Health Effects Of Spey/neuter: Labs And Goldies Differ

[sandgrubber]

One of the vet-related blogs I follow posted notes on a recent comparative study looking at longer term health effects of spey/neuter in Labradors and in Goldies. Surprise! The breeds were significantly different! Here's a summary:

First off, what does the study appear to show? Well, perhaps the most important finding is that there are significant differences in the pattern of possible effects of neutering between males and females, and also between the breeds. This latter point is extremely important. People tend to jump on the findings from a study in a single breed and want to generalize those findings to all dogs and make broad recommendations based on that single study. However, if the possible effects of neutering differ significantly between two breeds as closely related and similar as Golden retrievers and Labrador retrievers, likely these effects will be even more dissimilar in breeds less closely related. A study of Rottweilers or Viszlas, for example, may tell us little or nothing about what to expect in Beagles or Poodles. This is worth keeping in mind before we rush into making sweeping recommendations for all dogs.

The findings in this study suggest some increase in the risk of some diseases with neutering, but the patterns were not very consistent. Neutering before 6 months of age appeared to be associated with higher risk of cruciate ligament disease and elbow dysplasia in male Labradors but not with the risk of hip dysplasia. In females, however, the pattern was the opposite, with neutering at less than 6 months being associated with greater risk of hip dysplasia but not the other two orthopedic diseases. But in Golden retrievers, neutering before 6 months of age was associated with greater risk of cruciate ligament disease in both sexes, but with greater risk of hip dysplasia only in males, not females.

Source

http://skeptvet.com/Blog/2014/07/potential-risks-of-neutering-and-age-at-neutering-for-godlen-retrievers-and-labrador-retrievers/

reviewing

Hart BL, Hart LA, Thigpen AP, Willits NH. Long-Term Health Effects of Neutering Dogs: Comparison of Labrador Retrievers with Golden Retrievers. PLoS ONE 2014;9(7): e102241. doi:10.1371/journal.pone.0102241

Sigh! Once again, science tells us: "It's complicated."

[Dame Aussie]

Very interesting! This is why I don't like the push to desex anything and everything without even looking at the risks/benefits.

[Justrace]

It is complicated It really is like " how long is a piece of string" ?

[Christina]

One of the most relevant factors in credibility of studies & statistics is how many are used in the study & by how many different companies or individuals.

A study of 500 may well give very different results to a study of 10,000.

Always worth noting & finding out before giving complete credibility.

[denali]

One of the most relevant factors in credibility of studies & statistics is how many are used in the study & by how many different companies or individuals.

A study of 500 may well give very different results to a study of 10,000.

Always worth noting & finding out before giving complete credibility.

Very true. Something even i am guilty of overlooking and not checking at times.

[raineth]

One of the most relevant factors in credibility of studies & statistics is how many are used in the study & by how many different companies or individuals.

A study of 500 may well give very different results to a study of 10,000.

Always worth noting & finding out before giving complete credibility.

actually, there comes a point where increasing the sample size doesn't really contribute much at all :)

The statistical analysis also takes into account the size of the sample. When something is found to be 'statistically significant' it means that there is less than 5% chance that the findings were a result of chance. Sometimes it is less than a 1% chance, or even less than that.

Smaller samples require there to be larger differences in order to be found statistically significant.

There is a lot to critically examining a research article, and sample size is only one factor :)

Other things to look at is the study design, the effect size, how they chose to measure their variables and the representativeness of the sample. So if I was happy with these things listed above, I wouldn't disregard the findings just because they didn't have a massive sample.

[Dame Aussie]

One of the most relevant factors in credibility of studies & statistics is how many are used in the study & by how many different companies or individuals.

A study of 500 may well give very different results to a study of 10,000.

Always worth noting & finding out before giving complete credibility.

actually, there comes a point where increasing the sample size doesn't really contribute much at all :)

The statistical analysis also takes into account the size of the sample. When something is found to be 'statistically significant' it means that there is less than 5% chance that the findings were a result of chance. Sometimes it is less than a 1% chance, or even less than that.

Smaller samples require there to be larger differences in order to be found statistically significant.

There is a lot to critically examining a research article, and sample size is only one factor :)

Other things to look at is the study design, the effect size, how they chose to measure their variables and the representativeness of the sample. So if I was happy with these things listed above, I wouldn't disregard the findings just because they didn't have a massive sample.

Spot on. I'm studying statistics right now :laugh:

[The Spotted Devil]

My general issue with the spay/neuter studies is that they are retrospective and not experimental. Dogs are not randomly spayed/neutered and there may be other causal factors such as how the dog is raised, it's conformation, it's purpose, owner knowledge etc.

[sandgrubber]

Skeptvet's main reservation about the study was not sample size, but the potential bias of using dogs presenting at a University veterinary hospital.

Still, the bias should be the same for Goldies and Labs . . . so the difference between them is probably real, even if the figures are not representative of the two breeds as a whole.

As for experimental studies . . . cost prohibitive. Who would sponsor keeping 60 dogs for 10 yr under controlled conditions? And from the results of this study, it looks like you might have to do 60 dogs for each breed! (As a rule of thumb, you need something like 30 for control and 30 for experimental group to see statistically meaningful results. Larger samples needed if effects are subtle, smaller if effects are dramatic.)

[The Spotted Devil]

Oh I agree. But I think caution is warranted. It's like using the correlations between entire dogs and shelter admissions to say everything should be desexed.

[Mrs Rusty Bucket]

Golden Retriever cases with complete data for analyses totaled

1,015, with 543 males (315 neutered and 228 intact) and 472

females (306 neutered and 166 intact). Labrador Retriever cases

with complete data for analyses totaled 1,500 cases with 808 males

(272 neutered and 536 intact) and 692 females (347 neutered and

345 intact).

they only considered a disease significant if there were 15 cases or more...

And these were USA dogs in a uni vet clinic.

So their breeding would be a little different to what's available in Australia and there was already a reduced sample based on where these dogs were seen.

Would like to see a study where all vets could contribute data...

If this was a human study it would not be significant enough to make a decision. But I guess it's a start.

[sandgrubber]

Golden Retriever cases with complete data for analyses totaled

1,015, with 543 males (315 neutered and 228 intact) and 472

females (306 neutered and 166 intact). Labrador Retriever cases

with complete data for analyses totaled 1,500 cases with 808 males

(272 neutered and 536 intact) and 692 females (347 neutered and

345 intact).

they only considered a disease significant if there were 15 cases or more...

And these were USA dogs in a uni vet clinic.

So their breeding would be a little different to what's available in Australia and there was already a reduced sample based on where these dogs were seen.

Would like to see a study where all vets could contribute data...

If this was a human study it would not be significant enough to make a decision. But I guess it's a start.

I would guess they needed 15+ cases for a disease to be statistically significant.

A lot more data would be good. A random as opposed to opportunistic sample would be much preferred. But hell's bells. Most of the dog medical opinions we hear are based on anecdotal evidence: one vet's experience, one person's recollections of their dogs and the dogs owned by their friends and family. Memory data are distorted by selective memories and readily lead to conclusions based on spurious events. These guys, at least, were able to use statistical tests.

In human medicine, they register cause of death, and with a large fraction of the population either insured or in care of a government run program, there are lots of institutions with a financial interest in tracking morbidity.

IMO it's an enormous failing of the pedigree system that the KC's don't track, at least, date and cause of death; date of neutering would be helpful, too. It's extremely frustrating that little data are available on morbidity. Eg, I'd love to know not just hip and elbow scores, but whether or not the dog made it to old age without stiffness.

Edited September 23, 2014 by sandgrubber