Neuroimaging and Fluid-Based Biomarkers in Sport-Related Concussion in Female Athletes: A Scoping Review

Diffusion MRI (two female-only reports, five mixed-sex reports)

The most common neuroimaging modality studied was diffusion MRI, as it was in a recent review of biomarkers and emerging technologies in SRC by Tabor et al. ²⁴ Diffusion MRI is thought to reflect subtle microstructural changes that occur as a result of concussion, with research to date suggesting edema, axonal damage, and demyelination as possible pathological sources for diffusion differences.^54,55 Overall, diffusion MRI reports did identify some consistent areas of the brain as vulnerable to concussion, but varied in the nature of their findings. Multiple reports found diffusion differences in the corpus callosum at acute and subacute timeframes,^29,39,44,52 and two of those also saw differences in the corona radiata.^29,52 For female athletes, either separately or within a mixed-sex group, higher fractional anisotropy (FA),^39,44,52 lower mean diffusivity (MD) and axial diffusivity (AD),^39,44 and lower radial diffusivity (RD)^44,52 were the most common finding across studies, though lower FA ²⁹ and lack of significant changes to various diffusion metrics^29,37,45 were also seen. Two reports that included the sex-by-group interaction in analyses did not find it significant^37,44; however, Churchill et al. ²⁹ found lower FA and nonsignificantly lower MD for SRC females relative to controls while male SRC athletes had higher FA and MD than controls, suggesting the existence of sex-specific diffusion differences in response to injury. In addition, white matter features from diffusion MRI were important in classifying athletes with post-concussive syndrome and the predictive index was associated with the number of prior concussions for female but not male athletes. ⁵⁰ Variation among the reports with respect to direction of diffusion metric differences is consistent with findings in a systematic review by Lindsey et al. ⁵⁶ and may be partially attributed to timing of visits with respect to injury, variation in region of interests (ROIs), and differences in scanners and scanning parameters.

Summary: Diffusion data highlights the sensitivity of corpus callosum in female athletes following concussion, most typically characterized by higher FA and lower diffusivity, though variation across studies was observed. There was limited investigation of sex-dependent effects of injury on diffusion metrics, with one study identifying different effects in female athletes compared to male athletes.

MR spectroscopy (two female-only reports)

Magnetic resonance spectroscopy (MRS) can quantify metabolite concentrations that may change following concussion as a result of mitochondrial dysfunction and altered metabolism due to ionic flux and glutamate release. ⁵⁷ Among the most common metabolites measured are: N-acetyl aspartate (NAA), a neuronal marker synthesized by neurons; creatine (Cr), involved in energy homeostasis; glutamate (Glu), an excitatory neurotransmitter, and glutamine (Gln), an amino acid precursor to glutamate, which are often reported as Glx = Glu + Gln; and myo-inositol (mI), a neurotransmitter precursor involved in calcium signaling. ⁵⁸ Only two reports that utilized MRS met our study criteria and they had diverging results. One found higher Glx/Cr in the primary motor cortex and dorsolateral prefrontal cortex at 6 months post-injury, ²⁷ while the other reported lower Gln and Gln/Cr in prefrontal white matter acutely and around 3 months post-injury which resolved by 6 months. ⁴⁵ Both studies observed no differences in NAA ⁴⁵ and NAA/Cr ²⁷ in female athletes with SRC. Interestingly, these results conflict with those of a meta-analysis of MRI in all severity TBI where decrease in NAA and no effects for either Cr or Glx across brain regions were observed. ⁵⁸ The differences between these reports themselves and with the meta-analysis may be somewhat explained by regional variation of MRS results. ⁵⁸

Summary: Only two studies investigated MRS differences following concussion, both with a limited number of injuries. While this makes definitive conclusions difficult, it is noteworthy that results varied across studies and were inconsistent with a meta-analysis of MRS results from all severity TBI.

Arterial spin labeling (one female-only report, five mixed-sex reports)

Although the mechanism is not entirely understood, cerebral blood flow (CBF) may be affected by mTBI through decoupling from neuronal metabolism, disruption to the blood–brain barrier, and impaired autoregulation. ⁵⁹

ASL can be used to quantify CBF and potentially identify changes to the cerebrovascular system occurring as a consequence of concussion. Included ASL reports had varying results. Three found higher (or trending higher) CBF,^29,44,47 while two found lower (or trending lower) CBF post-SRC.^28,53 In addition, athletes reporting physical symptoms at 6 weeks post-SRC had higher CBF in the left dorsal ACC than asymptomatic athletes and controls. ⁴⁷ Other regions with significant differences for post-SRC female athletes included the caudate, ⁴⁴ left dorsal ACC, ⁴⁷ left insula, ⁴⁷ left superior and middle temporal gyri, ²⁸ and left superior parietal lobule, ²⁸ with no overlap between studies, although regions investigated across studies had little overlap. Higher CBF being slightly more common here is somewhat in contrast with the reports in a recent systematic review of ASL in mTBI, although their findings also varied between increased and decreased CBF. ⁵⁹ In terms of differing effects of injury on CBF based on sex, most studies found no sex difference ⁴⁷ or nonsignificant sex-by-group interactions,^44,53 though Churchill et al. ²⁹ found nonsignificantly higher CBF for female SRC athletes but lower CBF for post-SRC males versus controls. Similarly, male athletes had decreased occipital CBF during selective head cooling post-SRC, while females had stable or increased CBF, ⁵¹ though the sex-by-group interaction was not significant.

Summary: A small handful of studies have investigated ASL differences following SRC. Results are mixed in terms of direction of effects and impacted location, with elevated CBF being slightly more common. There is minimal evidence for sex-specific effect of concussion on CBF, though such analyses were likely underpowered to identify interactions.

Resting-state functional MRI (one female-only report, two mixed-sex reports)

Resting-state functional MRI (rs-fMRI) measures changes to the blood–oxygen-level dependent (BOLD) signal to evaluate the functional connectivity within the brain which could be disrupted by SRC. Included rs-fMRI studies showed that recently concussed athletes had altered functional connectivity.^37,39,44 Lower caudate connectivity to the pallidum and thalamus was present post-SRC, as well as a negative relationship between symptom severity and activity in the cerebellum and a positive relationship between days since injury and cerebellar-parietal connectivity. ⁴⁴ In a female-only report, altered connectivity (higher and lower) between the whole brain and the default mode, cerebellar, and executive networks was linked with diffusion tensor imaging (DTI) changes along with symptom severity. ³⁹ While one rs-fMRI report did not see a significant sex-by-group interaction, ⁴⁴ another found lower frontoparietal-visual association and default mode-visual II network connectivity for concussed males than for controls, with no group difference for females. ³⁷

Summary: Female athletes have altered functional connectivity post-SRC in various brain regions, with some evidence of potential sex differences.

Other neuroimaging modalities (five mixed-sex reports)

A handful of reports used other neuroimaging metrics, ranging from volumetric MRI to electroencephalography (EEG), with some evidence of sex-specific concussion effects across studies. For example, a study of EEG, which measures electrical currents to quantify neuronal function via electrodes on the scalp, found that the enhanced brain function index (eBFI; a percentile assigned based on EEG, neurocognitive measures, and clinical signs/symptoms of concussion) was lower for female versus male athletes overall. ³⁰ Moreover, split-by-sex analyses demonstrated that while SRC athletes had lower eBFI than controls in both sexes, effects were stronger in females (though no interactions with sex were significant). ³⁰ Similarly, sex differences in white matter hypointensity burden following concussion (higher in male vs. female) were observed using susceptibility weighted imaging. ³⁶ Using quantitative susceptibility mapping (QSM), researchers found a trend towards a significant sex-by-injury interaction driven by lower susceptibility in the straight sinus (a measure of cerebral venous oxygen saturation) in female SRC athletes versus controls. ⁵³ In contrast, another group saw higher susceptibility in the cerebellum post-SRC across a mostly male cohort, with no sex interaction observed. ⁴⁴ Finally, an analysis of adolescent female athletes post-SRC versus controls grouped by menarche status did not find any significant differences in volumetric measures. ³⁸

Summary: Some evidence across other neuroimaging modalities suggests the presence of sex-specific effects of SRC, including electrophysiological measures (worse in females), the presence of white matter hypointensities (greater in males), and a measure of cerebral venous oxygen saturation (lower in females). These findings were mostly informed through split-by-sex analyses rather than significant interactions.

Neurological injury markers (six mixed-sex reports)

Biomarkers with connection to neurological injury have gained attention for their potential in aiding diagnosis and prognosis for concussion. ²⁴ These include glial fibrillary acidic protein (GFAP), a marker of glial damage; neurofilament light (NfL) and tau, both axonal cytoskeletal proteins; and ubiquitin carboxyl-terminal esterase L1 (UCHL1), an enzyme found in neurons involved in protein degradation.^35,60,61

While post-SRC male athletes have high GFAP with respect to baseline ⁴⁰ or uninjured levels, ⁴⁹ female athletes either respond similarly ⁴⁹ or show no significant difference. ⁴⁰ Furthermore, an underlying sex difference exists for GFAP with higher levels for female subjects than males. ⁴⁹ Although baseline UCHL1 is lower in females,^25,40,49 post-SRC females and males may have similar levels of UCHL1.^35,49 While three studies indicated UCHL1 was not promising for distinguishing those with SRC from controls,^25,35,40 Tabor et al. ⁴⁹ found females had high post-SRC UCHL1 compared to uninjured levels, while males showed no significant difference. Sex differences appear to exist in the trajectory of NfL post-SRC with increases for male athletes, but no significant differences for female athletes.^40,49 Additionally, Bui et al. ²⁶ observed that a high percentage of female athletes with SRC had low NfL at baseline that remained low following injury, a trajectory that was less likely to be associated with loss of consciousness or post-traumatic amnesia, features which are less common for females than males.^62–64 Finally, reports on post-SRC trajectories of tau found evidence for sex differences, though the direction of these effects varied (e.g., no consistent pattern in elevated tau in either sex).^40,41,49

Summary: Evidence suggests that sex impacts baseline levels of neurologic blood biomarkers and that effects of SRC may be sex-dependent. Promise for using these markers to distinguish concussed female athletes from controls is mixed and may be dependent on timing from injury.

Cytokines/chemokines (four mixed-sex reports)

Inflammation has been shown to result from mild TBI and is suggested as a possible root of many concussion symptoms.^57,65 Measuring peripheral inflammatory cytokines and chemokines may offer a way to evaluate the effects of concussion and to identify possible mechanisms for these effects. ³³ Three of the four reports examining various cytokines and chemokines (three are from the same group) are suggestive of sex differences. Symptom severity was negatively correlated with interferon (IFN)-γ, myeloperoxidase (MPO), and tumor necrosis factor (TNF)-α and positively correlated with monocyte chemoattractant protein (MCP)-4 for female athletes, but in male athletes symptoms positively correlated with IFN-γ. ³³ In addition, MPO, macrophage inflammatory protein-1β, and eotaxin were lower, and IFN-γ-induced protein-10 was higher in female versus male SRC athletes, although these differences did not survive correction for multiple comparisons. ³³ Analyzing results with sex-controlled z-scores, Di Battista et al. ³² found higher post-SRC MCP-4 and MIP-1β compared to healthy controls and a positive relationship between days to recovery and higher subacute MCP-1 and MCP-4 in athletes with SRC, but not musculoskeletal injury. When examining immunoreactivity through stimulation of blood samples separately with lipopolysaccharide (LPS) and resiquimod (R848), significant effects of SRC on LPS reactivity were seen in male but not female athletes. ³⁴ Furthermore, the effects of history of concussion on post-SRC immunoreactivity also differed between male and female athletes in the same study. In contrast, O’Brien et al. ⁴² did not find a main effect of time, sex, or a sex-by-time interaction for IL-1β or IL-18 levels post-SRC.

Summary: Evidence supports sex may moderate the effects of concussion on peripheral inflammatory markers and their association with concussion symptoms, consistent with known sex differences in immune function. ⁶⁶

Other fluid-based biomarkers (two female-only reports; three mixed-sex reports)

The only fluid-based biomarker examined in female-only reports was progesterone. Sex hormones have been theorized to potentially play a role in the possible menstrual cycle influence on concussion vulnerability or outcomes and in sex differences in concussion.^19,67,68 Additionally, progesterone has been shown to provide a degree of neuroprotection in animal models and phase II clinical trials, but failed phase III trials indicate its role is not so straightforward. ¹⁹ Chen et al. ²⁸ found associations between progesterone, CBF, and perceived stress in female athletes, while perceived stress symptoms negatively correlated with progesterone in the SRC group. However, Ott et al. ⁴³ saw no significant associations between SRC and progesterone.

Research into other fluid-based biomarkers primarily showed no evidence of sex differences. One study found no significant findings or sex-by-time interactions for several markers including phosphoprotein enriched in astrocytes 15, von Willebrand factor, claudin-5, fibrinogen, high mobility group box protein 1, and 4-hydroxynoneal, though a decrease in brain lipid binding protein (BLBP) was seen for female athletes post-SRC compared to pre-injury baseline but not for males (no significant sex-by-time interaction). ⁴⁸ Similarly, no significant sex-by-time interactions post-SRC were found in an analysis of multiple microRNA in plasma, but plasma levels of miR-221-3p and miR-27a-3p were decreased post-SRC compared to baseline and were inversely correlated with symptom severity across the whole cohort. ⁴⁶ Finally, urinary insulin-like growth factor (IGF) 1 and IGF binding protein 5 were lower post-SRC compared to controls with no significant sex-specific differences. ³¹

Summary: Only two studies in this review directly measured endocrine markers in female athletes (progesterone in both cases), with mixed results. Investigation of additional fluid biomarkers does not support sex-specific effects but rather demonstrates limited associations of these markers with SRC status and symptom severity.