Are You A Dog Person or An Increased Risk of Schizophrenia Person?
Household pets and schizophrenia
The Frontier Psychiatrists is a newsletter, about health. Today, we are going to dig into the data on household pet exposure and subsequent development of schizophrenia. Given my recent coverage of BlackCat, I figured some more grim news about cats was in order.
The newsletter is written by Owen Muir, M.D., DFAACAP1, who does not have schizophrenia but has been keenly interested in the well-being of humans with that condition. I thought, for a hot minute, I would end up as a researcher2 into treatments3 for that illness, until
sucked me down the rabbit hole of treatment-resistant mood disorders4 and suicidal adolescents and adults!—amazon affiliate link)Cats Increase The Risk of Schizophrenia, Dogs Decrease It
The short version: “broadly defined cat exposure” is a real risk! The opposite appears to be the case for dogs. I was surprised too. This also allows us to discuss two concepts in research study design, cohort studies and case-control studies. To be clear—It is not the cat that is the problem, per se, it’s a parasite that reproduces in the gut of a cat. The etiology of the protectiveness of dogs also is suspected to be due to their microbial co-travelers.
There are some questions about human health that we can't answer with a randomized controlled trial. If you think something is harmful, you can't randomize people to exposure to that thing to determine causality. This would not be ethical. We are left with natural experiments, things that happen in the world that allow us to suss out a difference.
Case-Control Design
With a case-control study, you start two groups: individuals who have the condition, the “case” group, and individuals with characteristics other than having the condition, the “control” group. The job of the scientist is to see what is different between cases and controls, in their history. These studies look retrospectively, back in time, to try to figure out the difference.
Cohort Study Design
In a cohort study, the opposite approach is taken. You generally look at a very large group of people, starting at a point in time, and then follow them over time to see who is exposed, who is not, who ends up with a disease, and who does not. These studies take many years, in most cases, and require large samples and diligent work. Especially for rare conditions, it can require many, many people in the cohort to be able to determine who had a problem later, because you have to get enough people to develop the problem over to make sense of the data. The strength of a cohort Study is the prospective nature of the research: you don't have the bias of knowing the end of the story at the beginning of the study!
Whenever you see an article about a bad thing that causes a problem, generally, authors are referring to studies that look at large populations for the harmful impacts of this or that environmental exposure. This sort of study is part of a very patient field of medicine called epidemiology. The epidemiologists behind the study found that broadly defined cat ownership more than doubled the risk of developing schizophrenia later in life:
We found an association between broadly defined cat ownership and increased odds of developing schizophrenia-related disorders. The unadjusted pooled odds ratio (OR) was 2.35 (95% CI: 1.38–4.01), while the adjusted pooled estimate was 2.24 (95% CI: 1.61–3.12).5
See that “CI” there? That is a “confidence interval”— it's the range over which we believe the real effect might be found. Any given outcome in a research trial is subject to a degree of error, and the more with the confidence interval, the more confident we are that the actual result in the population is within the 95% range represented by those two numbers.
Let's dig into some of the science behind why any of this makes sense…
The exposure, in the case of cats, isn’t to the felines themselves. It’s a parasite that has a life cycle that starts in the gut of a cat, where it reproduces, and then its eggs are expelled in cat poop. Its life cycle in nature involves finding a rat to invade, and then creeping into its brain to make it less afraid of cats. The less fearful rats get eaten by cats. This is where reproduction can happen again. This same parasite, toxoplasma gondii, can also infect humans. There has long been a suspected link between toxoplasma gondii—a parasite—and schizophrenia6. A quick summary:
In animals, infection with Toxoplasma gondii can alter behavior and neurotransmitter function. In humans, acute infection with T. gondii can produce psychotic symptoms similar to those displayed by persons with schizophrenia. Since 1953, a total of 19 studies of T. gondii antibodies in persons with schizophrenia and other severe psychiatric disorders and in controls have been reported; 18 reported a higher percentage of antibodies in the affected persons; in 11 studies the difference was statistically significant. Two other studies found that exposure to cats in childhood was a risk factor for the development of schizophrenia. Some medications used to treat schizophrenia inhibit the replication of T. gondii in cell culture.
There is even a fantastic Radiolab episode on the topic.
A lot of people have pets though. A lot of people love the heck out of their pets. If there's a real difference, it's probably worth knowing. But since we can't randomize people ethically in a prospective manner to something that might cause a medical illness, with no benefit, the only way to determine whether cats or dogs are safer pets? Using the above Study Designs! To recap—Case-control studies work backward from known cases, and cohort studies move forward in time from known cohorts to evaluate what happens later.
My favorite thing about this paper is the following: they address Bonferroni correction for multiple comparisons:
The analyses involved 2 pets and 2 clinical diagnoses. Hence an alpha of p< = .0125 (.05/4) was employed as an indication of statistical significance. A value of .05< = p >.0125 was con- sidered to indicate a trend level of significance.
You can’t just use 0.05 as the cutoff when you have multiple comparisons! A breath of fresh scientific air!
Moving past the math…let’s take a look at what the study found. I warned you, the interactions are complicated. I have illustrated the walk-through of the data.
Starting with the survival analysis, which we call Kaplan Meyer, as I've covered previously, we are looking at “time to event”—the horizontal axis is different ages, and the vertical access is the fraction with a cat. So it's “time to get a cat in the household” vertically.
That above graph quite honestly notes that they didn't see a statistical difference in that analysis, but it gets more complicated. The survival analysis:
The survival analysis when we look at the development of schizophrenia demonstrates a statistically significant protective effect of dogs in the household of kids who might develop schizophrenia.
Things get complicated when we see the interaction of cats versus dogs, this first illustrated graph just tells you have to read the graph afterward, which is the same graph:
And now here's the actual outcome, annotated:
Let me evaluate the difference between schizophrenia and bipolar disorder and see a protective effect against the development of bipolar disorder later in life! Again, these were households that had early cat exposure, which seems to be the kindling for the fire, and dogs can be protective for older individuals. There's also some signal for increased risk for bipolar disorder from cats, which hadn't been noted in previous data that I have reviewed.
Cats are a risky decision, albeit about a disease that has a low base rate. Increasing the rate of a rare illness from 1% percent to 2% is a different framing from “cats double the risk of schizophrenia.” Both of those statements are true, in certain age groups, and depending on the household exposure to dogs, and when. My takeaway:
if you're going to have a cat, might as well get a dog also.
The basic science, however, suggests that, if these results are replicated, there's yet more data for the profound biology underlying the relationship between our microbiome, our mental health, and our environment, in determining who gets sick, when, and with what. We evolved with cats and dogs, and the parasites and microbes that evolve with them, too.
Other articles about schizophrenia and learning to read science like a pro include (yes, you’d need to become a paid subscriber):
Are Oral Medicines Driving Obesity?
Are Antipsyhchotics More Deadly than Fentanyl in Kids?
The Forgotten Horror of Tardive Dyskinesia
Schizophrenia Is More Dangerous than A Car Crash
…Become a subscriber today!
Yes, I got more letters after my name. I know, It’s a lot. It’s thanks to being named a Distinguished Fellow of the American Academy of Child and Adolescent Psychiatry.
Manu P, Sarpal D, Muir O, Kane JM, Correll CU. When can patients with potentially life-threatening adverse effects be rechallenged with clozapine? A systematic review of the published literature. Schizophr Res. 2012 Feb;134(2-3):180-6. doi: 10.1016/j.schres.2011.10.014. Epub 2011 Nov 22. PMID: 22113154; PMCID: PMC3318984.
Correll, C. U., Muir, O., Al-Jadiri, A., Kapoor, S., Carella, M., Sheridan, E., ... & Kane, J. (2013, December). Attitudes of children and adolescents and their caregivers towards long-acting injectable antipsychotics in a cohort of youth initiating oral antipsychotic treatment. In Neuropsychopharmacology(Vol. 38, pp. S280-S280). MACMILLAN BUILDING, 4 CRINAN ST, LONDON N1 9XW, ENGLAND: NATURE PUBLISHING GROUP.
Tendler, A., Goerigk, S., Zibman, S., Ouaknine, S., Harmelech, T., Pell, G. S., Zangen, A., Harvey, S. A., Grammer, G., Stehberg, J., Adefolarin, O., Muir, O., MacMillan, C., Ghelber, D., Duffy, W., Mania, I., Faruqui, Z., Munasifi, F., Antin, T., . . . Roth, Y. (2023). Deep TMS H1 Coil treatment for depression: Results from a large post-marketing data analysis. Psychiatry Research, 324, 115179. https://doi.org/10.1016/j.psychres.2023.115179
Yolken, R., Stallings, C., Origoni, A., Katsafanas, E., Sweeney, K., Squire, A., & Dickerson, F. (2019). Exposure to household pet cats and dogs in childhood and risk of subsequent diagnosis of schizophrenia or bipolar disorder. PLOS ONE, 14(12), e0225320. https://doi.org/10.1371/journal.pone.0225320
Torrey EF, Yolken RH. Toxoplasma gondii and schizophrenia. Emerg Infect Dis. 2003 Nov;9(11):1375-80. doi: 10.3201/eid0911.030143. PMID: 14725265; PMCID: PMC3035534.
So my first thought here is that dog owners and cat owners probably differ in some relevant ways--it's easier to keep a dog if you a) don't live in a city, b) don't live in subsidized or supportive housing, c) have more free time, d) have more resources, and e) have enough executive function to take care of said dog without drawing the attention of animal control. This is less true of cats, who are lower maintenance and acceptable in more housing situations. Looking over control group matching variables, I think some but not all of these concerns are mitigated. Insofar as children of schizophrenics and/or those with higher schizophrenia risk are less likely to fall into these categories, should we expect that dog exposure itself is causal? I'd be curious to compare children who grew up with dogs to the children of former dog-owners who didn't cohabitate with dogs during childhood.
I also love to see a Bonferroni correction. In grad school I did be telling people they needed to correct for multiple comparisons and they did be like "nah who even does that" and "stop reading over my shoulder" and "who appointed you emperor of statistics?" a disconcerting amount of the time (okay, the last two were just my husband).
Fascinating thank you. I wonder if bitd there was a survival advantage for those of us poor souls who are allergic to cats. Declaration of interest - I like both species and this is not a serious proposition.