Retrospective evaluation of continuous rate infusion of regular insulin intravenously for the management of feline diabetic ketoacidosis

Abstract

The use and efficacy of continuous rate infusion (CRI) of regular insulin intravenously for the treatment of feline diabetic ketoacidosis was retrospectively evaluated. The study focused on the rate of glucose decline, time to resolution of inappetence, time to long-term injectable insulin, and length of hospital stay. Review of medical records from 2009 to 2011 identified 10 cases that met the inclusion criteria. Six cats were existing diabetics, 3 of whom had recent insulin changes. Five cats had concurrent diseases. The mean time to long-term injectable insulin was 55 hours. The mean length of hospitalization was 3.8 days. Five cats survived to discharge. In 5 patients, an insulin CRI permitted a short hospital stay and transition to long-term injectable insulin. Many cats with diabetic ketosis or diabetic ketoacidosis are prior diabetics with concurrent disease and/or a history of recent insulin changes.

Introduction

Ketosis and ketoacidosis are severe complications of diabetes mellitus that can occur in both newly diagnosed and poorly controlled diabetic dogs and cats. Diabetic ketoacidosis (DKA) occurs when a relative deficiency of insulin and increased glucoregulatory hormones lead to an overproduction of ketones (1). The condition is characterized by dehydration, electrolyte derangements, osmolality imbalances, hyperglycemia, glucosuria, and ketonemia/uria (2–4). Patients often have concurrent illnesses such as pancreatitis, chronic renal failure, urinary tract infection, hepatic lipidosis, and/or neoplasia (2,5). Treatment goals include rehydration, correction of electrolyte and acid/base disturbances, decreasing blood glucose, reversing ketonemia, and treatment of any concurrent illnesses.

Along with fluid therapy, insulin therapy is one of the cornerstones of DKA management. Insulin is required to prevent lipolysis, halt hepatic ketogenesis and gluconeogenesis, and promote tissue metabolism of glucose and ketones (1). Continuous low dose regular insulin infusions are the standard of care in humans with DKA (6–11). Proposed insulin treatment methods for cats include intermittent regular insulin intramuscularly (IM) or subcutaneously (SC) (2), intermittent long-acting insulin IM (2,12), and continuous regular insulin intravenous infusions (13).

Regular insulin continuous rate infusions (CRI) are becoming the standard of care in veterinary medicine for the treatment of systemically ill DKA patients. This allows a slow and steady decline of blood glucose and resolution of electrolyte imbalances. It often requires intensive patient monitoring with 24-hour patient care and frequent blood sampling. A few studies have shown this method to be safe and effective for the treatment of feline DKA patients (13,14). Historically, intermittent intramuscular or subcutaneous injections of regular insulin were utilized.

The objective of this retrospective study was to document the use and efficacy of continuous infusion of regular insulin intravenously for the treatment of feline diabetic ketoacidosis, focusing on the rate of glucose decline, time to resolution of inappetence, time to long-term injectable insulin, and length of hospital stay.

Materials and methods

The study population consisted of cats admitted to a private specialty hospital between 2009 and 2011 that were diagnosed with diabetic ketosis (DK) or DKA and treated with a CRI of regular insulin intravenously. Diagnosis of DK was based on a blood glucose > 13.88 mmol/L, glucosuria, ketonemia, and/or ketonuria. Patients with DKA had the additional abnormality of acidosis. Acidosis was defined as a blood pH less than 7.35. For data analysis, a distinction between DK and DKA was not made in this study.

Medical records were reviewed and data were gathered for patients treated with continuous regular insulin IV infusions. Demographic information obtained from the medical records included age, gender, neuter status, and breed. Physical examination findings at presentation were recorded including mentation, rectal temperature, heart rate, respiratory rate, body condition, hydration status, and any significant abnormalities such as muscle wasting and heart murmurs. Clinical signs such as inappetence, vomiting, diarrhea, lethargy, polyuria, polydipsia, weight loss, and weakness were also recorded. Prior treatments were recorded, if available. Results of complete blood cell count, serum chemistry profile, urinalysis, thoracic radiographs, and abdominal ultrasound were noted. Results of blood gas were included when available. History of diabetes and concurrent diseases were recorded. Whole blood glucose was monitored every 2 h using a cage side glucometer (AlphaTRAK; Abbott Laboratories, Berkshire, United Kingdom) in all patients. The rate of blood glucose decline, time to resolution of inappetence, time to institution of long-term insulin, and length of hospital stay were determined.

All patients were treated with intravenous crystalloids (Veterinary Plasmalyte-A; Abbott Laboratories, Chicago, Illinois, USA) or 0.9% NaCl (0.9% NaCl; Braun, Irvine, California, USA) prior to and while receiving insulin treatment. Cats were treated with an insulin CRI that consisted of 250 mL of 0.9% NaCl to which 1.1 units of regular insulin (Humulin R; Lilly, Indianapolis, Indiana, USA) per kg body weight (BW) were added (1). To saturate binding of insulin to the IV tubing, 50 mL of the insulin solution were allowed to stand in the line for 30 min and then run through the IV line (15). The initial insulin CRI rate was based on the patient’s blood glucose at the time the CRI was started (16). The rate of the insulin CRI was adjusted every 2 h based on the patient’s blood glucose. Adjustments in the insulin CRI rate and the addition of dextrose were implemented at each clinician’s discretion loosely based on previously published guidelines (16). Long-term insulin was initiated when the patient was eating and appropriately hydrated.

Statistical methods

A computer spreadsheet program (Microsoft Excel 2007; Microsoft, Redmond, Washington, USA) was utilized for data entry. The data were analyzed using IBM SPSS Statistics version 21 (IBM Corp 2012; Armonk, New York, New York, USA). Means and standard deviations are reported for normally distributed data. Medians and ranges are reported for non-normally distributed data. Post hoc pairwise comparisons with a t-test were performed to evaluate the following variables between survivors and non-survivors: pH during first 24 h, body condition score, total bilirubin, serum osmolality before treatment, blood glucose at presentation, duration of treatment until the blood glucose was < 13.88 mmol/L, albumin, alanine aminotransferase (ALT) activity, mean units/kg per day of insulin, and rate of blood glucose decline over the first 14 h of treatment. A Chi-squared analysis was performed to determine statistical significance in survival between new and previously diagnosed diabetics. A P-value of < 0.05 was deemed statistically significant.

Results

Nineteen cats with a diagnosis of DK or DKA were identified. Five cats were treated with intermittent regular insulin injections and 4 cats were euthanized shortly after diagnosis. These 9 cats were excluded from the study. The 10 remaining cats were treated with an insulin CRI and met all criteria for inclusion in the study.

Four domestic short hair cats, 1 domestic medium hair cat, 2 domestic long hair cats, 1 Manx, 1 Maine Coon, and 1 Tonkinese were identified. There were 5 males (4 castrated, 1 intact) and 5 females (all spayed). The mean age was 10 y (range: 8 to 12 y). Mean body weight was 4.78 kg (SD: 1.53 kg). Significant physical examination findings included: dehydration (7/10), dull or obtunded mentation (5/10), generalized muscle wasting (5/10), hypothermia (3/10), and tachypnea (2/10). Clinical signs included inappetence (10/10), weight loss (7/10), lethargy (7/10), vomiting (5/10), polyuria/polydipsia (4/10), weakness (4/10), and diarrhea (1/10). Prior treatment included intermittent regular insulin (n = 3), intravenous fluids (n = 4), ampicillin (n = 1), and bicarbonate (n = 1). Six cats were existing diabetics. Four cats were treated with a single insulin, which included protamine zinc insulin (PZI VET; IDEXX laboratories, Westbrook, Maine, USA), porcine insulin zinc suspension (Vetsulin; Intervet, Merck Animal Health, Summit, New Jersey, USA), glargine (Lantus; Aventis Pharmaceuticals, Bridgewater, New Jersey, USA), and neutral protamine hagedorn (Humulin N; Lilly, Indianapolis, Indiana, USA). Two cats had been treated with multiple types of insulin. One had been treated with neutral protamine hagedorn, glargine, and protamine zinc insulin; the other had been treated with glargine and porcine insulin zinc suspension. These patients had been existing diabetics for a mean of 23 mo (SD: 22.7 mo). Three of the existing diabetics had recent changes in their insulin type or dosage. Thoracic radiographs were performed in 5 cats; no abnormalities were noted. Abdominal ultrasound was performed in 9/10 cats. Ultrasound abnormalities included hyperechoic liver (n = 6), renomegaly (n = 4), thickened hypoechoic pancreas with peripancreatic hyperechoic fat and mesentery consistent with pancreatitis (n = 3), pyelectasia (n = 3), and mild mesenteric lymphadenopathy (n = 1). Concurrent disease processes were identified in 5 cats. These included pancreatitis (n = 3) and hepatic lipidosis (n = 2). Pancreatitis was diagnosed based on ultrasonographic changes. While hepatic lipidosis may have been present in more cases, it was documented via liver cytology in only 2 cases.

Initial blood work was performed in all cats. There were no statistically significant differences in labwork findings between the patients that survived and those that were euthanized or died. All 10 patients had glucosuria. Eight patients were positive for ketones on urinalysis. Two patients did not have detectable urinary ketones, but were deemed to be ketotic based on characteristic scent (17).

Mean blood glucose at time of presentation was 22.26 ± 9.56 mmol/L. The mean rates of decline of blood glucose during the initial 2-hour time block of treatment are listed in Table 1. The mean time to resolution of inappetence was 23.6 h (SD: 28.35 h). One patient, which had an esophageal feeding tube placed due to concurrent hepatic lipidosis, was not eating at the time of discharge and was not included in this data point. The mean length of time to long-term injectable insulin from the start of intravenous fluids was 55 h (SD: 30.78 h). The mean daily dose of insulin for the first 24 h of treatment was 0.91 units/kg/day (SD: 0.15 unit/kg per day). Four patients received Plasmalyte-A while hospitalized. Six patients received 0.9% NaCl. Fluid additives included potassium chloride (n = 10), potassium phosphate (n = 6), and magnesium sulfate (n = 1). Additional treatments while hospitalized included standard doses of the following: ampicillin/sulbactam (Unasyn; SAGENT Pharmaceuticals, Schaumburg, Illinois, USA) (n = 4), hetastarch (Hespan; Braun, Irvine, California, USA) (n = 4), famotidine (Famotidine; West-ward, Eatontown, New Jersey, USA) (n= 3), mirtazapine (Mirtazapine; Greenstone, Peapack, New Jersey, USA) (n = 2), maropitant (Cerenia; Pfizer, New York, New York, USA) (n = 2), packed red blood cell transfusion (n = 1), furosemide (Salix; Intervet, Millsboro, Delaware, USA) (n = 1), dopamine (Dopamine HCL; Hospira, Lake Forest, Illinois, USA) CRI (n = 1), calcium gluconate (n = 1), ranitidine (Zantac; GlaxoSmithKline, Research Triangle Park, North Carolina, USA) (n = 1), vitamin E (Vitamin E oil; Nature’s Bounty, Bohemia, New York, USA) (n = 1), L-carnitine (L-carnitine; Elk River Naturals, Tulsa, Oklahoma, USA) (n = 1), vitamin B12 (Vitamin B12; Vet One, Boise, Idaho, USA) (n = 1), lactulose (Lactulose; Hi-Tech Pharmacal, Amityville, New York, USA) (n = 1), and oral potassium gluconate (Tumil K; Virbac Animal Health, Fort Worth, Texas, USA) (n = 1). Mean length of hospitalization was 3.8 d (SD: 1.49 d).

Table 1.

Mean rate of decline of blood glucose during treatment

Time block (hours)	Mean rate ± SD of blood glucose decline in mg/dL per hour and (mmol/L per hour)
0 to 2	−0.28 ± 33.79 (0.02 ± 1.88)
2 to 4	−2.33 ± 39.30 (0.13 ± 2.18)
4 to 6	−7.22 ± 35.54 (0.40 ± 1.97)
6 to 8	6.22 ± 56.24 (0.35 ± 3.12)
8 to 10	−8.5 ± 19.78 (0.47 ± 1.10)
10 to 12	0.71 ± 32.73 (0.04 ± 1.82)
12 to 14	6.07 ± −28.82 (0.34 ± 1.60)

Complications of treatment included hypoglycemia (n = 1) and severe anemia (PCV 13%) necessitating a packed red blood cell (pRBC) transfusion due to hypophosphatemia induced hemolytic anemia (n = 1). Five patients survived to discharge. Three patients died and 2 patients were euthanized. Two of the 3 patients died suddenly and the ultimate cause of death was unknown. Neither patient had been hypoglycemic during treatment. The other patient that died became hypotensive and was non-responsive to fluid resuscitation and a dopamine CRI. It is suspected that this patient succumbed to complications of necrotizing pancreatitis, although an abdominal ultrasound was not performed due to owner financial constraints. The hypoglycemic patient was ultimately euthanized due to complications of pancreatitis, which included persistent inappetence and worsening hyperbilirubinemia. The other patient that was euthanized had been under treatment for concurrent pancreatitis. This cat underwent respiratory arrest and initially responded to resuscitative measures, but the owners elected humane euthanasia. Necropsies were not performed on these patients. These deceased patients were removed from data analysis for resolution of inappetence, length of treatment, and length of hospitalization.

Post-hoc pairwise comparison via the t-test was utilized to further evaluate differences between the group of survivors and non-survivors. There were no statistically significant differences (P > 0.05) between these 2 groups with regards to the pH during first 24 h, body condition score, total bilirubin, serum osmolality before treatment, blood glucose at presentation, duration of treatment until the blood glucose was < 13.88 mmol/L, albumin, ALT activity, mean units/kg per day of insulin, and rate of blood glucose decline over the first 14 h of treatment. While not statistically significant (P = 0.09), the patients that did not survive had a mean albumin lower than the survivors, 0.16 mmol/L (SD 0.03 mmol/L) compared to 0.19 mmol/L (SD: 0.02 mmol/L), respectively.

Discussion

The use of an insulin CRI to manage feline DKA has previously been shown to be safe and effective (13,14). When comparing the results of the current study to historical controls, there are noteworthy similarities and differences. The signalment for this patient population is comparable to prior studies on feline DKA (2,14). The prevalence of concurrent diseases was similar to or lower than those noted in prior studies; prior studies had prevalences of 50% (14), 92% (2), and 96% (13), while this study showed a 50% prevalence. This study population had more existing diabetics, (60%) compared to 38%, 41%, and 50%, respectively (2,13,14) in prior studies. This patient population was discharged from the hospital sooner (mean 3.8 d) than patients in a prior insulin CRI study (mean 6, 7.1, or 7.3 d) (13). The most noteworthy difference is the survival rate in this patient population compared to that of prior patient populations. Our survival rate was 50%, which is significantly lower than in prior studies utilizing treatment with insulin CRIs. Prior survival rates have ranged from 90% to 93% (13,14). One study evaluating the use of intramuscular and subcutaneous insulin for DKA treatment had a survival rate of 73% (2). Also, a recent paper evaluating the use of intermittent glargine for the treatment of feline DKA found a survival rate of 100% (12). However, this patient population did not have any concurrent diseases, which could explain the more favorable outcomes. Possible explanations for the less favorable survival rate in this patient population include late referral resulting in increased severity of illness at presentation, more insulin-resistant patients, or inadequate insulin dosage. Patients in this study had a mean daily insulin dose of 0.91 u/kg per day during the first 24 h of treatment. This is in contrast to a mean of 0.3 u/kg per day in another group of cats scheduled to receive 1.1 u/kg per day, and 0.76 u/kg per day in a group of cats scheduled to receive 2.2 u/kg per day insulin CRI (13). This difference suggests that the population of cats in our study had a higher insulin demand, possibly from more significant illness.

A gradual decline in blood glucose is one of the initial goals of DKA treatment. The ideal rate of decline of blood glucose, extrapolated from human medicine, is 50 mg/dL per hour (2.78 mmol/L per hour) (7). Patients in this study had an acceptable rate of blood glucose decline, with only 3 patients having a time block with a rate of decline > 50 mg/dL per hour. This is important because consequences of a rapid blood glucose change include hypokalemia, hypophosphatemia, hypoglycemia, or cerebral edema (7,18). However, a prior study showed that cerebral edema, even with large changes in serum blood glucose, is uncommon in feline patients (14). Two of the 3 patients with a more rapid decline in blood glucose were euthanized or died. However, neither patient exhibited signs consistent with electrolyte abnormalities or cerebral edema due to the more rapid change in blood glucose. All 10 patients had at least 1 time point in which their blood glucose increased instead of decreased during treatment. This could indicate that cats with DKA may need an increased dose of insulin, either by increasing the units/kg per day in the insulin CRI or by increasing the rate of the CRI beyond 10 mL/h.

Fifty percent of existing diabetics in this study had recent changes in their insulin type or dosage. To the authors’ knowledge, this has not been documented in prior studies. Only 1 of these patients also had a concurrent disease. It is possible that the change in insulin type or dosage lead to the onset of DKA. However, it is more likely that these patients represent poorly controlled diabetics or insulin-resistant diabetics. These patients are prone to DKA due to their poor regulation, insulin deficiency, and consequently altered handling of fatty acids leading to the development of ketosis (1).

There are several limitations that must be considered when evaluating the study results. Because this was a retrospective study, we were unable to control variables between patients. Specifically, several of the patients received treatment prior to presentation at the referral institution. This may have influenced the end patient outcome. Also, because some patient files were incomplete, we were unable to analyze the time to resolution of ketosis or evaluate the time to resolution of electrolyte imbalances. The retrospective nature of the study also limited our ability to objectively assess the severity of illness through APPLE scoring (19), which may have clarified the poor survival time. The study results are also limited by the small number of cases, which may not be representative of the overall population. This study indicates the need for a prospective study to determine prognostic indicators and the optimal insulin CRI dosage for feline patients with DKA.

In conclusion, this study demonstrated that insulin CRI at a dose of 1.1 units/kg per day for the treatment of DK or DKA in feline patients was effective in 5 of 10 cats. In patients that responded favorably, an insulin CRI provided a gradual decline in blood glucose, permitting a short hospital stay and a transition to long-term injectable insulin once the patient was eating. The study confirmed that many cats with DK/DKA are prior diabetics with concurrent diseases and/or a history of recent insulin dosage or type changes. CVJ