Mild traumatic brain injury (mTBI) results in a constellation of symptoms commonly referred to as a concussion. It is unclear why certain individuals experience persistent symptoms. Given the growing evidence linking the microbiome with cognition and inflammation, we examined whether longitudinal microbiome patterns were associated with concussion symptoms. A cohort study of 118 children (aged 7–21 years) was conducted. Symptoms were assessed at three timepoints post-injury (4, 11, and 30 days) using the Post-Concussion Symptom Inventory. Saliva microbial activity was measured at each timepoint using RNA sequencing. A linear mixed model assessed the relationship between microbial activity and symptom burden while controlling for age, sex, and days post-mTBI. The participants’ mean age was 16 (±3) years. The symptom burden decreased across all three timepoints (25 ± 22, 13 ± 17, and 5 ± 12). The longitudinal symptom burden was associated with elevated activity of Lactobacillus (F = 5.47; adj. p = 0.020) and Saccharomyces (F = 6.79; adj. p = 0.020) and reduced activity of Micrococcus (F = 7.94, adj. p = 0.015). These results do not establish a causative relationship, or support the use of microbial measures as a concussion test. Further studies are needed to explore the role of the gut–brain axis in mTBI.

Poor sleep and psychological symptoms (depressive or anxiety symptoms) are common following concussion. Saliva microRNAs (miRNAs) are promising concussion biomarkers based on associations with the presence of a concussion in prior work. The purpose of this study was to identify saliva miRNAs that differ between participants with and without concussion and show correlations with sleep and/or psychological symptoms. We analyzed salivary miRNAs and concussion symptom data from n = 272 concussion cases (median: 2.5 days postinjury, interquartile range: 2.41–10.65) and n = 153 uninjured controls. Concussion symptoms were quantified with the Post-Concussion Symptom Scale (PCSS). miRNAs were quantified with high-throughput small RNA sequencing. Spearman rank correlations were used to evaluate the association between miRNAs and PCSS sleep (“Drowsiness,” “Trouble falling asleep”) and PCSS psychological (“Feeling More Emotional,” “Irritability,” “Sadness,” and “Nervousness”) symptoms. Wilcoxon rank sum tests evaluated differences in miRNAs between the concussion versus the uninjured controls. All tests were evaluated at the Benjamini-Hochberg adjusted alpha level of 0.05. We identified 14 miRNAs that were correlated with sleep symptoms ( n = 5 positively and n = 9 negatively correlated), 28 with psychological symptoms ( n = 12 positively and n = 16 negatively correlated), and 4 miRNAs with both (miR-361-5p, miR-4510, miR-193a-3p, miR-3074-5p). Gene targets for miRNAs correlated with both sleep and psychological symptoms were significantly enriched within “spinocerebellar ataxia,” “dopaminergic synapse,” “axon guidance,” and “circadian entrainment” pathways. Of the four miRNAs correlated with both psychological and sleep symptoms, miR-4510 and miR-193a-3p were also differentially expressed between the concussion cases and the uninjured controls. miR-4510 and miR-193a-3p were differentially expressed in individuals with concussion and were correlated with both sleep and psychological symptoms. Additional research is needed to confirm the temporal relationship between alterations in these miRNAs, concussion occurrence, and onset of sleep and/or psychological symptoms.

Dissonant approaches for measuring persisting symptoms after concussion (PSaC) make it difficult to predict who will experience prolonged symptoms. We sought to identify medical and sociodemographic characteristics associated with symptom burden and assess how such factors shape symptom evolution and PSaC classification after mild traumatic brain injury (mTBI). This analysis involved 1947 Post‐Concussion Symptom Scale reports from 1117 youths (11–21 years): 380 with mTBI, 737 without mTBI (194 healthy non‐mTBI; 543 non‐mTBI with underlying medical conditions). Multivariate regressions were used to assess the relationship of symptom burden with medical and sociodemographic factors among non‐mTBI youths, and interrogate how these factors impacted longitudinal symptom burden among mTBI participants. PSaC rates were characterized in the mTBI group 30 days after injury using 5 definitions: symptom burden, symptom severity, simple change in symptom burden and severity, and response to, ‘What percent of normal do you feel?’ In the absence of mTBI (i.e. baseline), symptom burden was associated with female sex, neuropsychiatric history, BMI and orthopaedic injury. In the 30 days following mTBI, age, sex and neuropsychiatric history were associated with symptom burden. Smaller household size, sports participation and parent education were protective. Rates of PSaC were 14.7–18.9%, displaying 84% agreement across the 5 definitions. However, PSaC misclassification was high among non‐mTBI youths with underlying medical conditions (37.2–50.6%), especially with the single‐item screener. Medical and sociodemographic factors affect concussion symptom reports and influence PSaC rates. A single‐item screener for PSaC may be useful but risks over‐detection among certain youths.