Materials and Methods

Dogs with surgically confirmed cranial cruciate disease that underwent lateral fabellar suture stabilization were recruited for the study. The study protocol received approval from the institutional animal care and use committee. Owner consent was obtained. Exclusion criteria included systemic illness, renal or hepatic dysfunction as determined on the basis of CBC and serum biochemical analysis, pregnancy, or NSAID or corticosteroid use within 1 week prior to surgery.

Each dog was randomly assigned to the treatment or control group. There were 17 dogs assigned to the control group and 18 dogs assigned to the treatment group. A random number generator was used by the hospital pharmacist to create a group assignment schedule prior to the start of the study. Measurements were performed for dogs in the order of the group assignment schedule. All dogs in the study received carprofen (2.2 mg/kg [1 mg/lb], PO, q 12 h) for the first 7 days after surgery. All dogs were hospitalized and received concentrated rehabilitation exercises by a single certified rehabilitation practitioner during weeks 3, 5, and 7 after surgery.

Dogs in the treatment group received carprofen (2.2 mg/kg, PO, q 12 h) during the weeks of concentrated rehabilitation only and did not receive any NSAID medication at any other time. Medication was administered by a veterinary student. Dogs in the control group did not receive any NSAID medication during the rehabilitation period.

Surgical procedure—Dogs were premedicated with morphine (0.5 mg/kg [0.23 mg/lb], IM), atropine (0.01 mg/kg [0.005 mg/lb], IM), and medetomidine (10 μg/kg [4.5 μg/lb], IM). Anesthesia was induced by use of propofol (3 to 6 mg/kg [1.4 to 2.7 mg/lb], IV) or thiopental (2 to 6 mg/kg [0.9 to 2.7 mg/lb], IV) and maintained with isoflurane. Dogs were administered cefazolin (22 mg/kg [10 mg/lb], IV) perioperatively. For postoperative analgesia, hydromorphone (0.1 mg/kg [0.05 mg/lb], SC, once) was administered immediately after surgery, followed by carprofen (2.2 mg/kg, PO, q 12 h) for 7 days.

Surgeries were performed by multiple surgeons including residents. The residents were supervised by a board-certified surgeon. Although the same surgeon did not perform all surgeries, the procedure was standardized except with regard to meniscal surgery. Standard lateral fabellar suture stabilization was performed after craniolateral arthrotomy and inspection of the joint.⁹ The cranial cruciate remnants were removed and the menisci inspected. Medial meniscal injury was treated with a partial meniscectomy. An abaxial medial meniscal release was performed at the surgeon's discretion. An 80-lb test monofilament nylon suture was passed around the lateral fabella, under the patellar ligament, and through a hole in the proximal tibia and fastened with a crimp clamp.^a The suture was tightened prior to crimping to a level that prevented a cranial drawer sign but allowed at least 90° of stifle joint flexion. A standard closure of the incision was performed.

Rehabilitation—Concentrated rehabilitation exercises were performed in all dogs during weeks 3, 5, and 7 after surgery by a single certified rehabilitation practitioner. An alternating weekly protocol was chosen on the basis of the literature^2,10,11 and to facilitate convenience for owners living considerable distances from the rehabilitation site. All dogs were hospitalized during concentrated rehabilitation weeks. All dogs received the same exercise protocol with regard to timing and type of exercises. Two rehabilitation sessions were performed daily; these included passive ROM exercises in both hind limbs and conventional and underwater treadmill exercises. Passive ROM exercises were performed with the dog in lateral recumbency. The stifle joint was held in alternating flexion and extension for 10 seconds each, for a total of 5 minutes on each hind limb. Dogs were then placed on a conventional treadmill and walked at 0.3 to 2 m/s for 10 minutes, followed by exercise on an underwater treadmill within the same speed range for 15 minutes. Each dog was maintained at a speed at which the perceived exertion score was between 4 and 6 on a children's scale of perceived exertion,¹² modified for use in dogs (Appendix). At this level, dogs have moderate signs of exertion, consistent panting but no labored breathing, mild signs of agitation, and no stumbling. The treadmill speed was recorded when the dog was observed to be performing at this level of perceived exertion.

On alternate weeks, the dog was exercised less intensely at home and dogs in the treatment group did not receive carprofen. Owners were instructed to perform passive ROM exercises for 5 minutes twice daily and were encouraged to walk the dog on a leash for up to 30 minutes as tolerated. Dogs were to remain in confinement when not directly supervised by the owner.

Outcome measures—Outcome was measured by use of gait analysis, goniometry, and thigh circumference measured before surgery and at the end of each intense rehabilitation week. Workout speed was also recorded during rehabilitation sessions. Gait analysis was performed by walking the dog over a pressure platform^b at a speed between 1.1 and 1.3 m/s and an acceleration and deceleration of 0.5 m/s². Peak vertical force and VI from 5 footfalls for the affected hind limb from each dog were measured. Mean values for each dog and PVF and VI expressed as percentages of body weight were used for group comparisons.

Thigh circumference was measured at the proximal quarter of the femur and expressed as the measurement of the affected hind limb and as a ratio of the measurement of the affected to unaffected (or contralateral) hind limb to account for morphological differences between dogs and potential weight loss over the study duration. Stifle joint goniometry was performed by use of previously published methods.¹³ The greater trochanter, femoral condyle, and lateral tibial malleolus were used as landmarks. Dogs were not sedated for measurements. Three measurements were taken for flexion and extension, the difference was calculated to determine ROM, and the mean of the 3 measurements for each time point was used for analysis.

Workout speed was calculated to provide an indirect measure of perceived exertion. This was done to eliminate the effect of fatigue from conventional treadmill exercise on subsequent underwater treadmill performance. The speed of the treadmill was set at a level where dogs had a perceived exertion score of 4 to 6 on the modified children's scale of perceived exertion.¹² The speeds of the underwater and conventional treadmills were recorded, and the mean value was calculated to determine the workout speed. A single investigator, who was blinded to group assignments, determined perceived exertion and performed measurements on all dogs.

In addition, information on potential confounding factors was recorded for each dog, including age, breed, body weight, body condition score (on a scale of 1 to 9), severity of preoperative lameness, presence of bilateral cruciate disease, meniscal injury and treatment, or postoperative surgical complications.

Statistical analysis—A matched-pairs t test was used to compare the change from baseline at each re-evaluation between groups. Groups were also compared at each time point by use of a Student t test except when there were unequal group variances (as determined on the basis of a Levene test), in which case the Welch ANOVA was used. Potential sources of bias described by nominal data were compared by use of a contingency table and χ² analysis with a Pearson test; numeric confounding factors were compared by use of a Student t test or a Wilcoxon rank sum test as appropriate. Summary statistics are reported as mean ± SEM. Values of P < 0.05 were considered statistically significant.

Results

Results are presented for the potential confounding factors as the mean ± SEM for the treatment and control group. There were no significant differences between groups with regard to age (treatment group, 3.9 ± 0.4 years old; control group, 4.7 ± 0.6 years old), breed (13 mixed-breed dogs and no other dominant breed), body weight (treatment group, 37.3 ± 2.7 kg [82.06 ± 5.94 lb]; control group, 33.8 ± 2.6 kg [74.36 ± 5.72 lb]), body condition score (treatment group, 6.2 ± 0.4; control group, 6.1 ± 0.3; on a scale of 1 to 9), presence of bilateral cruciate disease (treatment group, 3 dogs; control group, 4 dogs), meniscal injury and treatment (treatment group, 13 dogs; control group, 14 dogs), or postoperative complications (treatment group, 1 dog; control group, 0 dogs). There were no adverse effects attributable to drug administration. Peak vertical force and VI were also not significantly (P = 0.5) different before surgery between dogs with unilateral and bilateral cruciate disease. At 7 weeks after surgery, thigh circumference was greater in dogs with bilateral disease than in dogs with unilateral disease, but this difference was not significant (P = 0.09).

The mean ± SEM for the outcome measures are given (Table 1). There were no significant differences between the 2 groups in ground reaction forces, thigh circumference (regardless of normalization to the contralateral hind limb), or exertion (mean workout speed) over time or at any individual time point. However, both groups significantly (P < 0.01) improved over time for all outcome measures.

Although groups did not have a different ROM at any individual time point, the control group had significantly (P < 0.05) greater improvement over time in ROM from 0 to 5 weeks, 0 to 7 weeks, and 3 to 5 weeks. From weeks 3 to 5, dogs in the treatment group had a mean decrease in ROM by 2.2°, whereas the ROM of dogs in the control group steadily improved through week 5.

Discussion

Overall outcome measures assessed for treatment and control group dogs improved over the study duration. Dogs in the treatment group did not have greater improvement in hind limb function or thigh circumference, compared with dogs in the control group. This result is surprising given the previously published benefits of NSAIDs.^8,14

Nonsteroidal anti-inflammatory drugs in general and carprofen specifically have been effective in treating acute and chronic osteoarthritic pain.^8,15–18 The dogs in this study were treated with carprofen to improve performance at perceived exertion levels and reduce signs of pain and inflammation after exercise. This was theorized to improve outcome; however, there was no difference between groups in ground reaction forces or thigh circumference.

Although the number of dogs was small, which limited the power of the study, a post hoc power analysis was not performed because the differences in value between the groups were not clinically relevant (ie, a post hoc power analysis would only provide the power on the basis of the difference derived in this study, which is too small to be of consequence clinically).

The inclusion of dogs with bilateral disease may have impacted the results of this study by decreasing the sensitivity of the outcome measures. These dogs were not removed from the analyses because the study was designed to represent the clinical situation. It is reported that up to 48% of dogs rupture the contralateral cruciate within 10 months after evaluation for the first cranial cruciate ligament rupture. In the present study, 20% of the dogs ruptured their cruciate within the first 4 months, a rate consistent with previous findings.¹⁹ It is possible that dogs with bilateral cruciate disease may have had higher ground reaction forces in the hind limb being measured than did dogs with unilateral disease because a weight shift to the contralateral hind limb would be less comfortable. However, PVFs were not significantly (P = 0.5) different when compared between dogs with bilateral (n = 7) and unilateral (28) disease at any time point. It is possible that dogs with bilateral disease shifted weight to the forelimbs rather than to the contralateral hind limb so that no difference was seen in hind limb function between dogs with unilateral and bilateral disease.^10,11,20 However, this was only true in 2 of the 7 dogs with bilateral disease. The remaining 5 dogs had dramatic asymmetry in PVF between the hind limbs despite palpable effusion, signs of pain on extension, and a cranial drawer sign in the contralateral hind limb. Additionally, the results are the same even if the dogs with bilateral disease are removed from the analysis.

Thigh circumference comparison may have been affected by bilateral disease because it was analyzed as a measure of symmetry. At 7 weeks after surgery, thigh circumference was larger in the dogs with bilateral disease, compared with dogs with unilateral disease, but this difference was not significant. It is possible that dogs with bilateral disease have muscle atrophy in the contralateral hind limb, thus allowing a higher degree of symmetry sooner than in those dogs with unaffected contralateral hind limbs. These dogs were included because bilateral cruciate disease is common in veterinary medicine and the present study was designed to be clinically relevant.¹⁹ Additionally, dogs with bilateral disease were balanced across groups. Although symmetry is used to eliminate bias from changes in body condition over time, in this study, the data for the affected hind limb were also analyzed without normalization to the contralateral hind limb. However, there were still no differences between groups.

Range of motion improved faster in the control dogs. This result should be interpreted with caution. It would imply that NSAIDs should be avoided to maximize the ROM. However, it is possible that perceived exertion was lower in the carprofen treatment group and that dogs therefore worked harder, resulting in increased soreness associated with the intense rehabilitation exercises. If this were the case, the indirect measure of exertion used in this study (mean workout speed) did not detect the difference and the increased soreness did not lead to appreciable changes in the degree of lameness. Hind limb function as measured by PVF improved over each time point. If dogs in the carprofen treatment group were working more strenuously, it is logical to think that they should have gained muscle mass and restored symmetry quicker, but this was not the result. A more sensitive method of measuring muscle mass (eg, dual-energy x-ray absorptiometry scan) would be needed to investigate this hypothesis further.

Range of motion was determined at the extremes of flexion and extension at which no signs of pain were observed. Because this is a subjective interpretation, a single investigator performed goniometry. Some variation remained in the technique despite all efforts to eliminate bias. Range of motion for all dogs was much lower than that reported by Jaegger et al,¹³ likely because of differences in the populations of dogs. The observation of signs of pain is used as the endpoint for measuring flexion and extension. It is reasonable to assume that the ROM of clinically normal dogs reported by Jaegger et al¹³ represents the maximum ROM possible, whereas the population of dogs in the present study had signs of pain and inflammation associated with injury and surgery, which reduced the extent of flexion and extension that could be achieved without signs of pain. Jandi and Schulman²¹ and Hoelzler et al²² found decreases in flexion after stifle joint arthrotomy that improved over time. In addition, lateral fabellar suture stabilization may have also limited flexion in some dogs more than in others.

There were 4 dogs in the treatment group and 2 dogs in the control group with decreases in ROM from weeks 3 to 5. The decrease in ROM was primarily attributable to decreased flexion; however, 1 dog in the carprofen treatment group had marked decreases in flexion and extension amounting to 40°. The ROM results remained the same even if this dog was removed from the analyses. Additionally, none of these dogs had a diagnosis of any specific complication.

The perceived exertion scale used in the present study is modified from a human pediatric scale¹² and has not been validated for use in dogs. Unfortunately, it is highly subjective and based on observation because the dog cannot communicate its own perceived exertion level. The guidelines use respiration as the main sign of exertion. For instance, the dogs were maintained at a speed at which they were consistently panting but did not have labored breathing. There are many other reasons for panting besides exertion, including heat, anxiety, pain, and nasal occlusion, that can make determining the exertion level difficult. In an effort to be as accurate and precise as possible, only 1 evaluator determined the exertion level of the dogs. This evaluator was blinded to the group assignments. This method of determining exertion is insensitive for picking up the small changes expected from NSAID use. Other more sensitive and direct methods of determining exertion or quantifying metabolism, such as blood pressure, heart rate, tidal volume, and biochemical metabolic waste (ie, lactate), were not used because of the clinical nature of the study. These types of direct measures would have been preferable to the indirect method (speed at a static perceived exertion) used because indirect measures are subject to additional bias. However, it was not practical to attach invasive and unwieldy monitoring equipment to the dogs, which were uncooperative, during underwater exercises.

Signs of pain were not scored directly in this study. However, it is common to use ground reaction forces as an objective alternative to subjective pain scoring techniques. It is assumed that the degree of lameness is associated with the degree of pain.^{10,11,14,20,23,24} Peak vertical force and VI are the ground reaction forces most commonly used to quantify lameness. Peak vertical force alone is 89% accurate in identifying dogs with lameness attributable to cranial cruciate disease.²⁰ This method is commonly used to determine the efficacy of pharmaceuticals in reducing acute and chronic pain and has been shown to be more accurate than subjective lameness or pain scoring.^{14,15,17,20,23,24} Nonsteroidal anti-inflammatory drugs did not decrease soreness after exercise in several studies^25,26 involving human patients. Additionally, studies^14,27–29 evaluating inflammation after exercise have not proven the usefulness of NSAIDs. The dogs in the present study started intense rehabilitation 3 weeks after surgery. By this time, the signs of acute pain associated with the arthrotomy should be waning, and the rehabilitation protocol was carefully developed to minimize the potential for exacerbating osteoarthritis by incorporating low-impact exercises that promote muscle building, proprioception, and ROM.

Carprofen was administered only during the weeks of rehabilitation rather than during the entire study period because there was no placebo for owners to give the control group. A third party (ie, a veterinary student) administered the medication to the treatment group during hospitalization. Additionally, in human studies^26–29 evaluating muscle hypertrophy and NSAID use, the NSAID is given at the time of exercise rather than for prolonged periods after the end of the exercise session. In the present study, pain medication was not withheld between rehabilitations sessions; however, neither the owners nor the attending clinician believed that signs of pain in the dogs warranted medication.

The short study duration may not have elucidated a long-term effect of NSAID administration during rehabilitation in dogs. Investigation of a long-term effect was not a goal of the present study. Instead, this study focused on much more immediate results (ie, the end of each week of intensive rehabilitation). This is similar to human studies^26–29 that focused on the effects of NSAIDs on muscle hypertrophy. The results of the present study indicate that administration of carprofen during concentrated rehabilitation after lateral fabellar suture stabilization does not clinically improve hind limb function, ROM, or thigh circumference, compared with control dogs following the same rehabilitation protocol.