Plasma Electrolyte and Metabolite

Plasma Electrolyte and Metabolite Concentrations Associated with Pentobarbital or Pentobarbital-Propofol Anesthesia During Three Weeks’ Mechanical Ventilation and Intensive Care in Dogs

Gerald A. Gronert,1* Steve C. Haskins,2 Eugene P. Steffey,2 and Dennis Fung1

Abstract ͉ Propofol and pentobarbital were used for deep sedation during prolonged mechanical ventilation (3 weeks) and nutritional supplementation in 17 clinically normal dogs in an intensive care setting. Tolerance developed to both drugs. Propofol, in combination with pentobarbital, at an infusion rate of 75 ␮g/kg of body weight per minute was preferred. Pentobarbital infusion alone, begun at the rate of 5 to 6 mg•kg-1•h-1, was satisfactory. The combination of both drugs provided smooth, stable anesthesia and re- quired minimal interventions by intensive care unit personnel. Blood gas tensions and electrolyte, parathyroid hormone (PTH), and metabolite concentrations were generally stable throughout, unless condition of the dog deteriorated (e.g., infection, pneumothorax). Hematocrit and red blood cell count decreased with time, likely attributable principally to multiple blood sample collections. White blood cell count, alkaline phosphatase, phosphate, fibrinogen, cholesterol, and triglyceride values increased with time, in association with pentobarbital and the combination of pentobarbital and propofol. Some of these changes appear to have been related to generic responses to stress and inflammation, some to altered metabolism, and some to the lipid solvent of propofol. The increase in triglyceride concentration was greater when propofol was used. Mortality was 47%, with death occurring between days 2 and 18.

It is well recognized that sedation in the intensive care ing ICU care (e.g., infection, pneumothorax). In other words, unit (ICU) environment can facilitate management of the dogs should remain overtly healthy as long as daily fluid critically ill patient (e.g., nursing responsibilities or venti- and electrolyte replacement, change of position, bladder lation). However, to the authors’ knowledge, effects of such emptying, and ventilatory care were maintained, and spe- management per se (i.e., on healthy subjects) regarding cific problems were prevented or controlled. Documenta- plasma electrolyte, blood gas, and metabolite values have tion of this maintained homeostasis would be reflected by not been elucidated. Confounding factors that might alter stable values for plasma electrolytes, PTH, blood gases, and these properties during ICU care include duration of seda- metabolites. tion, types of drugs, and species. For example, not all species have the same reference ranges for electrolytes (e.g., Materials and Methods erythrocyte potassium [K+], but not plasma K+ concentra- The study reported here was approved by our animal use tion, is markedly lower in dogs and cats, compared with and care committee, and protocols conformed to the guide- pigs and humans [1]). Brief general anesthesia per se does lines of the National Institutes of Health. Dogs entering not appear to alter stability of plasma electrolyte concen- this protocol had been maintained in facilities with care trations and other variables in dogs (2–4). To our knowl- and use programs approved by AAALAC, International, and edge, long-term effects of anesthesia in the ICU environ- were housed individually in indoor-outdoor kennel runs ment have not been examined in healthy subjects. with automatic watering systems, fed a commercially avail- To further advance this area of knowledge, we examined able canine diet, provided with routine preventive veteri- the effects of deep sedation (no response to painful stimuli) nary care, and observed daily by trained animal techni- in clinically normal dogs during ICU care for 3 weeks. We cians for any signs of illness. After food had been withheld hypothesized that deep sedation would have unobtrusive overnight, 17 healthy mature beagles weighing 9 to 17 effects in dogs that did not develop specific problems dur- (mean Ϯ SEM, 10.5 Ϯ 2.2) kg were heavily sedated with either pentobarbital or a combination of pentobarbital and Department of Anesthesiology, School of Medicine1; Department of Surgi- propofol. When pentobarbital was used alone (n = 6), its cal and Radiological Sciences, School of Veterinary Medicine,2 University infusion rate was begun at 5 to 6 mg•kg-1•h-1 and was ad- of California, Davis, California *Address correspondence to Dr. Gerald A. Gronert, University of Califor- justed upward as tolerance developed (Model AS40A infu- nia, MED: Anesthesiology, TB 170, Davis, CA 95616. sion pump; Baxter Health Care Corp., Deerfield, Ill.). For

513 Vol 48, No 5 Laboratory Animal Science October 1998 the combination of pentobarbital and propofol, we initially parenteral nutrition was instituted via a percutaneous cen- used propofol (150 ␮g•kg-1•min-1; n = 3), but with further tral cannula. The veterinary hospital pharmacy provided experience, we modified the rate to 75 ␮g•kg-1•min-1 (n = parenteral nutrition as a mixture of 500 ml of 50% glucose, 8). Depth of sedation per anesthesia was evaluated by loss 450 ml of Intralipid, 900 ml of protein as Travasol (Intralipid of various signs: palpebral reflex, mandibular muscle tone, and Travasol; Baxter Health Care Corp.), 4 ml of multivi- spontaneous muscular activity, shivering, and appendicu- tamin B, and 22 mEq of potassium phosphate. lar skeletal muscle tone (when neuromuscular blocking Vital signs were closely followed and maintained via al- agents were used, paralysis was never total), and further, teration of the level of deep sedation and/or fluid balance, by changes in eye position, blood pressure, and heart rate. approximately as follows: blood pressure, systolic between The trachea was intubated (clear plastic; Mallinckrodt, 100 and 160, diastolic between 60 and 100, and mean be- Irvine, Calif.); the low-pressure, high-volume cuff was in- tween 80 and 120 mm Hg; heart rate between 90 and 150 flated sufficiently to prevent leakage of air around the tube; beats/min; rectal temperature between 37 and 39ЊC (main- and the lungs were mechanically ventilated (model MA1; tained by a circulating-water mattress and a heat lamp);

Puritan-Bennett Inc., Santa Monica, Calif.). Ventilation was pulse oximeter value >95%; PaO2 between 80 and 100 mm started with 21% inspired O2 at a tidal volume of 15 ml/kg Hg while breathing 21% O2, supplemented with added in- and 15 breaths/min; inspired O2 was increased if pulse spired O2 if PaO2 was <80 mm Hg; end-expired CO2 between oximeter saturation (shaved tail) was <95%. The cuff was 32 and 38 mm Hg; and PaCO2 between 35 and 42 mm Hg, deflated every 8 h, and the tube was moved slightly to maintained by adjusting minute ventilation. Plasma elec- change the pressure point on the tracheal epithelium. Use trolyte, packed cell volume, total protein, and glucose val- of an “artificial nose” (Humidivent; Louis Gibeck, Upplands, ues were monitored daily and were maintained within ca- Vasby, Sweden) aided in humidification. nine reference ranges. Nine dogs were given a competitive antagonist muscle Every 4 h, each dog’s position was changed (supine, left- relaxant (two received vecuronium [0.1 mg•kg-1•h-1], and lateral, sternal, right-lateral, dorsal); the trachea had lav- seven were given rocuronium [1 mg•kg-1•h-1]) to maintain age/suction (3 ml of normal saline aseptically instilled into one twitch in the train-of-four stimulus pattern. As several the trachea and suctioned, using a closed tracheal suction weeks passed and tolerance developed, these rates were catheter system, which, via its plastic sheath and access gradually increased by 50 to 200%. The train of four con- ports, was continuously attached to the endotracheal tube sists of four supramaximal stimuli (2 Hz for 2 sec) applied [Trach Care; Ballard Medical Products, Draper, Utah]); and via permanent stimulating wires (20 gauge) that were sur- the bladder was manually expressed. Urine output was gically placed around the sciatic nerve. Administration of measured, and daily input-output balances were measured. the muscle relaxant was continued for the 3-week period, The endotracheal tube was changed every 3 days, and with 12-h drug holidays each week for determination of parenteral lines were changed each week. Defecation was metocurine’s pharmacokinetics, and pharmacodynamics spontaneous or aided by enemas if palpation of the colon and change in potency associated with duration of whole- indicated constipation; perianal cleanliness was main- body immobilization (5). Intravascular catheters (cephalic tained. Antibiotics, usually ticarcillin-clavulanate, were or saphenous vein, occasionally external jugular) and stimu- administered when there was evidence of sepsis: fever, high lating wires were placed aseptically. A percutaneous ind- white blood cell (WBC) count, purulent tracheal secretions. welling femoral arterial catheter provided transduced pres- Backup antibiotics included ampicillin and gentamicin. sure data (model MX9505; Medex, Hilliard, Ohio) and blood Bacteriologic culture was not done, except for the propofol sample collection. A peripheral intravenous catheter pro- solution when there was suspicion of infection. vided the means for infusion of drugs by pumps (model Laboratory equipment immediately available in the ca- AS40A infusion pump; Baxter Healthcare and MedTrac nine ICU included blood gas machine (Corning 170 pH/blood model M9500 volumetric infusion pump; American Edwards gas analyzer; Ciba Corning Diagnostics Corp., Medfield, Laboratories, Irvine, Calif.) and of Plasmalyte 56 (Baxter Mass.), co-oximeter (IL 482 Co-Oximeter; Instrumentation Laboratories: [mEq/L] sodium, 40; potassium, 20 [increased Laboratories, Lexington, Mass.), microhematocrit centri- from 13]; magnesium, 3; chloride, 40; and acetate, 16, with os- fuge, lactate-glucose analyzer (2300 STAT glucose-lactate molality of 111 mOsm/L and pH between 4 and 6), at an infu- analyzer; Yellow Springs Instrument Co., Inc., Yellow sion rate of 3 mg•kg-1•h-1. Metoclopramide (0.04 mg•kg-1•h-1) Springs, Ohio), and an analyzer for potassium, sodium, was included in the infusion to improve gastric emptying. chloride, ionized calcium, and ionized magnesium (Electro- A nasogastric tube was placed, and canine semi-liquid en- lyte Analyzer 8; Nova Biomedical, Waltham, Mass.). Other riched enteral feed (Canine Clinicare; PetAg, Pet-Ag, Inc., routine tests, performed weekly, included platelet and red Hampshire, Ill.; 6 ml/kg of body weight) was administered blood cell (RBC) and WBC counts, calculated mean corpus- every 4 h. Stomach residuals were checked just prior to cular hemoglobin, and determination of fibrinogen, creati- feeding, and feedings were skipped if residuals exceeded nine, phosphate, blood urea nitrogen, alkaline phosphatase, 60 ml; for values <60 ml, the feeding was proportionately aspartate transaminase, creatine kinase, total bilirubin, reduced so that total volume did not exceed 100 ml. If gas- cholesterol, triglyceride, lactate dehydrogenase, total pro- tric residuals repeatedly exceeded 60 ml, motility was tein, albumin, and PTH values. stimulated, in addition to metoclopramide, by administra- This labor-intensive study mandated continuous, skilled, tion of cisapride, bethanecol, or erythromycin. If this failed, and conscientious care by two medical and two veterinary

514 ICU Stability in Sedated Ventilated Dogs

A B

(mm Hg)

C D

(mm Hg)

Pco

Base excess (mmol/L)

E Figure 1. Mean Ϯ SEM time (in days) versus arterial Po2 (A), pH (B), base excess (C), Pco2 (D), and lactate (E) values in dogs anesthetized with pentobarbital alone (5 to 6 mg•kg-1•h-1) initially (Pentobarb), pentobarbital plus propofol 150 ␮g•kg-1•min-1 (Propofol 150), or pentobarbital plus propofol (75 ␮g•kg-1•min-1).

a 24-h basis for participation in care or consultation. Care was provided on a continuous basis, generally with 4- to 6- h overlapping shifts. These were purposely terminal studies. Necropsy was performed to evaluate dogs that died prior

Lactate (mg/dl) to the 21st day. Statistics: Data are expressed as mean Ϯ SEM, and comparisons were done by use of ANOVA and the paired or unpaired Student’s t test, with P < 0.05 deemed signifi- cant. If data were not normally distributed, they were loga- rithmically transformed, and statistical comparisons were performed after normal distribution was confirmed. anesthesiologists (one with a primary focus in ICU care), one registered nurse experienced in ICU care, two full-time Results anesthesia technicians experienced in the care and han- Figures 1 and 2 and Table 1 display stability of the model. dling of dogs, and several temporary technicians taught Hematocrit steadily decreased in dogs given propofol, ear- and supervised by the preceding personnel. Two to three lier in those given 150 ␮g•kg-1•min-1 (Table 1). Decrease in dogs were studied simultaneously, with two people present hematocrit in pentobarbital-anesthetized dogs was slower for care and monitoring almost the entire time. The anes- to develop and was less pronounced. White blood cell, cho- thesiologists and nurse were either present or available on lesterol, alkaline phosphatase, and phosphate values in-

515 Vol 48, No 5 Laboratory Animal Science October 1998

A B

Figure 2. Mean Ϯ SEM time (in days) versus plasma sodium (Na+, A), potassium (K+, B), calcium (Ca++, C), and magnesium (Mg++, D) values in dogs anesthetized with pentobarbital alone 5 to 6 mg•kg-1•h-1 initially (Pentobarb), pentobarbital plus propofol 150 ␮g•kg-1•min-1 or pentobarbital plus propofol 75 ␮g•kg-1•min-1. creased, regardless of type of anesthesia. Triglyceride and high WBC count, most were stable and did not manifest cholesterol concentrations increased in all dogs, but to a signs of sepsis (e.g., shock, overall deterioration). greater extent in propofol-treated dogs. Fibrinogen con- Tolerance resulted in an increase in the pentobarbital centration initially increased by day 7, then decreased, al- infusion rate to 10 to 12 mg•kg-1•h-1 by the 21st day. Propofol though to values still greater than day-zero values (Table 1). infused at a rate of 150 ␮g•kg-1•min-1 was associated with The tube feedings and Plasmalyte 56 intravenous solution a trend toward a greater decrease in hematocrit than that provided appropriate electrolytes and free water for rou- associated with propofol 75, and caused concern regarding tine needs (Table 1, Figures 1 and 2). Two dogs received potential sepsis, although its risk, compared with that of parenteral nutrition. propofol 75, in that regard is unknown (the new stabilized Eight dogs died prior to day 21: on days 2, 5, 6, 8 (two preparation was not then available). dogs), 10, and 18 (two dogs). Causes included pneumotho- Hyperventilation occasionally developed; it altered blood rax (n = 5), sepsis (n = 2), and cardiovascular failure from CO2 values appropriately, and within a few minutes stimu- unknown causes (n = 1). Regarding pneumothorax, necropsy lated release of PTH, secondary to the associated decrease findings included pulmonary microemboli with surround- in ionized calcium concentration. This was observed early ing areas of necrosis (n = 4) and tracheal perforation by during the first day of study, when sample collection was the cuff of the endotracheal tube (n = 1). Anesthetics given frequent during initial stabilization. Oxygen was added to to dogs that died included pentobarbital alone in two dogs, the ventilator-inspired air when saturation was decreased; pentobarbital and propofol in six dogs. Six of the dogs that this accounts for the variations in Figure 1A. Mean pH (Fig- died had not received muscle relaxant, and two had received ure 1B), base excess (Figure 1C), and Pco2 (Figure 1D) var- rocuronium. In the two dogs with sepsis, results of culture ied little. Lactate values (Figure 1E) gradually increased of propofol were negative. Although many of the dogs had over the course of the experiment. There was no difference

516 ICU Stability in Sedated Ventilated Dogs

Table 1. Blood, electrolyte, and metabolite values, mean Ϯ SEM(N) Pentobarb- Pentobarb- Pentobarb- Pentobarb- Pentobarb prop 150 prop 75 Pentobarb prop 150 prop 75 Variable Day alone (␮g•kg-1•min-1)(␮g•kg-1•min-1) Variable Day alone (␮g•kg-1•min-1)(␮g•kg-1•min-1) Glucose (mg/dl) 0 76 Ϯ 4 (6) 80 Ϯ 1 (3) 49 Ϯ 6 (7) AST (U/L) 0 31 Ϯ 2 (6) 36 Ϯ 8 (3) 51 Ϯ 6 (7) 793 Ϯ 29 (4) 96 Ϯ 10 (3) 86 Ϯ 9 (4) 7 30 Ϯ 5 (4) 41 Ϯ 12 (3) 26 Ϯ 2 (4) 14 52 Ϯ 3 (4) 75 Ϯ 5 (2) 67 Ϯ 9 (4) 14 40 Ϯ 6 (3) 35 Ϯ 5 (2) 56 Ϯ 26 (3) 21 67 Ϯ 1 (3) NA 52 Ϯ 3 (3) 21 87 Ϯ 56 (3) NA 50 Ϯ 2 (2) Hematocrit (%) 0 36 Ϯ 2 (6) 35 Ϯ 2 (3) 33 Ϯ 1 (8) Total bili (mg/dl) 0 0.4 Ϯ 0.05 (6) 0.3 Ϯ 0.03 (3) 0.3 Ϯ0.02(7) 7a,b 31 Ϯ 1 (5) 19 Ϯ 2 (3) 24 Ϯ 2 (5) 7 0.5 Ϯ 0.2 (4) 0.6 Ϯ 0.03 (3) 0.3 Ϯ 0.07(4) 14a,b 28 Ϯ 2 (4) 20 Ϯ 1 (2) 20 Ϯ 1 (5) 14 0.4 Ϯ 0.03 (3) 1.0Ϯ 0.6 (2) 0.7 Ϯ 0.3 (3) 21 a 24 Ϯ 3 (4) 22 Ϯ 0 (1) 22 Ϯ 2 (4) 21 0.5 Ϯ 0.2 (3) NA 0.4 Ϯ 0 (2) RBC (x 106/␮l) 0 5.3 Ϯ 0.3 (6) 5.2 Ϯ 0.3 (3) 5.0 Ϯ 0.2 (8) CK (U/L) 0 311 Ϯ 37 (6) 413 Ϯ 173 (3) 1,034 Ϯ 282 (7) 7 a 4.5 Ϯ 0.1 (5) 2.8 Ϯ 0.3 (3) 3.7 Ϯ 0.4 (5) 7 333 Ϯ 116 (4) 283 Ϯ 58 (3) 252 Ϯ 61 (4) 14 a 4.1 Ϯ 0.3 (4) 2.8 Ϯ 0.3 (2) 3.0 Ϯ 0.2 (5) 14 145 Ϯ 13 (3) 207 Ϯ 47 (2) 352 Ϯ 147 (3) 21 a 35.8 Ϯ 0.4 (4) 2.9 Ϯ 0 (1) 2.9 Ϯ 0.2 (4) 21 231 Ϯ 74 (3) NA 212 Ϯ 8 (2) MCH (g/dl) 0 35.5 Ϯ 0.3 (6) 35.9 Ϯ 0.4 (3) 35.5 Ϯ 0.2 (8) Chol (mg/dl) 0 118 Ϯ 12 (6) 116 Ϯ 25 (3) 118 Ϯ 7 (7) 7 36.1 Ϯ 0.6 (5) 37.2 Ϯ 0.6 (3) 36.8 Ϯ 0.5 (5) 7 a 223 Ϯ 15 (4) 301 Ϯ 32 (3) 254 Ϯ 26 (4) 14 37.0 Ϯ 0.8 (4) 37.8 Ϯ 1.6 (2) 36.8 Ϯ 0.2 (5) 14 a 223 Ϯ 14 (3) 378 Ϯ 42 (2) 442 Ϯ 120 (3) 21 35.8 Ϯ 0.4 (4) 37.9 Ϯ 0 (1) 36.8 Ϯ 1.5 (4) 21 a 226 Ϯ 34 (3) NA 583 Ϯ 81 (2) WBC (x 103/␮l) 0 6.4 Ϯ 0.4 (6) 8.7 Ϯ 0.6 (3) 9.6 Ϯ 1.3 (8) Trigly (mg/dl) 0 13 Ϯ 0.8 (6) 13.7 Ϯ 2.2 (3) 17 Ϯ 0.9 (5) 7 a 14.0 Ϯ 1.3 (5) 11.8 Ϯ 2.2 (3) 14.1 Ϯ 1.2 (5) 7a,b 50 Ϯ 4.2 (4) 154 Ϯ 49 (3) 143 Ϯ 13 (4) 14 a 27.3 Ϯ 3.1 (4) 21.8 Ϯ 11.2 (2) 17.0 Ϯ 2.8 (5) 14a,b 88 Ϯ 5.5 (3) 302 Ϯ 82 (2) 630 Ϯ104 (3) 21 a 19.8 Ϯ 1.7 (4) 20.9 Ϯ 0 (1) 18.2 Ϯ 4.4 (4) 21a,b 124 Ϯ 33 (3) NA 764 Ϯ 118 (2) Fibrinogen (mg/dl) 0 125 Ϯ 25 (4) 167 Ϯ 67 (3) 150 Ϯ 19 (8) LD (U/L) 0 137 Ϯ 22 (6) 279 Ϯ 170 (3) 317 Ϯ 69 (7) 7 a 600 Ϯ 71 (4) 533 Ϯ 33 (3) 460 Ϯ 81 (5) 7 490 Ϯ 245 (4) 481 Ϯ 238 (3) 308 Ϯ 65 (4) 14 a 467 Ϯ 88 (3) 350 Ϯ 50 (2) 340 Ϯ 24 (5) 14 321 Ϯ 39 (3) 152 Ϯ 18 (2) 325 Ϯ 181 (3) 21 a 325 Ϯ 48 (4) 400 Ϯ 0 (1) 275 Ϯ 63 (4) 21 211 Ϯ 74 (3) NA 440 Ϯ 66 (2) Platelets (x 103/␮l) 0 174 Ϯ 32 (6) 295 Ϯ 19 (3) 356 Ϯ 28 (8) T.P. (mg/dl) 0 5.5 Ϯ 0.2 (6) 5.4 Ϯ 0.2 (3) 5.0 Ϯ 0.1 (7) 7 324 Ϯ 9.8 (5) 281 Ϯ 24 (3) 270 Ϯ 72 (5) 7 4.7 Ϯ 0.1 (4) 4.3 Ϯ 0.3 (3) 3.8 Ϯ 0.2 (4) 14 240 Ϯ 48 (4) 153 Ϯ 72 (2) 304 Ϯ 51 (4) 14 4.7 Ϯ 0.6 (3) 4.7 Ϯ 0.5 (2) 4.0 Ϯ 0.2 (2) 21 253 Ϯ 49 (4) 348 Ϯ 0 (1) 296 Ϯ 71 (4) 21 4.2 Ϯ 0.6 (3) NA 4.0 Ϯ 0.6 (2) Creatinine (mg/dl) 0 0.6 Ϯ 0.04 (6) 0.6 Ϯ 0.06 (3) 0.6 Ϯ 0.03 (7) PTH (pg/ml) These results are pooled for all anesthesia groups. 7 0.5 Ϯ 0.09 (4) 0.3 Ϯ 0.03 (3) 0.4 Ϯ 0.03 (4) 0 27.4 Ϯ 3.8 (17) 14 0.4 Ϯ 0.09 (3) 0.3 Ϯ 0 (2) 0.3 Ϯ 0.03 (3) 750 Ϯ 28.9 (16) 21 0.3 Ϯ 0.003 (3) NA 0.3 Ϯ 0.1 (2) 14 8.9 Ϯ 4.5 (12) BUN (mg/dl) 0 13 Ϯ 1.2 (6) 12 Ϯ 0.7 (3) 11 Ϯ 0.5 (7) 21 18.6 Ϯ 8.9 (9) 710 1.3 (4) 10 1.2 (3) 8 1.2 (4) Ϯ Ϯ Ϯ aValue for all 3 groups different from that for day zero, P < 0.05. 14 11 Ϯ 2.9 (3) 9 Ϯ 1 (2) 6 Ϯ 0.6 (3) bValue different from that for propofol groups, P < 0.05. 21 9 Ϯ 1.8 (3) NA 6.5 Ϯ 1.5 (2) NA = insufficient data; Pentobarb = pentobarbital; prop = propofol; RBC PO (mg/dl) 0 3.7 Ϯ 0.2 (6) 4.3 Ϯ 0.8 (3) 3.9 Ϯ 0.4 (7) 4 = red blood cells; MCH = mean cell hemoglobin; WBC = white blood cells; 7 a 5.9 0.3 (4) 4.6 0.7 (3) 7.2 0.3 (4) Ϯ Ϯ Ϯ BUN = blood urea nitrogen; PO = phosphate; ALP = alkaline phosphatase; 14 a 6.2 Ϯ 0.6 (3) 6.5 Ϯ 0.5 (2) 8.9 Ϯ 0.7 (3) 4 a AST = aspartate transaminase; bili = bilirubin; CK = creatine kinase; 21 5.0 Ϯ 0.2 (3) NA 10.2 Ϯ 1.4 (2) Chol = cholesterol; Trigly = triglycerides; LD = lactate dehydrogenase; ALP (U/L) 0 37 Ϯ 4 (6) 37 Ϯ 5 (3) 58 Ϯ 7 (7) T.P. = (serum) total protein; and PTH = parathyroid hormone. 7 a 302 Ϯ 156 (4) 234 Ϯ 45 (3) 128 Ϯ 7 (4) 14 a 340 Ϯ 107 (3) 366 Ϯ 32 (2) 419 Ϯ 78 (3) 21 a 734 Ϯ 317 (3) NA 574 Ϯ 130 (2) other was exposed to tidal volume of 45 ml/kg owing to tech- nical error. The dog with tracheal rupture was exposed to in these parameters associated with anesthesia regimen. tidal volume of 29 ml/kg. The remaining dogs with pneu- Sodium, potassium, calcium, and magnesium concentra- mothorax had been exposed to tidal volumes of 19.4 and 22 tions (Figure 2A-D) remained stable for all anesthesia regi- ml/kg, with these signs of pneumonia: PaO2 of 69 mm Hg, mens for the duration of the study. and purulent tracheal secretions. The inflammatory pro- cess likely weakened septal membranes and predisposed Discussion dogs to barotrauma. Mortality in this study was 47% (8 of 17), five attribut- Death was attributed to sepsis in two dogs, both on day able to pneumothorax, two to sepsis, and one to sudden 18, and both having high WBC count, hypotension unre- cardiac arrest. Necropsy findings of pulmonary microemboli sponsive to fluid therapy, recurrent metabolic acidosis, hy- associated with surrounding tissue necrosis established poglycemia, and gastrointestinal tract hemorrhage. Anes- increased likelihood of pneumothorax in four dogs, and in thetized dogs in recumbent position and instrumented with the fifth dog indicated tracheal perforation by the endotra- indwelling catheters and endotracheal tubes for a period cheal tube cuff. Ventilator-induced barotrauma may occur of days are susceptible to nosocomial infection from the when tidal volume or pressure is excessive. As described environment or the gastrointestinal tract (6). Infectious previously, tidal volume was set at 15 ml/kg and was ad- complications, as indicated by increasing WBC count or

justed according to values for expired CO2. Mean peak in- fever, developed in 14 of 17 dogs. Nine of the 14 survived the spiratory pressure for this tidal volume, in dogs that did 21-day study; three died of pneumonia-related complications, not develop pneumothorax, was 21.7 Ϯ 3.3 (range, 18.3 to and two died of sepsis-related cardiovascular failure.

29.8) cmH2O. One of the five dogs developing pneumotho- One dog died of sudden cardiovascular collapse on day 2. rax was exposed to tidal volume of 35 ml/kg, airway pres- Death was not associated with ventilator malfunction, pul-

sure of 47 cmH2O, and end-expiratory pressure of 8 cmH2O monary dysfunction, systemic sepsis, or drug administra- because of pneumonia and severe venous admixture. An- tion. The dog had one episode of hypotension on day 1 that

517 Vol 48, No 5 Laboratory Animal Science October 1998 responded to fluid therapy, and another episode just prior the intralipid solvent of propofol, and to ICU-imposed to cardiac arrest. A cause of death could not be ascertained. stresses that altered metabolism of the nutritionally bal- Although this is not a benign protocol, mortality was anced enteral feed. Increases in alkaline phosphatase and greater than anticipated or desired. We began with healthy phosphate values were likely related to the demineraliza- dogs, but the various stresses accumulated with time. Two tion of whole-body disuse in the absence of mobilization of dogs died of iatrogenic causes, and in the future we will limbs or physical therapy. Although calcium metabolism give more attention to details of the protocol. Five died of (e.g., turnover) likely was also altered, its levels are main- infectious complications despite antibiotic therapy. Bacte- tained normal through sensitive homeostatic mechanisms. riologic culture per antimicrobial susceptibility testing was Plasma electrolyte-metabolite findings were similar in this not done, and perhaps the organism was not susceptible to study whether or not muscle relaxants were used. the antibiotics chosen. All catheters were positioned by use Clinically normal mammals confined to bed and sedated of aseptic technique and were maintained using a daily care for prolonged periods can be expected to undergo changes protocol; there were few catheter-site infections, and such in physiologic functions and homeostasis, including muscle catheters were removed as soon as inflammation was no- disuse atrophy (7). Overall this study indicated stability, ticed. The trachea had lavage/suction and fluid and drug without increase in the muscle permeability enzyme cre- administration were performed aseptically, according to pro- atine kinase. Due to the gradual diminution of circulatory tocol, although breaks in technique could have occurred function, complications developed as the period of confine- without our general knowledge. Although propofol was a ment progressed: venous pooling despite periodic change possible source of infection, culture of the solution used in of body position; micro-emboli traveling to the lungs, with the two dogs that died of sepsis failed to yield growth. Mor- resultant small necrotic areas; occasional pneumothorax tality was similar (four of nine) in dogs undergoing gas- during continuous mechanical ventilation; irritated vascu- trointestinal tract sterilization by oral administration of an- lar cannula sites; and invasion by opportunistic bacteria. tibiotics (neomycin, amphotericin, or enrofloxacin), compared The fact that body weight was stable overall could be mis- with that in dogs in which it was not given (four of eight). leading by itself in evaluating fitness of individual dogs, The stresses of this protocol include prolonged depres- because sick dogs tended to retain fluids and develop tis- sion of vital functions related to heavy sedation, mechani- sue edema while otherwise losing tissue mass. It is vital to cal ventilation, and artificial management of various func- detect deterioration early, especially signs of infection, and tions (clearing of tracheal secretions, change in position, initiate prompt treatment; culture and susceptibility data maintenance of normal temperature, fluid balance, nutri- would likely aid in this. tion, clearance of body wastes). All this yields a vulnerable We conclude that definition of the effects of the inten- organism that can develop infection, thromboembolism, tis- sive care milieu on healthy dogs provides worthwhile in- sue irritation, and gastrointestinal tract disturbances, with formation regarding the altered homeostasis that devel- many potential problems. All in all the dogs were stable, if ops and is unrelated to intensive care disease processes sepsis or pneumothorax did not develop. per se. Many aspects of such care can go wrong and under- This study extended beyond earlier approaches that docu- mine the entire effort. Infection, pneumonia, pneumotho- mented electrolyte stability during brief anesthesia and rax, sepsis, thromboembolism, and gastrointestinal tract neuromuscular blockade (2–4). Use of this extended proto- dysfunction are ever-present potential complications. Long- col confirmed prior findings in dogs, as long as dogs did not term survival requires highly trained caregivers, superb develop the various complications. There was sufficient free nutrition, and great attention to details of care (e.g., endo- water and electrolyte content in maintenance fluids and tracheal tube cuff pressure, minor changes in vital signs, adequate gastrointestinal tract uptake to sustain normal and earlier intervention than that indicated during briefer functions in these dogs without continual adjustments in periods of care). We have documented that it is possible to flow rates or use of additive substances. successfully anesthetize and ventilate dogs for 3 weeks; Some plasma and blood values changed with time: he- when all goes well, it goes well indeed. Most, if not all, of matocrit decreased, and WBC, cholesterol, alkaline phos- the problems associated with this model can be managed phatase, fibrinogen, and triglyceride values increased. Ex- under the proper circumstances. This model may be appro- planations for these changes are hampered by the lack of priate for experimental investigation of the ICU milieu. ICU studies in normal subjects, because literature related to ICU care is focused on sick humans, and abnormal values have been related directly to disease processes. With Acknowledgements these considerations in mind, we explain the aforemen- We appreciate the cooperative and comprehensive efforts of tioned variations as follows: hematocrit was diminished due those who aided in meeting the challenges of these labor-inten- to multiple and continued blood sampling, although use of sive studies: our superb conscientious part- and full-time technicians, and our volunteers, consisting of veterinarians, students, propofol was associated with greater decreases. We do not physicians, and others. Brock Lewis; Neal Fleming, M.D., Ph.D.; know whether propofol has effects on canine erythropoie- Freda Hwang, M.S.; Kameron Chun, B.S.; and Renae Wurschmidt, sis in dogs. Increases in WBC numbers and fibrinogen con- R.N. contributed magnificently to this project. centration likely were due to continued stress and inflam- This work was supported in part by Zeneca Pharmaceuticals, Inc. mation, as an acute phase reaction. Increases in cholesterol and triglyceride concentrations may have been due to

518 ICU Stability in Sedated Ventilated Dogs

5. Fleming, N. W., G. A. Gronert, S. C. Haskins, et al. 1996. References Sedation and mechanical ventilation without paralysis pro- 1. Charbon, G. A., and M. H. Hoekstra. 1962. Mineral con- duces resistance to metocurine and increases muscle acetyl- tent in plasma and blood cells of various species. Acta Physiol. choline receptors in beagles. Anesth. Analg. 82:S112. Pharmacol. Neerlandica 10:209–214. 6. Bolton, C. F. 1996. Sepsis and the systemic inflammatory 2. Stevenson, D. E. 1960. Changes caused by anesthesia in the response syndrome: neuromuscular manifestations. Crit. Care blood electrolytes of the dog. Br. J. Anaesth. 32:353–363. Med. 24:1408–1416. 3. Stevenson, D. E. 1960. Changes in the blood electrolytes of 7. Faragher, M. W., B. J. Day, and X. Dennett. 1996. Critical anesthetized dogs caused by suxamethonium. Br. J. Anaesth. care myopathy: an electrophysiological and histological study. 32:364–371. Muscle Nerve 19:516–518. 4. Stevenson, D. E. 1960. Observations on the effects of d-tu- bo-curarine on the blood electrolytes of the dog. Br. J. Anaesth. 32:372–383.

519