Materials and Methods

Animals—Fifteen conditioned mixed-breed adult dogs were obtained from the central animal facility for use in this study. Each group had a mean body weight of 26 kg (range, 20 to 30 kg). Dogs were 0.7 to 3.5 years of age and were sexually intact; 5 were female, and 10 were male. All dogs were determined to be healthy on the basis of results of physical examination, CBC determination, serum biochemical analysis, and fecal analysis. The institutional animal care committee approved the experimental protocol, and dogs were cared for according to the Canadian Council for Animal Care and Use Guidelines.

Experimental protocol—Before the beginning of each procedure, food but not water was withheld for 12 hours. Dogs were randomly assigned to receive open surgical J-tube placement under general anesthesia (n = 5 dogs; group 1), laparoscopic-assisted J-tube placement under general anesthesia (5; group 2), or laparoscopic-assisted J-tube placement under CRI sedation with epidural and local anesthesia (5; group 3). In the dogs receiving general anesthesia (group 1 and 2 dogs), the surgical area was clipped prior to transferring to the surgical suite for either a ventral midline approach (group 1 dogs) or a left lateral approach (group 2 dogs). All dogs received 1 dose of cefoxitin (22 mg/kg, IV) at least 20 minutes prior to the first skin incision.

Anesthetic protocol (groups 1 and 2)—Group 1 and 2 dogs were premedicated with hydromorphone (0.1 mg/kg, IM) and glycopyrrolate (0.01 mg/kg, IM). A 20-gauge catheter was placed into the cephalic vein, and balanced electrolyte fluids were administered IV at 10 mL/kg/h. Anesthesia was induced with ketamine (7.5 mg/kg, IV) and diazepam (0.35 mg/kg, IV) at a 1:1 mixture, given to effect. Dogs were intubated with an appropriately sized cuffed endotracheal tube, placed in lateral recumbency, connected via the endotracheal tube to a semiclosed circle rebreathing system, and maintained on isoflurane (minimum alveolar concentration of 1.2) in 100% O₂ at 1.5 L/min. Positive-pressure ventilation was initiated at 10 breaths/min at a tidal volume of 15 mL/kg with adjustments as needed to achieve normocapnia (PETCO₂ of 35 to 45 mm Hg) throughout the procedure. Body temperature was maintained throughout the procedure with an electric recirculating water heating pad.

Epidural catheter placement and anesthetic protocol (group 3)—On the day prior to the scheduled laparoscopic-assisted J-tube placement, group 3 dogs were mask induced with isoflurane in a nitrous oxide–oxygen mixture, orotracheally intubated with an appropriately sized cuffed endotracheal tube, connected via the endotracheal tube to a semiclosed circle rebreathing system for instrumentation, and placed in sternal or lateral recumbency for epidural catheter placement. A cephalic IV catheter was placed and heparinized. A 19-gauge epidural catheter^a was placed in the lumbosacral space through palpation of the spinous processes by use of the loss-of-resistance technique.³⁴ A 17-gauge, 8.9-cm Tuohy needle was used and advanced approximately 4 to 5 cm in a cranial direction. Correct catheter position was verified in the epidural space between L4 and L5 by injecting 1.5 to 2.0 mL of iohexol^b into the catheter under fluoroscopy. Epidural catheters were flushed with 1 to 2 mL of preservative-free heparin (1,000 U/10 mL) to prevent clotting, sutured in place, and covered with an adhesive bandage to prevent contamination or inadvertent removal. To prevent stress associated with shaving the surgical site in conscious dogs, the left flank of each dog was clipped under anesthesia in preparation for laparoscopic-assisted J-tube placement. Dogs were recovered from inhalant anesthesia with an Elizabethan collar in place to prevent IV and epidural catheter removal.

On the morning of surgery, prior to sedation, epidural catheter position was again confirmed by use of fluoroscopy and dogs were premedicated with glycopyrrolate (0.01 mg/kg, IM). The cephalic catheter was flushed with heparinized saline (0.9% NaCl) solution to ensure patency, and balanced electrolyte fluids were administered IV at 10 mL/kg/h. Sedation was achieved with a bolus of fentanyl (5 μg/kg, IV) and midazolam (0.1 mg/kg, IV), followed by a CRI of fentanyl (30 to 60 μg/kg/h, IV) and midazolam (0.1 to 0.25 mg/kg/h, IV), titrated to effect. Doses were adjusted to the level of sedation that permitted dogs to lie in right lateral recumbency with minimal manual restraint. Cotton balls were placed within the external ear canals to minimize auditory stimulation. An oxygen mask delivering 100% O₂ was placed in the area of the dog's muzzle, as much as was tolerated, and dogs were allowed to breathe spontaneously. Immediately prior to initiating laparoscopic-assisted J-tube placement, preservative-free 2% lidocaine (4 mg/kg, maximum 6 mL) was injected into the epidural catheter over a 5-minute period and flushed with 2 mL of sterile saline solution.

Instrumentation and measurements—A multichannel patient monitor^c was used to monitor airway gases (group 1 and 2 dogs) and all physiologic parameters (group 1, 2, and 3 dogs). Prior to each experiment, all pressure transducers and pressure channels were calibrated with a mercury manometer and zero was set at the midsagittal level. The airway gas monitor was calibrated with a commercial gas mixture^d prior to each experiment. An endotracheal sample collection line was placed at the level of the carina for measurement of PETCO₂ via side-stream infrared absorption spectrometry at an aspiration rate of 200 mL/min. Heart rate and rhythm were monitored by use of a standard lead II ECG, and body temperature was monitored via an esophageal temperature probe (group 1 and 2 dogs).

A dorsal pedal artery was catheterized percutaneously with a 20-gauge, 4.8-cm catheter for determination of direct arterial blood pressures, arterial blood gas sample collection, and LiDCO.^35,36,e Blood samples were assessed for Hb and electrolyte concentrations and temperaturecorrected blood gas tensions.^f Preparation of the lithium sensor as well as assembly of the inlet port, flow regulator pump, and outlet port between the arterial catheter and the waste collection system were performed as described in the LiDCO operation manual.^g The cephalic vein catheter was connected to a 3-way stopcock allowing fluid administration by 1 port and to an extension tube connected to another port for the purpose of preloading the required lithium chloride in advance of CO measurements. A lithium chloride^h dose of 0.15 mmol (1 mL) was used for all LiDCO measurements throughout this study with a sensor constant of 10.5. A cephalic vein was used for the injection of the lithium sensor, as previously validated in dogs.³⁷ For each CO measurement, 2 mL of arterial blood was aspirated anaerobically for immediate blood gas tension analysis and sodium and Hb concentration measurements. The required values of sodium and Hb concentrations were entered into the LiDCO monitor. Lithium chloride was preloaded into the extension tube. Once a stable baseline was achieved on the LiDCO monitor, the ventilator was arrested at end expiration and 6 mL of saline solution was used to flush the lithium chloride into the vein.

For all dog groups, HR and respiratory rate, blood gas tension, serum electrolyte concentration, MAP, CO, and body temperature were measured at each time point; in addition, ETISO, PETCO₂, and dialed isoflurane percentage were measured in group 1 and 2 dogs. Following either sedation or induction of general anesthesia, dogs were instrumented and then stabilized for 20 minutes before baseline measurements were performed (time 0). Following baseline measurements, surgical draping and open surgical or laparoscopic-assisted J-tube placement were performed. Cardiovascular measurements were performed 5, 10, 20, and 30 minutes following the beginning of the open surgical or laparoscopic procedure (procedure period) and at 5 and 10 minutes after the procedure (after desufflation for group 2 and 3 dogs and after the start of abdominal closure for group 1 dogs).

General anesthesia (group 1 and 2 dogs) or sedation (group 3 dogs) scores were based on a scale of 0 to 3, and abdominal insufflation pressures (group 2 and 3 dogs) were recorded at each time point. General anesthesia was described as follows: inadequate, 0; mild anesthetic depth, 1; moderate anesthetic depth, 2; or profound anesthetic depth, 3. Sedation was described as follows: inadequate, 0; mild, 1; moderate, 2; and profound, 3. Duration of anesthesia and anesthetic recovery scores were also calculated.

Calculations—Stroke volume (L/heart beat) was calculated by dividing CO (L/min) by HR (beats/min) for each time point.³⁸ Cardiac index (L/min/m²) was calculated by dividing CO (L/min) by each dog's body mass (m²) converted from body weight (kg).³⁹ Oxygen delivery was calculated by use of the following formula:

View Expanded

Surgical protocol—Group 1 dogs were positioned in dorsal recumbency for a standard ventral midline approach with a 15-cm-long incision centered over the umbilicus. Open surgical J-tube placement was performed with an 8-F feeding tube inserted into the proximal portion of the jejunum through a stab incision in the left flank. A 4-point jejunopexy was performed, and the incision was closed routinely. Group 2 and 3 dogs were positioned in right lateral recumbency. Four sites were marked with a sterile pen (central [midflank]; cranioventral [caudal to the last rib]; caudoventral [cranial to the pubis]; and dorsal [midflank, ventral to the epaxial muscles]). Each site was injected with approximately 0.5 mL (10 mg) of 2% lidocaine 5 minutes prior to the insertion of the Veress needle and trocar-cannula units. The abdomen was insufflated with CO₂ to maintain IAPs between 10 and 15 mm Hg. Desufflation occurred once the jejunum was exteriorized through the caudal portal incision. Standard laparoscopic-assisted J-tube placement was performed with an 8-F feeding tube.^40,41 Closure was routine.

Postoperative procedures—Dogs recovered from anesthesia and were treated with acepromazine (0.03 mg/kg, IV) as needed for anxiety after anesthetic recovery. An investigator (SAH) and assistant who were aware of the procedure performed assigned postoperative pain scores by use of a previously described 0 to 10 scoring system⁴² at 30-minute intervals in the immediate postoperative period and then every 4 to 6 hours. Dogs received meloxicam (0.1 to 0.2 mg/kg, SC) when pain scores of > 3 were reported. Dogs received hydromorphone (0.05 mg/kg, SC, as necessary) if their pain scores did not improve 20 to 60 minutes following meloxicam administration. Dogs with pain scores of > 4 received a combination of meloxicam and hydromorphone. Dogs were monitored during the next 30 days as a part of a J-tube placement study.⁴¹

Statistical analysis—Heart rate, respiratory rate, MAP, PaO₂, PaCO₂, PETCO₂, ETISO, arterial pH, CO, and body temperature were recorded and SV, CI, and O₂ delivery values were calculated for each dog at each of the 6 time points described. Because group 3 dogs received a different anesthetic protocol than group 1 and 2 dogs, data were analyzed in 3 stages designated as follows: at baseline (before any surgical stimulation), during the procedure (measurement times 5 to 30 minutes [following insufflation or the start of the laparotomy and ending prior to desufflation or closure]), and after the procedure (5 and 10 minutes following the start of abdominal closure [group 1 dogs] or after desufflation [group 2 and 3 dogs]). Similar to previous reports,^17,43 statistical tests were performed by use of a commercial software program.ⁱ Comparisons were made among groups at baseline and also after the procedure. For each dog, the difference between postprocedural and baseline data was computed and compared within and among groups, treating the difference itself as the response. To determine whether desufflation had an effect on cardiopulmonary parameters, for each dog, the difference between postprocedural data and data at time 30 minutes during the procedure was computed and compared within group 2 and 3 dogs, treating the difference itself as a response. For each dog, the difference between procedural (at 5 to 30 minutes during the procedure) and baseline data was computed and compared within and among groups. Because the 4 time points contained in the procedure period were repeated measures, the mean value of these 4 values for each dog was computed to perform the analysis. Mean procedural data were compared to test whether cardiopulmonary parameters differed among groups. To determine whether any changes in cardiopulmonary parameters occurred over time within and among groups during the procedure, for each dog, slopes (measure of change over time) were computed and these slopes were treated as the response itself.

A 1-way ANOVA and post hoc t test were used to analyze all responses (data, means, differences, or slopes). Assumptions of the ANOVA were assessed via comprehensive analyses of residuals, including normality tests (Shapiro-Wilk test) and residual plots to look for outliers, unequal variances, or other patterns. A Kruskal-Wallis test was performed in instances where the residuals were not normally distributed and a transformation was not successful. A statistician (WCS) and his assistant performed all analyses. Values of P ≤ 0.05 were considered significant.

Results

The measured hemodynamic parameters CO, MAP, HR, and blood gas tensions remained within reference limits for all dogs during this study. All laparoscopic procedures performed under sedation with epidural and local anesthesia (group 3 dogs) were completed without requiring conversion to a general anesthesia protocol. One group 3 dog required a second administration of epidural lidocaine 90 minutes after the first injection as well as a temporary increase in the CRI dose of fentanyl (from 40 to 44 μg/kg/h) to enable completion of the laparoscopic procedure. The laparoscopic procedure in this dog was substantially longer as a result of technical difficulties. No complications occurred in association with epidural catheterization throughout the study.

The mean duration of each procedure was 39 minutes (median, 35 minutes; range, 30 to 60 minutes), 65.6 minutes (median, 70 minutes; range, 40 to 78 minutes), and 74 minutes (median, 74 minutes; range, 65 to 80 minutes) for group 1, 2, and 3 dogs, respectively. Because of the study design, measurements at times 5 to 30 minutes were completed prior to exteriorizing the chosen loop of jejunum. In instances where the wrong segment of intestine was exteriorized (2 dogs [1 dog each in groups 2 and 3]), the study design resulted in artificially lengthened procedure times.

Fentanyl infusion doses that were required to provide adequate sedation for the procedure ranged from 17 to 56 μg/kg/h (mean, 29 μg/kg/h) in group 3 dogs. Midazolam infusion doses ranged from 0.10 to 0.25 mg/kg/h (mean, 0.14 mg/kg/h).

Most data were normally distributed, and the assumptions of the ANOVA were adequately met. Examination of the residual plots revealed an occasional outlier that was only removed if examination of the raw data revealed apparent data acquisition abnormalities.

Group 1 (P = 0.009) and group 2 (P = 0.013) dogs had a significant increase in MAP during the procedure, compared with baseline values (Table 1), and a significant increase in MAP at 5 minutes following the initiation of closure (group 1; P = 0.002) or desufflation (group 2; P = 0.023), compared with baseline values (Table 2). In contrast, group 3 dogs had a significant decrease in HR (P = 0.006) and MAP (P = 0.001) and a significant increase in PaCO₂ (P = 0.001) during the procedure, compared with mean baseline values. Compared with baseline values for group 3 dogs, a significant decrease was also found in HR (P = 0.003), MAP (P = 0.005), CI (P = 0.048), and O₂ delivery (P = 0.006) and a significant increase in PaCO₂ (P = 0.005) was found at 5 minutes after the procedure; PaCO₂ remained significantly (P = 0.020) increased at the 10-minute postprocedure time point.

No significant differences were found between group 1 and 2 dogs for all time points and parameters, except for O₂ delivery and PETCO₂. A significant difference in O₂ delivery was observed between group 1 and 2 dogs, with a greater O₂ delivery detected in group 2 dogs, compared with group 1 dogs, during the procedure (Table 3; P = 0.043) and at 5 minutes after the procedure (Table 4; P = 0.041).

Ventilation was adjusted when required in an attempt to maintain PETCO₂ values at < 45 mm Hg. This was not achieved in all group 2 dogs, resulting in significantly (P = 0.012) higher PETCO₂ in group 2 dogs, compared with group 1 dogs, during the procedure (Table 3). No significant differences were found in mean dialed isoflurane concentrations or ETISO concentrations between group 1 and 2 dogs.

Group 3 dogs had significantly higher mean values for CI, compared with group 1 dogs, at baseline (Table 5; P < 0.001), during the procedure (Table 3; P = 0.007), and at 5 minutes after the procedure (Table 4; P = 0.001). Group 3 dogs had significantly higher mean values for CI, compared with group 2 dogs, at baseline (P < 0.001) and at 5 minutes after the procedure (P = 0.018); no significant difference in CI was found between group 3 and 2 dogs during the procedure (P = 0.061). No significant differences were found in CI among groups when procedural data were compared with baseline (Table 1) and when postprocedural data were compared with baseline (Table 2).

Group 3 dogs had significantly higher mean MAP at baseline, compared with group 1 (P < 0.001) and group 2 (P < 0.001) dogs, and significantly (P = 0.035) higher mean MAP during the procedure, compared with group 1 dogs. However, group 3 dogs had a significantly lower slope value (mean MAP decreasing over time) during the procedure, compared with group 1 (P = 0.002) and group 2 (P = 0.006) dogs. Mean MAP decreased over time in group 3 dogs (slope = −0.219) and increased over time in group 1 (slope = 0.952) and group 2 (slope = 0.734) dogs. In comparison of procedural data with baseline and postprocedural data with baseline, mean MAP in group 3 dogs was significantly different from that of group 2 dogs (P < 0.001 for both comparisons) and group 1 dogs (P < 0.001 and P = 0.001, respectively). Mean MAP values at baseline, compared with mean MAP values during and after the procedure, were significantly higher in group 3 dogs (P = 0.001 and P = 0.005, respectively) and lower in group 2 dogs (P = 0.013 and P = 0.023, respectively) and group 1 dogs (P = 0.009 and P = 0.002, respectively).

Although no significant differences were found in the mean HR among groups at baseline and during the procedure, group 3 dogs had a significantly lower HR after the procedure, compared with group 1 dogs (P = 0.039) and group 2 dogs (P = 0.003). Mean HR in group 3 dogs was significantly lower during (P = 0.006) and after (P = 0.003) the procedure, compared with baseline values. In comparison of values after the procedure with baseline values, the mean differences in HR were significantly (P = 0.018) different between group 2 and 3 dogs, with group 3 dogs having a significant decrease in HR over time (slope = −31.0) and group 2 dogs having an increased HR over time (slope = 1.8).

The SV was significantly (P = 0.008) higher at baseline in group 3 dogs, compared with group 1 dogs. The SV in group 3 dogs was significantly higher, compared with group 1 and 2 dogs, during (P < 0.001 for both comparisons) and after (P < 0.001 for both comparisons) the procedure.

Measurements of PaCO₂ in group 3 dogs were significantly (P = 0.036) higher at baseline, compared with group 1 dogs. Measurements of PaCO₂ in group 3 dogs were significantly higher, compared with group 1 and 2 dogs, during (P < 0.001 for both comparisons) and after (P < 0.001 for both comparisons) the procedure. Mean differences in PaCO₂ were significantly different between group 3 and 1 dogs in the comparison of procedural with baseline values (P = 0.005) and in the comparison of postprocedural with baseline values (P = 0.019). Similarly, mean differences in PaCO₂ were significantly different between group 3 (P = 0.005) and group 2 (P = 0.037) dogs.

The PaO₂ was significantly (P = 0.005) lower in group 3 dogs at baseline, compared with group 1 dogs. The PaO₂ was significantly lower in group 3 dogs during and after the procedure, compared with group 1 (P < 0.001 for both comparisons) and group 2 (P < 0.001 for both comparisons) dogs.

Significantly higher O₂ delivery was observed in group 3 dogs, compared with group 1 and 2 dogs, at baseline (P < 0.001 and P = 0.003, respectively), during the procedure (P < 0.001 for both comparisons), and after the procedure (P < 0.001 and P = 0.046, respectively). In addition, mean differences in O₂ delivery after the procedure in comparison with baseline values were significantly different in group 3 dogs, compared with group 1 (P = 0.030) and group 2 (P = 0.038) dogs. In fact, a significant decrease in O₂ delivery after the procedure, compared with baseline values, was observed in group 3 dogs (slope = −25.14), whereas O₂ delivery increased over time in group 1 (slope = 1.106) and decreased slightly in group 2 (slope = −0.208) dogs. A significantly lower arterial pH was observed in group 3 dogs, compared with group 1 and 2 dogs, at baseline (P = 0.008 and P = 0.003, respectively), during the procedure (P < 0.001 for both comparisons), and after the procedure (P < 0.001 for both comparisons). No significant differences in cardiopulmonary parameters were found within groups or among groups in comparison of the 5-minute postprocedural values with the 30-minute procedural values (Table 6).

Mean core body temperature decreased from 37.4 ± 0.47°C to 36.4 ± 0.38°C in group 1 dogs, from 38.2 ± 0.22°C to 37.2 ± 0.43°C in group 2 dogs, and from 37.1 ± 0.34°C to 36.2 ± 0.29°C in group 3 dogs. Body temperatures were not significantly (P = 0.860) different among groups.

All dogs recovered uneventfully after J-tube placement. Three group 1 dogs and 1 group 2 dog required treatment with acepromazine during anesthetic recovery to treat signs of dysphoria and agitation. Group 3 dogs did not require administration of acepromazine during this study. Most (4/5) group 3 dogs were alert and walking normally at < 30 minutes after the end of the procedure, whereas group 1 and 2 dogs took approximately 3 hours to return to an ambulatory state. The highest pain score recorded during this study was 5 (out of 10) in a group 1 dog. Group 1 dogs had a higher mean pain score (3.9) than group 2 (2.4) and group 3 (2.2) dogs. Pain scores were not significantly different between group 1 and 2 dogs (P = 0.090) and group 2 and 3 dogs (P = 0.800). Mean pain scores were significantly (P = 0.05) higher in group 1 dogs, compared with group 3 dogs. The F value for mean pain scores was 0.11, and the power of this study with 15 dogs was 51%.

Discussion

Results of this study indicate that laparoscopic-assisted J-tube placement in healthy dogs under general anesthesia with isoflurane is hemodynamically comparable to open surgical J-tube placement. Furthermore, laparoscopic-assisted J-tube placement in dogs under sedation with epidural and local anesthesia was possible and resulted in improved cardiovascular function, compared with procedures performed in dogs under general anesthesia.

A general anesthesia protocol similar to that used in previous laparoscopic studies^8-10,40 in dogs was selected for group 1 and 2 dogs. Ketamine and diazepam were used for anesthetic induction in group 1 and 2 dogs because thiopental-induced splenomegaly increases the risk of inadvertent splenic puncture during trocar insertion, potentially causing hemorrhage, gas embolism, and decreased organ visualization.³³ Following completion of our study, both ketamine and isoflurane have been shown to cause splenomegaly as well.⁴⁴ Regardless, splenic puncture did not occur during our study, although dogs were positioned in right lateral recumbency and underwent left flank trocar insertion for laparoscopic-assisted J-tube placement. Isoflurane was selected to maintain general anesthesia in our study because it is commonly used in clinical patients.

The fentanyl-midazolam sedation protocol selected in group 3 dogs was selected on the basis of previous studies performed in dogs¹⁰ and humans³ and our clinical preference in severely compromised patients. Advantages of these drugs for our study and potential clinical use with laparoscopic procedures include wide therapeutic ranges, high thresholds of toxicity, reversibility,⁴⁵ and rapid clinical responses with dose changes. Induction doses of fentanyl are between 3 to 10 μg/kg^46,j with maintenance of dosages between 10 to 40 μg/kg/h^46,j when used as adjunct to inhalant anesthesia. These doses are higher than reported doses used for postoperative pain management (1 to 6 μg/kg/h).^47,48 Fentanyl doses used in our study (30 to 60 μg/kg/h) were chosen to provide heavy sedation in conjunction with midazolam and epidural anesthesia.

Few reported doses exist for CRI of midazolam in dogs. Boluses of 0.1 to 0.5 mg/kg are described.^48,j Infusion rates of midazolam of 0.3 to 90 mg/kg/h have been described for studies^49,50 in dogs investigating midazolam pharmacokinetics and minimum alveolar concentration reduction for enflurane. In comparison, a typical dose for conscious sedation of adult humans is 1.26 mg/kg/h.⁵¹ In our study, midazolam was administered at an induction dose of 0.1 mg/kg followed by a 0.2 mg/kg/h infusion.

Overall, doses for fentanyl and midazolam were chosen on the basis of those found in the literature^10,46,49,50 and our clinical experience to allow dogs in our study to remain conscious but sedate and comfortable throughout the laparoscopic procedure. Adjustment of the fentanyl and midazolam doses in group 3 dogs was performed to maintain sedation scores between 1 and 2 with the lowest dose possible. Gentle manual restraint was required in all dogs to maintain recumbency. Adequate spontaneous ventilation was achieved in these healthy research dogs. It is important to note that with respect to both drugs, the doses used were required to achieve optimal sedation and maintain positioning in our healthy research dogs and doses required for debilitated dogs would likely be much less.

Because abdominal insufflation, intestinal manipulation, and administration of opioids can increase vagal tone, all dogs in both the general anesthesia (groups 1 and 2) and sedation (group 3) groups were premedicated with glycopyrrolate instead of treating bradycardia once it developed. This allowed accurate comparison of hemodynamic parameters among groups.

In addition to sedation, group 3 dogs also received epidural anesthesia. Lidocaine was chosen for our study because it has a rapid onset of action and duration of effect of 60 to 90 minutes,³⁴ which was adequate for 4 of 5 laparoscopic procedures in our study. Epidural effectiveness may be influenced by technique, local drug flow, and drug volume.^52–57 Epidural catheterization was performed in group 3 dogs, and catheter position was verified with contrast epidurograms prior to the procedure to ensure adequate administration of local anesthetic into the epidural space. Testing the cranial extent of the epidural by dermatome would have provided additional information requiring the necessity of additional local infiltration but was not performed. Lack of local anesthesia was not a problem in our study; however, temporary positioning of dogs in left lateral recumbency after lidocaine injection could improve anesthesia of the left abdominal wall.

With respect to cardiovascular measurements, no significant hemodynamic or inhalant differences were identified in dogs receiving general anesthesia (groups 1 and 2) with the exception of PETCO₂ and O₂ delivery. The significantly increase in PETCO₂ during insufflation in group 2 dogs, compared with group 1 dogs, was likely caused by exogenous CO₂ absorption and decreased ventilation as a result of abdominal insufflation. The significant increased O₂ delivery observed during and after the procedure in group 2 dogs, compared with group 1 dogs, was likely the result of the additive effects of mild increases (ie, nonsignificant changes). Because the anesthetic protocols were identical for group 1 and 2 dogs, the stimulating effects of capnoperitoneum (via neuroendocrine responses and hypercapnia) and differences in positioning may explain the greater cardiovascular function and therefore the increased O₂ delivery observed in group 2 dogs, compared with group 1 dogs.

Patient positioning for laparoscopic procedures can significantly affect hemodynamic function. Specifically, improvements in cardiac preload and MAP have resulted from elevating the head above the feet in humans^15,58 and pigs⁵⁹ and by performing standing procedures in horses.^17,18 The Trendelenburg position improves CO in dogs⁶⁰ and cardiovascular parameters in horses,¹³ but to our knowledge, studies comparing dorsal to lateral positioning have not been performed. It is possible that compression of the vena cava by abdominal organs while in dorsal recumbency (group 1 dogs) was greater than compression resulting from lateral recumbency (group 2 dogs). This may have potentially decreased venous return and cardiac preload and caused slight (nonsignificant) decreases in CO in group 1 dogs, compared with group 2 dogs.^16,17,58 Overall, the clinical importance of the improvement in O₂ delivery is considered minimal in these healthy research dogs; however, it is likely that the significantly higher O₂ delivery observed in group 2 dogs, compared with group 1 dogs, would have a clinical benefit in sick dogs.

Anesthetic requirements were not significantly different between the groups of dogs receiving general anesthesia (groups 1 and 2). Although the smaller incisions for laparoscopic-assisted J-tube placement should theoretically be less stimulating than the long abdominal incision used for open surgical J-tube placement, peritoneal nociceptor neuroendocrine stimulation caused by increased IAPs and chemical irritation with laparoscopy may have led to similar anesthetic requirements between group 1 and 2 dogs. Measurement of plasma catecholamine concentrations could have provided further insight into the degree of surgical stimulation in each group but was not performed in our study.

Group 3 dogs had improved hemodynamic function (CI, SV, and MAP) and O₂ delivery, compared with group 1 and 2 dogs, throughout the procedure. The doses of midazolam and fentanyl used in our study are unlikely to influence hemodynamic function as much as isoflurane inhalant anesthesia⁴⁵; this likely resulted in less depression of CI in group 3 dogs. Although myocardial depression produced by moderate hypercapnia and high IAPs can cause decreases in CO,²² the PaCO₂ and IAPs produced in our study do not reportedly affect CO in dogs,^9,10,16 horses,¹³ or humans^14,15 when adequate ventilation is provided. Lower intrathoracic pressures during spontaneous ventilation may also have resulted in better venous return to the heart and increased CO in group 3 dogs, compared with group 1 and 2 dogs. In addition, higher intrathoracic pressures resulting from mechanical ventilation may have decreased venous return in group 1 and 2 dogs.⁶¹

Values of MAP were significantly higher in group 3 dogs at baseline, compared with group 1 and 2 dogs. This was likely the result of their conscious state with epidural and local anesthesia, compared with general anesthesia. Following insufflation, MAP remained significantly greater in group 3 dogs, compared with group 1 dogs. This may have been caused by increases in SVR as a result of exogenous CO₂ absorption in addition to the conscious state of group 3 dogs. The progressive decrease in MAP throughout the procedure period (negative slope) in group 3 dogs, compared with an increase over time in group 1 and 2 dogs, may have been caused by the initial increase in venous return from compression of abdominal capacitance vessels and splanchnic vasculature during insufflation.^11,14,62 In addition, increases in MAP caused by surgical stimulation may have been prevented in group 3 dogs by epidural and local anesthesia, leading to a decrease in MAP over time.⁶³ Increases in MAP during and after the procedure in group 1 and 2 dogs, compared with baseline, may represent equivalent responses to nociception from surgical stimulation for group 1 dogs and stimulation from pneumoperitoneum in addition to surgical stimulation caused by intestinal manipulation in group 2 dogs.

Overall, the significantly higher MAP observed in group 3 dogs, compared with group 1 and 2 dogs, was not considered clinically important in these healthy research dogs because MAP remained > 60 mm Hg in all dog groups throughout the study period. However, the relatively higher MAP observed during laparoscopicassisted J-tube placement in dogs under sedation and local anesthesia might have clinical relevance in compromised patients because preexisting disease and hypovolemia from dehydration or inadequate IV administration of fluids may exacerbate insufflation-mediated hypotension.^2,19,30 It is interesting that the change in MAP between baseline and procedure and baseline and after the procedure was significantly greater in group 3 dogs than group 1 and 2 dogs. This change reflects a decline in MAP over time in group 3 dogs to a value that was equivalent to the other 2 groups by the 30-minute procedure and postprocedure time points.

Stroke volume has been shown to decrease significantly in dogs during laparoscopic procedures performed under general anesthesia.⁹ Decreases in SV were not identified in dogs (group 1 and 2) undergoing laparoscopy in our study. This is consistent with the maintenance of HR in all dog groups and may have resulted from premedication with glycopyrrolate. Although increases in PaCO₂ from exogenous absorption can increase SV and induce mild hypercapnia, as observed in our study, it has little clinical effect.¹⁴ Release of vasopressin and catecholamines in response to pneumoperitoneum and the direct effects of high IAPs is implicated in increased SV during laparoscopy.¹¹ The use of general anesthesia may have resulted in some myocardial depression and decreases in SV in group 2 dogs, compared with group 3 dogs.

Premedication with glycopyrrolate may have initially prevented the effects of fentanyl-induced bradycardia in group 3 dogs; HR in this group of dogs decreased over time, possibly as the effects of glycopyrrolate diminished. The absence of strong surgical stimulation caused by a combination of epidural and local anesthesia is also likely to have affected the HR of group 3 dogs. Neuroendocrine nociceptor stimulation caused by insufflation, intestinal manipulation, and hypercarbia may have increased plasma catecholamine and vasopressin concentrations^10,21 and helped to maintain HR in group 2 and 3 dogs. An increase in HR during laparoscopy under general anesthesia has been observed in some studies,^2,12 but not in others,^11,16 and was not observed in our study.

The cumulative effects of mildly improved cardiovascular parameters resulted in higher O₂ delivery in group 3 dogs at all time points, compared with group 1 and 2 dogs. However, O₂ delivery significantly decreased in the postprocedure period, compared with baseline values, in group 3 dogs, reflecting decreases in CO and PaO₂ observed after desufflation in this group. Although O₂ delivery did not change significantly over time in group 1 and 2 dogs, group 2 dogs had a significantly higher O₂ delivery during and after the procedure, compared with group 1 dogs; this may reflect the effect of body positioning on CO.

In contrast to hemodynamic parameters, pulmonary function (PaO₂ and PaCO₂) was significantly decreased in group 3 dogs, compared with group 1 and 2 dogs. Differences in anesthetic protocols, the effects of exogenous CO₂ absorption, and errors in ventilation technique contributed to these differences. Previous reports^9,16 document the need to increase VE in dogs undergoing general anesthesia to maintain PaCO₂ within reference range.^9,16 Spontaneously ventilating patients, such as group 3 dogs, must spontaneously increase VE, tidal volume, or both to maintain normocarbia.^19,22,25 Group 3 dogs developed a significantly higher PaCO₂, compared with group 1 and 2 dogs; this was likely in part from sedation-induced respiratory depression.²⁸ Increase in PaCO₂ in group 2 and 3 dogs, compared with group 1 dogs, was likely secondary to transperitoneal absorption of exogenous CO₂ and decreased pulmonary function from increased IAPs. These changes are consistent with those of previous reports in dogs,^5,9,10 horses,¹³ and children^6,14,32 and were not considered physiologically important because the PaCO₂ did not lead to cardiovascular compromise during our study. Group 3 dogs maintained a significantly higher PaCO₂ at 5 and 10 minutes after the procedure (after desufflation), compared with during the procedure and at baseline, whereas group 2 dogs undergoing mechanical ventilation had returned to baseline PaCO₂ within 5 minutes of desufflation. On the basis of these results, dogs receiving epidural anesthesia and sedation, especially those with preexisting cardiopulmonary compromise, would be likely to develop clinically relevant hypercarbia during lengthy procedures (even after desufflation) and might require conversion to general anesthesia to provide mechanical ventilation.

Measurements of PETCO₂ and PaCO₂ were possible in group 1 and 2 dogs and were used to guide adjustments in VE to maintain PETCO₂ at < 45 mm Hg. Despite this effort, a small but significant increase in PaCO₂ was observed during insufflation in group 2 dogs, compared with group 1 dogs. This was reflected by a small increase in PETCO₂ for group 2 dogs following insufflation and is consistent with previous reports in dogs⁹ and humans.^14,15 Although CO measurement with LiDCO is less time-consuming than with other dilution methods, delays in adjusting VE on the basis of blood gas analysis occurred because of the time required to perform each CO measurement. This fault in study design could have been eliminated with additional anesthetic personnel. Delays in adjusting VE likely affected group 2 more than group 1 dogs because the changes in PaCO₂ occurred more rapidly in this group secondary to abdominal insufflation. Although increases in PETCO₂ have been reported to follow abdominal desufflation,⁹ this was not identified in our study and may reflect shorter laparoscopic procedures (< 80 minutes), compared with those of the other study⁹ (ie, 180 minutes). Values of PETCO₂ in group 2 dogs were similar to group 1 dogs within 5 minutes following desufflation. Values of PETCO₂ in group 3 dogs had not returned to baseline at 5 and 10 minutes following desufflation.

During spontaneous ventilation in group 3 dogs, O₂ delivery resulted in a lower PaO₂, compared with intubated and mechanically ventilated group 1 and 2 dogs. Despite this difference, PaO₂ remained clinically adequate throughout the study (range, 180 to 244 mm Hg) for group 3 dogs.

Unsuccessful or poorly effective epidural anesthesia in clinically ill dogs may require conversion to general anesthesia. In our study, conversion to general anesthesia was not necessary for J-tube placement in any group 3 dog. Epidural injections have a reported clinical failure rate between 6.8% to 24% in human⁶⁴ and veterinary patients.^31,53 Although group 3 dogs did not react to trocar placement, it is possible that the lack of effective epidural anesthesia at the level of the cranial portal was masked by injection of lidocaine at the proposed portal sites. Epidural anesthesia may also provide unreliable visceral analgesia for intestinal manipulation in conscious patients, independent of adequate incisional anesthesia.^53,65 In another study,⁵³ 12 of 25 goats receiving sedation and either lidocaine or bupivacaine epidural anesthesia reacted to visceral manipulation via an incision in the left flank. In contrast, only 1 of 5 dogs in our study reacted to intestinal manipulation. This occurred 85 minutes after the initial epidural injection of lidocaine. A second epidural injection of lidocaine and a temporary increase in fentanyl CRI dose provided adequate anesthesia for the remainder of the procedure (31 minutes). In this dog, the failure of epidural anesthesia appeared to be related to a prolonged procedure and the limited duration of action of lidocaine.

Group 2 and 3 dogs required fewer postoperative administrations of analgesics than group 1 dogs; group 3 dogs appeared less dysphoric. Although these differences were not significant among group 1 and 2 dogs, the low F value suggests a possible difference. Further analyses revealed that the power for this part of the study was only 51%; 30 dogs would be required to achieve a power of 85%. Analgesic administration was required infrequently during the postoperative period and did not extend beyond 1 to 2 days for all dogs. Lower postoperative analgesic requirements in group 2 and 3 dogs likely reflect that smaller laparoscopic incisions result in less postoperative pain than traditional laparotomy.¹ In addition, group 2 and 3 dogs received effective local analgesia from portal site infiltration and group 3 dogs also received epidural anesthesia, which may have affected initial postoperative pain scores. In 1 report,⁶⁶ local trocar site infiltration with bupivacaine did not reduce postoperative analgesic requirements in human patients, but this has been disputed by others.^67–69 Similarly, intra- and postoperative peritoneal injection of local anesthetics in human patients undergoing laparoscopic procedures^70,71 has produced conflicting results and was not performed in our study. Limitations associated with pain assessment in our study include potential bias toward type of surgery and administration of analgesics to potentially dysphoric dogs. Blinding would have required separate investigators to assess pain scores and extensive bandaging to hide the clipped areas and various skin incisions.

On the basis of the results of our study, we conclude that with accurate patient monitoring and support, hemodynamic changes that occur during laparoscopic procedures performed under general anesthesia are comparable to those observed during laparotomy. In addition, the sedation, local infiltration, and epidural anesthesia protocol used in our study is adequate for laparoscopic-assisted J-tube placement in young, healthy dogs and results in significantly improved hemodynamic parameters. Maintenance of hemodynamic parameters similar to those obtained in awake dogs could be clinically important for geriatric or critical patients with illnesses that could be exacerbated by general anesthesia. Although the increase in PaCO₂ was tolerated well by healthy dogs in our study, special considerations for patients with compromised cardiopulmonary function or severe diseases are required. Our study also suggests that laparoscopic-assisted J-tube placement in dogs under sedation and local anesthesia offers a minimally invasive alternative to traditional procedures. However, further studies involving clinical patients are indicated.