Cardigan hip scores: A dose of orthopedic reality (Cardigan OFA, PennHIP, DI, etc.)

Gird your loins, because what I am about to say is (and should be) controversial. The only thing I can say is that I came into this research with a very strong bias toward what I’ve always thought of as the one incontrovertible rule in dog breeding: ALWAYS TEST THE HIPS AND ALWAYS SUBMIT THE TEST. What I will explain below was so striking to me that I find I don’t have anywhere near the same confidence in that unbreakable rule. So don’t dismiss or crucify me before you read and understand it too.

When I was hit with two pelvic “crises” in the same month (Clue’s injury and x-rays, including trying to interpret her eventual breeding soundness, and Bronte’s breeding date being moved up dramatically) I knew I had to get my mind around the hip issue in dwarfed dogs. I wanted to make sure that I was making good decisions and I understood the potential outcomes within various theoretical matings.

I started this in the same way I have always done: a pedigree analysis using OFA and some discreet inquiries to people who have been in the breed a lot longer than I have.

Here’s what I found:

1) There is a very troubling and almost complete lack of consistency in the OFA results for this breed. Ditto for PennHIP. I’m not talking about the fact that too few people are certifying; I mean that there is virtually no predictability in scores based on the pedigrees. I am computer-savvy enough to piece together a lot of information from OFA searches, so I didn’t find myself fatally hampered by the fact that there aren’t as many actual results in the pedigrees; I could always find a brother or a cousin or whatever and usually multiples of both.

What I saw was a situation that looked totally random and inconsistent. Two dogs with good ratings producing, within a single litter, everything from Excellent to Moderate. Stud dog lines that I know to have a moderate to high COI (which should indicate consistency) flipping from good to dysplastic and back again.

It didn’t make any sense. Even when you have a breed that is genuinely in trouble, the dysplasia tends to fall into family lines and at least a few people have consistent success. When you see the bad family lines doubled up on, you get worse hips again. You can get some unexpected stuff, but in general good hips make good hips and bad ones make bad ones. And so on. The OFA picture in Cardigans was one of seemingly random and completely unpredictable results.

2) Hip scores do not correlate anywhere near as well as they should with soundness or comfortable working lives. When I spoke to the orthopedic surgeon about Clue, I got some very good and very candidly given information. He said several things: This is not a breed he sees, as a surgeon who does lots of work to relieve pain in hips or to analyze x-rays. This is not a breed coming in with pain issues. It’s not a breed he associates with dysplasia symptoms with any regularity. He does not recommend OFA for corgis of any type, because he feels that the scoring is more or less guesswork unless the joint is clearly already arthritic or the socket just plain doesn’t exist, and EVEN THEN he rarely sees dogs come in with pain.

He said, and this is close to an exact quote, “These are dogs with weird hips, and they get along just beautifully on those weird hips.”

THIS IS SO IMPORTANT. I cannot make that in caps big enough.

There is a phenomenon called “a disease of numbers.” Good human docs know it well. A disease of numbers is a condition that causes a value on a test or chart to go high or low but may not actually correlate with any bad outcome.

Whenever you’re doing medical research, you have to be aware of this phenomenon and prove beyond a shadow of a doubt that the numbers are actually predictive. So nobody stops with “Your blood pressure is high. That’s bad.” It has to be “Your blood pressure is high. That’s bad because we know FOR SURE that high blood pressure is associated with these ten disorders, with precise and reliable predictive accuracy.”

My point here is that we have not satisfied that burden of proof when it comes to the classic OFA-type evaluation of Cardigan hips. We do not know at what point a hip will or will not cause quality of life issues. We are trying to evaluate them based on how close they are to the hips of a long-legged, short-backed dog (the greyhound–that’s the ideal hip and the one OFA is based on). Think about that: Does our breed look or function ANYTHING like a greyhound? They have entirely different biomechanics and centers of gravity, very different muscle formation, completely different cartilage, ligaments of a different shape and strength. What proof is there that making their hips more like a greyhound’s actually improves quality of life?

Conversely, what proof is there that looser or shallower hips mean a poorer quality of life FOR THIS BREED?

I would also add that we do not know how much the other things we classically do “wrong” with Cardi puppies (like keep them fat, feed them puppy foods, and flip them schizophrenically from hours on a soft couch to hours of play on hard surfaces) are affecting the eventual OFA-type view. We know that each one of those four things is bad for hips, but they continue to be the way most breeders and most owners raise puppies.

It’s at least theoretically possible that the giant variations within litters are the artifacts of these decisions: One owner keeps puppies lean to the point of being skinny, restricts calories and protein, has multiple dogs, and encourages vast amounts of hard free play in large areas for the majority of each day. The other has a soft, chubby puppy on Eukanuba puppy food; the dog is adored and carried much of the day and has a soft bed the rest of the time, gets two walks a day on a blacktop sidewalk, and on the weekends goes to puppy daycare (on a concrete surface, of course) or the dog park and runs to exhaustion.

We all know how adaptable Cardis are. It’s not like buying a terrier puppy who will make you pay for it if you don’t exercise them constantly–they’ll happily spend most of their day under your chair even as babies. And they’ll happily switch “on” and run for hours on the weekends. But that kind of love is MURDER on developing joints.

So, my research and the ortho vet said, relying on an OFA score is not a great idea and is not likely to get you where you want to be: Producing dogs who will predictably have a long, pain-free life. Remember, that’s what we care about. Long, pain-free life. Not a number, a life. The numbers serve us, not the other way around.

The ortho vet recommended that I look more closely at PennHIP because the database is breed-specific. I agree. Purely on that basis alone it is a better measure than OFA; it doesn’t compare Cardis to greyhounds but Cardis to Cardis.

But there I was stumped again. What I got from conversations with people was what virtually all breeders look at and remember: the percentile score. That magic little bar with the carat marking where your dog lies. Pray for over 50.

And AGAIN, percentile scores were CRAZY. Within individual litters everything from 95% to 20%.

With that kind of wild variation in scores, I began to ask to see the difference in the actual DI numbers.

Here’s the background: PennHIP establishes its percentages based on all dogs of that breed that are submitted. 0% is the highest percentile (the loosest hips in the breed database) and 99% is the lowest (the tightest hips in the breed database). With dogs that have thousands of submissions and that really do have a bell curve of hip health (where most are “OK,” some are great, a few are outstanding; some are borderline, a few are terrible), like Labs and Shepherds, that works out pretty well. Dogs with DIs below the breed average really do begin to represent “bad” hips.

PennHIP hedges all its bets and leaves almost every decision to the breeder. It makes only two real statements: Hips with scores LOWER than .3 represent a very, very small possibility of degenerative joint disease  (DJD) as the dog ages. Notice that they do not use the word “dysplasia.” That’s very smart on their part, because the problem with quality of life is not how loose, tight, shallow, or deep the hips are. The problem is DJD, which in layman’s terms is pain and arthritis. Instead of cartilage moving smoothly on cartilage, bone grates against bone. Range of motion is impaired. Weight-bearing surfaces become painful.

The second thing the PennHIP group says is that scores above .7 demonstrate a very high probability of the dog developing DJD.

Now, immediately, you have to realize that in defining the numbers they’re falling into the OFA trap of basing statements on long-legged, big dogs. The breeds that contributed to this statement were German Shepherds, Labs (a huge number of Labs), Golden Retrievers, and Rottweilers.

So begin the very long process of evaluating PennHIP’s predictive nature with that in mind.

Getting back to percentile scores, where things get wonky is when you have either very tight hips across the breed or very loose hips across the breed, or where there are only a small number of submissions.

For example, if you submitted 50 borzoi films, the DI would range from probably .2 to .26. That means that the ones with the DI of .25 or .26 get percentile scores of 5% or 10%. That could lead you to believe that those are dogs with bad hips. Not so–every single one has gorgeous tight hips. The entire breed has tight hips, so getting a score of 5% is meaningless. This is close to the case in Danes, at least it was when I was PennHIP testing four years ago. Median (50%) was a DI of about .33. 90% was .27 or so; 5% was maybe .4-something. I don’t know if there have been many more submissions since then so the median has shifted, but back then only the most cautious breeders were even using PennHIP so they already had an entire kennel of dogs with Good or better hips. So virtually the entire PennHIP population was healthy.

Similarly, if you submitted 50 pug films, the DI would probably range from .7-.9. So the .7 dogs would be in the 99%, but that doesn’t mean they have good tight hips. The entire breed has BAD loose hips. But, curiously, pugs almost never show any hip pain and they live their obnoxious little lives with their bad loose hips.

A similar thing happens when there have only been a few tens or a few hundred submissions. The submissions will tend to cluster and won’t form a good bell curve, so the percentile is not necessarily accurate.

So when you have a bizarre range of percentiles you have to look very critically at the actual DI scores.

The figure above is my attempt to illustrate the percentile versus DI conundrum. Greyhounds cluster at the very top end of the tightness scale, so a low percentage is meaningless. Labradors span a HUGE area, so if you have a 95% Lab it really does have very good tight hips and when you have a 5% Lab it really does have crappy loose hips and a high probability of pain and the need for surgery to fix the joints.

Now let’s consider our little group.

Cardigans have a median (50%) DI of something like .61 or .62 (which would indicate a dog very likely to be crippled or dead by age 4 if you were talking about a Shepherd or a Dane).

Based on looking at a lot of scores, 99% would be something like .51 or .52 and a dog with a really good percentile score would be, say, .53.

BUT (and this is a BIG but) 30%, at least as of two years ago, is .625.

Were you paying attention?

50% (remember, that’s your goal, 50% or higher) is .61…but 30% is .625.  So 20% would be something like .63, and 0% would be a few hundredths below that, right?

That means the ENTIRE RANGE of the breed is .52 to maybe .64 or .65 or .66. That’s a VERY small range; the entire range of the Labs is about .2 to .8!

SO, the question becomes–is a .61 (the average) really a better indication of a lifetime free from pain than a .625 (30%) or a .64? Keeping in mind that a .61 means a crippled dog in most other breeds? If .61 doesn’t cripple a Cardi, does .63?

The DI represents the percentage of the femoral head that can be moved out of the socket. A dog with a .55 has a femoral head that has moved .55 of its diameter out of the socket.

THINK ABOUT THAT. Consider the difference between a .2 and a .8–this is an easy mental exercise. A .2 barely moves out of the socket. A .8 moves almost entirely out. Clearly, one is going to be a better and sounder hip for a lifetime of pain-free work.

Are you prepared to make the same statements about the difference between .61 and .63?

Are you prepared to remove a dog or bitch from a breeding program because its femoral head moved two percent further than your other dog? Are you OK with other breeders telling you that you should do that or you’re a bad breeder? Because that’s exactly what’s happening. Make public that you are breeding a dog with a 20% PennHIP score and see how many love letters you get. Criticism falls like hail and people’s reputations are trashed based on a hip movement of two percent, even though we have no earthly idea and zero scientific evidence that that difference means ANYTHING when it comes to lifelong soundness.

So, then, where the heck does that leave us?

The fact that I’ve just written like fifty paragraphs on this should give you a hint about how conflicted I am about the genuine worth of EITHER method.

And so I have come to a conclusion that is not easy: This is up to you as a breeder. It’s not up to the score. ANY hip score given by an outside agency should be looked at with suspicion (in terms of it meaning the slightest amount about the actual health or life quality of your eventual puppies).

Guess what this also means: You don’t get to make the “She’s a bad breeder because she doesn’t submit hips to OFA” or “She’s irresponsible because she bred a bitch with one hip that was loose” statements. I pretty categorically hate those kind of statements anyway, since if we’re being honest with ourselves they very rarely come from a good place in the soul, but blanket pronouncements are sometimes at least somewhat reliable in other breeds. In this one, I think it’s inescapable that you can be a very, very good, very health-oriented breeder and not be submitting to either registry and you can have a flawless understanding of PennHIP and breed a 20% hip.

You know who the global authorities on Cardigan hip health are? The breeders who have been responsible for their own x-rays and making their own decisions for the last two or three decades. Personally, I’d love to see a Nationals panel of three or four breeders who have been x-raying hips and putting their own breeding programs on the line and ask THEM what a breedable hip is. They’ve probably seen a hundred times more Cardigan hip films than any ortho vet doing his rotation at OFA.

You’ve also got to stop leaning on OFA or PennHIP scores in pedigrees. It was already a iffy idea because of the spottiness of the data; if you face the fact that you have no idea at what point hips are genuinely functional, and that a whole bunch of breeders and owners follow an absolutely ideal recipe for creating bad joints, you need to stop assuming that lot of OFA scores mean a better risk than a few, or that poor scores in some offspring make the dog a bad stud dog.

I can feel the hackles rising, so let me assure you that I do not think that hips don’t matter. I think we must x-ray them all. But I think that analyzing hips is a unique and huge responsibility for the Cardigan (and other dwarfed dog) breeder and if you’re smart you will not leave it solely to outside agencies to decide.

I think the best we can do is to look at as many aspects of the picture as we can, make certain decisions about what we absolutely will not tolerate (keeping in mind that there are some Cardigans with a complete LACK of hip socket) and, if a dog comes back from analysis with looser or shallower hips than we like but is definitely pain free, we weigh whether that dog is actually worth breeding, with the weight against a “yes” decision depending on the severity of the hip “issue.” Call it a serious fault, like a bad bite or a fiddle front. For some dogs, that answer will be absolutely, if you breed to something with really wonderful hip structure (like you would breed a fiddle-fronted dog only to something with minimal turnout, and that may mean that you never find the right match). For others, that serious a fault will (and should) disqualify them from breeding.

My own personal limit would always be at the level of causing pain. If I x-ray a dog at 20 months and I can already see arthritis in both hips, that dog would have to be pretty freaking outstanding to ever consider breeding and then I would do so only for myself or with full and complete disclosure to any other breeders. I already have a replacement puppy clause for pet owners, so that wouldn’t change. If I x-ray at four years and the dog has weird-looking hips but no bony changes, I would feel no guilt about breeding him or her.

I will continue to use tools like PennHIP, and I will continue to get an OFA-style x-ray of each dog’s hips, but I will not allow the percentile score to make decisions for me and I very much doubt that I will waste money on getting an opinion from three guys at OFA who are staring at an x-ray and saying “Wow, corgis have weird hips–I dunno, what do you think? Should we treat it like a Basset hip?” I already have two x-rays to begin my library and I will continue to build that library and continue to refine my decisions. And when I face you, as other breeders and as potential buyers, you’ll know that whatever decision I made is one that I own, and one that I take full responsibility for.

And if anyone catches me making those lovely comments about so-and-so being a bad breeder, slap me.