Embedded Files
language (SQL) database. Data-fields collected included demographic (species, breed, date of birth, sex, neuter status, insurance status, weight and deceased status) and clinical information (free-form text clinical notes, summary diagnosis terms, treatment and deceased status with relevant dates). Sample size calculation estimated that a cross-sectional study design with 314 purebred cats and 2,826 crossbred cats would have 80% power to detect a one-year longevity difference between purebred and crossbred cats (at a = 0.05), assuming that overall longevity was normally distributed with standard deviation of 6 years and that 10% of cats were purebreds (24). Ethical approval of the project was granted by the RVC Ethics and Welfare Committee (URN 2010 1076). A random sample of potential deaths was selected from the ‘deceased status' field for detailed study (25). True death cases were identified from the clinical notes and the cause of mortality was extracted from the clinical note and VeNom diagnosis data relating to the death event. If a cause of mortality was not explicitly stated at the time of death, then no defined cause of mortality was included. The causes of mortality were grouped into appropriate pathophysiologic or organ-system categories. The mechanism of death (assisted [euthanasia] or non-assisted) (26) and mode of body disposal was noted. The age at death relied on the date of birth values recorded in the PMS; no age at 7 death was included for cats without birth date information. Cats with recognized breed names were grouped as ‘purebred’ while cats described as mixed-breed, breed-specified crosses or domestic cats were grouped as ‘crossbred’ (11). The neuter and insurance statuses at death were used. Neuter and sex status were combined to create a sex/neuter variable with four categories: female entire, female neutered, male entire and male neutered. The maximum bodyweights recorded after 6 months of age were categorised into six groups (0.0-2.9 kg, 3.0-3.9 kg, 4.0-4.9 kg, 5.0-5.9 kg, 6.0 kg and above, no weight recorded). Following data checking and cleaning in Excel (Microsoft Office Excel 2007, Microsoft Corp.), statistical analyses used Stata Version 11.2 (Stata Corporation). Overall and breed-specific (for breeds with 10 or more cats in the study) longevities were reported using median, interquartile range (IQR) and range. Median overall longevity values for purebred and crossbred cats were compared using the Mann-Whitney U test. The proportion of cats that were euthanased was compared between purebred and crossbred cats using the chi-squared test. Causes of mortality were separately tabulated for cats of all ages, cats dying before 5 years of age and cats dying at or after 5 years. General linear regression modelling was used to evaluate associations between risk factors (purebred/crossbred, sex/neuter, weight category, microchip and insured) and longevity in cats dying at or after 5 years of age. The 5 year cut-off was chosen because the longevity data after this age approximated the normal distribution that is a required 8 assumption for linear regression modelling (27). Univariable risk factors liberally associated with longevity (P < 0.2) were evaluated using multivariable models developed using backwards stepwise elimination. The final model was evaluated using the clinic attended as a random effect and for pairwise interaction effects (28). The predictivity of the final non-random effect model was evaluated using the adjusted r 2 value. Model diagnostics included visual inspection of residual and residual-versusfitted plots to assess normality and homoscedasticity, respectively (28). Statistical significance was set at P < 0.05. Results From 118,016 cats attending 90 practices in central and south-east England with 12,012 potential death cases, a study sample of 4,009 cats with confirmed deaths attending 87 practices was randomly selected. Data completeness varied between the variables: sex 98.9%, neutered 100.0%, breed 99.9%, date of birth 99.3%, insured 100.0%, microchipped 100.0%, weight 45.4%, cause of mortality 82.3%, mode of death 95.1% and mode of body disposal 84.2%. Demographic characterisation of deceased cats with information available indicated that 3,660 (91.7%) were crossbred, 2,009 (50.7%) were female, 2,599 (64.8%) were neutered, 543 (13.5%) were insured and 238 (5.9%) were microchipped. Adult bodyweights were distributed as follows: <3.0 kg (440 cats, 24.2%), 3.0-3.9 kg (612, 33.6%), 4.0-4.9 kg, (461, 25.3%), 5.0-5.9 kg (208, 11.4%), ≥ 6.0 kg, (98, 5.4%). The median (IQR) bodyweight was 3.7 (3.0-4.6) kg. Euthanasia 9 accounted for 3,265 (85.7%) deaths overall with a greater proportion of crossbred cats (3,006, 86.0%) being euthanased compared with purebred cats (254, 81.9%) (P = 0.050). Overall, 2,522 (74.7%) cats were cremated. Median overall longevity was 14.0 years (IQR 9.0-17.0; range 0.0-26.7) and was bimodally distributed, peaking at years 1 and 16 (Fig. 1). The median longevity of crossbred cats (14.0 years, IQR 9.1-17.0; range 0.0-26.7) was greater than purebred cats (12.5 years; IQR 6.1-16.4; range 0.0-22.0) (P >< 0.001). Female cats (15.0 years, IQR 11.0-17.4) had a higher median longevity than male cats (13.0 years, IQR 7.6-16.0) (P < 0.001). Neutered cats (15.0 years, IQR 11.8-17.0) had a higher median longevity than non-neutered cats (11.0 years, IQR 2.13-16.0) (P < 0.001). The longest-lived breeds were the Birman (n = 12; median 16.1 years; IQR 8.1-16.9) and Burmese (n = 31; 14.3 years; IQR 10.0-17.0). The shortest-lived breeds were the Bengal (n = 15; 7.3 years; IQR 2.2-11.5) and Abyssinian (n = 11; 10.0 years; IQR 1.1-18.1) (Table 1). A cause of mortality was specified for 3,309 (82.5%) cats. In cats of all ages with a cause of mortality specified, the most frequently attributed causes were trauma (n = 405; 12.2%), renal disorder (n = 399; 12.1%), non-specific illness (n = 370; 11.2%), neoplasia (n = 356; 10.8%) and mass lesion disorders (n = 336; 10.2%) (Table 2). Mass lesions described mass-associated disorders that did not have a more precise
Google Sites
Report abuse