A new article has just been published adding some information to a subject I’ve addressed before, the value of blood testing and other diagnostics in clinical health dogs and cats (discussed Here and Here). In the past, I have come to the following conclusions on this subject:
- If you test, you will find abnormalities
- The clinical significance of these abnormalities is often unclear
- You will find more abnormalities, and these will be easier to interpret, if pre-test probability is high, that is if there is a reason to suspect a real medical problem. For example,
- Some symptoms of a problem in the history
- Some identifiable abnormality in the physical exam
- There is, as yet, no evidence that most tests of healthy veterinary patients reduce the occurrence of disease or death
This paper seems to support these conclusions and the results of previous research.
Dell’Osa D, Jaensch S. Prevalence of clinicopathological changes in healthy middle-aged dogs and cats presenting to veterinary practices for routine procedures. Aust Vet J. 2016 Sep;94(9):317-23
The authors did a wide range of blood tests and a urinalysis on 406 dogs 5-8 years old and 130 cats 6-9 years old who came to the vet for routine examination, vaccination, parasite treatment, and other reasons but without an ongoing or new medical problem. This population is probable fairly representative of those animals who come to see their veterinarian when they are healthy, though of course it doesn’t represent well those who only come when ill or who don’t get regular veterinary care.
The study results certainly confirm that when you test healthy individuals, you will find abnormalities and most of those abnormalities will either not represent real disease or will be difficult to interpret.
- 86% of dogs had one or more abnormality. Only 6.2% of these were subjectively judged to justify further evaluation.
- 80% of cats had one or more abnormality. Only 19.2% of these were subjectively judged to justify further evaluation.
Common abnormalities that were judged to be errors, testing artefacts, or otherwise clinically unimportant included:
- 124 cats had abnormal platelet counts, none of which were thought to be real or important
- Electrolyte abnormalities and liver enzyme elevations were common in dogs and were generally mild and considered insignificant
- Low thyroid hormone levels were frequently seen without any clinical evidence of thyroid disease
- High red blood cell concentrations were common and attributed to excitement from the visit and blood sampling
- A number of dogs and cats had enzyme elevations that suggested the presence of pancreatitis, though none have clinical symptoms of this problem
The authors did not provide any information on the final diagnosis or clinical outcomes for the animals in this study, so it is impossible, in most cases, to tell how many might have had actual disease or been helped or harmed by the initial and subsequent testing and treatment. They did identify 3 cats (2%) and 5 dogs (1.2%) with Stage 2 or 3 kidney disease. There is evidence that diet change can prolong survival in asymptomatic animals with Stage 2 or higher kidney disease, so a case can be made that these results would likely have benefitted the patients if their owners followed recommended treatment guidelines.
However, there was no other information to identify any benefit or harm from this testing to the animals in the study. This is one of the common problems with research in this area. It is clear that performing diagnostic tests without any specific reason to do so will uncover abnormalities. What is not clear, is how often these results will lead to benefit to the patient through detection and treatment of disease.
As I have discussed previously, there is a real risk of overdiagnosis and overtreatment associated with diagnostic tests done without any specific clinical suspicion of a disease. Most will be meaningless, and this is, at least, a waste of money and resources. Some will be incorrect, as many of the results in this study likely were (especially platelet counts, thyroid testing, and pancreatitis testing). This can lead to unnecessary further tests or treatment, which not only wastes resources but can cause direct harm to patients.
It is also unclear how clients will react to abnormalities, whether or not the veterinarian believes them to be significant. If I tell an owner their older animal has Stage 1 kidney disease, mild liver enzyme elevations, or mild chronic anemia, it is quite possible they will come to view their pet is sick or “aging.” This can easily lead owners to decline testing and treatment that might actually benefit the pet. Comments like, “Well, he has kidney disease,” or “He already has a liver problem” are often made prior to declining dentistry, surgery for curable tumors, and other treatments I recommend for older pets.
Generally, screening tests like these of healthy individuals are very seldom recommended because the evidence has been developed to show they offer few benefits and some real risks. The authors of this paper make the fair observation that this might be different in veterinary medicine since we get less direct information from our patients about how they feel, so we may not be able to pick up signs of disease as soon as in humans. However, the evidence doesn’t exist to demonstrate this hypothesis, as reasonable as it is. And often, the evidence does not exist to show we can do anything useful to help animals with asymptomatic disease in most cases (kidney disease being an exception). So the value of a diagnosis in the face of no available treatment is somewhat questionable.
The subject of overdiagnosis is controversial, and discussions about it often become emotional. It is very difficult not to fall prey to anecdotal reasoning, especially in the absence of high-quality controlled research evidence. However, just as we frequently extrapolate in our testing and treatments from human medicine, it is reasonable to consider the evidence from human medicine that overdiagnosis is a real risk, and there may be times when not testing, especially in individuals who are clinically well, may be the best choice. I’m not arguing that such testing is not beneficial or is harmful, only that we don’t yet have the evidence to say for sure, most of our thinking on this subject is anecdotal, and we rarely consider the possible costs and risks of such testing.