The gonad chronicles part 1: Neutering's newest controversy

The gonad chronicles part 1: Neutering's newest controversy

Veterinarians and owners alike are troubled by a new study linking castration and ovariohysterectomy with an increased risk of disease—including some cancers. Should this study change the way we discuss and perform such common, recommended procedures?
Dec 01, 2013

Next >

This is the first article of a two-part dvm360 series focusing on neutering and the effect of hormones on the development of both physical and behavioral concerns (see the Related Links section below for a link to part 2). The following article will explore new research focusing on joint disorders and cancers in both neutered and intact golden retrievers.

All fields, even those governed by scientific principles, are affected by cultural shifts and trends. We see these effects daily in how pet dogs are treated. Dog parks and doggie daycare are modern inventions reflecting both our culture (no one is home for much of the day) and social trends (dog parks offer welcome opportunities for humans to socialize as well). These examples reflect another deeper change in our culture: that dogs are viewed as family members. Many owners believe their pets have social, cognitive and behavioral needs that require special attention.

So when a research team from the University of California-Davis (UCD) published a paper earlier this year titled, “Neutering dogs: effects on joint disorders and cancers in golden retrievers,”1 it’s no surprise that both veterinarians and dog owners felt unmoored from one of their cultural tenets—that we should neuter dogs for population control and for their health. Listservs and specialists have been inundated with questions about whether we should stop neutering dogs on the basis of this new information. To answer such questions, let’s step back, take a deep breath, critically review the data and use it for risk analysis.

< Back  |  Next >

Designing the study

Hormones’ effect on the development, progression and treatment outcomes for cancer has been a focus of intense research for decades, but very few populations of any species routinely castrate or perform ovariohysterectomy (OHE) on healthy animals. The exceptional population is dogs in the United States. Most dogs in this country are neutered, in contrast to Scandinavia, where most dogs are intact. Early prepubertal neutering has become more common here over the past 25 years, largely to control the homeless dog population. Additionally, routine veterinary recommendations for dog owners include neutering the pet if there is no intent to breed. These two factors together suggest that American dogs offer a good population for anyone interested in studying the time penetrance of hormonal factors on behavioral and physical health. Accordingly, the authors of the UCD PLOS One paper sought to investigate whether neutering (OHE and castration) created any real increased risk of cancer or other conditions—or offered protection from them—as had been proposed in a series of papers published mostly within the last two decades.

The historical data on hormones and behavioral and physical development suggest that the effects are complex and not always independent of social and physical environments. Hormonal effects on many cancers are also complex. Because hormones affect multiple organ systems, neutering represents an experimental manipulation of all of those systems. The PLOS One study included both male and female dogs and compared those that were intact and those that were neutered. The neutered population was further split into dogs neutered early (/= 12 months).

The authors sought to retrospectively examine the potential effects of neutering on a series of conditions within one breed (in an attempt to control for variation due to breed). They chose a common breed (golden retriever) that has a relatively high incidence of neoplasia. Golden retriever breed clubs also have an active interest in breed health. Because the study was retrospective, the authors sought to further control for differences in populations and diagnosis by using a single hospital database. Uncontrolled variation is a serious problem in epidemiological and risk studies because such variation interferes with interpretation of both statistical and biological effects.

The authors examined seven conditions that breeders identified as concerns for golden retrievers: hip dysplasia (HD), cranial cruciate ligament tear (CCL), lymphosarcoma (LSA), hemangiosarcoma (HAS), mast cell tumors (MCT), osteosarcoma (OSA) and elbow dysplasia. These conditions have been the subject of other papers—not always with the same conclusions—examining effects of neutering and hormones (see the references in the PLOS One paper). All of these conditions would be assessed in the context of the three treatment groups: intact, early neutered and late neutered retrievers.

< Back  |  Next >

Assembling the study group

All of the study dogs came from the veterinary teaching hospital database at UCD. Intact and neutered male and female golden retrievers were required to be between 1 and 8 years old, admitted between Jan. 1, 2000 and Dec. 31, 2009 and with the following data in their records: date of birth, date of neutering and date of diagnosis or onset. The authors also collected a body condition score when available. Researchers only included diseases that were present in a minimum of 15 cases (1.5 per year of the study). Furthermore, there had to be evidence linking clinical signs to the disease diagnosis or the dog was not included in the study. Once all of these criteria were met, the sample size for each group was determined (Table 1).

Dogs less than 1 year of age and those 9 years of age and older were not considered for the study, which means that we don’t know the outcomes for dogs who were disease-free before the age of 9, but not after. We also have incomplete follow-up for those flagged as affected once they aged out of the study, a finding that limits conclusions about survivorship and potential quality of life. In addition, age of onset may not have been noted or known. Sometimes this information was provided by the referring veterinarian or the owner and not a UCD pathologist.

An attempt was made to ensure that age of neutering preceded age of onset for all dogs included in the study, but because of the insidious ways in which many cancers develop and because of the highly variable times of neutering, this goal may not have been met. The authors are completely aware of these limitations and excluded some patients for one disease, but included them for another. The authors further note, “A case could be considered as intact for one disease if onset was prior to neutering and considered as late-neutered for another disease that may have occurred after neutering.”

For all of these reasons, collecting prospective data is preferred for establishing many risks and most mechanisms, but by definition, such studies take a long time. Retrospective studies, on the other hand, are often used to identify associations that may be further investigated in a prospective study, and the PLOS One paper is no exception (for more information, check out the Morris Animal Foundation’s project on golden retrievers at With a retrospective study, you hope that your sample size is large enough that, within the bounds of your specified conditions, you can still say something interesting and potentially helpful.

< Back  |  Next >

Reviewing the results of the study

Early versus late neutered dogs: When comparing early neutered (/= 12 months) males, there was a statistically significant difference in the number of dogs with CCL and LSA. A greater percentage of early neutered males were diagnosed with these conditions. Two statistical tests with slightly different assumptions were used for each analysis. Only one of these tests produced a statistically significant effect with respect to HD and more early neutered males were diagnosed with this condition. For females, there was a statistically significant difference only for those affected with CCL, reflecting a greater percentage among early neutered dogs.

Early neutered versus intact dogs: When comparing early neutered with intact male dogs, there was a statistically significant difference in the number of dogs diagnosed with HD, CCL and LSA. A greater percentage of neutered males was diagnosed with these conditions. For females, there was a statistically significant difference only for CCL, with more early neutered dogs being diagnosed.

Late neutered versus intact dogs: There was no statistically significant effect for any of these conditions when late neutered dogs were compared with intact dogs.

Using this data, we can now identify specific patterns of concern by comparing the numbers in each group. However, we cannot establish “risk” in a true statistical sense because of the data’s limitations. (The relevant data can be found in Tables 2 and 3.)

So what can we say? In this sample of golden retrievers, between the ages of 1 and 8, inclusive, and admitted to the veterinary teaching hospital at UCD over a decade of study, the number of male dogs with HD who were early neutered (N=16) was twice that of dogs who were intact (N=17). In early neutered dogs of both sexes, the early neutered dogs (N=9 for males and N=13 for females) were more likely to have CCL than either intact (N=0 for males and N=0 for females) or late neutered dogs (N=1 for males and N=0 for females). Early neutered males (N=17) had three times the occurrence of LSA compared with intact males (N=5), for the 22 total male dogs diagnosed over the decade of the study.

< Back  |  Start >

Putting the research into practice

The authors note that they cannot extend their results to other breeds, or even to dogs, in general. Imagine if all of the dogs that had no disease at the age of 8 developed it at the age of 10 and dropped dead? The analysis would look vastly different. We know cancer is a time-penetrant condition, and early cancers, especially those of pre-reproductive age, are rare in any species.

Any causal attribution for neutering and neoplasia and other non-behavioral pathologies is also premature. But we may be able to use breed to identify when there may be an increased genetic risk—and hence a putative mechanism—for any type of disease. The PLOS One study did not consider that some of the dogs in their study could be members of overlapping family lines and hence not independent cases. The study also did not address another type of breed canalization—that associated with work. Dogs that are bred for conformation versus work, whether the work is as a military working dog, a herding dog or a field trial dog, are different both behaviorally and genetically. In turn, these groups may produce different outcomes for such studies. Given these state-of-the-art data, we can conclude that there is no one recommendation that will fit each client and each dog.

If the dog is a shelter or rescue animal, neuter the dog immediately if it’s healthy enough to undergo the procedure. The risk here—especially in a country like the United States where unwanted, neglected dogs number in the millions—is that we must decrease the population entering shelters. And, although puppies are considered more adoptable, we must decrease the number of pups coming in from the street or being born into shelters. Too many of these young animals may have epigenetic effects from poor maternal and grandmaternal experiences, and too many of them have poor early or in utero nutrition.2

This is a formula for creating pathological reactivity, and I see a nonstop stream of these dogs in my patient population, in shelters and rescues, and through those who e-mail me for help. The Scandinavians do not seem to have the same unwanted dog population that we do, so they make different decisions. But we must intervene in this tragic situation.

Note: In the second part of this series, we will review data outlining neutering and hormonal effects on behavior and look at some more information about hormones and neoplasia.

Dr. Karen L. Overall is a researcher, editor of The Journal of Veterinary Behavior: Clinical Applications and Research, and author of more than 100 publications, dozens of chapters and a new book, The Manual of Clinical Behavioral Medicine for Dogs and Cats.


1. De la Riva GT, Hart BL, Farver TB, Oberbauer AM, Messam LL, Willits N, Hart LA. Neutering dogs: effects on joint disorders and cancers in golden retrievers. PLOS One 2013;8(2):e55937.

2. Overall KL. Editorial: Caring for the brains of young pups. Vet Rec 2011;169(18):465-466.