an equal volume of 30% polyethylene glycol (16). For insulin and GH, paired study samples were analyzed in the same run. Differences between treatment variables with a skewed distribution were tested with Mann-Whitney U tests, and differences in variables with a normal distribution were tested with multiple linear regression analyses. For, hypoglycemia frequency, and HBGM profiles, analyses were adjusted for baseline. Because CR hormone and hypoglycemic symptom responses are determined by many patient characteristics, including diabetes duration and glycemic control (17), these analyses were adjusted for confounders identified in univariate regression analyses (e.g., testing baseline value, age, diabetes duration, BMI, and ). For the CR hormones, the incremental area under the curve (AUC) was calculated with the trapezoidal rule. The adrenaline AUC was log transformed because of a skewed distribution. Hormone, symptom, and cognitive function thresholds were defined as the time at which an increase of 2 SD above the mean euglycemic level was observed on 2 consecutive samples. A commercial software package (SPSS 7.5 for Windows 95, Chicago) was used for all statistical analyses. RESULTS Baseline characteristics of the HM and human regular insulin randomization groups are shown in Table 1. The HM insulin group had higher than the regular insulin group (Table 1). As for the subgroups of patients participating in the euglycemic-hypoglycemic clamp tests, a similar difference in was observed at a level of significance between 0.05 and 0.1 (Table 1). During the lead-in period, decreased in both groups (mean change ± SD for the HM insulin group 0.3 ± 0.5 [P = 0.03] and for the regular insulin group 0.3 ± 0.7 [P = 0.08]). No difference in was observed for HM versus regular insulin treatment (Table 2). 630 DIABETES CARE, VOLUME 23, NUMBER 5, MAY 2000

3 Janssen and Associates Table 2, HBGM results, and hypoglycemia frequencies for the HM and regular insulin randomization groups HBGM results 2 h after breakfast and 2 h after lunch were significantly lower during HM insulin treatment. No other HBGM differences were observed (Table 2). Changes in total daily insulin dose and bedtime insulin dose between the last weeks of the lead-in and the treatment periods were similar for both groups (data not shown). With HM insulin treatment, the daily intermediate-acting insulin dose increased (HM vs. regular insulin 10.6 ± 3.6 vs. 0.4 ± 1.1; P = ), and the daily short-acting dose decreased (HM vs. regular insulin 7.3 ± 8.6 vs. 1.0 ± 5.3; P = 0.03). During the lead-in period, 1 patient from each randomization group experienced a severe hypoglycemic episode (i.e., hypoglycemia requiring glucagon or intravenous glucose therapy). Similarly, during the treatment period, another patient from each group experienced severe hypoglycemia. Frequencies of mild biochemical hypoglycemia were similar for both treatments. No differences in treatment satisfaction and emotional well-being were observed between the randomization groups at baseline or at the end of the treatment period (data not shown). During the euglycemic-hypoglycemic clamp studies, plasma glucose and insulin and total glucose requirements were similar for the randomization groups (data not shown). Adrenaline responses during the clamp studies at the end of the treatment period are shown in Fig. 1. In univariate analyses, the adrenaline AUC was associated with the baseline value and the level at the end of treatment but not Lead-in period Treatment period HM Regular HM Regular insulin insulin insulin insulin P* (%) 7.2 ± ± ± ± HBGM (mmol/l) Fasting 9.4 ± ± ± ± h after breakfast 10.7 ± ± ± ± Before lunch 6.8 ± ± ± ± h after lunch 7.5 ± ± ± ± Before dinner 8.3 ± ± ± ± h after dinner 8.8 ± ± ± ± Bedtime 9.9 ± ± ± ± Hypoglycemia frequency 8.5 ± ± ± ± Data are means ± SD. *P value for difference between HM and regular insulin treatment adjusted for baseline (lead-in period) result; reference range %; expressed as percentage of all HBGM measurements performed during the 6-week HBGM period. with age, diabetes duration, or BMI. A multivariate analysis adjusting for the baseline results revealed no difference in adrenaline AUC between HM and human regular insulin treatment. After further adjustment for, the natural log of the adrenaline AUC was 1.3 nmol l 1 min 1 lower after HM insulin treatment than after human regular insulin treatment, which corresponds to 3.7 nmol l 1 min 1 for the adrenaline AUC (Table 3). No differences between the treatments were observed for the adrenaline threshold, the noradrenaline and GH thresholds and AUCs and the neuroglycopenic symptom threshold (data not shown). The autonomic symptom thresholds for human regular insulin and HM treatment are shown in Fig. 1. The autonomic symptom threshold showed a univariate association with the baseline result and the level at the end of treatment but not with age, diabetes duration, or BMI. The multivariate analysis revealed that, on average, the threshold for autonomic symptoms occurred 38.1 min later with HM treatment than with human regular insulin treatment (Table 3). No differences between the 2 treatments were detected in any of the cognitive function test results (data not shown). CONCLUSIONS The results of the present study indicate that MIT with HM insulin before meals improves blood glucose 2 h after breakfast and lunch but has no favorable effect on level, fre- Figure 1 Adrenaline responses (means ± SEM) and autonomic symptom thresholds (mean values) during the euglycemic-hypoglycemic clamps after treatment with human regular and NPH insulin or HM and NPL insulin. The autonomic symptom threshold occurred at 233 min (at a blood glucose level of 3 mmol/l) after human regular insulin treatment and at 255 min (at a blood glucose level of 2.5 mmol/l) after HM insulin treatment. DIABETES CARE, VOLUME 23, NUMBER 5, MAY

4 Lispro high mixture therapy Table 3 Differences between HM and regular insulin treatment in adrenaline AUC and autonomic symptom threshold adjusted for baseline result (model 1) and for baseline result and (model 2) quency of hypoglycemia, emotional wellbeing, or treatment satisfaction. In addition, this regimen causes a small but significant attenuation of the adrenaline and autonomic symptom responses to hypoglycemia. Although it successfully prevented a blood glucose increase between meals, HM insulin therapy did not improve (Table 2). These results conflict with several studies that used MIT with the addition of NPH insulin to the premeal insulin lispro injections (5,8,18). The failure of HM insulin to improve cannot be ascribed to the baseline difference between the randomization groups because the statistical analyses were adjusted for baseline. A possible explanation is that both patient groups had already reached nearoptimal glycemic control during the relatively long lead-in period. An alternative explanation, however, is that individualized rather than fixed mixtures of insulin lispro and intermediate-acting insulin are necessary to improve glycemic control (19). After HM insulin treatment, the total adrenaline response to experimental hypoglycemia was reduced, and a lower blood glucose level was required to elicit autonomic symptoms. These findings were not because of an increase in the frequency of daytime biochemical hypoglycemia. With SD of independent Standardized regression variable coefficient (95% CI) P Adrenaline AUC* Model 1 (adjusted R 2 41%) Baseline AUC (nmol l 1 min 1 ) (0.3 to 1.2) HM (1) vs. regular (0) insulin 0.7 ( 1.5 to 0.2) 0.1 Model 2 (adjusted R 2 62%): Baseline AUC (nmol l 1 min 1 ) (0.1 to 0.9) 0.01 (%) (0.2 to 1.1) HM (1) vs. regular (0) insulin 1.3 ( 2.1 to 0.4) Threshold autonomic symptoms Model 1 (adjusted R 2 74%) Baseline threshold (min) (35.6 to 66.7) HM (1) vs. regular (0) insulin 26.7 ( 3.5 to 57.0) 0.08 Model 2 (adjusted R 2 77%) Baseline threshold (min) (35.7 to 65.4) (%) ( 27.8 to 3.7) 0.1 HM (1) vs. regular (0) insulin 38.1 (5.7 to 70.4) 0.02 *Adrenaline AUC was log transformed because of a skewed distribution. The standardized regression coefficient is the regression coefficient multiplied by the SD of the variable. The adjusted R 2 is the proportion of variation of the dependent variable explained by the model. The (1) and the (0) indicate the 2 values that can be inserted into the regression model, yielding the formulae for HM and regular insulin, respectively. acute administration of insulin lispro, no changes in the physiological responses to hypoglycemia have been reported (20). During MIT with insulin lispro, one study reported no change (6) and one study reported an improvement (5) in these responses. Analogous to the effect of HM on glycemic control discussed above, the difference in outcome between the latter study and our study may be explained by the use of a tailor-made basal insulin substitution in the latter study versus a fixed mixture of 75% lispro and 25% NPL insulin in our study. In particular, the predinner proportion of intermediate-acting insulin in our study is high compared with the proportion reported with the tailor-made basal insulin substitution (8). This may also explain why HM insulin therapy did not improve HBGM 2 h after dinner (Table 2). The combination of HM insulin before dinner and NPL insulin at bedtime may have caused an accumulation of insulin during the night that increased the frequency of nocturnal hypoglycemia. However, this explanation requires further research, including in-hospital studies to assess the frequencies of nocturnal hypoglycemia and clinical trials regarding the effects of different fixed insulin lispro/npl insulin mixture regimens on type 1 diabetic patients. Emotional well-being and treatment satisfaction are important parameters in the investigation of new insulin treatments. An important advantage of the WBQ and DTSQ is that these questionnaires have been validated and have proved useful with a wide range of treatments for diabetes (11). By using these questionnaires, we did not observe a change in treatment satisfaction or emotional well-being during HM insulin treatment. However, the possibility cannot be excluded that a larger patient group may be necessary to demonstrate improvements in these parameters. In summary, MIT with HM insulin before meals improves postprandial glycemia and is not associated with an increase in blood glucose between meals. However, this regimen offers no advantage over human regular insulin treatment regarding level, frequency of hypoglycemia, emotional well-being, or treatment satisfaction. Moreover, this regimen attenuates adrenaline and autonomic symptom responses to experimental hypoglycemia. Because we postulate that these latter effects resulted from an accumulation of intermediate-acting insulin during the night, a possible solution would be to replace the HM insulin injection before dinner with insulin lispro. At this point, however, further studies are necessary to establish the applicability of fixed rather than flexible mixtures of insulin lispro and intermediate-acting insulin for optimized basal-bolus therapy with insulin lispro in type 1 diabetes patients. Acknowledgments This work was supported by a study grant from Eli Lilly and Co., Indianapolis, IN. We thank Boris Kovatchev and Bob Smeets for their help with the HHC program and Mark Seubert and Petya Kamenova for their assistance during the clamps. We are most grateful to the patients for their participation. References 1. Howey DC, Bowsher RR, Brunelle R, Woodworth JR: [Lys(B28), Pro(B29)]-human insulin: a rapidly absorbed analogue of human insulin. Diabetes 43: , Kotsanos JG, Vignati L, Huster W, Andresjasch C, Boggs MB, Marrero D, Mathias SD, Patrick D, Zalani S, Anderson J: Healthrelated quality-of-life results from multinational clinical trials of insulin lispro: assessing benefits of a new diabetes therapy. Diabetes Care 20: , Anderson JH, Brunelle RL, DiMarchi R, 632 DIABETES CARE, VOLUME 23, NUMBER 5, MAY 2000

5 Janssen and Associates Koivisto VA, Pfutzner A, Trautmann ME, Vignati L, the Multicenter Insulin Lispro Study Group: Reduction of postprandial hyperglycemia and frequency of hypoglycemia in IDDM patients on insulin-analog treatment. Diabetes 46: , Garg SK, Carmain JA, Braddy KC, Anderson JH, Vignati L, Jennings MK, Chase HP: Pre-meal insulin analogue insulin lispro vs. Humulin R insulin treatment in young subjects with type 1 diabetes. Diabet Med 13: 47 52, Lalli C, Ciofetta M, Del Sindaco P, Brunetti P, Bolli GB: Long-term intensive treatment of type 1 diabetes with the short-acting insulin analog lispro in variable combination with NPH insulin at mealtime. Diabetes Care 22: , Heller SR, Amiel SA, Macdonald IA, Tattersall RB: Does insulin lispro preserve the physiological defences to hypoglycaemia during intensive insulin therapy (Abstract)? Diabetologia 41 (Suppl. 1):A241, Torlone E, Pampanelli S, Lalli C, Del Sindaco P, Di Vincenzo A, Rambotti AM, Modarelli F, Epifano L, Kassi G, Perriello G, Brunetti P, Bolli G: Effects of the short-acting insulin analog [Lys(B28), Pro(B29)] on postprandial blood glucose control in IDDM. Diabetes Care 19: , Del Sindaco P, Ciofetta M, Lalli C, Perriello G, Pampanelli S, Torlone E, Brunetti P, Bolli GB: Use of the short-acting insulin analogue lispro in intensive treatment of type 1 diabetes mellitus: importance of appropriate replacement of basal insulin and timeinterval injection-meal. Diabet Med 15: , Heise T, Weyer C, Serwas A, Heinrichs S, Osinga L, Roach P, Woodworth J, Gudat U, Heinemann L: Time-action profiles of novel premixed preparations of insulin lispro and NPL insulin. Diabetes Care 21: , Janssen MM, Casteleijn S, Devillé W, Popp- Snijders C, Roach P, Heine RJ: Nighttime insulin kinetics and glycemic control in type 1 diabetes patients following administration of an intermediate-acting lispro preparation. Diabetes Care 20: , Bradley C: Handbook of Psychology and Diabetes: A Guide to Psychological Measurement in Diabetes Research and Practice. Chur, Switzerland, Harwood Academic, Pouwer F, Snoek FJ, Van der Ploeg HM, Heine RJ, Brand AN: A comparison of the standard and computerized versions of the Well-Being Questionnaire (WBQ) and the Diabetes Treatment Satisfaction Questionnaire (DTSQ). Qual Life Res 7:33 38, Kanc K, Janssen MMJ, Keulen ETP, Jacobs MAJM, Popp-Snijders C, Snoek FJ, Heine RJ: Substitution of night-time continuous subcutaneous insulin infusion therapy for bedtime NPH insulin in a multiple injection regimen improves counterregulatory hormonal responses and warning symptoms of hypoglycaemia in IDDM. Diabetologia 41: , Cox DJ, Gonder-Frederick LA, Lee JH, Julian DM, Carter WR, Clarke WL: Effects and correlates of blood glucose awareness training among patients with IDDM. Diabetes Care 12: , Popp-Snijders C, Geenen B, Van der Heijden EAP: Serum noradrenaline is composed of plasma and platelet noradrenaline. Ann Clin Biochem 26: , Arnqvist H, Olsson PO, Von Schenck H: Determination of free and total insulin after precipitation with polyethylene glycol: analytical characteristics and effects of sample handling and storage. Clin Chem 33:93 96, Mokan M, Mitrakou A, Veneman T: Hypoglycemia unawareness in IDDM. Diabetes Care 17: , Ciofetta M, Lalli C, Del Sindaco P, Torlone E, Pampanelli S, Mauro L, Chiara DL, Brunetti P, Bolli GB: Contribution of postprandial versus interprandial blood glucose to in type I diabetes on physiologic intensive therapy with lispro insulin at mealtime. Diabetes Care 22: , Bolli GB, Di Marchi RD, Park GD, Pramming S, Koivisto VA: Insulin analogues and their potential in the management of diabetes mellitus. Diabetologia 42: , Jacobs MAJM, Salobir B, Popp-Snijders C, Ader H, Heine RJ: Counterregulatory hormone responses and symptoms during hypoglycaemia induced by porcine, human regular insulin, and Lys(B28), Pro(B29) human insulin analogue (insulin lispro) in healthy male volunteers. Diabet Med 14: , 1997 DIABETES CARE, VOLUME 23, NUMBER 5, MAY