VetInfoDigest

VetInfoDigest May 2007

Table of Contents:

Oral melanoma vaccine

Cancer prevention study

Anti-vomiting medication

Lineage by DNA

Microchip Wars

Evidence Based Medicine

This Month’s Note:

I set out this month to try to explain a little bit about how to evaluate statistics as they apply to medicine. Statistics are sometimes difficult to understand but they are important for proper analysis of medical data. As I wrote the newsletter, though, it occurred to me that it was also important to understand the types of scientific papers that are available and to be able to discern what isn’t scientific, despite claims to the contrary. This seemed like a more important starting point and so I put off trying to deal with statistics. After making this decision a number of interesting new medications and new information became available this month and I had to make a little room for them, as well.

New oral melanoma vaccine approved.

Merial ™ , working with the Animal Medical Center and Memorial Sloan-Kettering Cancer Center, have won approval for a therapeutic vaccination to be used when dogs have oral melanoma. The vaccine is not meant to be a preventative measure like most vaccines. It is given after the diagnosis is made in an effort to help the dog’s body fight the cancer. Oral melanomas are one of the most aggressive tumors in dogs. Life expectancy after diagnosis is between one and five months. The use of the vaccine provided an increase in life expectancy following diagnosis with vaccinated dogs living about 13 months (median survival time). This isn’t perfect but it is an improvement. In addition, this approach holds promise for people so data obtained from use in dogs could be helpful for that purpose, as well. The vaccine should be available by the middle of this month. Your veterinarian can contact Merial for more information if necessary. Hopefully none of you will have need for this information but it is good to know that there is at least some option for helping with this particular cancer.

Cancer Prevention Study

If you own a rottweiler whose age is 5 or 6 years of age, Proctor and Gamble and the Murphy Foundation would like you to consider enrolling in an study of the impact of anti-oxidants on cancer prevention. The foundation hopes to enroll 700 dogs in this study and to study them for 8 years. Rottweilers have been chosen due to the high prevalence of osteosarcoma in this breed, making it likely that there will be enough dogs who develop cancer to make the study valid. The study will be randomized, double-blind (neither the owners or the vets will know which dogs are getting placebos and which the anti-oxidants). This also could be an important study for both canine and human medicine because the role of anti-oxidants in cancer prevention has not been definitively proven at this time, to the best of my knowledge. The study announcement I drew this information from was in Veterinary Practice News, May 2007.

Maropitant ( Cerenia Rx), New Medication to Control Vomiting in Dogs

Maropitant is a new medication to control vomiting in dogs. Prior to the approval of this medication there has not been a really effective medication to control medication in dogs. Ondansetron (Zofran Rx) works well in dogs but is not approved for use in dogs. This medication is approved for use treating motion sickness and acute vomiting from other causes on dogs. It may prove to be very valuable in treating parvovirus, as control of vomiting in that disease can be difficult but dogs whose vomiting stops or can be controlled generally do better. It is made by Pfizer and should be available sometime this summer.

What Do You Think My Dog Is, Doc?

This is a common question in veterinary medicine. About 50% of dogs that come to veterinary practices are mixed breed dogs. It is surprising how different a puppy can look from either parent once interbreeding of purebreds occurs and lots of dogs have mixed parentage, as well. If you absolutely positively have to know what breeds are in your dog’s background Mars Veterinary will be marketing a DNA kit that can be used to sort out the breeds in a dog’s lineage. Over 100 breeds can be identified by genetic markers. I don’t have any pricing information but this is the URL of the company’s website: http://www.marsveterinary.com/

Prozac for Dogs is called Reconcile

Lilly has gained FDA approval for fluoxetine for dogs. Vets have been using fluoxetine pretty widely since it became available generically for humans so I’m not sure why Lilly went through the drug approval process but there may be some advantages to their product, other than the legal ones associated with using an approved drug. The tablets are chewable and the company plans to provide behavioral tips as well as access for your veterinarian to behavioral advice from the company. This may make it more palatable for some vets to use the medication who have been reluctant in the past. Reconcile is approved for use to control separation anxiety and the behavioral program is geared towards this disorder.

Microchip Wars

Some of you are probably aware of the fact that over the years there have been several types of microchips made for use in pets. The microchips that have been available in the U.S. have operated on different frequencies and readers did not always identify all types of chips. The international standard of 134 MHz for the microchips calls for a different frequency altogether. Bayer has announced a new microchip that operates on the international frequency but the accompanying reader is supposed to be able to read all types of microchips available today. This is re-opening the microchip wars, which were settled by lawsuit last time. I suspect that there may be litigation again. I wish that I knew the best approach to having a pet microchipped. It seems to make sense to use a chip that meets the international standards but the catch is that clinics and humane organizations with the old readers still won’t be able to read the new chips. Eventually I really think that everyone is going to end up having to go with the international standard but for right now it is still difficult to say which microchip is best to use. Bayer is providing readers to many shelters and some veterinary practices according to their press releases. Their microchip program will be marketed under the name ResQ. The major competition comes from Avid ™ and HomeAgain ™ chips, both of which read at 125 MHz. Avid has a history of aggressively defending the microchip turf but Bayer is a big company so it is likely that it will be able to fend off any legal challenges.

Understanding Why it is Important to Read Medical Reports Carefully

You may have heard of “evidence based medicine”. There is a push right now in the human medical community to change from the original model of medicine, which really was a combination of “practice” and “art” to a scientifically based model of medicine in which outcomes are examined carefully and it is possible to learn not only from one’s personal experience and mentors but also from the experiences of large numbers of doctors that have been compiled to allow statistical verification of the outcomes.

Compiling and examining valid data allows the medical community to share experiences in a meaningful way and to sort out which treatments actually work. You might think that practice experience would be enough to do that but it has major drawbacks.

The first drawback is the amount of time it takes to get experience and the toll that takes on some patients whose own experiences aren’t so great during the learning process.

The second drawback is that almost all humans are very flawed at remembering and organizing information naturally. We are biologically compelled to link a cause to an observed effect. This is an important survival instinct but it is very easy to mistakenly link any available “cause” with the observed effect. The only way to correct for this problem is to repeat the situation over and over until it is clear that whether or not there is a link between them.

A third drawback with traditional learning from practice experience is that humans are very often biased and almost as often reluctant to admit that to themselves. At my wife’s graduation ceremony from veterinary school the speaker said that it was a shame that humans spend almost all of their brainpower thinking up rationalizations to justify doing what their hormones were compelling them to do in the first place. Using some method to remove bias from observations is the best way to avoid this particular problem. It is difficult to accept blame for a patient’s death or to recognize that your actions caused harm. It is easier to rationalize that outside influences defeated your best efforts. If medicine is to advance properly it is necessary to remove rationalizations and biases to the greatest possible extent.

Knowing what the problems are and solving them are two entirely different issues. There are some inherent problems in veterinary medicine that prevent full utilization of an evidence based approach to veterinary care. The following paragraphs describe information sources and how they relate to an evidence based approach to medicine. I will try to show the differences between the ideal and the practical as they apply to the current situation in veterinary medicine as we go through them.

The highest standard of evidence based medicine is usually considered to be meta-analysis of a number of studies pertaining to the subject at hand. In order for a meta-analysis to be valid, though, the included studies should be similar enough that comparison is valid and should meet the standards expected of well designed scientific studies. In veterinary medicine there are frequently no large studies, the studies that are available vary significantly in their approach or design and it may be necessary to include poorly designed studies in order to include enough studies to make a meta-analysis possible.

A meta-analysis of poorly designed studies is no better than reviewing the results of the individual studies. In human medicine meta-analysis is becoming very important to researchers and clinicians but there may never be a time in veterinary medicine when it is possible to have great faith in meta-analysis. There just aren’t enough studies of large numbers of veterinary patients to provide valid results for most topics in veterinary medicine.

A literature review of existing studies can come close in value to a meta-analysis and is more appropriate to the level of statistical data available in veterinary medicine. Literature reviews must be carefully examined to ensure that they are not biased, that references relevant to the question being examined are used and that the studies chosen for review are themselves of the highest possible quality.

The second tier level for studies that contribute to evidence based medicine are well designed scientific studies that meet certain criteria. There are generally two categories of these, experimental studies and observational studies.

Experimental studies attempt to follow the general concept of scientific principles. . In science the ideal situation is to formulate an hypothesis, design a study that will prove or disprove that hypothesis, test the hypothesis with the study and then revise the process based on the information obtained, until it is possible to prove or disprove the original hypothesis or to formulate a new hypothesis that matches the data generated. This sounds complicated just to describe andit is a much more complicated task in medicine than it is in some other fields of endeavor. In many fields it is possible to isolate the subjects of the study so that only one variable changes in the environment, making it the sole reason that an action causes a reaction or a process occurs. In medicine it is often impossible to completely remove all variables except one. There are just too many variables that affect a living organism. This is one of the reasons that many people continue to argue that medicine is an art rather than a science. Despite that feeling, it is possible to come up with ways to study some aspects of medicine in a scientific manner. It seems unlikely that we will ever figure out a way to test all aspects of medicine using scientific principles, though. There are far too many variables to control all of them when studying live subjects.

It is generally agreed that the first step in producing a good scientific study is to ensure that the study is properly randomized. If it is a study of treatments for a particular condition the patients should be assigned to the treatment groups randomly. If it is a dietary trial then the patients getting a particular diet should be randomly selected. It can be hard to understand why random selection of patient is so important. In medicine it is almost impossible to study one single interaction between a patient and its environment in isolation. It is impossible to know if some or even all of the patients in a study will be affected by an outside factor, such as environmental temperature, dietary differences, pre-existing unrecognized disease conditions or anything else. The only way to balance the patients who might be affected negatively and those who have positive effects that are unrecognized is to include as many patients as possible and to select them randomly so that negative and positive effects will balance out as much as possible. While I don’t think that there are a huge number of unprincipled researchers one of the easiest ways to effect the outcome of a study is to select the subjects included in the study so that the outcome desired is more likely. This is the major reason to be suspicious of studies funded by drug manufacturers or others with an economic interest in the outcome.

One of the most important basic concepts is the population that the study is drawn from. Understanding how the researcher identified members of the study population is critical to understanding the value of the study. There is a big difference between studying the effects of a drug on neonates, or geriatrics, and studying it in the entire population of patients, for instance. Patient subsets within the population being studied can have a very big impact on the data. This is particularly true when studying dogs in veterinary medicine due to the effect that breed variations can have on studies. A very good study on arthritis in beagles may still not be representative of arthritis in all dog breeds or in some other particular breed, such as dachshunds. It may better represent arthritis in small dogs in general, than in large dogs in general. You must think about how the population of patients chosen for a study affect the results as they pertain to your pet. In particular it is important not to assume that a cat is a lot like a small dog.

Population variations are very vexing when dealing with some diseases. There are a number of studies of cancer for which no distinction is made between dogs or cats with different grades or stages of a particular cancer. When a pet owner has a collie with Grade 2, Stage 2 mast cell cancer it is hard to find a study that is specific to this particular breed, grade and stage of cancer. There just isn’t a good data set for patients this specific, most of the time. The closer the population in a study matches your pet, the better. Even with a near perfect match it is important to understand that statistics still don’t give an exact idea of what might happen with your pet. Statistics are great for trends within a population but they aren’t as useful for predicting the outcome of one particular individual within the population.

If we agree that random selection among patients who meet the criteria for a study is usually best, we still have a major problem in veterinary medicine. It is difficult to actually select patients at random in many instances. This is especially true for relatively rare conditions and even impossible for extremely rare conditions. If you are looking at a study of hip dysplasia, which is very common, it makes sense to restrict your studies to ones in which the population of dogs studied is randomly selected for inclusion in the study. If you are looking for information on pheochromocytomas, you might have to include any study that mentions the problem and accept that the quality of the data is not as good as you would like. While it is necessary to accept this limitation it is important to realize that despite the best intentions of the authors, such studies may contain invalid conclusions due to the small number of patients available for inclusion in the studies.

Observational studies are less rigid scientifically because the original population in the study can’t be randomized in most cases. There are a number of types of studies that fall into the category of observational studies. These studies usually involve groups of animals that are followed over a period of time and outcomes of treatment, disease exposure, diet or other factors observed in the group. In general larger groups and more tightly controlled environments during the studies produce the most trusted results.

Case studies are the lowest level of information that is really considered for evidence based medicine. The major problem with case studies, even those that include several animals, is that there are usually no controls, the patients are not randomly selected and there is usually no effort made to avoid biases a clinician may have.

There are a fairly large number of problems in veterinary medicine for which the only available literature references are case studies, unfortunately. To give you some idea of how this affects clinical decision making, I had a case in the office that seemed clearly to be a recognizable rare disease. After making the diagnosis I checked three or four text books for information. After reading the texts I checked the listed references. Despite two or three references in each text I was able to determine that all of the information in all four texts was actually based on a single case report. I didn’t follow the textbook suggestions for treatment because they didn’t seem appropriate to my patient and I knew that they were not much better than my personal experience, anyway. If you are going to use case reports to make decisions about health care for your pets, make sure that you see more than one case report, if at all possible.

The lowest level of literature, at least from the standpoint of evidence based medicine, is opinion. Even opinions come in more than one variety when thinking about health decisions, though. There is expert opinion, based on the practice knowledge of someone working in the field and hopefully a recognized expert. There is also anecdotal information, more along the lines of “my veterinarian put my dog on comet dust capsules and his hair coat is shiny, his gums are pink and he can leap small buildings in a single bound.”

So the order of information quality established for evidence based medicine is: 1) meta-analysis 2) systematic literature review 3) randomized controlled experimental trials 4) observational studies 5) cases studies 6) expert opinion and 7) anecdotal opinion.

By observing veterinary clients over the years (and note that in this case by observing I am referring more to expert opinion than to observational studies), I really believe that in order of influence upon client decision making the ranking would look more like this: 1) expert opinion when an expert is sufficiently well known 2) anecdotal opinion 3) observational studies 4) systematic literature review 5) experimental studies 6) case reports and 7) meta-analysis.

To be fair, I also think that among veterinarians the order would be different on the average, than the list for evidence based medicine. I think vets probably lean towards 1) randomized controlled experimental trials 2) literature review 3) expert opinion 4) observational trials 5)case reports 6) meta-analysis and 7) anecdotal opinion. I think that meta-analysis might rank higher among veterinary clinicians if the early use of it in veterinary medicine hadn’t been carried out poorly or used to justify therapies based on poor meta-analysis.

It is hard to come to grips with the concept that your individual observations, or those of your friends, are likely to be invalid much of the time. It makes sense to us that folks who love pets would be able to tell when a medication was helping, when dietary changes were working or when their pet was ill (or not ill). Unfortunately, well controlled studies show that people are often wrong and that this doesn’t just apply to pet owners. Veterinarians are fairly easy to fool, too.

Some disorders have a tendency to wax and wane on their own. Arthritis frequently varies in its impact on affected patients, for instance. It is not unusual for a patient with arthritis to have a period in which the discomfort or disability intensifies. It is also not unusual for these periods to end, with or without treatment. Any treatment that is given coincidentally with a period of relative comfort will appear to work. For studies involving this type of problem it is particularly important to have fairly large patient numbers and to study patients for enough time to account for at least some natural waxing and waning of symptoms. This is quite difficult to do when attempting to evaluate a single patient or even a group of patients over time, as veterinarians do.

If you don’t think that people can be fooled easily, consider the case of carprofen (Rimadyl, Rx). I’m not referring to any problems with the medication but rather to the original studies done in order to obtain FDA approval for the drug. In the original studies, in which neither the veterinarian or pet owner knew which dogs got the real medication and which dogs got the placebo, approximately 50% of pet owners thought their dogs improved when they were given the placebo rather than the carprofen in the controlled trials. That means that half of the pet owners were seeing improvement in their pet’s condition even though a known ineffective medication was being administered. In this same study 25% of veterinarians thought that their patients receiving the placebo got better. To be honest, I was surprised that the veterinarians did this well. It is just human nature to look for benefits when giving a medication.

Another example of the difficulty with anecdotal information comes from studies done on chronic diarrhea. Changing the type of diet, no matter what diet the pet was on or what diet it was changed to, appeared to correct the problem in approximately 25% of patients. High fiber, low fiber, high protein, low fat and several other dietary types were included in the literature review and some patients appeared to improve more than once when they relapsed and another alternative diet was introduced.

Studies of seizure medications are frequently complicated by the fact that seizures do not occur on a random basis over time in most patients. It is not unusual for dogs experiencing fairly frequent seizures for several months or more to suddenly stop having seizures for long periods of time. It can be extremely difficult to determine if medications are helping to control seizures due to this problem. For several years veterinary neurologists thought that phenytoin (Dilantin Rx) was effective for seizure control in dogs. Finally, a controlled study showed that the drug had a half life so short in dogs and was so poorly absorbed from the digestive tract that it seemed unlikely that it could be helping to control seizures. The natural changes in seizure frequency over time were just fooling people into thinking the medication was having an effect. Board certified veterinary internal medicine and neurology specialists were fooled by this drug, pointing out that even “expert” opinion can sometimes be wrong.

When you read a study it is a good idea to try to evaluate the selection process for patients included in the study and also to think about how the ones that were excluded or dropped from the study might have affected the data presented. I have actually read a study that purported to evaluate the safety of a pharmaceutical in which three dogs were “dropped” from the study because they died. The authors concluded that it wasn’t possible to confirm that the deaths were related to the medication and just left these dogs out of the data. That kind of decision making clearly has an impact on the meaning of the statistics that are presented. If a dog in a study on pharmaceuticals is dropped because it was hit by a car and died the meaning of the change in the data is much less than if it died of unknown causes in the middle of the study. It is also fairly important to note how much confirmation of a disease process was needed for a pet to qualify for a study. A review of corneal ulcers due to herpes virus in cats is more meaningful if herpes virus can be isolated from the lesion than if inclusion in the study is based on a veterinarian’s impression that the lesion was caused by herpesvirus, for instance.

It is worthwhile when reading a research paper to look at who collected the data and how they did it. This is generally described in the paper. Sometimes this also has a great impact on the usefulness of a study. As an example, the use of pain relief for declawing cats has been studied several times. In one of the studies the surgery was done by students at a veterinary school, in others by experienced surgeons. It seems reasonable to assume that this might have some impact on the length of surgery and the need for pain relief.

Perhaps the most important thing to cover is that you must take the time to read the entire paper if it is important to you. While you can almost always find abstracts of papers on PubMed or other web sites, it is difficult to really evaluate the paper if you don’t have access to the whole paper. It is worth paying for a copy of articles that look like they will be useful when researching a problem your pet might have. Many of the abstracts will have links to the publisher or to Loansome Doc, where you can obtain a copy of the documents you really need. By reading the whole paper you can really evaluate whether the abstract, which is often written by the author, really does support the conclusions the author draws from the research.

Thanks for your Support!

The VetInfo Digest is published by:

TierCom, Inc.

P.O. Box 476

Cobbs Creek, VA 23035.

The opinions expressed in this newsletter are those of Michael Richards, DVM., author.