In a search of viral agents in pulmonary macrophages of horses with chronic pulmonary disease, equine herpesvirus 2 was found to be unique. In 8 of 9 horses with chronic pulmonary disease, antigens of equine herpesvirus 2 were detected by indirect immunofluorescence staining of scattered foamy macrophages immediately after harvesting by bronchoalveolar lavage and fractionation on metrizamide gradients. In a healthy horse, antigens were not found. After 1 week of cultivation of bronchoalveolar lavage cells from a second group of 9 horses with chronic pulmonary disease, viral antigens were detected in 90% of the adherent pulmonary macrophages. In 2 of 3 healthy horses, viral antigens also were found in 90% of the adherent pulmonary macrophages. Antigens of equine herpesvirus 1, equine herpesvirus 4, parainfluenza virus 3, or adenovirus were not detected. Antigens of the 5 investigated viruses also were not detected in lung tissue slices from a third group of 14 horses, 4 healthy; 7 with varying degrees of bronchiolitis, 2 of which also had chronic intestitial pneumonia; 2 with eosinophilic bronchitis; and 1 with pulmonary hemorrhage. The exclusive presence of equine herpesvirus 2 in pulmonary macrophages was confirmed qualitatively by isolation of infective virus by cocultivation. In a fourth group of 12 horses with chronic pulmonary disease, infective virus could be isolated from pulmonary macrophages of 3 horses and from mixed-blood leukocytes of 5 horses. Virus isolations from 2 healthy horses were not successful from pulmonary macrophages, whereas 1 isolation was obtained from mixed-blood leukocytes. Other viruses were not detected by cocultivation.

In a search of viral agents in pulmonary macrophages of horses with chronic pulmonary disease, equine herpesvirus 2 was found to be unique. In 8 of 9 horses with chronic pulmonary disease, antigens of equine herpesvirus 2 were detected by indirect immunofluorescence staining of scattered foamy macrophages immediately after harvesting by bronchoalveolar lavage and fractionation on metrizamide gradients. In a healthy horse, antigens were not found. After 1 week of cultivation of bronchoalveolar lavage cells from a second group of 9 horses with chronic pulmonary disease, viral antigens were detected in 90% of the adherent pulmonary macrophages. In 2 of 3 healthy horses, viral antigens also were found in 90% of the adherent pulmonary macrophages. Antigens of equine herpesvirus 1, equine herpesvirus 4, parainfluenza virus 3, or adenovirus were not detected. Antigens of the 5 investigated viruses also were not detected in lung tissue slices from a third group of 14 horses, 4 healthy; 7 with varying degrees of bronchiolitis, 2 of which also had chronic intestitial pneumonia; 2 with eosinophilic bronchitis; and 1 with pulmonary hemorrhage. The exclusive presence of equine herpesvirus 2 in pulmonary macrophages was confirmed qualitatively by isolation of infective virus by cocultivation. In a fourth group of 12 horses with chronic pulmonary disease, infective virus could be isolated from pulmonary macrophages of 3 horses and from mixed-blood leukocytes of 5 horses. Virus isolations from 2 healthy horses were not successful from pulmonary macrophages, whereas 1 isolation was obtained from mixed-blood leukocytes. Other viruses were not detected by cocultivation.

Free, autogenous, full-thickness skin grafts were applied to 10 dogs; 5 dogs were given an iron chelator, deferoxamine-10% hydroxyethyl pentafraction starch (DEF-HES; 50 mg/kg of body weight, IV), and 5 dogs were given an equal volume of 10% hydroxyethyl pentafraction starch (HES) in 0.9% saline solution (5 ml/kg, IV). All dogs (DEF-HES/HBO- and HES/HBO-treated) were exposed to 60 minutes of hyperbaric oxygen (HBO) at 2 atmospheres absolute pressure twice daily for 10 days, beginning the day of surgery. The percentage of viable graft on day 10 was lower in HES/HBO-treated-dogs (mean +/- SD, 13.3 +/- 21.3%; median, 3.0%) than in DEF-HES/HBO-treated dogs (64.7 +/- 39.2%; 88.3%; P = 0.095, Mann-Whitney two-tailed test). There was a positive correlation between percentage of viable graft (on day 10) and percentage of haired skin on the graft site (on day 28) for all dogs (r = 0.91) and for HES/HBO-treated dogs (r = 0.97). The DEF-HES/HBO-treated dogs had less consistent correlation (r = 0.67). Perivascular aggregates of foamy cells were observed in the superficial and reticular portions of the dermis and in the subcutaneous tissue on both surfaces of the panniculus muscle in the graft sites of DEF-HES/HBO-treated dogs. These cells were also observed in the dermis, but not subcutaneous tissue of the control skin sections, and in some viscera of DEF-HES/HBO-treated dogs. Deferoxamine appears to attenuate the detrimental effect of HBO and HES on survival of free skin grafts. However, clinical use of HBO is not recommended as adjunct treatment for free skin grafts in dogs in the first 10 days after grafting. Administration of DEF-HES is not recommended because it has failed to improve the survival of free skin grafts, and the consequence of the cellular response seen in this study is undetermined.

Free, autogenous, full-thickness skin grafts were applied to 10 dogs; 5 dogs were given an iron chelator, deferoxamine-10% hydroxyethyl pentafraction starch (DEF-HES; 50 mg/kg of body weight, IV), and 5 dogs were given an equal volume of 10% hydroxyethyl pentafraction starch (HES) in 0.9% saline solution (5 ml/kg, IV). All dogs (DEF-HES/HBO- and HES/HBO-treated) were exposed to 60 minutes of hyperbaric oxygen (HBO) at 2 atmospheres absolute pressure twice daily for 10 days, beginning the day of surgery. The percentage of viable graft on day 10 was lower in HES/HBO-treated-dogs (mean +/- SD, 13.3 +/- 21.3%; median, 3.0%) than in DEF-HES/HBO-treated dogs (64.7 +/- 39.2%; 88.3%; P = 0.095, Mann-Whitney two-tailed test). There was a positive correlation between percentage of viable graft (on day 10) and percentage of haired skin on the graft site (on day 28) for all dogs (r = 0.91) and for HES/HBO-treated dogs (r = 0.97). The DEF-HES/HBO-treated dogs had less consistent correlation (r = 0.67). Perivascular aggregates of foamy cells were observed in the superficial and reticular portions of the dermis and in the subcutaneous tissue on both surfaces of the panniculus muscle in the graft sites of DEF-HES/HBO-treated dogs. These cells were also observed in the dermis, but not subcutaneous tissue of the control skin sections, and in some viscera of DEF-HES/HBO-treated dogs. Deferoxamine appears to attenuate the detrimental effect of HBO and HES on survival of free skin grafts. However, clinical use of HBO is not recommended as adjunct treatment for free skin grafts in dogs in the first 10 days after grafting. Administration of DEF-HES is not recommended because it has failed to improve the survival of free skin grafts, and the consequence of the cellular response seen in this study is undetermined.

Six horses were subjected to 3 hours of low-flow ischemia and 3 hours of reperfusion of the large colon. After induction of anesthesia, the large colon was exteriorized through a ventral midline celiotomy. Colonic blood flow was measured continuously, using Doppler ultrasonic flow probes placed on the colonic arteries supplying the dorsal and ventral colons and was allowed to stabilize for 15 to 30 minutes after instrumentation. Low-flow ischemia was induced by reducing colonic arterial blood flow to 20% of baseline (BL) flow. Colonic mucosal, seromuscular, and full-thickness blood flow were determined on a tissue-weight basis by injecting colored microspheres proximally into the colonic artery supplying the ventral colon. Reference blood samples were obtained at a known flow rate from the colonic artery and vein at a site more distal to the site of injection. Left ventral colon biopsy specimens were harvested at BL, 3 hours of ischemia, and 15 minutes of reperfusion. Blood and tissue samples were digested and filtered to collect the microspheres, and dimethylformamide was added to release the colored dyes. Dye concentration in blood and tissue samples was measured by use of spectrophotometry, and tissue-blood flow was calculated. Data were analyzed, using two-way ANOVA for repeated measures; statistical significance was set at P < 0.05. Doppler blood flow decreased to approximately 20% of BL, whereas microsphere blood flow ranged between 13.7 and 15.5% of BL at 3 hours of ischemia. Doppler-determined blood flow increased immediately on restoration of blood flow, reached 183% of BL at 15 minutes of reperfusion, and remained at or above BL throughout 3 hours of reperfusion. This reactive hyperemia was also detected, using the colored microspheres; blood flow increased to 242 and 327% of BL at 15 minutes of reperfusion in the mucosal and seromuscular layers, respectively. Mucosal blood flow was not different from seromuscular blood flow at any time, indicating relatively equal distribution of blood flow between these 2 layers. As determined from the venous reference samples, there was no evidence of arteriovenous anastomoses.

Six horses were subjected to 3 hours of low-flow ischemia and 3 hours of reperfusion of the large colon. After induction of anesthesia, the large colon was exteriorized through a ventral midline celiotomy. Colonic blood flow was measured continuously, using Doppler ultrasonic flow probes placed on the colonic arteries supplying the dorsal and ventral colons and was allowed to stabilize for 15 to 30 minutes after instrumentation. Low-flow ischemia was induced by reducing colonic arterial blood flow to 20% of baseline (BL) flow. Colonic mucosal, seromuscular, and full-thickness blood flow were determined on a tissue-weight basis by injecting colored microspheres proximally into the colonic artery supplying the ventral colon. Reference blood samples were obtained at a known flow rate from the colonic artery and vein at a site more distal to the site of injection. Left ventral colon biopsy specimens were harvested at BL, 3 hours of ischemia, and 15 minutes of reperfusion. Blood and tissue samples were digested and filtered to collect the microspheres, and dimethylformamide was added to release the colored dyes. Dye concentration in blood and tissue samples was measured by use of spectrophotometry, and tissue-blood flow was calculated. Data were analyzed, using two-way ANOVA for repeated measures; statistical significance was set at P < 0.05. Doppler blood flow decreased to approximately 20% of BL, whereas microsphere blood flow ranged between 13.7 and 15.5% of BL at 3 hours of ischemia. Doppler-determined blood flow increased immediately on restoration of blood flow, reached 183% of BL at 15 minutes of reperfusion, and remained at or above BL throughout 3 hours of reperfusion. This reactive hyperemia was also detected, using the colored microspheres; blood flow increased to 242 and 327% of BL at 15 minutes of reperfusion in the mucosal and seromuscular layers, respectively.

Effects of temperature and storage time on canine bone-transfixation pin specimens were tested by comparing pin pull-out forces. A total of 16 femurs from 8 mature dogs were tested. Five nonthreaded Steinmann pins were placed through both cortices in the diaphysis of each femur. The femurs were then sectioned transversely between each pin, with a bonepin specimen placed evenly into each of 5 groups prior to biomechanical testing. Four bone-pin specimen groups were stored at -20 or -70 C for 14 or 28 days, while 1 specimen group was immediately tested. Pull-out forces for frozen groups were compared with pull-out forces for the fresh group. Using two-way ANOVA, there was no statistical difference in mean axial-extraction forces among bonepin specimen in any of the tested groups. It is concluded that acute pin pull-out forces are not significantly affected by freezing temperature or time. However, specimens stored at -20 C for as few as 14 days had a trend for increased pull-out forces, compared with freshly harvested specimens. Therefore, the authors recommend storage of bone-pin specimens at -70 C when possible.

Effects of temperature and storage time on canine bone-transfixation pin specimens were tested by comparing pin pull-out forces. A total of 16 femurs from 8 mature dogs were tested. Five nonthreaded Steinmann pins were placed through both cortices in the diaphysis of each femur. The femurs were then sectioned transversely between each pin, with a bonepin specimen placed evenly into each of 5 groups prior to biomechanical testing. Four bone-pin specimen groups were stored at -20 or -70 C for 14 or 28 days, while 1 specimen group was immediately tested. Pull-out forces for frozen groups were compared with pull-out forces for the fresh group. Using two-way ANOVA, there was no statistical difference in mean axial-extraction forces among bonepin specimen in any of the tested groups. It is concluded that acute pin pull-out forces are not significantly affected by freezing temperature or time. However, specimens stored at -20 C for as few as 14 days had a trend for increased pull-out forces, compared with freshly harvested specimens. Therefore, the authors recommend storage of bone-pin specimens at -70 C when possible.

Twenty-four horses were randomly allocated to 3 groups. Horses were anesthetized, subjected to a ventral midline celiotomy, and the large colon was exteriorized and instrumented. Group-1 horses served as sham-operated controls. Group-2 horses were subjected to 6 hours of low-flow colonic arterial ischemia, and group-3 horses were subjected to 3 hours of ischemia and 3 hours of reperfusion. Baseline (BL) samples were collected, then low-flow ischemia was induced by reducing ventral colonic arterial blood flow to 20% of BL. All horses were monitored for 6 hours after BL data were collected. Blood samples were collected from the colonic vein and main pulmonary artery (systemic venous [SV]) for measurement of plasma endotoxin, 6-keto prostaglandin F1alpha (6-kPG), thromboxane B2 (TXB2), and prostaglandin E2 (PGE2) concentrations. Tumor necrosis factor and interleukin-6 activities were measured in colonic venous (CV) serum samples. Data were analyzed, using two-was ANOVA, and post-hoc comparisons were made, using Dunnett's and Tukey's tests. Statistical significance was set at P < 0.05 Endotoxin was not detected in CV or SV plasma at any time. There was no detectable tumor necrosis factor or interleukin-6 activity in CV samples at any time. There were no differences at BL among groups for CV or SV 6-kPG, PGE2, or TXB2 concentrations, nor were there any changes across time in group-1 horses. Colonic venous 6-kPG concentration increased during ischemia in horses of groups 2 and 3; CV 6-kPG concentration peaked at 3 hours in group-3 horses, then decreased during reperfusion, but remained increased through 6 hours in group-2 horses. Systemic venous 6-kPG concentration increased during reperfusion in group-3 horses, but there were no changes in group-2 horses. Colonic venous PGE2 concentration increased during ischemia in horses of groups 2 and 3, and remained increased for the first hour of reperfusion in group-3 horses and for the 6-hour duration of ischemia in group-2 horses. There were no temporal alterations in SV PGE2 concentration. There was no difference in CV or SV TXB2 concentration among or within groups across time; however, there was a trend (P = 0.075) toward greater CV TXB2 concentration at 3.25 hours, compared with BL, in group-3 horses. Eicosanoid concentrations were significantly lower in SV, compared with CV plasma. Prostaglandin E2 and 6-kPG concentrations were approximately 3 to 8 and 5 to 10 times greater, respectively, in CV than in SV plasma. The increased concentrations of 6-kPG and PGE2 in CV plasma were likely attributable to their accumulation secondary to colonic ischemia. The increased values of these vasodilator eicosanoids may have a role in the reactive hyperemia observed during reperfusion. The increased 6-kPG concentration in SV plasma may represent spillover from the colonic vasculature, but more likely reflects systemic production.

Twenty-four horses were randomly allocated to 3 groups. Horses were anesthetized, subjected to a ventral midline celiotomy, and the large colon was exteriorized and instrumented. Group-1 horses served as sham-operated controls. Group-2 horses were subjected to 6 hours of low-flow colonic arterial ischemia, and group-3 horses were subjected to 3 hours of ischemia and 3 hours of reperfusion. Baseline (BL) samples were collected, then low-flow ischemia was induced by reducing ventral colonic arterial blood flow to 20% of BL. All horses were monitored for 6 hours after BL data were collected. Blood samples were collected from the colonic vein and main pulmonary artery (systemic venous (SV) for measurement of plasma endotoxin, 6-keto prostaglandin F1alpha (6-kPG), thromboxane B2 (TXB2), and prostaglandin E2 (PGE2) concentrations. Tumor necrosis factor and interleukin-6 activities were measured in colonic venous (CV) serum samples. Data were analyzed, using two-was ANOVA, and post-hoc comparisons were made, using Dunnett's and Tukey's tests. Statistical significance was set at P < 0.05 Endotoxin was not detected in CV or SV plasma at any time. There was no detectable tumor necrosis factor or interleukin-6 activity in CV samples at any time. There were no differences at BL among groups for CV or SV 6-kPG, PGE2, or TXB2 concentrations, nor were there any changes across time in group-1 horses. Colonic venous 6-kPG concentration increased during ischemia in horses of groups 2 and 3; CV 6-kPG concentration peaked at 3 hours in group-3 horses, then decreased during reperfusion, but remained increased through 6 hours in group-2 horses. Systemic venous 6-kPG concentration increased during reperfusion in group-3 horses, but there were no changes in group-2 horses. Colonic venous PGE2 concentration increased during ischemia in horses of groups 2 and 3, and remained increased for the first hour of reperfusion in group-3 horses and for the 6-hour duration of ischemia in group-2 hors.

Because caffeine is metabolized by the hepatic P-450 cytochrome oxidase system, clearance of caffeine is an excellent quantitative test of hepatic function in human beings. It is currently used in much the same way that creatinine clearance is used to assess renal function. Caffeine clearance was measured in lactating dairy cows initially to determine the suitability of caffeine clearance as an indicator of hepatic function in cattle. Pharmacokinetic variables of caffeine were studied in 6 adult lactating dairy cows after IV administration of a single dose of caffeine sodium benzoate (2 mg of caffeine/kg of body weight). Caffeine concentration was analyzed by use of an automated enzyme immunoassay. The lower limit of detection of the assay for caffeine in serum was 0.079 micrograms/ml. Serum caffeine concentration-time curves best fit an open two-compartment pharmacokinetic model. Harmonic mean elimination half-life was 3.8 (range, 2.6 to 6.9) hours, and total clearance was 0.118 (range, 0.090 to 0.197) L/kg/h. Milk caffeine concentration was similar to serum concentration 1.5 to 24 hours after caffeine administration. Adverse effects were not observed in cows given caffeine.

Because caffeine is metabolized by the hepatic P-450 cytochrome oxidase system, clearance of caffeine is an excellent quantitative test of hepatic function in human beings. It is currently used in much the same way that creatinine clearance is used to assess renal function. Caffeine clearance was measured in lactating dairy cows initially to determine the suitability of caffeine clearance as an indicator of hepatic function in cattle. Pharmacokinetic variables of caffeine were studied in 6 adult lactating dairy cows after IV administration of a single dose of caffeine sodium benzoate (2 mg of caffeine/kg of body weight). Caffeine concentration was analyzed by use of an automated enzyme immunoassay. The lower limit of detection of the assay for caffeine in serum was 0.079 micrograms/ml. Serum caffeine concentration-time curves best fit an open two-compartment pharmacokinetic model. Harmonic mean elimination half-life was 3.8 (range, 2.6 to 6.9) hours, and total clearance was 0.118 (range, 0.090 to 0.197) L/kg/h. Milk caffeine concentration was similar to serum concentration 1.5 to 24 hours after caffeine administration. Adverse effects were not observed in cows given caffeine.

Immunoglobulin E is produced in response to parasitic nematodes that undergo blood and tissue migrations. Results of our previous studies indicated that IgE and IgG respond to Dirofilaria immitis in experimentally infected dogs. To determine the association between treatment with the larvicide, diethylcarbamazine (DEC), and antibody responses and to examine the potential influence of infection with a nonfilarid intestinal nematode on isotype-specific immune responses, we monitored, by use of isotype-specific ELISA, separate IgE and IgG responses against D immitis in 4 groups (A-D) of 8 dogs experimentally coinfected with D immitis and Ancylostoma caninum. All dogs were monitored from 2 weeks before inoculation with D immitis, through postinoculation (PI) week 20. Group-B dogs received a daily regimen of 6.6 mg of DEC/kg of body weight. Group-C dogs received 4.95 mg of oxibendazole/kg daily. Group-D dogs received DEC and oxibendazole, equivalent to the daily doses given to dogs of groups B and C. All dogs given oxibendazole had no A caninum at necropsy. Of the groups receiving DEC, 3 group-B dogs each had 1 to 2 D immitis at necropsy. When results of chronologic IgE determination for all groups were statistically compared, only groups B and C had significant (P = 0.0148 and P << 0.00005, respectively) increases in IgE values. Group-C dogs had the highest IgE values from PI week 10 until the end of the study, whereas IgG values were statistically identical to those of group-A dogs. Group-B dogs given only DEC and having the least number of D immitis of all groups, had IgE values that peaked at PI week 6; values were significantly (P = 0.0002) higher than those for all other groups. In Group-B dogs, IgG values increased significantly (P << 0.00005) only at PI week 20 and were significantly (P << 0.00005) decreased after PI week 6, compared with values for all other groups. Group D containing 6 dogs infected with 1 to 18 D immitis found at necropsy had IgE values between those of group-B dogs and those of non-DEC-treated groups at PI week 6. There was no difference in IgG values between 3 groups at PI week 6, and IgE values were found to be a better correlator than were IgG values to the number of D immitis larvae killed in the tissues during this period. All differences in IgG and IgE values not only correlated with treatment status and number of D immitis adults found at necropsy, but also with the developmental stage of D immitis commonly present in the groups at each time point and the number of adult D immitis found at necropsy.

Immunoglobulin E is produced in response to parasitic nematodes that undergo blood and tissue migrations. Results of our previous studies indicated that IgE and IgG respond to Dirofilaria immitis in experimentally infected dogs. To determine the association between treatment with the larvicide, diethylcarbamazine (DEC), and antibody responses and to examine the potential influence of infection with a nonfilarid intestinal nematode on isotype-specific immune responses, we monitored, by use of isotype-specific ELISA, separate IgE and IgG responses against D immitis in 4 groups (A-D) of 8 dogs experimentally coinfected with D immitis and Ancylostoma caninum. All dogs were monitored from 2 weeks before inoculation with D immitis, through postinoculation (PI) week 20. Group-B dogs received a daily regimen of 6.6 mg of DEC/kg of body weight. Group-C dogs received 4.95 mg of oxibendazole/kg daily. Group-D dogs received DEC and oxibendazole, equivalent to the daily doses given to dogs of groups B and C. All dogs given oxibendazole had no A caninum at necropsy. Of the groups receiving DEC, 3 group-B dogs each had 1 to 2 D immitis at necropsy. When results of chronologic IgE determination for all groups were statistically compared, only groups B and C had significant (P = 0.0148 and P << 0.00005, respectively) increases in IgE values. Group-C dogs had the highest IgE values from PI week 10 until the end of the study, whereas IgG values were statistically identical to those of group-A dogs. Group-B dogs given only DEC and having the least number of D immitis of all groups, had IgE values that peaked at PI week 6; values were significantly (P = 0.0002) higher than those for all other groups. In Group-B dogs, IgG values increased significantly (P << 0.00005) only at PI week 20 and were significantly (P << 0.00005) decreased after PI week 6, compared with values for all other groups. Group D containing 6 dogs infected with 1 to 18 D immitis found at necropsy had IgE values betwee.

Intraocular pressure (IOP) was measured by use of Mackay-Marg applanation tonometry in 8 normal, manometrically controlled, enucleated, canine eyes with and without 1 of 2 piano therapeutic soft contact lenses (1 and 2) covering the cornea. Differences were not significant between measurements made without a contact lens and those made through either lens at manometer IOP < 30 mm of Hg. At manometer IOP greater than or equal to 30 mm of Hg, use of a contact lens tended to result in a statistically greater (P < 0.05) estimate of IOP than when a lens was not used. This difference, however, achieved only a maximum of 2.6 mm of Hg at the 80 mm of Hg value, and was not regarded as clinically important. Measurements obtained through lens 1 were not significantly different from those obtained through lens 2. The IOP can be accurately estimated in dogs; using the Mackay-Marg tonometer, without removing either type of bandage soft contact lens, thereby avoiding potential disruption of an already compromised cornea.

Intraocular pressure (IOP) was measured by use of Mackay-Marg applanation tonometry in 8 normal, manometrically controlled, enucleated, canine eyes with and without 1 of 2 piano therapeutic soft contact lenses (1 and 2) covering the cornea. Differences were not significant between measurements made without a contact lens and those made through either lens at manometer IOP < 30 mm of Hg. At manometer IOP greater than or equal to 30 mm of Hg, use of a contact lens tended to result in a statistically greater (P < 0.05) estimate of IOP than when a lens was not used. This difference, however, achieved only a maximum of 2.6 mm of Hg at the 80 mm of Hg value, and was not regarded as clinically important. Measurements obtained through lens 1 were not significantly different from those obtained through lens 2. The IOP can be accurately estimated in dogs; using the Mackay-Marg tonometer, without removing either type of bandage soft contact lens, thereby avoiding potential disruption of an already compromised cornea.

Vestibulotoxic and ototoxic effects often are seen after long-term, high-dose systemic treatment with gentamicin, but toxic effects after topical use have not been reported in animals, to the authors' knowledge. Vestibular and auditory effects of twice daily otic gentamicin treatment for 21 days were evaluated in 10 dogs with intact tympanic membranes and in the same 10 dogs after experimental bilateral myringotomy. Each dog served as its own control; 7 drops of gentamicin sulfate (3 mg/ml in a buffered aqueous vehicle) were placed in 1 ear, and 7 drops of vehicle were placed in the opposite ear. Treatment and control ears were reversed after myringotomy. Vestibular function was evaluated daily by neurologic examination and behavioral assessment Auditory function was evaluated twice weekly by determination of brain stem auditory evoked potentials. Gentamicin sulfate placed in the ear of clinically normal dogs with intact or ruptured tympanic membranes, in the quantities used in this study, did not induce detectable alteration of cochlear or vestibular function. Serum gentamicin concentration after 21 days of treatment was detectable in only 2 dogs and was an order of magnitude below documented toxic concentrations.

Vestibulotoxic and ototoxic effects often are seen after long-term, high-dose systemic treatment with gentamicin, but toxic effects after topical use have not been reported in animals, to the authors' knowledge. Vestibular and auditory effects of twice daily otic gentamicin treatment for 21 days were evaluated in 10 dogs with intact tympanic membranes and in the same 10 dogs after experimental bilateral myringotomy. Each dog served as its own control; 7 drops of gentamicin sulfate (3 mg/ml in a buffered aqueous vehicle) were placed in 1 ear, and 7 drops of vehicle were placed in the opposite ear. Treatment and control ears were reversed after myringotomy. Vestibular function was evaluated daily by neurologic examination and behavioral assessment Auditory function was evaluated twice weekly by determination of brain stem auditory evoked potentials. Gentamicin sulfate placed in the ear of clinically normal dogs with intact or ruptured tympanic membranes, in the quantities used in this study, did not induce detectable alteration of cochlear or vestibular function. Serum gentamicin concentration after 21 days of treatment was detectable in only 2 dogs and was an order of magnitude below documented toxic concentrations.