The relationship between symptom burden and systemic inflammation differs between male and female athletes following concussion

Sport-related concussion (SRC) is a complex injury that can lead to somatic, cognitive, visual, sleep and emotional disturbances. Symptoms typically abate within weeks but can persist for months to years in a subset of individuals [1]. Importantly, symptom burden is a key tool for clinicians in guiding patients through the recovery process; the resolution of symptoms is a prerequisite for medical clearance and return to sport participation [1]. While the biological mechanisms underlying symptom presentation following injury remain elusive, the gap in our understanding comes as no surprise, as there is significant heterogeneity in symptom presentation between individuals, both in severity and type. In addition, symptoms commonly observed following SRC have also been reported in a variety of other conditions such as polytrauma, infection, and mental illness [2‐5]. In views of this, inflammation presents as a unifying concept, as it has not only been associated with concussion-like symptoms across numerous medical conditions [5], but is increasingly recognized as an important and prominent feature of concussion secondary injury [6‐10]. Therefore, investigating the relationship between inflammation and symptom burden following concussion may help elucidate clinically meaningful pathophysiological mechanisms that mediate patient recovery [11, 12].

Rathbone and colleagues suggested inflammation as a common mechanism underlying the constellation of symptoms observed following concussion [5]. This assertion was made in light of the evidence linking inflammatory mediators to numerous symptoms commonly observed following concussion that are seen in medical maladies such as headache, chronic fatigue syndrome, in response to immunomodulatory medical treatments, and psychological conditions such as depression and anxiety [5]. Importantly, evidence of these relationships was commonly characterized by the presence of symptom(s) and altered blood concentrations (typically an increase) of inflammatory cytokines and/or chemokines [5].

Systemic inflammatory mediators can communicate with the central nervous system (CNS) via several mechanisms. Primary afferent and efferent nerves innervating the CNS can respond to cytokine and chemokine signalling from either the brain or periphery [13, 14], CNS-derived neuroendocrine hormones can interact with their respective receptors on circulating leukocytes in the periphery [15‐18], and the blood brain barrier (BBB) permits passage of leukocytes and their mediators (cytokines and chemokines) via receptor-mediated transport, endothelial transmigration, and/or diffusion [19]. As a result of this bi-directional communicative network, peripheral inflammatory indices may be used to 1) evaluate the systemic consequences of brain-related maladies, or 2) indirectly elucidate processes occurring in the CNS. An example of this can be seen in sickness behaviour, whereby infection or illness results in fatigue, malaise, decreased appetite, lack of concentration, and feelings of depression and lethargy [20‐22]. This constellation of behavioural and physiological changes is due primarily to systemic and centrally produced cytokines, namely tumor necrosis factor (TNF)-α, interleukins (IL)-1β, IL-6, and interferons (IFNs) [13, 22‐26].

Our group recently identified a unique inflammatory profile in the peripheral blood of athletes following SRC that was distinct from non-head injury and characterized by elevations in the circulating chemokines monocyte chemoattractant protein (MCP)-4 and macrophage inflammatory protein (MIP)-1β [6]. In addition, higher concentrations of MCP-1 and -4 were associated with prolonged recovery [6]. Similarly, altered inflammatory gene expression in leukocytes [8, 9], and increased brain extracellular vesicles containing TNF-α and IL-8 have also been observed in the peripheral blood of subjects following SRC [27]. Despite these findings and the strong linkage between inflammation and symptomology across numerous clinical conditions, human data characterizing the potential role of inflammation on symptom burden following concussion is scarce. Su and colleagues observed that a higher concentration of the circulating inflammatory index C-reactive protein (CRP) was associated with a greater symptom burden in subjects in the months following mTBI [28]. Likewise, Nitta and colleagues recently identified a positive relationship between serum interleukin IL-6 concentrations and symptom duration following SRC [10]. While this is intriguing, the relationship between concussion symptoms and inflammation has yet to be evaluated across the range of inflammatory mediators previously studied in TBI [6, 29‐31]. Furthermore, probable sex differences have not been accounted for; despite the profound differences observed in male versus female immune responses as a consequence of genetic, environment and hormonal influences [32, 33], few studies evaluating inflammation stratify their cohorts by sex [33].

The present study builds upon prior works by examining the relationships between symptom burden and inflammation in males and females following SRC. We hypothesized that symptoms reported in the subacute period following injury would correlate with systemic inflammatory cytokines and chemokines, and that this relationship would differ between males and females. This hypothesis was evaluated in a cohort of university-level athletes assessed within seven days post-injury using a set of multivariate PLS models. These analyses were performed with the goal of improving our understanding of how inflammation may mediate symptom burden in males and females following SRC.

In this study, we identified a significant relationship between symptom burden and systemic inflammation following SRC, with differences between males and females. Within the first week of injury, reported symptom severity was inversely correlated with inflammatory cytokines in the peripheral blood of female athletes, yet positively correlated in males. Importantly, these differences were noted despite male and female groups having a comparable symptom burden and time to recovery. This suggests that inflammation is an important and clinically relevant component of secondary injury following SRC but may present differently in males and females.

We observed greater symptom burden was associated with lower concentrations of the classical inflammatory cytokines IFN-γ and TNF-α, and the innate immune function marker MPO in female athletes following SRC; a positive relationship was observed with the chemokine MCP-4. The somatic symptom cluster displayed the greatest effect, while the cognitive cluster displayed the lowest, and the overall relationship between symptoms and inflammation was comparatively more generalizable in females than males (PLS cross-correlation R² = 0.07 in males, R² = 0.29 in females). While we hypothesized a sex-difference in the relationship between inflammation and symptom burden following injury, the direction of the association observed in female athletes is seemingly non-intuitive. The preponderance of evidence linking concussion symptoms to inflammation suggests a positive relationship (for review see [5]), and females typically have a greater inflammatory response to challenge compared to males [33]. However, in a recent meta-analysis evaluating peripheral chemokine and cytokines in depression, while the majority of studies found positive relationships, lower concentrations of TNF-α and IFN-γ were also observed (for review see [34]). Furthermore, a number of perturbations to inflammatory genes in peripheral leukocytes have been observed following SRC, displaying both upregulation and downregulation [8, 9]. Hence, it is possible that both the complexity of the inflammatory response to concussion as well as potentially divergent responses in males and females contribute to these mixed findings. In addition, it is conceivable that the menstrual cycle may have had an effect; immune function can vary widely with hormonal changes during the menstrual cycle, with potential immunosuppression in the luteal phase [35‐37]. While it is still unclear why we observed that symptom burden was correlated with lower levels of inflammatory cytokines in females following SRC, the results of this investigation provide supportive evidence that inflammation following concussion should not be looked at without dichotomizing groups by sex.

As opposed to females, in the current study symptom severity was positively correlated with blood concentrations of IFN-γ in male athletes following SRC. This observation is generally supported by literature that has shown a positive association between the family of IFN proteins and symptoms such as headache, fatigue, irritability, hostility, and depression [34, 38, 39] in numerous maladies. While we found IFN-γ was associated with all symptom clusters, cognitive symptoms displayed the highest effect; in females cognitive symptoms displayed the lowest effect. Furthermore, we observed a small, stable effect in males between the emotion symptom cluster and the chemokines IL-8, MCP-4 and TARC. These mediators are primarily responsible for neutrophil [40], monocyte [40] and lymphocyte chemoattraction [41], respectively, and chemokines have been broadly implicated in leukocyte trafficking to the brain following TBI [29, 40, 42, 43]; indeed, higher concentrations of IL-8 in the blood acutely following severe TBI have been correlated with unfavorable outcome [30]. Regarding symptom burden, although chemokines have been broadly implicated in depression, the specific relationship(s) between MCP-4, IL-8 and TARC and the behaviours and mood states comprising the “emotion” symptom cluster in the current study (emotional, irritability, sadness, nervous/anxious) are not well defined. Furthermore, it is important not to overstate this relationship, as these findings were found on the second LV comprising only 25% of the covariance in the PLS model and are comparatively small relative to the magnitude of the relationship observed between symptoms and IFN-γ. Taken together, further research is warranted both to replicate these findings and to investigate the mechanisms mediating the relationship between inflammation and symptom burden in male athletes following SRC.

The mechanism(s) underlying the relationship observed in the current study between symptom burden and inflammation following SRC are unknown. However, given the strong influence of the neuroendocrine system on inflammation [44‐46], its role in both brain injury [9, 17, 47‐49] and related symptomology [50‐53] along with the differences in neuroendocrine biology between males and females [54‐56], it is plausible that the stress-immune axis mediates the relationship between symptoms and inflammation following concussion. The two major arms of the body’s stress system, the sympathetic nervous system (SNS) and the hypothalamic pituitary adrenal (HPA) axis can augment the immune system in numerous ways. For example, the HPA-axis, through the actions of glucocorticoids can have strong immunosuppressive effects [57‐59], while the SNS can be pro- or anti-inflammatory depending on the nature of the stimulus and subsequent catecholamine-adrenergic receptor pairing [15, 32, 31], and have recently observed that neuroendocrine hormones in the blood following sport concussion are associated with both symptom burden and time to medical clearance [60]. In support of the latter, Merchant-Borna and colleagues observed a change in lymphocyte transcription of genes regulating HPA-axis activity and inflammation at seven days post-SRC [9]. However, the complex interactions between sex, neuroendocrine-immune signalling, their effect on symptom burden, and the temporal shifts in these processes that likely occur to restore homeostasis throughout recovery are not understood. Lastly, given the pleiotropicity, biological redundancy, and dynamic nature of the inflammatory response, it is difficult to control for the vast number of confounds when conducting a human study. As a result, a normative healthy group can be a potential source of biological noise in statistical analyses. The purposeful exclusion of a healthy control sample in the current study allowed for a direct evaluation of the relationship between symptom burden and inflammation post-injury. However, this approach also allows for the possibility that pre-injury biological conditions and/or phenotypes affecting the inflammatory system or that contribute to symptom burden, impacted our findings. In view of this, in the current study we observed a relationship between inflammatory biomarkers and concussion history in male athletes following SRC (Supplementary Fig. 1), and it is unclear how this may have impacted post-injury inflammation. Hence future studies aimed at uncovering the intricacies of these interrelationships, including both pre- and post-injury biology and the potential mediating effect of concussion history, are necessary.

The results of the study must be interpreted within the context of its limitations. While the sport concussion assessment tool (SCAT) symptom profile is a well-studied and useful tool utilized by clinicians and concussion researchers, investigation into specific symptoms may have benefited from more directed measures. For example, clinical screening tools for depression (i.e., the Beck Depression Inventory) and anxiety disorders (i.e., the general anxiety disorder questionnaire-7) may have provided utility to the current analysis, as would physiological evaluations of sleep (i.e., heart rate variability monitoring), and more objective neuropsychological evaluations of cognitive function. Furthermore, a larger study sample would have permitted a greater degree of subgroup analysis (i.e., prior concussion history, menstrual cycle in females), and baseline evaluations would have allowed for the consideration of potential pre-injury differences in symptom presentation between males and females [61, 62]. Finally, a more narrow and acute sample timepoint may have helped capture a more profound inflammatory response to injury; despite there being no relationship between the time elapsed from SRC to blood draw and either symptom severity (spearman rho = 0.05, p = 0.75), or time to medical clearance (spearman rho = 0.17, p = 0.29), we observed a correlation between inflammatory biomarkers and days from SRC to blood draw (Supplementary Fig. 2). Lastly, our study focused on evaluating the relationship between symptom burden and systemic inflammation at a single time point post-injury, therefore follow-up studies are required to characterize this relationship throughout recovery; given our groups’ prior findings of persistent symptoms at medical clearance utilizing both advanced neuroimaging [63, 64] and blood biomarkers [65], it remains unclear how symptom burden tracks biological recovery.

Symptom burden is associated with unique inflammatory profiles in male and female athletes following SRC. Symptom severity is associated with elevated blood concentrations of IFN-γ in males, yet lower levels of IFN-γ, TNF-α, and MPO in females: the relationship between symptoms and inflammation appears to be more generalizable in females. Future investigations into inflammation as a clinically meaningful secondary injury process following concussion should not ignore the potentially divergent pathophysiology underlying symptom presentation in males and females.