Abstract
Elective ovariohysterectomy was performed on 66 cats. Surgical approach was flank (group F) or midline (group M) allocated by block randomisation. Pre-anaesthetic medication was acepromazine (0.1 mg/kg) via intramuscular injection. Anaesthesia was induced with intravenous thiopentone, and maintained with halothane in 100% oxygen. Carprofen (4 mg/kg) was administered by the subcutaneous route immediately after induction of anaesthesia. Postoperative pain and wound tenderness were assessed at 1, 3, 6, 9, 11–12 and 20–24 h after the end of surgery, and the assessment outcome marked on visual analogue scales (VAS). Intervention analgesia (if pain VAS was >40 mm) was pethidine 4 mg/kg via intramuscular injection. Area under the curve (AUC) for VAS for pain and VAS for wound tenderness for each cat were calculated. AUC for wound tenderness was significantly greater for group F (P=0.007). There was no significant difference for AUC for pain between the groups. In conclusion, wounds after flank ovariohysterectomy are significantly more tender than after midline ovariohysterectomy in the cat. This indicates that interactive methods, including wound palpation, must be used to assess postoperative pain and the findings should be appropriately weighted in the overall assessment.
Ovariohysterectomy (OHE) produces considerable postoperative pain in cats (Slingsby and Waterman-Pearson 1998). In a survey of 2000 British veterinary surgeons, Lascelles et al (1999) found that 99.9% of respondents believed that cats are in pain after OHE, but only 26% provide analgesia. This may suggest that recognition of pain in cats may be difficult, a view which was supported by Slingsby and Waterman-Pearson (1998).
A questionnaire survey of 183 UK veterinary surgeons found that 96% use a flank approach for OHE in cats (Coe et al in press) but several USA standard surgical texts describe only the midline approach. Two of these texts (Fingland 1998, Stone 2003) advise avoiding the flank approach for OHE in the cat, citing inadequate surgical access and difficulty exteriorising the uterus. Comparative studies of various approaches to feline OHE, including midline, oblique flank, horizontal flank and vertical flank have been performed (Hoque 1991, Ghanawat and Mantri 1996) but none of these studies attempted to assess the severity of postoperative pain.
A variety of techniques have been used to assess pain in small animals including the use of visual analogue scale (VAS) numerical ratings scale and simple descriptive scale (Holton et al 1998, Slingsby and Waterman-Pearson 1998). A dynamic and interactive approach to pain scoring (DIVAS) uses physical interactions with the patient in addition to mere observation of appearance and behavioural responses (Lascelles et al 1997, Slingsby and Waterman-Pearson 1998). The addition of the physical interactions (including wound palpation) are claimed to enhance the discriminative sensitivity of the VAS scale (Lascelles 1995, Cambridge et al 2000). Some investigators have additionally employed nocioceptive threshold testing devices (Slingsby and Waterman-Pearson 2000, Slingsby et al 2001, Robertson et al 2003), to quantify pain more objectively, but such equipment was not used in the current study because in the UK such assessments are deemed regulated procedures under the Animals (Scientific Procedures) Act 1986.
The aim of this study was to compare the relative severity of postoperative pain and wound tenderness following either flank or midline OHE in cats using behavioural measures. The relative merits of the two surgical approaches in surgical terms investigated in this same cohort of cats are reported elsewhere (Coe et al in press).
Materials and method
Animals
Sixty-six female cats admitted for elective OHE in the School of Clinical Veterinary Science between October 2002 and October 2003 were included in this study. Informed owner consent was obtained. The cats were examined clinically to ensure that they were healthy and in American Society of Anaesthesiologists (ASA) category 1. Exclusion criteria were evidence of pregnancy, oestrus or lactation found during the pre-anaesthetic examination, age <6 months or >10 years, ASA>1 including evidence of dehydration or a cat that could not be easily handled without being distressed.
Immediately before induction of anaesthesia the cats were assigned to one of two groups (flank (group F) or midline (group M) surgical approach) by blocked randomisation.
Anaesthesia and surgical protocol
Water was freely available until the time of pre-anaesthetic medication. Food was withheld for 8–12 h before anaesthesia. Pre-anaesthetic medication was acepromazine maleate (ACP injection: Novartis Animal Health) at 0.1 mg/kg given by intramuscular injection into the epaxial muscles. Anaesthesia was induced 30 min later with 2.5% thiopentone sodium (Thiovet; Novartis Animal Health) administered via the cephalic vein, without prior catheterisation. Thiopentone was administered until jaw relaxation was sufficient to allow intubation of the trachea. This was performed after application of lidocaine hydrochloride (Xylocaine 4% Topical; Astra Zeneca) to the vocal cords. If perivascular injection of thiopentone was suspected, 2 ml of sterile saline was injected subcutaneously in the affected area and then distributed by gentle massage. Anaesthesia was maintained using halothane (Halothane; Rhodia Organique Fine), delivered as required to maintain a surgical plane of anaesthesia. It was vaporised in 100% oxygen, and delivered via a non-rebreathing system (Ayre's T-piece with Jackson-Rees modification). The animals were allowed to breath spontaneously throughout anaesthesia. Carprofen (Rimadyl; Pfizer) 4 mg/kg was administered by subcutaneous injection immediately after induction of anaesthesia. Appropriate antibiotic therapy was also administered. Intravenous fluid therapy was not considered necessary because these were young, healthy animals, undergoing a short procedure. If additional drug or fluid administration was required in any of the cats, an intravenous catheter was inserted, and the cat was removed from the study.
Heart rate, respiratory rate and vaporiser setting were recorded every 5 min. The operating theatre was maintained at an ambient temperature of between 22 and 24°C and the cats' rectal temperature was monitored and appropriate remedial action taken if their rectal temperature decreased below 36°C. This close control of ambient temperature was to permit additional observations to be made regarding heat loss under anaesthesia, which will form part of another report.
Surgery was carried out by a final year veterinary student with limited experience of either approach, under the direct and continuous supervision of a veterinary surgeon. For group M, a 2.5 cm incision was made from midway between the umbilicus and pubis, caudally. For group F, a 2.5 cm incision was made vertically from the third vertex of a triangle formed by the wing of the ilium and the greater trochanter of the femur. Any extension of the incision was measured and recorded. The duration of surgery was also recorded. At the end of surgery, the vaporiser was turned off; this time was taken as time zero.
The first six cats studied had their abdomens bandaged so that the observer would be unaware of the treatment group. This was abandoned for the subsequent 60 cats. Water was offered to the cats within 2–3 h of the end of surgery.
Scoring sedation, pain and wound tenderness
Before pre-anaesthetic medication, the cats were examined and their level of sedation was assessed by a single, trained observer who allocated a sedation score, using a simple descriptive scale (see Table 1 for details).
Simple descriptive scale used to describe sedation
The cats' level of pain was assessed and a pain score (VASpain) assigned using a visual analogue scale, 100 mm in length anchored by ‘no pain’ and ‘worst possible pain following ovariohysterectomy’. This was carried out as described by Slingsby and Waterman-Pearson (1998). The cat was observed from a distance without the kennel door open and the cat's posture and any attention the cat was paying to its the wound, including licking, biting and chewing were noted. The cage door was then opened and the observer interacted with the cat, calling the cat to the front of the cage, noting their demeanour, willingness to move and willingness to interact. Any vocalisation including purring, hissing, or crying at any stage was also noted.
Wound tenderness was assessed by palpation of the surgical site and a wound tenderness score (VASwound) also assigned on a VAS scale (100 mm in length, anchored by ‘no wound tenderness’ and ‘worst possible wound tenderness following ovariohysterectomy’). The method used was standard for all cats and took the form of gentle digital palpation using the flat part of the middle three fingers of the observer's right hand. The site palpated was immediately adjacent to the line of incision, but not directly over the incision. The observer palpated the site up to four times and took into account not only the severity of the cats reaction to the palpation (ranging from looking towards the incision, to hissing and trying to bite the observer) but also the number of palpations performed before the cat reacted. Pre-operatively, both possible incision sites were palpated. Postoperatively only the actual surgical site was palpated.
Sedation, pain and wound tenderness scores were reassessed 25 min after pre-anaesthetic medication and at 1, 3, 6, 9, 11–12 and 20–24 h after time zero. Intervention analgesia was administered if a VAS pain score greater than 40 mm was recorded (pethidine (Pethidine; Arnolds) 4 mg/kg by intramuscular injection).
Statistical analysis
A cumulative total sedation score was calculated for each cat. The area under the curve for pain (AUCpain) and area under the curve for wound tenderness (AUCwound) were calculated for each cat using block squares method.
Mean VASpain and VASwound at each time point were plotted for the two groups (see Figs 1 and 2).
Mean VASpain over time. Error bars=total range.
Mean VASwound over time. Error bars=total range.
A sample size calculation was performed using the area under the curve for wound tenderness data from the first 20 cats. A two-group t-test was used assuming a power of 90% (nQuery advisor version 4).
The ages, weights, mean vaporiser settings for each anaesthetic period, incision length, duration of surgery, total sedation score, AUCpain and AUCwound were plotted as histograms. Data that were normally distributed were compared using a two-tailed t test, and results are reported as the mean (standard deviation). Data that were not normally distributed were compared using a Mann–Whitney test, and results are reported as median (range).
Statistical tests were performed using SPSS (v11.5). Significance was taken as P<0.05.
Results
It was calculated that a sample size of 23 cats in each group would have a 90% power to detect a 150 unit difference between areas under the curve for wound tenderness.
The first six cats in an initial pilot study were not included in any statistical analysis. They exhibited such abnormal behaviour after bandaging of their abdomens that their results were discarded. No other cats were removed from the study.
There were 30 cats in each group with similar breed distribution (Table 2). The mean weight and median age of each group of cats were not significantly different (Table 2). The median total sedation score for each group was 1. Perivascular injection of thiopentone was suspected in five out of the 66 cats, and treated appropriately. No further problems were reported in these cats following discharge. No significant difference was found between the two groups with respect to mean vaporiser setting.
Distribution of age, weight and breed
Sixteen cats in group M had incisions extended, compared with six cats in group F. The mean incision length for group M was 3.1 cm (0.6), and for group F was 2.6 cm (0.2). This difference was significant (P=0.001). The duration of surgery was not significantly different (P=0.21) between groups (group M 43.8 min (11.0), group F 41.1 min (1.2)).
The mean AUCpain for group M was 324 (178) and for group F was 357 (202). There was no significant difference between these groups with P=0.516. However, the AUCwound mean was significantly higher for group F (381 (169)) than group M [256 (177)] with P=0.007.
Two cats in group M (one cat at zero+11 h, one cat at zero+1 h) and two cats in group F (one cat at zero+3 h and one cat at zero+1 and zero+3 h) required intervention analgesia. The VASpain and VASwound score for each cat requiring intervention analgesia was maintained at that level for all subsequent time points.
Discussion
This was a clinical study to compare the severity of pain and wound tenderness between midline and flank approaches for OHE in the cat. The findings indicate that postoperative pain scores between the two groups were similar, but there was significantly more wound tenderness following the flank approach.
A major limitation of this study was that it was not a blind trial. Abdominal bandaging, in the initial study of six cats, led to markedly abnormal behaviour by the cats. This included cats exhibiting extreme reluctance to move and abnormal postures with hindlimb extension and evasive behaviours. This made pain assessment almost impossible. Glerum et al (2001) described using a ‘light wrap’ over cats' thoraxes to conceal the absence or presence of fentanyl patches but no mention was made of any behavioural changes in the cats. A light wrap would not have been sufficient to blind the observer to the surgical approach used in this study. An alternative to bandaging could have been the clipping of both surgical sites, with subsequent placement of adhesive dressings. However, we would have been very unlikely to obtain owner consent for this strategy.
A control group (ie, no analgesia) was not included in this study. Slingsby and Waterman-Pearson (1998) studied a control group of cats post OHE with no analgesia, and all cats required intervention analgesia; therefore, we felt that such a study need not be repeated.
While administering thiopentone via a hypodermic needle rather than a pre-placed cannula increases the risk of perivascular injection, at the concentration of thiopentone used (2.5%) necrosis and sloughing are rare (Hall et al 2001). No animals suffered adverse consequences of this technique of drug administration. Catheterisation is not always performed in veterinary practice and teaching students to undertake intravenous injections so that they are confident to do this in an emergency is part of the teaching philosophy at Bristol. The use of intra-operative fluid administration during ovariohysterectomy in cats is rare in practice and it was deemed important that the protocol used should parallel as closely as possible what is the norm in general practice. Any animal that might have required supplementary drugs or fluids would have been appropriately treated and withdrawn from the study, in the event no such need arose.
The use of parasympatholytic agents as a routine part of pre-anaesthetic medication is no longer recommended due to increasing myocardial oxygen demand and risk of cardiac dysrhythmias (Thurman et al 1996). Excessive salivation and bradydysrhythmias are not frequently observed in cats undergoing acepromazine–thiopentone–halothane anaesthesia. Hence, these drugs were not administered.
Recognition of pain in cats is challenging as the signs often are more subtle than in dogs (Cambridge et al 2000). Cats show mainly postural signs, without vocalisation (Sanford et al 1986). In our study postoperative pain scores were allocated using a VAS that may allow greater sensitivity to detect pain compared with a simple descriptive score (Holton et al 1998, Cambridge et al 2000). Our method of pain assessment was subjective, based entirely on behavioural measures. Physiological measures such as blood cortisol, glucose and β endorphin levels, heart rate, respiratory rate and rectal temperature are poor indicators of postoperative pain in cats (Smith et al 1999, Cambridge et al 2000). Nocioceptive threshold testing might have provided more objective results, for example Slingsby et al (2001) used a finger-mounted device to measure mechanical nocioceptive thresholds in cats following castration. Their study also demonstrated, however, that the threshold measurements were mirrored by the VAS pain scores, suggesting good correlation between the two methods of pain scoring.
Pain scores (VASpain) were similar between the two groups, however wound tenderness scores (VASwound) were significantly higher in group F, even though the midline incisions cross several dermatomes (T12–L2), compared to the one dermatome (L2) affected by a flank incision (Hekmatpanah 1961), and the incision length in group M was longer in this group of cats. The flank incisions may be more tender because more muscle fibres are severed by the incision to gain access to the abdomen. Free nerve ending nocioceptors are present near small blood vessels in muscle (Graven-Nielsen and Mense 2001), and connect to both Aδ and C fibres (Lloyd 1943). Conversely, deep tissues such as the connective tissue which forms the linea alba are sparsely supplied (Guyton and Hall 1996). In addition to the lack of nocioceptors, the linea alba lacks the blood supply needed to transport inflammatory mediators to sites of tissue damage.
Movement evoked pain is more intense than pain at rest due to the recruitment of more nocioceptive sensory inputs during movement (Gilron et al 2002), and the site of the flank incision was just cranial to the left hind leg, and thus the surgical site moved more than the midline incision as the cats moved, thus recruiting more nocioceptor activity at the surgical site.
The lack of significant difference in pain scores between the two groups may have been due to the administration of carprofen as a perioperative analgesia, masking the ability to pick up subtle differences between the groups. In two other studies using the same anaesthetic protocol for OHE, intervention analgesia was administered to two cats out of 57 (Balmer et al 1998) none out of 10 (Slingsby and Waterman-Pearson 2000), one out of 40 (Slingsby and Waterman-Pearson 2002), and four cats out of 60 in this study. The fact that few cats needed intervention analgesia suggests either that carprofen is an effective analgesic for post OHE in the cat, or that the assessment system was not sensitive enough to pick up subtle signs of pain. Interestingly, Slingsby and Waterman-Pearson (2000) found that carprofen was not able to prevent the development of postoperative wound tenderness after feline OHE and we also noticed this. This could explain why a difference became apparent between the two groups regarding wound tenderness, but not for pain.
This study indicates that using behaviour scoring systems to assess postoperative pain between flank and midline approaches, there was no apparent difference. However, when the wound tenderness was evaluated, a significant difference was observed. This suggests two conclusions; in any assessment mere observation is insufficient to assess pain and interactive means such as palpation must be used and properly weighted, in arriving at an overall score. Secondly, that wounds are significantly more tender after flank incisions compared with the midline approach even when analgesia is provided by the administration of a single dose of a non-steroidal anti-inflammatory agent.
Footnotes
Acknowledgements
The authors would like to thank the following who assisted in this study: Dr Louisa Slingsby, Mr Colin Capner, Ms Katy Evans and Ms Emma Love for their anaesthesia skills; Professor Peter Holt, Mr Alasdair Hotston-Moore, Professor Gary Ellison, Mrs Rosa Ragni and Mr Andy Moores for their surgical skills, all veterinary students, and nursing staff involved, and Dr Toby Knowles for his assistance with statistical analysis.