Plasma lipid components have dynamic roles in the physiologic responses needed to maintain systemic homeostasis during periods of critical illness in people, and certain lipid components are beneficial in adapting to physiologic distress. For example, endogenous HDL, LDL, and cholesterol may be beneficial in adapting to a spectrum of human diseases.1–6 Other plasma lipid components may be antagonistic to recovery, as are plasma nonesterified fatty acids that may contribute to progressive cardiac dysfunction in people with sepsis.7 Furthermore, high plasma triglycerides concentration may be a risk factor for development of acute pancreatitis and cardiac disease in hospitalized people.1,8,9 The net effect of lipids during a patient's adaptation to severe disease is likely dependent on plasma lipid concentrations, body lipid metabolism, and total adipose stores at the time of the critical illness.
In people, systemic disease can significantly alter lipid metabolism, thereby changing the lipid profile in blood and availability in tissues.1,2,10–16 Specifically, inflammatory cytokines TNF-α, interleukin-1β, and interleukin-6 released during severe inflammation suppress hepatic production of HDL apolipoproteins and cholesterol.2,3,17 Lipopolysaccharides, TNF-α, interleukin-1β, and interleukin-6 increase hepatic triglyceride synthesis and release but concurrently downregulate endothelial lipoprotein lipase activity.14,16–20 In addition, TNF-α causes an increase in plasma nonesterified fatty acids concentration by stimulating lipolysis in adipocytes.21 Lower concentrations of the immunomodulatory lipid lysophosphatidylcholine, possibly from attenuated phospholipase A2 activity, are observed in human ICU patients with sepsis; sepsis is associated with higher concentrations of VLDL and triglycerides.1,22 Without therapeutic intervention, critical illness can promote a blood lipid profile with lower concentrations of protective lipids (eg, cholesterol) and higher concentrations of proinflammatory lipids (eg, triglycerides and nonesterified fatty acids). Additionally, because severe systemic disease may alter blood lipid profiles, the serum lipid concentrations can be useful prognostic markers for survival in critically ill people. In particular, development and persistence of hypocholesterolemia is a negative prognostic indicator for survival in people with blunt trauma, sepsis, organ failure, acute renal failure, and SIRS,2,11,23,24 whereas hypertriglyceridemia is associated with increased disease severity and is a negative prognostic indicator for survival in people with critical illness and sepsis.11,18
Despite abundant evidence in the human medicine literature that lipid metabolism is altered during many disease states and that these alterations are associated with patient prognosis, the prognostic importance of dyslipidemias in veterinary medicine is minimally reported. Giunti et al25 observed no difference in serum cholesterol concentration between healthy control dogs and dogs with SIRS or between hospitalized dogs that survived or died. However, hypertriglyceridemia was identified in 100% of horses with SIRS and was associated with progressive renal disease in horses.26 Therefore, the primary aims of the study reported here were to evaluate the lipidemia status and serum concentrations of cholesterol and triglycerides of dogs when initially examined for hospitalization in the ICU of a veterinary teaching hospital and to determine whether these variables were predictive of survival to hospital discharge. We hypothesized that sick dogs hospitalized in the ICU would have lower serum concentrations of cholesterol and a higher prevalence of hypocholesterolemia than would healthy dogs. We also expected that sick dogs hospitalized in the ICU would have higher serum concentrations of triglycerides and a higher prevalence of hypertriglyceridemia than would healthy dogs. Further, we anticipated that sick dogs not surviving to hospital discharge (nonsurvivors) would have lower serum cholesterol concentrations and higher serum triglycerides concentrations than would sick dogs that survived to discharge (survivors) and that these variables could be predictive of survival to hospital discharge.
Materials and Methods
Animals
The electronic medical record database of the Iowa State University Lloyd Veterinary Medical Center was searched to identify dogs that had been hospitalized in the ICU between January 1, 2012, and September 1, 2015. Dogs were eligible for inclusion if they had been hospitalized in the ICU for an illness requiring treatment; a complete physical examination, CBC, and serum biochemical profile had been performed during the initial evaluation; the underlying diagnosis had been confirmed; and survival status (discharged vs died or euthanized) was known. Each diagnosis was categorized by primary disease process as an autoimmune, cardiovascular, hepatobiliary, pancreatic, or renal disease; coagulopathy; endocrinopathy; gastroenteropathy; infectious or sterile inflammation; neoplasia; or trauma. Dogs that were hospitalized in the ICU strictly for monitoring purposes (eg, postoperative care following elective surgical procedures) were excluded from the study.
The medical record database was also searched to identify healthy dogs examined during the same period. Healthy dogs were eligible for inclusion in the study if they lacked evidence of systemic disease and a complete physical examination, CBC, and serum biochemical profile had been performed during the initial evaluation. The reason each healthy dog had been examined was categorized as elective dental procedure, elective orthopedic procedure, or wellness examination.
For each dog included, data collected from medical records for the initial evaluation were signalment (age, breed, sex, and neuter status), physical examination findings (body condition score on a scale of 1 to 9, heart and respiratory rates, periodontal disease grade on a scale of 1 to 3, rectal temperature, and body weight), and results of clinicopathologic analyses (eg, WBC count, band neutrophil count and fraction, serum concentrations of cholesterol and triglycerides, serum CTR, and lipidemia status). Lipidemia status of each dog was determined by comparing its serum concentrations of cholesterol and triglycerides to the reference intervals of 3.4 to 7.8 mmol/L and 0.27 to 1.3 mmol/L, respectively.27 Dogs with serum cholesterol concentrations < 3.4 mmol/L, from 3.4 to 7.8 mmol/L, and > 7.8 mmol/L were classified as hypo-, normo-, and hypercholesterolemic, respectively. Dogs with serum triglycerides concentrations < 0.27 mmol/L, from 0.27 to 1.3 mmol/L, and > 1.3 mmol/L were classified as hypo-, normo-, and hyper-triglyceridemic.
For sick dogs hospitalized in the ICU, additional information was collected regarding glucocorticoid administration, nutritional support, duration of hospitalization, SIRS status, and survival to hospital discharge. Dogs were considered to have received nutritional support if they received enteral or parenteral nutritional treatment. Sick dogs were classified as having SIRS if they exhibited > 2 of the following criteria at admission: rectal temperature < 38.1°C (100.6°F) or > 39.2°C (102.6°F; reference range, 37.9° to 39.2°C [100.2° to 102.5°F]), heart rate > 120 beats/min (reference range, 70 to 120 beats/min), respiratory rate > 20 breaths/min (reference range, 18 to 35 breaths/min), WBC count < 6,000 WBCs/μL or > 16,000 WBCs/μL (reference range, 3,000 to 11,400 WBCs/μL), and percentage of band neutrophils > 3% (reference range, 0% to 2.5%).27 Sick dogs with < 2 of these criteria at admission were classified as not having SIRS. Sick dogs that survived to hospital discharge were classified as survivors, whereas those that died or were euthanized before hospital discharge were classified as nonsurvivors.
Laboratory methods
On the basis of medical records and standard operating procedures of the veterinary teaching hospital, it was known that blood samples for each patient had been collected from a peripheral vein and placed into plain tubes and tubes containing EDTA during the initial examination. These samples were analyzed < 24 hours after collection, with the WBC count and band neutrophil count and fraction determined with an automated hematologic analyzera and manual blood smear evaluation.28,29 Serum concentrations of cholesterol and triglycerides were measured with an automated clinical biochemical analyzerb by colorimetric, reflectance spectrophotometry assays validated for use in dogs. The CTR was calculated by dividing each dog's serum cholesterol concentration by its serum triglycerides concentration, and the reference range used for CRT in dogs was 0 to 3. All measurements were performed by the Iowa State University Veterinary Clinical Pathology Laboratory.
Statistical analysis
Data were assessed for normality with the D'Agostino-Pearson test. Because most data were nonparametric, the Wilcoxon rank sum test was used to compare continuous variables between 2 groups, and a Kruskal-Wallis ANOVA with the Dunn test for post hoc comparisons was used to compare continuous variables among > 2 groups. The Fisher exact test was used to evaluate differences in categorical variables between groups, and the Spearman rank-order test was used to assess correlations between continuous variables. Correlations were considered present when r > 0.2 or r < −0.2 and P < 0.05. Univariate logistic regression analysis was performed to initially identify continuous and categorical variables associated with survival to hospital discharge. Variables that were significant in the univariate analysis were then incorporated into a multivariate logistic regression model with backward removal. Significance was set as P < 0.05 for all statistical tests. Data were presented as median and range or total number and percentage, where appropriate. Statistical evaluations were performed with available software.c–e
Ethics and consent
At the time of the study, the Institutional Animal Care and Use Committee of Iowa State University did not require ethical approval of retrospective studies that collected data generated through routine clinical assessment and care of privately owned animals. The authors obtained permission from the Veterinary Teaching Hospital and Veterinary Clinical Pathology Laboratory of Iowa State University to use the involved clinical and laboratory patient data, such that the private information of the owners and patients remained anonymous.
Results
Animals
The initial record search identified 568 dogs, of which 549 dogs (151 [27.5%] healthy dogs and 398 [72.5%] hospitalized sick dogs) satisfied inclusion criteria. The primary reasons for which the 151 healthy dogs were evaluated were categorized as elective orthopedic procedures (n = 123 [81.5%]), wellness examinations (24 [15.9%]), or elective dental procedures (4 [2.6%]). The primary disease processes affecting the 398 sick dogs hospitalized in the ICU were categorized as infectious or sterile inflammation (n = 101 [25.4%]), neoplasia (86 [21.6%]), autoimmune disease (63 [15.8%]), gastroenteropathy (42 [10.6%]), endocrinopathy (37 [9.3%]), pancreatic disease (26 [6.5%]), hepatobiliary disease (18 [4.5%]), renal disease (8 [2.0%]), trauma (7 [1.8%]), coagulopathy (6 [1.5%]), and cardiovascular disease (4 [1.0%]). In addition, 319 of the 398 (80.2%) hospitalized sick dogs had signs meeting the criteria for having SIRS, whereas 79 (19.8%) did not. Further, 282 of the 398 (70.9%) sick dogs survived to hospital discharge, whereas 116 (29.1%) did not.
On initial evaluation, healthy dogs had significantly (P < 0.01) higher median body weight (32.3 kg [71.1 lb]; range, 0.9 to 90.9 kg [2.0 to 200.0 lb]), body condition score (6; range, 3 to 9), rectal temperature (38.9°C [102.2°F]; range, 37.5° to 40.6°C [99.5° to 105.8°F]), and respiratory rate (60 breaths/min; range, 20 to 240 breaths/min) than did sick dogs (Table 1). Sick dogs, however, had a significantly (P < 0.01) higher median heart rate (125 beats/min; range, 44 to 280 beats/min) and WBC count (15,235 WBCs/μL; range, 800 to 95,180 WBCs/μL) than did healthy dogs. Although the median band neutrophil count was 0 for healthy and sick dogs, the range was higher for sick dogs (0 to 8,561 cells/μL), compared with healthy dogs (0 to 168 cells/μL). Additionally, a significantly (P < 0.01) higher proportion of sick dogs (64/398 [16.1%]) were sexually intact, compared with that of healthy dogs (10/151 [6.6%]). There were no meaningful differences between healthy and sick dogs for median age, periodontal disease grade, or sex. The proportion of Labrador Retrievers was significantly (P < 0.01) higher in the healthy group (35/151 [23.2%]) than in the sick group (38/398 [9.5%]); however, there was no meaningful difference between the groups in the representation of other breeds (Supplementary Table S1, available at avmajournals.avma.org/doi/suppl/10.2460/javma.254.6.699). Of the sick dogs, nonsurvivors had significantly higher median age (8 years; range 1 to 16 years; P < 0.01), heart rate (130 beats/min; range, 70 to 230 beats/min; P = 0.02), and band neutrophil count (108 cells/μL; range, 0 to 8,561 cells/μL; P < 0.01), compared with survivors (Table 2). There was no meaningful difference in the proportion of breeds represented between survivors and nonsurvivors.
Cholesterol and triglycerides concentrations and CTR
There was no meaningful difference in median serum cholesterol concentration between healthy and sick dogs (Table 1; Figure 1), nor between sick dogs with and without SIRS. However, sick dogs did have a significantly (P < 0.01) higher median serum triglycerides concentration (0.8 mmol/L; range, 0.2 to 105.0 mmol/L) than did healthy dogs (0.6 mmol/L; range, 0.1 to 5.6 mmol/L), and sick dogs with SIRS had a significantly (P < 0.01) higher median serum triglycerides concentration (0.84 mmol/L; range 0.2 to 105.0 mmol/L) than did sick dogs without SIRS (0.6 mmol/L; range 0.23 to 36.1 mmol/L). Median CTR was significantly (P < 0.01) lower for sick dogs (6.2; range, 0.1 to 35.4), compared with healthy dogs (9.0; range, 0.9 to 42.5). In addition, the median CTR was significantly (P = 0.02) lower for sick dogs with SIRS (6.0; range, 0.1 to 35.4), compared with sick dogs without SIRS (7.8; range, 0.2 to 26.5).
Descriptive statistics for select variables for 398 sick dogs hospitalized in the ICU of a veterinary teaching hospital between January 1, 2012, and September 30, 2015, and for 151 healthy dogs evaluated at the teaching hospital during the same time frame.
| Variable | Healthy dogs | Sick dogs | P value |
|---|---|---|---|
| Sex | 0.18 | ||
| Male | 65 (43.1) | 197 (49.5) | |
| Female | 86 (56.9) | 200 (50.5) | |
| Sexual status | < 0.01 | ||
| Sexually intact | 10 (6.6) | 64 (16.1) | |
| Neutered | 141 (93.4) | 334 (83.9) | |
| Age (y)* | 7 (0–15) | 7 (0–16) | 0.10 |
| Weight (kg) | 32.3 (0.9–90.9) | 17.5 (1.4–74.0) | < 0.01 |
| Body condition score (scale of 1–9) | 6 (3–9) | 5 (1–9) | < 0.01 |
| Periodontal disease grade (scale of 1–3) | 2 (1–3) | 2 (1–3) | 0.29 |
| Rectal temperature (°C) | 38.9 (37.5–40.6) | 38.7 (35.1–41.6) | < 0.01 |
| Heart rate (beats/min) | 108 (42–222) | 125 (44–280) | < 0.01 |
| Respiratory rate (breaths/min) | 60 (20–240) | 44 (10–240) | < 0.01 |
| WBCs (cells/μL) | 8,060 (4,670–23,370) | 15,235 (800–95,180) | < 0.01 |
| Band neutrophils (cells/μL) | 0 (0–168) | 0 (0–8,561) | < 0.01 |
| Serum cholesterol concentration (mmol/L) | 5.7 (2.8–12.5) | 5.7 (0.4–17.7) | 0.82 |
| Serum triglycerides concentration (mmol/L) | 0.6 (0.1–5.6) | 0.8 (0.2–105.0) | < 0.01 |
| CTR | 9.0 (0.9–42.5) | 6.2 (0.1–35.4) | < 0.01 |
| Duration of hospitalization (d) | — | 2 (0–21) | — |
| Outcome for sick dogs | |||
| Survivors | — | 282 (70.9) | — |
| Nonsurvivors | — | 116 (29.1) | — |
Continuous variables are given as median (range), and categorical variables are given as number (percentage) of dogs. The Wilcoxon rank sum test and Fisher exact test were used to evaluate differences between groups.
Age range lower limit of 0 indicates dogs < 1 year of age.
— = Not applicable.
Descriptive statistics for select clinical and clinicopathologic variables for the sick dogs in Table 1 that did (survivors; n = 282) or did not (nonsurvivors; 116) survive to hospital discharge.
| Variable | Survivors | Nonsurvivors | P value |
|---|---|---|---|
| Sex | 0.32 | ||
| Male | 145 (51.4) | 53 (45.7) | |
| Female | 137 (48.6) | 63 (54.3) | |
| Sexual status | 0.09 | ||
| Sexually intact | 51 (18.1) | 13 (11.2) | |
| Neutered | 231 (81.9) | 103 (88.8) | |
| Age (y)* | 7 (0–15) | 8 (1–16) | < 0.01 |
| Weight (kg) | 18.2 (2.0–66.8) | 15.6 (1.4–74.0) | 0.46 |
| Body condition score (scale of 1–9) | 5 (1–9) | 5 (1–9) | 0.60 |
| Periodontal disease grade (scale of 1–3) | 2 (1–3) | 2 (1–3) | 0.58 |
| Rectal temperature (°C) | 38.7 (35.11–41.2) | 38.6 (35.8–41.6) | 0.74 |
| Heart rate (beats/min) | 124 (44–280) | 130 (70–230) | 0.02 |
| Respiratory rate (breaths/min) | 44 (16–204) | 44 (10–240) | 0.41 |
| WBCs (cells/μL) | 14,930 (1,010–76,130) | 16,100 (800–95,180) | 0.60 |
| Band neutrophils (cells/μL) | 0 (0–6,607) | 108 (0–8561) | < 0.01 |
| Serum cholesterol concentration (mmol/L) | 5.7 (1.7–17.7) | 5.6 (0.4–15.0) | 0.44 |
| Serum triglycerides concentration (mmol/L) | 0.7 (0.2–105.0) | 1.0 (0.2–31.7) | < 0.01 |
| CTR | 7.2 (0.1–35.4) | 5.0 (0.2–26.8) | < 0.01 |
| Duration of hospitalization (d) | 2 (0–21) | 2 (0–14) | 0.23 |
Age range lower limit of 0 indicates dogs < 1 year of age.
Box-and-whisker plots of serum concentrations of cholesterol (A) and triglycerides (B) and the CTR (C) at initial evaluation of 398 sick dogs hospitalized in the ICU at a veterinary teaching hospital between January 1, 2012, and September 30, 2015, and of 151 healthy dogs evaluated at the teaching hospital during the same time frame. Each box represents data from the 25th to 75th percentile, the line in the box represents the median, and the whiskers represent the range. No significant difference in median serum cholesterol concentration was observed between healthy and sick dogs; however, sick dogs had a significantly (P < 0.01) higher median serum triglycerides concentration and a significantly (P < 0.01) lower CTR than did healthy dogs.
Citation: Journal of the American Veterinary Medical Association 254, 6; 10.2460/javma.254.6.699
Sick dogs with an endocrinopathy had significantly (P = 0.01 and P < 0.01, respectively) higher median serum concentrations of cholesterol and triglycerides than did dogs with a primary gastroenteropathy, and sick dogs with an endocrinopathy also had significantly (P = 0.02) higher median serum cholesterol concentration than did dogs with neoplasia (Figure 2). Further, dogs with pancreatic disease had a significantly (P < 0.01) higher median serum triglycerides concentration than did dogs with a gastroenteropathy. No substantial differences in median serum cholesterol or triglycerides concentration were detected among dogs in the remaining disease categories. Further, no substantial differences in median CTR were detected among dogs grouped by disease categories. When sick dogs with a primary endocrinopathy or pancreatic disorder were removed from the analysis, sick dogs still had higher median serum triglycerides concentration and lower CTR, compared with such in healthy dogs.
No difference in median serum cholesterol concentration was observed between survivors and nonsurvivors (Table 2; Figure 3). However, survivors had a significantly (P < 0.01) lower median serum triglycerides concentration (0.7 mmol/L; range, 0.2 to 105.0 mmol/L), compared with nonsurvivors (1.0 mmol/L; range 0.2 to 31.7 mmol/L). Survivors also had a significantly (P < 0.01) higher median CTR (7.2; range, 0.1 to 35.4), compared with nonsurvivors (5.0; range, 0.2 to 26.8).
Lipidemia status
Hypertriglyceridemia and hyper- and hypocholesterolemia were significantly (P < 0.01) more common in sick dogs, compared with healthy dogs, whereas normotriglyceridemia and normocholesterolemia were significantly (P < 0.01) more common in healthy dogs (Table 3). There was no meaningful difference in the proportions of healthy and sick dogs with hypotriglyceridemia.
In sick dogs, hypocholesterolemia was significantly (P = 0.04) more common in nonsurvivors (23/116 [19.8%]) than in survivors (33/282 [11.7%]), whereas normocholesterolemia was significantly (P = 0.02) more common in survivors (192/282 [68.1%]) than in nonsurvivors (64/116 [55.2%]; Table 3). Similarly, hypertriglyceridemia was significantly (P < 0.01) more common in nonsurvivors (50/116 [43.1%]) than in survivors (54/282 [19.1%]), whereas normotriglyceridemia was significantly (P < 0.01) more common in survivors (220/282 [78.0%]) than in nonsurvivors (64/116 [55.2%]). Further, all 10 dogs with concurrent hypocholesterolemia and hypertriglyceridemia were in the nonsurvivor group, and dogs with concurrent hypocholesterolemia and hypertriglyceridemia were far more likely (Fisher exact test; OR, 55.7; 95% CI, 3.2 to 959.6) to be nonsurvivors than were dogs without concurrent hypocholesterolemia and hypertriglyceridemia.
Box-and-whisker plots of the serum concentrations of cholesterol (A) and triglycerides (B) and the CTR (C) for the sick dogs in Figure 1 grouped according to primary disease process category. See Figure 1 for key.
Citation: Journal of the American Veterinary Medical Association 254, 6; 10.2460/javma.254.6.699
Box-and-whisker plots of the serum concentrations of cholesterol (A) and triglycerides (B) and the CTR (C) at the time of initial evaluation for the sick dogs in Figure 1 that did (survivors; n = 282) or did not (nonsurvivors; 116) survive to hospital discharge. Survivors had a significantly (P < 0.01) lower median serum triglycerides concentration and a significantly (P < 0.01) higher median CTR, compared with nonsurvivors. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 254, 6; 10.2460/javma.254.6.699
Lipidemia status for the 151 healthy and 398 sick (282 survivors and 116 nonsurvivors) dogs in Table 1.
| All dogs | Sick dogs | |||||
|---|---|---|---|---|---|---|
| Lipidemia status | Healthy | Sick | P value | Survivors | Nonsurvivors | P value |
| Hypercholesterolemia | 10 (6.6) | 86 (21.6) | < 0.01 | 57 (20.2) | 29 (25.0) | 0.29 |
| Normocholesterolemia | 139 (92.1) | 256 (64.3) | < 0.01 | 192 (68.1) | 64 (55.2) | 0.02 |
| Hypocholesterolemia | 2 (1.3) | 56 (14.1) | < 0.01 | 33 (11.7) | 23 (19.8) | 0.04 |
| Hypertriglyceridemia | 20 (13.2) | 104 (26.1) | < 0.01 | 54 (19.1) | 50 (43.1) | < 0.01 |
| Normotriglyceridemia | 127 (84.1) | 285 (71.6) | < 0.01 | 220 (78.0) | 65 (56.0) | < 0.01 |
| Hypotriglyceridemia | 4 (2.7) | 9 (2.3) | 0.76 | 8 (2.8) | 1 (0.9) | 0.17 |
| Concurrent hypocholesterolemia and hypertriglyceridemia | 0.07 | < 0.01 | ||||
| Yes | 0 (0.0) | 10 (2.5) | 0 (0.0) | 10 (8.6) | ||
| No | 151 (100.0) | 388 (97.5) | 282 (100.0) | 106 (91.4) | ||
Data are given as number (percentage) of dogs in the group.
Results of univariate and multivariate logistic regression analyses for variables associated with sick dogs in Table 1 not surviving to hospital discharge.
| Variable | r | OR | 95% CI of OR | P value |
|---|---|---|---|---|
| Univariate logistic regression analysis | ||||
| Band neutrophil count | 0.32 | 1.37 | 1.09–1.74 | < 0.01 |
| CTR | –0.10 | 0.90 | 0.86–0.95 | < 0.01 |
| Gastroenteropathy | –1.02 | 0.36 | 0.15–0.88 | 0.03 |
| Heart rate | 0.01 | 1.01 | 1.00–1.01 | 0.02 |
| Hypertriglyceridemia | 1.16 | 3.20 | 2.00–5.13 | < 0.01 |
| Hypocholesterolemia | 0.62 | 1.87 | 1.04–3.34 | 0.04 |
| Neoplasia | 0.67 | 1.96 | 1.19–3.22 | < 0.01 |
| Normocholesterolemia | –0.55 | 0.58 | 0.37–0.90 | 0.02 |
| Normotriglyceridemia | –1.02 | 0.36 | 0.23–0.57 | < 0.01 |
| Renal disease | 2.03 | 7.64 | 1.52–38.41 | 0.01 |
| Multivariate logistic regression model | ||||
| Band neutrophil count | 0.32 | 1.37 | 1.09–1.74 | < 0.01 |
| Heart rate | 0.01 | 1.01 | 1.00–1.01 | 0.02 |
| Hypertriglyceridemia | 1.16 | 3.20 | 2.00–5.13 | < 0.01 |
| Hypocholesterolemia | 0.62 | 1.87 | 1.04–3.34 | 0.04 |
| Neoplasia | 0.67 | 1.96 | 1.19–3.22 | < 0.01 |
| Renal disease | 2.03 | 7.64 | 1.52–38.41 | 0.01 |
Glucocorticoid treatment
Administration of glucocorticoid treatment was noted in medical records of 102 of the 398 (25.6%) sick dogs, and when stratified by primary disease process included glucocorticoid administration to 55 of the 63 (87.3%) dogs with autoimmune disease, 19 of the 86 (22.1%) dogs with neoplasia, 11 of the 101 (10.9%) dogs with inflammatory disease, 8 of the 37 (21.6%) dogs with an endocrinopathy, 4 of the 42 (9.5%) dogs with a gastroenteropathy, 2 of the 8 (25.0%) dogs with renal disease, 2 of the 4 (50.0%) dogs with cardiovascular disease, and 1 of the 18 (5.6%) dogs with hepatobiliary disease. No sick dogs in the primary disease categories of trauma or coagulopathy were treated with glucocorticoids. Glucocorticoid administration was significantly (P < 0.01) more common in dogs with an autoimmune disorder. The proportions of dogs receiving glucocorticoids did not differ substantially between survivors (73/282 [25.9%]) and nonsurvivors (29/116 [25%]).
Nutritional support
Information on nutritional support was available in the medical records for 308 of the 398 (77.4%) sick dogs (survivors, n = 237/282 [84.0%]; nonsurvivors, 71/116 [61.2%]), with 231 of the 308 (75.0%) dogs receiving and 77 (25.0%) not receiving nutritional support. A significantly (P < 0.01) higher proportion of survivors (196/237 [82.7%]) received nutritional support, compared with nonsurvivors (35/71 [49.3%]).
Predictors of survival to hospital discharge
In the univariate logistic regression analysis, band neutrophil count (P < 0.01), heart rate (P = 0.02), hypocholesterolemia (P = 0.04), hypertriglyceridemia (P < 0.01), neoplasia (P < 0.01), and renal disease (P = 0.01) were associated with an increased probability of death, whereas gastroenteropathy (P = 0.03), high CTR (P < 0.01), normocholesterolemia (P = 0.02), and normotriglyceridemia (P < 0.01) were associated with a decreased probability of death (Table 4). Serum concentrations of cholesterol and triglycerides were not associated with survival in the univariate analysis. In the multivariate logistic regression model, band neutrophil count (P < 0.01), heart rate (P = 0.02), hypocholesterolemia (P = 0.04), hypertriglyceridemia (P < 0.01), neoplasia (P < 0.01), and renal disease (P = 0.01) remained significantly associated with an increased probability of death. Dyslipidemia factors associated with increased odds of sick dogs not surviving to hospital discharge were hypocholesterolemia (OR, 1.87; 95% CI, 1.04 to 3.34) and hypertriglyceridemia (OR, 3.20; 95% CI, 2.00 to 5.13).
Correlations between variables
With the Spearman rank-order correlation analysis, a strong, inverse correlation (r = −0.84; P < 0.01) was observed between serum triglycerides concentration and CTR. Weak, positive correlations were observed between serum triglycerides concentration and serum cholesterol concentration (r = 0.29; P < 0.01), WBC count (r = 0.21; P < 0.01), and band neutrophil count (r = 0.26; P < 0.01). Serum cholesterol concentration had a weak, positive correlation (r = 0.29; P < 0.01) with CTR, as did periodontal disease grade with serum triglycerides concentration (r = 0.35; P < 0.01), respiratory rate with body temperature (r = 0.20; P < 0.01), and WBC count with band neutrophil count (r = 0.36; P < 0.01). No other correlations were identified between variables.
Discussion
Results of the present study indicated that dyslipidemias were frequently present in sick dogs hospitalized in the ICU at the Iowa State University veterinary teaching hospital between January 1, 2012, and September 1, 2015. To various extents, lipid alterations were associated with disease severity and patient survival in sick dogs of the present study. Specifically, sick dogs hospitalized in the ICU were more frequently hyper- or hypocholesterolemic than were healthy dogs, and nonsurvivors were more likely to be hypocholesterolemic than were survivors. In the present study, median serum triglycerides concentration was higher in sick dogs versus healthy dogs, in sick dogs with SIRS versus without SIRS, and in nonsurvivors versus survivors. Although the exact numeric serum concentrations of cholesterol and triglycerides were not predictive of death, results indicated that sick dogs with either hypertriglyceridemia or hypocholesterolemia at initial evaluation were more likely to die before discharge from the hospital than were sick dogs without either of those conditions. Further, dogs with concurrent hypertriglyceridemia and hypocholesterolemia had poor outcomes in that all 10 dogs in the present study with this lipid profile at initial evaluation died during hospitalization.
Findings of the present study were similar to findings that dyslipoproteinemias and dyslipidemias are frequently documented in critically ill people.2,11 In particular, a pattern of low serum HDL, LDL, or serum cholesterol concentration, alone or in combination, is commonly observed in people with SIRS,2,10 sepsis,12,14,15 severe trauma,30,31 and MODS.23,32 Conversely, systemic inflammation may also promote increased serum concentrations of triglycerides.14,16,18–20 Dyslipidemia is associated with death in critically ill people, and hypocholesterolemia and hypertriglyceridemia at the time of initial examination are independent predictors of nonsurvival in people with sepsis.33,34 In an observational study35 of 85,000 people hospitalized in ICUs, patients with serum cholesterol concentration < 100 mg/dL were 10 times as likely to die and all patients with serum cholesterol concentration < 45 mg/dL died.35 Dyslipidemia was also independently associated with an increased risk of death within a 1-year period in elderly patients admitted to an ICU.36 A recent prospective study37 of hospitalized people with severe sepsis shows that patients who died had lower serum concentrations of cholesterol and HDL than did those that survived to day 3 of hospitalization. In addition, minimum serum cholesterol concentration and maximum serum total bilirubin concentration are the most prognostically useful serum biochemical analytes measured in human ICU patients with sepsis.38
Abnormalities in serum cholesterol and lipoprotein concentrations have been reported less frequently in hospitalized or ill dogs. Specifically, dogs with chronic kidney disease have dyslipoproteinemias characterized by a low HDL fraction and high LDL and VLDL fractions, compared with healthy control dogs.39 Dogs naturally infected with Babesia canis have altered lipid profiles at certain stages of illness in that infected dogs had lower HDL concentrations before treatment and higher serum cholesterol concentrations after treatment, compared with healthy controls.40 In that study,40 infected dogs were further categorized as meeting the criteria for SIRS or meeting the criteria for MODS. Dogs meeting the criteria for SIRS, but not the dogs meeting the criteria for MODS, had higher serum cholesterol concentration at day 2 of treatment for infection, compared with controls; however, no differences in serum concentrations of HDL were noted between the SIRS, MODS, and control groups.40 In another study,41 dogs with parvoviral enteritis had lower serum concentrations of cholesterol, HDL, and LDL, compared with controls, and nonsurvivors had lower serum concentrations of cholesterol and HDL, compared with survivors.
Curiously, in the present study, we observed no difference in median serum cholesterol concentration between sick and healthy dogs nor between survivors and nonsurvivors. Rather, hypocholesterolemia appeared predictive of nonsurvival and was more common in sick versus healthy dogs and in nonsurvivors versus survivors. This finding could have been related to the fact that cholesterol metabolism in dogs is different from that in many other species. In people, the enzyme CETP facilitates interlipoprotein lipid exchange in serum42 and transfers triglycerides from VLDL and chylomicrons to HDL and cholesteryl esters from HDL to VLDL and LDL. As such, CETP permits dynamic movement of cholesterol between lipoproteins. Dogs have markedly lower quantities of CETP, compared with many other animal species, and move cholesterol esters from HDL to LDL and VLDL at a relatively slower rate.43 Consequently, HDL may become saturated with cholesterol and function as a cholesterol depot in serum.44 This depot effect may offset decreased hepatic production of cholesterol associated with inflammatory disease and make serum cholesterol concentration more stable in dogs with critical illness. Only with very severe clinical disease might serum cholesterol concentration actually decrease, and this could possibly explain why in the present study hypocholesterolemia was observed more frequently in nonsurvivors versus survivors and was predictive of nonsurvival.
In addition to hypocholesterolemia, findings of the present study indicated that hypertriglyceridemia could be predictive of sick dogs not surviving to hospital discharge. Specifically, median serum triglycerides concentration was higher in sick versus healthy dogs, in dogs with versus without SIRS, and in nonsurvivors versus survivors. Although serum triglycerides concentration was not identified as a prognostic factor for survival in the multiple logistic regression model of the present study, hypertriglyceridemia was predictive of nonsurvival. These findings were similar to observations previously reported for dogs40,42 and horses.26 Hypertriglyceridemia is also considered a serious complication of hepatic disease in horses and is prognostic for survival with certain hepatopathies.45 Overall, results of the present study indicated that serum triglycerides concentration may be useful in assessing illness severity and predicting nonsurvival to hospital discharge in dogs; however, prospective studies are needed to further evaluate the clinical importance of these findings.
In the present study, the CTR was lower in sick versus healthy dogs, in dogs with versus without SIRS, and in nonsurvivors versus survivors. In the univariate analysis, the CTR was associated with survival to hospital discharge. A better predictor of nonsurvival in the present study, however, was the presence of concurrent hypocholesterolemia and hypertriglyceridemia. Although this was an uncommon finding, with only 10 of 398 (2.5%) sick dogs and none of the healthy dogs having this profile, all affected dogs died. However, the proportion of dogs with this profile of dual dyslipidemias was not significantly different between sick and healthy dogs, likely because of the profile's uncommon occurrence. Nonetheless, dogs with this lipid profile had higher odds (OR, 55.7; 95% CI, 3.2 to 959.6) of not surviving to hospital discharge in the present study. These observations suggested that concurrent hypocholesterolemia and hypertriglyceridemia was a marker of severe clinical disease and probable death. Similarly, concurrent hypocholesterolemia with hypertriglyceridemia is associated with nonsurvival in critically ill people.11,46
Results of the present study indicated that the observed prognostic potential of hypocholesterolemia and hypertriglyceridemia for nonsurvival to hospital discharge should be added to a growing list of prognostic biomarkers for dogs with critical illness, including those predictive for survival in dogs with SIRS or sepsis. For instance, a study25 evaluating the prognostic potential of routinely collected clinical and clinicopathologic features in 36 hospitalized dogs with SIRS shows that the APPLE score is the best predictor of survival. That study used a 5-variable model, termed the APPLEfast score, that considers serum concentrations of albumin, lactate, and glucose as well as platelet count and mentation score. An APPLEfast score > 27 had an OR of 1.7 for death, with a sensitivity of 80% and specificity of 90.4% for death in dogs.25 A separate study47 of 1,237 sick dogs with various disease processes shows that the hazard ratio is 1.9 for death in dogs with versus without a degenerative left shift leukogram on initial evaluation. The prognostic value of hypocholesterolemia (OR, 1.87; 95% CI, 1.04 to 3.34) and hypertriglyceridemia (OR, 3.20; 95% CI, 2.00 to 5.13) for sick dogs to not survive to hospital discharge in the present study appeared similar to the prognostic potentials of the APPLEfast score and degenerative left shift. Further, hypocholesterolemia and hypertriglyceridemia are also easily identified on a routine serum biochemical profile, making them readily available prognostic variables for veterinarians managing sick dogs hospitalized in an ICU.
Regarding relationships between clinical and clinicopathologic variables, one of the more interesting findings of the present study was that initial serum triglycerides concentration was positively correlated with WBC count and band neutrophil count. Although the correlation was weak, this observation could support a possible link between serum triglycerides concentration and systemic inflammation in dogs and warrants further investigation because WBC count and band neutrophil count generally increase with inflammatory conditions. Many of the nonlipid differences observed between healthy and sick dogs of the present study were likely a reflection of the study populations because the higher median WBC count and band neutrophil count in sick dogs was consistent with these dogs having inflammation. Sick dogs in the present study also had a higher median heart rate and lower median rectal temperature, compared with healthy dogs, findings that could occur with SIRS.27
Administration of glucocorticoids or nutritional support, alone or in combination, was common in sick dogs in the present study. However, these treatments would not have affected serum lipid measurements because blood samples were routinely collected during initial evaluation and before such treatments had begun, unless emergency treatment was warranted. Glucocorticoid treatment and nutritional support could have theoretically influenced patient outcome, although there was no difference in the proportions of survivors and nonsurvivors receiving glucocorticoid treatment, and glucocorticoid administration was not identified as a predictor of survival in the logistic regression models of the present study. Nutritional support, however, was more common in survivors versus nonsurvivors in the present study, and early nutritional support has been associated with decreased duration of hospitalization or increased probability of survival in dogs with a number of diseases, including septic peritonitis and pancreatitis.48,49 As such, the nutritional support received by some sick dogs in the present study may have been a confounding variable with regard to predictors of patient survival.
The retrospective nature of the present study was the main source of limitations in the investigation. In particular, the study could have been strengthened by the addition of a true control group, with matched controls having similar diet, diagnosis, body weight, body condition score, age, and breed that had undergone similar screening for occult comorbidities (eg, hyperadrenocorticism). Additionally, clinicopathologic analyses were limited to those obtained during initial evaluation, and convalescent serum biochemical profiles were not performed on all patients. As such, we were unable to determine whether any dogs corrected or developed dyslipidemias during hospitalization. In people, serial monitoring of lipids (serum concentrations of triglycerides, cholesterol, HDL, and LDL, alone or in combination) is useful for monitoring disease progression and predicting survival.1,11,17 Serum concentrations of HDL, LDL, and VLDL were not clinically measured at our institutions; therefore, we were unable to determine whether dyslipoproteinemias existed in dogs of the present study. Another limitation was that although treatments administered at our institution were unlikely to affect lipid concentrations because blood samples were collected prior to initiation of treatment, owners or referring veterinarians might have provided treatments before referral to our institution that could have affected lipid concentrations. For instance, drugs that are known to induce dyslipidemias include corticosteroids, loop diuretics, antiseizure medications, estrogen, cyclosporine, and β-adrenoceptor blockers.50 Another factor that could have affected lipid concentrations at initial evaluation was whether dogs had recently eaten; however, we were unable to determine retrospectively which dogs (healthy or sick) had food withheld before blood was collected.
Another limitation was that unassessed factors (eg, known prognostic factors [eg, APPLE score or whole blood lactate concentration] or the number of sick dogs that died of disease-related reasons after hospital discharge) could have been confounding variables in survival analyses of the present study. A prospective, observational study of sick dogs hospitalized in an ICU with case-matched controls and serial assessment of serum lipid components is warranted to fully assess impacts that these confounding factors may have on lipid concentrations in dogs.
In conclusion, we observed that dyslipidemias were present during initial evaluation of sick dogs hospitalized in the ICU of a veterinary teaching hospital. Hypocholesterolemia and hypertriglyceridemia were observed more commonly in sick versus healthy dogs and in nonsurvivors versus survivors. In multivariate logistic regression model analyses, the presence of either of these lipid derangements at initial evaluation was associated with dogs not surviving to hospital discharge. All 10 sick dogs with concurrent hypocholesterolemia and hypertriglyceridemia died before discharge from the hospital. Our findings indicated that dogs with a lipid profile of concurrent hypocholesterolemia and hypertriglyceridemia at initial evaluation for hospitalization have high odds of death and warrant relatively comprehensive clinical monitoring and treatment for their underlying diseases.
Acknowledgments
Supported in part by a 2015 Veterinary Student Summer Scholar Program stipend from the College of Veterinary Medicine, Iowa State University. The authors declare that there were no other conflicts of interest.
Presented in part in poster form at the 2015 Iowa State University Veterinary Student Research Symposium, Ames, Iowa, August 2015; and the 2015 Merial–National Institutes of Health National Veterinary Scholars Symposium, Davis, Calif, August 2015.
The authors thank Yaxuan Sun for statistical consultations.
ABBREVIATIONS
| APPLE | Acute patient physiologic and laboratory evaluation |
| CETP | Cholesteryl ester transfer protein |
| CI | Confidence interval |
| CTR | Cholesterol-to-triglycerides ratio |
| HDL | High-density lipoprotein |
| ICU | Intensive care unit |
| LDL | Low-density lipoprotein |
| MODS | Multiple organ dysfunction syndrome |
| SIRS | Systemic inflammatory response syndrome |
| TNF-α | Tumor necrosis factor-α |
| VLDL | Very-low-density lipoprotein |
Footnotes
Advia 2120, Siemens Corp, Washington, DC.
Vitros, Ortho Clinical Diagnostics, Raritan, NJ.
Graphpad Prism, GraphPad Software Inc, La Jolla, Calif.
NCSS10, NCSS LLC, Kaysville, Utah.
MedCalc, MedCalc Software, Ostend, Belgium.
References
1. Green P, Theilla M, Singer P. Lipid metabolism in critical illness. Curr Opin Clin Nutr Metab Care 2016;19:111–115.
2. Chiarla C, Giovannini I, Giuliante F, et al. Hypocholesterolemia in critical illness. Crit Care Med 2009;37:2681–2682.
3. Vyroubal P, Chiarla C, Giovannini I, et al. Hypocholesterolemia in clinically serious conditions—review. Biomed Pap Med Fac Univ Palacky Olomouc Czech Repub 2008;152:181–189.
4. Levine DM, Parker TS, Donnelly TM, et al. In vivo protection against endotoxin by plasma high density lipoprotein. Proc Natl Acad Sci U S A 1993;90:12040–12044.
5. Netea MG, Demacker PN, Kullberg BJ, et al. Low-density lipoprotein receptor-deficient mice are protected against lethal endotoxemia and severe gram-negative infections. J Clin Invest 1996;97:1366–1372.
6. Pajkrt D, Doran JE, Koster F, et al. Antiinflammatory effects of reconstituted high-density lipoprotein during human endotoxemia. J Exp Med 1996;184:1601–1608.
7. Nogueira AC, Kawabata V, Biselli P, et al. Changes in plasma free fatty acid levels in septic patients are associated with cardiac damage and reduction in heart rate variability. Shock 2008;29:342–348.
8. Murad MH, Hazem A, Coto-Yglesias F, et al. The association of hypertriglyceridemia with cardiovascular events and pancreatitis: a systematic review and meta-analysis. BMC Endocr Disord 2012;12:2.
9. Carr RA, Rejowski BJ, Cote GA, et al. Systematic review of hypertriglyceridemia-induced acute pancreatitis: a more virulent etiology? Pancreatology 2016;16:469–476.
10. Bonville DA, Parker TS, Levine DM, et al. The relationships of hypocholesterolemia to cytokine concentrations and mortality in critically ill patients with systemic inflammatory response syndrome. Surg Infect (Larchmt) 2004;5:39–49.
11. Chiarla C, Giovannini I, Giuliante F, et al. Severe hypocholesterolemia in surgical patients, sepsis, and critical illness. J Crit Care 2010;25:361.e361.e7–361.e312.
12. Chiarla C, Giovannini I, Siegel JH. The relationship between plasma cholesterol, amino acids and acute phase proteins in sepsis. Amino Acids 2004;27:97–100.
13. Giovannini I, Boldrini G, Chiarla C, et al. Pathophysiologic correlates of hypocholesterolemia in critically ill surgical patients. Intensive Care Med 1999;25:748–751.
14. Rival T, Cinq-Frais C, Silva-Sifontes S, et al. Alteration of plasma phospholipid fatty acid profile in patients with septic shock. Biochimie 2013;95:2177–2181.
15. Wilson RF, Barletta JF, Tyburski JG. Hypocholesterolemia in sepsis and critically ill or injured patients. Crit Care 2003;7:413–414.
16. Kitchens RL, Thompson PA, Munford RS, et al. Acute inflammation and infection maintain circulating phospholipid levels and enhance lipopolysaccharide binding to plasma lipoproteins. J Lipid Res 2003;44:2339–2348.
17. Marik PE. Dyslipidemia in the critically ill. Crit Care Clin 2006;22:151–159.
18. Cetinkaya A, Erden A, Avci D, et al. Is hypertriglyceridemia a prognostic factor in sepsis? Ther Clin Risk Manag 2014;10:147–150.
19. Ilias I, Vassiliadi DA, Theodorakopoulou M, et al. Adipose tissue lipolysis and circulating lipids in acute and subacute critical illness: effects of shock and treatment. J Crit Care 2014;29:e5–e9.
20. Wendel M, Paul R, Heller AR. Lipoproteins in inflammation and sepsis. II. Clinical aspects. Intensive Care Med 2007;33:25–35.
21. Green A, Dobias SB, Walters DJ, et al. Tumor necrosis factor increases the rate of lipolysis in primary cultures of adipocytes without altering levels of hormone-sensitive lipase. Endocrinology 1994;134:2581–2588.
22. Park DW, Kwak DS, Park YY, et al. Impact of serial measurements of lysophosphatidylcholine on 28-day mortality prediction in patients admitted to the intensive care unit with severe sepsis or septic shock. J Crit Care 2014;29:882.e5–882.e11.
23. Dunham CM, Fealk MH, Sever WE. Following severe injury, hypocholesterolemia improves with convalescence but persists with organ failure or onset of infection. Crit Care 2003;7:R145–R153.
24. Elliott DC, Wiles CE III. Low lipid concentrations in critical illness: hypocholesterolemia among trauma patients. Crit Care Med 1997;25:1437–1439.
25. Giunti M, Troia R, Bergamini PF, et al. Prospective evaluation of the acute patient physiologic and laboratory evaluation score and an extended clinicopathological profile in dogs with systemic inflammatory response syndrome. J Vet Emerg Crit Care (San Antonio) 2015;25:226–233.
26. Dunkel B, McKenzie HC III. Severe hypertriglyceridaemia in clinically ill horses: diagnosis, treatment and outcome. Equine Vet J 2003;35:590–595.
27. Friedrichs KR, Harr KE, Freeman KP, et al. ASVCP reference interval guidelines: determination of de novo reference intervals in veterinary species and other related topics. Vet Clin Pathol 2012;41:441–453.
28. Bauer N, Nakagawa J, Dunker C, et al. Evaluation of the automated hematology analyzer Sysmex XT-2000iV™ compared to the ADVIA® 2120 for its use in dogs, cats, and horses: part I—precision, linearity, and accuracy of complete blood cell count. J Vet Diagn Invest 2011;23:1168–1180.
29. Bauer N, Nakagawa J, Dunker C, et al. Evaluation of the automated hematology analyzer Sysmex XT-2000iV ™ compared to the ADVIA ® 2120 for its use in dogs, cats, and horses. Part II: accuracy of leukocyte differential and reticulocyte count, impact of anticoagulant and sample aging. J Vet Diagn Invest 2012;24:74–89.
30. Gordon BR, Parker TS, Levine DM, et al. Relationship of hypolipidemia to cytokine concentrations and outcomes in critically ill surgical patients. Crit Care Med 2001;29:1563–1568.
31. Dunham CM, Chirichella TJ. Attenuated hypocholesterolemia following severe trauma signals risk for late ventilator-associated pneumonia, ventilator dependency, and death: a retrospective study of consecutive patients. Lipids Health Dis 2011;10:42.
32. Peng Y-S, Chen Y-C, Tian Y-C, et al. Serum levels of apolipo-protein A-I and high-density lipoprotein can predict organ failure in acute pancreatitis. Crit Care 2015;19:88.
33. Lee SH, Park MS, Park BH, et al. Prognostic implications of serum lipid metabolism over time during sepsis. Biomed Res Int 2015;2015:789298.
34. Das S, Bhargava S, Manocha A, et al. The prognostic value of hypocholesterolemia in sepsis. Asian J Pharm Biol Res 2011;1:41–46.
35. Windler E, Ewers-Grabow U, Thiery J, et al. The prognostic value of hypocholesterolemia in hospitalized patients. Clin Investig 1994;72:939–943.
36. Puchades R, González B, Contreras M, et al. Cardiovascular profile in critically ill elderly medical patients: prevalence, mortality and length of stay. Eur J Intern Med 2015;26:49–55.
37. Lekkou A, Mouzaki A, Siagris D, et al. Serum lipid profile, cytokine production, and clinical outcome in patients with severe sepsis. J Crit Care 2014;29:723–727.
38. Behling-Kelly E. Serum lipoprotein changes in dogs with renal disease. J Vet Intern Med 2014;28:1692–1698.
39. Yamano S, Shimizu K, Ogura H, et al. Low total cholesterol and high total bilirubin are associated with prognosis in patients with prolonged sepsis. J Crit Care 2016;31:36–40.
40. Mrljak V, Kučer N, Kuleš J, et al. Serum concentrations of eicosanoids and lipids in dogs naturally infected with Babesia canis. Vet Parasitol 2014;201:24–30.
41. Yilmaz Z, Senturk S. Characterisation of lipid profiles in dogs with parvoviral enteritis. J Small Anim Pract 2007;48:643–650.
42. Xenoulis PG, Steiner JM. Lipid metabolism and hyperlipidemia in dogs. Vet J 2010;183:12–21.
43. Tsutsumi K, Hagi A, Inoue Y. The relationship between plasma high density lipoprotein cholesterol levels and cholesteryl ester transfer protein activity in six species of healthy experimental animals. Biol Pharm Bull 2001;24:579–581.
44. Bauer JE. Lipoprotein-mediated transport of dietary and synthesized lipids and lipid abnormalities of dogs and cats. J Am Vet Med Assoc 2004;224:668–675.
45. Radcliffe RM, Buchanan BR, Cook VL, et al. The clinical value of whole blood point-of-care biomarkers in large animal emergency and critical care medicine. J Vet Emerg Crit Care (San Antonio) 2015;25:138–151.
46. Chiarla C, Giovannini I, Siegel JH. Hypotransferrinemia and changes in plasma lipid and metabolic patterns in sepsis. Amino Acids 2009;36:327–331.
47. Burton AG, Harris LA, Owens SD, et al. The prognostic utility of degenerative left shifts in dogs. J Vet Intern Med 2013;27:1517–1522.
48. Liu DT, Brown DC, Silverstein DC. Early nutritional support is associated with decreased length of hospitalization in dogs with septic peritonitis: a retrospective study of 45 cases (2000–2009). J Vet Emerg Crit Care (San Antonio) 2012;22:453–459.
49. Jensen KB, Chan DL. Nutritional management of acute pancreatitis in dogs and cats. J Vet Emerg Crit Care (San Antonio) 2014;24:240–250.
50. Yuan G, Al-Shali KZ, Hegele RA. Hypertriglyceridemia: its etiology, effects and treatment. CMAJ 2007;176:1113–1120.
