Comparison of Peripheral and Portal (Via the Umbilical Vein) Routes of Insulin Infusion in IDDM Patients PETER I. SHISHKO, PAVEL A. KOVALEV, VALENTIN G. GONCHAROV, AND IGOR U. ZAJARNY
Twelve subjects with insulin-dependent diabetes mellitus were treated using continuous subcutaneous insulin infusion (CSII) and intraportal insulin infusion (IPII) via the umbilical vein for 4 mo. Glucose control improved in both CSII and IPII groups, but a decrease in glucose and HbAlc was more rapid and more significant in the IPII group than in CSII, even though insulin requirement was lower during IPII than CSII (40 ± 2 vs. 50 ± 2 U/day, P < 0.05). The insulin plasma fasting levels were different (88 ± 10.7 in the IPII group vs. 263 ± 23 pM in CSII, P < 0.001). High plasma levels of lactate, pyruvate, alanine, cortisol, and growth hormone were decreased in both groups, with their full normalization only in the IPII group. Glucagon concentrations were low in both groups at the beginning of the study (30.0 ± 4.1 in the CSII group and 32.3 ± 1 . 8 ng/L in IPII); they were equalized to control values in the IPII group and were low in the CSII group at the study's end (46.0 ± 3.7 in IPII vs. 31.7 ± 3.1 ng/L in CSII, P < 0.05. We conclude that intraportal administration of insulin via the umbilical vein at rates of 0.01-0.05 U • kg"1 hr 1 reduces plasma levels of glucose, three carbon precursors, cortisol, and growth hormone by a direct action on the liver, and the hepatic action of peripherally administered insulin is manifested only when the infusion rate is increased to 0.1-0.3 U • kg'1 • hr 1 . Thus the intraportal route via the umbilical vein may offer improved physiological insulin delivery over the subcutaneous route. Diabetes 41:1042-49,1992
From the Institute of Nutrition, Clinical Department, USSR Academy of Medical Sciences, and the Surgical Department, Vladivostok Medical Institute, Moscow, Russia, CIS. Address all correspondence and reprint requests to Dr. P.I. Shishko, Clinical Department, Institute of Nutrition, USSR Academy of Medical Sciences, 2/14 Ustinsky Proezd, 109240 Moscow, Russia, CIS. Received for publication 10 April 1991 and accepted in revised form 11 March 1992.
1042
N
ormal blood glucose control is the aim of new intensive treatment regiments for insulin-dependent diabetic mellitus (IDDM) patients. New modes of insulin delivery (closed- and openloop) systems have been used to provide better physiological insulin replacement. Normalization of blood glucose is possible with an artificial p-cell (Biostator; 1 Life Science Instruments, Miles, Elkhart, IN). However, during treatment with the Biostator, insulin is administered via peripheral route, therefore several metabolic parameters remain pathological (2,3). On the other hand, the functional significance of insulin infusion into the portal vein has not been established. Numerous studies indicate that vascular connection between pancreatic (3-cells and the liver is not important for the metabolic actions of insulin (4-6). The hormone's direct infusion into the liver via the hepatic portal vein had no different effect on glucose metabolism than did its delivery into the peripheral vein (6). Nevertheless, some studies performed on animals suggest that portal insulin delivery should be required to achieve normalization of metabolic disturbances (7-10). Thus, we compared glucose, serum free insulin levels, intermediary metabolites, and hormone profiles achieved with portal or peripheral insulin infusion over 4 mo. We also compared the metabolic status in similar glycemic circumstances after a breakfast of 650 cal and insulin infusion at 0.05 U • kg"1 • hr"1 in IDDM subjects and compared these results with those for nondiabetic control subjects. In this study, we used a catheterization of the portal vein via the umbilical vein and implantation of the PortA-Cath system (Pharmacia, Piscataway, NJ) for chronic insulin infusion.
RESEARCH DESIGN AND METHODS Twelve IDDM patients participated in this study. Their clinical characteristics are shown in Table 1. Mean age of
DIABETES, VOL. 41, SEPTEMBER 1992
P.I. SHISHKO AND ASSOCIATES
TABLE 1 Clinical characteristics of patients on entry to the study and after 4 mo of therapy
Patients (n) Control (5) CSII (6) 0.0 mo 1.0 mo 2.0 mo 3.0 mo 4.0 mo IPII (6) 0.0 mo 1.0 mo 2.0 mo 3.0 mo 4.0 mo
HbA 1c
(%)
5.5 ± 0.8
Glucose (mM)
Total insulin dose (U/day)
4.5 ± 1.5
IRI (pM)
C-peptide (nM)
86.3 ± 10.8
3.6 ± 0.36
Duration of diabetes (yr)
Age (yr)
Weight (kg)
28 ± 2
74 ± 2
15.0 10.2 8.2 7.8 6.8
± 1.5* ± 1.2t ± 0.8 ± 0.4|| ± 0.211
53 ± 4 52 ± 2 52 ± 2 52 ± 2 50±2§
260 ± 22.3* 263 ± 23* 259 ± 22 250 dt 13 248 dt 6.5**
0.6 0.6 0.6 0.6 0.6
± 0.06* ± 0.15 ±0.12 ± 0.24 ± 0.21
18±2
24 ± 2
1 4 . 2 : t 1.1 13.5 dt 0.5 12.2 dfcO.2§ 10.0 dt 0.4§t
66 66 66 67 67
dt 3 dt 3 dt 3 dfc2 db 2
13.8 dt 1.5* 13.0 dt 1.0 11.1 dt 0.5|| 6.9 dfc0.4 5.5 dbO.4f
14.5 6.0 5.5 5.0 4.8
± 1.7* ± 2.4||§1I ± 1.5 ± 0.5 ± 0.51
52 ± 3 48 ± 2 40±2t 40 ± 2 40±2§
270 dt8.6* 88 dt 10.7U 73 dt 7.1 79 dt8.6 82 ± 5.7
0.6 0.6 0.6 0.6 0.6
± 0.06* ±0.15 ± 0.09 ± 0.06 ±0.12
18±2
24 ± 2
67 69 70 70 70
db 2 dt 2 dt 1 dt 1 dt 1
14.8 i t 1.5*
Values are mean ± SE. IRI, immune-reactive insulin, n, number of patients. CSII, diabetic patients during continuous subcutaneous insulin infusion. IPII, diabetic patients during intraportal insulin infusion. *P < 0.001 in comparison with control values. t P < 0.05 in comparison with 0 mo within one group. *P < 0.001 comparison between IPII and CSII groups. §P < 0.05 comparison between IPII and CSII groups. \\P < 0.05 in comparison with control values. IIP < 0.001 in comparison with 0 mo within 1 group.
the patients was 24 yr (range 22-26 yr), mean weight was 66 ± 3 kg. These patients were selected for study as having been diabetic for >10 yr (mean 18 ± 2 yr, range 15-20 yr). All patients were ketosis prone; 10 had retinopathy (proliferative in 7), and 5 had clinically apparent nephropathy. All were being treated with one or two doses of mixed intermediate and short-acting insulin before the study, had fasting C-peptide of 0.6 ±0.10 nM, and an average HbA,c of 14.2% (range 16.5-11.8%). All patients were explained the study's aim, nature, and risk, in detail. Each gave written consent to the study, which itself was approved by the ethical committee of the Declaration of Helsinki and performed according to its principles. After careful investigation, patients were divided into two groups: six subjects for continuous subcutaneous insulin infusion (CSII) and six for intraportal insulin infusion (IPII). Five healthy, young, nonobese, nondiabetic men (mean age 28 ± 2 yr, mean weight 74 ± 2 kg, mean height 175 ± 2 cm) also were studied. None of this control group had a family history of IDDM, nor was" taking any drug known to affect carbohydrate metabolism or liver function. These subjects were determined nondiabetic according to the National Diabetes Data Group criteria for fasting plasma glucose and 2-h oral glucose-tolerance test (11). For 4 mo, patients were studied once a week, including 24-h profiles of plasma free insulin and glucose, values of HbA,c, C-peptide, regiment in total insulin dosage, weight, hormone concentrations (glucagon, cortisol, growth hormone), and intermediary metabolites (lactate, pyruvate, alanine, glycerol, nonesterified fatty acids [NEFA]). A catheter for withdrawal of blood samples was placed in an antecubital vein. All tests were conducted after an overnight fast. Intermediate or long-acting insulin had
DIABETES, VOL 41, SEPTEMBER 1992
been discontinued for 36 h before the study, and regular insulin was administered for >12 h before data collection. The insulin used for this study was Actrapid HM, U-40 (Novo, Copenhagen). Dosages of administered insulin were adjusted by measuring glucose levels, and their levels were ranged in near-normal values. Throughout the study, patients received five meals daily that corresponded with time of breakfast (0800), lunch (1200), dinner (1400), supper (1800), and a snack (2200), as a precaution of hypoglycemia. The meals comprised 2500 cal (50% carbohydrates, 30% proteins, 20% lipids). An AS6MP syringe pump (Travenol, Savage, MD) was used for insulin infusion in the CSII group. The infusion algorithm was basal insulin of 26 U/day—boluses of 6 U before breakfast, 2 U before lunch, 8 U before dinner, 6 U before supper, and 2 U before snack. Patients with IPII had the same algorithm, i.e., 50% of dosage requirement for meals and 50% basally, as is found physiologically). Surgical procedures. Six IDDM patients requiring abdominal exploration and hepatic portography were selected to verify presence or absence of liver pathology that could have influenced valid information in patients who were to receive intraportal insulin. After this study and an overnight fast, patients were prepared for surgical procedure. Under sterile conditions, with local anesthesia of 1% Novocain (Minmedbioprom, Lvovpharm, CIS), the umbilical vein was exposed through an extraperitoneal upper abdominal midline incision and cannulated with a Port-A-Cath catheter. Then the catheter was passed into the left branch of the portal vein just past its entrance into the liver. After the catheter was secured in place, Port-A-Cath was subcutaneously implanted on the lamina anterior vagina musculi recti abdominis and secured by four nonabsorbable ligatures. The implanted
1043
COMPARISON OF INSULIN INFUSIONS Sn
g
OSE
1210"
o
3
8
s
6 "
1
9
14
w
12
8
10
-
4 "
450 ^
400
400
P.
~
%
35C •
3*0-
3
3C0 •
| °° 8 25 °
2C0
S
200 •
c
10 10C 10
12 14
20
22 24
2
8.00
8.30
9.00
9.30
10.00 10.30 11.00 11.30 12.00
TIME (hours) 'i'lMi! (hours)
FIG. 1. Levels of glucose and serum free insulin concentrations during 24-h monitoring after 4 mo of the CSII and IPII. B, breakfast; L, lunch; D, dinner; S, supper; Sn, snack; (•—•) nondiabetic control subjects; (• •) diabetic patients with intraportal insulin infusion; (O) diabetic patients with continuous subcutaneous Insulin Infusion.
system was rinsed with solution heparin before taking in the wound. Immediately after implantation, the needle of the AS6MP syringe pump was put transcutaneously into the implanted Port-A-Cath, and the pump was started. Six subjects undergoing this procedure encountered neither complication nor infection. Assay methodology. Plasma glucose concentration was measured by a glucose- oxidase method (P02 technique Beckman analyzer, Fullerton, CA) six times daily, and plasma free insulin was measured by double-antibody radioimmunoassay after precipitation with 25% polyethylene glycol to remove endogenous insulin antibodies (12). Glucagon samples were collected in aprotinin EDTA and immediately centrifuged and stored at -20°C until assay was studied using wick chromatography (13). Cortisol and growth hormone were measured by radioimmunoassay (14). Metabolite samples were collected into weighted tubes containing chilled 0.5-M perchloric acid. Lactate, pyruvate, alanine, and glycerol were assayed using an automated enzymatic fluorometric technique, and NEFAs were assayed by radiocobalt dilution (15). HbA,c was measured by cation-exchange chromatography (Bio-Rad, Richmond, CA). C-peptide was assayed with a kit from Byk-Mallinkrodt (Germany). Results are expressed as mean ± SE. Statistical significance of the differences in glucose, serum free insulin, metabolite values, and hormone concentrations was analyzed by using Student's t test.
1044
FIG. 2. Mean plasma glucose and serum free insulin concentrations for 4-h monitoring In diabetic patients. Insulin infusion was begun at 0.2 U/kg at 0800. Simultaneously with the start of insulin infusion, all subjects began to eat a 650-cal breakfast. • — • , nondiabetic control subjects; • •, diabetic patients with Intraportal Insulin Infusion; O O, diabetic patients with continuous subcutaneous insulin Infusion.
RESULTS There were no significant variations in body weight during the CSII study, but weight was elevated 3 ± 1 kg during IPII study over 4 mo. The 24-h mean plasma glucose concentrations are shown in Table 1. Glucose control in the two treatment groups did not differ significantly at the beginning of study (IPII, 14.5 ± 1.7 vs. CSII, 15.0 ± 1.5 mM), but the IPII group had lower glucose values after 1 mo of therapy (IPII, 6.0 ± 2.4 vs. CSII, 10.2 ± 1.2 mM, P < 0.05) and after 4 mo (IPII, 4.8 ± 0.5 vs. CSII, 6.8 ± 0.2 mM, P < 0.05). When insulin was infused via the umbilical vein, normoglycemia was reached in the 1st mo, but'CSII led to near normoglycemia at the end of 4 mo. The mean 24-h blood glucose profile in both groups at 4 mo is shown in Figure 1. Although glycemic control was satisfactory in both groups, mean glucose levels were lower at several comparable points during IPII. Glucose excursion was increased more significantly after the breakfast of 650 cal in the CSII group (60 min after ingestion, 12.8 ± 0.9 mM; 90 min, 10 ± 1.4 mM; 120 min, 9.0 ± 0.4 mM) than in the IPII group (60 min after ingestion, 7.8 ± 0.5 mM; 90 min, 8.5 ± 0.3 mM; 120 min, 6.5 ± 0.2 mM, P < 0.05, Figure 2). The rise was most rapid during CSII; this difference became significant at 30 min after breakfast (11.0 ±0.2 vs. 6.0 ± 0.5 mM, P < 0.05). Intraportal administration of insulin at rates of
DIABETES, VOL 41, SEPTEMBER 1992
P.I. SHISHKO AND ASSOCIATES
0.01-0.05 U • kg"1 • hr 1 significantly reduced plasma glucose levels (from 14.0 ± 1.5 to 5.0 ± 0.6 mM, P < 0.05), and this reduction is manifested only when the rate of the peripheral insulin infusion is increased to 0.2-0.3 U-kg" 1 • hr 1 . Before initiation of CSII, HbA,c was 14.8 ± 1.5% (range 12.8-16.9%), and it was 13.8 ±1.5% (range 12.216.0%) before IPII. After CSII was begun, HbA,c declined to 10.0 ± 0.4%, but after IPII, HbA|C declined more significantly to 5.5 ± 0.4% after 4 mo, P < 0.05. The fall in HbA,c values occurred earlier during IPII than CSII (Table 1)At baseline 0 mo, there was no difference in the mean 24-h plasma free insulin concentration between patients with CSII (260 ± 22.3 pM) and those with IPII (270 ± 8.6 pM), Table 1. After 2 days of therapy, mean 24-h plasma free insulin concentration was significantly lower in the IPII group (100 ± 12.6 pM) than in CSII (260 ± 22.3 pM), P < 0.005), Figure 1. In the CSII group, the insulin plasma fasting level was 250 ± 23 pM, then rose to 388 ± 37 pM during 120 min after the start of a high-dose insulin injection (0.2 U/kg) infused before breakfast. After bolus insulin injection, there was a slow decline in plasma free insulin concentration, and it remained elevated at 280-350 pM for 4 h, Figure 2. In the IPII group, the insulin plasma fasting level was 88 ± 10.7 pM and rose to a maximum 240 ± 14.2 pM at 60 min of postinitiation of the 0.2-U/kg insulin infusion followed by a decline for 120 min, reaching a level of
150 ±7.1 pM. Further decline in plasma free insulin concentration was observed to 112 ± 12.6 pM at the 4th h after insulin infusion initiation and was not different from concentrations of the control subjects (100 ± 8.6 pM), Figure 2. Total insulin dosage in the CSII and IPII groups before study (0 mo) did not differ (53 ± 4 and 52 ± 3 U/day, respectively), although there tended to be a slightly lower insulin requirement during IPII (at 2 mo, 10% less than CSII; at 4 mo, 20-25% less; Table 1) The mean total insulin dosage in the IPII group differed from the baseline compared with the CSII group at 4 mo (IPII group at 4 mo, 40 ± 2 U/day and at baseline, 52 ± 3 U/day, P < 0.05; CSII group at 4 mo, 50 ± 2 U/day, P < 0.05, Table 1). Plasma C-peptide concentrations were at 0.6 ±0.18 nM in both groups at any observation point during this study. Dynamic measurement of intermediary metabolites showed that lactate, pyruvate, and alanine concentrations remained elevated throughout the 4 mo in the CSII group and were decreased to normal concentrations at 1 mo in the IPII group (Table 2). At 1 mo, lactate was CSII, 2.50 ± 0.03 vs. IPII, 0.82 ± 0.07 mM, P < 0.001; pyruvate was CSII, 0.370 ± 0.031 vs. IPII, 0.071 ± 0.012 mM, P < 0.001; and alanine was CSII, 2.21 ± 0.20 vs. IPII, 0.26 ± 0.10 mM, P < 0.001. A decrease in glycerol and NEFA concentrations was observed in both groups, but was more significant in CSII. At 4 mo, glycerol was CSII, 0.042 ±0.004 vs. IPII, 0.072 ±0.005 mM, P < 0.05;
TABLE 2 Intermediary metabolites for 4 mo of study Month (mM) Lactate Control CSII IPII Pyruvate Control CSII IPII Alanine Control CSII IPII Glycerol Control CSII IPII MPPA INCrM
Control CSII IPII
0
1
2
3
4
0.80 ± 0.03 2.80 ± 0.05* 2.70 ± 0.04*
0.78 ± 0.04 2.50 ± 0.03* 0.82 ± 0.07§
0.78 ± 0.03 2.30 ± 0.05* 0.78 ± 0.06
0.80 ± 0.04 2.00 ± 0.04 0.76 ± 0.02||t
0.78 ± 0.02 1.72±0.03t* 0.75 ± 0.03||H
0.076 ± 0.004 0.437 ± 0.027* 0.439 ± 0.029*
0.079 ± 0.005 0.370 ±0.031** 0.071 ± 0.012§H
0.074 ± 0.002 0.35 ±0.015* 0.070 ± 0.02§
0.079 ± 0.007 0.324 ±0.021* 0.072 ± 0.011§
0.075 ± 0.004 0.251 ± 0.025* 0.072 ± 0.008§H
0.25 ± 0.02 2.85 ± 0.23* 2.76 ±0.15*
0.23 ± 0.04 2.21 ± 0.20* 0.26 ± 0.10§H
0.25 ± 0.04 1.95 ±0.21** 0.28 ± 0.17§
0.21 ± 0.03 1.51 ±0.10* 0.27 ± 0.07§
0.22 ± 0.04 1.47 ±0.20** 0.24 ± 0.04§U
0.042 ± 0.005 0.151 ± 0.034* 0.164 ±0.041*
0.044 ± 0.003 0.136 ±0.012* 0.151 ±0.020*
0.046 ± 0.004 0.095 ± 0.020t* 0.120 ±0.20||
0.044 ± 0.003 0.056 ± 0.007 0.109 ± 0.017t*
0.040 ± 0.004 0.042 ± 0.004 0.072 ± 0.005t|
0.78 ± 0.02 1.47±0.19t 1.61 ±0.27f
0.75 ± 0.02 1.01 ±0.04t 1.49±0.09t||
0.74 ± 0.04 0.89 ± 0.09 1.21 ± 0.04t||*
0.72 ± 0.05 0.58 ± 0.06 0.96 ± 0.05t||
0.75 ± 0.03 0.62 ± 0.02 0.85 ± 0.03||*
Values are means ± SE. CSII, diabetic patients during continuous subcutaneous insulin infusion. IPII, diabetic patients during intraportal insulin infusion; NEFA, nonesterified fatty acid. *P < 0.001 in comparison with control values. fP < 0.05 in comparison with control values *P < 0.05 in comparison with 0 mo within 1 group. §P < 0.001 comparison between IPII and CSII groups. \\P < 0.05 comparison between IPII and CSII groups. IIP < 0.001 in comparison with 0 mo within 1 group.
DIABETES, VOL. 41, SEPTEMBER 1992
1045
| COMPARISON OF INSULIN INFUSIONS
TABLE 3 Intermediary metabolite values in response to breakfast of 650 cal and insulin infusion at 0.05 U•kg" 1 • hr~1 in patients Time (min) Rlnod LJI\J\J\J metabolite Lactate (mM) Control CSII IPII Pyruvate (mM) Control CSII IPII
0
30
60
90
120
0.78 ± 0.03 1.98 ± 0.08* 0.72 ± 0.02§
0.80 ± 0.02 2.15 ±0.04* 0.80 ± 0.04§
0.91 ± 0.05 2.48 ± 0.03* 0.94 ± 0.03§t
0.85 ± 0.06 2.10 ± 0.06* 0.87 ± 0.04§t
0.72 ± 0.04 2.10 ±0.02* 0.70 ± 0.05||
0.079 ± 0.008 0.248 ± 0.012* 0.072 ± 0.004||
0.088 ± 0.008 0.295 ± 0.025* 0.087 ± 0.005
0.095 ± 0.008t 0.348 ± 0.020t* 0.099 ± 0.002||t
0.087 ± 0.005 0.300 ± 0.009t* 0.088 ± 0.005
0.075 ± 0.004 0.295 ± 0.008* 0.070 ± 0.008||
0.20 ± 0.01 1.50 ± 0 . 0 1 * 0.25 ± 0.03||
0.25 ± 0.02 1.70 ± 0 . 0 1 * 0.26 ±0.02
0.28 ± 0.02 1.85 ± 0.02f* 0.30±0.01||
0.30 ± 0.02t 1.80 ±0.02* 0.32 ± 0.02
0.25 ± 0.02 1.75 ±0.02* 0.28 ± 0.02||
0.048 ± 0.003 0.030 ± 0.004* 0.045 ±0.001*
0.035 ± 0.002 0.034 ± 0.003 0.043 ± 0.002
0.028 ± 0.008t 0.025 ± 0.004 0.035 ± 0.001
0.025 ± 0.002t 0.025 ± 0.004 0.039 ± 0.004*
0.028 ± 0.005 0.027 ± 0.004 0.035 ± 0.005
0.75 ± 0.02 0.40 ± 0.07* 0.76 ± 0.07t
0.52 ± 0.04 0.45 ± 0.02 0.77 ± 0.05*
0.44 ± 0.05 0.35 ± 0.04 0.64 ± 0.02*
0.35 ± 0.02t 0.30 ± 0.04 0.57 ± 0.04*t
A lonlno
Mianine (mM) Control CSII IPII Glycerol (mM) Control CSII IPII NEFA (pi) (mM) Control CSII IPII
0.40 ± 0.02f 0.30 ± 0.04 0.53 ± 0.05*t
Values are means ± SE. CSII, diabetic patients during continuous subcutaneous insulin infusion; IPII, diabetic patients during intraportal insulin infusion; NEFA (pi), nonesterified fatty acid, plasma. *P < 0.05 in comparison with control values. t P < 0.05 in comparison with 0 min within 1 group. *P < 0.001 in comparison with control values. §P < 0.05 comparison between IPII and CSII groups. \\P < 0.001 comparison between IPII and CSII groups.
NEFA was CSII, 0.62 ± 0.02 vs. IPII, 0.85 ± 0.03 mM, P
Twelve subjects with insulin-dependent diabetes mellitus were treated using continuous subcutaneous insulin infusion (CSII) and intraportal insulin infusion (IPII) via the umbilical vein for 4 mo. Glucose control improved in both CSII and IPII groups, but a decrease in glucose and HbAlc was more rapid and more significant in the IPII group than in CSII, even though insulin requirement was lower during IPII than CSII (40 ± 2 vs. 50 ± 2 U/day, P < 0.05). The insulin plasma fasting levels were different (88 ± 10.7 in the IPII group vs. 263 ± 23 pM in CSII, P < 0.001). High plasma levels of lactate, pyruvate, alanine, cortisol, and growth hormone were decreased in both groups, with their full normalization only in the IPII group. Glucagon concentrations were low in both groups at the beginning of the study (30.0 ± 4.1 in the CSII group and 32.3 ± 1 . 8 ng/L in IPII); they were equalized to control values in the IPII group and were low in the CSII group at the study's end (46.0 ± 3.7 in IPII vs. 31.7 ± 3.1 ng/L in CSII, P < 0.05. We conclude that intraportal administration of insulin via the umbilical vein at rates of 0.01-0.05 U • kg"1 hr 1 reduces plasma levels of glucose, three carbon precursors, cortisol, and growth hormone by a direct action on the liver, and the hepatic action of peripherally administered insulin is manifested only when the infusion rate is increased to 0.1-0.3 U • kg'1 • hr 1 . Thus the intraportal route via the umbilical vein may offer improved physiological insulin delivery over the subcutaneous route. Diabetes 41:1042-49,1992
From the Institute of Nutrition, Clinical Department, USSR Academy of Medical Sciences, and the Surgical Department, Vladivostok Medical Institute, Moscow, Russia, CIS. Address all correspondence and reprint requests to Dr. P.I. Shishko, Clinical Department, Institute of Nutrition, USSR Academy of Medical Sciences, 2/14 Ustinsky Proezd, 109240 Moscow, Russia, CIS. Received for publication 10 April 1991 and accepted in revised form 11 March 1992.
1042
N
ormal blood glucose control is the aim of new intensive treatment regiments for insulin-dependent diabetic mellitus (IDDM) patients. New modes of insulin delivery (closed- and openloop) systems have been used to provide better physiological insulin replacement. Normalization of blood glucose is possible with an artificial p-cell (Biostator; 1 Life Science Instruments, Miles, Elkhart, IN). However, during treatment with the Biostator, insulin is administered via peripheral route, therefore several metabolic parameters remain pathological (2,3). On the other hand, the functional significance of insulin infusion into the portal vein has not been established. Numerous studies indicate that vascular connection between pancreatic (3-cells and the liver is not important for the metabolic actions of insulin (4-6). The hormone's direct infusion into the liver via the hepatic portal vein had no different effect on glucose metabolism than did its delivery into the peripheral vein (6). Nevertheless, some studies performed on animals suggest that portal insulin delivery should be required to achieve normalization of metabolic disturbances (7-10). Thus, we compared glucose, serum free insulin levels, intermediary metabolites, and hormone profiles achieved with portal or peripheral insulin infusion over 4 mo. We also compared the metabolic status in similar glycemic circumstances after a breakfast of 650 cal and insulin infusion at 0.05 U • kg"1 • hr"1 in IDDM subjects and compared these results with those for nondiabetic control subjects. In this study, we used a catheterization of the portal vein via the umbilical vein and implantation of the PortA-Cath system (Pharmacia, Piscataway, NJ) for chronic insulin infusion.
RESEARCH DESIGN AND METHODS Twelve IDDM patients participated in this study. Their clinical characteristics are shown in Table 1. Mean age of
DIABETES, VOL. 41, SEPTEMBER 1992
P.I. SHISHKO AND ASSOCIATES
TABLE 1 Clinical characteristics of patients on entry to the study and after 4 mo of therapy
Patients (n) Control (5) CSII (6) 0.0 mo 1.0 mo 2.0 mo 3.0 mo 4.0 mo IPII (6) 0.0 mo 1.0 mo 2.0 mo 3.0 mo 4.0 mo
HbA 1c
(%)
5.5 ± 0.8
Glucose (mM)
Total insulin dose (U/day)
4.5 ± 1.5
IRI (pM)
C-peptide (nM)
86.3 ± 10.8
3.6 ± 0.36
Duration of diabetes (yr)
Age (yr)
Weight (kg)
28 ± 2
74 ± 2
15.0 10.2 8.2 7.8 6.8
± 1.5* ± 1.2t ± 0.8 ± 0.4|| ± 0.211
53 ± 4 52 ± 2 52 ± 2 52 ± 2 50±2§
260 ± 22.3* 263 ± 23* 259 ± 22 250 dt 13 248 dt 6.5**
0.6 0.6 0.6 0.6 0.6
± 0.06* ± 0.15 ±0.12 ± 0.24 ± 0.21
18±2
24 ± 2
1 4 . 2 : t 1.1 13.5 dt 0.5 12.2 dfcO.2§ 10.0 dt 0.4§t
66 66 66 67 67
dt 3 dt 3 dt 3 dfc2 db 2
13.8 dt 1.5* 13.0 dt 1.0 11.1 dt 0.5|| 6.9 dfc0.4 5.5 dbO.4f
14.5 6.0 5.5 5.0 4.8
± 1.7* ± 2.4||§1I ± 1.5 ± 0.5 ± 0.51
52 ± 3 48 ± 2 40±2t 40 ± 2 40±2§
270 dt8.6* 88 dt 10.7U 73 dt 7.1 79 dt8.6 82 ± 5.7
0.6 0.6 0.6 0.6 0.6
± 0.06* ±0.15 ± 0.09 ± 0.06 ±0.12
18±2
24 ± 2
67 69 70 70 70
db 2 dt 2 dt 1 dt 1 dt 1
14.8 i t 1.5*
Values are mean ± SE. IRI, immune-reactive insulin, n, number of patients. CSII, diabetic patients during continuous subcutaneous insulin infusion. IPII, diabetic patients during intraportal insulin infusion. *P < 0.001 in comparison with control values. t P < 0.05 in comparison with 0 mo within one group. *P < 0.001 comparison between IPII and CSII groups. §P < 0.05 comparison between IPII and CSII groups. \\P < 0.05 in comparison with control values. IIP < 0.001 in comparison with 0 mo within 1 group.
the patients was 24 yr (range 22-26 yr), mean weight was 66 ± 3 kg. These patients were selected for study as having been diabetic for >10 yr (mean 18 ± 2 yr, range 15-20 yr). All patients were ketosis prone; 10 had retinopathy (proliferative in 7), and 5 had clinically apparent nephropathy. All were being treated with one or two doses of mixed intermediate and short-acting insulin before the study, had fasting C-peptide of 0.6 ±0.10 nM, and an average HbA,c of 14.2% (range 16.5-11.8%). All patients were explained the study's aim, nature, and risk, in detail. Each gave written consent to the study, which itself was approved by the ethical committee of the Declaration of Helsinki and performed according to its principles. After careful investigation, patients were divided into two groups: six subjects for continuous subcutaneous insulin infusion (CSII) and six for intraportal insulin infusion (IPII). Five healthy, young, nonobese, nondiabetic men (mean age 28 ± 2 yr, mean weight 74 ± 2 kg, mean height 175 ± 2 cm) also were studied. None of this control group had a family history of IDDM, nor was" taking any drug known to affect carbohydrate metabolism or liver function. These subjects were determined nondiabetic according to the National Diabetes Data Group criteria for fasting plasma glucose and 2-h oral glucose-tolerance test (11). For 4 mo, patients were studied once a week, including 24-h profiles of plasma free insulin and glucose, values of HbA,c, C-peptide, regiment in total insulin dosage, weight, hormone concentrations (glucagon, cortisol, growth hormone), and intermediary metabolites (lactate, pyruvate, alanine, glycerol, nonesterified fatty acids [NEFA]). A catheter for withdrawal of blood samples was placed in an antecubital vein. All tests were conducted after an overnight fast. Intermediate or long-acting insulin had
DIABETES, VOL 41, SEPTEMBER 1992
been discontinued for 36 h before the study, and regular insulin was administered for >12 h before data collection. The insulin used for this study was Actrapid HM, U-40 (Novo, Copenhagen). Dosages of administered insulin were adjusted by measuring glucose levels, and their levels were ranged in near-normal values. Throughout the study, patients received five meals daily that corresponded with time of breakfast (0800), lunch (1200), dinner (1400), supper (1800), and a snack (2200), as a precaution of hypoglycemia. The meals comprised 2500 cal (50% carbohydrates, 30% proteins, 20% lipids). An AS6MP syringe pump (Travenol, Savage, MD) was used for insulin infusion in the CSII group. The infusion algorithm was basal insulin of 26 U/day—boluses of 6 U before breakfast, 2 U before lunch, 8 U before dinner, 6 U before supper, and 2 U before snack. Patients with IPII had the same algorithm, i.e., 50% of dosage requirement for meals and 50% basally, as is found physiologically). Surgical procedures. Six IDDM patients requiring abdominal exploration and hepatic portography were selected to verify presence or absence of liver pathology that could have influenced valid information in patients who were to receive intraportal insulin. After this study and an overnight fast, patients were prepared for surgical procedure. Under sterile conditions, with local anesthesia of 1% Novocain (Minmedbioprom, Lvovpharm, CIS), the umbilical vein was exposed through an extraperitoneal upper abdominal midline incision and cannulated with a Port-A-Cath catheter. Then the catheter was passed into the left branch of the portal vein just past its entrance into the liver. After the catheter was secured in place, Port-A-Cath was subcutaneously implanted on the lamina anterior vagina musculi recti abdominis and secured by four nonabsorbable ligatures. The implanted
1043
COMPARISON OF INSULIN INFUSIONS Sn
g
OSE
1210"
o
3
8
s
6 "
1
9
14
w
12
8
10
-
4 "
450 ^
400
400
P.
~
%
35C •
3*0-
3
3C0 •
| °° 8 25 °
2C0
S
200 •
c
10 10C 10
12 14
20
22 24
2
8.00
8.30
9.00
9.30
10.00 10.30 11.00 11.30 12.00
TIME (hours) 'i'lMi! (hours)
FIG. 1. Levels of glucose and serum free insulin concentrations during 24-h monitoring after 4 mo of the CSII and IPII. B, breakfast; L, lunch; D, dinner; S, supper; Sn, snack; (•—•) nondiabetic control subjects; (• •) diabetic patients with intraportal insulin infusion; (O) diabetic patients with continuous subcutaneous Insulin Infusion.
system was rinsed with solution heparin before taking in the wound. Immediately after implantation, the needle of the AS6MP syringe pump was put transcutaneously into the implanted Port-A-Cath, and the pump was started. Six subjects undergoing this procedure encountered neither complication nor infection. Assay methodology. Plasma glucose concentration was measured by a glucose- oxidase method (P02 technique Beckman analyzer, Fullerton, CA) six times daily, and plasma free insulin was measured by double-antibody radioimmunoassay after precipitation with 25% polyethylene glycol to remove endogenous insulin antibodies (12). Glucagon samples were collected in aprotinin EDTA and immediately centrifuged and stored at -20°C until assay was studied using wick chromatography (13). Cortisol and growth hormone were measured by radioimmunoassay (14). Metabolite samples were collected into weighted tubes containing chilled 0.5-M perchloric acid. Lactate, pyruvate, alanine, and glycerol were assayed using an automated enzymatic fluorometric technique, and NEFAs were assayed by radiocobalt dilution (15). HbA,c was measured by cation-exchange chromatography (Bio-Rad, Richmond, CA). C-peptide was assayed with a kit from Byk-Mallinkrodt (Germany). Results are expressed as mean ± SE. Statistical significance of the differences in glucose, serum free insulin, metabolite values, and hormone concentrations was analyzed by using Student's t test.
1044
FIG. 2. Mean plasma glucose and serum free insulin concentrations for 4-h monitoring In diabetic patients. Insulin infusion was begun at 0.2 U/kg at 0800. Simultaneously with the start of insulin infusion, all subjects began to eat a 650-cal breakfast. • — • , nondiabetic control subjects; • •, diabetic patients with Intraportal Insulin Infusion; O O, diabetic patients with continuous subcutaneous insulin Infusion.
RESULTS There were no significant variations in body weight during the CSII study, but weight was elevated 3 ± 1 kg during IPII study over 4 mo. The 24-h mean plasma glucose concentrations are shown in Table 1. Glucose control in the two treatment groups did not differ significantly at the beginning of study (IPII, 14.5 ± 1.7 vs. CSII, 15.0 ± 1.5 mM), but the IPII group had lower glucose values after 1 mo of therapy (IPII, 6.0 ± 2.4 vs. CSII, 10.2 ± 1.2 mM, P < 0.05) and after 4 mo (IPII, 4.8 ± 0.5 vs. CSII, 6.8 ± 0.2 mM, P < 0.05). When insulin was infused via the umbilical vein, normoglycemia was reached in the 1st mo, but'CSII led to near normoglycemia at the end of 4 mo. The mean 24-h blood glucose profile in both groups at 4 mo is shown in Figure 1. Although glycemic control was satisfactory in both groups, mean glucose levels were lower at several comparable points during IPII. Glucose excursion was increased more significantly after the breakfast of 650 cal in the CSII group (60 min after ingestion, 12.8 ± 0.9 mM; 90 min, 10 ± 1.4 mM; 120 min, 9.0 ± 0.4 mM) than in the IPII group (60 min after ingestion, 7.8 ± 0.5 mM; 90 min, 8.5 ± 0.3 mM; 120 min, 6.5 ± 0.2 mM, P < 0.05, Figure 2). The rise was most rapid during CSII; this difference became significant at 30 min after breakfast (11.0 ±0.2 vs. 6.0 ± 0.5 mM, P < 0.05). Intraportal administration of insulin at rates of
DIABETES, VOL 41, SEPTEMBER 1992
P.I. SHISHKO AND ASSOCIATES
0.01-0.05 U • kg"1 • hr 1 significantly reduced plasma glucose levels (from 14.0 ± 1.5 to 5.0 ± 0.6 mM, P < 0.05), and this reduction is manifested only when the rate of the peripheral insulin infusion is increased to 0.2-0.3 U-kg" 1 • hr 1 . Before initiation of CSII, HbA,c was 14.8 ± 1.5% (range 12.8-16.9%), and it was 13.8 ±1.5% (range 12.216.0%) before IPII. After CSII was begun, HbA,c declined to 10.0 ± 0.4%, but after IPII, HbA|C declined more significantly to 5.5 ± 0.4% after 4 mo, P < 0.05. The fall in HbA,c values occurred earlier during IPII than CSII (Table 1)At baseline 0 mo, there was no difference in the mean 24-h plasma free insulin concentration between patients with CSII (260 ± 22.3 pM) and those with IPII (270 ± 8.6 pM), Table 1. After 2 days of therapy, mean 24-h plasma free insulin concentration was significantly lower in the IPII group (100 ± 12.6 pM) than in CSII (260 ± 22.3 pM), P < 0.005), Figure 1. In the CSII group, the insulin plasma fasting level was 250 ± 23 pM, then rose to 388 ± 37 pM during 120 min after the start of a high-dose insulin injection (0.2 U/kg) infused before breakfast. After bolus insulin injection, there was a slow decline in plasma free insulin concentration, and it remained elevated at 280-350 pM for 4 h, Figure 2. In the IPII group, the insulin plasma fasting level was 88 ± 10.7 pM and rose to a maximum 240 ± 14.2 pM at 60 min of postinitiation of the 0.2-U/kg insulin infusion followed by a decline for 120 min, reaching a level of
150 ±7.1 pM. Further decline in plasma free insulin concentration was observed to 112 ± 12.6 pM at the 4th h after insulin infusion initiation and was not different from concentrations of the control subjects (100 ± 8.6 pM), Figure 2. Total insulin dosage in the CSII and IPII groups before study (0 mo) did not differ (53 ± 4 and 52 ± 3 U/day, respectively), although there tended to be a slightly lower insulin requirement during IPII (at 2 mo, 10% less than CSII; at 4 mo, 20-25% less; Table 1) The mean total insulin dosage in the IPII group differed from the baseline compared with the CSII group at 4 mo (IPII group at 4 mo, 40 ± 2 U/day and at baseline, 52 ± 3 U/day, P < 0.05; CSII group at 4 mo, 50 ± 2 U/day, P < 0.05, Table 1). Plasma C-peptide concentrations were at 0.6 ±0.18 nM in both groups at any observation point during this study. Dynamic measurement of intermediary metabolites showed that lactate, pyruvate, and alanine concentrations remained elevated throughout the 4 mo in the CSII group and were decreased to normal concentrations at 1 mo in the IPII group (Table 2). At 1 mo, lactate was CSII, 2.50 ± 0.03 vs. IPII, 0.82 ± 0.07 mM, P < 0.001; pyruvate was CSII, 0.370 ± 0.031 vs. IPII, 0.071 ± 0.012 mM, P < 0.001; and alanine was CSII, 2.21 ± 0.20 vs. IPII, 0.26 ± 0.10 mM, P < 0.001. A decrease in glycerol and NEFA concentrations was observed in both groups, but was more significant in CSII. At 4 mo, glycerol was CSII, 0.042 ±0.004 vs. IPII, 0.072 ±0.005 mM, P < 0.05;
TABLE 2 Intermediary metabolites for 4 mo of study Month (mM) Lactate Control CSII IPII Pyruvate Control CSII IPII Alanine Control CSII IPII Glycerol Control CSII IPII MPPA INCrM
Control CSII IPII
0
1
2
3
4
0.80 ± 0.03 2.80 ± 0.05* 2.70 ± 0.04*
0.78 ± 0.04 2.50 ± 0.03* 0.82 ± 0.07§
0.78 ± 0.03 2.30 ± 0.05* 0.78 ± 0.06
0.80 ± 0.04 2.00 ± 0.04 0.76 ± 0.02||t
0.78 ± 0.02 1.72±0.03t* 0.75 ± 0.03||H
0.076 ± 0.004 0.437 ± 0.027* 0.439 ± 0.029*
0.079 ± 0.005 0.370 ±0.031** 0.071 ± 0.012§H
0.074 ± 0.002 0.35 ±0.015* 0.070 ± 0.02§
0.079 ± 0.007 0.324 ±0.021* 0.072 ± 0.011§
0.075 ± 0.004 0.251 ± 0.025* 0.072 ± 0.008§H
0.25 ± 0.02 2.85 ± 0.23* 2.76 ±0.15*
0.23 ± 0.04 2.21 ± 0.20* 0.26 ± 0.10§H
0.25 ± 0.04 1.95 ±0.21** 0.28 ± 0.17§
0.21 ± 0.03 1.51 ±0.10* 0.27 ± 0.07§
0.22 ± 0.04 1.47 ±0.20** 0.24 ± 0.04§U
0.042 ± 0.005 0.151 ± 0.034* 0.164 ±0.041*
0.044 ± 0.003 0.136 ±0.012* 0.151 ±0.020*
0.046 ± 0.004 0.095 ± 0.020t* 0.120 ±0.20||
0.044 ± 0.003 0.056 ± 0.007 0.109 ± 0.017t*
0.040 ± 0.004 0.042 ± 0.004 0.072 ± 0.005t|
0.78 ± 0.02 1.47±0.19t 1.61 ±0.27f
0.75 ± 0.02 1.01 ±0.04t 1.49±0.09t||
0.74 ± 0.04 0.89 ± 0.09 1.21 ± 0.04t||*
0.72 ± 0.05 0.58 ± 0.06 0.96 ± 0.05t||
0.75 ± 0.03 0.62 ± 0.02 0.85 ± 0.03||*
Values are means ± SE. CSII, diabetic patients during continuous subcutaneous insulin infusion. IPII, diabetic patients during intraportal insulin infusion; NEFA, nonesterified fatty acid. *P < 0.001 in comparison with control values. fP < 0.05 in comparison with control values *P < 0.05 in comparison with 0 mo within 1 group. §P < 0.001 comparison between IPII and CSII groups. \\P < 0.05 comparison between IPII and CSII groups. IIP < 0.001 in comparison with 0 mo within 1 group.
DIABETES, VOL. 41, SEPTEMBER 1992
1045
| COMPARISON OF INSULIN INFUSIONS
TABLE 3 Intermediary metabolite values in response to breakfast of 650 cal and insulin infusion at 0.05 U•kg" 1 • hr~1 in patients Time (min) Rlnod LJI\J\J\J metabolite Lactate (mM) Control CSII IPII Pyruvate (mM) Control CSII IPII
0
30
60
90
120
0.78 ± 0.03 1.98 ± 0.08* 0.72 ± 0.02§
0.80 ± 0.02 2.15 ±0.04* 0.80 ± 0.04§
0.91 ± 0.05 2.48 ± 0.03* 0.94 ± 0.03§t
0.85 ± 0.06 2.10 ± 0.06* 0.87 ± 0.04§t
0.72 ± 0.04 2.10 ±0.02* 0.70 ± 0.05||
0.079 ± 0.008 0.248 ± 0.012* 0.072 ± 0.004||
0.088 ± 0.008 0.295 ± 0.025* 0.087 ± 0.005
0.095 ± 0.008t 0.348 ± 0.020t* 0.099 ± 0.002||t
0.087 ± 0.005 0.300 ± 0.009t* 0.088 ± 0.005
0.075 ± 0.004 0.295 ± 0.008* 0.070 ± 0.008||
0.20 ± 0.01 1.50 ± 0 . 0 1 * 0.25 ± 0.03||
0.25 ± 0.02 1.70 ± 0 . 0 1 * 0.26 ±0.02
0.28 ± 0.02 1.85 ± 0.02f* 0.30±0.01||
0.30 ± 0.02t 1.80 ±0.02* 0.32 ± 0.02
0.25 ± 0.02 1.75 ±0.02* 0.28 ± 0.02||
0.048 ± 0.003 0.030 ± 0.004* 0.045 ±0.001*
0.035 ± 0.002 0.034 ± 0.003 0.043 ± 0.002
0.028 ± 0.008t 0.025 ± 0.004 0.035 ± 0.001
0.025 ± 0.002t 0.025 ± 0.004 0.039 ± 0.004*
0.028 ± 0.005 0.027 ± 0.004 0.035 ± 0.005
0.75 ± 0.02 0.40 ± 0.07* 0.76 ± 0.07t
0.52 ± 0.04 0.45 ± 0.02 0.77 ± 0.05*
0.44 ± 0.05 0.35 ± 0.04 0.64 ± 0.02*
0.35 ± 0.02t 0.30 ± 0.04 0.57 ± 0.04*t
A lonlno
Mianine (mM) Control CSII IPII Glycerol (mM) Control CSII IPII NEFA (pi) (mM) Control CSII IPII
0.40 ± 0.02f 0.30 ± 0.04 0.53 ± 0.05*t
Values are means ± SE. CSII, diabetic patients during continuous subcutaneous insulin infusion; IPII, diabetic patients during intraportal insulin infusion; NEFA (pi), nonesterified fatty acid, plasma. *P < 0.05 in comparison with control values. t P < 0.05 in comparison with 0 min within 1 group. *P < 0.001 in comparison with control values. §P < 0.05 comparison between IPII and CSII groups. \\P < 0.001 comparison between IPII and CSII groups.
NEFA was CSII, 0.62 ± 0.02 vs. IPII, 0.85 ± 0.03 mM, P
