- Email: [email protected]
Surgical experience with implantable insulin pumps
Surgical Experience with Implantable Insulin Pumps Jon S. Thompson, MD, William C. Duckworth, MD, Christopher D. Saudek, MD, Anita Giobbie-Hurder, MS, for the Department of Veterans Affairs Implantable Insulin Pump Study Group, Omaha, Nebraska
BACKGROUND: A recent Veterans Affairs cooperative trial demonstrated that intensive insulin therapy via an implantable pump with intraperitoneal insulin delivery reduced glycemic variability and improved quality of life compared with multiple daily insulin injections. Our aim was to determine perioperative morbidity and assess long-term function of the implantable insulin pump. METHODS: Fifty-one adult patients with type 2 diabetes had infusion pumps placed over a 2-year period at seven VA Medical Centers as part of a randomized prospective study. RESULTS: All pumps were placed successfully. There were two (4%) perioperative complications. There were no wound complications. Duration of pump use ranged from 12 to 25 months (mean 20). Catheter obstruction (57%) and pump malfunction (25%) were the most common reasons for pump explantation. Catheter occlusions increased after 12 months. Catheter occlusion was treated by percutaneous rinse procedure in 75% and revisional procedures in 31% of patients. CONCLUSIONS: Implantable insulin pumps can be placed with minimal surgical morbidity. Attention to surgical detail and infusion protocol permits satisfactory long-term function. Pump/catheter complications increase with time but are usually resolvable by either operative or percutaneous manipulations. Am J Surg. 1998;176:622– 626. © 1998 by Excerpta Medica, Inc.
A
recent Department of Veterans Affairs cooperative trial demonstrated that intensive insulin therapy via an implantable pump with intraperitoneal insulin delivery had significant advantage over multiple daily insulin injections in reducing glycemic variability, clinical hypoglycemia, and weight gain and improving quality of life in patients with type 2 diabetes mellitus.1 Other clinical trials have also demonstrated that insulin
From the Omaha Veterans Affairs Medical Center, Omaha, Nebraska. This work was supported by the Cooperative Studies Program of the Department of Veterans Affairs Medical Research Service. Requests for reprints should be addressed to Jon S. Thompson, MD, University of Nebraska Medical Center, Department of Surgery, 600 S. 42nd Street, Omaha, Nebraska 68198-3280. Presented at the 50th Annual Meeting of the Southwestern Surgical Congress, San Antonio, Texas, April 19 –22, 1998.
622
© 1998 by Excerpta Medica, Inc. All rights reserved.
pumps can safely achieve good glycemic control in diabetic patients.2– 4 These devices are currently in the investigational stage in the United States and information about the safety and long-term function of the implantable pump is necessary before widespread clinical use is approved. Important surgical issues in the clinical utility of this therapy are risk of complications related to placement of the pump, long-term function, and the need for revisional procedures.5,6 The aim of this study was to determine perioperative morbidity and assess long-term function of the implanted insulin pump in the above-mentioned cooperative trial.
METHODS Fifty-one patients underwent placement of an implantable insulin pump at seven VA Medical Centers as part of a randomized clinical trial evaluating the insulin pump versus multiple dose insulin for type 2 diabetes mellitus between October 1991 and May 1993.1 Only the patients undergoing pump implantation are included in the present study. All patients were men between 40 and 69 years of age with a mean age of 56 6 8 years. Average body mass index was 30 6 5 kg/m2. Eligible patients had diabetes for 2 to 15 years documented clinically and by fasting C peptide level greater than 0.21 pmol/mL, inadequate glycemic control documented by hemoglobin A1C greater than 8% despite one or more insulin injections daily, and no history of hypoglycemia unawareness, life-threatening conditions, or medications that would interfere with the study protocol. All patients were initially admitted for diabetic education, initiation of therapy, and pump implantation. After discharge, visits were scheduled every 2 weeks for the first 2 months and then monthly until pump explantation. Minimum follow-up of 12 months was planned. Data were collected prospectively related to surgical placement of the pump, postoperative complications (wound infection, fluid collections, erosion, skin necrosis) and long-term function and need for surgical intervention (pump failure, catheter occlusion, infection). Data are presented as mean 6 SD. The Minimed implantable pump (Minimed Technologies, Inc., Sylmar, California) was employed in all patients (Figure 1). This system consists of the implantable pump, the delivery catheter, and the patient pump communicator. The pump device is an 8.1-cm diameter disk that is 1.9 cm thick and weighs 220 g when full. The pump is powered by a lithium cell battery designed to provide approximately 3 years of service depending on delivery rates. The reservoir holds up to 6,000 units of U-400 insulin. After subcutaneous placement of the pump, a 12-cm catheter with a free-floating tip is introduced into the abdominal cavity. 0002-9610/98/$19.00 PII S0002-9610(98)00273-6
EXPERIENCE WITH IMPLANTABLE INSULIN PUMPS/THOMPSON ET AL
TABLE Reasons for Pump Explanation at Termination of Study Catheter obstruction Pump malfunction Battery depletion Other Total
Figure 1. The implantable insulin pump with attached catheter.
29 13 5 4 51
exited the pocket at an extreme edge of the pocket running lateral or deep to the pump. An incision was made in the outer fascia, the muscle was spread apart, and the catheter introduced via a peritoneal incision. The catheter was secured via a flange sutured beneath the fascia. The pocket was irrigated, and meticulous hemostasis achieved. Both subcutaneous tissue and skin were closed as separate layers. Catheter position was verified by radiographs. Insulin underdelivery was investigated to determine the cause (Figure 2). Programming failure and errors were corrected. Insulin precipitation in the pump mechanism was treated by percutaneous rinsing of the pump with dilute sodium hydroxide. Catheter occlusion was treated initially by rinsing with catheter replacement reserved for intractable occlusion. Revisional procedures were performed by entering the pump pocket from the medial side to avoid injuring the catheter. Dense fibrous tissue normally encapsulated the pump. If the catheter was being displaced temporarily for replacement, an 8 French pediatric feeding tube was inserted in the catheter site to maintain the tract during the procedure.
RESULTS
Figure 2. Evaluation and management of insulin underdelivery.
The patient uses an external telemeter system to regulate a basal insulin infusion rate and preprogrammed boluses. The pump was refilled transcutaneously with a syringe every 4 to 12 weeks, depending on insulin requirements. The pump implant site and incision was determined by the patient’s body habitus, lifestyle, and previous procedures. Either a transverse or longitudinal incision was made adjacent to the planned pocket so that the incision was not directly over the pump. A subcutaneous pocket was developed for the pump and lipectomy performed, if necessary, so that the pump was within 4 cm of the skin surface. This permits proper radio wave telemetry with the programming units as well as facilitating pump refills. Pump migration, rotation, or inversion is prevented by creating a tight pocket and securing the pump to the abdominal fascia via three suture tabs in the pump. The peritoneal catheter
Pump implantation was performed with general anaesthesia in 45 (88%) patients and local anesthesia in 6 (12%) patients. Forty-one (80%) pumps were placed in the left upper quadrant, 6 (12%) in the left lower quadrant, and 4 (8%) in the right upper quadrant. Subcutaneous tissue overlying the pump ranged 0.6 to 4 cm in thickness (mean 1.8 cm) with 20 (39%) patients undergoing lipectomy to achieve the desired 4 cm maximal thickness. All pumps were placed successfully. Two (4%) patients had procedure-related complications, one with a hypotensive episode in the recovery room and the other with a difficult intubation. There were no incidences of wound infection, pump pocket infection, erosion of the pump, or skin necrosis. Transient small seromas in the pump pocket did not require treatment or interfere with use of the pumps. No catheters were dislodged. Duration of pump use ranged from 12 to 25 months with a mean of 20 months. Reasons for pump explantation at the end of study follow-up are shown in the Table. Catheter obstruction was the most common reason for terminating the study after the 12-month mandated follow-up period. Revisional procedures (pump/catheter replacement) were performed in 17 (33%) patients. The majority of these were for catheter occlusion. No catheter occlusions occurred in the first 8 months but there were 32 instances at 24 months with a marked increase after 12 months (Figure 3). There were 15 catheter replacements for occlusion in 14 patients. Rinse procedures were utilized in 38 (75%)
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
623
EXPERIENCE WITH IMPLANTABLE INSULIN PUMPS/THOMPSON ET AL
Figure 3. Cumulative incidence of catheter occlusion during 24month follow-up period.
The incidence of pump/catheter malfunction increases with time but such problems are usually resolvable by operative and nonoperative manipulations. Insulin underdelivery is the most frequent long-term problem. The majority of these episodes are correctable technical problems or related to insulin precipitation, which can be cleared by percutaneous rinse procedures. However, persistent catheter obstruction or pump failure require operation. We performed 18 revisional procedures during 100 patient years on pump therapy (one per 5.5 patient years). However, experience in other centers with longer follow-up indicates that revisional procedures are more frequently required, averaging one per 1.5 to 2 patient years.5,6 Although such procedures generally are associated with minimal morbidity, further technological advances will hopefully improve the durability of implantable pump function in the future.
ACKNOWLEDGMENT patients, with 19 (50%) patients undergoing more than one rinse. Three pumps were replaced, two for malfunction early in the study and one in a patient who developed acute appendicitis necessitating removal of the initial pump.
COMMENTS Implantable insulin pumps can be placed with minimal surgical morbidity.5,6 There were no significant surgical complications in the present study. Although the majority of pumps in this study were implanted under general anesthesia, with experience the procedure is certainly feasible under local anesthesia.5 Patients were hospitalized to initiate the study, but the operative procedure itself could be performed on an outpatient basis.6 Attention to surgical detail and infusion protocols permits satisfactory long-term function of implantable pumps. Although our study design mandated a defined period of pump use, our mean duration of pump use was approaching that of other centers where the goal was to maintain function indefinitely. We found no pump site complications. Udelsman et al5 reported a 4% incidence of pump site infections in 77 pump placements. These usually occurred several weeks after the procedure and occurred in patients who had undergone several revisional procedures. This potentially serious complication requires removal of the pump. Waxman et al6 reported a pump site infection due to extravasation in a deep, difficult to palpate port. Placement of the pump above the fascia and judicious use of subcutaneous lipectomy should minimize this problem. Waxman et al6 also had difficulty with pump migration and skin irritation necessitating pump replacement in 1 patient. Udelsman et al5 reported one instance of intestinal obstruction. Peritonitis, a common problem with peritoneal dialysis catheters, has not been reported with intraperitoneal insulin catheters.5–7 We did remove one pump in a patient who developed peritonitis related to appendicitis.
DISCUSSION Romano Delcore, MD (Kansas City, Kansas): The authors surgically implanted insulin infusion pumps in 51 patients with type 2 diabetes as part of a prospective multicenter cooperative VA study over a 2-year period. 624
The following persons participated in this study: Cochairmen: William C. Duckworth, MD, and Christopher D. Saudek, MD. Offices of the co-chairmen: Timothy D. Hagerty, MA, and Kimberly E. Loman, RN, CDE. Principal investigators: Bruce P. Hamilton, MD, Thomas Donner, MD, M. Sue Kirkman, MD, James W. Anderson, MD, Robert J. Anderson, MD, David E. Kelley, MD, Franklin J. Zieve, MD, PhD, Robert A. Adler, MD, Robert R. Henry, MD, and Steven V. Edelman, MD. Surgeons: Lois A. Killewich, MD, PhD, John A. Steers, MD, John P. Grant, MD, Richard W. Schwartz, MD, Jon S. Thompson, MD, Michel Makaroun, MD, Hunter H. McGuire, Jr., MD, Bruce E. Stabile, MD.
REFERENCES 1. Saudek CD, Duckworth WC, Giobbie-Hurder A, et al. Implantable insulin pump vs multiple-dose insulin for non-insulin dependent diabetes mellitus: a randomized clinical trial. JAMA. 1996; 276:1322–1327. 2. Saudek CD, Selam JL, Pitt HA, et al. A preliminary trial of the programmable implantable medication system for insulin delivery. NEJM. 1989;321:574 –579. 3. Point Study Group. One year trial of a remote-controlled implantable insulin infusion system in type I diabetic patients. Lancet. 1988;21:866 – 869. 4. Nathan DM, Dunn FL, Bruch J, et al. Postprandial insulin profiles with implantable pump therapy may explain decreased frequently of severe hypoglycemia compared with intensive subcutaneous regimens in insulin dependent diabetes mellitus patients. Am J Med. 1996;100:412– 417. 5. Udelsmen R, Chen H, Loman K, et al. Implanted programmable insulin pumps: one hundred fifty-three patient years of surgical experience. Surgery. 1997;122:1005–1011. 6. Waxman K, Turner D, Nguyen ST, et al. Implantable programmable insulin pumps for the treatment of diabetes. Arch Surg. 1992;127:1032–1037. 7. Sanderson MC, Swartzendruber DJ, Fenoglio ME, et al. Surgical complications of continuous ambulatory peritoneal dialysis. Am J Surg. 1990;160:561–566.
The purpose of their report today was to determine (1) the perioperative morbidity associated with the procedure and (2) to assess the long-term function of the implantable pump and catheter. Their surgical results are excellent, with not a single failed attempt at implantation and an
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
EXPERIENCE WITH IMPLANTABLE INSULIN PUMPS/THOMPSON ET AL
overall morbidity of only 4%. The authors have thus demonstrated that this procedure can be easily and safely performed. As to the second part of their objectives, that is, to determine the long-term function of the implantable pump and catheter, their results are less encouraging. In their 51 patients, catheter occlusion occurred in 32 instances by the end of the second year. Catheter rinsing was required in 38 instances. Surgical replacement of the catheter was required in 15 cases, and surgical replacement of the pump was required in 3 cases. In all, at least one revisional surgical procedure was required for every 5 patient years. However, the authors acknowledge in their manuscript that even this is an overly optimistic result, since reports from other centers, with longer follow-up, indicate that revisional procedures are required even more frequently, requiring at least one revisional procedure for every 1 and a half to 2 patient years. Despite these shortcomings, this method of insulin delivery appears to offer a safe and effective alternative to conventional subcutaneous insulin therapy. I have the following questions for the authors. First, is this form of insulin delivery more effective in providing tight blood glucose control? If so, is this tighter control good enough to lead us to expect a significant reduction in the incidence of diabetic retinopathy, neuropathy, nephropathy, as well as cardiac and vascular complications? Second, do you have any data on what the costs associated with this procedure and technology will be, especially when you consider the frequent need for revisional procedures? In one of your slides, it seemed to me that there were other previous abdominal incisions. I wonder if there are contraindications similar to peritoneal dialysis catheter placement, where low abdominal and pelvic previous surgery present a problem? And finally, this looks like a huge inflatable pump. What is the level of comfort for the patients with this 8-cm, half-pound pump in their abdominal wall? Carey P. Page, MD (San Antonio, Texas): I’m pretty impressed that over this many centers a group could put any kind of device in 51 diabetic patients without any complications. I think that’s really a superior accomplishment. When you took these things out, was there global protest among the patients? In other words, did they believe that this device represented a significant improvement in their life? How did it compare to them having to give themselves shots or wearing a belt-driven pump with a subcutaneous needle that they move? Victor J. Zannis, MD (Phoenix, Arizona): I’d like to know how many of these have been implanted across the country or world? And secondly, I’d like to know where Dr. Thompson thinks this device is going to be 10 years from now? F. Charles Brunicardi, MD (Houston, Texas): There are 14 million diabetics worldwide, and the annual cost of the disease is billions. It’s interesting to note that surgeons have led to some of the greatest advances in the treatment for diabetes. A pancreatectomy in a dog by Von Meeting 100 years ago led to the concept that the pancreas controlled glucose regulation. Dr. Banting, an orthopedic sur-
geon, discovered insulin in 1921. Drs Kelly and Lillehie performed the first pancreas transplant in 1966. Drs Ricordi and Sharp, both surgeons, did the first successful human islet transplants in the 1980s. The essential problem with pancreas and islet transplants is that they require human organ donation. Since there are only 7,000 donors per year in this country, even if these technologies were perfected, only a fraction of the millions of diabetics could receive treatment. Based upon your experience, do you feel that the insulin pump therapy has the potential to be the best treatment for the millions of patients that require insulin? Kelly L. Banks, MD (Bozeman, Montana): If there are no further questions, I have just a couple. I implant pacemakers, so have some experience with that, but, of course, none with insulin pumps. With pacemakers, we have some idea by phone interrogation when the pacemaker life is going to come up and when to replace the battery. How do you know this with an insulin pump? And also, is there a need to add insulin to the pump intermittently?
CLOSING Jon S. Thompson, MD: Dr. Delcore was on target with many of the questions. Perhaps we’re a little optimistic in terms of how we look at the complication rates here. Clearly, the overriding issue is whether this tight glucose control will lead to a lower incidence of complications, particularly with this therapy. That is going to be the focus, hopefully, of a future study by this study group to follow up a group of patients out for a longer term and to address that issue. As you know, that takes a long time to determine if you actually are changing the incidence of complications. I think this clearly would need to be demonstrated. Why the effect of this therapy should be better could rest on several factors. Obviously, we’re giving insulin continuously. We’re able to adjust boluses physiologically throughout the day. Also, we’re delivering it into the peritoneal cavity, which will lead to any benefits in terms of the portal uptake of the insulin and how this would modulate glucose metabolism. So there are a lot of reasons to think that this might actually be a better way to deliver it. Certainly a lot of the other pump-type therapies have been delivered systemically, and that may lead to different results. You always get concerned when you’re going to put a peritoneal catheter in patients who have had previous operations. That didn’t pose any difficulties in terms of the insulin delivery in these patients, but, of course, you’d have to evaluate that on a case-by-case basis in these patients. This is a large pump. I’m sure as the technology progresses, the pump will become smaller. The reservoir capacity is not that large in these pumps, but that’s one of the issues that will drive the size. You need to place it high in the abdomen generally to try to get it away from the belt line. Clearly, most of these patients are aware of the pump, although in some of them who are more obese, it’s hard to tell that the pump’s there. Again, we’re looking at a more elderly male population here, and this may be different in the more generalized population. I think most of the patients like the pump. The quality of life questionnaire would suggest that it did enhance their quality of life. Dr. Page asked whether the patients wanted
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
625
EXPERIENCE WITH IMPLANTABLE INSULIN PUMPS/THOMPSON ET AL
to continue. I think a lot of the patients who had the pumps taken out would have liked to continue on the studies. Others didn’t seem to mind that much going back to what they were doing. This is not new therapy. People have been looking at these investigational pumps for many years, initially given systemically with different types of pumps. This was approved for clinical use in Europe, so there is a fair experience with hundreds of patients there that would indicate that this does work well. Dr. Brunicardi asked about the future of this therapy. I suspect in the long run that perhaps pancreas transplant may be the way to go in more of these patients. Certainly as this technology advances, particularly if we can get down to a smaller pump, and can develop a pump that would sense the glucose levels and make automatic adjustments, this would be an appealing approach. Dr. Banks, we are able to interrogate the pump and see
626
what the delivery pattern has been. That’s another unit that the investigators had available and should be able to detect the pump life is well. Depending on how much insulin is required in these patients, the pump only has to be filled every 1 to 3 months. Compared with other devices that we put in, it’s not being accessed very frequently, and this is probably responsible for some of the low infection rates. A question was raised about the incidence of peritonitis. Other than the patient who had appendicitis, we didn’t see that problem, but certainly it’s been reported with peritoneal dialysis catheters with some regularity. I think the difference here is that the volume being infused is very small, and it’s not being directly infused from the outside. The pump’s being loaded sterilely and then it’s very slowly discharged over a period of months. I suspect that has something to do with the low rate of peritonitis that was observed in this study.
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
BACKGROUND: A recent Veterans Affairs cooperative trial demonstrated that intensive insulin therapy via an implantable pump with intraperitoneal insulin delivery reduced glycemic variability and improved quality of life compared with multiple daily insulin injections. Our aim was to determine perioperative morbidity and assess long-term function of the implantable insulin pump. METHODS: Fifty-one adult patients with type 2 diabetes had infusion pumps placed over a 2-year period at seven VA Medical Centers as part of a randomized prospective study. RESULTS: All pumps were placed successfully. There were two (4%) perioperative complications. There were no wound complications. Duration of pump use ranged from 12 to 25 months (mean 20). Catheter obstruction (57%) and pump malfunction (25%) were the most common reasons for pump explantation. Catheter occlusions increased after 12 months. Catheter occlusion was treated by percutaneous rinse procedure in 75% and revisional procedures in 31% of patients. CONCLUSIONS: Implantable insulin pumps can be placed with minimal surgical morbidity. Attention to surgical detail and infusion protocol permits satisfactory long-term function. Pump/catheter complications increase with time but are usually resolvable by either operative or percutaneous manipulations. Am J Surg. 1998;176:622– 626. © 1998 by Excerpta Medica, Inc.
A
recent Department of Veterans Affairs cooperative trial demonstrated that intensive insulin therapy via an implantable pump with intraperitoneal insulin delivery had significant advantage over multiple daily insulin injections in reducing glycemic variability, clinical hypoglycemia, and weight gain and improving quality of life in patients with type 2 diabetes mellitus.1 Other clinical trials have also demonstrated that insulin
From the Omaha Veterans Affairs Medical Center, Omaha, Nebraska. This work was supported by the Cooperative Studies Program of the Department of Veterans Affairs Medical Research Service. Requests for reprints should be addressed to Jon S. Thompson, MD, University of Nebraska Medical Center, Department of Surgery, 600 S. 42nd Street, Omaha, Nebraska 68198-3280. Presented at the 50th Annual Meeting of the Southwestern Surgical Congress, San Antonio, Texas, April 19 –22, 1998.
622
© 1998 by Excerpta Medica, Inc. All rights reserved.
pumps can safely achieve good glycemic control in diabetic patients.2– 4 These devices are currently in the investigational stage in the United States and information about the safety and long-term function of the implantable pump is necessary before widespread clinical use is approved. Important surgical issues in the clinical utility of this therapy are risk of complications related to placement of the pump, long-term function, and the need for revisional procedures.5,6 The aim of this study was to determine perioperative morbidity and assess long-term function of the implanted insulin pump in the above-mentioned cooperative trial.
METHODS Fifty-one patients underwent placement of an implantable insulin pump at seven VA Medical Centers as part of a randomized clinical trial evaluating the insulin pump versus multiple dose insulin for type 2 diabetes mellitus between October 1991 and May 1993.1 Only the patients undergoing pump implantation are included in the present study. All patients were men between 40 and 69 years of age with a mean age of 56 6 8 years. Average body mass index was 30 6 5 kg/m2. Eligible patients had diabetes for 2 to 15 years documented clinically and by fasting C peptide level greater than 0.21 pmol/mL, inadequate glycemic control documented by hemoglobin A1C greater than 8% despite one or more insulin injections daily, and no history of hypoglycemia unawareness, life-threatening conditions, or medications that would interfere with the study protocol. All patients were initially admitted for diabetic education, initiation of therapy, and pump implantation. After discharge, visits were scheduled every 2 weeks for the first 2 months and then monthly until pump explantation. Minimum follow-up of 12 months was planned. Data were collected prospectively related to surgical placement of the pump, postoperative complications (wound infection, fluid collections, erosion, skin necrosis) and long-term function and need for surgical intervention (pump failure, catheter occlusion, infection). Data are presented as mean 6 SD. The Minimed implantable pump (Minimed Technologies, Inc., Sylmar, California) was employed in all patients (Figure 1). This system consists of the implantable pump, the delivery catheter, and the patient pump communicator. The pump device is an 8.1-cm diameter disk that is 1.9 cm thick and weighs 220 g when full. The pump is powered by a lithium cell battery designed to provide approximately 3 years of service depending on delivery rates. The reservoir holds up to 6,000 units of U-400 insulin. After subcutaneous placement of the pump, a 12-cm catheter with a free-floating tip is introduced into the abdominal cavity. 0002-9610/98/$19.00 PII S0002-9610(98)00273-6
EXPERIENCE WITH IMPLANTABLE INSULIN PUMPS/THOMPSON ET AL
TABLE Reasons for Pump Explanation at Termination of Study Catheter obstruction Pump malfunction Battery depletion Other Total
Figure 1. The implantable insulin pump with attached catheter.
29 13 5 4 51
exited the pocket at an extreme edge of the pocket running lateral or deep to the pump. An incision was made in the outer fascia, the muscle was spread apart, and the catheter introduced via a peritoneal incision. The catheter was secured via a flange sutured beneath the fascia. The pocket was irrigated, and meticulous hemostasis achieved. Both subcutaneous tissue and skin were closed as separate layers. Catheter position was verified by radiographs. Insulin underdelivery was investigated to determine the cause (Figure 2). Programming failure and errors were corrected. Insulin precipitation in the pump mechanism was treated by percutaneous rinsing of the pump with dilute sodium hydroxide. Catheter occlusion was treated initially by rinsing with catheter replacement reserved for intractable occlusion. Revisional procedures were performed by entering the pump pocket from the medial side to avoid injuring the catheter. Dense fibrous tissue normally encapsulated the pump. If the catheter was being displaced temporarily for replacement, an 8 French pediatric feeding tube was inserted in the catheter site to maintain the tract during the procedure.
RESULTS
Figure 2. Evaluation and management of insulin underdelivery.
The patient uses an external telemeter system to regulate a basal insulin infusion rate and preprogrammed boluses. The pump was refilled transcutaneously with a syringe every 4 to 12 weeks, depending on insulin requirements. The pump implant site and incision was determined by the patient’s body habitus, lifestyle, and previous procedures. Either a transverse or longitudinal incision was made adjacent to the planned pocket so that the incision was not directly over the pump. A subcutaneous pocket was developed for the pump and lipectomy performed, if necessary, so that the pump was within 4 cm of the skin surface. This permits proper radio wave telemetry with the programming units as well as facilitating pump refills. Pump migration, rotation, or inversion is prevented by creating a tight pocket and securing the pump to the abdominal fascia via three suture tabs in the pump. The peritoneal catheter
Pump implantation was performed with general anaesthesia in 45 (88%) patients and local anesthesia in 6 (12%) patients. Forty-one (80%) pumps were placed in the left upper quadrant, 6 (12%) in the left lower quadrant, and 4 (8%) in the right upper quadrant. Subcutaneous tissue overlying the pump ranged 0.6 to 4 cm in thickness (mean 1.8 cm) with 20 (39%) patients undergoing lipectomy to achieve the desired 4 cm maximal thickness. All pumps were placed successfully. Two (4%) patients had procedure-related complications, one with a hypotensive episode in the recovery room and the other with a difficult intubation. There were no incidences of wound infection, pump pocket infection, erosion of the pump, or skin necrosis. Transient small seromas in the pump pocket did not require treatment or interfere with use of the pumps. No catheters were dislodged. Duration of pump use ranged from 12 to 25 months with a mean of 20 months. Reasons for pump explantation at the end of study follow-up are shown in the Table. Catheter obstruction was the most common reason for terminating the study after the 12-month mandated follow-up period. Revisional procedures (pump/catheter replacement) were performed in 17 (33%) patients. The majority of these were for catheter occlusion. No catheter occlusions occurred in the first 8 months but there were 32 instances at 24 months with a marked increase after 12 months (Figure 3). There were 15 catheter replacements for occlusion in 14 patients. Rinse procedures were utilized in 38 (75%)
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
623
EXPERIENCE WITH IMPLANTABLE INSULIN PUMPS/THOMPSON ET AL
Figure 3. Cumulative incidence of catheter occlusion during 24month follow-up period.
The incidence of pump/catheter malfunction increases with time but such problems are usually resolvable by operative and nonoperative manipulations. Insulin underdelivery is the most frequent long-term problem. The majority of these episodes are correctable technical problems or related to insulin precipitation, which can be cleared by percutaneous rinse procedures. However, persistent catheter obstruction or pump failure require operation. We performed 18 revisional procedures during 100 patient years on pump therapy (one per 5.5 patient years). However, experience in other centers with longer follow-up indicates that revisional procedures are more frequently required, averaging one per 1.5 to 2 patient years.5,6 Although such procedures generally are associated with minimal morbidity, further technological advances will hopefully improve the durability of implantable pump function in the future.
ACKNOWLEDGMENT patients, with 19 (50%) patients undergoing more than one rinse. Three pumps were replaced, two for malfunction early in the study and one in a patient who developed acute appendicitis necessitating removal of the initial pump.
COMMENTS Implantable insulin pumps can be placed with minimal surgical morbidity.5,6 There were no significant surgical complications in the present study. Although the majority of pumps in this study were implanted under general anesthesia, with experience the procedure is certainly feasible under local anesthesia.5 Patients were hospitalized to initiate the study, but the operative procedure itself could be performed on an outpatient basis.6 Attention to surgical detail and infusion protocols permits satisfactory long-term function of implantable pumps. Although our study design mandated a defined period of pump use, our mean duration of pump use was approaching that of other centers where the goal was to maintain function indefinitely. We found no pump site complications. Udelsman et al5 reported a 4% incidence of pump site infections in 77 pump placements. These usually occurred several weeks after the procedure and occurred in patients who had undergone several revisional procedures. This potentially serious complication requires removal of the pump. Waxman et al6 reported a pump site infection due to extravasation in a deep, difficult to palpate port. Placement of the pump above the fascia and judicious use of subcutaneous lipectomy should minimize this problem. Waxman et al6 also had difficulty with pump migration and skin irritation necessitating pump replacement in 1 patient. Udelsman et al5 reported one instance of intestinal obstruction. Peritonitis, a common problem with peritoneal dialysis catheters, has not been reported with intraperitoneal insulin catheters.5–7 We did remove one pump in a patient who developed peritonitis related to appendicitis.
DISCUSSION Romano Delcore, MD (Kansas City, Kansas): The authors surgically implanted insulin infusion pumps in 51 patients with type 2 diabetes as part of a prospective multicenter cooperative VA study over a 2-year period. 624
The following persons participated in this study: Cochairmen: William C. Duckworth, MD, and Christopher D. Saudek, MD. Offices of the co-chairmen: Timothy D. Hagerty, MA, and Kimberly E. Loman, RN, CDE. Principal investigators: Bruce P. Hamilton, MD, Thomas Donner, MD, M. Sue Kirkman, MD, James W. Anderson, MD, Robert J. Anderson, MD, David E. Kelley, MD, Franklin J. Zieve, MD, PhD, Robert A. Adler, MD, Robert R. Henry, MD, and Steven V. Edelman, MD. Surgeons: Lois A. Killewich, MD, PhD, John A. Steers, MD, John P. Grant, MD, Richard W. Schwartz, MD, Jon S. Thompson, MD, Michel Makaroun, MD, Hunter H. McGuire, Jr., MD, Bruce E. Stabile, MD.
REFERENCES 1. Saudek CD, Duckworth WC, Giobbie-Hurder A, et al. Implantable insulin pump vs multiple-dose insulin for non-insulin dependent diabetes mellitus: a randomized clinical trial. JAMA. 1996; 276:1322–1327. 2. Saudek CD, Selam JL, Pitt HA, et al. A preliminary trial of the programmable implantable medication system for insulin delivery. NEJM. 1989;321:574 –579. 3. Point Study Group. One year trial of a remote-controlled implantable insulin infusion system in type I diabetic patients. Lancet. 1988;21:866 – 869. 4. Nathan DM, Dunn FL, Bruch J, et al. Postprandial insulin profiles with implantable pump therapy may explain decreased frequently of severe hypoglycemia compared with intensive subcutaneous regimens in insulin dependent diabetes mellitus patients. Am J Med. 1996;100:412– 417. 5. Udelsmen R, Chen H, Loman K, et al. Implanted programmable insulin pumps: one hundred fifty-three patient years of surgical experience. Surgery. 1997;122:1005–1011. 6. Waxman K, Turner D, Nguyen ST, et al. Implantable programmable insulin pumps for the treatment of diabetes. Arch Surg. 1992;127:1032–1037. 7. Sanderson MC, Swartzendruber DJ, Fenoglio ME, et al. Surgical complications of continuous ambulatory peritoneal dialysis. Am J Surg. 1990;160:561–566.
The purpose of their report today was to determine (1) the perioperative morbidity associated with the procedure and (2) to assess the long-term function of the implantable pump and catheter. Their surgical results are excellent, with not a single failed attempt at implantation and an
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
EXPERIENCE WITH IMPLANTABLE INSULIN PUMPS/THOMPSON ET AL
overall morbidity of only 4%. The authors have thus demonstrated that this procedure can be easily and safely performed. As to the second part of their objectives, that is, to determine the long-term function of the implantable pump and catheter, their results are less encouraging. In their 51 patients, catheter occlusion occurred in 32 instances by the end of the second year. Catheter rinsing was required in 38 instances. Surgical replacement of the catheter was required in 15 cases, and surgical replacement of the pump was required in 3 cases. In all, at least one revisional surgical procedure was required for every 5 patient years. However, the authors acknowledge in their manuscript that even this is an overly optimistic result, since reports from other centers, with longer follow-up, indicate that revisional procedures are required even more frequently, requiring at least one revisional procedure for every 1 and a half to 2 patient years. Despite these shortcomings, this method of insulin delivery appears to offer a safe and effective alternative to conventional subcutaneous insulin therapy. I have the following questions for the authors. First, is this form of insulin delivery more effective in providing tight blood glucose control? If so, is this tighter control good enough to lead us to expect a significant reduction in the incidence of diabetic retinopathy, neuropathy, nephropathy, as well as cardiac and vascular complications? Second, do you have any data on what the costs associated with this procedure and technology will be, especially when you consider the frequent need for revisional procedures? In one of your slides, it seemed to me that there were other previous abdominal incisions. I wonder if there are contraindications similar to peritoneal dialysis catheter placement, where low abdominal and pelvic previous surgery present a problem? And finally, this looks like a huge inflatable pump. What is the level of comfort for the patients with this 8-cm, half-pound pump in their abdominal wall? Carey P. Page, MD (San Antonio, Texas): I’m pretty impressed that over this many centers a group could put any kind of device in 51 diabetic patients without any complications. I think that’s really a superior accomplishment. When you took these things out, was there global protest among the patients? In other words, did they believe that this device represented a significant improvement in their life? How did it compare to them having to give themselves shots or wearing a belt-driven pump with a subcutaneous needle that they move? Victor J. Zannis, MD (Phoenix, Arizona): I’d like to know how many of these have been implanted across the country or world? And secondly, I’d like to know where Dr. Thompson thinks this device is going to be 10 years from now? F. Charles Brunicardi, MD (Houston, Texas): There are 14 million diabetics worldwide, and the annual cost of the disease is billions. It’s interesting to note that surgeons have led to some of the greatest advances in the treatment for diabetes. A pancreatectomy in a dog by Von Meeting 100 years ago led to the concept that the pancreas controlled glucose regulation. Dr. Banting, an orthopedic sur-
geon, discovered insulin in 1921. Drs Kelly and Lillehie performed the first pancreas transplant in 1966. Drs Ricordi and Sharp, both surgeons, did the first successful human islet transplants in the 1980s. The essential problem with pancreas and islet transplants is that they require human organ donation. Since there are only 7,000 donors per year in this country, even if these technologies were perfected, only a fraction of the millions of diabetics could receive treatment. Based upon your experience, do you feel that the insulin pump therapy has the potential to be the best treatment for the millions of patients that require insulin? Kelly L. Banks, MD (Bozeman, Montana): If there are no further questions, I have just a couple. I implant pacemakers, so have some experience with that, but, of course, none with insulin pumps. With pacemakers, we have some idea by phone interrogation when the pacemaker life is going to come up and when to replace the battery. How do you know this with an insulin pump? And also, is there a need to add insulin to the pump intermittently?
CLOSING Jon S. Thompson, MD: Dr. Delcore was on target with many of the questions. Perhaps we’re a little optimistic in terms of how we look at the complication rates here. Clearly, the overriding issue is whether this tight glucose control will lead to a lower incidence of complications, particularly with this therapy. That is going to be the focus, hopefully, of a future study by this study group to follow up a group of patients out for a longer term and to address that issue. As you know, that takes a long time to determine if you actually are changing the incidence of complications. I think this clearly would need to be demonstrated. Why the effect of this therapy should be better could rest on several factors. Obviously, we’re giving insulin continuously. We’re able to adjust boluses physiologically throughout the day. Also, we’re delivering it into the peritoneal cavity, which will lead to any benefits in terms of the portal uptake of the insulin and how this would modulate glucose metabolism. So there are a lot of reasons to think that this might actually be a better way to deliver it. Certainly a lot of the other pump-type therapies have been delivered systemically, and that may lead to different results. You always get concerned when you’re going to put a peritoneal catheter in patients who have had previous operations. That didn’t pose any difficulties in terms of the insulin delivery in these patients, but, of course, you’d have to evaluate that on a case-by-case basis in these patients. This is a large pump. I’m sure as the technology progresses, the pump will become smaller. The reservoir capacity is not that large in these pumps, but that’s one of the issues that will drive the size. You need to place it high in the abdomen generally to try to get it away from the belt line. Clearly, most of these patients are aware of the pump, although in some of them who are more obese, it’s hard to tell that the pump’s there. Again, we’re looking at a more elderly male population here, and this may be different in the more generalized population. I think most of the patients like the pump. The quality of life questionnaire would suggest that it did enhance their quality of life. Dr. Page asked whether the patients wanted
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998
625
EXPERIENCE WITH IMPLANTABLE INSULIN PUMPS/THOMPSON ET AL
to continue. I think a lot of the patients who had the pumps taken out would have liked to continue on the studies. Others didn’t seem to mind that much going back to what they were doing. This is not new therapy. People have been looking at these investigational pumps for many years, initially given systemically with different types of pumps. This was approved for clinical use in Europe, so there is a fair experience with hundreds of patients there that would indicate that this does work well. Dr. Brunicardi asked about the future of this therapy. I suspect in the long run that perhaps pancreas transplant may be the way to go in more of these patients. Certainly as this technology advances, particularly if we can get down to a smaller pump, and can develop a pump that would sense the glucose levels and make automatic adjustments, this would be an appealing approach. Dr. Banks, we are able to interrogate the pump and see
626
what the delivery pattern has been. That’s another unit that the investigators had available and should be able to detect the pump life is well. Depending on how much insulin is required in these patients, the pump only has to be filled every 1 to 3 months. Compared with other devices that we put in, it’s not being accessed very frequently, and this is probably responsible for some of the low infection rates. A question was raised about the incidence of peritonitis. Other than the patient who had appendicitis, we didn’t see that problem, but certainly it’s been reported with peritoneal dialysis catheters with some regularity. I think the difference here is that the volume being infused is very small, and it’s not being directly infused from the outside. The pump’s being loaded sterilely and then it’s very slowly discharged over a period of months. I suspect that has something to do with the low rate of peritonitis that was observed in this study.
THE AMERICAN JOURNAL OF SURGERY® VOLUME 176 DECEMBER 1998