Anesthetic agents affect urodynamic parameters and anesthetic depth at doses necessary to facilitate preclinical testing in felines

Urodynamic studies, used to understand bladder function, diagnose bladder disease, and develop treatments for dysfunctions, are ideally performed with awake subjects. However, in small and medium-sized animal models, anesthesia is often required for these procedures and can be a research confounder. This study compared the effects of select survival agents (dexmedetomidine, alfaxalone, and propofol) on urodynamic (Δpressure, bladder capacity, bladder compliance, non-voiding contractions, bladder pressure slopes) and anesthetic (change in heart rate [∆HR], average heart rate [HR], reflexes, induction/recovery times) parameters in repeated cystometrograms across five adult male cats. The urodynamic parameters under isoflurane and α-chloralose were also examined in terminal procedures for four cats. Δpressure was greatest with propofol, bladder capacity was highest with α-chloralose, non-voiding contractions were greatest with α-chloralose. Propofol and dexmedetomidine had the highest bladder pressure slopes during the initial and final portions of the cystometrograms respectively. Cats progressed to a deeper plane of anesthesia (lower HR, smaller ΔHR, decreased reflexes) under dexmedetomidine, compared to propofol and alfaxalone. Time to induction was shortest with propofol, and time to recovery was shortest with dexmedetomidine. These agent-specific differences in urodynamic and anesthetic parameters in cats will facilitate appropriate study-specific anesthetic choices.


Results
Five cats were anesthetized at least 3 times with each agent (see Supplementary Table S1 online). At least 2 CMG trials were conducted per session. We only collected terminal (isoflurane and α-chloralose) trials from four of five cats due to one unexpected death.
Non-voiding contraction rate was highest with α-chloralose (0.03 ± 0.005 NVCs/s) and significantly higher than all other agents (p < 0.01 for all comparisons; Fig. 1c, also see Supplementary Tables S2 and S3 online). Alfaxalone (0.01 ± 0.003 NVCs/s) had the second highest rate of NVCs, and was significantly greater than dexmedetomidine which had the lowest rate of NVCs (0.002 ± 0.003 NVCs/s, p = 0.008). The amplitude of non-voiding contractions was not different between agents (p > 0.05).
The trends of the CMG bladder pressure responses changed with agent, with different agents having variable periods of passive filling compared to active contraction over the trial. Notably, propofol CMG responses were characterized by an early and gradual rise (longer period of active contraction), compared to dexmedetomidine responses which were characterized by a long period of low pressure (passive contraction) and a rapid rise immediately before voiding (Fig. 2a, also see Supplementary Table S2 online). The bladder pressure trends for other agents fell between the propofol and dexmedetomidine extremes. This was further described through quantification of slopes at select time points (Fig. 2b). During slope 1 (infusion start to 100 s before void), propofol (0.06 ± 0.01 cm H 2 O/s) was significantly increased compared to alfaxalone (0.03 ± 0.01 cm H 2 O/s, p < 0.0001) and dexmedetomidine (0.02 ± 0.01 cm H 2 O/s, p < 0.0001). During slope 2 (100 to 50 s before void), α-chloralose had the highest slope (0.5 ± 0.08 cm H 2 O/s), significantly higher than dexmedetomidine (0.2 ± 0.05 cm H 2 O/s, p = 0.0004) and alfaxalone (0.3 ± 0.05 cm H 2 O/s, p = 0.03). During slope 3 (last 50 s before void) dexmedetomidine (1.5 ± 0.1 cm H 2 O/s) was significantly greater than all other agents: alfaxalone ( Anesthetic parameters. Continuous HR readings from select survival sessions averaged by agent visually demonstrate that mean HR and mean ΔHR varied by agent (Fig. 3a). Two trials (both propofol) were excluded from the averaged continuous HR plot due to suspect artifacts.
Mean HR and mean ΔHR taken from all trials confirmed variation between agents (Fig. 3b, also see Supplementary Tables S4 and S5 online). Mean HR, compared across all five agents, was highest under alfaxalone (215.2 ± 5.7 bmp), and significantly greater than all other agents: α-chloralose (163.3 ± 9.5 bmp, p < 0.0001), dexmedetomidine (110.7 ± 5.8 bpm, p < 0.0001), isoflurane (130.7 ± 8.2 bmp, p < 0.0001), and propofol (176.7 ± 5.8 bmp, p < 0.0001). All other agents fell in between the two extremes. While there was some animal-specific www.nature.com/scientificreports/ variation, across animals the mean HR maintained these trends (see Supplementary Fig. S2 online). Mean change in heart rate (ΔHR), compared across only non-terminal agents, was lowest with dexmedetomidine (13.8 ± 6.8 bmp), and significantly different from propofol (37.5 ± 7.1 bmp, p = 0.008) and alfaxalone (45.4 ± 6.7 bmp, p < 0.0001). Mean ΔHR did not differ significantly between propofol and alfaxalone (p = 0.8). Time to lateral recumbency and time to head up and walking differed significantly between the non-terminal agents (see Supplementary Tables S5 and S6 online). The time to lateral recumbency was shortest under propofol due to IV administration (2.4 ± 1.6 min), and significantly shorter than alfaxalone (9.5 ± 1.5 min, p = 0.009). The time to head up and walking were shortest under dexmedetomidine (1.7 ± 6.1, 2.5 ± 10.6 min) due to use of the reversal agent atipamezole compared to alfaxalone ( (Fig. 3c). Despite variation between individual cats (see Supplementary Fig. S3 online), reflexes showed greater anesthetic depth under dexmedetomidine compared to alfaxalone and propofol, although jaw tone was tighter under dexmedetomidine (Fig. 4). In general, average HR was influenced by both animal and session, and ΔHR was influenced only by animal (see Supplementary Adverse consequences. Adverse consequences associated with VAP placement included surgical dehiscence (1/5) and port failure (1/5). In the cat with port failure, the last propofol trial was performed through isoflurane induction, in order to facilitate IV catheter placement. Dexmedetomidine caused vomiting immediately  Tips of whiskers extending below and above box represent 1st quartile -1.5*IQR and 3rd quartile + 1.5*IQR respectively. Significance: *p < 0.05, **p < 0.01. www.nature.com/scientificreports/ after injection in 7/15 (46.7%) of sessions. Cats did not vomit under other agents. Each cat vomited at least once under dexmedetomidine. One cat had an independent incidence of lethargy and vomiting unassociated with an anesthetic event, and was diagnosed with disseminated lymphoma at necropsy. One cat displayed significant bradycardia with 2nd degree atrioventricular (AV) block arrhythmia under dexmedetomidine in 2 separate anesthetized events. Echocardiogram and subsequent necropsy of this cat suggested evidence of left ventricular diastolic dysfunction and myofibril degeneration and atrophy. Finally, 1/13 (7.7%) alfaxalone trials resulted in unexpected death. During recovery, after the constant rate infusion (CRI) had been discontinued, one cat became apneic and was not able to be resuscitated.

Discussion
In this study we demonstrated differences in urodynamic parameters and anesthetic depth at minimal doses needed to perform CMG testing in male cats. To our knowledge, this was the first study to examine CMG parameters under alfaxalone, propofol in isolation without another agent, or dexmedetomidine in comparison to the other agents used here. Our study incorporates a novel method of CMG assessment (bladder pressure slopes) in conjunction with traditional CMG parameters and careful evaluation of anesthetic depth. Together, our results examine the interaction between urodynamic parameters and anesthetic depth, and provide recommendations for anesthetic choice based on study parameters of interest and logistic feasibility for CMG studies in male cats. www.nature.com/scientificreports/ Due to commonalities between species in response to anesthetic agents, select information from this study may be able to be guide preliminary anesthetic choice to other species, including humans. However, feasibility and effect on urodynamic and anesthetic parameters should be confirmed for the species in question. Bladder capacity increases due to inhibition of central nervous system reflexes under anesthesia 3 . In this study, the bladder increases seen under terminal agents (isoflurane and α-chloralose) (Fig. 1a) were strongly correlated with prolonged anesthesia (isoflurane [2-6 h before CMG trials, 8-10 h total], followed by α-chloralose [12-22 h before CMG trials]) and repeated bladder fills for other experimental objectives. Although we did not see significant differences between survival agents, dexmedetomidine has been reported to cause polyuria through blockage of arginine-vasopressin release 20,21 .
Maximum detrusor pressure [Pdet(max)] can also be used to assess bladder function 18 . We reported Δpressure to account for any positioning differences between trials that may have affected baseline pressures. Consistent with previous studies examining Pdet(max) 2,9 , we found that Δpressure was higher with propofol, and lower with isoflurane ( Fig. 1b). While this may be attributed to the lighter plane of anesthesia under propofol (general anesthetics depresses Pdet(max) in a dose-dependent manner 2,9 ), a human study showed that even under comparable anesthetic depth, propofol showed less suppression of cortical somatosensory evoked potentials compared to isoflurane 22 . Alternatively, the increase in Δpressure may be related to loss of function and initiation of the voiding reflex under propofol, as evidenced by bladder slopes drastically different from the awake state. While the results of propofol are difficult to interpret, isoflurane should be avoided in studies where Pdet(max) or Δpressure is a key parameter of interest. Our Δpressure values trended slightly higher than previously reported values both in cats under α-chloralose 5,16,23 , isoflurane 9 , propofol 9 , and dexmedetomidine 16 . Individual animal variation or experimental logistics could explain these trends. The placement of a transurethral catheter is likely a primary factor in this difference, as it provided a partial blockage of the bladder outlet. Transurethral catheters have been documented to increase voiding pressure in rats compared to suprapubic catheters 24 . Our catheterization approach was consistent across all agents, so relative pressure trends are generalizable.
Non-voiding contractions are phasic pressure increases seen during bladder filling 25 that may be linked to sensations relating to bladder volume 26 . They are more likely and more frequently occur with urinary bladder pathologies 25,27,28 . Consistent with our understanding that α-chloralose maintains spinal reflexes relative to other anesthetic agents 5 , the greatest number of NVCs were observed with α-chloralose (Fig. 1c). Alfaxalone trials had the next greatest number of NVCs, and the most out of the non-terminal agents studied. Notably, dexmedetomidine almost completely eliminated NVCs compared to other agents. The α 2 mediated analgesic properties of dexmedetomidine may block the sensation of bladder filling at the spinal or supra-spinal level 19,29 . Clinically, dexmedetomidine has been used for analgesia and prevention of post-operative catheter-related bladder discomfort 30 . However, awake CMG trials have significantly less NVCs than anesthetized trials 16 , so absence of NVCs is not necessarily pathological. When NVCs are desired (ex: overactive bladder models), α-chloralose (non-survival) and alfaxalone (survival) can be considered, and dexmedetomidine should be avoided.  www.nature.com/scientificreports/ In this study, CMG bladder pressure slopes differed between agents (Fig. 2), interpreted as previously described variable periods of passive filling and active contraction 31 . Pressure slopes under dexmedetomidine (characterized by a flat filling period, followed by a sharp rise to the void event) are most similar to awake or decerebrate animal cystometrograms 5,10,32,33 . In contrast, pressure slopes under propofol had the most obvious change from the awake slope, characterized by an early rapid rise and less obvious void event. Although there are no standardized criteria for defining passive filling and active contraction, we found propofol consistently reached a higher normalized pressure compared to dexmedetomidine across normalized time, with the most obvious separation noted between 900 and 950 normalized time, where the mean normalized pressure ± standard deviation range of dexmedetomidine is distinctly separate from the corresponding range for propofol (Fig. 2a). These findings suggest dexmedetomidine was associated with more awake-like bladder responses compared to GABA-ergic agonists.
Anesthetic depth was increased under alfaxalone and propofol compared to dexmedetomidine as demonstrated by reflexes, HR, and ΔHR. The rise in bladder pressure was closely associated with a rise in heart rate for propofol and alfaxalone (around 200 normalized time units in both measures for both agents). This phenomenon is likely driven by sympathetic activity in response to urinary bladder distension, as previously described in cats by the viscerovascular reflex 34 , and was notably not observed in the dexmedetomidine trials.
The exact significance of bladder pressure slopes and their association with anesthetic depth is unclear. In a previous rat study comparing varying doses across a single anesthetic agent (tiletamine-zolazepam), lighter anesthesia was associated with more awake-like bladder curves 32 . In this experiment, comparing across anesthetic agents, we found the opposite, likely due to differences in mechanism of action between agents. We recommend further investigation of bladder pressure response trends in future studies in association to anesthetic depth, pharmacokinetics, preservation of spinal reflexes, and potential bladder pathology.
In addition to affecting CMG and urodynamic parameters, each agent offers unique logistical benefits and constraints. Dexmedetomidine (α 2 agonist), conveniently administered through IM boluses (CRI only recommended for analgesia, not for sedation or anesthesia 19 ) and rapidly reversed through atipamezole, provides a convenient method of anesthetizing cats. Due to the nature of IM bolus administration, depth of anesthesia and therefore bladder function can vary as plasma levels of dexmedetomidine increase or decrease. Common adverse effects of dexmedetomidine include bradycardia, arrhythmias (AV block), and vomiting 19 , all of which were noticed in our experiments.
Propofol and alfaxalone (GABA A agonistics) are used clinically to both induce (IV) and maintain (CRI or repeat IV bolus) anesthesia 15 . Maintenance of anesthesia using a CRI allows for steady-state drug plasma levels and a more stable plane of anesthesia. While propofol is limited to IV administration, alfaxalone can also be given IM for single-agent anesthesia. Common side effects of both propofol and alfaxalone are apnea, although one study showed alfaxalone had less respiratory depression compared to propofol 35 . Unexpectedly, we experienced a death during the recovery phase (most common time of anesthetic-related death 36 ) of one alfaxalone trial. To our knowledge, this was an isolated incident and not representative of the general safety of alfaxalone.
In this study, we were able to decease concentrations of GABA-ergic agonists below recommended labelled dosing in part due to the stable plane of anesthesia associated with CRI administration. This could be one explanation for the lighter plane of anesthesia (more preserved reflexes, higher average heart rate, and greater change in heart rate over trial) observed in GABA-ergic agonistics compared to dexmedetomidine (Figs. 3 and  4). However, another concurrent explanation for increased heart rate under the GABA-ergic agonists is hypotension with reflex tachycardia 37 . This is in contrast with dexmedetomidine, which causes vasoconstriction, leading to hypertension and reflex bradycardia 19 . Consistent with previous studies in dogs showing alfaxalone causes less cardiovascular depression than propofol 37 , we observed the mean heart rate under alfaxalone (at or slightly above the upper limit of awake reference range 38 ) was significantly higher than under propofol (Fig. 3).
Isoflurane (exact mechanism of action unknown 19,39 , but suspected to affect GABA receptors 40 ) is the primary inhalant anesthetic used in veterinary medicine, especially in prolonged invasive procedures 19 . Although in this study it was only used during terminal procedures, in clinical veterinary medicine isoflurane is commonly used during recovery procedures. Its inhaled route of administration allows dosing to be easily titratable to adjust the depth of anesthesia. However, inhaled agents are logistically more challenging, either requiring technical expertise to intubate the animal, or risking personnel exposure. Since recovery from isoflurane is rapid, and isoflurane has been previously used to facilitate propofol administration for urodynamic procedures 41 , the effects of isoflurane on subsequent α-chloralose and propofol trials in this study are assumed to be minimal.
The non-survival hypnotic α-chloralose (mechanism unknown) provides light long-lasting anesthesia with minimal analgesia, and minimal effect on normal spinal and sympathetic reflex function 5,15 . Unfortunately, α-chloralose is not recommended for survival studies due to involuntary excitement and prolonged recovery 15 although several studies have successfully recovered animals following α-chloralose anesthesia 42,43 . In this study, isoflurane and α-chloralose were only used in terminal surgeries, and thus had fewer test sessions than for the other agents. Given this relative disparity, results from isoflurane and α-chloralose should be interpreted with caution.
The independent effects of animal, session, and trial number of session are confounders to this study (see Supplementary Fig. S1 and Supplementary Table S7 online). Animal-specific changes in compliance and slope are linked to animal-specific changes in bladder capacity and Δpressure from which compliance and slope are calculated. The differences in slope and average HR over sessions can be attributed to the specific agents tested during that session, rather than an independent session-specific effect. The variation in NVCs, Δpressure, and slope may be attributable to changes in plane of anesthesia during the experiment. Although differences in urodynamic and anesthetic parameters were found between the examined variables, the mixed model analysis, as well as similar number of trials and sessions per agent and per animal control for this variation. www.nature.com/scientificreports/ Limitations of this study include the following: (1) lack of direct comparison to awake CMG trials, (2) use of transurethral catheterization, which may cause obstructive changes to voiding response 24 , (3) inability for full randomization due to logistical constraints (drug availability, ability to secure IV access, etc.), although effort was made to spread out agents, so that each agent was tested once within the first 3 trials, once within the next 3 trials, and once within the final 3 trials, (4) inability to extrapolate data to female cats, (5) limited data for terminal agents (isoflurane and α-chloralose) due to nature of the experiments and the unexpected death of one animal, (6) possible influence of previously administered agents on future trials (especially analgesics which may decrease the sensation of bladder filling), both between and within sessions, especially during the terminal studies, (7) confounding effects of prolonged anesthesia, repeat CMG trials on subsequent CMG trials, and an invasive laminectomy for terminal studies, (8) lack of blinding in data analysis and experiments, especially for qualitative measures such as anesthetic reflexes, (9) lack of monitoring data other than heart rate due to measurement artifact and animal movement during trials, (10) inability to control for different doses and routes of administration of a single drug due to logistical considerations and sample size, and (11) confounding effects of animal-specific and anesthetic-specific factors, though attempts were made to control this variability using a mixed model statistical analysis.
Despite these limitations, this study compares and contrasts urodynamic and anesthetic parameters between survival agents, including a novel agent alfaxalone (previously unstudied in cats in the context of urodynamic parameters), and gold standard terminal agents. The differences identified between agents in this study can help (1) interpret the results of existing urodynamic studies by differentiating unintended anesthetic artifact from intentional experimental manipulation, and (2) guide selection of anesthetic agents based on logistic or experimental considerations. Future studies to extend this work and address the limitations of this study include additional routes and doses of single agents, concurrent CMG in awake cats for direct comparison to anesthetized cats, observing bladder contractions outside of the active fill period, and filling of ex-vivo or post-mortem bladders to provide a truly passive bladder control.

Animals.
Five adult, intact male domestic shorthair cats (7-18 months, 4-6 kg, Marshall BioResources, Inc., North Rose, NY) were included in this study. These animals were available from another study, and the cohort size is similar to other relevant studies 2,9,44 . Animals were group housed in a temperature (72 ± 2 °F) and humidity (30-70%) controlled room on a 12:12 h light/dark cycle with ad libitum access to water and fed a complete diet for adult cats (Purina One Urinary Tract Health Formula and Purina ProPlan EN, Nestle Purina, Neenah, WI). Cats were free of feline herpesvirus, calicivirus, panleukopenia, coronavirus, feline immunodeficiency virus, chlamydia, and toxoplasmosis. All experimental procedures were performed at an AAALAC accredited institution, according to the standards established by the Guide for the Care and Use of Laboratory Animals, and were approved by the University of Michigan Institutional Animal Care and Use Committee. Vascular access port. Each cat was implanted with a VAP (ClearPort Mid VAPs 5Fr with a rounded tip, 24″ catheter; Access Technologies, Skokie, IL) to guarantee immediate IV access. IV access was essential for avoiding a second agent in propofol experiments, and useful in providing a constant rate infusion (CRI) in the alfaxalone and propofol experiments to maintain anesthesia. Cats were induced with dexmedetomidine (0.02 mg/kg, Dexmedseed, Dechra, Overland Park, KS), ketamine (6.6 mg/kg, Zetamine, Vet One, Boise, ID), and butorphanol (0.66 mg/kg, Torbugesic, Zoetis, Kalamazoo, MI), and maintained on isoflurane (0-2.5%, Fluriso, Vet One, Boise, ID). Incisions were made on the left ventral neck and dorsal neck or back. A catheter was inserted and secured in the left jugular vein, and tunneled to a subcutaneous port secured to the muscle between the shoulder blades 45,46 . The port was accessed with a non-coring Huber needle (Access Technologies, Skokie, IL), and flushed regularly (weekly, or after each use) with a locking solution (TCS Lock Solution, Access Technologies, Skokie, IL) or heparinized saline. Post-operative medications (buprenorphine [0.01-0.02 mg/kg, Par Pharmaceutical, Chestnut Ridge, NY], robenacoxib [6 mg/kg, Onsior, Elanco, Greenfield, IN] and famotidine [0.5 mg/kg, Westward, Eatontown, NJ]) were given for 24 h post-operatively. The cats were monitored daily until suture removal 7-14 days later.
Anesthesia. Each cat was anesthetized at least three times ("sessions") with each of three agents: (1) dexmedetomidine (0.02-0.04 mg/kg IM bolus; reversed with matched volume of atipamezole [Antisedan, Orion Corporation, Espoo, Finland] after session), (2) alfaxalone (5 mg/kg IM bolus + 0.08 mg/kg/min IV CRI, Alfaxan Multidose, Jurox Inc. Kansas City, MO), and (3) propofol (2 mg/kg IV bolus + 0.15 -0.20 mg/kg/min IV CRI, Hospira, Inc., Lake Forest, IL). Effort was made to randomize sessions (each agent was tested once within the first 3 sessions, once within the next 3 sessions, and once within the final 3 sessions), but full randomization and consistency among animals could not be achieved due to logistical constraints (agent availability, repeat sessions, etc.). Each session was on a different day, at least 2 days after the previous anesthetized session. Sessions ranged from 2-5 h in length, with CRIs running continuously during alfaxalone and propofol sessions, and intermittent IM bolus injections given to maintain anesthesia during dexmedetomidine sessions. Occasionally, 1 mL IV boluses of alfaxalone and propofol were given in addition to the CRI to keep the cats at an appropriate plane of anesthesia.
During terminal procedures in conjunction with an unrelated research study, cats were anesthetized with the same dexmedetomidine/ketamine/butorphanol combination as for the VAP surgery, and transitioned to isoflurane. Surgical procedures for the other study were performed under isoflurane (2-4 h)  Anesthetic parameters Urodynamic parameters T-100 Figure 5. Experimental overview. Cats were implanted with VAPs and anesthetized at the lowest doses necessary to facilitate urodynamic testing. At least 3 sessions of propofol, dexmedetomidine, and alfaxalone each were conducted, in addition to a terminal experiment under isoflurane and α-chloralose in conjunction with another study. During anesthetized sessions at least 2 CMG trials were performed in each session to assess urodynamic parameters. Heart rate and reflexes were used to assess anesthetic depth, and time to induction/ recovery was noted.  (Fig. 5). The bladder was emptied and a break of at least 10 min was given between trials to allow the bladder to relax. Bladder capacity volume was calculated using two methods: (1) the total volume infused and (2) the sum of urine collected from the bladder and collecting tray. The greater of these two values was used in subsequent analysis. The variables analyzed for each CMG were bladder capacity, change in pressure from baseline to maximum (Δpressure), compliance (capacity/Δpressure), number and amplitude of non-voiding contractions (NVCs), and slopes of the bladder pressure ("pressure slopes") over three intervals as defined below.
Anesthetic parameters. The time to lateral recumbency after initial agent injection and times to head up and walking (ten consecutive steps without falling) were recorded for each anesthetic session with non-terminal agents. During each session, vitals (heart rate [electrocardiogram or pulse oximetry], O 2 hemoglobin saturation (pulse oximetry; tongue, lip, or ear), non-invasive blood pressure [forelimb] and temperature [rectal]) were monitored once every 15 min using a vitals monitor (Surgivet, Smiths Medical, Minneapolis, MN). Respiratory rate was taken manually. The average heart rate (mean HR) for a trial appeared to be a consistent indicator of anesthetic plane and was later included in statistical analysis. Other vitals either did not fluctuate across sessions or were subject to artifact and were not analyzed across sessions. During non-terminal procedures for three of the five cats, a data logger was used for continuous vitals recording once per second and the calculation of change in heart rate over trial (ΔHR). Reflexes (palpebral, withdrawal, jaw tone, pupil dilation) were checked before each CMG trial. Palpebral and withdrawal reflexes were categorized as "yes (strong withdrawal with paw pinch)", "some (weak withdrawal with paw pinch)" or "no (no withdrawal, unresponsive to paw pinch)". Jaw tone was categorized as "tight (significant pressure required to open jaw), " "mid (intermediate pressure required to open jaw)", or "loose (lack of muscle tone, jaw opened easily)". Pupil dilation was categorized as "constricted, " "mid (intermediate)", or "dilated". In general, as anesthetic depth increases, palpebral and withdrawal reflexes decrease, jaw tone loosens, and pupils become more dilated 48 .
Data acquisition and analysis. Bladder pressure data was processed using MATLAB (R2019a, Math-Works, Natick, MA) and Excel (Microsoft, Redmond, WA). Bladder pressure was smoothed using a 1 s moving average. NVCs were identified in MATLAB using the findpeaks function, defining a peak as at least 3 cm H 2 O, at least 0.5 s away from another peak, and then manually confirmed. To account for the range of trial lengths, NVC counts were divided by the total length of the trial to calculate NVCs per second. To quantify overall bladder pressure trends seen between agents, pressure slopes were calculated for three distinct time periods: (1) infusion start to 100 s before void, (2) 100 to 50 s before void, and (3) last 50 s before void. To visualize the changes seen during slope calculations over different pressure and time ranges across cats and test sessions, the bladder pressure was normalized from 0 (start) to peak/void (1,000) on the x (time) axis. After noting that pressures did not significantly fluctuate from resting baseline pressure until about 675 normalized time units, the 675-1,000 normalized time unit range was targeted, and the pressure was normalized from 0 to 1 within that window. Continuous heart rate readings during non-terminal trials for the last 3 cats were also processed similarly, but over the entire length of the trial. Statistics were analyzed using JMP Pro 14.2.0 (SAS Institution Inc. Cary, NC), using a mixed model approach, setting anesthetic agent as a fixed variable and animal as a random variable. For analysis of bladder capacity, the cat weight was set as an additional fixed variable. Nominal categorical data was converted to ordinal categorical data in order to fit the mixed model, for reflex data. Yes/some/no, tight/mid/loose, and constricted/mid/dilated were converted to 1/0.5/0 respectively. Pairwise comparisons were made using Tukey's HSD to examine the effect of anesthetic agents on the bladder. Pairwise comparisons using a Steel-Dwass test (non-parametric version of Tukey's HSD) was used to determine the effect of confounding factors (session, trial number of session, and animal) on urodynamic and anesthetic parameters, using only survival agents, the first 9 sessions of each animal, and first 3 trials of each session. Significance was defined as p < 0.05. Where relevant, data is reported as mean ± standard deviation.

Data availability
The datasets generated and analyzed during the current study and the MATLAB code used for data analysis are available in the Open Science Framework repository, (https ://osf.io/8zjkp / [https ://doi.org/10.17605 /OSF. IO/8ZJKP ]).