Association between Hyperglycemia and Canine Serum Pancreatic Lipase Immunoreactivity Concentration in Diabetic Dogs
ABSTRACT
It has been reported that hypertriglyceridemia can partially mediate between diabetes mellitus (DM) and pancreatitis in dogs, implying that another mediator, such as chronic hyperglycemia, might exist. Therefore, this study aimed to evaluate the relationship between hyperglycemia and serum canine pancreatic lipase immunoreactivity (cPLI) concentration in diabetic dogs. This retrospective cohort study included 26 client-owned diabetic dogs, divided according to their serum fructosamine levels (<500 μmol/L = well-controlled DM group; ≥500 μmol/L = untreated or poorly controlled DM group). Five of the 26 DM dogs (19.2%) had serum cPLI concentrations consistent with pancreatitis, among which two showed ultrasonographic evidence of pancreatitis without clinical signs. The serum cPLI concentrations (median [interquartile range]) were significantly higher in the untreated or poorly controlled group (520 μg/L [179.76–1000 μg/L]) than in the well-controlled group (77 μg/L [32.22–244.6 μg/L], P = 0.0147). The serum fructosamine concentration was positively correlated with the serum cPLI concentration (r = 0.4816; P = 0.0127). Multivariate analysis revealed serum triglyceride and fructosamine concentrations were associated with the serum cPLI concentration. In conclusion, this study suggests that chronic hyperglycemia may induce pancreatic inflammation in diabetic dogs; however, the clinical significance of increased cPLI concentration is unknown.
Introduction
Diabetes mellitus (DM) is a relatively common disorder of the endocrine pancreas in dogs and is associated with chronic hyperglycemia, with a prevalence of approximately 0.26–0.56%.1–3 DM results from an absolute or relative deficiency of insulin, increased insulin resistance, or both, thus leading to clinical signs of polyuria, polydipsia, polyphagia, and weight loss.4
Because of the anatomical proximity between the exocrine and endocrine pancreas, the concurrent occurrence of pancreatitis and DM has been described in many studies.1,5 DM, like acute pancreatitis (AP) and exocrine pancreatic insufficiency, is an established consequence of chronic pancreatitis (CP). The suggested mechanisms by which pancreatitis leads to DM are through bystander damage, which is the spread of inflammation to the endocrine pancreas.1 The reverse has also been reported in humans and cats, although seldom. Furthermore, several studies have suggested that diabetic human patients have an increased risk of developing AP.6,7 A previous study reported that serum feline pancreatic lipase immunoreactivity (fPLI) concentrations were considerably higher in diabetic cats than in nondiabetic cats, and serum fPLI concentrations were positively correlated with serum fructosamine concentrations.8 However, this relationship has not been elucidated in dogs with DM.
A recent study showed that hypertriglyceridemia (HTG) partially mediates the relationship between DM and pancreatitis in dogs, which implies the presence of another association between these two conditions.9 Another possible mechanism is glucotoxicity, which is associated with prolonged hyperglycemia. Fructosamine is a glycated protein that undergoes irreversible nonenzymatic reactions in proportion to blood glucose concentration. In dogs, serum fructosamine concentration is determined by the average blood glucose concentration over the previous 1–2 wk; therefore, an increased fructosamine concentration is due to prolonged hyperglycemia.4,10
Clinically, pancreatitis is diagnosed based on the clinical signs, abnormalities consistent with pancreatitis on ultrasonography, and serum canine pancreatic lipase immunoreactivity (cPLI) in dogs.11,12 The serum cPLI revealed a good performance characteristic for diagnosing histopathologic lesions of pancreatitis in dogs.11 Furthermore, 73% of dogs with diabetic ketoacidosis (DKA) had evidence of concurrent pancreatic injury based on the serum cPLI activity.13 Theoretically, pancreatitis may develop in relation to chronic hyperglycemia associated with DM in dogs, like in humans and cats. Conversely, it could be possible that proceeding CP leads to pancreatic islet dysfunction in dogs.14 However, this has not been investigated in dogs with DM.
Thereore, it is hypothesized that serum fructosamine concentration is positively associated with serum cPLI in dogs. This study aimed to evaluate the relationship between hyperglycemia and serum cPLI concentration in diabetic dogs.
Materials and Methods
Case Selection
Clinical records from the Veterinary Teaching Hospital database were searched for dogs diagnosed with DM between May 2015 and August 2022. This study was approved by an Institutional Animal Care and Use Committee (CBNAU-1999-22-01). In total, 48 diabetic dogs were initially included in this study. The diagnosis of DM was verified by documentation of compatible clinical signs (polyuria, polydipsia, polyphagia, and weight loss), persistent glucosuria, and persistent hyperglycemia.4 Dogs were excluded from the study if they had apparent risk factors for pancreatitis other than DM as follows:15 (1) dietary factors such as a high-fat diet and dietary indiscretion, (2) severe HTG (serum triglyceride concentration >862 mg/dL),16 (3) drugs (such as L-asparaginase, phenobarbital, potassium bromide, azathioprine, potentiated sulfonamides, corticosteroids, furosemide, atovaquone/proguanil, and clomipramine), (4) endocrinopathy, such as hyperadrenocorticism and hypothyroidism, and (5) tick-borne infections (such as babesiosis and ehrlichiosis). Dogs with a concurrent disease that could potentially affect cPLI elevation other than pancreatitis, such as renal disease, cardiac disease, DKA, hyperadrenocorticism, and enteritis, were also excluded.17 In addition, dogs that had ≥2 previous episodes of AP, although a single AP episode was not exclusionary, were also excluded to clarify the relationship between DM and pancreatitis.1 Dogs with either hypoalbuminemia, azotemia, or hypothyroidism were excluded because of the potential effect on the fructosamine, and samples that were visibly hemolyzed or lipemic were also excluded from the study.10,18 Serum biochemistries were performed using a biochemical analyzera. Finally, 26 dogs with DM were included in the present study.
Animal Care and Use Statement
Informed consent from pet owner was not required because investigating new/unproven/comparative treatments and procedures of a client-owned animal was not required in this; medical records were collected during clinical management. All patients described in this study were clinically managed according to contemporary standards of care.
Analysis of Serum cPLI and Fructosamine Concentrations
Blood samples were collected from the jugular or peripheral veins into plain tubes and centrifuged at 1200g for 10 min within 1 hr of blood collection. For canine-specific pancreatic lipase analysisb, separated sera were frozen on dry ice and sent to a reference laboratoryc. The laboratory reference interval is 0–200 μg/L; however, 201–399 μg/L is a questionable range, and ≥400 μg/L is consistent with pancreatitis.17
Fructosamine was measured in the reference laboratoryc using a colorimetric assay. In diabetic dogs, fructosamine concentrations <500 μmol/L indicate good control, whereas concentrations of ≥500 μmol/L indicate poor control.19 In the present study, 23 dogs were previously treated with insulin, and 3 were newly diagnosed with DM; the latter had serum fructosamine concentrations corresponding to those of the poorly controlled group and were grouped with the poor controls.
Statistical Analyses
A flow chart of the statistical analyses is presented in Supplementary Figure 1. Commercially available statistical softwared,e were used for statistical analyses, and statistical significance was set at P < 0.05. Quantitative data were assessed for normality using the Kolmogorov-Smirnov test. Correlations between the concentrations of fructosamine, cPLI, triglyceride, albumin, and creatinine were evaluated between two variables using Pearson or Spearman’s correlation tests, depending on normality. The association between these variables and fructosamine and cPLI concentrations was also examined, accounting for the confounding effect with partial correlation. In addition, cPLI concentration was compared between the well-controlled and untreated or poorly controlled DM groups using the Mann-Whitney U test, based on serum fructosamine concentration. In addition, fructosamine concentration was compared between diabetic dogs with cPLI concentration <400 μg/L and those with cPLI concentration ≥400 μg/L using the Mann-Whitney U test. Multiple linear regression analysis was used to identify the association between serum cPLI concentration (dependent variable) and other variables, including serum fructosamine, albumin, and creatinine concentrations (independent variables). A multiple forward stepwise regression analysis was performed to determine whether two or more variables could predict serum cPLI concentration.
Results
A total of 26 dogs with DM, 19 with well-controlled DM and 7 with untreated or poorly controlled DM, were included in this study between May 2015 and August 2022. Data on age at sample collection, sex, breed, body condition score, and comorbidities were collected. Of the 26 diabetic dogs, 16 (61.5%) were castrated males, and 8 (30.8%) were spayed females. The DM group included four Maltese, four miniature poodles, four mixed-breed dogs, three Yorkshire terriers, two Pomeranians, two miniature schnauzers, and one each of the following breeds: cocker spaniel, beagle, bichon frise, spitz, Pekingese, Chihuahua, and shih tzu.
Serum fructosamine concentrations were positively correlated with serum cPLI concentrations (r = 0.4816; P = 0.0127; Figure 1) in dogs with DM. The correlation remained significant after adjusting for the potential confounding effects of serum triglyceride, albumin, and creatinine levels (Table 1). In contrast, serum fructosamine concentrations were not correlated with serum triglyceride, albumin, or creatinine concentrations. Similarly, serum cPLI concentrations were not significantly associated with serum albumin or creatinine levels (Figure 2). These correlations were also not found after adjusting for albumin and creatinine levels using partial correlation analysis. There was also no simple correlation between serum cPLI and triglyceride concentrations; however, after adjusting for fructosamine using partial correlation analysis, a significant positive correlation was observed (partial r = 0.502; P = 0.011; Table 1). In multiple stepwise regression analysis, serum triglyceride and fructosamine concentrations were predictors of serum cPLI concentrations (P = 0.002; R2 = 0.426; Table 2). Albumin and creatinine concentrations did not affect serum cPLI concentrations.
Citation: Journal of the American Animal Hospital Association 59, 5; 10.5326/JAAHA-MS-7365
Citation: Journal of the American Animal Hospital Association 59, 5; 10.5326/JAAHA-MS-7365
Of the 26 diabetic dogs, 10 (38.5%) had cPLI concentrations >200 μg/L, and 5 (19.2%) had cPLI concentrations ≥400 μg/L. The serum cPLI concentration (median [interquartile range]) was significantly higher in the untreated or poorly controlled DM group (520 μg/L [179.76–1000 μg/L]) than in the well-controlled DM group (77 μg/L [32.22–244.6 μg/L]; P = 0.0147; Figure 3A). Additionally, serum fructosamine concentration was significantly higher in diabetic dogs with cPLI concentrations ≥400 μg/L (569 μmol/L [472–727 μmol/L]) than in those with cPLI concentrations <400 μg/L (423 μmol/L [370.5–475 μmol/L]; P = 0.0124; Figure 3B).
Citation: Journal of the American Animal Hospital Association 59, 5; 10.5326/JAAHA-MS-7365
Discussion
This study evaluated serum cPLI and fructosamine concentrations in diabetic dogs to examine the association between chronic hyperglycemia and pancreatic injury. Dogs with DM were evaluated and classified based on circulating fructosamine concentrations. Circulating fructosamine concentrations were positively correlated with cPLI concentrations. This is consistent with the results of a study on cats, in which a linear association was found between serum fructosamine and fPLI concentrations.8 In the present study, we conducted further analyses, in addition to simple correlation analysis, to determine a cause-and-effect relationship between the serum cPLI and fructosamine concentrations. Serum fructosamine and cPLI concentrations maintained a significant positive correlation after adjustment for serum triglyceride, albumin, and creatinine levels. In addition, multiple linear regression analysis was performed to determine the association with increased cPLI concentrations. Multivariate analysis revealed an association between serum fructosamine and cPLI concentrations. Furthermore, multivariate analysis revealed that serum fructosamine and triglyceride concentrations were independent factors that influenced cPLI concentrations. It has been reported that HTG secondary to DM is a risk factor for AP.9 Consistent with this previous study, the present study supports the possibility that hyperglycemia might be another potential mediator for pancreatitis in dogs with DM. These findings suggest chronic hyperglycemia can induce pancreatic injury in dogs with DM.
According to a previous report, human patients with DM who had a history of hyperglycemic crisis episodes had a higher risk of AP than those without hyperglycemic crisis episodes.7 In the present study, a significant difference of serum cPLI was observed when the dogs were divided based on a circulating fructosamine concentration of 500 μmol/L. There was also a significant difference in serum fructosamine concentration when the diabetic dogs were divided based on a circulating cPLI concentration of 400 μg/L, suggesting that diabetic dogs with cPLI concentrations consistent with pancreatitis had been chronically affected by higher average blood glucose concentrations than those without cPLI concentrations consistent with pancreatitis.
In the DM group in the present study, 38.5% of the dogs had cPLI concentrations >200 μg/L, whereas 19.2% had cPLI concentrations consistent with pancreatitis. This is consistent with a recent study20 that reported that 33% of the diabetic dogs had cPLI concentrations >200 μg/L, whereas 20% had cPLI concentrations ≥400 μg/L when dogs with DKA were excluded because DKA may induce pancreatitis through HTG- or hypovolemia-induced hypoxic damage rather than glucotoxicity itself.13 However, compared with our study, the previous study20 had different inclusion criteria, although both studies included a small number of dogs with DM; therefore, additional strict and large studies would be required to estimate the prevalence of serum cPLI elevation in diabetic dogs.
The role of hyperglycemia in the pathogenesis of pancreatitis is not completely understood; however, several mechanisms have been proposed. In humans, chronic hyperglycemia induces increased reactive oxygen species generation and lipid peroxidation, which may be associated with the development of AP.7,21 Chronic hyperglycemia-induced oxidative stress was also identified in feline DM.22 This oxidative damage or hypovolemia resulting from persistent hyperglycemia may lead to pancreatic injury. Protracted hyperglycemia impairs insulin production and induces insulin resistance, which might play a major role in pancreatic inflammation because insulin attenuates hyperglycemia-mediated oxidative stress and stimulates an anti-inflammatory cascade.23,24 In addition, insulin resistance and diabetes state impair communication between the endoplasmic reticulum and mitochondria, leading to an imbalance in calcium homeostasis and signaling in the pancreas.25 This calcium signaling imbalance subsequently induces AP by early activation of trypsinogen and mitochondrial dysfunction.15 All these reports supported that glucotoxicity can induce inflammation of the exocrine pancreas. Increased serum cPLI concentrations in the present study’s results might reflect glucotoxicity-induced pancreatic inflammation in dogs with DM; however, clinical pancreatitis was not identified in the present study.
There is a relationship between DM and changes in pancreatic structure in humans. Furthermore, some studies have shown that human patients with DM have smaller pancreas than non-DM controls.26 Diabetic patients also had islet histological lesions such as fibrosis, fatty infiltration, lymphocytic infiltration, and atherosclerosis in the pancreas.26 Pancreatic fibrosis has been reported to be aggravated by hyperglycemia, and β-cell mass decreases as fibrosis progresses.27 In a study on 37 diabetic cats, postmortem examination revealed exocrine abnormalities in 73% of the cats, islet abnormalities in 89%, and both exocrine and endocrine abnormalities in 57%. Of the cats with exocrine pancreatic abnormalities, AP was present in 5% of the cats, whereas CP was present in 46%.28 Another study reported that hyperglycemia increased neutrophils histologically in the exocrine pancreas of cats, suggesting pancreatic inflammation.29 Because histological examination was not performed in the present study, whether histological changes were indeed present in the pancreas of diabetic dogs could not be confirmed. However, a study reported that histological changes associated with AP or CP were identified in 6 of 18 dogs with spontaneous diabetes, and only 2 had a history of tentatively diagnosed pancreatitis.30 Collectively, given these previous results and the present study’s finding of a positive correlation between serum fructosamine and cPLI concentrations in diabetic dogs, histological alterations in dogs with DM could be associated with chronic hyperglycemia.
cPLI concentrations consistent with pancreatic injury were measured in five diabetic dogs; however, none showed clinical signs of pancreatitis, such as hyporexia, abdominal pain, and vomiting. Furthermore, two dogs showed ultrasonographic evidence of pancreatitis, including enlarged, irregular, and hypoechoic pancreas and hyperechoic mesentery;17 however, no dogs showed clinical signs. Therefore, elevated cPLI concentrations in diabetic dogs may be associated with subclinical pancreatic injury/inflammation. Although such diabetic dogs could have subclinical pancreatitis, only high cPLI result may not indicate a clinical diagnosis of pancreatitis in dogs with DM. However, further study will be necessary to clarify this assumption in dogs with DM demonstrating high cPLI concentrations without ultrasonographic evidence of pancreatitis.
The serum concentrations of albumin and renal function could affect the fructosamine and cPLI concentrations in dogs.18,31 However, the correlation between the serum fructosamine and cPLI concentrations persisted after adjusting for the confounding effect of the serum concentrations of triglyceride and albumin. The effect of the serum albumin concentration could be excluded by the exclusion of diabetic dogs with hypoabuminemia because serum fructosamine could be highly correlated to albumin in only dogs with hypoalbuminemia.10 Furthermore, the serum creatinine concentrations did not correlate with the serum fructosamine and cPLI concentrations in dogs with DM. The serum fructosamine concentrations could be affected by the presence of azotemia in dogs.18 However, in the present study, the exclusion of azotemic dogs with DM could minimize the effect of decreased renal function on the serum fructosamine level. In addition, decreased glomerular filtration rate and urinary clearance are unlikely to affect the serum cPLI concentrations in dogs.32 Therefore, the possibility of the effect of serum albumin and renal function on our results could be low.
Our study had several limitations. First, because of the retrospective nature of this study, the duration of DM until referral to a veterinary teaching hospital varied. Whereas some human studies have identified an inverse correlation between the duration of DM and pancreatic size,26 others could not find such a correlation.33,34 Furthermore, a recent study on diabetic dogs reported no correlation between the duration of DM and canine trypsin-like immunoreactivity concentrations.20 Moreover, inferring a cause-and-effect relationship between chronic hyperglycemia associated with DM and increased cPLI is challenging owing to the retrospective design. AP could be attributed to transient insulin deficiency, whereas CP could result in permanent loss of pancreatic β-cells.1 However, we excluded the dogs that had ≥2 previous episodes of AP to minimize the possibility that CP preceded DM. Therefore, chronic hyperglycemia associated with DM could be attributed to increased cPLI in dogs of the present study. Nevertheless, a cause-and-effect relationship between the elevated serum fructosamine and cPLI concentrations in dogs with DM could not be clearly inferred owing to the retrospective study design. To confirm the cause-and-effect relationship, it would be necessary to histopathologically examine the pancreas before, during, and after optimizing diabetic control in dogs with DM; however, the procedure could be challenging in clinical practice.
Second, confirmation using the histopathological examination was not performed on any dogs in this study; therefore, it is impossible to distinguish between AP and CP, and it is also impossible to document whether the pancreatic histological alteration is more severe in the untreated or poorly controlled DM group. Considering that the degree of agreement between cPLI and histopathology was relatively high and that pancreatic inflammation was localized and overlooked in biopsy,17 even if the high cPLI concentrations alone could not determine a definitive diagnosis of pancreatitis, it could indicate an inflammation of the pancreas if there was no other disease affecting serum cPLI concentration.
Third, our results could not confirm whether there was an increased risk of pancreatitis in diabetic dogs with high blood glucose concentrations and poorly controlled DM. Instead, our results suggest that chronic and severe hyperglycemia may induce pancreatic acinar cell injury in diabetic dogs. It is conceivable that additional insults associated with the activation of nuclear factor-κ β and the release of interleukin-6 in acinar cell injury may predispose dogs to pancreatitis.15 Further longitudinal research is necessary to confirm the higher risk of pancreatitis in dogs with poorly controlled DM.
Fourth, the serum creatinine concentration was used as a substitute for the glomerular filtration rate35 owing to the retrospective nature of the study. However, serum creatinine could be an insensitive marker for detecting an early decrease in the kidney function in dogs, which could be affected by muscle mass. In addition, the serum creatinine concentration is strongly correlated with age and weight as well as the glomerular filtration rate and moderately correlated with the body condition score in dogs,36 which could complicate certain results of the present study in dogs with DM, which is a muscle-wasting disease. Thus, some potential effect of signalment such as age, weight, and body condition score on the results of the present study could not be completely excluded. However, age, body weight, and body condition score did not correlate with the serum fructosamine and cPLI concentrations (data not shown), and mild decreased renal function, which cannot be detected by serum creatinine, may not affect the serum fructosamine and cPLI concentration in dogs.18,32
Finally, the small sample size in this study might have caused type 2 statistical errors. We excluded more cases than expected to minimize the possibility that CP preceded DM; however, a cause-and-effect relationship between elevated serum fructosamine and cPLI concentrations in dogs with DM could not be elucidated. Further studies with larger sample sizes will be beneficial to achieving a solid conclusion.
Conclusion
This study provides evidence to support a positive correlation between the serum fructosamine and cPLI concentrations in diabetic dogs. Therefore, this study’s findings suggest that prolonged hyperglycemia could induce subclinical pancreatic injury in diabetic dogs, especially in those with poorly controlled DM and evaluating serum cPLI concentrations in diabetic dogs could contribute to case management. However, serum cPLI concentrations in diabetic dogs need to be considered in light of clinical signs and managed accordingly, with the aim of optimizing diabetic control because only high serum cPLI concentration may not indicate a clinical diagnosis of pancreatitis in dogs with DM. Furthermore, pre-existing or subclinical pancreatitis in dogs with DM were not completely excluded in the present study. Further studies with a larger sample size are necessary to achieve better results.
The authors thank the owners of the dogs included in this study. Informed consent was obtained from the pet owners.
Analysis of the correlation analysis between serum fructosamine and cPLI concentrations in diabetic dogs. The serum cPLI concentration was positively (weakly) correlated with fructosamine concentration (r = 0.4816, P= 0.0127). cPLI, canine pancreatic lipase immunoreactivity.
Analysis of the correlation analysis between serum fructosamine concentration and other variables such as (A) TG, (B) albumin, and (C) creatinine concentrations. No association was found between fructosamine concentration and the other variables. No association was also found between cPLI concentration and (D) TG, (E) albumin, and (F) creatinine concentrations. cPLI, canine pancreatic lipase immunoreactivity; TG, triglyceride.
(A) Comparison of serum cPLI concentration between dogs with well-controlled and untreated or poorly controlled DM, based on serum fructosamine concentration of 500 μmol/L. A significant difference was found when the dogs were divided based on the fructosamine concentration of 500 μmol/L (Mann-Whitney test, P = 0.0147). (B) Comparison of serum fructosamine concentration between dogs with cPLI concentration ≥400 μg/L and <400 μg/L. Median fructosamine concentration was also significantly higher in dogs with cPLI concentration ≥400 μg/L than in those with cPLI concentration <400 μg/L (Mann-Whitney test, P = 0.0124). cPLI, canine pancreatic lipase immunoreactivity; DM, diabetes mellitus.
Contributor Notes
The online version of this article (available at www.jaaha.org) contains supplementary data in the form of one figure.
