Questions study on dietary supplementation in dogs with osteoarthritis

As a reader trying to interpret the results of the March 1, 2010, JAVMA report¹ of the effect of dietary supplementation on the carprofen dosage requirement in dogs with osteoarthritis, I have four general areas of concern.

Did the investigators have a hypothesis? Given the study design, it seems logical that an alternative-to-null hypothesis would have been something along these lines: “In dogs with signs of pain from osteoarthritis despite treatment with carprofen, supplementation with fish oil will allow a greater reduction in the dosage of carprofen with no increase in estimated pain than supplementation with a placebo.” In that case, this hypothesis is rejected, as there was not a significant difference between groups at the end of the treatment. So, while the study confirmed that clinicians will reduce the dosage of carprofen when asked to do so if signs permit, it did not demonstrate an independent treatment effect.

Participating clinicians were instructed to reduce the dosage of carprofen in response to improved pain scores or according to their “overall perception.” No pain score data were presented, so I am left wondering what the basis was for any attempts at dosage reduction. Was there a threshold of pain score reduction that was used to trigger a reduction in dosage, or were clinicians instructed to reduce the dosage if the dog's condition was no worse? Because “overall perception” was allowed to influence the decision, was the dosage reduced in any dogs that had no improvement in pain scores? Was it increased in any dogs with reduced pain scores?

If the daily carprofen dosage was “standardized… to approximately 4.4 mg/kg,” then why was the baseline dosage range 2.19 to 6.8 mg/kg? One wonders how the dosage was arrived at during the acclimation period because there is no indication it was in fact standardized to any systematically applied endpoint and there is no information provided regarding owner or clinician assessments. The study design limited enrollment to dogs that had failed carprofen monotherapy (as it required lameness despite ongoing treatment); how did this affect the standardization of the dose? There was no effort made during recruitment to populate the two treatment arms with dogs receiving identical baseline dosages of carprofen. Given the potential importance of baseline carprofen dosage on willingness to reduce the dosage further, was the difference in baseline (after-adjustment) carprofen dosage between groups significant?

Since two very different diets were used and one may have smelled like fish, did the investigators ask owners to guess which diet they were using at study completion?

Positive results attributed to the test diet and emphasized in this report were a steeper slope of carprofen dosage reduction (on the way to an insignificant final difference) and a different pattern of response when dosage changes were parsed into six levels of range. However, as a clinician, I'm interested to know if this diet has a sparing effect on the dosage requirement for NSAIDs, which sadly was not demonstrated.

Bernie Hansen, dvm, ms, dacvecc, dacvim

College of Veterinary Medicine, North Carolina State University, Raleigh, NC

The author responds:

The authors appreciate Dr. Hansen's interest in our report¹ but disagree with his interpretation and conclusions.

The hypothesis of this randomized, prospective study was to determine whether feeding a diet supplemented with fish oil would permit a greater reduction in carprofen dosage in arthritic dogs, compared with feeding a control food, while maintaining or improving clinical relief of signs. All dogs were evaluated by a veterinarian every three weeks for 12 weeks and were regularly assessed by their owners. The appropriate statistical analysis for this study design is a repeated-measures ANOVA, with specific focus on the diet-by-time interaction. As such, we reported the rate of change in carprofen dosage over the duration of the study, which takes into account all available information. Focusing on the mean difference in carprofen dosage at the final time point is not appropriate because it ignores all other information from the study. We observed a two-fold difference in the change in carprofen dosage between the test group (−0.49 mg/kg) and control group (−0.25 mg/kg) after 12 weeks of feeding (P = 0.010), supporting our hypothesis.

Decisions regarding the reduction of carprofen dosage were not based on signs of pain alone but rather the overall arthritic condition as assessed by the veterinarian, in conjunction with the pet owner. Pain was only one of the clinical signs considered in the assessment. For instance, a slight increase in signs of pain during palpation could have been offset by improvements in severity of lameness or the owner's perception of the arthritic condition overall. In our study, carprofen dosage was actually reduced in seven dogs with an increase in signs of pain during palpation.

The standardized dosage of carprofen at the beginning of the study was approximate because labeling instructions specify that tablet dosages should be calculated in half-tablet increments. Because the smallest dose of carprofen was a 25-mg tablet, variation in dosage was most pronounced in the smallest dogs, which explains the range in baseline dosages.

This study was not limited to dogs in which carprofen monotherapy had failed. Some dogs had lameness that was improved but not completely resolved with carprofen. The degree of success with previous treatment with carprofen did not affect the dosage standardization. There was no difference between groups in the baseline carprofen dosage after standardization (P = 0.11).

It is unlikely that pet owners would have been able to guess which food they were feeding to the dogs. Pet owners were introduced to only one of the two diets at the beginning of the study and, as such, could never have noticed a difference between the foods. Additionally, pet owners did not rate the aroma pleasantness of the foods differently (P = 0.20).

The authors respectfully disagree with Dr. Hansen's conclusions. Rather, our findings clearly demonstrate that feeding a diet supplemented with fish oil allows for a reduction in carprofen dosage while maintaining comfort and function of dogs with osteoarthritis. These results furthermore concur with our previous studies^2,3 showing that this food improves the comfort and function of dogs with osteoarthritis.

Kevin A. Hahn, dvm, phd

Hill's Pet Nutrition, Topeka, Kan