Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Serum vitamin D concentrations in hospitalized critically ill dogs

  • Jared A. Jaffey ,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Writing – original draft, Writing – review & editing

    jaffeyj@missouri.edu

    Affiliation Department of Veterinary Medicine and Surgery, Veterinary Health Center, University of Missouri, Columbia, Missouri, United States of America

  • Robert C. Backus,

    Roles Data curation, Investigation, Resources, Writing – review & editing

    Affiliation Department of Veterinary Medicine and Surgery, Veterinary Health Center, University of Missouri, Columbia, Missouri, United States of America

  • Kaylyn M. McDaniel,

    Roles Data curation, Investigation, Writing – review & editing

    Affiliation Department of Veterinary Medicine and Surgery, Veterinary Health Center, University of Missouri, Columbia, Missouri, United States of America

  • Amy E. DeClue

    Roles Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Supervision, Writing – review & editing

    Affiliation Department of Veterinary Medicine and Surgery, Veterinary Health Center, University of Missouri, Columbia, Missouri, United States of America

Serum vitamin D concentrations in hospitalized critically ill dogs

  • Jared A. Jaffey, 
  • Robert C. Backus, 
  • Kaylyn M. McDaniel, 
  • Amy E. DeClue
PLOS
x

Abstract

Hypovitaminosis D has been extensively documented in critically ill humans. However, whether or not critically ill dogs have alterations in vitamin D concentrations remains unconfirmed. The primary aims of our study were to compare serum 25-hydroxycholecalciferol [25(OH)D] concentrations in critically ill dogs with healthy control dogs, determine the prognostic utility of serum 25(OH)D concentration as a biomarker in critically ill dogs, and to assess if serum 25(OH)D concentrations in critically ill dogs are associated with length of stay in the intensive care unit or illness severity. Serum concentrations of 25(OH)D together with a range of other clinical, biochemical, and hematological parameters, were measured in 99 dogs within 24 hours of admission to the Intensive Care Unit (ICU). Critically ill dogs (P = 0.001) and dogs with sepsis (P = 0.002) had significantly lower serum 25(OH)D concentrations compared to healthy control dogs. In addition, serum 25(OH)D concentration was an independent predictor of in-hospital and 30 day survival. Using a cut-off of 33 ng/mL, serum 25(OH)D concentrations had excellent sensitivity (0.94; 95% CI, 0.71–1.00), but poor specificity (0.41; 95% CI, 0.31–0.53) for detection of survival. Serum 25(OH)D concentrations were inversely associated with acute patient physiologic and laboratory evaluation (APPLE) fast score but were not associated with ICU length of stay. Hospitalized dogs with critical illness have decreased serum 25(OH)D concentrations compared to healthy dogs and can be used to predict survival in this cohort.

Introduction

Hypovitaminosis D has been extensively investigated in critically ill people. The prevalence of hypovitaminosis D in critically ill people is as high as 82% [1] and is associated with increased mortality [19], length of hospitalization [110], sepsis [8,9,11,12], and illness severity [13]. While vitamin D is classically thought of in the context of calcium, phosphorous, and skeletal homeostasis [1416], identification of vitamin D receptors and 1-alpha hydroxylase in tissues outside of the kidney suggests more extensive effects than previously thought [16]. The pleiotropic effects of vitamin D extend to regulation of cell proliferation, angiogenesis, blood pressure, as well as the innate and adaptive immune systems [17].

The importance of hypovitaminosis D in the context of critical illness may relate to its immunomodulatory properties. Dogs with naturally developing critical illness might be an ideal pre-clinical model for studying hypovitaminosis D during critical illness in people since dogs have similar environmental exposures, co-morbidities and undergo similar medical and surgical interventions. Further, vitamin D augments the innate immune response in people by increasing phagocytosis [1820] and induction of antimicrobial peptides [2125] as well as decreases leukocyte production of pro-inflammatory cytokines such as tumor necrosis factor (TNF)-α and interleukin (IL)-6 while increasing production of the anti-inflammatory cytokine IL-10 [2629]. Likewise, vitamin D blunts leukocyte production of TNF without compromising toll-like receptor 4 expression or phagocytosis in neutrophils and monocytes in dogs [30]. In addition, vitamin D increases lipopolysaccharide stimulated leukocyte production of IL-10 in endotoxin primed whole blood in dogs, in vitro (Jaffey et al., unpublished work) further suggesting the dog has similarities to people. Dogs are not able to synthesize adequate amounts of vitamin D in response to ultraviolet light from the sun, and thus are dependent on dietary intake to meet their vitamin D needs [31]. Therefore, studies assessing the effects of vitamin D supplementation in dogs will not be limited by confounding variables (e.g., latitude [32], seasonality [33], age [33], ethnicity [34], obesity [35], or gender [36]) inherent to similar studies in humans.

The primary aims of our study were to determine if dogs, like people, have decreased serum concentrations of vitamin D during critical illness and to determine if decreased serum concentrations of vitamin D were associated with illness severity and outcome. We hypothesized that critically ill dogs and dogs with sepsis would have significantly lower serum 25-hydroxycholecalciferol [25(OH)D] concentrations than healthy control dogs. Further, we hypothesized that critically ill dogs with higher serum 25(OH)D concentrations would be more likely to survive to discharge and be alive 30 days after discharge. Secondary aims were to compare serum 25(OH)D concentrations in dogs with sepsis and healthy control dogs as well as to compare serum 25(OH)D concentrations in dogs with sepsis that survived to discharge with those that did not survive. We hypothesized that dogs with sepsis would have lower serum 25(OH)D concentrations than healthy control dogs and that dogs with sepsis that survived to discharge would have greater serum 25(OH)D concentrations than dogs that did not survive.

Materials and methods

Dogs admitted to the University of Missouri Veterinary Health Center Intensive Care Unit (ICU) between October 2016 and January 2017 were eligible for inclusion in this prospective observational study. Client consent was obtained. This study was conducted in accordance with guidelines for clinical studies and was approved by the University of Missouri Animal Care and Use Committee (protocol 7334). Exclusion criteria dogs included, pregnancy, lactating, hypercalcemia of malignancy, hyperparathyroidism, hypoparathyroidism, known history of chronic kidney disease, supplementation with vitamin D or calcium, and if a dog had already been admitted to the ICU during the study period. Furthermore, dogs were excluded if blood glucose, albumin, lactate, platelet count, and mentation score were not obtained within 24 hours of ICU admission in order to determine the acute patient physiologic and laboratory evaluation (APPLE) fast score for each dog. A second population of dogs were also enrolled as a control population. The control dogs were determined to be healthy based on history and physical examination. Control dogs could not have any illnesses, vaccination or administration of medications, with the exception of routine parasitic prevention, within 30 days of enrollment.

The APPLE fast scoring system was used to assess illness severity. The APPLE fast score for each dog was calculated using blood glucose, albumin, lactate, platelet count, and mentation score obtained within 24 hours of admission to the ICU (Table 1) [37]. Criteria for mentation scoring are summarized in (Table 2). All mentation scores were performed by a single individual (JAJ). Dogs were considered to have sepsis if they met 2 or more of the criteria for systemic inflammatory response syndrome (heart rate ≥160 bpm, respiratory rate ≥ 40 bpm, temperature ≤ 37.8 or ≥ 39.7°C, and white blood cell count ≤ 4 or ≥ 12 x 103/μL or ≥ 10% band neutrophils) [38] and had either cytological or microbiological confirmation of bacterial infection or related to intestinal bacterial translocation.

thumbnail
Table 1. APPLE fast score.

Calculated by adding the values in the upper left corner of each cell for the 5 parameters listed, with a maximum score of 50 [37]. Adapted from Hayes et al., 2010.

https://doi.org/10.1371/journal.pone.0194062.t001

Sample collection

Medical records were reviewed for each dog enrolled. The age, sex and breed were recorded for each dog. The following clinical information was extracted for each dog within 24 hours of ICU admission: white blood cell count, platelet count, albumin, glucose, lactate, and serum 25(OH)D concentrations. In addition, the duration of hospitalization, reason for ICU admission, and survival were recorded. Survival data was obtained from clinical records or follow up telephone calls to clients for each dog for survival to discharge and day 30 after ICU discharge. All hematology and biochemical parameters were measured using the same analyzers. Blood lactate concentrations were measured either on a blood gas analyzer or a handheld device previously validated in dogs [39]. Blood samples were collected into evacuated glass tubes and were centrifuged (1,500 x g, 7 minutes) and serum harvested within 1 hour of sample collection. The serum was placed in an airtight, freezer-resistant plastic tube and stored for a maximum of 4 months at -80°C for batch analysis of serum 25(OH)D concentrations [40]. Serum concentrations of 25(OH)D were determined in thawed serum using a modification of a HPLC method previously reported [41]. Tubes were coded so that the identity of the sample was only known to the investigator and not to the laboratory measuring serum 25(OH)D. In addition, serum 25(OH)D concentrations were evaluated after all dogs had been discharged from the ICU, died, or were euthanized. Therefore, serum 25(OH)D results did not impact the clinical management of the dogs.

Statistical analyses

Statistical analysis was performed by commercial software. A Shapiro-Wilk test was used to assess normality. The data with a non-normal distribution were assessed using Mann-Whitney U test and the median, Q1, Q3, and range given. Categorical variables are presented as proportions. The length of stay was dichotomized based on its median value (3 days) into 2 categories: short stay (≤ 3 days) and long stay (> 3 days). Dogs that died or were euthanized in the ICU were not included in the assessment of length of stay. A receiver-operating characteristic (ROC) curves were used to determine the area under the curve (AUC) and select the optimum cut-off value for detection of death that maximized the Youden’s J statistic for sensitivity and specificity reporting. Multivariate logistic regression analyses were performed for the whole study population, choosing ICU mortality and 30 day post ICU discharge mortality as the dependent variables. Independent variables for the multivariate logistic model were APPLE fast score, serum 25(OH)D concentration, and the presence of sepsis. The correlation between serum 25(OH)D concentration with length of stay and APPLE fast scores were tested by the Spearman test. The level of significance was set at P < 0.05.

Results

Two-hundred and sixteen dogs fulfilled the inclusions criteria. One-hundred and seventeen dogs were excluded from analysis for the following reasons, lack of hematologic or biochemical data needed to calculate an APPLE fast score (n = 58), owner objection (n = 30), hospitalization for > 24 hours before sample acquisition (n = 11), chronic kidney disease (n = 4), malignancy associated hypercalcemia (n = 4), euthanized before sample acquisition (n = 3), clinician objection (n = 3), primary hyperparathyroidism (n = 2), and 1 each for pregnancy and a dog that had already been enrolled from a previous ICU admission.

Animal population

Ninety-nine dogs were enrolled. There were 84 purebred dogs and 15 mixed breed dogs included. Breeds were Labrador Retriever (n = 12), Golden Retriever (4), Miniature Dachshund (4), German Shepherd Dog (3), Border Collie (3), Miniature Schnauzer (3), Beagle (2), American Staffordshire Terrier (2), Maltese (2), Jack Russell Terrier (2), Boxer (2), French Bulldog (2), and 1 each of Akita, Australian Shepherd, Bloodhound, Brussels Griffon, Bullmastiff, Chihuahua, Dogue De Bordeaux, Great Dane, Mastiff, Miniature Poodle. The median age at presentation was 7.6 years (Q1, Q3, range; 4.0, 10.6, 0.2–17.5). The median weight was 17.0 kg (7.7, 29.8, 1.6–58.7). There were 48 neutered males, 42 spayed females, 5 intact males, and 4 intact females.

Dogs were admitted to the ICU for the following reasons postoperative supportive care (n = 22); neoplasia (15), septic peritonitis (6), vomiting and diarrhea (5), trauma (5), respiratory distress (4), diabetic ketoacidosis with pancreatitis (3), hypoadrenocorticism (2), gastric dilatation (2), immune-mediated hemolytic anemia (2), seizures of unknown etiology (2), pancreatitis (2), vomiting (2), bacterial pneumonia (3), aspiration pneumonia (3), acute kidney failure (2), and 1 each of subcutaneous abscess, anaphylaxis, cholangiohepatitis, nephrotic syndrome, myxedema coma, diskospondylitis, diabetic ketoacidosis, fever of unknown origin, laryngeal paralysis, unclassified type of pneumonia, lung lobe torsion, lymphangiectasia, meningoencephalitis of unknown origin, pancytopenia and myelodysplasia, protein losing enteropathy of unknown etiology, steroid responsive pneumonia (suspected bronchiolitis obliterans with organizing pneumonia), regurgitation, suspected xylitol toxicity.

A total of 16/99 dogs were diagnosed with sepsis. The source of sepsis was classified as abdominal (8), respiratory (2), unknown (3), and 1 each of following central nervous system, musculoskeletal, and dermatologic/subcutaneous. Infectious agents identified on either cytology, culture, or histopathology included non-specific bacilli (2), rods (2), cocci (1), Staphylococcus hemolyticus (2), Corynebacterium species (1), Streptococcus canis (beta hemolytic Streptococcus)(1), Serratia marcescens (1), Enterobacter cloacae (1), and Enterococcus faecium (1).

Seventeen of the 99 critically ill dogs died (4) or were euthanized (13) for an in hospital mortality rate of 17.1%. The survival status for dogs 30 days after hospital discharge was available in 93/99 dogs. Twenty-three critically ill dogs died (4) or were euthanized (19) 30 days after discharge for an overall mortality rate of 24.7%. Eight of the 16 dogs with sepsis died or were euthanized in the hospital.

Seventeen healthy control dogs were included. There were 13 purebred dogs and 4 mixed breed dogs. Breeds were Miniature Dachshund (3), Pit-bull Terrier (2), and 1 each of Miniature Poodle, Brittany Spaniel, Catahoula, Labrador Retriever, Golden Retriever, German Shepherd Dog, Pomeranian, Great Dane. The median age was 6.0 years (Q1, Q3, range; 2.5, 9.0, 0.6–12.0). There were 9 spayed females, 6 neutered males, and 2 intact males.

Serum 25(OH)D concentration and survival

Healthy control dogs had significantly greater serum 25(OH)D concentrations than critically ill dogs (Fig 1). Dogs that survived to discharge had significantly greater serum 25(OH)D concentrations than dogs that died or were euthanized (Fig 2A). Furthermore, dogs that were alive 30 days after discharge had significantly greater serum 25(OH)D concentrations than dogs that were not alive (Fig 2B).

thumbnail
Fig 1. Box and whiskers plot comparing serum 25(OH)D concentrations in critically ill dogs and healthy control dogs.

The boxes represent the 25th and 75th quartiles with the horizontal line representing the median. The whiskers represent the range of the data. The black circles represent results for individual dogs. Healthy control dogs (n = 17) had significantly greater serum 25(OH)D concentration than critically ill dogs (n = 99).

https://doi.org/10.1371/journal.pone.0194062.g001

thumbnail
Fig 2. Box and whiskers plot comparing serum 25(OH)D concentrations in survivor and non-survivor critically ill dogs.

The boxes represent the 25th and 75th quartiles with the horizontal line representing the median. The whiskers represent the range of the data. The black circles represent results for individual dogs. (A) Dogs that survived (n = 82) to discharge had significantly greater serum 25(OH)D concentration than non-survivors (n = 17). (B) Dogs that were alive (n = 23) 30 days after discharge had significantly greater serum 25(OH)D concentrations than non-survivors (n = 70).

https://doi.org/10.1371/journal.pone.0194062.g002

Next, serum 25(OH)D was evaluated in a multivariate logistic regression model to determine if it was an independent predictor of mortality. Diagnosis of sepsis and APPLE fast scores were included in the model as covariates because of their potential association with serum 25(OH)D and mortality. The logistic regression analysis for ICU mortality indicated that only serum 25(OH)D concentrations and sepsis remained significant predictors of survival (Table 3). Importantly, this result indicates that for every 1 ng/ml increase in serum 25(OH)D concentration, the risk for mortality decreased by 6.0%. Serum 25(OH)D concentration was the only independent predictor for survival 30 days after discharge (Table 4). Moreover, for every 1 ng/ml increase in serum 25(OH)D concentration, the risk for mortality 30 days after discharge decreased by 5.0%.

thumbnail
Table 3. Multivariate logistic regression model for survival in the intensive care unit.

https://doi.org/10.1371/journal.pone.0194062.t003

thumbnail
Table 4. Multivariate logistic regression model for survival 30 days after discharge.

https://doi.org/10.1371/journal.pone.0194062.t004

The area under the ROC curve for serum 25(OH)D concentrations in relation to ICU survival was 0.70 (95% CI, 0.57–0.83). The optimal cutoff was > 33 ng/ml, with a sensitivity of 0.94 (95% CI, 0.71–1.00) and a specificity of 0.41 (95% CI, 0.31–0.53) for the detection of survival (Fig 3; Table 5). The area under the ROC curve for serum 25(OH)D concentrations in relation to 30 days post ICU discharge survival was 0.72 (95% CI, 0.60–0.84). The optimal cutoff was >30 ng/ml, with a sensitivity of 0.87 (95% CI, 0.66–0.97) and a specificity of 0.53 (95% CI, 0.41–0.65) for detection of survival (Fig 4; Table 6).

thumbnail
Fig 3. Receiver-operating characteristic (ROC) curve and dot plot comparing the diagnostic sensitivity and 1-specificity of serum 25(OH)D concentration for determining survival in the intensive care unit.

(A) The diagonal red line represents the reference line (B) The horizontal red line represents the optimal serum 25(OH)D concentration cutoff of 33 ng/ml.

https://doi.org/10.1371/journal.pone.0194062.g003

thumbnail
Fig 4. Receiver-operating characteristic (ROC) curve and dot plot comparing the diagnostic sensitivity and 1-specificity of serum 25(OH)D concentration for determining survival 30 days after discharge.

(A) The diagonal red line represents the reference line. (B) The horizontal red line represents the optimal serum 25(OH)D concentration cutoff of 30 ng/ml.

https://doi.org/10.1371/journal.pone.0194062.g004

thumbnail
Table 5. Sensitivity and specificity of serum 25(OH)D concentrations for detection of survival in the ICU using different cutoff values.

https://doi.org/10.1371/journal.pone.0194062.t005

thumbnail
Table 6. Cutoff serum 25(OH)D concentrations and resulting sensitivities and specificities for detection of survival 30 days after discharge from the intensive care unit.

https://doi.org/10.1371/journal.pone.0194062.t006

Serum 25(OH)D concentration and ICU length of stay

The median length of stay for critically ill dogs in the ICU was 3 days (Q1, Q3, range; 3.0, 5.0; 1.0–15). There was not a significant difference in serum 25(OH)D concentrations in dogs that were hospitalized for ≤ 3 days compared to > 3 days (Fig 5). Furthermore, no significant correlation was found between serum 25(OH)D concentration and ICU length of stay when length of stay was analyzed as a continuous variable (n = 82, rs = 0.03, P = 0.79).

thumbnail
Fig 5. Box and whiskers plot comparing serum 25(OH)D concentrations in dogs that stayed ≤ 3 days in the intensive care unit to those that stayed > 3 days.

The boxes represent the 25th and 75th quartiles with the horizontal line representing the median. The whiskers represent the range of the data. There was not a significant difference in serum 25(OH)D concentrations in dogs that stayed ≤ 3 days (n = 42) and those that stayed > 3 days (n = 40, P = 0.14).

https://doi.org/10.1371/journal.pone.0194062.g005

Serum 25(OH)D concentration and illness severity

The evaluation of the linear relationship between serum 25(OH)D concentrations and illness severity defined by APPLE fast scores resulted in an inverse, significant correlation (Fig 6).

thumbnail
Fig 6. Correlation between serum 25(OH)D concentration and APPLE fast score.

Serum 25(OH)D concentrations had a weak, inverse, yet significant correlation with APPLE fast scores.

https://doi.org/10.1371/journal.pone.0194062.g006

Serum 25(OH)D concentration and sepsis

Dogs with sepsis had significantly lower serum 25(OH)D concentrations than healthy control dogs (Fig 7). There was not a significant difference (P = 0.67) in serum 25(OH)D concentration in dogs with sepsis that survived to discharge (n = 8 dogs, 25 ng/ml, 17, 37; 17–69) compared to those that died or were euthanized (n = 9 dogs, 22 ng/ml, 15, 29; 10–55).

thumbnail
Fig 7. Box and whiskers plot comparing serum 25(OH)D concentrations in dogs with sepsis and healthy control dogs.

The boxes represent the 25th and 75th quartiles with the horizontal line representing the median. The whiskers represent the range of the data. The black circles represent results for individual dogs. Healthy control dogs (n = 17) had significantly greater serum 25(OH)D concentration than dogs with sepsis (n = 17).

https://doi.org/10.1371/journal.pone.0194062.g007

Discussion

In our investigation, we found that critically ill dogs had significantly lower serum 25(OH)D concentrations than healthy control dogs. Serum 25(OH)D concentration was an independent predictor of survival for dogs in the ICU as well as at 30 days after discharge. In addition, a serum 25(OH)D concentration cutoff of >33 ng/ml yielded a sensitivity of 94% and a specificity of 41% for predicting survival of critically ill dogs in our study. Serum 25(OH)D concentration had a significant correlation with illness severity but was not associated with length of stay in the ICU. Dogs with sepsis had significantly lower serum 25(OH) D concentrations than healthy control dogs but there was no difference in serum 25(OH)D concentration in dogs with sepsis that survived to discharge and those that died or were euthanized.

We found that serum 25(OH)D concentration was lower in critically ill dogs than in healthy control dogs. Furthermore, serum 25(OH)D was an independent predictor of survival in the ICU as well as at 30 days after discharge. This relationship is commonly found in critically ill people [19]. The pleiotropic effects of vitamin D in immunity, endothelial and mucosal functions as well as glucose and calcium homeostasis, may be fundamental to common illnesses encountered by critically ill dogs as they are in humans [42, 43]. More specifically, the role of vitamin D as a potent modulator of the immune system may best explain the increased risk of mortality in critically ill dogs and people. Vitamin D increases the innate immune response by enhancing phagocytosis [1820] and induction of the antimicrobial peptides, cathelicidin and β-defensins [2125]. Furthermore, immunologic production of calcitriol, the hormonally active form of vitamin D, down-regulated leukocyte production of pro-inflammatory cytokines such as IL-1, IL-2, TNF-α, and IL-6 while enhancing production of the anti-inflammatory cytokine IL-10 in humans [2629]. Similarly, calcitriol blunts leukocyte production of TNF in dogs [30] and increases lipopolysaccharide stimulated leukocyte production of IL-10 in endotoxin primed whole blood in dogs, in vitro (Jaffey et al., unpublished work). Therefore, it is reasonable to postulate that during critical illness, the decreased circulating 25(OH)D reservoir needed for paracrine production of calcitriol is diminished resulting in compromised innate immune function and an exaggerated pro-inflammatory milieu.

The area under the ROC curve for serum 25(OH)D concentration in relation to ICU and 30 day survival was 0.70 and 0.72, respectively. The optimal cutoffs for ICU and 30 day survival were >33 ng/ml and >30 ng/ml yielding sensitivities and specificities for survival of 95%, 87% and 41%, 53%, respectively. These results indicate that serum 25(OH)D concentrations obtained within 24 hours of ICU admission has excellent sensitivity to detect survival in critically ill dogs. In contrast to its excellent sensitivity, serum 25(OH)D concentrations had a poor specificity for survival. In fact, only 1 of the 17 dogs that died or was euthanized in the ICU had a serum 25(OH)D concentration above 33 ng/ml. The only dog with a serum 25(OH)D concentration greater than the cutoff that died had acute fulminant liver failure. In people, liver dysfunction results in increased serum 25(OH)D concentrations [1, 44]. A conclusion of a recent meta-analysis was that vitamin D supplementation reduced mortality in critically ill people [45]. Additional studies are needed to assess if supplementation of vitamin D is associated with a survival benefit in critically ill dogs. Until those studies are available it is not recommended to apply our results in a clinical setting.

Length of stay in the ICU is associated with serum 25(OH)D concentrations in critically ill humans [1, 10]. This relationship makes sense because serum 25(OH)D concentration correlates well with illness severity in people. It stands to reason that more time and support is necessary to recover from a profound illness. Similarly, our study found that serum 25(OH)D concentration had a significant correlation with illness severity. However, there was not a significant correlation between serum 25(OH)D concentration and length of stay in our study. In addition, there was not a significant difference in serum 25(OH)D concentration when length of stay was dichotomized around the median (3 days) into 2 groups and compared. The exclusion of dogs that died or were euthanized from the length of stay analysis may have impacted our results.

Dogs with sepsis in our study had significantly lower serum 25(OH)D concentrations than did healthy control dogs. However, there was not a significant difference in serum 25(OH)D concentration in dogs with sepsis that survived to discharge and those that died or were euthanized. Our results are in contrast with numerous survival studies in humans with sepsis that indicate that hypovitaminosis D is predictive of ICU survival [8, 12, 29, 46, 47]. A limitation of this sub analysis is that it was performed with results from a small population of dogs with sepsis and may not have been powered to identify a difference if one had been present. Future, sufficiently powered studies are needed to investigate if serum 25(OH)D concentrations are independent predictors of survival in dogs with sepsis.

A number of limitations of our study should be considered when interpreting the results presented. Firstly, we cannot completely rule out the potential confounding influence of unassessed variables such as the volume infused in the initial hours after admission, vitamin D binding protein (VDBP) concentrations, and enteral nutrition. Intravenous fluid resuscitation in people lowers serum 25(OH)D concentration for up to 24 hours [48]. Additional studies are needed to know if fluid shifts within 24 hours of ICU admission influence serum 25(OH)D concentrations in critically ill dogs. We were unable to control for VDBP because a validated assay for use in dogs was not available during the time of our study. A decrease in this protein during critical illness could impact 25(OH)D concentrations. Vitamin D binding protein concentrations decrease in humans during times of diffuse inflammation, including in sepsis, severe trauma, and after surgery [49]. The association of vitamin D with its binding protein in humans is complex and non-linear [50]. Because of this complexity, future studies evaluating hypovitaminosis D and survival in dogs would benefit from also measuring VDBP concentrations. In addition, we did not investigate alternative vitamin D pathways or metabolites that may influence 25(OH)D concentrations or represent novel biomarkers for survival. An alternative vitamin D3 metabolism pathway utilizes steroidogenic enzyme cytochrome P450scc (CYP11A1) and 1α-hydroxylase (CYP27B1) to convert vitamin D3 to 20(OH)D3 and 1,20(OH)2D3, respectively, in people, pigs, and mice [51, 52]. Metabolites from this alternative pathway have comparable antiproliferative and anti-inflammatory properties as the classical 1,25(OH)2D3 system [52]. We did not record nutritional intake for the dogs in our study. Therefore, it is possible dogs provided assisted nutrition via tube feedings or voluntarily consumed food increased their serum 25(OH)D concentration. However, this is unlikely because the half time to equilibration of serum 25(OH)D for a given level of vitamin D intake is several weeks. Although serum 25(OH)D concentrations were obtained one time within 24 hours of admission, the duration of illness prior to sample collection varied. Vitamin D concentrations in people decrease in the acute phase of critical illness and increase during recovery [17]. It is possible that sampling at a different time point would have yielded different results. Lastly, we did not identify a causal organism in 4 dogs with sepsis. Two of these dogs had small intestinal foreign bodies with concomitant perforations identified at the time of surgery. The remaining 2 dogs were believed to have gut-derived sepsis [53].

Conclusions

Our findings indicate that critically ill dogs and dogs with sepsis have significantly lower serum 25(OH)D concentrations than do healthy control dogs. In addition, serum 25(OH)D concentration obtained within 24 hours of hospital admission was an independent predictor of survival. Serum 25(OH)D concentration had excellent sensitivity, but poor specificity for detection of death. Serum 25(OH)D was inversely associated with APPLE fast score, but was not associated with ICU length of stay. Our results support the initiation of future clinical trials to investigate the effects of vitamin D (e.g., calcidiol or calcitriol) supplementation in critically ill dogs and indicates that dogs with severe illness may represent a valuable animal model to explore the effect vitamin D supplementation has on people in a critical care setting.

Acknowledgments

The authors thank Matt Haight, Savannah Smith, and Kaoru Tsurua for their technical support and Kate Anderson for literature review assistance.

References

  1. 1. Higgins DM, Wischmeyer PE, Queensland KM, Sillau SH, Sufit AJ, Heyland DK. Relationship of vitamin d deficiency to clinical outcomes in critically ill patients. J Parenter Enteral Nutr. 2012;36(6):713–720.
  2. 2. Venkatram S, Chilimuri S, Adrish M, Salako A, Patel M, Diaz-Fuentes G. Vitamin D deficiency is associated with mortality in the medical intensive care unit. Crit Care Res Pract. 2011;15(6):R292.
  3. 3. Dickerson RN, Van Cleve JR, Swanson JM, Maish GO, Minard G, Croce MA, et al. Vitamin D deficiency in critically ill patients with traumatic injuries. Burns Trauma. 2016;4:28. pmid:27833924
  4. 4. Quraishi SA, Bittner EA, Blum L, McCarthy CM, Bhan I, Camargo CA Jr. Prospective study of vitamin D status at initiation in critically ill surgical patients and risk of 90 day mortality. Crit Care Med. 2014;42(6):1365–1371. pmid:24557421
  5. 5. Braun AB, Gibbons FK, Litonjua AA, Giovannucci E, Christopher KB. Low serum 25-hydroxyvitamin D at critical care initiation is associated with increased mortality. Crit Care Med. 2012;40(1):63–72. pmid:21926604
  6. 6. Guan J, Karsy M, Brock AA, Eli I, Ledhard H, Hawryluk GW, et al. A prospective analysis of hypovitaminosis D mortality in 400 patients in the neurocritical care setting. J Neurosurg. 2016. pmid:27367248
  7. 7. Braun A, Chang D, Mahadevappa K, Gibbons FK, Liu Y, Giovannucci E, et al. Association of low serum 25-hydroxyvitamin D levels and mortality in the critically ill. Crit Care Med. 2011;39(4):671–677. pmid:21242800
  8. 8. Amrein K, Zajic P, Schnedl C, Waltensdorfer A, Fruhwald S, Holl A, et al. Vitamin D status and its association with season, hospital and sepsis mortality in critical illness. Crit Care. 2014;18(2):R47. pmid:24661739
  9. 9. De Pascale G, Vallecoccia MS, Schiattarella A, Di Gravio V, Cutuli SL, Bello G, et al. Clinical and microbiological outcome in septic patients with extremely low 25-hydroxyvitamin D levels at initiation of critical care. Clin Microbiol Infect. 2016;22(5):456.e7–456.e13.
  10. 10. Matthews RL, Ahmed Y, Wilson KL, Griggs DD, Danner OK. Worsening severity of vitamin D deficiency is associated with increased length of stay, surgical intensive care unit cost, and mortality rate in surgical intensive care unit patients. Am J Surg. 2012;204(1):37–43. pmid:22325335
  11. 11. de Haan K, Groeneveld ABJ, de Geus HRH, Egal M, Struijs A. Vitamin D deficiency as a risk factor for infection, sepsis and mortality in the critically ill: Systemic review and meta-analysis. Crit Care. 2014;18(6):660. pmid:25475621
  12. 12. Arnson Y, Gringauz I, Itzhaky D, Amital H. Vitamin D deficiency is associated with poor outcomes and increased mortality in severely ill patients. QJM. 2012;105(7):633–639. pmid:22331959
  13. 13. Anwar E, Hamdy G, Taher E, Fawzy E, Abdulattif S, Attia H. Burden and outcome of vitamin D deficiency among critically ill patients: A prospective study. Nutr Clin Pract. 2016;32(3):378–384. pmid:28537519
  14. 14. Penna G, Adorini L. 1α,25-dihydroxyvitamin D3 inhibits differentiation, maturation, activation, and survival of dendritic cells leading to impaired alloreactive T cell activation. J Immunol. 2000;164(5):2405–2411. pmid:10679076
  15. 15. Szymczak I, Pawliczak R. The active metabolite of vitamin D3 as a potential immunomodulator. Scand J Immunol. 2016;83(2):83–91. pmid:26678915
  16. 16. Altieri B, Muscoqiuri G, Barrea L, Mathieu C, Vallone CV, Mascitelli L, et al. Does vitamin D play a role in autoimmune endocrine disorders? A proof of concept. Rev Endocr Metab Disord. 2017;18(3):335–346. pmid:28070798
  17. 17. Czarnik T, Czarnik A, Gawda R, Gawor M, Piwoda M, Marszalski M, Maj M, et al. Vitamin D kinetics in the acute phase of critical illness: a prospective observational study. J Crit Care. 2018;43:294–299. pmid:28968524
  18. 18. Tokuda N, Levy RB.1,25-dihydroxyvitamin D3 stimulates phagocytosis but suppresses HLA-DR and CD13 antigen expression in human mononuclear phagocytes. Proc Soc Exp Biol Med. 1996;211(3):244–250. pmid:8633104
  19. 19. Chandra G, Selvaraj P, Jawahar MS, Banurekha VV, Narayanan PR. Effect of vitamin D3 on phagocytic potential of macrophages with live mycobacterium tuberculosis and lymphoproliferative response in pulmonary tuberculosis. J Clin Immunol. 2004;24(3):249–257. pmid:15114055
  20. 20. Motlagh BM, Ahangaran NA, Froushani SMA. Calcitriol modulates the effects of bone marrow-derived mesenchymal stem cells on macrophage functions. Iran J Basic Med Sci. 2015;18(7):672–676. pmid:26351558
  21. 21. Liu PT, Stenger S, Li H, Tan BH, Krutzik SR, Ochoa MT, et al. Toll-like receptor triggering of a vitamin D-mediated human antimicrobial response. Science. 2006;311(5768):1770–1773. pmid:16497887
  22. 22. Jeng L, Yamshchikov AV, Judd SE, Blumberg HM, Martin GS, Ziegler TR, et al. Alterations in vitamin D status and anti-microbial peptide levels in patients in the intensive care unit with sepsis. J Trans Med. 2009;7:28. pmid:19389235
  23. 23. Baeke F, Takiishi T, Korf H, Gysemans C, Mathieu C. Vitamin D: Modulator of the immune system. Curr Opin Pharmacol. 2010;10(4):482–496. pmid:20427238
  24. 24. Baeke F, Gysemans C, Korf H, Mathieu C. Vitamin D insufficiency: Implications for the immune system. Pediatr Nephrol. 2010;25(9):1597–1606. pmid:20180136
  25. 25. Leaf DE, Raed A, Donnino MW, Ginde AA, Waikar SS. Randomized controlled trial of calcitriol in severe sepsis. Am J Respir Crit Care. 2014;190(5):533–541.
  26. 26. Villaggio B, Soldano S, Cutolo M. 1,25-dihydroxyvitamin D3 downregulates aromatase expression and inflammatory cytokines in human macrophages. Clin Exp Rheumatol. 2012;30(6):934–938. pmid:23253631
  27. 27. Harishankar M, Afsal K, Banurekha VV, Meenakshi N. 1,25-Dihydroxy vitamin D3 downregulates pro-inflammatory cytokine response in pulmonary tuberculosis. Int Immunopharmacol. 2014;23(1):148–152. pmid:25194676
  28. 28. Neve A, Corrado A, Cantatore FP. Immunomodulatory effects of vitamin D in peripheral blood monocyte-derived macrophages from patients with rheumatoid arthritis. Clin Exp Med. 2014;14(3):275–283. pmid:23824148
  29. 29. Grubczak K, Lipinska D, Eljaszewicz A, Singh P, Radzikowska U, Miklasz P, et al. Vitamin D3 treatment decreases frequencies of CD16-positive and TNF-α-secreting monocytes in asthmatic patients. Int Arch Allergy Immunol. 2015;166(3):170–176. pmid:25872112
  30. 30. Jaffey J, Amorim J, DeClue AE. In vitro effects of vitamin D on phagocytosis, TLR4, and cytokine production [abstract]. J Vet Intern Med. 2017;31(4):775.
  31. 31. How KL, Hazewinkel HA, Mol JA. Dietary vitamin D dependence of cat and dog due to inadequate cutaneous synthesis of vitamin D. Gen Comp Endocrinol. 1994;96(1):12–18. pmid:7843559
  32. 32. Leary PF, Zamfirova I, Au J, McCracken WH. Effect of latitude on vitamin D levels. J Am Osteopath Assoc. 2017;117(7):433–439. pmid:28662556
  33. 33. Klenk J, Rapp K, Denkinger MD, Nagel G, Nikolaus T, Peter R, et al. Seasonality of vitamin D status in older people in southern Germany: Implications for assessment. Age Ageing. 2013;42(3):404–408. pmid:23542723
  34. 34. Gordon CM, DePeter KC, Feldman HA, Grace E, Emans SJ. Prevalence of vitamin D deficiency among healthy adolescents. Arch Pediatr Adolesc Med. 2004;158(6):531–537. pmid:15184215
  35. 35. Walsh JS, Bowles S, Evans AL. Vitamin D in obesity. Curr Opin Endocrinol Diabetes Obes. 2017;24(6):389–394. pmid:28915134
  36. 36. Carnevale V, Modoni S, Pileri M, Di Giorgio A, Chiodini L, Minisola A, et al. Longitudinal evaluation of vitamin D status in healthy subjects from southern Italy: Seasonal and gender differences. Osteoporos Int. 2001;12(12):1026–1030. pmid:11846328
  37. 37. Hayes G, Mathews K, Doig G, Kruth S, Boston S, Nykamp S, et al. The acute patient physiologic and laboratory evaluation (APPLE) score: A severity of illness stratification system for hospitalized dogs. J Vet Intern Med. 2010;24(5):1034–1047. pmid:20629945
  38. 38. DeClue AE, Osterbur K, Bigio A, Sharp CR. Evaluation of serum NT-pCNP as a diagnostic and prognostic biomarker for sepsis in dogs. J Vet Intern Med. 2011;25(3):453–459. pmid:21457321
  39. 39. Nye CJ, Musulin SE, Hanel RM, Mariani CL. Evaluation of the Lactate Plus monitor for plasma lactate concentration measurements in dogs. J Vet Emerg Crit Care. 2017;27(1):66–70.
  40. 40. Young LR, Backus RC. Oral vitamin D supplementation at five times the recommended allowance marginally affects serum 25-hydroxyvitamin D concentrations in dogs. J Nutr Sci. 2016;5:e31. pmid:27547394
  41. 41. Lensmeyer GL, Wiebe DA, Binkley N, Drezner ML. HPLC mthod for 25-hydroxyvitamin D measurement: comparison with contemporary assays. Clin Chem. 2006;52(6):1120–1126. pmid:16574756
  42. 42. Atalan HK, Gucyetmez B. Serum vitamin D level at ICU admission and mortality. Turk J Anaesthesiol Reanim. 2017;45(4):193–196. pmid:28868165
  43. 43. Lee P, Nair P, Eisman JA, Center JR. Vitamin D deficiency in the intensive care unit: An invisible accomplice to morbidity and mortality? Intensive Care Med. 2009;35(12):2028–2032. pmid:19756497
  44. 44. Alizadeh N, Khalili H, Mohammadi M, Abodollahi A. Serum vitamin D levels at admission predict the length of intensive care unit stay but not in-hospital mortality of critically ill surgical patients. J Res Pharm Pract. 2015;4(4):193–198. pmid:26645025
  45. 45. Putzu A, Belletti A, Cassina T, Clivio S, Monti G, Zangrillo A, et al. Vitamin D and outcomes in adult critically ill patients. A systematic review and meta-analysis of randomized trials. J Crit Care. 2017;38:109–114. pmid:27883968
  46. 46. Vipul P, Shuchi C, Avinash A, Manish G, Sukriti K, Ved P. Correlation of serum vitamin D level with mortality in patients with sepsis. Indian J Crit Care Med. 2017;21(4):199–204. pmid:28515603
  47. 47. Moromizato T, Litonjua AA, Braun AA, Gibbons FK, Giovannucci E, Christopher KB. Association of low serum 25-hydroxyvitamin D levels and sepsis in the critically ill. Crit Care Med. 2014;42(1):97–107. pmid:23982028
  48. 48. Krishnan A, Ochola J, Mundy J, Jones M, Kruger P, Duncan E, et al. Acute fluid shifts influence the assessment of serum vitamin D status in critically ill patients. Crit Care. 2010;14(6):R216. pmid:21110839
  49. 49. Madden K, Feldman HA, Chun RF, Smith EM, Sullivan RM, Agan AA, et al. Critically ill children have low vitamin D-binding protein, influencing bioavailability of vitamin D. Ann Am Thorac Soc. 2015;12(11):1654–1661. pmid:26356094
  50. 50. Barnett N, Zhao Z, Koyama T, Janz DR, Wang CY, May AK, et al. Vitamin D deficiency and risk of acute lung injury in severe sepsis and severe trauma: A case control study. Ann Intensive Care. 2014;4(1):5. pmid:24559079
  51. 51. Slominski AT, Kim TK, Li W, Postlethwaite A, Tieu EW, Tang EKY, et al. Detection of novel CYP11A1-derived secosteroids in the human epidermis and serum and pig adrenal gland. Sci Rep 2015;5:14875. pmid:26445902
  52. 52. Slominski AT, Kim TK, Shehabi HZ, Semak I, Tang EKY, Nguyen MN, et al. In vivo evidence for a novel pathway of vitamin D3 metabolism initiated by P450scc and modified by CYP27B1. FASEB J 2012;26(9):3901–3915. pmid:22683847
  53. 53. Deitch EA. Gut-origin sepsis: evolution of a concept. Surgeon. 2012;10(6):350–356. pmid:22534256