The PennHIP Radiograph Rationale, Technique, Differences, and Value

Revised 2012.

As the author of “Canine Hip Dysplasia,” and an international lecturer on orthopedic disorders, as well as a dog show judge, I am frequently asked to comment on similarities and differences in the procedures used and information obtained when radiographs are taken for OFA and other leg-extended positions as compared to the PennHIP evaluation, which you will see is an improved diagnostic technique.

The Methods

By now, you know that the acronym stands for (University of) Pennsylvania Hip Improvement Program. This program arose from scientific inquiry, which in turn had its roots in the two related parents of invention: need and curiosity. The need was the desire of breeders and buyers for an earlier idea of how good were the hips of their canine “products.” For several years in the beginning of OFA, breeders who got into the program in earnest made some progress, and many were able to avoid high incidence of severe HD in their lines. But a plateau was reached before total satisfaction could be attained, and they started to look for a means to progress beyond where they were, especially regarding early identification of the most likely carriers of the most “bad genes.” Curiosity is the very heart of science, the “need to know;” in this case the question was “What must we learn to do in order to provide that early information in a valid and reliable manner?” The 30-plus years’ history of the older hip dysplasia control programs had not resulted in satisfactory progress, so by the early 1990s researchers at that veterinary college in Philadelphia developed equipment and techniques to satisfy both breeder and scientist needs. Times change: what was acceptable in the past is not enough now; the bar has been raised, and to perform today we must jump higher, do better.

The Orthopedic Foundation for Animals was established in the mid 1960s to collect radiographic data on hip dysplasia (abnormal hip joint development) and to register and publicize those dogs with more normal joint appearance so breeders could avoid the worst ones, which also might be the worst “carriers.” The American Veterinary Medical Association (AVMA) developed guidelines for positioning the dog for its radiograph in order to show the maximum number and extent of bony growths and remodeling of bone contours. In doing so, vets discovered an important principal: there was a correlation between those abnormalities and laxity (loose fit). Both for the individual’s risk of affliction and the risk of bestowing the causative genes upon future descendants, the phrase “Tighter Is Better” became an obvious truth.
The AVMA position, adopted by OFA and foreign breed clubs, is that of a dog lying on its back in a similar way that we bipedal humans stretch out in our beds or coffins. It is certainly not a “natural” position for a quadripedal animal – one that travels on all four limbs of approximately equal lengths. In order to make a dog assume this supine humanoid position, the legs must be pulled (extended) with some force and restraint, or the dog would pull the knees up (flex them forward toward the chest and head). Conversely, the “neutral/natural” position for the standing or moving dog is with the vertical femurs (nearly 90 degrees from horizontal) making an angle with the pelvis of somewhere near 120 degrees. For Homo sapiens it is the erect position when standing or moving. Neutral means that position in which there is the greatest state of relaxation in the muscles used to extend or flex the limb. Not only are the muscles and ligaments most relaxed, but also the joints are then the loosest they will ever be. When the quadripedal dog or bipedal man is standing at ease, a very few nerve impulses are all that are needed to maintain balance by triggering a very few muscle fibers on all sides of the joint. The contractions in the rear parts of our legs keep us from falling forward, for example, while at the same time the momentary contraction of a few “front” muscle fibers counteract their effect.

It is very important to understand this stasis or position of most neutrality, this balance of forces, in order to understand one of the significant differences in AVMA’s current protocol and the position used by PennHIP. The AVMA-OFA position stretches (tightens) the muscles on the belly side and front of thigh while not letting those on the back side operate in contraction and balance. Using this view with legs extended unnaturally, we “wind up” the muscles, tendons, and ligaments in and around the hip joint and tighten the joint capsule. The soft tissues closest to the joint are primarily the white-tissue, high-collagen types such as tendons and ligaments, and these do not extend (change length) to the degree that muscle fiber can. Thus, the twisting of white-tissue fibers is like twisting a nylon rope with two sticks turning in opposite directions, but in this case it tends to cause bones to be pushed closer together – the femoral head deeper into the socket than it would otherwise be. This artificially tighter-than-natural aspect contributes to the high false-negative rates in the OFA-certified dogs, as pertaining to laxity. Remember, both degenerative joint disease (DJD) and joint space are grounds for diagnosing HD in this method. Penn makes a semantic distinction between DJD as the definition of HD, and laxity as being a risk factor for eventual DJD.

While the hip-extended position is best for discovering DJD, it is not best for uncovering latent laxity, or what I call “covert laxity.” False-negative means that a passing grade is given because the true laxity was not observed, and that is the biggest drawback of the hip-extended methods worldwide. There are some individuals (usually of certain giant mastiff-family breeds) that do not develop DJD but are OFA-assessed as dysplastic because of laxity at two years’ age. But even more importantly, there are a greater number of dogs of other breeds that are adjudged “normal” at one or two years but later develop DJD or produce an unacceptably high percentage of dysplastic descendants. Thus, the accuracy of the hip-extended methods is gravely flawed. The gene pool is hurt most by these false negative diagnoses.

Latest Improvements

Two movements in America arose in the past decade or two that promise better progress than does adherence to OFA numbers as the way to coxofemoral nirvana. One is the proposal to use a voluntary “open registry,” promulgated by the Institute for Genetic Disease Control (GDC). The other is PennHIP (University of Pennsylvania Veterinary School Hip Improvement Program). I had the pleasure of working with the OFA’s first “program director,” Penn’s Dr. Wayne Riser, when I was researching and preparing my book, Canine Hip Dysplasia, and I also have had the good fortune to visit Dr. Gail Smith (PennHIP) in Philadelphia in the late 1980s. I reviewed his methods, philosophy, and results, and am increasingly a supporter of this protocol. At present, only PennHIP has the accuracy, repeatability, precision, and scientific foundation for real and rapid progress in producing better hips. The Seeing Eye, Inc. has turned to the distraction index (PennHIP) as a means of assessing hip quality.
In the PennHIP technique, the dog is placed in a position that is even more neutral than standing naturally because the small effect of gravity is diminished. While under chemical relaxants sufficient to prevent resistance to manipulation, the dog’s femurs are spread apart (distracted) with the force applied as close to the hip joints as possible. One of three radiographic exposures is made at that time, and the actual displacement is measured. An index is calculated in order to take into account the various sizes of dogs and their femoral heads/acetabulums. Any dog with an index of lower than 0.3 is practically guaranteed to never get HD. So far there have only been a few “semi-exceptions” in the many thousands of dogs evaluated. PennHIP does not make breeding recommendations, only evaluations; it leaves the decisions up to you, and counseling up to your veterinarian and peers.

It should not be surprising to anyone that the looser the hips, the less accurate a prediction of a specific grade or severity might be, especially in the hip-extended method. HD is developmental (DJD might not show up right away), progressive (it’ll eventually be worse), and multifactorial (environment has a part to play in the expression of the bad genes). Some young dogs will get worse than others even with the same DI.

Other Differences

Other differences exist. There are three radiographs used in the PennHIP procedure, and only PennHIP-certified vets may submit them. Every dog’s films enter the database, so there is not the skew or bias as found with the OFA-type registries. The “first” film (actually, it doesn’t much matter in which order they are made) is identical to that used by the older method: the traditional extended-leg picture for the study of bone abnormalities – in some cases, especially the worst ones, laxity is also apparent here. The second film is of the knees-up neutral position with a very small compressive force pushing the femoral heads into the sockets. While not as important as the other two, this view allows an evaluation of congruity, how neatly the round head fits into the curve of the socket. It is the third view that really makes all the difference. While the dog is deeply “under,” the patented distractor unit is placed between the legs at the groin, roughly parallel to the pelvis. Twin bars in this device that is shaped like the Roman numeral II act as the fulcrum, and when the lower legs are held near the hocks and pressed together, the vet leverages the femoral heads away from each other and outward (laterally) from the sockets. No covert laxity escapes this view.

The films are sent to the PennHIP Analysis Center, where a handful of people evaluate them (OFA uses a panel of radiologists that rotates or varies constantly). DJD presence or absence is noted on the first film, and circle gauges are laid on the third radiograph for use in objectively measuring the displacement. It is here where the paths diverge markedly: OFA, AVMA, SV, and most foreign hip registries or breed clubs use only the subjective hip-extended view, while PennHIP adds the objective view. At Penn, the results are added to those already in the database and compared. A report is issued that gives the Distraction Index, which can be thought of as expressing the percentage that the head is out of the socket. Another part states where this particular dog stands in relation to the average (mean) for its breed, expressed as “percentile.” For example, if the mean DI for GSDs is 0.41, your Shepherd with a DI of 0.53 will be in a percentile between 50 and zero (worse than half of the breed). A percentile of 80 means that your dog has tighter (better) hips than about 80% of those in the breed. The mean can vary a little with time, especially when there is a low initial number of dogs in the database. However, there is no escaping the facts that “tighter is better” and that a relative threshold of safety of 0.3 exists.

PennHIP-certified vets have to pass a training and subsequent testing regimen. For OFA, any local practitioner may submit films, even if all she or he has ever X-rayed for in the past has been fractures. Some clubs, such as the SV (GSD club in Germany) have a list of approved vets who may submit films.

PennHIP researchers and method do not show estrus to be a factor in the distraction view. In fact, there appears to be no veterinary literature yet, to support the idea that it is so, even in the leg-extended view. Furthermore, a study performed at the veterinary school at U of PA definitively showed that hip laxity, whether on the distraction view or the hip extended view, was not affected by estrus. Their conclusion is that that scientific evidence refutes the purported relationship of estrus to hip laxity.


The great value of PennHIP is the higher accuracy and reliability of evaluations done at an early age, so owners don’t spend more money than necessary in training for more demanding work, or even breed a dog that has a relatively high risk of later transmitting many bad genes to progeny, or itself developing DJD. The accuracy and repeatability of DI is just about as valid at six months age as throughout life; in fact about 95% reliability is seen in pups even as young as four months. The report by OFA that they too, now have equal predictive value (JAVMA, 1997) was refuted by a University of Wisconsin study published later and has not been confirmed by other independent research. Similarly, the OFA claim of progress in the past quarter-century has not been supported by data or experience elsewhere. The claims in their news release were reduced to just a 2.83% increase by the time the article was reviewed and then published in JAVMA in 1997; that would indicate that the inflated numbers in the OFA mailings to clubs might not be all that impressive. All that our reliance on OFA numbers has done is to allow very slow, perhaps almost imperceptible, progress in some lines of some breeds and, in a statistically insignificant amount, the “excellent” ratings in a few breeds. In almost all others, more than thirty-five years of partial use of OFA for breeding decisions has resulted in no progress, and in a few breeds the situation may actually have worsened. Breeders complain of a plateau reached in rates of progress when relying solely on OFA certification.

During the seminar on HD and other orthopedic disorders that I have presented in many countries, I recommend a few points to keep in mind when comparing the methods:

  1. PennHIP is the hip-extended view plus two more radiographs that show different things,
  2. PennHIP has performed biomechanical studies on its radiographic positioning while others have not,
  3. PennHIP has performed much research in general and these have been published in refered journals to prove the science is valid. Those who quote old information and say that “School is still out on the PennHIP method” simply have been skipping classes in the past several years.

To replace the old combined-approach program of Bardens/Ortolani palpation, wedge X-ray, and OFA-Good or Excellent, today I recommend PennHIP’s improved technology at 4-6 months or at any other time before breeding, as a viable and more accurate evaluation than all of those. Incidentally, the Bardens and wedge techniques are no longer recommended simply because they have been supplanted by better techniques such as PennHIP and Zuchtwert. For breeding, I advise my audiences to breed only to a partner with PH higher than 50th percentile and a lower DI than the mean, or at least, a lower DI than their own dog has; if they really want to accelerate progress, to breed dogs with 0.3 or better. Get as close to that threshold as possible, consistent with preservation of breed type and character.

What Does This Mean? — The Situation Today: Slow Progress and Why

Why, after nearly some four decades of awareness, breeding changes, and study, do we continue to hear from disgruntled or dissatisfied dog buyers and breeders? Knowing that orthopedic disorders are almost all genetic, one might think that it would be a simple answer to just breed non-carriers of HD or ED (elbow disorders), or those with the best genetic bank for good joints, but it is discovering these dogs that is the challenge. Now that we have good diagnostic tools and effective hip registries, the next step toward progress is for each breeder to develop a breeding program. Fortunately, some breed clubs and other organizations have already done the greater part of laying a foundation. We have already potentially removed one of the two major obstacles to progress, lack of understanding – or in other words, a lack of good diagnostic guidance. After decades of using the hip-extended method, most or all of those agencies have not generated a reliable heritability figure for hip phenotype, nor has the method used in North America reduced the incidence of HD as an average, across the breed populations. Even when we look at subsets of canine populations in the serious hobbyist world, whether we speak of individual or group (club) efforts, we find that discontinued progress. One reason is the failure to adopt the better diagnostic techniques.

The other reason for insufficient progress in reducing and ultimately eliminating canine HD is non-compliance: the failure of most breeders to stick with a really vigorous program of control and reduction. As you might think, some breeders do their best to provide an environment that causes the least dysplasia. However, genes that induce HD will thus be masked and therefore retained in the stock. Few breeders are likely to provide knowingly the adverse eugenics environment that would reveal such genes. Part of that second reason (breeding practices) for slow progress is the win-at-all-costs attitude maintained by many of the more prolific breeders and leaders of breed clubs. In 1986 John Bardens, a friend and a widely respected veterinary researcher who perfected the puppy palpation procedure, wrote to me, “Many of the breeding [genetic] defects do not hit the breeder in the pocketbook, and winning in the show ring is all that’s important.”

In some parts of the world, organizations and individuals have made greater strides than those in North America have, but there is still a way to go. The requirement in Germany for all radiographs to be recorded and dogs’ results made known, is admirable. In later years, they added the requirement that stud dogs had to be re-radiographed after having produced a certain number of offspring. The “sometime-pressure” in the UK for vets to cooperate by sending in all films and getting the results posted in the GSD database founded by Dr. Malcolm Willis has helped a little. “Kiwis and Aussies” down-under use the UK system, but also have room for improvement, as the following example would indicate: I received a request for advice and counsel from a breeder in Australasia who sold a pet-price bitch (no guarantees), paid for the 12-month radiograph, and got a BVA-type score of 11 in the Australian hip scheme. When the bitch was approximately 3 years old, the buyers decided they wanted to breed her, had her re-radiographed, and the score was 81. Now, 11 is pretty good but 81 certainly is not. There are two likely reasons for the two different readings, and I suspect both are involved, even though the bitch had no clinical signs. One is the nearly-notorious inaccuracy of that supine, legs-extended procedure used in diagnosis in the bulk of the world. The other is the rule rather than the exception that loose hips at a young age (even if undetected) can appear even looser at an older age when examined by the old method, and that DJD (degenerative joint disease, arthritis, remodeling) is more likely then. On the other hand, the experience with the great majority of cases evaluated with the PennHIP method tells us that true laxity does not change significantly after 4 months of age. At least, it is a rare occurrence. If the prevailing culture and conventional wisdom amongst breeders and vets in New Zealand and Australia ignores the newer, improved, more accurate techniques, can they rightly blame the average breeder? However, using a method shown to be not the best available opens the door to litigation if defects should appear.

Progress in the United Kingdom

England, Scotland, Wales, and to a lesser extent countries with historical ties to England, such as Ireland, Singapore, “OZ and NZ,” South Africa, and a few others have the potential for making great strides in reducing HD. Part of the mechanism is in place; what breeders need to do is use it. However, it may be difficult to accomplish without government legislation or regulation by breed clubs and The Kennel Club. They certify hips at one year of age; whether by government force or voluntary peer pressure, I would like to see a reconfirmation of phenotype normalcy after 2 years of age. Where the UK scheme continues to fall short of being ideal, besides certifying at an early age, is in not requiring all films to be submitted for the statistical study.

The BVA system concerns nine features; values of zero (no irregularities) to 6 (horrible) are given to both left and right hips joints, and the columns added. Most good breeders refuse to use any dog with a grand total of anything more than 10. My late friend Dr. Malcolm Willis, had long reported, for many breeds, results with dogs’ identities, in a form useful to breeders. The British Veterinary Association’s scheme was adopted or copied in several countries historically connected to the old Empire. Besides giving a quantitative score, the BVA/GSDL/KC scheme also has another important advantage for breeders over the American OFA and some other systems: it does produce information on progeny for several breeds. Computer-retrievable data by kennel name, sex, birthdate, age at time of radiography, and numerical value for each hip are used for genetic analyses and for your own conclusions on with whom to breed Schatzie, or whether to breed at all in deference to waiting to buy a better dog. Say you like the looks of that dog that placed in the Top Ten at his breed’s national specialty show the past two or three years. You look up his published hip scores, the mean score of his offspring who are old enough to be assessed, and scan the column that tells you whether and by how much he improved on the hip scores of bitches he previously bred. If your breed club doesn’t have that information, and it’s likely it doesn’t, then it isn’t doing all it can to serve you and your breed. That’s where “politics” can have a rare, beneficial effect on purebred dogs and the sport. Get into or start a movement to require your national club to hire a geneticist and give instructions to set up a scheme similar to that now employed by BVA/KC. Yes, you can go it alone, but your choices of breeding animals will be more limited than if you were backed with the power of a club like the U.K.’s GSD League or BAGS, or the GSD Council of Australia.

However, despite one of the most advanced information and control schemes in the world, the mean scores for GSD males and females born in the UK since 1959 have not changed a whit. About 45% of the UK’s GSDs have scores of 10 or below, with most considering the really “normal” ones as being in the 0-5 range and the 6-10s being equivalent to what we might call “near-normal.” BVA scores as high as 20 could encompass the level of quality in dogs given the ‘A’ stamp in Australia (not the same meaning as the FCI’s “A” designation for normal hips), but allowing that many dogs to breed will slow the progress, regardless of breed or country. Much better to make the requirements more strict each year until something approaching the Swedish model can be had. Progeny data are often seen in tables published in breed magazines. Obviously, those sires that produce higher percentages in the 0-5 score category and (of slightly lesser importance) a close second-high percentage in the 6-10 column, are the most desirable for improvement in hips and should be preferentially bred to, as long as they also produce other important good features.

Progress in Australasia

I was an honored guest and minor judging participant at Australia’s 1991version of a “Sieger Show,” the only foreigner to have been so honored up to that date. It is called the “Main Breed Assessment” rather than a “show,” to avoid problems with the quasi-governmental Australian National Kennel Council over practices allowed at regular shows, such as pedigrees and catalogs in the judges’ hands, gun sureness testing, and especially information on what problems and good features the dog being examined has passed on to its pups. I was very impressed that, in coming to the placement decisions, the judges of the adult classes took into account such things as the Australian ‘A’ stamp hip status (they capitalize the letter there) of the individual as well as of siblings and offspring, and other genetic factors as well as a full and expert evaluation of the dog in question. The GSD people in Australia modified the BVA scheme in conjunction with their own system. But I think they give the ‘A’ stamp to too many animals for fast enough progress. The 6 grades are: N, NN, A, BL, III, and IV. Dogs are considered eligible for the ‘A’ stamp if they have one of the four top grades of the six, and this includes A (acceptable) and Borderline (many of which have what OFA would call mild to moderate HD). As in Germany, this allows too many to breed, and tends to act as a brake on progress. However, they have what we in North America don’t have, to any appreciable amount: progeny data. This tends to offset part of the failings of less-strict radiograph requirements, at least when comparing those schemes to OFA’s. According to an issue of the Australian GSD club’s newsletter, almost all of the Normals and 61.4% of the Near-Normals score 0-5. While GSD hip quality has not increased as dramatically as quality of breed type, there are hip requirements for breeding and, in time, increased strictures will produce faster improvement. By limiting breedings to animals with the ‘A’ stamp, the Aussies and New Zealanders would exclude about a third of the breed, better than what was done in England, but far inferior to Sweden and what had been required in East Germany. Since 1981, the percentage of Australian GSDs receiving the ‘A’ stamp has risen from 60% to 80%, while grades III and IV (roughly equivalent to moderate and severe HD in the United States or the BVA scores of 0-10) have declined by half.

Judging at the Australian “Sieger Show”

Japan and Pacific Rim

In the modern, dog-loving portion of Japan’s society, control of hip dysplasia is progressing. I have judged and lectured there, and long ago found great interest in improving many areas. The Japan Kennel Club adopted PennHIP as the official and preferred HD diagnostic procedure in the late 1990s. In Taiwan, dog shows and interest in improved breeding—including for better hips—are on the increase. When I lectured in Malaysia and the Philippines, I found the progress and awareness at a lower level, but at least they know enough to ask about hip status when they import dogs for their breeding programs.

Comparing America to the World

In the Americas, the oldest hip registry is the OFA, but there is a better one: PennHIP. It’s a good thing that OFA requires a minimum age of 24 months for certification of “normalcy”; otherwise the situation in most breeds in the USA would be dismally poorer. In most breeds it is not that great, anyway, if you look at over-all breed statistics instead of individual breeders’ accomplishments. Paradoxically, the greatest rates of progress are in some of those countries where dogs are radiographed and certified for breeding as soon as they pass their first year’s birthdate, although they would be even better if approval were to be delayed at least 6 more months. The reason for this paradoxically better situation, though, is that many breed clubs outside America control authorization for breeding and registering, denying such privileges to dogs with the worst hips. In America, while the AKC gleefully registers anything that comes with money and the specified paperwork.

Compare progress in the U.S. with that in Germany, for example, and specifically the most popular breed there and in the world, the German Shepherd Dog. There has been a shift toward normalcy that came about in spite of the practice of forbidding breeding rights only to those with severe HD. As time went on, requirements for the VA (excellent-select) class at the world Sieger Show in Germany were tightened more and more. Not only must current highly placing show dogs have advanced training degrees, they must also have the better hips and produce a good number of normal hips as well as structurally desirable progeny. Today a dog with a Noch Zugelassen (“still permissible”) rating might make it into the VA class of some eight or ten dogs out of hundreds of competitors, but he or she will not win the top title of Sieger or Siegerin, and there is now pressure to keep the bad producers (with high ZW numbers) from being honored with the Sieger title. There is annually increasing emphasis that the very top be Normal, not just Fast (nearly) Normal. So the dogs that get the most breedings in most of Europe will generally have the best hips. There is no similar restriction in the GSDCA’s sizeable Select class for GSDs at either the American (U.S.) national specialty show nor at the smaller but similar “Canadian National”. Nor is there anything similar in the other AKC- or CKC-affiliated breed clubs. In America we have neither the strict rules nor the peer pressure nor strong suggestions to judges. We judges certainly aren’t allowed to officially “know” the hip status or other information important to the breed when we are in the AKC/CKC/KC/UKC rings, though we either have been given the documents or can ask, when we judge SV type shows. In the latter, knowledge is more important to the breed than feigned ignorance.

Even faster progress could have been made by the SV if they would award the ‘a’ stamp only to dogs radiographed after 18 or 24 months of age. And in other countries, we could see an increase in the progress rates if all dogs were to be radiographed and evaluated, even if they had poor hips and would never be bred. It would give valuable data for progeny testing.

Improve Your Breed by Improved Breeding

I recommend that breeders use this triad: evaluating mature dogs for DJD, using PennHIP for early risk detection, and following a Breed Value/Zuchtwert program. If, as is certainly indicated, the DI gives a better picture of future hip quality in your dog, then deductive reasoning would lead you to think of it as a reasonably accurate indicator of the genotype of your dog. That means a better idea of the proportion of bad hip genes to good hip genes, which in turn means relatively how many bad genes are likely to be transmitted to the next generation Now that, dear friends, is really revolutionary. The lack of further progress we have seen in modern times, with ratings by BVA, OFA, SV, ADRK, OVC, and other breed and veterinary organizations is a direct result of their inability to indicate those hidden genes via the limits of the old technique. A dog that has a good picture in the extended-leg view yet still produces an unacceptably high number of dysplastic offspring has too many of those hidden genes. Since OFA would be the first to tell you of the link between laxity and HD (remember, they actually use that as a definition), the only reason for the poor progress is the covert laxity I mentioned earlier. Therefore, using a logical process of thought, if PennHIP shows more of this laxity than shows up in the AVMA-type view, it better shows us the effects of more “hip genes.” Since mapping the dog’s genome (at least finding markers for enough of the polygenic perpetrators) is many decades away, the DI evaluation as promoted by PennHIP is by far the best tool in our tool chest.

Fred Lanting

Fred Lanting is an internationally respected show judge, approved by many registries as an all-breed judge, has judged numerous countries’ Sieger Shows and Landesgruppen events, and has many years experience as one of only two SV breed judges in the US. He presents seminars and consults worldwide on such topics as Gait-&-Structure, HD and Other Orthopedic Disorders, and The GSD. He conducts annual non-profit sightseeing tours of Europe, centered on the Sieger Show (biggest breed show in the world) and BSP.

Books by Fred Lanting