The relationship between the length of gonadal exposure and bladder cancer incidence in Scottish terriers: Risk analysis, modification of risk, and sex-specific differences

Only in recent years have studies emerged that highlight the medical consequences of spay or neutering an animal. However, the method used to categorize dogs in a majority of studies impedes the progress to discovery: veterinary literature documents the neuter status of animals as neutered or intact to study an outcome. In the context of disease outcome, this method assumes that all neutered dogs are created equal. In contrast, a small number of veterinary literature report that the length of exposure to hormones determines the outcome for a disease, thus, not all neutered dogs are created equal. A previous study showed that categorizing dogs in this fashion, called dichotomous binning (DB), gives way to misleading assumptions about the consequences of spay or neuter. It is therefore likely that most studies err by implying there is no difference between a dog neutered at four years and neutered at fourteen years when in fact a difference actually exists. In 2009, our laboratory reported that female Rottweiler dogs, which kept their ovaries longer, lived longer. By this merit, we hypothesized that in our data set of Scottish Terriers, 1) keeping gonads longer is associated with lower risk for bladder cancer (TCC). The first objective of this research was to investigate whether there was a window of time of gonadal exposure that was protective against bladder cancer. The second objective was to determine if there were sex-specific differences in the relationship between length of gonad exposure and bladder cancer risk. To test this hypothesis, we grouped dogs into three based on their length of gonadal exposure (LGE). Using logistic regression analyses in SPSS Statistics Data Editor, we studied associations between length of gonadal exposure (LGE) and TCC. We performed the same analysis using the same dogs but categorized by DB, and compared findings derived from using DB and LGE classification methods. After adjusting for covariates, the DB model yielded no significant difference in TCC risk between neutered and intact dogs, while the LGE model revealed that gonadal exposure may be a risk factor for TCC: LGE dogs with the shortest length of gonadal exposure had a borderline significant increased risk of TCC (OR, 3.39; 95% CI, 0.92 to 12.49; P = 0.07) compared to those with the longest gonadal exposure. The LGE model further revealed potential sex-specific differences, where an increased risk in dogs with the shortest length of gonadal exposure occurred among males but not in females. The LGE model also showed a protective gonadal exposure window of 1 year especially in male dogs, where neutering before then elevated the risk for TCC 4.5 times that of dogs with the longest exposure. The DB method could not detect this. Overall, the LGE model appears to be the favorable and more clinically applicable model of grouping dogs than the DB model. A previous study showed that Scotties exposed to lawns or gardens treated with herbicides had increased risk of TCC compared to those exposed to non-treated lawns or gardens, and those fed any vegetables at least 3 times/wk had decreased risk compared to those fed vegetables less than 3 times/wk. Literature suggests that factors like circulating hormones in pre- and post-menopausal women, and sex influence the relationship between vegetable consumption and cancer risk. Given these reports, we hypothesized that 2) gonadal exposure would modify the relationship between vegetable consumption and TCC and 3) gonadal exposure would modify the relationship between herbicide exposure and TCC. One objective was to investigate if modification effects of gonadal exposure on vegetable-TCC or herbicide-TCC relationship differed between the sexes. We performed logistic regression analyses separately for both hypotheses. The findings suggested that in dogs categorized by LGE increasing gonadal exposure regardless of the frequency of vegetables consumed per week resulted in a 52% reduction in risk. Increasing the length of gonadal exposure results in a larger reduction in risk among male dogs (75%; P = 0.06) than among female dogs (31%; P = 0.50) indicating that there may be sex-specific differences in modification effects of LGE. Among dogs categorized as DB, there were not any clear or reliable modification effects. There were also no clear modification effects of gonadal exposure on the relationship between herbicide exposure and TCC risk in both DB and LGE dogs. The vegetable study highlights the benefits of using the length of gonadal exposure over dichotomous binning in determining risk for TCC and risk modification of factors influencing TCC incidence. Additionally, the likelihood test showed that statistically, categorizing dogs using the LGE method significantly improved the model fit compared to the DB method. (Abstract shortened by UMI.