Quantification of Induced Hypothermia from Aseptic Scrub Applications during Rodent Surgery Preparation

Abstract

Laboratory mice (Mus musculus) are prone to develop hypothermia during anesthesia for surgery, thus potentially impeding anesthetic recovery, wound healing, and future health. The core body temperatures of isoflurane-anesthetized mice are influenced by the choice of supplemental heat sources; however, the contribution of various surgical scrubs on the body temperatures of mice under gas anesthesia has not been assessed. We sought to quantify the effect of using alcohol (70% isopropyl alcohol [IPA]) compared with saline to rinse away surgical scrub on the progression of hypothermia in anesthetized mice (n = 47). IPA, room-temperature saline, or warmed saline (37 °C) was combined with povidone–iodine and then assessed for effects on core (rectal) and surface (infrared) temperatures. Agents were applied to a 2×2-cm shaved abdominal area of mice maintained on a water-recirculating blanket (at 38 °C) under isoflurane anesthesia (1.5% to 2.0% at 0.6 L/min) for 30 min. Although all scrub regimens significantly decreased body temperature at the time of application, treatments that included povidone–iodine led to the coldest core temperatures, which persisted while mice were anesthetized. Compared with room-temperature saline and when combined with povidone–iodine, warming of saline did not ameliorate heat loss. IPA alone demonstrated the most dramatic cooling of both surface and core readings at application but generated an unanticipated warming (rebound) phase during which body temperatures equilibrated with those of controls within minutes of application. Although alcohol is inappropriate as a stand-alone agent for surgical skin preparation, IPA is a viable alternative to saline-based rinses in this context, and its use should be encouraged within institutional guidance for rodent surgical procedures without concern for prolonged hypothermia in mice.

Rodent surgery is one of the most common procedures performed in comparative medical animal modeling. For surgical experiments, mice are anesthetized to minimize movement, enhance restraint, and ameliorate discomfort or pain during manipulations.¹³ Isoflurane is the most popular inhalant agent for veterinary care in small mammals and is typically administered without the complement of other anesthetics.¹⁶ Confounding outcomes after exposure to gas anesthesia are disruptions to normal physiologic processes, depression of the cardiopulmonary system, decreases in metabolism, and altered thermoregulation; surgical exposure of opened body cavities to ambient room air may also contribute to loss of body heat.^8,16,38 Previous studies have assessed thermal support options for rodents and responses to different types of anesthesia;^8,21 therefore, we performed the current study to quantitatively assess the potential effects of the skin preparation process on body temperatures of laboratory mice during gas anesthesia for surgical intervention.

Preparation of laboratory rodents for surgical procedures follows standards of veterinary patient care for larger species, including provision of thermal support, removal of hair, skin surface disinfection, and supportive care during and after anesthesia. All related efforts are aimed to contribute to uneventful recovery and return to consciousness. Aseptic techniques are applied for survival procedures, with the rationale that “aseptic surgical procedures… [will] prevent postsurgical infection” and “aseptic technique results in decreased inflammation … enhances recovery and reduces postoperative complication; [further] prevention of infection improves the welfare of the animal and eliminates a source of uncontrolled variation in the experimental results”.^3,4,20

When placed under general anesthesia, laboratory mice and rats typically require external heat supplementation to maintain body temperatures and avoid the development of hypothermia.^3,8 Hypothermia is undesirable in rodents due to consequences that include depressed cardiopulmonary and respiratory function, decreased metabolism of targeted drugs, delayed anesthetic recovery, decreased wound healing, and diminished efficacy of immunologic responses involving T cells and neutrophils.^8,16,38 Antiseptic and disinfectant agents typically used on skin have been presumed to decrease the temperature of rodents after application; however, this decrease may occur in the context of the surgical patient that is overly ‘soaked’ through the application of excessive amounts of scrub liquids.³⁶ The recommendation regarding laboratory rodents is the application of only small volumes of surgical scrub fluids, which should be limited to the specific area required for the surgical procedure.^10,19,23

Common scrub disinfectants used in rodent medicine include iodophors and chlorhexidine. Iodophors (for example, Betadine, composed of povidone–iodine) inactivate a wide range of microbes, but their action is reduced in the presence of organic matter; chlorhexidine (for example, Novalsan) is rapidly bactericidal, persistent, active against many viruses, and effective despite the presence of blood.²² It is not uncommon for institutional guidelines to recommend triplicate applications or prolonged contact times of alternating surgical prep liquids on skin. In essence, disinfectant is repeatedly applied and then removed with a ‘rinse,’ usually a neutral substance (typically an alcohol-based solution or sterile water or saline) to best achieve asepsis.^18,32,43,45 Although alcohols have broad-spectrum antimicrobial, antiviral, and antifungal effects,^24,28 their evaporative cooling effect is predicted to be aversive for use as a rinse in laboratory mice. Biomedical research institutions, including our own, have stated that alcohols (for example, 70% isopropyl alcohol [IPA] or ethanol) should be avoided for skin preparation in small rodents due to concerns regarding exacerbation of hypothermia.^{3,7,22,27,35,44} Thus, sterile saline rinse has been deemed as an alternative to alcohol for use in patient skin preparation.³

The characteristics of and basis for aseptic surgical preparation remain an area of dispute among scientists, veterinarians, and IACUC in the creation of rodent surgical models and for the prevention of surgery-associated infections within these models. Therefore, laboratory animal surgical practices may not have undergone rigorous evidence-based evaluations of efficacy. In this context, our current objective was to quantify the hypothermic effect of IPA applied alone and in combination with povidone–iodine (P-I) antiseptic, as compared a sterile saline rinse of P-I, for surgical preparation of rodent skin. Given the contention that alcohol applications to rodent skin potentiate hypothermia despite being recommended at numerous institutions as part of surgical protocols, we sought to evaluate quantitatively whether 70% isopropyl alcohol (IPA) in fact contributes to undesirable hypothermic outcomes during surgical skin preparations.

Materials and Methods

Enrolled mice (n = 48; male; age, 5 to 11 mo; weight, 31 to 45 g) were donated from an inhouse breeding colony. These mice were the progeny of mice that lacked tryptophan hydroxylase (Tph1^−/−) bred with commercially available C57BL/6J mice; among the enrolled mice, 53% were Tph1^+/–, and 47% were wildtype C57BL/6J. Tryptophan hydroxylase influences serotonin levels, such that Tph1 ^−/− mice have reduced levels of peripheral serotonin but adequate amounts in classic serotonergic brain regions.⁴⁶ Phenotypic alterations in the heterozygous mice have not been reported by the donating laboratory; furthermore, TPH1 heterozygotes are not described in the literature with regard to effects on metabolism and thermoregulation.^2,29 After transfer from the donor colony, mice were acclimated to their housing location for a minimum of 72 h, according to institutional recommendations, prior to the start of the study. Two mice were used in a pilot study to verify the skin preparation and temperature monitoring plan as described in the protocol, which was approved by the IACUC at Michigan State University. For the purposes of experimental consistency, all study manipulations, scrub applications, and temperature recordings were performed by the primary author (AS) to eliminate bias and to eradicate confounding variables.

Mice were housed in accordance with the Guide for the Care and Use of Laboratory Animals.²⁰ Briefly, mice were maintained on a 12:12-h light:dark cycle within individual static cages (75 in.², Ancare, Bellmore, NY) containing bedding (Aspen chips; Nepco, Warrensburg, NY) and enrichment (Bed-r'Nest, Andersons Lab Bedding, Maumee, OH). Mice received reverse-osmosis–purified water in bottles and had unrestricted access to chow (diet no. 8940, Envigo, Indianapolis, IN). Housing rooms were maintained at 71 to 72 °F (21.7 to 22.2 °C), with recorded relative humidity of 54% to 65%; the surgery room temperatures were 68 to 73 °F (20.0 to 22.8 °C), and humidity ranged between 19% to 33% during the time of experimental procedures. Sentinel mice were maintained within the housing room, tested quarterly, and found to be free of cilia-associated respiratory bacillus, ectromelia virus, Encephalitozoon cuniculi, lymphocytic choriomeningitis virus, minute virus of mice, mouse adenovirus, mouse coronavirus, Theiler murine encephalomyelitis virus, mouse hepatitis virus, mouse parvovirus, mouse rotavirus, mouse thymic virus, Mycoplasma pulmonis, Myobia musculi, Pasteurella pneumotropica, pneumonia virus of mice, polyoma virus, reovirus 3, Streptococcus pneumoniae (Charles River Laboratories Diagnostic Services, Wilmington, DE). In addition, the rodent colony was tested on site for the presence of fur mites and pinworms; none of these organisms were detected.

Mice (n = 8 per group) were assigned to one of 6 experimental groups. Mice were randomly distributed so that heterozygous and WT mice were mixed across groups. Mice were weighed at time of anesthesia; thus weight distribution was not a factor in group placements. Groups were randomized so that approximately half of the mice were anesthetized in the morning (0800 to 1200) and the other half in the afternoon (1200 to 1600), to account for potential fluctuations in body temperatures throughout the light cycle. Group assignments comprised: (1) control animals that received no scrub but were placed under isoflurane anesthesia and shaved; (2) skin preparation with P-I (10%; Betadine Solution, Purdue) alternating with 70% IPA; (3) P-I alternating with 0.9% NaCl saline; (4) P-I alternating with 0.9% NaCl saline that had been heated in a warm water bath to 37 °C, removed only during the application process, and then promptly returned to the heating bath; (5) 70% IPA only (applied 3 times); and (6) P-I only (applied 3 times). The P-I–only group was added after all initial mice had been assigned, therefore, this group contained exclusively B6/J WT mice from the original donor colony; animals were similarly segregated into morning and afternoon treatment subgroups. For all groups, the final liquid agent (whether saline, IPA, or P-I) was allowed to air dry on the skin.

Isoflurane (3% in O₂ at 0.6 L/min) was delivered initially by using an induction chamber (Parkland Scientific, Coral Springs, FL). Mice were deemed fully anesthetized when righting reflex was lost and no response was elicited by toe pinch. Isoflurane was reduced to 1.5% to 2% for maintenance of anesthesia with a nose cone at a flow of 0.6 L/min for approximately 38 min. Mice were then transferred onto a heating pad (T/Pump Localized Therapy System Circulating Water Warming Pad, Stryker, Kalamazoo, MI) maintained at 38 °C setting and placed in dorsal recumbency, with the nose cone positioned for continuous anesthetic gas delivery. Mice were not placed under surgical draping to avoid confounding temperature assessments of the scrubs’ effects. A temperature probe (TH5 Thermalert Monitoring Thermometer, Braintree Scientific, Braintree, MS) was placed intrarectally in each mouse, secured by medical tape to the tail, and elevated 1 cm from the surface of the heating pad to minimize interference due to heat from the pad itself. This device provided core rectal temperatures in real time. Surface temperatures were collected by pointing the capture light of the infrared scanner (model Ir-101, Wireless Infra-red Thermometer, La Crosse Technology, La Crosse, WI) at the ventral flank at a distance of approximately 1 cm above the skin surface. Thermometry measurements were collected beginning at time 0 (that is, immediately after rectal probe placement). Temperatures were recorded each minute from that point forward, with the rectal temperature recorded prior to the corresponding surface temperature at each collection point.

After the temperature baseline readings were collected (minutes 0 through 3), mice were shaved (no. 40 blade) to remove hair and expose a 2×2-cm area of skin over the abdominal region where a laparotomy incision might be placed. Additional recordings were taken (at minutes 4 through 7) prior to application of the scrub treatment. Scrub treatments were applied at minute 8 with a single-use cotton swab across the shaved skin. Swabs absorbed solutions to saturation, and then excess droplets were removed by rolling the swab against the container to allow excess to drain before application to the shaved abdominal area. They were applied 3 times in a counterclockwise manner, as is typical for skin surgery preparation, and alternated between scrubs and rinses as needed. With the mouse prepped and under isoflurane anesthesia, core and surface temperatures were collected from minutes 8 to 38, with no further manipulation. Core temperatures were not permitted to fall below 27 °C or exceed 41 °C, consistent with previous studies.⁸ Had temperatures fallen outside of the established parameters, mice were to be removed from anesthesia immediately and then warmed or cooled accordingly.

Isoflurane was discontinued at the 38-min time point, and the final temperature data were collected. Mice were kept on 100% oxygen at 0.6 L/min until movement was observed, at which time the rectal probe was removed. Once mice regained their righting reflex, they were transferred to a prewarmed cage for additional recovery. Mice were returned to their home cage once normal ambulation was observed.

Statistical analysis.

Statistical analyses were conducted by the Michigan State University Center for Statistical Training and Consulting Service (CSTAT). A piecewise multilevel longitudinal model³⁹ was used to capture the trajectory of temperature change and investigate the effect of different treatment groups, as previously described.³⁰ The Level-1 model was defined as

$$\begin{array}{c}{Y}_{it}={\mathrm{\pi }}_{0i}+{\mathrm{\pi }}_{1i}TIME{1}_{it}+{\mathrm{\pi }}_{2i}TIME{2}_{it}+{\mathrm{\pi }}_{3i}TIME{2}_{it}^2\\ +{\mathrm{\pi }}_{4i}TIME{2}_{it}^2+{\mathrm{\epsilon }}_{it}\end{array}$$

and the Level-2 model was defined as

$${\mathrm{\pi }}_{\mathit{pi}}={\mathrm{\beta }}_{\mathit{p}0}+{\mathrm{\beta }}_{\mathit{p}1}\mathrm{\times }TR{T}_i+{\mathrm{\zeta }}_{\mathit{it}}.$$

In these models, i = 1,…,N, t = 1,…,T, p = 1,…,4; N is the total number of mice in the experiment; and is the number of measurements for each individual. γ_it is the temperature measured for mouse i at time t; ε_it and ζ_it are normally distributed random errors. The variable TRT_i represents the treatment groups for individual.

Time was divided into 2 phases: TIME 1 and TIME 2. The variable TIME 1 was created to capture the trajectory of temperature prior to beginning of treatment; TIME 2 was used for the posttreatment temperature trajectory. Through likelihood ratio test modeling,⁴⁷ the relationship between time and temperature was further assessed; it was determined to use a linear function of TIME 1 in the Level-1 model to capture the trajectories before treatment was given. For posttreatment trajectories, we used orthogonal polynomials,⁹ which contained linear, quadratic, and cubic functions of TIME 2 in the Level-1 model. The variables TIME 1 and TIME 2 were coded such that the intercept was located at minute 7, the last time point prior to onset of treatment (calculations not shown).

To investigate the effect of body weight of mice on the trajectory of temperature changes, the following approach was used such that the Level-1 model was defined as

$$\begin{array}{c}{Y}_{it}={\mathrm{\pi }}_{0i}+{\mathrm{\pi }}_{1i}TIME{1}_{it}+{\mathrm{\pi }}_{2i}TIME{2}_{it}+{\mathrm{\pi }}_{3i}TIME{2}_{it}^2\\ +{\mathrm{\pi }}_{4i}TIME{2}_{it}^3+{\mathrm{\epsilon }}_{it}\end{array}$$

and the Level-2 model was defined as

$$\begin{array}{c}{\mathrm{\pi }}_{\mathit{pi}}={\mathrm{\beta }}_{\mathit{p}0}+{\mathrm{\beta }}_{\mathit{p}1}\mathrm{\times }TR{T}_i+{\mathrm{\beta }}_{\mathit{p}2}\mathrm{\times }WEIGH{T}_i+{\mathrm{\beta }}_{\mathit{p}3}\\ \times TR{T}_i\times WEIGH{T}_i\mathrm{+}{\mathrm{\zeta }}_{\mathit{it}}.\end{array}$$

where the variable WEIGHT_i denotes the body weight for mouse i. Body weight across all study animals ranged from 31 to 45 g. The variable WEIGHT was centered at the mouse median weight (36 g), which is where the intercept was positioned.

Of interest in this study was the review of any predictive relationship between core temperature and surface temperature data measurements and their trajectories during the study phase. The core and surface temperatures for each mouse represent 2 related domains, thus a cross-domain model³⁹ was selected to evaluate the relationship among the growth model parameters by assessing the 2 domains simultaneously and estimating the correlation among all the parameters in the growth curves. Strong correlations (|ρ| ≥ 0.5) imply that the 2 yielded very similar information, whereas small values (|ρ| ≤ 0.1) indicate that the 2 types provide unique data.

To visualize the data, graphic figures were designed to show mean temperatures for each treatment group every 4 min. Error bars on the graphs identify the pointwise 95% confidence intervals for the mean temperature at each time point, based on the sampling distribution of sample mean (Student t distribution). When the 95% confidence interval for a treatment group did not overlap with the 95% confidence interval for the control group, the mean for this treatment group was significantly different from the mean of control group at the 0.05-level of significance.

Results

Control and experimental groups are discussed in the following sections, with changes in core (Figure 1) and surface (Figure 2) temperatures depicted graphically; average temperatures and body weights are summarized in Table 1. Estimates of coefficients in the model and SE, t, and P values were collated in tabular form (data not shown) and are instead summarized in the sections that follow. It was determined that the control animals’ starting temperature represented ‘normothermia’ for all mice under anesthesia, and the use of the term ‘normothermia’ refers to comparisons to control baseline temperatures throughout the Results sections.

Figure 1.

Core (rectal) temperatures showing changes across the baseline phase (minutes 0–3) and during removal of hair (minutes 4–7), application of scrub treatment (at minute 8; TX arrow), and through an additional 30 min of isoflurane anesthesia. Average temperature recordings with 95% confidence intervals are shown every 4 min. Temperatures in the control group (no-scrub treatment) are indicated by the red solid line, the uppermost line on the graph; *, significant (P < 0.05) difference from control. At time of first treatment application, core temperatures in the groups prepped with povidone–iodine and isopropyl alcohol (P-I/IPA) or povidone–iodine and saline (P-I/saline) were significantly cooler than the control group. At minutes 12 and 16, all treatment groups were significantly cooler than the control value. At minutes 20, 24, 28, 32, and 36, temperatures in all treatment groups except IPA were significantly lower than controls. W, warmed.

Figure 2.

Surface (infrared) temperatures showing changes across the baseline phase (minutes 0–3) and during removal of hair (minutes 4–7), application of scrub treatment (minute 8; TX arrow), and an additional 30 min of isoflurane anesthesia. The coldest average core temperature of 31.2 °C is highlighted (dashed blue horizontal line). Average temperature recordings with 95% confidence intervals are shown every 4 min; *, significant (P < 0.05) difference from control value (red solid line). At minute 8, all treatment groups except that prepped with povidone–iodine and warmed saline (P-I/W.Saline) were significantly cooler than the control group. At minute 12, all treatment groups were significantly cooler than the control group. At minute 16, all treatment groups except isopropyl alcohol (IPA) were significantly lower than the control. At time 20, only the groups P-I/Saline, P-I/W.Saline, and P-I were lower than the control. At minute 24, whereas the temperatures of P-I/Saline and P-I were significantly lower than the control, IPA is significantly higher. At minute 28, treatment groups P-I/Saline, P-I/W.Saline, and P-I were significantly cooler than the control value. Of note, at minutes 32 and 36, the IPA group was significantly warmer than the control surface temperatures.

Table 1.

Core (rectal) and surface (infrared) temperatures (°C; mean ± SD) and body weight (g) across experimental groups

	Core temperature	Surface temperature	Weight
Control	34.5 ± 0.48	31.1 ± 1.2	38.5 ± 3.8
IPA	33.9 ± 0.84	31.1 ± 1.7	39.2 ± 3.5
P-I only	33.0 ± 1.3	28.2 ± 1.9	34.5 ± 1.8
P-I/IPA	32.3 ± 1.3	29.2 ± 1.8	35.0 ± 3.5
P-I /saline	32.7 ± 1.3	28.6 ± 1.9	35.4 ± 1.4
P-I/warmed saline	32.6 ± 1.6	28.7 ± 1.9	37.4 ± 3.7

IPA, isopropyl alcohol; P-I, povidone–iodine

Means were calculated from treatment application (minute 8) until the last measurement was recorded (approximately minute 37).

Core temperature.

Control.

Mice to which no scrub treatment was applied exhibited a slight decline in core temperature throughout the experimental phase (Figure 1). The control group baseline showed a drop of 1.29 °C prior to other experimental groups receiving their respective scrub treatments. Between minutes 8 through 12 (the 4 min immediately during and after scrub–rinse application for experimental animals), the control group had the smallest change—a 0.18 °C loss. The average control core temperature was 34.50 ± 0.48 °C, with a temperature drop over the experimental time course of 2.025 °C. The linear rate-of-change showed a decline (β = –9.318, P = 0.001), without the dramatic drop that occurred in groups given scrub treatments. The acceleration component (β = 1.482, P = 0.012) was positive, indicating overall minimal loss of body heat over the entire experimental period; however through the course of the experiment, these animals did not rewarm after the slow cooling that occurred during anesthesia.

IPA.

Mice prepped with 70% IPA had a core temperature baseline change of 1.28 °C. The first 4 min after application resulted in a 1.75 °C drop, one of the largest changes at application across groups (Figure 1). IPA was the only agent that had no significant differences compared with core temps of controls within minutes after application through to the end of the experimental phase. The average IPA group core temperature was 33.9 ± 0.84 °C, similar to the control group average. The quadratic term (β = 10.608, P < 0.001) was positive, indicating that the IPA group showed a significantly faster recovery to the baseline temperature than the control group; furthermore IPA mice returned to normothermia more rapidly than the control group (β = –12.834, P < 0.001).

P-I only.

Core temperatures for the P-I only group dropped 1.25 °C during the baseline phase, with an additional decrease of 1.28 °C after scrub application. The core body temperatures continued to decrease for approximately 24 min under anesthesia, when temperatures began to rebound from the nadir (Figure 1). The overall drop in core temperature totaled 3.01 °C, with an average temperature of 33.0 ± 1.3 °C. The linear rate- of-change showed a more dramatic decline in temperature after administration of treatment (β = –21.616, P < 0.001) as well as faster recovery to the baseline temperature (β = 22.5, P < 0.001) than the control group.

P-I/IPA.

The P-I/IPA group had a core temperature baseline change of 1.24 °C, with a 1.76 °C drop within the first minutes after scrub application (Figure 1). The temperature began to rebound around 16 min after the treatment was applied. The lowest average temperature recorded was 31.28 °C. The average core temperature for P-I/IPA was 32.3 ± 1.28 °C. The linear rate-of-change in temperature after administration of treatment showed a dramatic decline (β = –17.945, P < 0.001) compared with temperature of the control group, and the P-I/IPA group had a faster recovery toward the baseline temperature (β = 27.948, P < 0.001).

P-I/saline.

The P-I/saline group experienced the smallest drop in core temperatures at the time of application (0.73 °C) and had a further 0.90 °C drop during the first 4 min after application. However, core temperatures continued to decrease in this group throughout the study period, to an overall drop of 4.09 °C. The average core temperature was 32.7 ± 1.32 °C, and the lowest average temperature was 31.4 °C. The linear rate-of-change in temperature after administration of treatment showed a more dramatic decline in temperature (β = –33.940, P < 0.001) than in the control group. The P-I/saline group displayed faster recovery to the baseline temperature than the control group (β = 13.553, P < 0.001) and was on a trajectory to achieve normothermia (β = –3.091, P < 0.001).

P-I/warmed saline.

The P-I/warmed (37 °C) saline group had a 1.27 °C drop soon after application. The core body temperature continued to decrease throughout the study period, with an overall temperature drop of 4.46 °C and the lowest temperature reading from the initiation to end of the study (31.26 °C) compared with all other groups. The average core temperature was 32.6 ±1.56 °C. This group showed a more dramatic decline in temperature (β = –42.521, P < 0.001) and a faster trajectory to baseline recovery (β = 23.672, P < 0.001) than the control group. The P-I/warmed saline group lost one mouse to follow up, such that the final analyses were performed on a group of 7 mice.

Effect of body weight on core temperatures.

Mouse body weight had significant effects on temperature responses in particular treatment groups: (1) heavier mice in the P-I group took longer (β = –0.872, P = 0.015) to recover from the initial drop of temperature after treatment application than the control group; (2) for the IPA group, heavier mice had a significantly faster (β = 0.842, P < 0.001) recovery from the initial drop in temperature than the control group; and (3) in the P-I/warmed saline group, heavier mice achieved normothermia more rapidly (β = –0.75, P < 0.001) than the control group.

Surface temperature.

Control.

Mice to which no scrub treatment was applied exhibited a slight increase in surface temperatures (0.533 °C) prior to anesthesia (Figure 2). Under anesthesia, the surface temperature dropped 2.27 °C. The average surface temperature for control animals was 31.1 ± 1.16 °C. The linear rate-of-change for surface heat loss was significant (β = –24.035, P = 0.022), but the rebound phase did not differ from other groups.

IPA.

The IPA group had a surface temperature drop of 5.12 °C on application. This group was the only one that rebounded to exceed control surface temperatures by 16 min after anesthesia was started and then remained warmer than control values (Figure 2). Overall, the average surface temperature for IPA was 31.1 ± 1.65 °C, which was virtually equivalent to the control average surface temperature (Table 1). The linear rate-of-change in temperature after administration of treatment showed a less dramatic decline in temperature (β = 49.272, P < 0.001) than the control group. Surface temperatures in the IPA group returned to normothermia more rapidly (β = –27.763, P < 0.001) than in control mice.

P-I only.

Surface temperatures dropped at application of P-I (4.74 °C) and then began to rebound and appeared to equilibrate to a final reading of 29.5 °C, similar to the controls (30.1 °C), by the end of the study. Overall, the surface temperature change from start to end in the P-I group was 0.04 °C. The average surface temperature was 28.2 ± 1.86 °C. The P-I group had a significantly faster recovery to baseline temperatures (β=59.084, P < 0.001) and was on trajectory to achieve normothermia (β = –40.389, P < 0.001) more rapidly than the control group.

P-I/IPA.

Similar to other treatment groups, the P-I/IPA group experienced a drop in surface temperatures at the time of application (5.90 °C); however, values rebounded to a final reading higher than controls. Overall, the surface temperature decreased 0.71 °C across the experimental time period, yielding an average surface temperature was 29.2 ± 1.8 °C. The P-I/IPA group displayed faster recovery to the baseline temperature (β = 28.798, P < 0.001) and achieved normothermia more rapidly (β = –36.629, P < 0.001) than the control group.

P-I/saline.

Surface temperatures in the P-I/saline group displayed the same trend as other saline groups, with a decrease in temperature at application (5.36 °C) and continued lowering of temperature throughout the study. Overall, the surface temperature dropped 2.89 °C, generating an average surface temperature of 28.6 ± 1.94 °C. The P-I/saline group displayed faster recovery to the baseline temperature (β = 44 .959, P < 0.001) and achieved normothermia more rapidly (β = –19.572, P < 0.001) than the control group.

P-I/warmed saline.

Surface temperatures for the P-I/warmed saline group did not decrease as markedly as in other groups, with only a 2.86 °C drop at time of scrub application. The average surface temperature was 28.7 ± 1.93 °C. The P-I/warmed saline group displayed faster recovery to the baseline temperature (β = 57.9, P < 0.001) and was on trajectory to achieve normothermia more rapidly (β = –26.928, P < 0.001) than the control group.

Effect of body weight on surface temperature.

As described for the effects on core temperatures, body weight influenced surface temperature in particular treatment groups: heavier mice from the IPA group (β = 3.913, P = 0.006), P-I/IPA group (β = 6.148, P < 0.001), and P-I/warmed saline group (β = 4.151, P = 0.004) recovered from the initial drop of temperature more rapidly than control mice. Furthermore, heavier animals in the P-I/saline group took longer for temperature recovery (β=5.291, P = 0.045) than the controls. For surface temperatures, the warming of saline may have helped to sustain surface temperatures in this group.

Cross-domain analysis.

The quadratic components for each temperature domain (core compared with surface temperature) were strongly and positively related (r = 0.56), indicating the acceleration in one type of temperature was reflected in measures of the other (R² ≈ 0.31). The linear rate-of-change in surface temperature showed a strong, positive relationship to the linear rate-of-change for core temperature (r = 0.77). Although average core and surface temperatures differed by several degrees across groups (Table 1), the trajectories followed similar paths (but could not be interchanged directly) through the course of the experiment.

Discussion

Antiseptic use for surgical procedures dates back to the 1800s, beginning with investigations on hand-washing and reduced surgical morbidity, followed by lowered rates of surgical site infections.¹² Alcohol is one of the most effective and fast-acting antiseptics, but it is rarely used as a single agent due to its lack of continued antimicrobial effects.^11,17,31 Alcohol, as part of the surgical scrub preparation protocol, has been reviewed in laboratory animal literature,^10,19,20 and its use is supported by AAALAC.¹ However, without clear evidence-based studies, alcohol continues to be variably recommended at contemporary institutions and is prohibited at others, and typically its use is accompanied by caveats for limited application prior to initiation of surgery to avoid exacerbation of hypothermia. Beyond animal research, there is a perceived absence of high-quality medical research to provide guidance on approaches for skin sterility and preparation, leaving clinicians anxious for evidence that better informs practice.⁴²

The described study was undertaken to quantitatively assess the effect of 70% isopropyl alcohol (IPA) and saline, in combination with the common scrub agent povidone–iodine (P-I), on core and surface temperatures in anesthetized mice. As predicted, skin application of IPA—both alone and in combination with P-I—resulted in the steepest drops in core and surface temperatures. Unexpectedly and significantly, both the core and surface temperatures in IPA-treated mice recovered within minutes after application, such that core temperature were near baseline levels and surface temperatures actually exceeded those of control animals by the end of the anesthetic study period. Because IPA in an aqueous solution does not leave a residue,²⁴ we predicted that, after evaporation, no additional fluid remained on the skin of IPA-treated mice to influence the temperature measurements. A similar phenomenon has been reported in conscious haired mice exposed to menthol, which returned to baseline temperatures approximately 15 min after treatment; the authors postulated that mice were able to mitigate the cold stress of menthol with autonomic and behavioral heat-gain responses.⁴¹

The mechanism by which the mice responded to IPA with rebounding core temperatures might involve (1) vasodilation of skin, with an increase in blood flow due to a peripheral site of action or (2) direct effects on central vasomotor control mechanisms after skin absorption. In support of the peripheral mechanism, the consumption of ethanol is thought to increase heat loss through the skin by increasing peripheral blood flow.^14,23 This response is inferred from available yet limited information from humans, wherein rubbing alcohol (IPA) has been used historically to attempt to reduce fevers; notably, this practice is now denounced.^33,40 The classic alcohol-bath treatment was believed to work temporarily, as skin would cool after the evaporation of the alcohol and leave a perception of relief. In reality, the physiologic response after the absorption of alcohol is to signal the body to raise internal temperatures (typically by shivering), often resulting in unwanted spikes in core and skin temperatures. IPA is quickly absorbed through human skin, and when used in a dousing situation to lower a fever, large amounts can be absorbed and can be inhaled inadvertently, resulting in alcohol poisoning and cardiac and neurologic problems.^25,26,33 The possibility of appreciable absorption of alcohol through skin preparation in animals should serve as a precaution to the veterinary community to review whether this practice, during which unregulated volumes are applied to skin, has the potential to lead to overdosing, neurologic sequellae, and potential anesthetic complications in laboratory animal species. Further studies are warranted to assess this potential complication of alcohol baths and topical application in laboratory animal medicine.

Regarding the perception that saline and alcohol are virtually interchangeable for scrub preparations,³ our results demonstrated that the mice in which P-I was wiped away with saline had prolonged hypothermic effects that did not rebound over the time course of anesthesia. In groups that were treated with P-I followed by either saline or IPA, the average core and surface temperatures remained the coldest over the experimental period (Table 1). We considered that, after the evaporation of IPA or with the removal of saline, residual moist P-I on the skin (potentially entrapped at the skin surface in haired areas) led to continued decreases in body temperature. Interestingly, there was no appreciable difference in stabilization or rebounding of core temperatures regardless of whether the saline was warmed to 37 °C prior to application or applied at ambient temperature to the skin surface. Given the chilling effect of P-I, further heat loss in surgery patients is likely once the peritoneal cavity is open to the room environment.

In this study, we used the practice standard of rectal thermometry for internal (core) temperature assessment and compared these data with those from noninvasive infrared scanning of the surgical prep site. We undertook this comparison to evaluate options for rodent thermometry, given that noninvasive methods of temperature detection in mice are variably effective in rodent research.^15,37 As we expected and as shown previously,¹⁵ overall surface temperatures were approximately 3 °C cooler than core measurements; however, there was high statistical correlation between changes in surface temperature and core temperatures. For those treatment groups that received IPA alone or in combination, one can appreciate the effect of the microvasodilatory responses at the skin level and subcutaneous heat release, resulting in the thermogenic effect that warmed mice to temperatures exceeding control group temperatures, as captured by surface thermography.

In addition, our study design enabled us to evaluate the effect of body weight in mice and potential hypothermic reactions to aseptic scrubs. Heavier mice had a tendency to experience a more rapid recovery of body temperature, although this effect varied by skin prep treatment. We surmise that this effect is attributable to the fact that heavier mice have a smaller surface area:volume ratio, potentially predisposing them to counter hypothermia while under anesthesia. On exposure to temperatures below their thermoneutral zone (29 to 31˚C), mice typically attempt to maintain body temperatures by energetically inexpensive means (for example, vasoconstriction, piloerection) and by changes in posture to decrease surface area.⁶ Nonshivering thermogenesis, modulated through brown adipose tissue, generates heat;⁵ therefore perhaps relative to lighter mice, heavier mice have more brown fat to assist with maintaining heat after cold exposure,⁶ such as the scrub–rinse treatments in the current study. Our mice had an average weight of 36 g as adult animals; therefore, the effects of scrub agents in younger (or lighter) mice might result in prolonged temperature changes that are more challenging to overcome after delivery of gas anesthesia.⁶ The mice treated with P-I only had the lowest average body weight but were not the animals for which the coldest temperatures were recorded; therefore, the combination of ambient environmental or room temperatures, scrub types, body weight, and background genotype must all be considered as potential influences on thermoregulation. Ultimately, with so many factors that might contribute to heat loss, providing supplemental thermal support, in addition to selection of scrub agents, should be prioritized to minimize hypothermia and maximize rodent surgery survival rates.

In conclusion, the application of alcohol (IPA) during surgical preparation resulted in the greatest cooling effect in mice at the time of application, with the greatest decreases in initial core and surface temperatures; however, this effect was offset within minutes by an unanticipated thermogenic rebound of temperatures to baseline (control) levels. We noted significant differences in the mouse perioperative body temperatures between saline- and IPA-based skin preparations; specifically, saline rinse alternating with scrub agents may represent a source of prolonged heat loss during rodent skin preparation protocols. Retention or pooling of these aqueous agents may be difficult to appreciate during surgical prep procedures but has been demonstrated to occur in certain species.³⁴ Our current data did not substantiate concerns that using alcohol-based rinse agents, as part of a tripicate skin prep protocol, were overly influential on hypothermia. Alcohol is not appropriate as a stand-alone agent for surgical skin preparation; however, alcohol is a viable alternative to saline-based rinses for surgical prep under inhalant anesthesia and its use should be encouraged within institutional guidance for rodent surgical procedures. Although all of the scrub agents tested caused temperatures to drop upon application, it was determined that over a 30-min anesthetic period, the inclusion of an IPA rinse during surgery preparation appears to be protective against prolonged hypothermia in mice.