| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Service de Diabétologie, Maladies Métaboliques et Nutrition (B.G., L.M., A.S., O.Z., P.D.), Centre d’Investigation Clinique-INSERM/Centre Hospitalier Universitaire de NANCY-Hôpital Jeanne d’Arc, 54201 Toul Cedex; Hôpital de l’Antiquaille (A.C.), Service de Médecine Nucléaire, 69000 Lyon; and Laboratoire de Biochimie (B.D.), Hôpital Central, 54000 Nancy, France
Address all correspondence and requests for reprints to: Bruno Guerci, M.D., Diabétologie, Maladies Métaboliques et Nutrition, Hôpital Jeanne d’Arc, CIC-INSERM/Centre Hospitalier Universitaire de NANCY, B.P. 303, 54201 Toul Cedex, France. E-mail: cic{at}chu-nancy.fr
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Continuous sc insulin infusion (CSII) provides the most physiological pattern of insulin delivery currently available (8, 9), and recent studies suggest that LP can be used in insulin pump therapy to improve blood glucose control (10) and provide greater stability without increasing the risk of hypoglycemia (11). However, technical problems with CSII (pump failure, catheter occlusion, and skin infection) may reduce insulin delivery or insulin absorption and cause diabetic ketoacidosis (12). A rapid drop in plasma free insulin and an increase in blood glucose and plasma 3ß-hydroxybutyrate (ß-OHB) occurred when CSII with regular insulin was interrupted for 6 h during the night (13). Given the pharmacokinetic characteristics of LP, the interruption of CSII may result in a more rapid metabolic deterioration in type 1 diabetic patients (14).
We, therefore, compared the hormonal and metabolic changes occurring in type 1 diabetic patients given LP and those given regular insulin Velosuline (VE) during a 5-h interruption of CSII. We also evaluated the efficacy of a therapeutic scheme to restore adequate blood glucose control after resetting the pump with LP or VE.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The study was carried out on 10 type 1 diabetic patients, diagnosed according to American Diabetes Association criteria (15), who were C peptide negative (<0.3 nmol/L, 6 min after 1 mg glucagon, iv). The patients were treated by external pump (Infusor 506, MiniMed, Inc., Sylmar, CA) and deconnectable catheter (Tender catheter, Disetronic Medical Systems, Inc., Plymouth, MN). These patients were selected from 170 insulin-dependent diabetic patients treated by CSII at our out-patient clinic. Their inclusion criteria were: receiving CSII for more than 1 yr (range, 2–12 yr) and awareness of hypoglycemia. Insulin was infused into the abdomen, and the infusion site was changed every 3 days. None had had a recent episode of ketoacidosis or required a daily insulin dose greater than 1.5 U/kg·day. Patients with hyperlipidemia, thyroid or liver disease, diabetic or nondiabetic renal disease, urinary tract infection, pregnancy, or acute or chronic inflammatory syndrome were excluded. None had proteinuria or microalbuminuria (defined as an albumin excretion rate of 20–200 µg/min). None of the patients had macrovascular complications, and none was affected with retinopathy. Finally, patients were instructed to follow a weight-maintaining diet (15% of calories as protein, 35% as fat, and 50% as carbohydrate) taken as three main meals and one to three snacks per day throughout the study. Two patients were treated for hypertension by angiotensin-converting enzyme inhibitor (perindopril, 2 mg/day) and one by calcium antagonists (nicardipine, 100 mg/day). One premenopausal women was taking the oral contraceptive pill (mini-pill, levonorgestrel, 3 mg/day), and one postmenopausal women was receiving hormonal replacement therapy [17ß-estradiol using the transdermal route (1.25 mg/25 days) plus progesterone (200 mg/14 days)].
The patients had been taught to measure their capillary blood glucose using a One Touch Profile memory meter (Lifescan, Roissy, France) to calculate mean blood glucose levels, mean SD of blood glucose, postprandial glucose levels, and incidence of hypoglycemia. Blood glucose concentrations less than 3.5 mmol/L were considered to indicate hypoglycemia. The practical aspects of home blood glucose monitoring and pump handling were checked, as were the guidelines for adjusting insulin doses in response to fluctuations of blood glucose (16, 17).
Study protocol
The study was a randomized, cross-over, open label design to compare the effects of interruption of CSII on treatment with LP or VE over a period of 3 months. Each patient served as his own control. Patients were given a CSII with VE (Velosuline HM, U-100, Novo-Nordisk, Boulogne-Billancourt, France) for a 1-month run-in period and were then randomly assigned to treatment with VE (five patients) or LP (Humalog, U-100, Lilly France, Saint-Cloud, France; five patients) for 1 month. The patients were then switched to the alternate insulin therapy (VE or LP) for the second 1-month period.
The patients came to the Clinical Research Center (INSERM-Centre Hospitalier Universitaire de Nancy) the day before the interruption of CSII at 2000 h for a calibrated meal (800 Cal; 31% fat, 19% protein, and 50% carbohydrate). No hypoglycemic episode occurred during the 24 h before each test. The subjects fasted for 11 h, until 0700 h (no breakfast), at which time the pump was stopped, and the catheter was disconnected. The pump of each patient was reactivated at his usual basal rate at 1200 h (300 min), at which time the patients ate a similar calibrated lunch and activated their usual prelunch insulin boluses. Additional insulin was given each hour according to the following criteria: 4 U if blood glucose was 11.1 mmol/L or higher (and/or whether there was major ketonuria), 2 U if blood glucose was below 11.1 mmol/L (and/or whether ketonuria was moderate), and 1 U if blood glucose was below 11.1 mmol/L (without ketonuria). No additional bolus was given if blood glucose was below 8.3 mmol/L (without ketonuria). All of the type 1 diabetic patients ate a snack at 1700 h and had an additional 2-U insulin bolus. They left the Clinical Research Center when their blood glucose levels were below 16.6 mmol/L without ketonuria. We planned to shorten the protocol if the plasma glucose level during the insulin withdrawal period was 3 times the plasma glucose concentration at 0700 h or when there was severe ketonuria with symptoms of insulinopenia. This did not occur, and all patients completed the second phase of the protocol.
Informed written consent was obtained from all the patients. The study protocol was approved by the ethics committee of Centre Hospitalier Universitaire de Nancy, France.
Biochemical determinations
Blood samples were taken every hour, from 0700 h (time zero) to 1700 h (600 min). Plasma glucose was measured by the glucose oxidase method (glucose analyzer, Beckman Coulter, Inc., Fullerton, CA). Glycated hemoglobin (HbA1c) was measured by high performance liquid chromatography on Biorex resins (Bio-Rad Laboratories, Inc. Richmond, CA; normal range, 4.3–6%). Plasma free insulin was measured by immunoradiometric assay (Bi-Insulin IRMA, Sanofi Pharmaceuticals, Inc., Pasteur, Marnes la Coquette, France). This assay used two monoclonal antibodies, mAb 19 and mAb P10, which recognize two distinct regions of the insulin molecule (18). Plasma samples were treated with polyethylene glycol immediately after separation to precipitate antibody-bound insulin. The limit of detection for plasma free insulin was 1.2 pmol/L. Antiinsulin antibodies were measured by RIA with a commercial kit (Biosource Technologies, Inc., Fleurus, Belgium). The intra- and interassay coefficients of variation were 2.9% and 4.8%, respectively. Plasma glucagon was measured by RIA using a selective two-site assay with two antibodies (normal range, 11–37 pmol/L; Behring, Rueil Malmaison, France). The limit of detection for glucagon was 3.7 pmol/L. Blood samples for determining ketone bodies and nonesterified fatty acids (NEFA) were collected on crushed ice, and plasma was immediately obtained by centrifugation at 4 C. ß-OHB was determined by an enzymatic end-point spectrophotometric method using 3-hydroxybutyrate (normal range, 60–170 µmol/L; KONE Delta Automatic Analyzer, Wako Biohemicals, Osaka, Japan). The intraassay coefficient of variation was 4.9% (19). NEFA were measured using an enzymatic colorimetric assay (normal range, 150–450 µmol/L; KONE Delta Analyzer) (20).
Statistical analyses
Data are expressed as the mean ± SD. The data
in Figs. 1
and 2 are the mean ± SEM. The distribution
of variables was tested for approximation to a Gaussian distribution
(normality) using the kurtosis and skewness test. Data were compared
using the nonparametric Wilcoxon test for two paired groups (LP and VE)
and within a given test, vs. the value recorded at 0700
h (time zero). ANOVA was used for a cross-over study, with period,
treatment group, and interaction factors. Statistical significance is
implied by P < 0.05. Statistical analyses were
performed using the StatView program (StatView IV, Brain Power, Inc. Calabasas, CA).
|
|
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
and 2.
All of the patients were treated with VE during this run-in period.
ANOVA used for the cross-over study showed no period or carry-over
effect and no treatment-period interaction for any of the parameters
studied. Therefore, the individual data for the two periods of each
treatment were pooled for analyses. Table 3 shows the baseline data before the
start of each of the two insulin preparations. There were no
differences between the two insulin preparations.
|
|
|
compares the main metabolic
control parameters before CSII interruption for the two insulin
preparations studied. Mean plasma HbA1c, blood glucose, and
mean SD of blood glucose during the LP or VE periods were
not significantly different. The 1.5-h postprandial blood glucose was
lower with LP than with VE (P < 0.05). There were
fewer hypoglycemia episodes with LP than with VE (P <
0.05). Seven of the 10 patients had decreased rates of hypoglycemia
when switched to LP (15.6 ± 10.9 with VE and 6.4 ± 5.0 with
LP, number/30 days), whereas the 3 others had increased rates (5.7
± 3.5 with VE and 8.7 ± 3.8 with LP, number/30 days). The daily
insulin doses during the VE and LP periods were similar. However, the
daily bolus insulin dose was significantly lower with LP than with VE,
whereas the basal insulin delivery rate was significantly higher with
LP than with VE. Nine of the 10 patients required smaller boluses when
switched to LP.
|
The plasma glucose levels (Fig. 1
) at time zero during the two
treatments were similar (VE, 7.38 ± 3.38; LP, 7.82 ± 3.33
mmol/L) and remained unchanged during the first hour after stopping the
pump. The plasma glucose concentrations then increased significantly
from 120 until 300 min and were significantly higher during LP than
during VE from 180 min (13.93 ± 3.77 vs. 10.77 ±
4.38 mmol/L; P < 0.05) to 300 min (17.04 ± 3.27
vs. 12.98 ± 4.33 mmol/L; P <
0.01).
The plasma free insulin levels (Fig. 1)
at baseline for the two treatments were not significantly different
[VE, 58.44 ± 24.24 (range, 28.02–93.0); LP, 50.82 ± 20.16
(range, 25.14–87.6) pmol/L]. The plasma free insulin significantly
decreased from baseline during the interruption of CSII during both VE
and LP, but the decrease was more rapid for LP than for VE. The plasma
free insulin levels were significantly lower with LP than with VE from
60–300 min. (P < 0.05–0.01), with minimums at 300
min of 8.22 (LP) and 21.96 (VE) pmol/L (P < 0.05).
The plasma glucagon levels (Fig. 1) at time zero during the two
treatments were similar [VE, 21.2 ± 8.3 (range, 12–39.1); LP,
19.7 ± 8.1 (range, 13.2–39.2) pmol/L], and no patients had low
or undetectable plasma glucagon. The plasma glucagon remained unchanged
after stopping the pump and was similar for VE and LP from 0–300 min.
During VE treatment, plasma glucagon concentrations were slightly lower
at 60 min than at 0 min (P < 0.01).
The plasma NEFA levels (Fig. 2) during LP
treatment were higher after 60 min than at 0 min, whereas they were
significantly higher after 180 min during VE treatment, indicating that
the increase in plasma NEFA occurred more rapidly for LP than for VE.
Plasma NEFA was also significantly higher for LP than for VE from 120
min (805 ± 223 vs. 638 ± 225 µmol/L;
P < 0.01) to 240 min (1401 ± 510 vs.
993 ± 353 µmol/L; P < 0.05).
The plasma ß-OHB levels (Fig. 2) increased rapidly after the pump was
stopped in the two treatments, but the increase began earlier during
the LP treatment (at T60 min) than during the VE treatment (at 120
min). The ß-OHB levels in LP and VE were similar from 0–300 min.
We looked for correlations between metabolic control at baseline and the changes that occurred during CSII interruption. Basal free insulin concentrations were not correlated with HbA1c in VE or LP. In VE, the plasma free insulin levels at baseline were negatively correlated with alterations in plasma glucose (r = -0.86; P < 0.01) and ß-OHB (r = -0.69; P < 0.05). The same trend was observed in LP, but the correlations were not statistically significant (P = 0.09 for plasma glucose and P = 0.06 for ß-OHB). There was no correlation between plasma glucagon at baseline and alterations in ketone bodies or between HbA1c at baseline and any of the parameters studied throughout the interruption of CSII.
Restarting the pump
Plasma glucose levels (Fig. 1) remained stable from 300 min
(17.04 ± 3.27 mmol/L) to 420 min (18.15 ± 3.22 mmol/L)
during LP and then decreased from 480–600 min. Plasma glucose levels
during VE continued to increase during the first 2 h after
restarting the pump (300 min, 12.98 ± 4.33 mmol/L; 420 min,
17.26 ± 3.05 mmol/L; P < 0.01) and then
decreased from 420–600 min. The plasma glucose level during LP was
significantly higher than that during VE only at 360 min.
The plasma free insulin peak (Fig. 1) was significantly higher and
occurred faster with LP than with VE (224.4 ± 120.6
vs. 96.6 ± 30.6 pmol/L; P < 0.05;
from 1 h after restarting the pump) despite the fact that the
usual insulin boluses administered with LP were significantly smaller
than those given with VE (LP, 7.3 ± 4.5 U; VE, 8.9 ± 3.6 U;
P < 0.05). Plasma free insulin at 540 min was
significantly higher during VE than during LP (153.0 ± 44.4
vs. 103.8 ± 57.0 pmol/L; P < 0.05)
The insulin supplements delivered by the pump from 360–600 min were
similar for the two insulin preparations (LP, 6.5 ± 4.7 U; VE,
7.3 ± 3.6 U).
The plasma glucagon levels (Fig. 1) from 300–600 min were similar for
the two treatments. The plasma glucagon level during LP was slightly,
but significantly, higher at 360 min than at 0 min.
The plasma NEFA level (Fig. 2) during LP treatment decreased
dramatically after restarting the pump and was no longer different from
baseline at 360 min. VE treatment produced a similar, but less
pronounced, decrease. There was no difference between the LP and VE
treatments at any time from 300–600 min. The plasma NEFA concentration
during VE was below baseline from 480–600 min (P <
0.05–0.01).
The plasma ß-OHB levels (Fig. 2) at T360 min were higher with LP than
with VE (709 ± 364 vs. 462 ± 371 µmol/L;
P < 0.05). ß-OHB levels returned to the normal range
2 h after restarting the pump and thereafter were not
statistically different from the baseline levels.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Rapid metabolic deterioration is more frequent in users of insulin pumps because of catheter occlusions, disconnections, or pump failures causing low insulin levels (21). LP is also more rapidly absorbed and has a shorter action than regular insulin, which could place type 1 diabetic patients at risk of diabetic ketoacidosis if the insulin input fails. Two recent studies have demonstrated that LP provides better blood glucose control and stability when used in external pumps (11), without increasing the incidence of hypoglycemic episodes (10). These results suggest that LP should be more widely used with CSII. Therefore, specific guidelines of safety and efficacy for type 1 diabetic patients treated with LP by CSII must be evaluated.
Our results are slightly different from those of a recent study that compared human regular and LP insulins after a 6-h interruption of CSII (22). The researchers concluded that interrupting the basal insulin infusion did not result in more rapid metabolic decompensation in patients treated with LP than in those treated with regular insulin. These discrepancies may be due to differences in methodology, such as the different times at which insulin was stopped (0300–0900 h vs. 0700–1200 h). Each of our patients served as his own control, which ruled out the potential influence of diabetes duration. None of our patients had low or undetectable glucagon levels at baseline, which might influence the increment of ketone bodies in response to insulin withdrawal.
The dramatic fall in plasma free insulin with LP is probably linked to the more rapid disappearance of LP from sc tissue. The rise in plasma NEFA appears to be the earliest marker of metabolic deterioration when LP infusion is discontinued and is probably due to lipolytic activity that is poorly inhibited by insulin, resulting in increased NEFA in the portal circulation and plasma (23). ß-OHB levels increased after stopping the pump, but later than NEFA levels. This could be linked to the delay needed for an increased flux of NEFA to the liver and the hepatic breakdown of NEFA.
With regard to plasma glucagon, stopping LP treatment caused a slight, but significant, increase at 360 min compared with the baseline value. This is consistent with the lower plasma free insulin levels during LP treatment and consequently less inhibition of plasma glucagon by insulin. Krzentowski et al., who studied the hormonal consequences of nocturnal interruption of CSII with regular insulin (13), demonstrated that plasma glucagon was above baseline only 6 h after stopping the pump and remained significantly elevated until 4 h after restarting the pump. The lack of a significant change in glucagon in our study was probably due to the shorter interruption of CSII.
This study was also performed to provide patients receiving CSII with guidelines for restoring adequate metabolic control after an incidental interruption of their insulin infusions. The scheme used included insulin bolus supplements to the usual prelunch doses. These guidelines were effective, as the NEFA and ß-OHB levels (markers of insulinopenia) were controlled in the two insulin groups and reached the normal range 2 h after restarting the pump. The insulin levels rose more rapidly with LP; consequently, plasma glucose levels plateaued during LP, whereas they further increased during VE (from 300–420 min). These results are in accordance with the pharmacokinetic characteristics of LP (3) and with the study of Holleman et al. (24), who demonstrated that LP is more potent at controlling incidental hyperglycemia.
The total daily insulin doses in the VE and LP treatments were similar. However, the basal insulin delivery rate of LP was significantly higher than that of VE. This is in agreement with a previous study using basal-bolus insulin regimens by multiple daily injections (25). The daily bolus insulin dose with LP was significantly lower than that with VE, as previously reported with CSII (11). This important reduction in bolus doses remained statistically significant when three patients with elevated bolus dose and bolus to basal insulin ratio during VE due to high carbohydrate intake were removed from the analysis. Two of these three patients were also slightly overweight. Conversely, the significant reduction of the rate of hypoglycemia in our patients after 1 month of LP treatment was no longer significant when these three patients were removed from the analysis.
In summary, a short interruption of Lispro in CSII is associated with an earlier and greater metabolic deterioration. However, Lispro, due to its more rapid absorption, is more effective in correcting this metabolic deterioration after restarting the pump. Consequently, the use of insulin Lispro in CSII requires specific education with careful self-monitoring of blood glucose and ketone bodies.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received October 30, 1998.
Revised May 4, 1999.
Accepted May 10, 1999.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. W. Ponder, J. S. Skyler, D. F. Kruger, D. Matheson, and B. W. Brown Unexplained Hyperglycemia in Continuous Subcutaneous Insulin Infusion: Evaluation and Treatment The Diabetes Educator, March 1, 2008; 34(2): 327 - 333. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Guerci, M. Benichou, M. Floriot, P. Bohme, S. Fougnot, P. Franck, and P. Drouin Accuracy of an Electrochemical Sensor for Measuring Capillary Blood Ketones by Fingerstick Samples During Metabolic Deterioration After Continuous Subcutaneous Insulin Infusion Interruption in Type 1 Diabetic Patients Diabetes Care, April 1, 2003; 26(4): 1137 - 1141. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
) or VE (
). Type 1
diabetic patients (n = 10) were randomly assigned to treatment
with VE (n = 5) or LP (n = 5) for 1 month. The patients were
then switched to the alternate insulin therapy (VE or LP) for the
second 1-month period. Results are the mean ± SEM. *,
P < 0.05; **, P < 0.01 (LP
vs. VE). 
•, P < 0.05
[vs. 0700 h (time zero)].