Materials and Methods

Dogs—Forty client-owned sexually intact healthy female dogs (age range, 7 months to 10 years) that were admitted for neutering at the Faculty of Veterinary Medicine in Utrecht, The Netherlands, were included in the study. The dogs were of several breeds. These included mixed breed (n = 12), Labrador Retriever (8), German Shepherd Dog (4), Belgian Shepherd (2), Golden Retriever (2), American Bulldog (2), and 1 of each of the following breeds: Beagle, Heidewachtel, Stabyhoun, American Staffordshire Terrier, Boxer, Rottweiler, Münsterlander, English Cocker Spaniel, Viszla, and Hungarian Mudi. Dogs were determined to be healthy on the basis of medical history and physical examination, and enrollment criteria included a requirement that the most recent estrus had ended ≥ 6 weeks prior to the date of surgery. Age was recorded and a BCS^a was determined for each dog (range, 1 [emaciated] to 5 [obese]).^15,16 Written consent was obtained from all owners before each dog was randomly assigned to either the OVE or OVH surgical treatment groups. The study protocol was reviewed and approved by the Ethics and Research Committee of the Department of Clinical Sciences of Companion Animals.

The randomization procedure for group assignment was as follows: before the start of the study, 40 sheets of paper marked OVH (20 sheets) or OVE (20) were placed in 40 identical envelopes, and the envelopes were sealed. The envelopes were mixed and thereafter numbered from 1 to 40. Numbers 1 through 40 were assigned to each animal consecutively at the time of enrollment in the study. After induction of anesthesia, the corresponding numbered envelope was opened.

Anesthesia and perioperative pain management— Hair was clipped over a cephalic vein and the skin was aseptically prepared before an IV catheter^b was placed. Dogs were premedicated with medetomidine hydro-chloride^c (1 mg/m², IV) and carprofen^d (4 mg/kg [1.82 mg/lb], IV). Anesthesia was induced with propofol^e (1 to 2 mg/kg [0.45 to 0.91 mg/lb], IV, to effect). Dogs were intubated, and anesthesia was maintained with isoflurane^f (sufficient to maintain the required plane of anesthesia in each patient) in oxygen and air. Intermittent positive pressure ventilation was provided to ensure normocapnea (end-tidal partial pressure of carbon dioxide, 4.5 to 5.5 kPa). Lactated Ringer's solution^g was infused IV at a rate of 10 mL/kg/h (4.5 mL/lb/h) until the patient was extubated. Intraoperative monitoring included ECG, capnography, and assessments of body temperature via esophageal probe and concentrations of oxygen and isoflurane^f determined by use of a sensor located inside the ventilator. If the duration of a surgical procedure was > 1 hour, half of the originally administered dose of medetomidine^c was given IV 1 hour after the initial dose was administered. After surgery, atipamezole hydrochloride^h (2.5 μg/m², IM) was administered as a medetomidine antagonist.

Blood and urine sample collection—After the induction of anesthesia, hair was clipped over a jugular vein; the skin was aseptically prepared, and an IV jugular catheterⁱ was placed. A urinary catheterj was also placed at that time. Immediately before the midline abdominal skin incision was begun and immediately prior to its closing, blood (10-mL) and urine (10-mL) samples were obtained via these catheters for another research study. To confirm the assessment that dogs were healthy, the following biochemical and CBC variables were analyzed in heparinized plasma and blood samples obtained for the other study: BUN concentration (reference range, 3.0 to 12.5 mmol/L); alkaline phosphatase activity (reference range, < 73 U/L) and creatinine (reference range, 50 to 129 μmol/L), bile acids (reference range, < 10 μmol/L), phosphorus (reference range, 0.65 to 2.12 mmol/L), sodium (reference range, 141 to 150 mmol/L), and potassium (reference range, 3.6 to 5.6 mmol/L) concentrations; plasma total calcium concentration (reference range, 1.98 to 2.97 mmol/L); Hct (reference range, 42% to 61%); total leukocyte count (reference range, 4.5 × 10⁹ cells/L to 14.6 × 10⁹ cells/L); and platelet count (reference range, 144 ×10⁹ platelets/L to 603 × 10⁹platelets/L).

Surgical procedures—All surgeries were performed by 1 board-certified surgeon with > 25 years of experience performing canine OVH and OVE procedures (MEP), with the help of 1 assistant (various veterinary students). Standardized surgical protocols were used for all procedures.

Hair was clipped at the surgical site, and the skin was aseptically prepared. A ventral median celiotomy was performed with the incision starting at the cranial border of the umbilicus. The initial length of the incision was estimated by the surgeon and extended during surgery if it was deemed necessary according to the following criteria: after cutting the suspensory ligaments, both ovarian pedicles had to be ligated with as little traction as possible; for OVH procedures, the uterine body had to be ligated without traction; and visual inspection of the ovarian and uterine pedicles in their normal anatomic positions had to be possible before the abdominal incision was closed. Each skin incision was made with a standard No. 10 scalpel blade, and subcutaneous tissues were dissected at the midline and separated from the abdominal fascia with Mayo scissors. Hemostasis was achieved by use of monopolar electrosurgical equipment.^k

A small incision was made in the linea alba and extended in either direction with scissors to access the abdominal cavity. The left uterine horn was located and retracted caudoventrally to expose the ovarian bursa and suspensory ligament. The suspensory ligament was coagulated and then cut with scissors approximately 5 mm cranial to the ovarian bursa. The broad ligament was perforated at the level of the ovarian bursa, and a Baby-Mixter clamp (185 mm) was placed on the proper ligament of the ovary, including the anastomosing vessels between the ovarian and uterine blood supplies. The mesovarium (including the ovarian artery and vein) was double ligated with a single length of size-0 polyglactin 910 suture^l dorsal to the ovarian bursa; after a classic surgical knot was tied, the 2 suture ends were used to encircle the pedicle again, and a second surgical knot was made at the same location. To ensure complete resection of ovaries, each ovarian bursa was opened with scissors to allow visual inspection of the ovary before the mesovarium was cut. For OVE, the cranial tip of the uterine horn was single ligated together with the uterine artery and uterine vein by use of a size-0 polyglactin 910 suture^l immediately distal to the proper ligament. The proper ligament was cut approximately 2 mm from the suture with scissors to complete the OVE. The left uterine horn was returned to its anatomic position, and the procedure was repeated on the right side.

For OVH, the ligation technique and the dissection of the mesovarium were performed as for OVE. After the mesovarium was cut, the broad ligaments were torn parallel to the uterine vessels toward the cervix. The round ligaments were cut with scissors, and small vessels were cauterized if bleeding occurred. The uterine body, uterine artery, and uterine vein were double ligated with a single length of size-0 polyglactin 910 suture^l immediately cranial to the cervix. A Baby-Mixter clamp (185 mm) was applied cranial to the ligature site, and the uterine body was divided with scissors between the suture and the clamp. The uterus was removed, and the uterine pedicle was repositioned in its normal anatomic position.

All ligated vessels were inspected (in their normal anatomic positions) for bleeding, and surgical sponges were counted before the abdominal incision was closed. The abdominal fascia was closed with size-0 polydioxanone suture^m material in a simple continuous pattern. The subcutaneous and subdermal layers were closed with 3–0 poliglecaprone 25 sutureⁿ in a continuous pattern.

Postoperative care—All dogs were monitored until fully recovered from anesthesia and remained at the clinic for 24 to 32 hours after surgery. Buprenorphine^o (10 μg/kg [4.5 μg/lb], IV) was administered for acute pain management approximately 40 minutes before administration of atipamezole^h and was subsequently given SC every 6 hours during the 24 hours after surgery. Dogs hospitalized for > 24 hours were administered carprofen^d (2 mg/kg, PO, q 12 h) after buprenorphine was discontinued. Long and short forms of the GCMPS^o were used to evaluate the severity of postoperative pain; these are modifications^17,18 of a prototype developed by Holton et al.¹⁹ Both forms were used to ensure that estimation of pain (via assessment of demeanor, posture, comfort, vocalization, attention to the surgical wound, mobility, and response to touch) was as thorough as possible. Descriptor weights^17,o were applied for each answer on the long form and were totaled to obtain an overall pain score. Rescue analgesia was provided according to established protocols in patients with pain scores > 15 (range, 0 to 24) on the short form by increasing the dose of buprenorphine^p to 20 μg/kg (9.1 μg/lb). When dogs were discharged from the hospital, owners were instructed to administer carprofen^d (2 mg/kg, PO, q 12 h) for the next 2 days.

Measurement of surgical variables—Blood loss was estimated by determining the weight of used surgical sponges after surgery.²⁰ Dry cotton gauze sponges^q that were weighed before use were removed from the operating table immediately after use and kept in a sealed airtight bag^r to be weighed after surgery was completed. The BL_rel was defined as (blood loss [mL])/(BW [kg] × 0.1), where 1 g was accepted as the equivalent of 1 mL; the result was expressed as a percentage.²¹

The time between predetermined stages of the surgical procedure was measured by use of a stopwatch. The TST was measured as the time from the start of the skin incision until closure of the incision was completed. The TSS was measured from the start of the skin incision until counting of the used surgical sponges was started. The TWC was measured from the start of placement of the first suture in the abdominal fascia until the incision was completely closed (including the fascia, subcutaneous tissues, and skin).

The SP (measured in centimeters from the cranial tip of the manubrium to the cranial rim of the pubis at the midline) and the length of incisions in the skin and in the abdominal fascia were measured for each dog in dorsal recumbency by use of a sterilized metal ruler with millimeter markings. The S_rel was defined as the measured length of the skin incision (in mm)/SP. Simirellarly, the F_rel was defined as the measured length of the linea alba incision (in mm)/SP, and both measurements were expressed as percentages.

Pain score and wound score assessment—Postoperative pain scores were determined by use of the GCMPS long and short forms at 2, 5, 6, 12, and 24 hours after the start of placement of the first suture in the abdominal fascia (ie, time 0). Long-form descriptor weights were added together to determine a total pain score. Results from all sections of the short form were added together, and the total score of the short form was used to determine whether rescue analgesia was needed; the criterion for this was a score > 15. Postoperative characteristics of the surgical wound were scored at 2, 6, 12, and 24 hours from time 0 by subjective assessment of swelling, redness, dehiscence, discharge, and pain at palpation. Scores were assigned for each of these features from 0 to 4 as follows: 0, not detected; 1, barely detectable; 2, mild; 3, moderate; 4, severe. The scores were evaluated separately and were added together to obtain an overall wound score for each time point. Four senior veterinary students who were unaware of the surgical technique that was used and were trained by the surgical staff to correctly assess pain and wound scores performed these assessments. Each student evaluated 10 of the 40 patients.

Statistical analysis—Surgical variables were evaluated by use of commercially available statistical analysis software.^s The effect of surgery type on dependent variables was compared by use of ANOVA. The BW, BCS, and SP were used as covariates, and the TST, TSS, TWC, BL_rel, F_rel, and S_rel, were used as dependent variables. A Pearson correlation analysis was used to examine the influence of BW, BCS, and SP on the dependent variables. A quantile-quantile plot was used to assess normality of the data. To evaluate pain scores and wound scores, a statistical program^t was used. Individual dog effect (random effect) and procedure (OVE or OVH), time, and their interactions (fixed effects) were used together in a mixed-effect model to evaluate log-transformed pain scores. For wound scores, logistic regression analysis was used with individual dogs considered as a random effect and the type of surgical procedure (OVE vs OVH), time, and their interaction as fixed effects. For all analysis, values of P < 0.05 were accepted as significant.

Results

The results of biochemical and hematologic analysis were within normal limits for all dogs. The age, BW, BCS, and SP of dogs that underwent OVE were compared with those of dogs that underwent OVH (Table 1). No significant differences were detected for these variables between the 2 groups.

Surgical variables—Analysis of the quantile-quantile plot indicated that all data were normally distributed. The results of correlation analysis revealed no significant relationship between BW and BL_rel, S_rel, and F_rel, but a significant positive correlation between BW and TST (r = 0.517; P < 0.001), TSS (r = 0.473; P = 0.002), TWC (r = 0.436; P = 0.005), and SP (r = 0.84; P < 0.001) was detected. No significant effect of BCS was determined for any variable. A significant positive correlation was observed between SP and TST (r = 0.418; P = 0.007), TSS (r = 0.338; P = 0.033) and TWC (r = 0.397; P = 0.011). No significant correlation was found among SP, S_rel, F_rel, and BL_rel. Mean values for F_rel and S_rel were significantly (P = 0.000 for both) increased for dogs that underwent OVH, compared with those that underwent OVE; however, no difference was detected for TST, TSS, TWC or BL_rel between these 2 groups (Table 2).

Pain scores and wound characteristics—Rescue analgesia was not needed for any dog because none had postoperative pain scores > 15 as determined by use of the GCMPS short form. No significant differences were detected between dogs that underwent OVH and those that underwent OVE for any variables evaluated by use of the short and long forms of the GCMPS. Similarly, no significant differences were found for total wound scores at any time point between dogs that underwent OVH and those that underwent OVE.

Discussion

Differences in surgical variables, wound characteristics, and short-term complications (including blood loss, erythema, swelling, discharge, dehiscence, and degree of postoperative pain) were evaluated in a prospective clinical trial of healthy dogs that underwent either OVE or OVH. The dogs in each group were considered comparable because no significant differences were detected for age, BW, BCS, or SP.

The TST, TSS, and TWC durations were positively correlated with BW, which is in agreement with the results of another report¹⁰ that indicated OVH procedures in large dogs were more time-consuming than those in small dogs. The preoperative BCS had no significant influence on the TST, TSS, TWC, BL_rel, F_rel, or S_rel. The BCS reflects an estimation of the amount and distriburelrelreltion of body fat; visible and palpable features and the silhouette of the dog are used in a semiquantitative but subjective 1 to 5 scoring system, with 1 indicative of emaciation and 5 indicative of obesity.^15,16 Ovariohysterectomy was shown²² to be more difficult in obese dogs. The amount of intra-abdominal fat and its relationship with the BCS were not investigated in our study. Therefore, it remains unclear whether the amount of intra-abdominal fat had an effect on the variables evaluated in the present study that was not reflected in the results of external BCS evaluation.

Although the F_rel and S_rel were shorter in dogs that underwent OVE, compared with those that underwent OVH, no difference was observed between dogs that underwent the 2 surgical procedures for TST, TSS, TWC, or BL_rel. On the basis of those results, the hypothesis that surgical time would be decreased for OVE, compared with that required for OVH when both procedures were performed by an experienced surgeon, was rejected.

In several studies,^2,10–12 intra-abdominal hemorrhage has been described as the most common cause of death after OVH in dogs. The incidence of intra-abdominal hemorrhage after OVH has also been associated with BW of the patients (2% vs 79% in dogs that weighed < or ≥ 22.7 kg [50 lb], respectively).⁸ In a study²² in which investigators analyzed the complications observed during and after OVH of 142 dogs at a veterinary teaching hospital, 9 had intra-abdominal hemorrhage from the ovarian arteries during surgery, and 4 had intra-abdominal hemorrhage after surgery. These results suggested that complications such as hemorrhage may occur more frequently in surgeries performed by relatively unskilled students; the exteriorization of the ovaries and ligation of the ovarian pedicles were reported to be the most difficult parts of the procedure. Seven of the 9 dogs with intra-abdominal hemorrhage during those surgeries had bleeding from the right ovarian pedicle. The right ovary is positioned more cranially in the abdomen than the left ovary and would be more difficult to exteriorize adequately.²² In a review¹² of complications following elective surgeries at 5 private general veterinary practices, 11 of 62 dogs that underwent OVH had surgical complications, an unspecified number of which included hemorrhage. Although the surgical techniques and materials used for OVH were similar among the 5 practices, the procedures were performed by different surgeons, and the data collection procedures and definitions of surgical complications varied substantially; therefore, it is difficult to compare the results of both studies^12,22 with the results of the strictly standardized study reported here. We used a study design in which 1 experienced veterinary surgeon performed all surgeries using a standardized protocol to eliminate the problem of interindividual variability among surgeons.

Life threatening, intra-abdominal hemorrhage associated with OVH and OVE may occur from the ovarian or uterine pedicles or from the suspensory and broad ligaments.^1,2 Blood loss can be quantified via gravimetric (ie, weight analysis) or colorometric methods.^20,23 Gravimetric methods are easy to conduct and were determined to be reliable.²⁰ In the study reported here, analysis by use of a gravimetric method detected no difference in BL_rel between dogs that underwent OVH and those that underrelwent OVE. To prevent intra-abdominal hemorrhage, the ovarian and uterine pedicles were double ligated with 1 strand of suture material, and electrosurgical equipment was used routinely in all procedures. The double-ligation technique performed by an experienced surgeon was reliable; no hemorrhage from the ovarian or uterine pedicles was observed. The use of electrosurgical equipment was not mentioned in the other studies^{2,8,10–12,22} discussed here. In the present study, the suspensory ligament was coagulated before it was cut with scissors. It is possible that this prevented hemorrhage from the suspensory ligament vessels. Hemorrhage was considered minor and not life threatening in any of the dogs of the present study, and no adverse consequences developed.

The recognition and evaluation of pain in animals are challenging.²⁴ The assessment of human pain is conducted by means of self-reporting of pain by the individual, and the visual analogue scale is considered to be the preferred standard.^25,26 The recognition of pain in animals relies mainly on the interpretation of their behavior by an observer. Behavior-based methods used to assess pain in animals were historically based on the scales used for humans; however, these were found to be unreliable in the assessment of pain in dogs in a hospital setting.²⁷ Investigators have used pain scales in dogs to evaluate pain in the postoperative period after OVH.^28–31 The GCMPS is a behavior-based assessment tool to evaluate acute pain in dogs.¹⁹ A modified version that was tested at our veterinary teaching hospital was determined to be a useful tool for the evaluation of perioperative pain in a clinical setting.^17,32 Although the 4 student observers in the study reported here were extensively trained and very familiar with the application of the long and short forms of the GCMPS, interobserver variability could not be completely excluded as a bias of the study. For organizational reasons, it was not possible for 1 student to observe all 40 dogs. The advantage of the GCMPS is that it is standardized and easy to use, which should decrease the incidence of variation among observers.³² Additionally, when pain scales are used to evaluate differences between surgical procedures, it is important to be unaware of the technique that was performed. At the end of the study, the incisions in dogs that underwent OVE were noticeably shorter than those in dogs that underwent OVH. As a consequence, although the students that performed the pain assessment were not informed of the surgical procedure at the time of the evaluation, we could not guarantee complete unawareness.

No significant differences were detected in pain scores (evaluated by use of the long and short forms of the GCMPS) between dogs that underwent OVH and those that underwent OVE. It is likely that more tissue was traumatized during OVH procedures than during OVE because the former required larger incisions and resulted in a larger wound surface. The sensation of pain in both procedures apparently was mostly blocked by the pain medication that was provided before, during, and after surgery. The effectiveness of the analgesics had the disadvantage of limiting the range of pain behavior that could be observed, thus obscuring a possible difference in the detection of pain between the 2 procedures. The degree of postoperative pain in 426 dogs that underwent elective castration or OVH was evaluated in another study,³³ in which 1 of 3 different perioperative analgesic treatments (morphine, nalbuphine, and ketoprofen) was administered or no analgesic was given. Significant differences were revealed in postoperative pain scores among the 4 treatment groups, which suggested that treatment with morphine or ketoprofen was superior to treatment with nalbuphine and to no perioperative analgesic administration. Results of that study indicated that the degree of postoperative pain assessed in the dogs was not dependent on the length of the incision, duration of surgery, or experience of the surgeon. Because a standardized analgesic protocol was used in the study reported here, we could not exclude that differences in pain scores would be detected between dogs that underwent OVH and those that underwent OVE if other analgesic protocols had been used.

In another study¹⁴ performed in our department, Van Goethem et al recommended OVE as the preferred surgical procedure for neutering female dogs because OVH was thought to be more time-consuming and to have more intraoperative and postoperative complications. However, significant differences were not detected between dogs that underwent OVH and those that underwent OVE in the present study with respect to surgical time, postoperative short-term complications (including blood loss and dehiscence), pain scores, or surgical wound scores. The results of the present study suggest that the uses of OVH and OVE procedures for neutering healthy female dogs have equivalent outcomes in regard to these measures.