Methods

This prospective, randomized, cross-over experimental study was conducted at the Scott Ritchey Research Center at Auburn University College of Veterinary Medicine. The study was approved by and adhered to IACUC guidelines (protocol No. 2023-5221). The cats were originally sourced from a colony under protocol No. 2021-3909. Between May and June 2023, 10 healthy, adult, purpose-bred domestic shorthair cats were enrolled. The cats consisted of 5 intact males and 5 intact females. The median age was 17 months (range, 14 to 81 months), and the mean body weight was 3.80 ± 0.97 kg. Each cat was deemed healthy based on physical examination, CBC, serum biochemistry profile, and N-terminal pro-B-type natriuretic peptide SNAP test. Housing and enrichment were provided to the subjects following approved IACUC protocols.

Data collection

For each experiment, Hb concentration and PCV were measured at the following time points (Figure 2): after cephalic catheter placement, at the start and end of target temperature stabilization, and immediately before, during, and after the fluid bolus administration. During the fluid bolus, Hb was measured every 5 minutes. Afterward, Hb was measured every 10 minutes for 60 minutes and then every 20 minutes for an additional 80 minutes. Total anesthesia time ranged from 250 to 300 minutes.

View Full Size

Schematic of the experimental design showing premedication (A), cephalic catheterization (B), induction of anesthesia and initiation of instrumentation (C), jugular catheter placement (D), achievement and stabilization of target temperature (E), and administration of fluid bolus (F). Arrows indicate specific time points for blood sample collection.

Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.24.09.0279

• Download Figure
• Download figure as PowerPoint slide

For each blood sampling, 0.5 mL of whole blood was collected into a 500-μL EDTA microtainer tube. Blood samples were collected from the jugular catheter using the following technique: aspiration of 1 mL of sample from the jugular catheter, collection of the blood sample (0.5 mL of whole blood), return of the 1-mL initial sample (diluted blood) through the cephalic catheter, followed by a 0.25-mL flush of heparinized saline (2 units/mL) through both the cephalic and jugular catheters.

Hemoglobin was measured with a point-of-care analyzer (Hb 201+; HemoCue) that had been previously validated for use in cats.¹¹ Duplicate measurements of Hb and PCV were performed to assess the coefficient of variance and reduce measurement error.

Twenty experimental datasets (10 from the normothermic group and 10 from the hypothermic group) were acquired from 10 cats on 2 separate days, with at least 1 week between sessions.

Results

All 10 cats completed the study. Analysis of body temperature, HR, and ETISO showed significant effects of both time (P < .001) and treatment (P < .001). The EMMs and estimated differences for each treatment are summarized (Table 1). For body temperature, pairwise comparisons between time points revealed significant differences for some comparisons; however, these differences were < 0.3 °C and are therefore not reported. Heart rate was significantly higher at 5 and 15 to 30 minutes postinduction compared to 100 minutes and at 30 minutes compared to 120 and 160 minutes. The ETISO was significantly lower from 0 to 40 minutes compared to 160 minutes and at 10 to 20 minutes compared to 120 minutes.

Analysis of direct blood pressure showed significant treatment effects on SAP (P = .014), MAP (P < .001), and DAP (P < .001), with a significant effect of time on DAP (P = .041). The normothermic group had higher EMMs and estimated differences in all cases, and these data are summarized (Table 1). Pairwise comparisons between time points for DAP did not reveal any statistically significant differences after adjusting for multiple comparisons.

Urine output for the cats is summarized (Table 1), with no statistically significant difference found between groups (P = 0.155).

Volume kinetics analysis

Twenty data sets were obtained; however, direct blood pressure measurements were unavailable for 2 cats in the hypothermic group due to failure of arterial catheter placement. A subgroup analysis of the remaining 16 data sets revealed that direct blood pressure was not a significant covariate. Consequently, all 20 data sets were analyzed together, excluding the direct blood pressure values entirely. Of the 600 paired serial Hb concentration measurements, 32 (5.3%) were excluded due to a lighter plane of general anesthesia and the administration of alfaxalone.

Both 1-VOFS and 2-VOFS kinetic models were evaluated to determine the best statistical fit. The 2-VOFS kinetic model had the lowest Akaike information criterion value and was selected as the base model for covariate analysis, which identified several significant covariates. In the final model, body weight was a significant covariate for both V_c and k₁₀, body temperature was a significant covariate for V_c, and ETISO was a covariate for both V_c and k₁₂ (Table 2). The goodness-of-fit and residual plots showed slight improvements after incorporating these covariates (Figures 3 and 4).

Table 2

Volume kinetic (VK) parameter estimates from the 2-volume of fluid space kinetic model and representative covariates reflecting interindividual variabilities of VK parameters.

VK parameter	Covariate	Estimate	2.5% CI	97.5% CI	CV (%)
Vc (mL)		216.0	200.6	231.4	3.63
k10 (L/min)		0.0013	0.0011	0.0014	6.39
k12 (L/min)		0.0312	0.0210	0.0413	16.55
k21 (L/min)		0.0109	0.0015	0.0202	43.62
Covariate effect
Vc (mL)	Body weight	1.143	0.7944	1.491	15.48
Vc (mL)	Hypothermia	0.2122	0.0886	0.3357	29.58
k12 (L/min)	ETISO	1.629	0.6606	2.598	30.21
k10 (L/min)	Body weight	–1.241	–2.072	–0.4107	–34.00
Vc (mL)	ETISO	0.3907	0.0919	0.6896	38.86

CV = Coefficient of variance. k₁₀ = Elimination rate constant. k₁₂ = Distribution rate constant from the central to the peripheral compartment. k₂₁ = Distribution rate constant from the peripheral to the central compartment. V_c = Volume of central compartment.

View Full Size

Model-predicted versus observed plasma dilution plots for the 2-volume fluid space kinetic model before (A and B) and after (C and D) accounting for significant covariate effects, including hypothermia, body weight, and end-tidal isoflurane concentration.

Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.24.09.0279

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Conditional weighted residuals versus predicted values for the 2-volume fluid space kinetic model before (A and B) and after (C and D) accounting for significant covariate effects, including hypothermia, body weight, and end-tidal isoflurane concentration.

Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.24.09.0279

• Download Figure
• Download figure as PowerPoint slide

In summary, the VK parameter estimates for individual cats in the final model can be expressed as follows:

$$\begin{array}{l}{\text{V}}_{\text{c}}\text{individual}\hfill \\ =({\text{V}}_{\text{c}}\text{population}\times {\text{e}}^{\text{Group}})\hfill \\ \times (\text{body}\text{weight}\text{individual}\hfill \\ ÷\text{mean}\text{population}\text{body}\text{weight}{)}^{1.1426}\hfill \\ \times (\text{ETISO}\text{individual}\hfill \\ ÷\text{mean}\text{population}\text{ETISO}{)}^{0.3907}\hfill \\ {\text{k}}_{10}\text{individual}\hfill \\ ={\text{k}}_{10}\text{population}(\text{body}\text{weight}\text{individual}\hfill \\ ÷\text{mean}\text{population}\text{body}\text{weight}{)}^{-1.2413}\hfill \\ {\text{k}}_{12}\text{individual}={\text{k}}_{12}\text{population}\hfill \\ {\left(\text{ETISO}÷\text{mean}\text{ETISO}\right)}^{1.6293}\hfill \end{array}$$

where the group exponent is 0 for the normothermic group and 0.2122 for the hypothermic group.

Discussion

This study investigated the effects of mild hypothermia on the VK of an IV crystalloid fluid bolus in healthy anesthetized cats. The 2-VFOS kinetic model provided a better fit than 1-VFOS kinetic model. Covariate analysis identified hypothermia, body weight, and ETISO as significant covariates affecting V_c, k₁₀, and k₁₂. Mild hypothermia was associated with a larger V_c. Higher body weight was associated with a larger V_c, as might be expected, and a slower k₁₀, while higher ETISO correlated with both an increased V_c and a faster k₁₂. Furthermore, hypothermia resulted in reduced HR, lower blood pressure, and lower ETISO. Contrary to the initial hypothesis, hypothermia did not significantly affect the k₁₀ of the fluid bolus or induce cold diuresis. Additionally, hypothermia was not associated with slower fluid k₁₂.

The superior fit of the 2-VOFS model aligns with the findings of Yiew et al,¹⁴ who used VK modeling to explore fluid disposition in healthy conscious cats. The study of Yiew et al¹⁴ reported a V_c of 139 mL (95% CI, 83 to 195), whereas this study found a higher V_c of 216 mL (95% CI, 201 to 231) in normothermic anesthetized cats. Differences in methodology, cat populations, and the effects of anesthesia likely contributed to these variations. Anesthesia, possibly through mechanisms such as vasodilation and splenomegaly, may lead to larger V_c. It is important to note that in VK modeling, the central compartment represents an apparent or functional volume, similar to the volume of distribution in traditional PK analysis, rather than a fixed anatomical space or actual blood volume. The intercompartmental rate constants (k₁₂ and k₂₁) in the study of Yiew et al¹⁴ were 0.753/min (95% CI, 0.199 to 1.308) and 0.161/min (95% CI, 0.089 to 0.233), respectively. These values indicate a more rapid fluid exchange between compartments in conscious cats compared to the anesthetized subjects in this study, where the values were lower at 0.0312/min (95% CI, 0.0210 to 0.0413) for k₁₂ and 0.0109/min (95% CI, 0.0015 to 0.0202) for k₂₁. Additionally, k₁₀ was slower in this study, with a value of 0.0013/min (95% CI, 0.0011 to 0.0014), compared to the k₁₀ in the study of Yiew et al¹⁴ of 0.007/min (95% CI, 0.004 to 0.010). It should be emphasized that the comparison between this study and the study of Yiew et al¹⁴ involves comparing anesthetized and conscious states. Another study¹⁵ in healthy, normothermic, anesthetized cats found a V_c of 78.8 mL/kg (236 mL for a 3 kg cat) and a k₁₂ of 0.0419/min. These findings are consistent with the results of this study, suggesting a level of consistency across different conditions in anesthetized cats. Without data on awake subjects from this study, a direct comparison with the study of Yiew et al¹⁴ cannot be made. Therefore, further research comparing awake and anesthetized states is needed to fully understand these differences.

The lower ETISO observed in hypothermic cats (1.22% vs 1.46%) suggests that hypothermia may reduce anesthetic requirements, consistent with previous studies^16–18 in both humans and animals. Specifically, Liu et al¹⁶ demonstrated a reduction of approximately 5.1% in isoflurane MAC per degree Celsius drop in body temperature in children, which aligns with the findings in hypothermic cats. Contrary to our hypothesis, hypothermia did not significantly affect the elimination rate or increase urine output in anesthetized cats, diverging from reports^5,6,19 of cold diuresis observed in other mammalian species such as rats and humans. Although there is a lack of feline-specific data, this slower elimination and lack of cold diuresis observed in anesthetized cats may be attributed to the effects of anesthesia on renal function and overall fluid distribution, retention, and elimination.^20–23 In conscious humans, mild hypothermia typically triggers sympathetic vasoconstriction and decreases ADH secretion, leading to increased urine output.⁶ However, anesthesia appears to significantly alter these responses. Studies^20,21 have shown that anesthesia can increase plasma ADH concentrations, promoting water retention and reducing urine output in humans and dogs. This increase in ADH during anesthesia likely counteracts the diuretic effects usually seen in hypothermia. Furthermore, anesthesia-induced vasodilation can offset the sympathetic vasoconstriction that normally occurs as an initial response to hypothermia, thereby reducing the stimulus for increased urine production.²⁰ In addition, a previous VK study²² demonstrated that isoflurane reduces the k₁₀ by 50% in anesthetized, spontaneously breathing human subjects compared to when awake. A subsequent study⁴ that combined data from several publications in humans attributed the reduction in k₁₀ to MAP and age. It has been proposed that decreased MAP due to anesthesia leads to the unloading of baroreceptors, which subsequently activates renal sympathetic nerves. This activation enhances sodium and water reabsorption and induces renal vasoconstriction, decreasing urine output.²³ Although arterial blood pressure was not a significant covariate in this study, the influence of ETISO on V_c and k₁₂ suggests that anesthesia-related changes could be relevant in cats.

Covariate analysis in this study demonstrated that hypothermia and higher ETISO were associated with increased V_c. Both inhalant anesthesia and hypothermia can induce vasodilation and splenomegaly, which may partly explain the larger V_c.^24–27 Furthermore, higher ETISO were linked to a faster k₁₂. The dose-dependent vasodilation from isoflurane could contribute to this higher k₁₂ by increasing capillary hydrostatic pressure due to the relaxation of arterioles.^24,28

Cats with higher body weights had a larger V_c and a slower k₁₀. The positive relationship between weight and V_c is consistent with the understanding that body weight correlates with blood volume. This relationship is supported by another study²⁹ demonstrating a linear correlation between blood volume and body weight in cats up to 4 kg, beyond which the increase in blood volume becomes less apparent. The authors proposed that weight gain above 4 kg is primarily due to an increase in fat, which is metabolically inactive and contains less blood volume. This same mechanism may also explain the negative relationship between body weight and k₁₀. Recent human studies^30,31 suggest that using actual body weight to calculate urine output can lead to an overestimation of acute kidney injury in obese patients, as the additional body fat contributes minimally to metabolic activity. In this study, despite all cats having body condition scores of 4 out of 9, body composition analysis was not performed. Therefore, whether the negative relationship between k₁₀ and body weight was due to excessive body fat or represents a spurious finding remains unclear.

In the study by Ahn et al,³² surface cooling in cats led to a progressive decrease in HR, blood pressure, and myocardial contractility. In that study, both SAP and DAP decreased significantly as body temperature fell, with DAP showing a more pronounced reduction. This study replicated these findings, demonstrating that hypothermia significantly lowered HR and blood pressure, with MAP and DAP showing a greater reduction than SAP.

The reduction in DAP observed during hypothermia may be attributed to several factors. Hypothermia has been shown to reduce norepinephrine release and diminish the effectiveness of catecholamines on organs and vessels.^25–27 Therefore, hypothermia induces vasodilation, which decreases peripheral resistance, directly lowering DAP since it is primarily influenced by arterial resistance.^33,34 Additionally, hypothermia leads to a significant reduction in HR, which extends the diastolic period. This extended diastolic period allows more time for blood to leave the arterial circulation, contributing further to the reduction in DAP.^33,35

This study offers valuable insights into the effects of mild hypothermia on VK in anesthetized cats; however, several important limitations should be considered. The small sample size, which consisted only of healthy, purposed-bred, anesthetized adult cats, may limit the generalizability of these findings to a broader feline population, including those of varying ages, health conditions, and body condition scores. Additionally, challenges with arterial catheter placement affected direct blood pressure measurements in 2 cats in the hypothermic group, potentially influencing the results. Despite this, blood pressure was assessed in 8 cats, and no significant covariate effect was found, with similar microconstant parameter estimates for the base model. Another limitation is the lack of cardiac output measurements, which would have provided a more comprehensive understanding of hemodynamic changes during hypothermia and fluid administration. Furthermore, other relevant factors affecting VK may not have been captured. Further studies should explore additional covariates and their interactions to gain a deeper understanding of the complex interplay between hypothermia, anesthesia, and fluid kinetics in cats.

This study investigated the effects of mild hypothermia on the VK of an IV crystalloid fluid bolus in healthy anesthetized cats. A 2-VOFS kinetic model best fits the data. Significant covariates affecting VK parameters included body weight, body temperature, and ETISO. Contrary to our hypothesis, mild hypothermia did not significantly impact urinary output but was associated with an increased V_c. This increase in V_c may represent decreased fluid potency and plasma volume expansion from the fluid bolus compared to normothermic states. The observed negative relationship between body weight and k₁₀ suggests the need to base fluid therapy on ideal rather than actual body weight to avoid potential overdosing. Additionally, the dose-dependent increase in V_c and k₁₂ with higher isoflurane concentrations suggests that higher isoflurane doses may reduce the potency and duration of the fluid bolus. Further studies are warranted to reproduce these findings and explore their implications.

Disclosures

The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.

Funding

The authors have nothing to disclose.

ORCID

X. T. Yiew https://orcid.org/0000-0002-3661-4763

S. Bateman https://orcid.org/0000-0002-0468-0256

K. K. Gerken https://orcid.org/0000-0001-8577-685X

K. W. Kuo https://orcid.org/0000-0003-4279-1966