Abstract
The objective of this study was to document the resolution rate of mild Traumatic Brain Injury (mTBI) symptoms at various time points in a nonathletic adult population and identify prognostic factors influencing recovery. Sixteen prospective cohort studies were included, focusing on participants aged 18–65 with acute mTBI, followed for a minimum of 1 month. The recovery criterion was the resolution of symptoms not attributable to pre-existing conditions. Risk of bias was assessed using the Quality in Prognostic Studies tool, with most studies rated as moderate risk, highlighting variability in methodological rigor. Symptom resolution was reported in 49.0% to 69.5% of patients at 1 month, 40.8% to 84.4% at 3 months, 38.3% to 72.2% at 6 months, and 58.1 to 68.3% at 12 months. These findings emphasize the first 6 months as a critical period for evaluating the risk of symptom chronicity. The most commonly reported prognostic factors was baseline symptom severity, including higher intensity of symptoms such as headaches, nausea, and dizziness, as well as elevated scores on validated symptom scales. Psychiatric history, such as pre-existing anxiety or depression, was also a significant predictor of prolonged symptoms. Biomarkers, including NSE and S-100B levels, and reduced blood-derived neurotrophic factors, were associated with poorer recovery at 6 months. Demographic factors, including age, gender, and education level, showed mixed results. While some studies associated female gender, older age, and lower education with poorer recovery, others found no significant correlations. These discrepancies highlight the complexity of mTBI prognosis. Overall, more than half of patients recover within 6 months, but persistent symptoms can have a profound impact on quality of life and functional status. Identifying patients at higher risk of prolonged recovery is crucial for targeted management strategies, emphasizing the importance of individualized, evidence-based care in mTBI populations.
Introduction
The World Health Organization defines mild traumatic brain injury (mTBI) as an acute brain injury resulting from a transfer of energy from an external source to the skull and underlying structures. 1 In Canada, approximately 100–300/100,000 cases of mTBI are reported each year. 2 Following a direct or indirect blow to the head, mTBI diagnosis is considered when a score between 13 and 15 on the Glasgow Coma Scale 3 is recorded with the presence of one or more of the following symptoms: (1) loss of consciousness of <30 min; (2) post-traumatic amnesia of <24 h; (3) an altered mental state at the time of the accident, confusion, disorientation, etc.; and (4) transient neurological deficits. 1 The diagnosis criteria have been recently revised and now also include the presence of two acute symptoms (e.g., headache, dizziness, and nausea) combined with one acute clinical sign (e.g., vestibulo-ocular impairment and balance impairment). 4
Full recovery from symptoms following mTBI usually takes an average of 10 days to 3 months. 5 However, some factors can negatively influence recovery, i.e., persistence of symptoms beyond 3 months. Individual factors such as age, gender, or neuropsychological history have an impact on recovery, but the number, severity, and duration of symptoms present initially and a few days after trauma are the most important predictors of recovery. 5 Persistent symptoms can have a negative impact on the daily functioning of individuals and have significant functional consequences, thus making it an important issue in terms of public health. 6 It is therefore interesting to identify factors that can negatively influence the prognosis in order to help clinicians in their management strategies.
A recent overview of systematic reviews 6 included 12 systematic reviews on prognosis of recovery from mTBI and mentioned that normal resolution of symptoms occurs within 3 months, with symptoms transitioning to the chronic state beyond this period defined as prolonged recovery from mTBI. Financial compensation awarded to individuals following mTBI, absenteeism from work due to the incident, and post-traumatic symptoms of nausea or memory impairment were associated with slower recovery or persistence of symptoms. Studies from other authors suggest that the gender of the individual is a mixed factor; it was not a prognostic factor influencing post-traumatic recovery according to some, 7 but others 8 demonstrated that age and female gender are vulnerability factors to the persistence of postconcussion symptoms (PCS).
There is also variability in conclusions about the time required to recover from an mTBI. A systematic review by Caroll et al. reported that the recovery time for PCS is between 3 and 12 months for the adult population with age and gender as important prognostic factors, 2 whereas another study by Machamer et al. claims that most adults recover within 3 months. 9 There is also variability in the proportion of people who remain symptomatic over time. Indeed, the same study by Machamer et al. determined that 78% of participants in the mTBI cohort reported at least one symptom that was new or worse at 3 months than before the trauma. 9 Nevertheless, according to a study by Skandsen et al., only 20.8% reported having postconcussion syndrome or at least three symptoms rated as moderate (score of ≥3 or a total of ≥13 on the British Columbia Postconcussion Symptom Inventory questionnaire). 10 The many differences between the definition of recovery time and prognostic factors create an issue in the effective management of people with mTBI for clinicians.
The Amsterdam International Consensus Statement on Concussion in Sports 2023 is a universally renowned reference for sport-related concussion. 11 It defined prolonged recovery as persistence of symptoms over a month or more and describes that most athletes of all ages (93%) have a full recovery at 1 month. 12 Recovery time is within 1 month for children, adolescents, and adults, with an average of 19.8 days. There are factors that tend to lead to poorer recoveries, such as histories of migraines and anxiety, a worsening condition, the presence of psychological stress, or social factors. 12 The topic of prognosis of mTBI in athletes is therefore well known and widely described in the literature. However, it is possible that nonathlete population sustaining an mTBI recovers differently than athlete population because they sustained different type of injuries (e.g., intimate violence, and motor vehicle accident) or because they are more likely to have comorbidities. The access to care could also be different in athlete versus nonathlete population.
Considering the high variability previously established and the fact that the prognosis for the nonathletic population affected by mTBI is significantly less well documented than that for athletes, it is necessary to clarify the prognosis of the nonathletic population. Thus, the first objective of this systematic review was to determine the prevalence of symptom resolution in adults who have undergone mTBI in a nonsport setting, at different time points. The assumption regarding the recovery of these individuals is that a majority will recover at 3 months. The second objective was to explore the factors that influence recovery of symptoms.
Methodology
A literature search was conducted for relevant articles in the Embase, OVID Medline, and CINHAL databases. The keywords used are presented in Supplementary Table S1. Articles published before May 31st, 2024, were considered. From the identified articles, the titles and abstracts were first reviewed by the four authors to determine their relevance to the research question. After this initial screening, the full texts of the remaining articles were evaluated, and the decision to include the article or not was made according to the inclusion and exclusion criteria. Covidence software was used for the screening and selection phases. All authors participated to the screening phase with a moderate level of agreement (Cohen’s Kappa = 0.41–0.48).
Eligible articles met the following inclusion criteria: (1) the study population was adults between 18 and 65 years of age, with Traumatic Brain Injury (TBI) diagnosis, of which at least 95% were mTBI; (2) prospective cohort studies with follow-up of at least 1 month; (3) more than 50% of participants had undergone their mTBI in a nonsporting setting and/or should not be considered athletes as the most recent consensus statement on sport-related concussion included articles with a population of more than 50% of concussion sustained in sports, 13 it was the intent to retrieve different sources articles; (4) the initial assessment was completed within 2 weeks of the mTBI; and (5) solely PCS were used as a measure of recovery. Note that for studies with some moderate or severe TBI, data were analyzed with mTBI participants only when it was possible. Similarly, articles with a proportion of patients under 18 years of age or over 65 years of age were analyzed considering only patients in the targeted adult age group (18–65 years old). Articles were excluded if they met the following criteria: 1 patients with associated conditions such as other traumatic pathologies or other neurological disorders 2 ; studies with other types of designs (randomized clinical trials and retrospective studies) 3 ; and studies on military populations.
In data extraction, the proportion of patients considered recovered during the various predetermined follow-ups and the prognostic factors were collected. Other collected characteristics included study location (country); year of publication; sample size; average age; proportion of male or female participants; mean time from mTBI to intake; possible prognostic factors measured at baseline; number and timing of follow-ups; and the authors’ definition of “recovery” and symptom-related outcome measures at each follow-up (Table 1). Given the significant variability in the definitions of recovery from TBI reported in the articles, it was necessary to establish the definition of recovery used in this systematic review. Therefore, consistent with the systematic review published by Putukian et al., 12 and considering that this study is primarily interested in symptom resolution, the recovery criterion was defined as the time to the absence of symptoms produced by the current mTBI that cannot be attributed to a pre-existing injury. This definition aligns with the definition of persistent PCS presented by Lagacé-Legendre et al., 30 which defines it as “the presence of any symptom that cannot be attributed to a pre-existing condition and that appeared within hours of a head injury, and is still present every day 3 months after the injury, and impacts at least one area of a person’s life.”
Characteristics of Included Studies
mTBI, mild traumatic brain injury; N, number of participants; NR, not reported.
Subsequently, each article was subjected to a risk of bias assessment by at least two authors, using the Quality in Prognostic Studies (QUIPS) Risk of Bias Assessment Tool for Prognostic Factor Studies. 31 This tool divides the assessment of risk of bias into six domains of potential bias: study participation, study attrition, measures of prognostic factors, outcome measures, study confounding factors, and statistical analysis. Each domain was assigned a low, moderate, or high risk of bias rating based on whether or not the characteristic was present in the study. Following the assessment of the six domains, a judgment was made on the overall risk of bias of the study using the same ratings.
By combining the results of each study, while also taking into consideration the determined risk of bias as well as the characteristics of each study (e.g., the study population), conclusions could be drawn to determine the most significant prognostic factors and the proportions of patients with resolved symptoms at different times of recovery following mTBI. To support this, the collected data were analyzed using the mean and standard deviation (SD) of population size of the included studies, as well as the average number of follow-ups per study, the range, and total proportion of asymptomatic patients at 1, 3, 6, and 12 months. A qualitative analysis was also performed to describe the prognostic factors reported by the different studies and considered significant.
Results
The literature search identified 1162 eligible studies (603 in Embase, 423 in Medline, and 136 in Cinhal; Supplementary Table S1). After removal of duplicates, 721 studies met the search criteria. Following revision of titles and abstracts, 117 articles were reviewed in full and finally, 16 articles14–29 were selected according to the inclusion and exclusion criteria (Fig. 1).
Study flow chart.
Characteristics of included studies
Table 1 presents the characteristics of the 16 included studies. Half of the studies included a cohort of mTBI participants, while the other half included a comparison group. Ten studies included a population of <115 participants, five studies included between 150 and 500 participants, and only one study included more than 500 participants (Table 1). On average, 175.1 (SD = 163.1) participants with mTBI took part in the studies, demonstrating a very wide distribution of populations. Follow-up time points vary between 2 and 52 weeks. The studies performed an average of 2.4 follow-ups. Regarding symptom measurements at follow-up, the most commonly used tools were the Rivermead Postconcussion Questionnaire (RPQ; 10 studies) and the Brief Symptom Inventory (BSI; 4 studies). The Beck Depression Inventory was used in three studies, while the following tests were used by two: the Global Severity Index, a pain scale (visual analogue scale or numeric scale), the Post-Traumatic Stress Disorder Checklist, and the Symptom Severity Score (SSS) from the Sport Concussion Assessment Tool-2. Ten other tools were also used once, either independently or in combination with an aforementioned test (Table 1).
Risk of bias assessment
The risk of bias of each study is presented in Table 2. Ten studies14–17,19,22,24,26,27 were classified as moderate risk, i.e., had no major bias but were of high or moderate risk in enough domains to raise doubts about the bias present in the study and the reliability of their results. Four studies18,21,23,25 were categorized as low risk of bias, which required for the studies to have taken steps to minimize the risk of bias in most of the domains of the QUIPS tool. In contrast, two studies28,29 had a “high” overall risk of bias, having either a major bias in one or more domains, or “moderate” or “high” risk in a majority of domains.
Risk of Bias Assessment of Studies Included with the QUIPS Tool
QUIPS, Quality in Prognostic Studies.
In addition, some areas of the QUIPS tool were more frequently identified as “high risk” or “low risk.” First, most studies were found to have a low risk in the “Prognostic Factor Measurement” (14 out of 16 studies) and “Outcome Measurement” (15 out of 16 studies) domains. Indeed, the use of valid and reliable tools was widespread in the studies. In both cases, data collection was conducted in person, by telephone, or through self-administered questionnaires. In addition, clear definition of the prognostic factors or outcome measures was consistently used. In contrast, high risk was often noted for the domains “Cconfounding factors” (10 out of 16 studies) and “Study attrition” (9 out of 16 studies) as data on participants who dropped out during follow-up were often missing. Few studies described attempts to collect information from these participants or the reasons for their dropout. Only a few articles reported key characteristics of lost participants or noted that there were no significant differences between those who completed the study and those who did not. Regarding attrition, although the proportion of participants who completed the studies was sometimes satisfactory, several pieces of information on participants attrition were often missing. It was also rare to find a description of attempts to collect information from them or the reasons for the loss to follow-up.
Proportions of recovered patients
As different studies used different outcome measures, it was not possible to pool the data across studies. The results of each study with proportion of patients recovered at 1, 3, 6, and 12 months are shown in Table 3.
Proportions of Participants Who Recovered at 1, 3, 6, and 12 Months with the Overall Risk of Bias of the Study Represented with Colors
Risk of bias legend: green = low; yellow = moderate; red = high.
Four studies reported a proportion of asymptomatic patients at 1 month, and proportions of asymptomatic participants ranged from 49.0% to 69.5%. Six studies reported a proportion of asymptomatic patients at 3 months, with a range of participants with full recovery between 40.8% and 84.4%. Subsequently, six studies presented prognostic data at 6 months and the proportion of recovered patients at 6 months ranged from 38.3% to 72.2%. Finally, at 12 months post-mTBI, two studies reported data to establish a proportion of recovered patients ranging between 58.1% and 68.3%.
Prognostic factors
A wide variety of factors were reported to be predictive of symptom persistence (Table 1). The most common factor, reported by four studies, was baseline symptomatology, i.e., the number and intensity of symptoms present at baseline. However, this was expressed differently in different articles. Predictive symptoms can therefore vary; these include high pain intensity, 14 headache, nausea, or dizziness, 17 or high scores on symptom questionnaires such as the RPQ or BSI.19,24 Notably, all four studies that identified symptom severity as a prognostic factor measured it within the first week postinjury. In contrast, six studies16,20,21,23,25,26 did not assess symptoms at baseline, and two studies18,29 did not report prognostic factors. Three additional studies15,22,28 measured symptoms at baseline but focused on other specific prognostic factors. Only one other study 27 assessed symptom severity within 72 h postinjury and did not find it to be a prognostic factor.
Another common factor is female gender, reported by three studies which identified that a larger proportion of female patients had persistent symptoms at 324,27 and 6 months. 20 Education is also a demographic characteristic mentioned by some authors. Indeed, two studies report that low education is a factor of poorer prognosis at 12 14 and 6 months. 20 However, it is important to note that these two factors, female gender and education, are challenged by the study of Losoi et al. Rather, their results indicate that at 12 months, there was no difference in age, education, or gender between patients with mild PCS and those without symptoms. 21
Older age is another factor on which there is disagreement between authors. As mentioned, Losoi et al. determined that age is not a significant factor in poor prognosis at 12 months. 21 In contrast, three other studies reported the opposite result.20,24,27
History of psychiatric problems prior to the mTBI was also identified. Patients with preaccident psychiatric problems were more likely to complain of PCS, at 3 27 and 6 months. 20 The incidence of stressful events in a person’s life, as assessed by the Stressful Life Events questionnaire, was also a predictor of scores on the BSI depression and anxiety subscales at 3 months. 25
Blood biomarkers were part of the initial assessment in three studies. One considered the presence of NSE and S-100B markers in serum within 6 h of mTBI, 17 while another investigated the presence of blood-derived neurotrophic factors (BDNFs) within the first 24 h. 26 Both conclude that blood biomarker levels influence the 6-month prognosis, i.e., elevated NSE and S-100B or low BDNF levels are predictive of poorer recovery. The third article analyzed 420 proteins in blood samples and concluded that 31, 43, 15, and 26 proteins were significantly associated with the poor recovery of neuropsychological symptoms at 6 months when measured at <72 h, 1 week, 1 month, and 3 months after the injury, respectively. 22 However, no unique and common protein biomarker was identified. Finally, functional magnetic resonance imaging (fMRI) was a baseline measure used by two articles, which both determined that decreased connectivity between various and nonconsistent brain regions on fMRI is a predisposing factor for poorer recovery at 3 28 and 12 months. 14
Discussion
This systematic review shows that symptom resolution has a heterogeneous prevalence according to the different studies, but the total proportion of recovered patients varies little between 1 and 6 months. Prognostic factors for poor recovery were identified, but some are not universally accepted.
The intensity and number of symptoms following an mTBI and the individual’s history of psychiatric problems are the most significant predictors of delayed recovery reported by the largest number of mostly moderate-quality studies.14,17,19,20,24,27 There is no clear consensus on the impact of gender, low education, and advanced age on recovery. The conclusions of the four moderate-risk articles citing these factors (three articles report female gender,20,24,27 three report age,20,24,27 and two report education14,20) are challenged by one article of low risk of bias. 21 However, this study has only 75 participants with mTBI, which is not enough to contradict the set of four articles which total a much larger number of participants (n = 580 mTBI). It can therefore be assumed that age, female gender, and education probably have an impact on the prognosis of symptom recovery, but the literature is not yet sufficient on this subject. Remarkably, most included studies did not examine symptom severity as prognostic factor, either because they did not measure symptoms at baseline or because their primary objectives focused on other predictors. This suggests that symptom severity and intensity might have emerged as significant prognostic factors in more than just four studies had they been assessed more consistently. Additionally, all four studies linking symptom severity to prognosis measured it within the first week, suggesting timing may influence its predictive value. In contrast, one study measuring symptom severity within 72 h postinjury did not find it to be a prognostic factor, further highlighting the need for future research on the influence of early symptom assessment on recovery predictions. Regarding biomarkers and fMRI, studies seem to indicate that they are significant factors, but there were not enough high-quality studies to draw a strong conclusion.14,17,22,26,28
The prevalence of complete symptom resolution in patients is surprising. Indeed, as shown in Table 3, homogeneity in the total proportion of asymptomatic individuals can be observed. In other words, there is no recovery time that stands out where most participants had a full recovery.
At 1 month, the risk of bias of the total proportion can be considered moderate since two out of four have a moderate risk and the other two are opposites, i.e., low and high.15,21,22,29 Thus, because the proportion is based only on four articles totaling a relatively small population of 455 participants, our confidence level for the 1-month prognosis is rather low.
Subsequently, at 3 months, the risk of bias of the calculated proportion was judged to be low to moderate. Even though the article with the largest number of participants has a low risk of bias, two are of high risk, and the other are of moderate risk of bias.15,22–24,28,29 However, the total population of 665 participants across six articles allows for a possibly more robust conclusion. That said, our level of certainty about the strength of the results is moderate.
At 6 months, since the majority of articles have a moderate risk of bias and the only article with a low risk of bias has the smallest population among the six articles, the risk of bias of this grouping of results is considered moderate.15,17,20–22,25,26 With its 774 participants, this group has the largest total population among the four follow-up times, which increases the strength of its results. Despite this, our level of confidence is moderate given the quality of the articles included in this grouping.
Only two articles looked at the 12-month follow-up. This group is at moderate risk of bias because the study with the majority of the 165 participants (attrition rate of 0% 14 and 20% 21 ) in this grouping is at moderate risk of bias.14,21 Our level of confidence in this grouping is low due to the few studies and small sample that limits generalization to a larger mTBI population.
Several hypotheses can explain the similarity in the total proportions reported at 1, 3, 6, and 12 months. First, the large range in the proportions of recovered participants reported at each follow-up is one possible explanation (Table 3). For example, at 6 months, the study with the highest proportion of asymptomatic patients reports almost double the proportion of the lowest rate. However, this variability is not visible in the sum of recovered participants in this grouping. In addition, all articles included in this systematic review do not use the same methodology. The total proportion of recovered patients at each time point was obtained using data from different studies. Additionally, there may be differences in what each study considers as “asymptomatic,” given that they do not all use the same tools. For example, what is considered a “recovered” patient when measured with the SSS is not necessarily the same as with the RPQ. Differences in study designs may therefore have contributed to the similarity of total proportions.
Another hypothesis to explain these results could be an attrition bias due to a high lost to follow-up of recovered individuals. 32 Indeed, it is possible that when participants no longer had symptoms, they stopped to participate in the study and did not return to study follow-ups. In addition, as mentioned above, study attrition is one of the domains most consistently rated as high risk of bias of the included studies. Therefore, it is possible that there were lost data, which had a direct impact on the results. Indeed, a high attrition phenomenon could explain why the prevalence of recovered patients seems to stagnate between 1 and 6 months, and only slightly increase at 12 months.
Comparison with the literature
Several other studies have commented on the prognosis of symptom recovery following mTBI. A study by Cancelliere et al. showed that the presence of symptoms between 3 and 6 months using the RPQ or similar tools ranged from 18% to 31%. 32 Symptom prevalence values vary depending on the use of a strict PCS definition (moderate PCS; 18%) compared with a broader definition (mild PCS; 31%). This represents a proportion of asymptomatic individuals that varies between 82% and 69%. These results are not consistent with those of the current study, which found a much lower proportion of asymptomatic patients at 3 and 6 months.
The differences in results can be explained by several reasons. Differences in methodology, attrition, mechanism and severity of trauma, and eligibility criteria can be some examples. Indeed, this systematic review 32 used a population of adults presenting to emergency departments or trauma centers but does not specify what the exact population is, i.e., whether it is a nonathletic population. A large population of 11,742 participants across 43 studies were evaluated, which allows for greater generalization of the results.
The systematic review of the Amsterdam International Consensus Statement on Concussion in Sports 2023, which only included sport-related mTBI, found the same predictors of delayed recovery as the current study, namely the intensity and number of initial symptoms as the strongest predictors. 12 Also, in agreement with this review, age and gender do not appear to be a significant factor in the prolonged recovery of individuals. It therefore seems that the prognostic factors predicting poorer recovery are similar between the athletic and nonathletic populations. However, the consensus’ conclusions on the prevalence of asymptomatic participants highly differ from those of this systematic review. Indeed, they suggest that symptom recovery is 14.0 days, 12 highlighting the large difference in recovery between the two populations.
For studies conducted with an athletic population, other reasons may explain the differences in results. It can be hypothesized that sports teams with protocols post-mTBI would offer better management to athletes and thus a better recovery time, especially when they are supported by multidisciplinary teams to ensure return to play. The articles included in our literature review did not discuss protocols or interventions to optimize recovery time in the nonathletic population with mTBI. Thus, this may explain the difference in recovery time expressed by the consensus and that determined by this systematic review. It is also possible that younger healthy and fit sports individuals, with an active lifestyle and minimal number of comorbid health problems, were quicker to recover than the nonsport adult with a higher likelihood of comorbidities.
Clinical relevance of findings
The identification of the proportion of asymptomatic patients during the various post-mTBI follow-ups also provides interesting insights. Indeed, although the literature commonly reports that normal recovery is <3 months, 5 our findings suggest that about half of patients may have symptoms that persist for up to 6 months and up to 12 months for about 40% of them. These data can be used by clinicians to reassure patients that recovery extending to more than 3 months is a possible reality. Indeed, the presence of persistent symptoms following mTBI can be anxiety provoking in nearly 30% of patients, and therefore it may be relevant to present them with this type of statistic to explain that recovery varies greatly from one person to another. 33
In addition, by highlighting the differences in the recovery of an athletic patient to that of a nonathlete patient, it is possible to see that the prognosis is not the same. Therefore, clinicians need to be able to manage patient expectations accordingly, by adjusting their message to the type of patient they are treating.
Prognosis after mTBI is important to help clinicians identify individuals at risk of a delayed recovery and adapt their management to optimize their recovery. Therefore, the identification of factors of delayed recovery such as the intensity and high number of initial symptoms and the presence of a psychiatric history should be considered by clinicians particularly during the initial questionnaire.
Limitations
First, an important limitation of this study is that no recognized method has been used to determine the levels of evidence for our findings. Indeed, the use of a method such as the Grading of Recommendations Assessment, Development and Evaluation (GRADE) method could have supported the evidence found in relation to prognostic factors predicting poorer recovery after mTBI. The GRADE approach is used to define the quality of the evidence in terms of the certainty of the estimation of the effects of the prognostic factor(s) in question. 34 As this step was not performed in this systematic review, it is not possible to determine the certainty of the evidence regarding the risk posed by the presence of a prognostic factor on recovery from mTBI.
Subsequently, the choice to use only PCS as a measure of recovery is also a limitation of this study. The presence of persistent symptoms following a concussion is certainly an important issue to consider, but recovery from mTBI can also be considered in a broader sense. Indeed, several studies use a different recovery measure and therefore had to be eliminated from this study. This is the case for several studies in the Transforming Research and Clinical Knowledge in Traumatic Brain Injury cohort, a longitudinal study between 2014 and 2018 aimed at demystifying the diagnosis and prognosis following a concussion. These studies are of high methodological quality and often include large cohorts of patients; as a result, their results are often of good quality. For example, three studies had to be excluded as they use the Extended Glasgow Outcome Scale as an outcome measure, and this tool does not measure symptoms but rather mTBI-related disabilities.35–37 It might have been relevant to include both disabilities and PCS to have a more holistic view of recovery after mTBI.
In addition, as previously noted, a common weakness of the included articles is participant attrition, which was often assessed with a high risk of bias; this raises doubts about the quality of our results. It is possible that some patients may not have completed their follow-up if they considered themselves recovered. This would strongly bias the results, particularly on the proportions of patients recovered at the various follow-ups established in our study. In fact, the overall risk of bias of studies could also be a possible source of error. To establish the proportion of recovered patients, only 11 of the 16 included studies had data that could be used, with most being of moderate risk of bias. Thus, it is a significant possibility that some biases have crept in and skewed the results. Finally, the heterogeneity of included studies, particularly in terms of methodologies, sample sizes, and outcome measures, makes direct comparisons challenging and prevents the performance of a meta-analysis. The discrepancies observed in predictors such as female gender, older age, and lower education may stem from these methodological differences, including variations in sample size and study design.
Suggestions for future
There are a few possible solutions to improve scientific research in this field. First, it would be interesting to use a clear and consistent definition of recovery. Indeed, as reported in the Amsterdam Consensus, the wide variety of definitions of recovery in the different articles is a significant limitation to the conclusions drawn. 10 Standardizing the definition of recovery would allow for more accurate and unanimous conclusions related to recovery time and prognostic factors. In addition, the use of the same follow-up points over time, for example to always assess the number of people recovered at 1, 3, and 6 months, could make it possible to distinguish exactly the precise time of recovery of individuals since the studies would be easily comparable with each other. Subsequently, it would be relevant for more high-quality longitudinal and prospective studies to be carried out, to minimize the risk of bias and thus optimize the quality of the results. In addition, the sample size should be large enough and, most importantly, have low attrition to ensure the applicability of their findings to the target population.
Conclusion
This study concludes that the proportions of patients with complete resolution of symptoms after mTBI in the nonathletic population range from 55.6% to 61.8% between 1 and 12 months and reach a minimum proportion of 52.3% at 6 months. Thus, this suggests that about half of patients will become asymptomatic in the first 6 months following mTBI. The most significant prognostic factors in predicting poorer recovery in this population are the intensity and number of symptoms as well as psychological history. Further studies are needed to determine with certainty the impact of other prognostic factors reported in the literature. Individualized patient care should be prioritized to maximize the recovery of individuals with mTBI who have risk factors for prolonged recovery.
Transparency, Rigor, and Reproducibility
This study demonstrates a strong commitment to transparency by thoroughly documenting the methodology used for the systematic review. The search strategy, which involved the Embase, OVID Medline, and CINHAL databases, is clearly described, along with the specific search strategy employed (Supplementary Table S1). The inclusion and exclusion criteria, as well as the step-by-step screening process using Covidence software, are explicitly detailed to ensure that the selection of studies is reproducible. Additionally, the risk of bias for each included study was systematically evaluated using the QUIPS Risk of Bias Assessment Tool, with domain-specific and overall ratings provided in Table 2. Comprehensive data tables and supplementary materials further enhance transparency by allowing other researchers to scrutinize the underlying evidence and decisions made throughout the review.
The rigor of this review is evident in the adherence to established guidelines for systematic reviews and the use of validated tools to assess recovery outcomes and prognostic factors. Data extraction focused on standardized outcome measures, such as the Rivermead Post-Concussion Questionnaire and the Brief Symptom Inventory, ensuring the reliability of the findings. Covidence software was employed to manage the screening and selection phases, reducing variability, and improving the consistency of study inclusion decisions. The assessment of risk of bias across six domains using a validated framework ensured that methodological limitations were carefully considered in the synthesis of results. This methodological rigor underpins the validity and reliability of the study’s conclusions about recovery trajectories and prognostic factors following mTBI.
To facilitate reproducibility, the study provides detailed descriptions of the search strategy, inclusion/exclusion criteria, and data extraction processes, enabling other researchers to replicate the review. A consistent definition of recovery, aligned with established literature, was used to standardize outcome assessments across included studies. The analytical methods used to calculate pooled recovery proportions and identify significant prognostic factors are transparently reported, with results presented in tables and supplementary materials. By offering clear documentation and access to key data, this study ensures that its findings can be independently verified and utilized as a foundation for further research in mTBI recovery and prognosis.
Footnotes
Authors’ Contributions
V.D., G.L., and R.P.: study selection, data extraction, risk of bias assessment, writing. P.L.: Overview of all phases, data analysis, writing.
Author Disclosure Statement
The authors declare that they have no competing interests. Although P.L. is one of the owners of a private Concussion Clinic, the clinic and the authors did not financially support this project, nor did they gain or lose financially from the publication of this article. The project was completely independent from the clinic. P.L. is also remunerated to provide postgraduate course on mTBI which course do not support this project, nor did gain or lose financially from the publication of this article.
Funding Information
This study was not funded.
Abbreviations Used
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
