Diabetic remission in a cat treated with an implantable pump to deliver insulin

Abstract

A diabetic cat was referred because of poor metabolic control and difficulties the owner experienced injecting insulin. A pump, telemetrically controlled with a smartphone, was implanted subcutaneously to deliver insulin. Before implantation, the pump reservoir was filled with a rapid-acting human recombinant insulin. The insulin was administered through continuous infusion or periodic boluses over 2 weeks while the cat was hospitalized and over another 2 weeks after discharge from the hospital. Adjustments of insulin dosage were performed based on blood glucose concentrations measured with a continuous blood monitoring system (CGMS). The cat achieved diabetic remission that is still lasting after 1 year. The treatment protocol adopted in this cat contributed to achieving remission. The owner’s unwillingness to inject insulin into an uncooperative cat was circumvented with the implantable pump.

Key clinical message:

The implantable subcutaneous pump, telemetrically controlled by a smartphone, easily allowed the clinician to modify the type of administration and the amount of insulin delivered; the concurrent use of a CGMS allowed detection of sudden changes in blood glucose while limiting stress to the cat.

The mainstay of treatment in diabetic cats is subcutaneous injection of insulin and feeding a low-carbohydrate diet. With adequate therapy, approximately half of the cats with diabetes mellitus (DM) achieve remission and, therefore, do not need insulin injections to maintain normoglycemia. When remission occurs, it is within 6 mo of diagnosis in more than 90% of the cases (1). This favorable outcome is more likely if hyperglycemia, causing β-cell dysfunction and loss in cats (2), is promptly and strictly controlled allowing reversal of glucose toxicity (1,2). However, if owners are unwilling to inject insulin or cats are not amenable to injections, remission is unlikely and poorly controlled DM ultimately leads to early death.

Insulin pumps, external or implantable, have been developed for diabetes treatment in humans. External pumps have a display that allows the user to enter dosage information, and usually deliver insulin through a cannula inserted into the subcutaneous tissues through a hand-held controller to adjust rates (3,4). Implantable pumps have been used in humans with type 1 DM in which an external pump failed to achieve acceptable glycemic control due to an erratic or limited absorption of insulin from the subcutaneous tissue. These pumps are surgically implanted into the subcutaneous tissue and insulin is delivered into the peritoneal cavity via a catheter. Studies have shown that the use of external or implantable insulin pumps is superior to multiple daily subcutaneous insulin injections for glycemic control in humans with type 1 or type 2 DM (3,4).

To date, insulin pumps have not been used in diabetic cats. Implantable pumps may be more practical and could provide owners with a method that eliminates restraint of cats and injection of insulin.

The present study describes the use of an implantable pump telemetrically controlled through a smartphone, to deliver insulin in a diabetic cat that had become uncompliant to subcutaneous injections.

Case description

A 10-year-old, spayed female, domestic shorthair cat was referred because of poorly controlled DM. Persistent polyuria and polydipsia were reported by the owner, and large swings of blood glucose concentrations [i.e., between 2.5 and 30 mmol/L; reference range (RR): 3.5 to 7.8 mmol/L] and glycosuria were documented by the referring veterinarian. The cat had been diagnosed with DM 4 mo earlier and, at the time of presentation, was receiving subcutaneous injections of porcine insulin zinc suspension (Caninsulin: MSD Animal Health, Segrate, Michigan, USA), at the dosage of 4 IU, twice daily. In the last month the owner was unable to inject insulin regularly because the cat had become uncompliant.

On admission, the cat was overweight, with a body weight of 7 kg. The biochemical profile and urinalysis showed hyperglycemia (blood glucose concentration: 27.4 mmol/L), increased serum fructosamine concentration (520 μmol/L; RR: 190 to 365 μmol/L), and glycosuria. Blood tests and urinalysis were otherwise unremarkable. Abdominal ultrasonography and radiographs of the thorax were normal.

Due to the frantic behavior of the cat and the owner’s difficulty in injecting insulin, the subcutaneous implantation of a pump for telemetrically controlled insulin release was proposed and written informed consent was obtained. Furthermore, because it has been shown that strict glycemic control can decrease the subsequent dosage of insulin and chances of remission are more likely during the first 6 mo from diagnosis of DM (1,5), the aim of using the pump was to perform an intensive insulin treatment protocol in order to achieve this favorable outcome.

We used an implantable pump (Ithetis; Antlia SA, Lausanne, Switzerland) that had been developed to deliver drugs in laboratory animals and was smaller than the pump originally employed in healthy cats (6). The new pump has a diameter of 2.5 cm (the original pump had major axis of 4.2 cm), a reservoir of 1 mL and a battery lasting for up to 2 mo (Figure 1). For implantation of the pump, a 3 × 5 cm area of skin in the dorsal neck was clipped and surgically prepared. A small incision of 1.5 cm was made under short general anesthesia, and a 3-cm subcutaneous pouch was created to accommodate the pump, with the reservoir placed ventrally; the pump was not secured to the subcutaneous tissue through sutures. The skin was closed with 4 absorbable sutures. The whole procedure lasted 20 min. Before implantation, the pump reservoir was filled with 1 mL of a human rapid-acting recombinant insulin (Insuman Implantable 400 IU/mL; Sanofi-Aventis, Frankfurt am Main, Germany), through a 1-mL syringe connected to a 28-G needle, under aseptic conditions. The above insulin is produced via recombinant DNA technology in Escherichia coli and was developed for continuous intra-peritoneal insulin infusion in humans with type 1 DM through an implantable pump (7). Insuman Implantable is a sterile, neutral (pH 7.5) solution supplied in vials of 10 mL with 400 IU of insulin per mL. The solution contains phenol as an antimicrobial preservative, trometamol as a buffering agent to stabilize the pH, and poloxamer 171, which is a surfactant that prevents aggregation of insulin and increases its stability. The product has a shelf life of 24 mo when stored at 2°C to 8°C, protected from light, and it is stable for 8 wk at 37 ± 2°C (7). The vial was taken from the refrigerator 6 h before use, and immediately before being injected in the pump, it was gently shaken 3 times for 30 s each time, as advised by the manufacturer.

Figure 1.

Implantable pump to deliver insulin (top left). The blue area on the lower part of the pump is the diaphragm for filling the 1-mL reservoir; a 1-mL syringe was used for this purpose. The tube on the right side of the pump is the catheter that delivers insulin; the catheter was cut before implantation to avoid kinking and obstruction. The diameter of the pump is 2.5 cm, similar to a 1-euro coin (bottom right).

Following implantation of the pump, a continuous glucose monitoring system (CGMS) (G4 Dexcom: Medtronic, Minneapolis, Minnesota, USA), that had been validated at the authors’ institution, was used. The sensor was injected in the subcutaneous tissue of the thorax, at the 6th intercostal space, midway between the sternum and the spine, by means of a 22-G needle (8). The transmitter was connected to the sensor and fixed to the skin with tape. For calibration, blood glucose concentration was measured from the inner pinna with a portable blood glucose meter (AlphaTRAK2: Abbott Animal Health, Abbott Park, Illinois, USA). The sensor and the transmitter of the CGMS as well as the surgical suture were covered with a soft bandage; the monitor of the CGMS was fixed to the cage door (the sensor and the monitor have to be within a distance of 6 m). The CGMS displays blood glucose measurements every 5 min. Notably, the monitor shows glucose concentrations between 2.2 and 22.2 mmol/L; concentrations beyond this range are correctly recorded by the CGMS but need to be downloaded to be checked. The CGMS is calibrated every 12 h.

After anesthesia, the cat was hospitalized in the intensive care unit. Insulin infusion was started approximately 2 h after implantation of the pump, as soon as the cat was completely awake and blood glucose with the CGMS had reached a concentration of 22 mmol/L. The insulin pump is telemetrically controlled by an application installed on a smartphone. The device allows the setting up of 3 different administration protocols, including continuous or periodical rate infusion, or single boluses. The maximum flow rate is 100 μL/h and the minimum is 0.01 μL/bolus. Drug delivery from the pump can be modified whenever it is necessary.

The insulin administration was started with a continuous rate infusion of 1.25 μL/h (0.5 IU/h). However, after 5 h from the beginning of the infusion, blood glucose had fallen to 3 mmol/L; insulin infusion was stopped, and food was provided to the cat. The continuous rate infusion of insulin was restarted after 3 h, and gradually reduced to 0.62 μL/h (0.25 IU/h) for 16 h and then to 0.4 μL/h (0.16 IU/h) for 6 h. Thereafter, boluses of 0.15 to 0.35 μL (0.06 to 0.14 IU) were provided during the next 11 d using the following protocol: i) if blood glucose concentration was between 5 and 9 mmol/L, boluses of insulin with equal dosage were provided every hour; ii) if blood glucose concentration was < 5 mmol/L or > 5 mmol/L but dropping rapidly and expected to decrease to < 5 mmol/L within 2 to 3 h, boluses of insulin were stopped; iii) if blood glucose concentration increased to > 9 mmol/L following a previous interruption, boluses of insulin with the last dosage were restarted; and iv) if blood glucose concentration was > 9 mmol/L and continued to increase, the dosage of each bolus was increased by 0.05 μL (0.02 IU) every 2 to 3 h (Figure 2). During the last 2 d, boluses of 0.20 μL (0.08 IU) were provided every 4 h. Blood glucose concentrations were assessed every 15 to 45 min on the monitor of the CGMS. The CGMS was calibrated more frequently than every 12 h if the glucose curve showed rapid or unexpected swings. During the whole period of hospitalization, the cat underwent only 1 episode of moderate hypoglycemia (blood glucose of 2.5 mmol/L) requiring infusion of a glucose bolus (dose: 250 mg). The episode was not associated with clinical signs.

Figure 2.

Blood glucose curve achieved with the continuous glucose monitoring system during day 3. The 2 red arrows show when the insulin boluses were stopped, the blue arrow shows when the insulin boluses were restarted (at 0.08 IU) and the 2 green arrows show when there were stepwise increases of insulin dosage (at 0.10 IU and 0.12 IU). Food was administered at 8 am and 8 pm.

Following implantation of the pump, the cat was monitored twice daily for discomfort (pruritus, vocalization, pain during neck palpation, increased respiratory rate), local and systemic signs of inflammation (erythema, erosions, bleeding, purulent discharge, increased rectal temperature, lethargy, anorexia) possibly associated with the device. After 2 wk of hospitalization, the cat was discharged with blood glucose concentrations between 6 and 8 mmol/L. At the time of discharge the CGMS was removed but the pump was left in place. The owner’s daughter was taught how to measure blood glucose from the inner pinna with a portable glucose meter, at least twice daily. The owner was contacted every day in order to receive all glucose measurements and information about the cat. During the following 2 wk the cat’s capillary blood glucose concentrations varied between 3 and 7 mmol/L; therefore, the pump was not programmed to administer boluses of insulin.

One month after implantation of the pump, the cat did not show any detectable signs of discomfort; well-being, thirst, appetite, and physical examination, including skin and subcutaneous tissue at the surgical site, were normal. A complete blood cell count and biochemical profile yielded normal blood glucose (6 mmol/L) and fructosamine (265 μmol/L) concentrations. The insulin pump was removed under short general anesthesia and the surgical wound was closed with non-absorbable sutures that were removed after 10 d. During the procedure, a biopsy of the subcutaneous pouch surrounding the pump was collected. The histopathological examination showed mild fibrosis with only mild signs of inflammation (Figure 3).

Figure 3.

Histology of the subcutaneous surgical pouch (hematoxylin and eosin, magnification ×100). The luminal side of the cystic space is delineated by mesenchymal fusiform cells without signs of atypia and moderate amounts of fibrous tissue (pseudocapsule). Mild multifocal perivascular lymphocytic infiltrates are present in the fibrous tissue. Normal adipose tissue surrounds the pouch (top left and bottom right).

During removal of the pump, the reservoir was inadvertently damaged by the surgeon. An unknown amount of insulin leaked in the subcutaneous tissue causing severe hypoglycemia (2 mmol/L) and weakness. Glucose infusion was provided for 10 h in order to maintain euglycemia (total dose: 1.25 g). Immediately after removal of the pump, the functionality of the device was confirmed by programming a bolus that yielded a visible drop of insulin solution. After 2 days of hospitalization, blood glucose values were normal, and the cat was discharged without insulin treatment. Following implantation of the pump and its removal antibiotic prophylaxis and anti-inflammatory treatment were not required.

After 1 y of follow-up the cat is still in diabetic remission.

Discussion

Current cornerstone of DM management in cats consists of subcutaneous insulin injections. However, injecting insulin in cats may be challenging for some owners due to the difficulty in restraining the cat, lack of time, or aversion to injections (6). In the present case, the owner had become reluctant to inject and the cat was uncooperative. As less than 6 mo had passed from diagnosis of DM, the cat was a suitable candidate for an implantable pump to deliver insulin. This would allow the cat to receive an intensive insulin protocol aimed at optimally controlling blood glucose concentrations and, possibly, achieving remission.

Previous studies in humans with type 1 and type 2 DM have shown that insulin pumps provide better glycemic control compared to multiple subcutaneous injections (3,4,9). In cats, only 1 study has been carried out on the use of implantable pumps (6). In particular, the administration of insulin glargine to 10 healthy cats by an earlier generation of the present implantable pump was followed by an increase in plasma insulin glargine levels only following the first bolus (6). Successive boluses scheduled during 2 mo from implantation of the pump could not be performed in 5 cats because the machinery of the pump stopped working. In the other 5 cases no significant change in insulin glargine concentrations was recorded and capillary blood glucose did not decrease; this failure was probably caused by the loss of stability of insulin glargine over time. It was assumed that some part of the pump had caused an inactivation of the insulin glargine molecule (6).

In the diabetic cat herein, implantation of the pump was quick and easy to perform, and the device was well-tolerated. No adverse reactions were documented locally or systemically during the hospitalization period and during the 2 wk following discharge. Moreover, no overt inflammation was detected microscopically in the subcutaneous tissues at the site of implantation. In humans, abscess formation, skin ulceration, hematoma, and pain at the implantation site have been described (10). In addition, the pump implanted did not malfunction during the treatment period and at the time of its removal the machinery was still functioning correctly. The technical improvements recently made to the pump were probably responsible for the success. Still, it is worth noting that, in contrast to the previous study, the pump was used for only 1 mo. It cannot be excluded that problems may have occurred if it were to be used for a longer period or if refilling of the reservoir had become necessary.

An interesting feature of the new pump used in this cat is the possibility of controlling insulin delivery by an application installed on a smartphone. Therefore, it was easy for the attending clinician to modify the type of administration or the amount of insulin delivered at any time. The concurrent use of a CGMS limited the stress to the cat and identified episodes of hypoglycemia and hyperglycemia.

In this report a type of human recombinant insulin never tested in cats was used. This insulin remains stable at body temperature for more than 1 mo, partly due to the surfactant in the solution. In humans with type 1 DM this type of insulin has been used for more than a decade without significant side effects (7). The most commonly reported issues related to its use along with implantable pumps are infection of the surgical pocket, skin erosion, abnormal healing of the implantation site, and formation of insulin antibodies. In the cat in this report, no side effects were documented with either the insulin or the pump. Investigations were not carried out to determine whether insulin antibodies had developed; however, based on the fact that relapse of DM had not occurred, if antibodies were present, they may not substantially reduce the action of endogenous insulin.

In conclusion, this report is the first to describe the use of an implantable pump in a diabetic cat for subcutaneous delivery of an insulin designed to remain stable at body temperature for a long period. The pump worked correctly and insulin was effective during the entire treatment period. The owner’s unwillingness to inject insulin and the uncooperative behavior of the cat were circumvented by using the implantable pump. The treatment protocol adopted in this cat contributed to achievement of remission. The results should encourage further studies on a larger number of cats in order to assess the safety and efficacy of the device and of the insulin preparation in this species. Cost may limit the use of this treatment option, as the 2-week period of hospitalization and care, and pump and insulin may approximate $3000 USD. In addition, the current pump is not suitable for long duration of treatment and it is unclear whether refilling would be possible with the pump still in place.