AAN 2026: How Biological Sex and Gender Independently Shape TBI Survival and Recovery

Tatyana Mollayeva, MD, PhD, a scientist from the University of Toronto’s Rehabilitation Sciences Institute, continued her discussion with Pharmacy Times at the 2026 American Academy of Neurology (AAN) annual meeting, presenting findings from a large Ontario population-based study separating the effects of biological sex and gender on traumatic brain injury (TBI) outcomes. She shared that sex emerged as the stronger predictor of early excess mortality, while gender more strongly predicted discharge destination—home versus rehabilitation or supportive care.

Pharmacy Times: What does current research tell us about how biological sex influences TBI outcomes, and should it change how we treat these patients?

Tatyana Mollayeva, MD, PhD: That’s an extremely important question, because I wish I could say yes or no, black or white, but in fact, that’s just the matter—we have only started to touch the surface and bring evidence that these factors matter. The goal is largely to bring attention and to continue this line of research forward. I’ll speak a little about some of our research in which we tried to separate the effect of sex from the effect of gender at the population level in Ontario, Canada—the largest province in Canada, covering almost 40% of the Canadian population. Ontario is extremely diverse—probably the most diverse of all provinces—representing a number of cultures, ethnicities, and races. We sought to understand whether we would be able to detect an effect of sex separately from an effect of gender on certain outcomes that are of extreme relevance for sustaining our publicly funded healthcare system.

We had hypotheses in mind, because researchers before us—including people we greatly admire, who initiated this line of research, first in Europe and then in the US before it came to Canada—had done foundational work. It started with the DEMON study, and our hypotheses were largely based on studies that originated in the US in traumatic brain injury because complex studies had been conducted there. Other hypotheses were built on observations of excess mortality—specifically, that females appear to survive severe injury better than males. One hypothesis is that this is driven by physiology and certain hormones believed to be anti-inflammatory and protective for survivorship.

To separate sex from gender in a living person is almost impossible. Even in animal models, it is very difficult, because animals are also social creatures—they form colonies and have social structures. In human beings, we hardly understand all the ways people behave, who supports them, and what their health status is. It is very difficult to capture. So what did we do? We knew that gender is a larger construct: as human beings, we have our roles, responsibilities, and relationships, and then we have our physiology—physiology is, in a sense, nested within gender. Our hypothesis was that if we positioned ourselves within a large dataset using what we had—largely binary sex constructs, because the transgender population is a very limited sample to study separately, though we attempted this as well using methodologies developed in the US—we could make progress. We had binary sex in population-based data, but we did not have a direct measure of gender.

We reasoned that traumatic brain injury is a very gendered event—the circumstances of injury and location of injury differ based on your roles, responsibilities, and relationships in society. So we thought that if we looked at TBI, we would be able to see what separates binary males and females in terms of injury markers and other conditions captured in population-based data through ICD-10 diagnostic codes, across our sample of the whole population, all of whom had the ability to visit healthcare facilities free of charge. That is a particular strength of our system—it covers everyone and is not a selection bias artifact, which is less of a concern in Canada compared to other parts of the world precisely because care is free at the point of access for all citizens.

What we found—with the support of advanced data-driven techniques, necessary because the data comprised millions of data points requiring dimensionality reduction—was that after false discovery rate correction, separation, and internal validation, certain codes loaded strongly and were associated more with males and others more with females. We did not want to enter with a strong hypothesis, but we suspected we would see something reflecting relationships in society and roles and responsibilities, because that is why we started this work in the first place. And indeed, at the top of the factors most associated with being male was falling from scaffolding and other occupational injury domains. On the most extreme end associated with females was assault and abuse—intimate partner violence. It is heartbreaking, but it was internally validated. We do not feel well about this finding, but we sought to see whether this gender—conceived as a continuum, not a binary—would matter in outcomes.

We created a continuous gender score based on effect sizes and code associations. We wanted to see whether this score would matter in outcomes—otherwise, if it simply reflected the codes, it would have limited practical implications. Our hypothesis was that if we looked at excess mortality and severe injury, it would be biological sex—as a binary variable—that played a greater role, based on the prior literature from which we built our hypotheses. But when we looked at function-related or societally related outcomes, we expected gender to be more prominent. We conducted this analysis and investigation, and in fact, we found that when examining early excess mortality after injury, sex played a greater role than the gender score—across several models we built. However, when we looked at discharge destination—whether people were discharged home, to a rehabilitation facility, or to supportive care—it was gender that was the more determining and associated factor.

For us, this is a perfect hypothesis confirmation that it matters. It is not the end—we understand that we are capturing binary sex and that there is great variability, including transgender individuals who are not separately represented due to small sample sizes in our Canadian system. But it is a beginning that says it matters, and it matters to the point where we must take care of this—not only for proper care but also because of what we observe in outcomes.

We then decided to examine whether it matters in cognition, because we know that traumatic brain injury is considered a risk factor for dementia and adverse brain health outcomes. We sought to conduct meta-research, because our Ontario population is one specific group—Canada is not the whole world. So we conducted systematic reviews and reanalyzed findings, because we had conducted many systematic reviews in the past. When done properly, each systematic review leaves you with more questions than answers and a great deal of uncertainty about why the evidence is so heterogeneous. We implemented machine learning approaches to see whether these factors truly matter.

That work required considerable effort. We had longitudinal data from studies published since inception up to one year prior, encompassing around 32 cohorts with both moderate and severe traumatic brain injury. We examined whether cohort characteristics predicted recovery course, where recovery was standardized based on rate of change in cognitive markers using standardized neuropsychological tests—which we consider the gold standard tools for understanding cognitive domains. The data was extremely heterogeneous, with very little description of participants’ social characteristics. However, at least two variables were reported consistently: sex proportion and injury severity, and most also reported age with standard deviation. Some additionally reported education, which is an important factor for cognition given its strong association with cognitive scores.

Because the evidence came from around the world across different publication years, we also looked at the Gender Inequality Index—a structural parameter assigned to each country annually based on maternal mortality rate, representation in political offices, and educational attainment. The score varies based on societies’ progress globally. We assigned this index to each published study as a continuous variable and built three machine learning models. What we found was that without the Gender Inequality Index, sex proportion and gender proportion mattered. But when we assigned the Gender Inequality Index to the models, sex and gender proportion dissipated—the structural index became the dominant variable. The key finding from this meta-analysis is that structural, societal factors matter for how people progress and their rate of change in cognitive recovery. It is preliminary, but it says that you are not independent from your community, your country, or global considerations. That is an important lesson—it tells us that if we want to understand why certain societies are more prone to dementia than others, we may need to look at these topics more precisely. In addition to physiology, we are prone to responding to stresses that accompany our roles and responsibilities in society.

Pharmacy Times: Is there anything you would like to add?

Mollayeva: I think your question was initially about the clinical brain injury matrix, and it is an amazing tool. It is a first step toward personalized medicine, I would say, and it is extremely relevant for traumatic brain injury. It captures a great deal of clinical information and also attends to social parameters as modifiable factors. Social parameters in long-term outcomes probably outweigh many clinical parameters, based on our observations in research—and this tool represents an extraordinary step toward understanding traumatic brain injury and its courses. People come to traumatic brain injury not the same—they are very diverse in terms of their health status, social positioning, risk factors, and frequency of prior injuries, all of which will be reflected in this clinical matrix. Documentation is key for future progress in traumatic brain injury.