.
`
`.
`
`.
`
`I
`
`.
`
`.
`
`The Intravenous, intrape_ri~oneal, and
`Subcutaneous Routes of Insulin
`Delivery in; Diabetic Man
`
`.
`
`•
`
`DAVID S. SCHADE, R. PHILIP EATON, NEAL FRIEDMAN, AND WILLIAM SPENCER
`
`SUMMARY
`Successful implantation of an artificial pancreas re(cid:173)
`quires the infusion of insulin into an appropriate ana(cid:173)
`tomic site. Three sites bein:g actively investigated in(cid:173)
`clude (1) intravenous (i.v.), ,(2) intraperitoneal (i.p.),
`and (3) subcutaneous (s.c.). This s'tudy compared the
`rate, magnitude, and clurati.on of insulin absorption
`from these three absorption sites as assessed by the
`appearance of "free" in·sul!n into the plasma of 10 in(cid:173)
`sulin-dependent dia betic sbbjects. The biologic effec(cid:173)
`tiveness of insulin was as~essed by the suppression
`of plasma glucose concentration following a 750-calo(cid:173)
`rie meal.
`Our results suggest that i.v. delivered insulin pro-·
`vides the ITIOSt rapid increase in plasma free insulin
`concentration, followed by ·the i.p. and s.c .. routes, re(cid:173)
`spectively. In contrast, the ielevatioh of plasma free in(cid:173)
`sulin concentration was most prolonged with the.s.c.
`route, followed by i.p. and j.v. routes, respectively.
`Compared with the i.v. and: s.c. routes of insulin deliv(cid:173)
`ery, only 50% of the i.p. deiivered insulin appeared in
`the plasma. The O!'!Set of tbe biologic activity of the in(cid:173)
`sulin delivere_d by the three different routes during the
`4½-h observation period was most rapid for the i.v.
`and least rapid for the s.c. ·route. These results sug(cid:173)
`gest that all three routes m.ay be appropriate sites for
`delivery of insuliri frQm an artificial panc_reas.
`However, because of the d_ifference in absorption ki(cid:173)
`netics and the onset of biologic effectiveness of tt,e
`delivered insulin, different :quantities and timing of In(cid:173)
`sulin delivery may be nee~ed. DIABETES 28 :1069- 1072 ,
`December 1979.
`
`. .
`
`.
`
`.
`
`I
`
`.
`
`.
`
`T he successful devel:opment and implantation of an
`
`artificial insulin delivery system into diabetic man
`requires an appropriate site for the infusion of insu(cid:173)
`lin. The major anatomic sites being actively inves-
`
`From the University of New Mexico School of Medicine. Albuquerque. New
`Mexico. and Sand ia Laboratories, Livermore. Ca lifornia.
`Address reprint requests to David S. Schade, M.D., University of New Mexico
`SchOol of Medicine, Albuquerque. New Mexico 87131.
`Received for publicat ion 19 July 1979.
`
`ligated include (1) intravenous (i.v.), '-i (2) intraperitoneal
`(i. p. ),_3 and (3) subcutaneous (s:c.). •.s Although there are ad(cid:173)
`vantages -and disadvantages to each administration site. to
`date no studies are available that compare the rate and de•
`gree of insulin absorption as measured by the appearance
`of insulin into the peripheral circulation.
`This study examined the increase in plasma free insulin
`concentration in diabetic subjects receiving a prepro~
`grammed infusion of exogenous insulin. Re.gular insulin (40
`U/ml) was infused inio either the i.v. , i.p., or s.c. site by
`means of a portable. rotary solenoid microliter-infusion
`pump. An assessment of the rapidity, magnitude, and dura(cid:173)
`tion of insul in absorption from these three sites was made
`using a randomized paired-study protocol in 10 diabetic
`men lacking endogenous insu lin secretion (no C-peptide
`secretion). In addition, the change in plasma glucose con(cid:173)
`centration following ingestion of identical meals was also
`examined as a measure of ihe onset of biologic activity of
`the insulin del ivered during the 4½-h observation period.
`
`METHODS
`.
`Subject population. Ten healthy, normal weight, male, in(cid:173)
`sulin-dependent diabetic subjects participated
`in
`ttie
`13½-h studies. Ages ranged from 22 to 39 yr and all sub(cid:173)
`jects were receiving eithe~ one or two injections per day of
`isophane insulin with or without the add ition of regular irisu(cid:173)
`lin. On the evening before study. the subjects were admitted
`to the Clinical Research Center of the University of New
`Mexico School of Medicine. No subject received insulifl
`between 7:30 p.m. and the start of the studies at 7:30 a.m.
`the next day. Before initiating these studies, informed
`written consent was obtained from each volunteer, and all
`studies were approved by the University of New Mexico
`·
`Human Research Review Committee.
`Protocol. On the day of study, each subject received three
`identical meals, which corresponded in time to breakfast,
`lunch, and supper at 7:30 a.m.: noon, and 4:30 p.m. Each
`mea l consisted of ham, fruit, bread, and milk comprising
`750 calories (40% carbohyqrate, 40% fat, and 20% protein).
`
`DIABETES. VOL 28, DECEMBER 1979
`
`1069
`
`FRESENIUS EXHIBIT 1052
`Page 1 of 4
`
`

`

`INSULIN DELI VE RY RO UTES
`
`Between meals, free access to water was allowed but no
`food was ingested.
`The 13½-h protocol was divided into three equar time
`periods of 4 ½ h each, during which each subject received
`one i.v .. one i.p., and one s.c. insulin infusion study in ran(cid:173)
`domized sequence. The protocol used in all studies em(cid:173)
`ployed an initial 30-min high dose insulin infusion of 5 U of
`regular insulin beginning simultaneously with the start of the
`meal. Following the 30-min high dose insulin infusion, the
`infusion pump reverted to the low dose insulin infusion rate
`of 2 U/h for three additional hours. Al the end of _the 31/? h, the
`insu lin delivery catheter was removed. The subsequent 60-
`min "no insulin" period permitted an assessment of contin(cid:173)
`ued insulin absorption from the infusion site.
`All insulin protocols used. a mini, peristaltic rotary sole(cid:173)
`noid pump to deliver the insulin.6 This pump was portable
`(6.0 x 7.5 x 10 cm) and battery operated. The pump was
`driven in a pulsati le fashion by a rotary solenoid, each pulse
`rotating the pump head by 10 degrees and delivering a vol(cid:173)
`ume of 2 µI/pulse. Reproducibility of ihe pump was ± 5% for
`a delivery volume of 72 µI. In all infusion studies, the insulin
`was stored in a 50-ml plastic reseNoir connected by a 10-
`cm plastic tube to the rotary solenoid pump. Before starting
`the individual infusion studies, 1 ml of the U40 insulin in the
`reservoir was pumped through the p iastic tubing to saturate
`insulin absorption sites.
`Infusion technique. insulin in all three studies was deliv(cid:173)
`ered through a 22-gauge plastic lntracath (Deseret Pharma(cid:173)
`ceutical Company, Sandy, Utah), 20 cm in length, attached
`to the insulin pump via a 40-cm connecting .tube. In the i.v.
`delivery of insulin study, the 19-gauge trocar was inserted
`into an antecubital vein, permitting placement of the plastic
`catheter approximately 5 cm into the vein.
`In the i.p. insulin infusion studies, the 22-gauge lntracath
`was inserted 3 cm below the umbilicus into the i.p .. space
`through a locally anesthetized area. The flexible plastic ln(cid:173)
`tracath catheter was "threaded" through a 19-gauge trocar
`into the peritoneal space to a depth of approximately 5 cm.
`This catheter was not rigid enough to pierce either a b lood
`vessel or the wall of the intestine. Once the catheter was
`threaded through the trocar, the trocar was removed and
`taped to the abdominal skin. Sterile technique was used
`throughout the procedure. Following the 3½-h peritoneal in(cid:173)
`fusion, the catheter was removed and cultured. In the 10
`subjects undergoing this procedure, no complications or in(cid:173)
`fections were encountered. Our experience indicates that
`this procedure is both safe and pain less and agrees with
`published reports in the literature on the safety of this tech(cid:173)
`nique.'
`In the s.c. insulin infusion study, an identical lntracath
`was p laced subcutaneously approximately 4 cm above the
`umbil icus. The trocar was inserted approximately 5 cm into
`the s.c. tissue and ihen withdrawn, leaving the plastic cath(cid:173)
`eter approximatel_y 4 cm within the s.c. space.
`Assay methodology. Plasma glucose concentration was
`measured by glucose oxidase and plasma free insulin by
`double antibody radioimmunoassay foll owing precipitation
`with 25% polyethylene glycol lo remove endogenous insulin
`antibodies as previously described. 3 C-peptide was as(cid:173)
`sayed with a kit from Calbiochem (Van Nuys, California).8
`Statistical .assess.ment. The changes in plasma free ,insu(cid:173)
`lin concentration in these diabetk subjects w.ere assessed
`
`125
`
`100
`
`75
`
`50
`
`25
`
`0
`20
`
`15
`
`10
`
`..,.
`e
`'3 :,
`-= a
`-=
`! ..
`
`ci.
`
`<:: g
`
`E
`
`:i:
`0
`~
`a,
`~ 5
`"' .,
`a
`
`0
`
`, - - - - low - - - - - - ott -
`
`- - intravenous delivery
`-
`-
`intraperltoneal
`..
`- ---·· subcutaneous
`
`l
`
`0 r MEAL
`
`2
`HOURS
`
`3
`
`4
`
`Figure 1. The changes In plasma free Insulin concentration (top) a1,1d
`glucose concentration (bottom) during programmed delivery of 11 U of
`regular Insulin ~Y three different routes in diabetic man. The Infusion
`of Insulin (U40-regular) was begun at a high dose (1 0 U/h) at time 0
`and continued for 30 min, when the Infusion pump automatically
`reverted to the low dose (2 U/h) for an additional 3 h. Simultaneously
`with the start of 'the high dose Insulin Infusion, each cif the 10 diabetic
`subjects bega_n to eat a 750-calorie meal (arrow). At the end of the
`inhi sloh period (3½ h) the insulin infusion was stopped for 60 min
`(off). Each of the 10 diabetic subjects participated in an intravenous,
`intraperitoneal, and subcutaneous insulin infu sion study iri a
`randomized sequence. The mean :t SEM of these 10 subjects is
`shown in the figure.
`
`by frequent venous blood sampling from the arm contra(cid:173)
`lateral to that which received the i.v. insulin. Four-milliliter
`blood samples were withdrawn via a #19 scalp vein needle
`every 15 or 30 min throughout the study. as shown in Figure
`1. Patency of this scalp vein needle was maintained by con(cid:173)
`tinuous infusion of isotonic saline (1 ml/min).
`The mean ::!: SE for insulin and glucose for all studies are
`plotted in Figure 1. To quantitate the amount of insulin ap(cid:173)
`pearing in the peripheral circulation during the 4 ½-h
`period, the increase in insulin concentration above the ini(cid:173)
`ti~I starting concentration at time O was assessed by inte(cid:173)
`gration. To estimate the rapidity of onset of biologic effec(cid:173)
`tiveness of the administered insulin, the increase of plasma
`glucose concentration over baseline levels was compared
`with Student's t test for paired data.
`
`RESULTS
`Mean plasma free insulin concentration (Figure 1 ). In the
`i.v. insulin infusion study, mean plasma free insulin concen(cid:173)
`tration rose to a maximum of 119 ± 13 µ U/ml by 15 min
`post initiation of infusion and then decreased rapidly fol(cid:173)
`lowing termination of the high dose infusion to an approxi(cid:173)
`matE! background level of 30 µU/ml. When the insulin infusion
`
`1070
`
`DIABETES, VOL: 2!), DEC.EMBER 1979
`
`FRESENIUS EXHIBIT 1052
`Page 2 of 4
`
`

`

`0/\VID S SCHADE /\NO ASSOCIATES
`
`was stopped at 3½ h, the p lasma tree insulin concentration
`decreased further to approximately 18 µU/ml, 5 µU/ml
`below the initial starting basal tree insulin concentration of
`22 ± 3 µ,U/ml (P < 0.0 1 ).
`In the i.p. insulin infusion study, mean plasma free insulin
`concentration rose to a maximal concentration of 37 ± 6
`µ,U/ml by 45 min postinitiation of the high dose insulin infu(cid:173)
`sion followed by a slow decline over a period of approxi(cid:173)
`mately 1 h, attaining a level of 26 µU/ml during the low dose
`insulin infusion. Following termination of the infusion. a fur(cid:173)
`ther decline in p lasma free insulin concentration was ob(cid:173)
`served to a level of 18 ± 1 µU/ml, not different from the ini(cid:173)
`tial starting basal free insulin concentration of 19 ± 3 µU/ml
`(P > 0.05).
`In the s.c. insulin infu sion study, mean plasma free insul in
`concentration rose to 40 ± 4 µU/ml by 2 h postinitiation of
`high dose infusion followed by a slow decline in plasma free
`insu lin concentration. After the infusion of insulin had
`ceased for 60 min, the plasma free insulin concentration re(cid:173)
`mained elevated at 29 ± 2 µU/ml (P < 0.001 ).
`In each study, plasma free insulin concentration demon(cid:173)
`strated a rise and subsequent fall. However, the timing and
`magnitude of this change depended on the route of insulin
`delivery. At 30 min following the initiation of insulin infusion,
`plasma free insulin concentration in the i. p. study (34 ± 4
`µU/ml) was significantly greater than in the s.c. one (21 ± 2
`µU/ml) and significantly less than in the i.p. study (117 ± 10
`µ U/ml) (P < 0.05) (Figure 1 ). By 2 h, however, the plasma
`free insulin concentration in the s.c. study (41 ± 3 µU/ml)
`had exceeded the concentration in both the i.p. (26 ± 3
`µU/ml) and i.v. studies (29 ± 2 µU/ml) (P < 0.05). This late
`elevation in plasma free insulin concentration persisted
`throughout the remainder of the study such that at 4 ½ h, the
`insulin concentration in the s.c. study (29 ± 2 µU/ml) was
`sign ificantly greater than in the i. p. (18 ± 1 µU/ml) and the
`i. v. ones ( 16 ± 2 µU/ml) (P < 0.01 ).
`Integrated insulin concentraiion (Table 1 ). Although the
`same quantity of exogenous insulin was infused into all sub(cid:173)
`jects in each study (11 U), the quantity of insulin appearing
`in the plasma significantly differed, depending on the route
`of administration. During the 4\/2-h i.v.
`infusion study,
`
`TABLE 1
`Integrated area of free insulin above basal concentration
`(µU/ml · h/100)
`
`Subject
`
`Intravenous
`
`lntraperitoneal
`
`Subcutaneous
`
`1. N.V.
`2 E.G.
`3. J.A.
`4. M.P.
`5. J.G.
`6. J.0 .
`7. 0.8.
`8. J.S
`9. L.P.
`10. 0.8.
`Mean
`Sem
`
`4,825
`12,837
`5,437
`13,975
`4,437
`8, 175
`9,262
`8,850
`11,637
`8,000
`+ 8,743
`-
`1,046
`
`2,275
`5,262
`1,040
`5,625
`3,537
`2, 11 2
`3,050
`4,862
`4,775
`3,850
`+ 3,638
`-
`480
`
`8,187
`8,300
`9,362
`9,162
`3,262
`3,525
`5,650
`8,637
`9,150
`8.600
`+ 7,383
`-
`743
`
`.-- P < 0.00 1 - - P < 0.001 --.
`
`- - - - - P > 0.05 - -- --+
`
`8743 ± 1046 µU/ml · h/100 of insulin appeared in the
`plasma. This quantity was twice that of insulin appearing in
`the plasma when delivered via the i.p. route (3636 ± 480
`µU/ml · h/100) (P < 0.001 ). When the 11 U of insulin was
`delivered via the s.c. route, 7383 ± 743 µU/ml · h/100 ap(cid:173)
`peared in the plasma, significantly greater than that during
`the i.p. route (P < 0.00 1} but not different from that quantity
`appearing during the i.v. infusion of insulin (P > 0.05).
`Plasma glucose concentration (Figure 1). Plasma glucose
`concentration rose in all studies following ingestion of the
`750-calorie meal. However, the rise was most rapid during
`the s.c. administration of insulin and least rapid during i. v.
`administration. Statistically, this difference became signifi(cid:173)
`cant at 45 min when plasma glucose concentration had
`risen by 4.1 ± 1.2 mrnol/L in the s.c. study, 1.6 ± 0. 7 mmol/L
`in the i.p. study, and declined by -0.9 ± 0.8 mmol/L in the
`i.v. study (P < 0.05 for all comparisons). This statistical dif(cid:173)
`ference remained throughout the initial 75-min observation
`period. This rate of rise corresponded inversely to the rate of
`rise in plasma free insulin concentration in the three differ(cid:173)
`ent delivery route studies (Figure 1, top).
`Plasma C-peptlde concentration. Plasma C-peptide con(cid:173)
`centration was assayed in all plasma samples. In no subject
`was C-peptide detectable in the plasma at any observation
`point during these studies.
`
`DISCUSSION
`This study compared the rapidity, duration, and magnitude
`of insulin absorption during insulin infusion via the i.v., i.p.,
`and s.c. delivery routes. Our results demonstrate that i.v. in(cid:173)
`su lin attains maximal plasma concentration by 15 min, fol(cid:173)
`lowed by the i.p. (45 min) and the s.c. routes (120 min), re(cid:173)
`spectively. With
`respect
`to
`the duration of
`insulin
`absorption, absorption from the s.c. site was most pro(cid:173)
`longed in contrast to the very rapid decline in p lasma insu(cid:173)
`lin concentration following termination of i.v. insulin deliv(cid:173)
`ery. The magnitude of the integrated rise in plasma insulin
`was comparable in both the i.v. and s.c. infusion routes, but
`was reduced by 50% with peritoneal delivery of insulin. In
`spite of this "Joss," i.p. delivery of insulin was as effective, if
`not more effective, than s.c. delivery in suppressing the
`meal-related hyperglycemia.
`This is the first study demonstrating that i.p. insulin deliv(cid:173)
`ery may control postprandial hyperglycemia in healthy dia(cid:173)
`betic men. lntraperitoneal insulin delivery has several char(cid:173)
`acteristics that make it an attractive alternative to i.v. and
`s.c. insulin delivery. First, the absorption of insulin is rela(cid:173)
`tively rapid such that rapid control of hyperglycemia may be
`achieved. Second, the peritoneal space is extravascular
`such that catheter thrombosis is minimized. 9 Third, the peri(cid:173)
`toneal space is a potentially large anatomic space into
`which an artificial pancreas could be implanted without dis(cid:173)
`figurement of the recipient.
`In these studies, only short-term insulin absorption kinet(cid:173)
`ics were examined, which may not necessarily reflect kinet(cid:173)
`ics of long-term delivery. Furthermore, because our protocol
`was not designed to optimize glucose control by the three
`routes of insu lin delivery, additional studies will be required
`to determine the most effective route in controlling plasma
`glucose concentration. However, since a relationship exists
`between the timing of the insulin increase in the plasma and
`the rise of plasma glucose concentration, 10 our results sug-
`
`DIABETES, VOL 28, . DECEMBER 1979
`
`1071
`
`FRESENIUS EXHIBIT 1052
`Page 3 of 4
`
`

`

`INSULIN DE LI VERY ROUTES
`
`gest that the infusion of s.c. insulin be begun before the start
`of the meal, as has been employed. 4
`5 Furthermore, since
`•
`both our resu lts and the results of others" demonstrate a
`prolonged absorption of insulin at the s.c. site following ter(cid:173)
`mination of the insul in infusion, a reduced rate of delivery
`between meals should be evaluated when using this route.
`Our results also suggest that improved glucose control
`might be achieved during the i.p. route if the insulin is ad(cid:173)
`ministered at a greater concentration for a shorter period of
`time following the initiation of the meal. Whatever the route
`of insulin delivery used by an artificial pancreas, knowledge
`of the kinetics of insulin absorption and the onset of biologic
`activity should result in improved control of meal-induced
`hyperglycemia.
`
`ACKNOWLEDGMENTS
`This investigation was supported by grants from the Clinical
`Research Center Program, ORR, NIH grant RR-00997-02,
`the Kroc Foundation, NIAMD-19998-01 , NICHD grant IP50-
`HD11327-01, and the Department of Energy. David S.
`Schade is the recipient of Research Career Development
`Award 1 KZ4AM00260-02.
`The technical assistance of Enid Pinero and Jerry Towle
`is appreciated. The mechanical and electrical engineering
`expertise of Raymond Bair, Gary Carlson, and Jerry Love in
`the design, fabrication, and testing of the ins_ulin mini infu-
`
`sion pump used in these studies is gratefully acknowl(cid:173)
`edged.
`
`REFERENCES
`' Albisser, A M .. Botz. C. K., and Leibel, 8 . S.: Blood g lucose regula(cid:173)
`tion using an open loop insulin delivery system in pancreatectomized dogs
`given glucose infusions. Diabetologia 16:129- 33, 1979.
`' lrsigler, K .. and Kritz, H.: Long-term continuous intravenous insulin
`therapy with a portable insulin dosage-regulating apparatus. Diabetes
`·
`28:196- 203. 1979.
`3 Schade. D. S., Eaton, R. P., Spencer, W., et al.: The peritoneal ab(cid:173)
`sorption of insulin in diabetic man: a potential site for a mechanical insulin
`delivery system. Metabolism 28: 195- 97. 1979.
`• Pickup, J. C., Keen, H .. Parsons, J. A, et al.: Continuous subcutane(cid:173)
`ous insulin infusion: an approach to achieving normoglycaemia. Br. Med. J.
`I 204 - 07. 1978.
`• Tambcrlane, W. V. , Sherwin, R. S., Genet, M., et al.: Reduction to nor(cid:173)
`mal of plasma glucose in juvenile d iabetes by subcutaneous administration
`of insulin with a portable infusion pump. N. Engl. J. Med. 300:573- 78. 1979.
`• Carlson, G. A., Shafer, 8. D., Urenda, R. S., et al.: A new low-power
`high-reliability infusion pump. Presented at the 7th Annual New England
`(Northeast) Bioengineering Conference, March 22-23. t 979.
`' Berger, W. J ., Jr.: Evaluation of " intracath" method of abdomi nal
`paracentesis. Am. Surg. 35:23- 26, 1969.
`• Kuzuya, H., Blix, P. M., Horwitz, D. L., et al.: Determination of free and
`total insulin and C-peptide in insulin-treated diabetics. Diabetes 26:22-29,
`1977.
`
`• Yoffey, J. M., and Courtice, F. C.: Lymphatics, Lymph and the Lym(cid:173)
`phOmyeloid Complex. New York, Academic Press, 1970, p. 295.
`•• Eaton, R. P., Spencer, W .. Schade, D. S., et al.: Diabetic glucose
`control: matching plasma insulin concentration to dietary and stress hyper(cid:173)
`glycemia. Diabetes Care I :40-44, 1978.
`" Slama, G .. Buu, K. N. P .. Tchobroutsky. G .. et al.: Plasma insulin and
`C-peptide levels during continuous subcutaneous insulin infusion. Diabetes
`Care 2:251 - 55. 1979.
`
`1072
`
`DIABETES, VOL. 28, DECEMBER 1979
`
`FRESENIUS EXHIBIT 1052
`Page 4 of 4
`
`