Establishment of a reference interval for thiamine concentrations in healthy dogs and evaluation of the prevalence of absolute thiamine deficiency in critically ill dogs with and without sepsis using high performance liquid chromatography

Abstract

Objective:

To determine the normal reference interval for thiamine concentrations in healthy dogs and investigate the prevalence of thiamine deficiency in critically ill dogs with and without sepsis.

Design:

Prospective, observational, multicenter study, conducted between 2019 and 2021.

Setting:

Two veterinary university teaching hospitals.

Animals:

A total of 109 dogs were enrolled into 3 groups: 40 healthy dogs, 33 dogs with suspected or confirmed sepsis and evidence of tissue hypoperfusion (Doppler blood pressure ≤90 mm Hg or plasma lactate ≥3 mmol/L), and 36 dogs with other critical illnesses and evidence of tissue hypoperfusion.

Interventions:

For each dog, CBC, serum biochemistry, plasma lactate concentration, whole blood thiamine concentration, blood pressure, vital parameters, Acute Patient Physiologic and Laboratory Evaluation (APPLE)-fast score, and outcome were recorded, alongside basic patient parameters and dietary history. Whole blood thiamine pyrophosphate (TPP) concentrations were measured using high-performance liquid chromatography.

Measurements and Main Results:

The reference interval for whole blood TPP in healthy dogs was 70.9 to 135.3 μg/L. Median TPP concentrations were significantly lower in septic dogs compared to healthy controls (P = 0.036). No significant difference in median TPP concentrations was found between septic dogs and nonseptic critically ill dogs, or between healthy dogs and nonseptic critically ill dogs. Thiamine pyrophosphate concentrations were below the normal reference interval in 27.3% of septic dogs, compared to 19.4% of nonseptic critically ill dogs (P = 0.57). No correlations were found between TPP concentrations and lactate concentrations, age, body condition scores, time since last meal, RBC count, serum alanine aminotransferase, APPLE-fast scores, or patient outcomes.

Conclusions:

Thiamine pyrophosphate concentrations were significantly lower in septic dogs compared to healthy controls, with an absolute thiamine deficiency found in 27.3% of septic dogs. The established TPP reference interval allows for further investigation of thiamine deficiency in critically ill dogs.

INTRODUCTION

Thiamine (vitamin B1) is an essential component of cellular metabolism and antioxidant pathways.^1,2 In its phosphorylated form, thiamine pyrophosphate (TPP) plays a critical role in carbohydrate metabolism and ATP production by acting as a cofactor for pyruvate dehydrogenase and α-ketoglutarate dehydrogenase in the Krebs cycle.^1–3 TPP is also an integral cofactor of transketolase in the pentose phosphate pathway, which is essential for the generation of NADPH and, consequently, glutathione cycling, an important antioxidant pathway.^1,2 As a water-soluble vitamin, thiamine stores in the body are limited, making adequate dietary intake critical for maintaining normal concentrations.⁴ Without adequate thiamine concentrations, pyruvate cannot enter the Krebs cycle, causing a shift from aerobic to anaerobic metabolism and leading to a type B lactic acidosis, inability to utilize oxygen and produce ATP, and ultimately, cardiovascular collapse.^1,2,5

Thiamine deficiency is common among people with sepsis, ranging in prevalence from 10% to 70%, and is considered a potential contributor to mitochondrial dysfunction, refractory lactic acidosis, and multiple organ dysfunction syndrome (MODS) in these patients.^2,3,5,6 Thiamine administration has been suggested to improve lactate clearance and decrease mortality in people with septic shock, although it has not been demonstrated in additional large-scale studies and meta-analyses.^7–10

In dogs with sepsis, MODS significantly increases the odds of death, but therapeutic interventions to prevent or mitigate the development of MODS are poorly understood.^11,12 Thiamine administration has been shown to improve lactate clearance, oxygen consumption, and arterial blood pressure in an experimental model of canine septic shock but, to date, has not been evaluated in clinical patients.¹³

There has been limited work investigating thiamine concentrations in veterinary medicine.^4,14–17 Published values for thiamine concentrations in healthy dogs are scarce, with no established reference interval (RI) to date.^4,15,16 In dogs, thiamine deficiency is most commonly seen secondary to insufficient dietary intake, associated with malnutrition or an unbalanced diet.^4,15,17 In dogs with normal dietary intake, thiamine deficiency has been reported in various illnesses, such as chronic gastrointestinal disease associated with severe malabsorption or congestive heart failure.⁴ The prevalence of thiamine deficiency in dogs with critical illness and with sepsis, in particular, is unknown.^a,4,14

The objectives of this study were to determine the normal RI for whole blood TPP concentrations in healthy dogs using high-performance liquid chromatography (HPLC) and to investigate the prevalence of thiamine deficiency in critically ill dogs with and without sepsis. Thiamine concentrations were hypothesized to be lower in critically ill and septic dogs compared to healthy controls. Additionally, to further characterize thiamine deficiency in critically ill dogs with and without sepsis, TPP concentrations were analyzed in light of demographic, clinical, and laboratory variables.

METHODS

This prospective, multicenter, observational study was conducted in 2 veterinary university teaching hospitals. Dogs were enrolled into 3 groups: healthy dogs, critically ill dogs with sepsis, and dogs with nonseptic critical illness. All dogs were enrolled after obtaining informed owner consent. The study was approved by the respective Institutional Animal Care and Use Committee (IACUC) at both participating facilities, as well as by the Clinical Science Research Committee (CSRC) at Tufts University.

Healthy adult dogs (>1 years old) were recruited from the student and staff population of 1 of the participating university teaching hospitals. Dogs were determined to be healthy based on medical history, physical examination, and a normal CBC and serum biochemistry profile. All healthy dogs enrolled were fed a balanced commercial diet (not home cooked or raw), received no thiamine-containing supplements in the last 14 days, and were fasted for 12 hours prior to sample collection to minimize dietary-related variability in thiamine concentrations for the purpose of an RI establishment.

Critically ill dogs were recruited from the emergency services and intensive care units of the participating hospitals. Dogs were enrolled after a clinical diagnosis of sepsis or other critical illness with evidence of tissue hypoperfusion (Doppler blood pressure ≤90 mm Hg or blood lactate ≥3 mmol/L). Sepsis was defined as life-threatening organ dysfunction caused by a dysregulated host response to infection.¹⁸ Enrollment criteria for septic dogs included clinical evidence of infection, fulfillment of the systemic inflammatory response syndrome (SIRS) criteria, and evidence of tissue hypoperfusion as previously described.^19,20 Critical illness was defined as any illness that is potentially life threatening, with evidence of organ dysfunction and hypoperfusion, as determined by the enrolling clinicians.²¹

For each dog, the signalment, a complete medical history, and a full dietary history (including specific type and amounts of diet, treats, dietary supplements including thiamine-containing supplements) and time of last meal was recorded, as well as the body condition score (BCS).²² Dogs that received any thiamine supplementation within the last 14 days were not included. Vital parameters (temperature, respiratory rate, heart rate), arterial blood pressure, Acute Patient Physiologic and Laboratory Evaluation (APPLE)-fast scores, diagnosis, and clinical outcomes were recorded for all critically ill dogs (with and without sepsis).²³ Fluid resuscitation status or any drugs other than thiamine supplementation were not recorded, nor were they controlled for in these patients.

Four mL of whole blood was collected via venipuncture or IV catheter into a syringe and divided into an EDTA tube (1.5 mL), heparin tube (0.5 mL), and a serum separator tube (2 mL) within 12 hours of enrollment. Samples were not collected within 1 hour of administration of an IV fluid bolus of ≥10 mL/kg to avoid sample dilution. Whole blood anticoagulated with EDTA was used to obtain a CBC and PCV.^b An 0.5-mL aliquot of the EDTA-anticoagulated whole blood was transferred into Eppendorf tubes that were frozen at −80°C and submitted batched on dry ice to a specialized referral laboratory for whole-blood TPP measurement using HPLC.^c,16,24 Whole blood TPP concentrations were analyzed as described by Talwar et al.^24,c Samples were stored at −80°C for 7 months or less prior to analysis to ensure sample stability.^16,24 Quality control (QC) and calibration samples were prepared in-house by the referral laboratory by weighing thiamine and preparing stock solutions separate from calibration materials. The QC target was defined as 50 μg/L. Heparinized blood was used to measure plasma lactate concentrations.^d Serum was used for evaluation of biochemical parameters and to specifically assess correlation between alanine aminotransferase (ALT), TPP, and plasma lactate concentrations, as liver injury has been identified as a confounding factor in human studies.^5,e,f

Statistical methods

A 1-sided 2-sample means test was performed to determine the number of dogs required to detect a decrease of 35 μg/L in critically ill dogs (with or without sepsis) compared to healthy dogs with an assumed mean TPP concentration of 311.4 ± 46.8 μg/L in healthy dogs.¹⁶ Additional assumptions included a power of 0.8 and type I error rate of 0.05. This calculation resulted in a required sample size of 30 TPP concentration measurements for each group of dogs. Enrollment continued for the duration of allocated research time, and 40 samples were obtained from healthy dogs for the purpose of establishing a normal RI for canine thiamine concentration, by adding and subtracting 2 SDs from the mean TPP concentration of these samples (mean ± 2 SD).²⁵ The established RI was then applied to the groups of sick dogs for further analysis.

Variables were visually assessed for normality and using the skewness/kurtosis tests for normality, by implementing the test described by D’Agostino, Belanger, and D’Agostino with the empirical correction developed by Royston.^26,27 For the healthy dogs’ group, TPP data were also visually inspected to ensure there were no outliers. For normally distributed data, a 1-way ANOVA was chosen to assess differences between the groups. For abnormally distributed data, the nonparametric 2-sample Wilcoxon rank-sum (Mann–Whitney) test was chosen for comparison of medians from 2 independent samples. To further characterize thiamine concentrations in critically ill dogs with and without sepsis and to assess for potential confounding factors, the Spearman correlation was used to examine the association between thiamine concentrations and several demographic, clinical, and laboratory variables. These variables included plasma lactate concentrations, age, BCS, time since last meal, RBC count (as TPP is measured in RBCs), serum ALT, APPLE-fast scores, and patient outcomes. Fisher’s exact test was used to determine if there is a relationship between 2 categorical variables because some cells had fewer than 5 observations. P-values < 0.05 were considered significant for all comparisons. All statistical analyses were performed using a statistical software package.^g

RESULTS

A total of 109 dogs were enrolled between August 2019 and January 2021, including 40 healthy dogs, 33 critically ill dogs with sepsis, and 36 critically ill dogs without sepsis.

The healthy dog group included 25 females (2 intact, 23 neutered) and 15 males (1 intact, 14 neutered) of various breeds, as listed in Table 1. The median age was 7 years (range: 1–12 y). The median BCS was 5 (range: 4–7). There were no physical examination or laboratory abnormalities in any of the dogs in this group.

Table 1.

Summary of age, sex, breed, and BCS of all 109 dogs, as well as diagnosis, APPLE-fast scores, and outcomes of 69 critically ill dogs enrolled in the study

	Healthy dogs (n=40)	Septic dogs (n=33)		Critically ill nonseptic dogs (n=36)
Age in years (median, range)	7 (1–12)	8.5 (0.9–14.7)		9.4 (0.9–16.1)
Sex (number of dogs, %)	FI (2, 5%) FN (23, 57.5%) MI (1, 2.5%) MN (14, 35%)	FI (8, 24.2%) FN (14, 42.4%) MI (3, 9.1%) MN (8, 24.2%)		FN (13, 36.1%) MI (5, 13.9%) MN (18, 50%)
Breed (number of dogs, %)	Mixed breed (14, 35%) Golden Retriever (4, 10%) Chihuahua (3, 7.5%) Labrador (3, 7.5%) German Shorthaired Pointer (2. 5%) Shetland Sheepdog (2, 5%) Other breeds (11, 25%)*	Mixed breed (9, 27.3%) American Bulldog (3, 9.1%) Labrador Retriever (3, 9.1%) Great Dane (2, 6.1%) Pit Bull (2, 6.1%) Other breeds (14, 42%)*		Mixed breed (14, 38.9%) Labrador Retriever (4, 11.1%) English Bulldog (2, 5.6%) German Shepherd (2, 5.6%) Other breeds (14, 39.2%)*
BCS (median, range)	5 (4–7)	5 (2–9)		5 (2–7)
	Septic dogs (n=33)		Critically ill nonseptic dogs (n=36)
Diagnosis (number of dogs, %)	Septic peritonitis (17, 51.5%) Bite wounds/integument (4, 12.1%) Pneumonia (4, 12.1%) Reproductive (3, 9.1%) Gastroenteritis (2, 6.1%) Chemosepsis (1, 3%) Endocarditis (1, 3%) Neoplastic (1, 3%)		Hemorrhagic shock (9, 25%) Gastrointestinal obstruction/torsion (8, 22.2%) Cardiac (5, 13.9%) Hepatic/biliary (3, 8.3%) Anaphylaxis (2, 5.6%) Bite wounds/integument (2, 5.6%) Gastroenteritis (2, 5.6%) Neoplasia (2, 5.6%) Blunt trauma (1, 2.8%) Heat stroke (1, 2.8%) Pneumonia (1, 2.8%)
APPLE-fast score (median, range)	29 (15–39)		26.5 (12–48)
Outcome (number of dogs, %)	Survived to discharge (10, 30.3%) Euthanized (11, 33.3%) Died (11, 33.3%) Unknown (1, 3%)§		Survived to discharge (21, 58.3%)** Euthanized (13, 36.1%) Died (1, 2.8%)** Unknown (1, 2.8%)||

Note: APPLE-fast score ranges from 0 to 50. The value correlates with illness severity.

Abbreviations: APPLE, Acute Patient Physiologic and Laboratory Evaluation; BCS, body condition score; FI, female, intact; FN, female, neutered; MI, male, intact; MN, male, neutered; N/A = not applicable.

^*All breeds with only 1 dog per group were grouped under “Other breeds.”

^§Discharged against medical advice for hospitalization at another faciality but failed to arrive at said facility and was lost to follow-up.

^║Discharged against medical advice for hospice care or home euthanasia; lost to follow-up.

^**P < 0.05 compared to septic dogs.

The critically ill septic dogs included 22 females (8 intact, 14 neutered) and 11 males (3 intact, 8 neutered) of various breeds. The median age was 8.5 years (range: 0.9–14.7 y). The median BCS was 5 (range: 2–9). Common origins of sepsis included septic peritonitis (17 dogs, 51.5%), pneumonia (4 dogs, 12.1%), bite wounds/integument (4 dogs, 12.1%), reproductive (3 dogs, 9.1%), gastroenteritis (2 dogs, 6.1%), and other sources, as detailed in Table 1. The median APPLE-fast score was 29 (range: 15–39). Within this group, 10 (30.3%) dogs survived to discharge, 11 dogs (33.3%) were euthanized, and 11 dogs (33.3%) died. One dog was discharged against medical advice for hospitalization at another facility but failed to arrive at said facility and was subsequently lost to follow-up.

Among the 36 nonseptic critically ill dogs, 13 were neutered females, 5 were intact males, and 18 were neutered males, of various breeds, as detailed in Table 1. The median age was 9.4 years (range: 0.9–16.1 y). The median BCS was 5 (range: 2–7). Common conditions included hemorrhagic shock (9 dogs, 25%), gastrointestinal obstruction/torsion (8 dogs, 22.2%), cardiac disease (5 dogs, 13.9%), and hepatobiliary disease (3 dogs, 8.3%). A full list of diagnoses is listed in Table 1. The median APPLE-fast score was 26.5 (range: 12–48), which was not significantly different from the median score of the septic group (P = 0.45). Twenty-one (58.3%) of the dogs in this group survived to discharge, 13 (36.1%) dogs were euthanized, 1 dog died, and 1 dog was discharged against medical advice for hospice care and at-home euthanasia but was lost to follow-up. The mortality rate (excluding euthanasia) was significantly lower in this group compared to the septic group (2.8% vs 33.3%, respectively, P < 0.001), but the euthanasia rate was similar (36.1% vs 33.3%, respectively, P = 0.81), resulting in an overall higher survival rate among nonseptic critically ill dogs compared to septic dogs (58.3% vs 30.3%, respectively, P = 0.02).

TPP concentrations were normally distributed within the healthy dog group (mean ± SD: 103.1 ± 16.1 μg/L, median: 102.5 μg/L) and were used to determine the normal RI, which was found to be 70.9 (90% CI: 66.6–75.2) to 135.3 (90%CI: 131.0–139.6) μg/L. In addition to determining that TPP concentrations were normally distributed within the healthy dog group by skewness and kurtosis testing, data were also visually inspected to ensure there were no outliers. QC measures for TPP HPLC analysis revealed a mean accuracy of 98.5%, with a variance of 2.4% and an interday variance of 2.4%.

TPP concentrations were normally distributed within the nonseptic critically ill dogs but not in the septic dogs. Median TPP concentrations were 99.2 μg/L and 87.1 μg/L for the nonseptic and septic critically ill dogs, respectively. TPP concentrations were below the normal RI in 27.3% (9/33) of septic dogs and in 19.4% (7/36) of nonseptic critically ill dogs (P = 0.57).

Median TPP concentrations were significantly lower in the septic dogs compared to healthy controls (87.1 μg/L vs 102.5 μg/L, respectively; P = 0.036). There was no significant difference in median TPP concentrations between septic dogs and nonseptic critically ill dogs (87.1 μg/L vs 99.2 μg/L, respectively; P = 0.36), nor between healthy dogs and nonseptic critically ill dogs (102.5 μg/L vs 99.2 μg/L, respectively; P = 0.68). TPP concentrations per group are shown in Figure 1.

Figure 1.

Dot plot of fasted thiamine pyrophosphate concentrations (TPP, μg/L) measured using high-performance liquid chromatography of healthy, septic, and nonseptic critically ill dogs within 12 hours of hospital admission. The dot plot displays a point for each individual dog among the groups. The median is displayed as a dotted line, while the first and third quartiles are displayed as a dashed line. Median TPP concentrations were significantly lower in septic dogs compared to healthy controls (P = 0.036). No significant difference was found between septic dogs and nonseptic critically ill dogs or between healthy dogs and nonseptic critically ill dogs. Abbreviation: ns = no significant difference.

No correlations were found between TPP concentrations and any of the following: age, BCS, time since last meal, plasma lactate concentration, RBC count, ALT, APPLE-fast score, or patient outcome. Analyses were done separately for each group, as well as for all groups combined.

DISCUSSION

In this investigation, a normal RI for TPP using HPLC was established and was then used to investigate thiamine deficiency in critically ill dogs with and without sepsis. In critically ill dogs with sepsis, median TPP concentrations were significantly lower when compared to healthy controls, with an absolute thiamine deficiency found in 27.3% of those dogs.

A potential link between sepsis and thiamine deficiency has been reported in human literature, with a prevalence of 10% to 70%.^3,5 The high variability in the reported prevalence is suggested to be related to patient characteristics, illness severity, underlying conditions, and even the timing and technique of TPP analysis.^3,5 In dogs, systemic illness is often accompanied by anorexia, which, along with a catabolic state and other comorbidities, could potentially contribute to the development of thiamine deficiency.^4,14 For these reasons, we hypothesized that median TPP concentrations would be lower in all critically ill dogs (with or without sepsis) compared to healthy controls. However, while thiamine deficiency was observed in both septic and nonseptic critically ill dogs, it was more prevalent in the septic group, and median TPP was significantly lower only in the septic group. Moreover, there were no significant differences in age, BCS, or APPLE-fast scores between the 2 critically ill groups in our study, which further supports a possible link between sepsis and lower TPP concentrations in dogs. Additionally, no correlation was observed between TPP concentrations and age, BCS, or time since last meal in any of the groups, thus reducing the potential for their effect as potential biases.⁴ That being said, other factors which may affect TPP concentrations, such as specific medications and comorbid conditions, were not assessed in this study.^3,4,13 Larger-scale studies are warranted in order to further evaluate these correlations due to conflicting results in both the human and veterinary literature, with very little data available in the latter.^4,a Such studies are also warranted to confirm the observed prevalence of thiamine deficiency in critically ill dogs with sepsis, given the high variability in reported prevalence in human patients with sepsis.^3,5 Additionally, while the results in this study are statistically significant, their clinical significance remains unclear and warrants further research, especially given the overlap between the TPP ranges of the various groups.

Thiamine deficiency is a known cause of hyperlactatemia due to its key role in aerobic metabolism as an essential cofactor in the Krebs cycle.¹ However, no correlation between lactate and TPP concentrations were found in the current study. One possible explanation could be the underlying cause for hyperlactatemia in our patient population: while thiamine deficiency might contribute to hyperlactatemia in some of our patients, many other contributing factors could be involved, including type A hyperlactatemia resulting from significant hypoperfusion. Fluid resuscitation and volume status were not controlled in this study, which may present a limitation in interpreting these findings.²⁸ In a study by Donnino et al, a significant negative correlation between lactic acidosis and thiamine concentrations was observed only in a subset of septic patients with no concurrent liver disease, thus identifying liver dysfunction as a potential confounder.⁵ Those observations were not noted in this study, as no correlations were found between serum ALT and TPP concentrations, or between serum ALT and plasma lactate concentrations. Further studies should be considered that focus on patients with sustained hypotension or hyperlactatemia despite adequate fluid resuscitation.

TPP concentrations were not associated with illness severity or with patient outcomes. This finding coincides with reports in human medicine but conflicts with a recent veterinary report that found a positive correlation between thiamine concentrations and APPLE-fast scores in critically ill dogs with and without sepsis.^a,29 Further, large-scale studies are warranted, due to the relatively small size of both veterinary studies. Interestingly, despite a similar median APPLE-fast score (29 in septic dogs vs 26.5 in nonseptic critically ill dogs, P = 0.45), septic dogs had a higher rate of death (excluding euthanasia) compared to nonseptic critically ill dogs (33.3% vs 2.8%, respectively, P < 0.001) and an overall lower survival rate (58.3% of septic dogs vs 30.3% of nonseptic critically ill dogs survived to discharge, P = 0.02). This could potentially be attributed to the different disease etiologies within the nonseptic group, some of which are known to have better short-term outcomes compared to sepsis (eg,, trauma, hemorrhagic shock, gastrointestinal obstruction without sepsis, anaphylaxis).^11,30–33 Additionally, some of the nonseptic critical conditions may be associated with a shorter duration of illness or a better nutritional status prior to hospital admission, which may not be reflected in the data collected. Furthermore, humane euthanasia may impact the measured mortality rate in both groups.

Thiamine supplementation is regarded as safe, with a wide dosing range and few adverse effects when given orally, intramuscularly, or subcutaneously, although adverse events are associated with IV administration in dogs and cats.^4,14 In human medicine, thiamine supplementation has been investigated in patients with severe sepsis, with mixed results regarding its clinical benefit and ability to improve survival.^2,7–10 Differences in thiamine dosage and in patient populations may contribute to the reported discrepancies. Specifically, it has been suggested that thiamine administration is more beneficial in a subset of septic patients who are thiamine-deficient or are more prone to thiamine deficiency.^3,7 This hypothesis is supported by a recent small-scale veterinary study in which thiamine was significantly lower only in a subset of septic dogs requiring surgery.^a Thiamine supplementation may improve outcome in critically ill dogs with sepsis, but further research is warranted. Furthermore, large-scale observational studies to try to identify subpopulations of septic dogs that may benefit from thiamine administration should be considered.

This study has several limitations. As a feasibility investigation, a relatively small number of dogs were enrolled using relatively wide inclusion criteria for both critically ill dog groups, with the intent of casting a wide net at this preliminary stage. While this strategy is suited for a preliminary investigation, large-scale and more focused studies are still needed to better understand the epidemiology of thiamine deficiency in critically ill dogs with sepsis. Specifically, focusing on a subset of patients with septic shock can be considered.¹⁸ Additional large-scale studies are also required due to conflicting reports in recent veterinary literature, including a recent study that found no significant difference in thiamine concentrations between healthy and in critically ill dogs with or without sepsis.^a However, lower thiamine concentrations were identified in a subset of septic dogs that also required surgery.^a This finding reiterates the need for further research to identify subpopulations of septic dogs that may require thiamine supplementation. Another possible limitation is the use of the SIRS criteria in the definition of sepsis.¹⁹ Although this is still the accepted definition of sepsis in veterinary medicine, it has been replaced by the use of the Sequential Organ Failure Assessment (SOFA) score or quick SOFA (qSOFA) score in human medicine, due to insufficient sensitivity and specificity of the SIRS criteria for sepsis screening.^18,34 Unfortunately, neither the SOFA nor the qSOFA has been successfully validated as a veterinary sepsis screening tool, even though an increase in serial SOFA scores has been associated with mortality in dogs with sepsis.^34–36 Finally, correction for multiple testing was not performed because, in early exploratory studies such as this one, the benefit of initial discoveries that can pave the way for future studies outweighs the risk of a type I statistical error in which a false-positive finding is reported.³⁷

The RI for whole blood TPP using HPLC established in this study has several limitations that warrant caution in its interpretation and its application to other populations. While the use of HPLC to measure whole-blood TPP concentrations is considered the most specific, sensitive, and stable method for thiamine analysis, differences in laboratory methods, instruments, protocols, and reagents may cause variability between different laboratories.^14,38 Additionally, differences in the local canine populations due to differences in geographic regions and in regional population genetics are expected. Thus, establishing an institutional RI or running control samples for further quality assurance is recommended. Furthermore, due to financial limitations, this RI was established using 40 animals only, which is considered the minimal acceptable sample size for a RI.²⁵ This relatively small sample size may enhance the aforementioned limitations, thus increasing the potential impact of population-related differences and individual characteristics (such as age, breed, diet, or genetics), which may limit the transferability of this RI to other populations. Moreover, the relatively small sample size may contribute to the overlap observed between the TPP ranges of the healthy animals and the sick animals, which may impair the potential diagnostic utility of the established RI.

In the current study, TTP was only measured once per dog (within 12 hours of hospital admission), but thiamine deficiency may also develop during hospitalization due to prolonged anorexia, ongoing losses, and increased demand as a result of ongoing illness.^4,14 In fact, thiamine deficiency has been shown to develop over the first 72 hours of hospitalization in septic patients who were not thiamine deficient upon hospital admission.⁵ These findings were not observed in a recent veterinary report that measured thiamine concentrations over 72 hours in a small group of both healthy and critically ill dogs with or without sepsis.^a Future studies to further assess thiamine concentrations over time should be considered, as serial thiamine measurements during hospitalization may expose a higher prevalence of thiamine deficiency than currently observed, especially given the relatively small study populations.

In conclusion, thiamine deficiency was identified in 27.3% of septic dogs in this study. Further research is warranted to determine the clinical significance of this finding. The normal TPP RI measured by HPLC established in the current study will enable further investigation of thiamine deficiency and thiamine administration in critically ill dogs.

Abbreviations

ALT

alanine aminotransferase

APPLE

Acute Patient Physiologic and Laboratory Evaluation

BCS

Body Condition Score

BPM

Beats per Minute

CRI

Continuous Rate Infusion

CI

Confidence Interval

HPLC

High Performance Liquid Chromatography

MODS

Multiple Organ Dysfunction Syndrome

RI

Reference Interval

SOFA

Sequential Organ Failure Assessment

TPP

Thiamine Pyrophosphate