Friends with malefit. The effects of keeping dogs and cats, sustaining animal-related injuries and Toxoplasma infection on health and quality of life

Many studies show that keeping cats and dogs has a positive impact on humans’ physical and mental health and quality of life. The existence of this “pet phenomenon” is now widely discussed because other studies performed recently have demonstrated a negative impact of owning pets or no impact at all. The main problem of many studies was the autoselection–participants were informed about the aims of the study during recruitment and later likely described their health and wellbeing according to their personal beliefs and wishes, not according to their real status. To avoid this source of bias, we did not mention pets during participant recruitment and hid the pet-related questions among many hundreds of questions in an 80-minute Internet questionnaire. Results of our explorative study performed on a sample of 10,858 subjects showed that liking dogs has a weak positive association with quality of life. However, keeping pets, especially cats, and even more being injured by pets, were strongly negatively associated with many facets of quality of life. Our data also confirmed that infection by the cat parasite Toxoplasma had a very strong negative effect on quality of life, especially on mental health. However, the infection was not responsible for the observed negative effects of keeping pets, as these effects were much stronger in 1,527 Toxoplasma-free subjects than in the whole population. Any cross-sectional study cannot discriminate between a cause and an effect. However, because of the large and still growing popularity of keeping pets, the existence and nature of the reverse pet phenomenon deserve the outmost attention.


Introduction
population, keeping pets, and especially dogs, was shown to have positive effects on the subjects' 48 community connectedness, and on their mental health and wellbeing [7][8][9]. It was also shown that 49 keeping dogs, but often not cats [10], increased the chance for survival after heart diseases including 50 myocardial infarction [10] and even decreased the total mortality rate in various populations [11][12][13][14].

51
Simply engaging with animals, e.g. stroking animals or just thinking about pets, was shown to reduce 52 blood pressure and stress [15] and to provide relief from social rejection [16]. Pet owners were shown 53 to spend a smaller fraction of their income on healthcare than non-owners [17].

54
However, in the past twenty years, more and more studies were performed that failed to 55 reproduce the results of these older studies. They either observed no effects of keeping pets on human 56 wellbeing and health [18], or actually observed negative effects of keeping dogs, and even more of 57 keeping cats [19], on health, survival of patients and members of general population and their 58 wellbeing [13,[20][21][22][23][24][25][26][27][28][29][30]. The question of the existence of positive or negative effects of keeping pets is 59 still open; however, more and more authors have reached the conclusion that many of the older results 60 may be strongly biased by autoselection of participants of studies and selective reporting of only 61 positive results, or a priori expected results of studies [17]. Also, even many highly cited studies were 62 performed on rather small or imbalanced samples of participants and most of them reported only the 63 effects of dogs, despite the fact that data on the effects of cats were probably also studied [8, 9, 31, 64 32]. It is sometimes stated that no differences in economic situation exists between the pet keepers and nontrivial problem of earlier studies was that the participants in the studies were informed in advance 75 that the effects of pets (or dogs and cats, or animals) on quality of life will be the subject of the study.

76
It is highly probable that people who keep pets also love them and believe that pets make their life 77 better (otherwise they would get rid of them) and therefore answer the questions in agreement with 78 their a priori beliefs, regardless of the real effects of pets on their lives. This is in agreement with a 79 common observation that, when directly asked, people often report a positive effect of their pets on 80 wellbeing while the results of detailed questionnaires suggest an absence of such effects [30].

81
To avoid this problem we ran a large Internet study that was advertised without mentioning The internet questionnaire was distributed as a Qualtrics survey. Subjects were invited to 97 participate in the study using a Facebook-based snowball method [35]. Potential volunteers, mostly 98 members of the "Lab bunnies" community, an 18,000-member group of Czech and Slovak nationals 99 willing to take part in evolutionary psychology experiments, and their Facebook friends, were invited The responders were asked about their sex, age, education (ordinal scale 1-8: 1-elementary, 8-PhD or 127 MD), body height, body weight, and the size of the communities where they currently live (ordinal 128 variable urbanization: 0-less than 1000 inhabitants, 1-1-5 thousand inhabitants, 2: 5-50 thousand 129 inhabitants, 3: 50-100 thousand inhabitants, 4: 100-500 thousand inhabitants, 5: more than 500 130 thousand inhabitants). The anamnestic part of the questionnaire contained several questions about the 131 intensity and nature of the person´s contact with dogs and cats and about sustained animal-related 132 injuries. Subjects were asked to rate how much they like dogs (cats) using the scale 0-100 (the 133 preference of dogs to cats was calculated as the difference between liking dogs and liking cats), and 134 then rate the intensity of their past and current contacts with dogs and cats using a 8-points pseudo- i.e., how many cigarettes they smoke a day (ordinal scale: 0-0, 1-0-0.1, 2-0.1-1, 3-1.1-3, 4-3.1-10, 5-11-20, 6-21-40, 7-more than 40), frequency of consuming alcohol "not to be allowed to drive a car 155 for a while for this reason" (ordinal scale: 0-never, 1-maximally 1× a months, 2-maximally 2× a 156 months, 3-maximally 4× a months, 4-maximally 2× a week, 5-every second day, 6-every day, 7-157 nearly all the time), and frequency of consuming illegal drugs "not to be allowed to drive a car for a 158 while for this reason" (ordinal scale: 0-never, 1-maximally 1× a months, 2-maximally 2× a months,

161
In another part of the questionnaire we collected information concerning the following undiagnosed disorders of the responder's partner). Using a 0-100 scale they also answered the questions -How intensely they are sexually attracted to men, How intensely they are sexually 183 attracted to women, (Z-score of the higher of these two responses was considered as the intensity of 184 sexual desire). They were also asked with how many men (women) they had sex ("vaginal, oral or 185 anal"). To respond to these two and the next two questions, the participants used a 0-9 ordinal scale (0-186 0, 1-1, 2-2, 3-3, 4-4, 5-5-6, 6-7-9, 7-10-19, 8-20 and more Tables 2-3). Therefore, we used 298 nonparametric partial Kendall regression tests controlled for age, size of place of living, and education 299 to search for associations between the animal-related variables and wellbeing and health, separately 300 for male (Table 4) and female (Table 5)

327
In men, keeping a dog had small, mostly non-significant effects on mental and physical health;   (77), not all correlations, however, were statistically significant suggesting that keeping dogs and cats is not responsible for most of the observed associations between 428 liking animals and wellbeing-and health-related variables. Similarly, the negative associations between keeping animals and mental health could theoretically just be a secondary effect of much 430 stronger negative association between sustaining animal-related injuries and mental and physical 431 health problems. Again, the pattern of association was very similar for the subjects who reported not to 432 have been injured by cats and dogs (Supplementary tables S7-S10).

433
It is known, and also our current data strongly suggests, that latent infection with the parasite associations observed on these subsets were much stronger than the same associations observed in the 446 whole set (Tables 6, 7, S11, S12). This suggests that toxoplasmosis was not responsible for the 447 observed associations between the animal-related variables and human wellbeing, health, and fitness.   subjects. Regardless of these obvious methodological problems, even the data from observational 482 studies could render some models/hypotheses more and some less feasible. For example, our data 483 strongly contradict the model that transmission of Toxoplasma by a cat-related injury rather than the 484 cat-related injuries themselves could be responsible for the negative association between cat-related 485 injuries and mental health because this negative association was also detected in subjects who were admit mental illness. The latter explanation is, however, contradicted by the fact that number of cats in 502 the home correlated more strongly (in men in fact only) with the number of partner's mental health disorders than with the number of the responder's mental health disorders. It is indicative that this is 504 not true for the cat-related injuries, which, in contrast to the number of cats in the home, affect only 505 the responder and not his/her partner. Here, the association of injuries with the number of mental 506 health disorders was much stronger than with the number of partner's mental health disorders; in 507 women, the latter association did not exist at all (Table S2).

508
The analyses performed on subjects who never kept a dog showed that, in men, a positive 509 association between liking dogs and mental health and wellbeing (and negative with number of 510 children, see below) could be the result of keeping a dog by people who like dogs. This conclusion, 511 however, is only preliminary, as the positive association between liking dogs and wellbeing and with 512 mental health still existed in this subpopulation but was much weaker than in the whole population 513 (wellbeing Tau=0.011 vs 0.063; mental health Tau= -0.001 vs -0.026). On the contrary, the negative 514 association of liking dogs with physical health, and also the negative association with number of 515 children in women, also occurred in subjects who never kept a dog. Here, the situation was opposite 516 than it was in men as these two associations were even stronger in women who never kept a dog than

524
Different patterns showed in the analyses performed on subjects who never kept a cat. In men, 525 a negative association between liking cats and mental health also existed in this subpopulation and was 526 only slightly weaker than that observed in the whole population (Tau=0.039 vs 0.044). Unexpectedly, 527 a positive association between liking cats and physical health was observed in this subpopulation

528
(Physical health problems score Tau=-0.039), which was reflected by an even stronger positive 529 association between preferring dogs to cats and physical health problems (Tau=0.054). These two associations were not detected in the whole population (liking cats Tau=-0.004; preferring dogs to cats 531 Tau=0.008). In women, the strength of the association between liking cats and mental health problems

542
All analyses were also performed on subpopulations of people who had never been injured by

547
Again, a slightly different pattern was revealed in the analyses of subjects who had never been injured 548 by a cat. In men, the negative association between keeping a cat and mental health was absent in the 549 cat injury-free subpopulation (mental health problems score: Tau=0

570
The causality relation between the Toxoplasma infection and impaired mental health has been 571 confirmed by prospective cohort study [47,48] and experimental infections of laboratory animals.

572
Such confirmation of the direction of causality between animal-related injuries and mental health is 573 not currently available. However, according to three of nine Bradford Hill criteria of causation [49] 574 (Specificity, Biological gradient, and Analogy), the animal-related injuries seems to be more probably 575 the cause then the effect of impaired mental health: (1) As was already published, the spectra of dog mental health [45,51,52]. In the present study we showed that Toxoplasma was not the cause of the 583 observed associations between the injuries and impaired health of the responders [45]. However, the participants were not tested for, e.g., Bartonella henselae, the agent of cat-scratch disease. This disease 585 probably has many effects on the physical and mental health of patients and on the incidence of some 586 mental health disorders, most strongly on the risk of major depression [53][54][55][56]. Many other, both 587 known and unknown, pathogens could be transmitted from animals to humans either by biting and 588 scratching, or by other forms of close contact. Of course, the infection-based explanation of the 589 association of contact with cats and dogs with mental health problems must be considered just as a 590 working hypothesis that should be tested in future studies.

591
Liking dogs and cats was strongly negatively associated with number of children. Most 592 probably, having (and liking) one's own children has a negative effect on the relative intensity of 593 liking pets. An alternative and more alarming explanation, i.e. the existence of a negative effect of 594 liking pets on, e.g., parental instincts, should be tested in future cohort studies by searching for an 595 association between liking pets and intensity of desire to have children in childless people. Also, we 596 plan to repeat our study after five years to see whether young childless women, registered members of 597 the Lab Bunnies community, who originally reported the more intensive love for cats would report 598 fewer children in the follow-up study than those who originally reported less intensive love for cats.

599
The number of children was also lower in subjects who presently keep animals and even more 600 so in subjects who were more seriously injured by animals. Here, the decrease of biological fitness 601 could be the result of impaired physical and mental health of affected subjects. Despite the negative 602 effects on the number of children, animal-keepers, and even more so subjects who were more seriously 603 injured by dogs or cats, were more sexually active. This could be explained by the adoption of a

604
"quick life strategy" by people in bad mental and physical health conditions. The slow life strategy,

605
i.e., the preferential investment in quality of offspring, to carefully search for the best available sexual 606 partners and to accumulation of resources, including education, before starting the process of 607 reproduction, pays more to healthy subjects. For the subjects with lower life expectancy it pays to 608 adopt the quick strategy, to have a sex soon and, especially in men, with as many partners as possible.

609
In the affected (animal-keeping or animal-injured) men, high sexual activity was accompanied by high 610 sexual desire. In contrast, in women, the affected subjects had lower sexual desire regardless of their higher sexual activity. This suggests that psychological mechanisms responsible for the transition from 612 a slow to fast life strategy probably differ in men and women. The slow and fast life strategies were 613 outside of the scope of present study; however, the existence of a highly significant positive 614 correlation between both physical and mental health problems and sexual activity was detected in 615 women, regardless of their contacts with dogs and cats.

616
A major limitation of the present study is its observational character. increase the risk of failure to detect existing effects but not the risk of detecting non-existing effects.

647
We asked the subjects whether a dog (a cat) had ever been kept in their family. However, most 648 subjects had already two families, the family to which they were born and their own new family. It is 649 possible that keeping pets in original and new family are associated differently with the subjects' 650 wellbeing and health. Any future study should discriminate between these two kinds of families.

651
It is widely believed that keeping cats and dogs has a positive impact on the subject's quality participants. The studies were advertised as studies of the effect of pets on human life. It is rather probable that people who like pets, keep them and believe that their pets positively influence many 666 aspects of their lives preferentially participate in such studies. People who like pets and believe that 667 pets positively influence the quality of life of pet-keepers are also present in controls, however, here 668 their frequency is lower than in the pet-keepers. This flaw occurs even in studies that explicitly claim 669 the opposite. For example, Bennett et al. [8] wrote "We were careful not to indicate that we were 670 particularly interested in pet owners," however, the study had been advertised as "being concerned 671 with investigating the daily activities, emotions and perceptions of older adults living independently, 672 either alone or with other persons or pets, in the community." To eliminate this source of bias, our 673 study was advertised as a study on "mystical thinking, superstitions, prejudices, religion and relation 674 between various environmental factors and health and wellbeing." The pet-related questions were 675 buried in many hundreds of other questions of the more than 80-minute questionnaire. It is probable 676 that the relation to pets had only a small impact, if any, on how participants responded to our questions 677 on quality of life and mental and physical health. Therefore, our study probably more objectively 678 reflects the situation in general nonclinical population than most previously published studies. It is 679 also rather probable that most students of the effects of pet-keeping on quality of life like dogs and 680 cats and therefore a priory believe that keeping pets has a positive effect on quality of life [17,31]

681
(which was true also for us when we registered this study in our grant proposal). It is possible that 682 most people are reluctant to publish results that contradict their expectations [17]. Similarly, many 683 editors and referees are probably less willing to endorse the publication of manuscripts that contradict 684 general opinion, and very often their personal beliefs too. variability in the mental health problems score in the total population (but more than 10% of variability in the Toxoplasma-free subpopulation). It is important to remember, however, that these 693 effects are of similar strength as those of other risk factors -smoking, alcohol, illegal drugs, and high 694 BMI. Most importantly, more than 50% of households in developed countries keep a dog or a cat and 695 this number is continuously rising. Therefore, the real impacts of formally weak effects of keeping of 696 pets on public health and wellbeing could be enormous and growing.