Abstract
Background:
Prognosticating recovery times for individual athletes with a concussion remains a challenge for health care providers. Several preinjury and postinjury factors have been proposed to be predictive of prolonged return-to-play (RTP) times, but the data in this area are still sparse.
Purposes:
This study aimed to identify risk factors associated with prolonged recovery times and determine which are most predictive of prolonged recovery times in a head-to-head comparison.
Study Design:
Case-control study; Level of evidence, 3.
Methods:
All concussions occurring between September 2017 and August 2020 at a single National Collegiate Athletic Association Division I institution were reviewed and included in this study. Preinjury modifiers including age, sex, sport, concussion history, and past medical problems were collected from the electronic medical records. Postinjury modifiers analyzed included initial and follow-up Sport Concussion Assessment Tool 5th Edition scores, vestibular evaluation findings, and eye tracking results.
Results:
A total of 159 athletes and 187 concussion cases were included. Preinjury factors that were correlated with prolonged RTP times included a history of concussions (P = .015), a history of migraines (P = .013), and whether an athlete participated in an individual sport (P = .009). Postinjury factors correlated with prolonged RTP times included the total number (P = .020) and severity (P = .023) of symptoms as well as abnormal Vestibular Ocular Motor Screening findings (P = .002). Overall, 6 different symptoms (balance problems, difficulty concentrating, light sensitivity, drowsiness, fatigue/low energy, and difficulty remembering) were significantly correlated with prolonged RTP times. The study also found that the number and severity of symptoms were additive in a dose-dependent fashion. On multivariable analysis of all these factors, a history of concussions was found to be the most predictive of prolonged RTP times, while participation in an individual sport had the largest effect on recovery times.
Conclusion:
Several preinjury and postinjury risk factors were identified as being correlated with prolonged recovery times. Many of these risk factors were also found to be additive in nature. This information provides clinicians with a valuable tool in prognosticating and estimating recovery times for athletes. The study also revealed that athletes participating in individual sports had longer RTP times compared with athletes in team sports, which is a novel finding that requires further research.
Keywords
Concussions are common among athletes and have received much attention because of concerns about the long-term effect that they may have on cognitive function. Much of the focus surrounding concussion management has been on the risk of return to play (RTP) before complete resolution of a concussion. Premature RTP and subsequent head reinjuries can lead to second impact syndrome and other catastrophic consequences.7,12
Estimating the time to RTP after a concussion remains challenging. Most athletes will demonstrate clinical recovery from a concussion within 2 to 4 weeks.6,8,18 However, some athletes take longer, and a small subset can have symptoms for months after the initial injury.4,8,10,16-18 Identifying athletes who are at risk for delayed recovery is crucial, not only for prognostication and early intervention but also to mitigate the emotional and physical toll associated with prolonged RTP times.
In terms of preinjury factors, a history of concussions has been associated with prolonged recovery.8,18 However, much of this same literature has reported inconsistent results regarding the correlation of female sex, a history of migraines, or a history of attention-deficit/hyperactivity disorder (ADHD) with prolonged RTP times after a concussion.1,4,9,10,12,13,16,18 Furthermore, several systematic reviews6,9,13 have investigated these as well as other risk factors associated with prolonged recovery from a concussion but with inconsistent findings. The recent 2019 American Medical Society for Sports Medicine position statement on concussions specifically called for further research to identify risk factors for prolonged recovery and identified this as a key area for future studies. 7
The most consistent postinjury predictor of prolonged RTP times is the number and severity of acute symptoms on initial evaluation.6-8,18 However, the literature is mixed regarding whether specific symptoms can be predictive of prolonged recovery. Many studies have sought to identify specific self-reported symptoms or symptom patterns that are correlated with prolonged RTP times, but findings have been inconsistent.3,8,10,15,18 Recent literature has also examined the effect of concussions on the vestibular system. The Vestibular Ocular Motor Screening (VOMS) is a validated, easy-to-perform assessment that is increasingly used in the diagnosis and management of concussions.6,11 Although studies have proven it to be a valuable diagnostic tool, studies on its ability to predict prolonged RTP times are lacking. However, preliminary studies have shown that abnormal results on these vestibular tests, as well as symptom provocation with the VOMS, have been linked with longer RTP times.2,17
There is a clear need to better synthesize the evidence regarding preinjury and postinjury factors that can lead to prolonged RTP times for athletes who sustain a concussion. Furthermore, it is important to understand which risk factors are most predictive of prolonged RTP times and whether these risk factors are additive. This study is one of the first to analyze a comprehensive collection of preinjury and postinjury factors in a single cohort. The aims of this study were to (1) identify which risk factors are associated with prolonged RTP times, (2) assess whether these risk factors are additive, and (3) determine which risk factors are most predictive of prolonged RTP times by conducting a head-to-head comparison in the same cohort.
Methods
All concussions occurring between September 2017 and August 2020 at a single National Collegiate Athletic Association Division I institution were retrospectively reviewed and included in this study. This study was approved by the Stanford University Institutional Review Board (IRB) and granted a waiver of informed consent. All athletes who sustained a concussion underwent a standard postinjury protocol consisting of the Sport Concussion Assessment Tool 5th Edition (SCAT5), 5 a vestibular evaluation by a trained physical therapist, and formal eye tracking testing via a virtual reality headset using EYE-SYNC technology (a US Food and Drug Administration–approved assessment tool to track eye movement impairment, that analyzes smooth pursuit, saccades, and the vestibulo-ocular reflex; SyncThink). It should be noted that all concussions, not only concussions that occurred while playing sports, were included in this study. The same evaluation and treatment protocol was used for all concussions, regardless of whether the injury occurred during sports. All athletes were evaluated by a team physician as well as a neurosurgeon who specializes in sports concussion management. In some cases, additional consultations with a neuropsychologist, neuro-optometrist, sports psychologist, or neurologist were utilized. During the postinjury phase, all athletes who had abnormal vestibular evaluation findings received ongoing vestibular therapy and were monitored by sports medicine staff. Patients who did not demonstrate any vestibular abnormalities did not continue this rehabilitation program. Vestibular evaluation findings were considered abnormal if there were objective results (eg, nystagmus or poor balance) or if the athlete endorsed subjective worsening of headaches, nausea, dizziness, or fogginess. In addition, athletes were encouraged to start early symptom-limited physical activity, which typically began 24 to 48 hours after the injury. This consisted of stationary bicycling with a goal heart rate of 150 beats per minute, barring an exacerbation of symptoms. All athletes completed a standard stepwise RTP protocol before being cleared to return to competition. Final full clearance was determined by a sports medicine physician using input from the various consultants and was based on the resolution of symptoms, vestibulo-ocular abnormalities, and eye tracking abnormalities.
Preinjury modifiers including age, sex, sport, concussion history, and past medical problems (eg, a history of migraines, motion sickness, or ADHD) were collected from the electronic medical records. Postinjury modifiers analyzed included initial and follow-up SCAT5 scores, vestibular evaluation findings, and eye tracking results. The SCAT5 included symptom scoring, the Standardized Assessment of Concussion, neurological screening, and the modified Balance Error Scoring System (mBESS). Recovery time was defined as days from the injury until final clearance by the team physician.
Descriptive statistics are presented as counts and percentages for categorical variables and as means and standard deviations or as medians and interquartile ranges for continuous variables. Individual multilevel regression models were used to find variables associated with prolonged RTP times, accounting for athletes who sustained >1 concussion. Variables that were independently associated with RTP times (P < .05) were subsequently added to a multilevel, multivariable linear regression model, allowing for adjustments among multiple predictor variables. Post hoc standardized parameter estimates were calculated to compare the effect sizes of each predictor. All analyses were conducted using RStudio Version 1.1.456 (RStudio) using complete cases for relevant variables and a 2-sided level of significance of .05.
Results
A total of 159 athletes with 187 concussions that occurred during the study period were included. Any case without a listed RTP time was excluded (n = 3). Of the athletes, 52% were male and 48% were female, representing 19 different sports (Tables 1 and 2). The median time missed because of a concussion was 16 days. The most time missed was 158 days.
Concussion Characteristics a
IQR, interquartile range.
Concussion Incidence by Sport
When analyzing preinjury modifiers, we observed that each previous concussion increased the estimated RTP time by 3.7 days (95% CI, 0.7-6.7 days; P = .015). In addition, a history of migraines increased the RTP time by an estimated 10.7 days (95% CI, 2.4-19.1 days; P = .013). Participation in an individual sport was also associated with an estimated 11.1 days longer RTP time compared with participation in team sports (95% CI, 2.9-19.3 days; P = .009). Other preinjury factors, including a history of ADHD and a history of motion sickness, were not associated with prolonged recovery. There was no difference in RTP times when comparing sex or age. Furthermore, injury occurrence during practice versus competition and during sports versus outside of sports was not associated with increased RTP times (Table 3).
Preinjury Variables and Their Effect on RTP a
ADHD, attention-deficit/hyperactivity disorder; RTP, return to play.
Statistical significance (P < .05).
When analyzing postinjury modifiers, we observed that both the total number of symptoms (P = .020) and the severity of symptoms (P = .023) were associated with longer RTP times. For every additional symptom reported, the RTP time increased by 0.7 days (95% CI, 0.1-1.3 days), and for every additional symptom severity point, the RTP time increased by 0.2 days (95% CI, 0.0-0.4 days). Furthermore, the severity of 6 specific symptoms was significantly associated with prolonged RTP times: balance problems, difficulty concentrating, light sensitivity, drowsiness, fatigue/low energy, and difficulty remembering. Difficulty concentrating was found to be the most statistically significant risk factor (P = .003), but the number of days missed was greatest in athletes with balance problems (3.4 days for each 1-point increase in severity) (Table 4). Other objective measures from the SCAT5, including the change in delayed 10-item recall, the change in counting digits backward, and the mBESS score, were not associated with prolonged RTP times (Table 5).
Symptoms and Their Effect on RTP a
RTP, return to play.
Statistical significance (P < .05).
Eye Tracking and Clinical Variables and Their Effect on RTP a
mBESS, modified Balance Error Scoring System; RTP, return to play.
An abnormal VOMS finding was associated with increased RTP times by an estimated 11.3 days (95% CI, 4.3-18.4 days; P = .002). When further analyzing the separate components of the VOMS we observed that, abnormal scores on vertical and horizontal saccades, vertical and horizontal visual motion sensitivity, horizontal vestibulo-ocular reflex, and convergence were all significantly associated with longer RTP times (Table 6).
Abnormal VOMS Findings and Their Effect on RTP a
RTP, return to play; VOMS, Vestibular Ocular Motor Screening.
There were no abnormal findings.
Statistical significance (P < .05).
The variables that were identified as being significantly associated with prolonged RTP times were then run together in a multivariable model. Based on these results, 3 factors were identified to be most strongly associated with prolonged RTP times: (1) an increased number of previous concussions (P = .002), (2) abnormal vertical saccades on the VOMS (P = .004), and (3) participation in an individual sport versus a team sport (P = .008). The effect size in terms of the standardized parameter estimate was greatest for athletes who participated in an individual sport (19.7 days longer compared with athletes participating in team sports) (Table 7).
Multivariable Model Comparing All Statistically Significant Risk Factors and Their Effect on RTP a
RTP, return to play.
Statistical significance (P < .05).
Discussion
This study furthers our understanding of the risk factors for prolonged recovery from a concussion by comprehensively analyzing multiple preinjury and postinjury variables in a single cohort. In terms of preinjury factors, a history of migraines and a history of concussions were correlated with prolonged recovery, which is consistent with conclusions drawn by multiple previous studies and systematic reviews.1,7,8,10 However, sex and a history of ADHD had no significant effect on recovery time. This is contrary to several studies that have reported prolonged recovery in female athletes4,9,10 or in patients with a history of ADHD,1,8,16 but these findings have been inconsistent.12-14,18 Thus, our results suggest, as do several of the aforementioned studies, that female sex and a history of ADHD are not consistently associated with prolonged RTP times after a concussion.
Interestingly, this study identified that athletes who participated in individual sports were more likely to have prolonged RTP times compared with athletes who participated in team sports. In fact, participating in an individual sport had the largest effect on recovery time, leading to an estimated 19.7 added days until recovery compared with participating in team sports on average. The underlying cause for this novel finding is unclear and warrants further investigation. It is possible that athletes in individual sports may share characteristics that predispose them to prolonged recovery from a concussion. It is also possible that social pressures that arise from participation in a team sport may influence an athlete’s recovery. Although RTP protocols are designed to incorporate objective assessments such as physical examinations, repeat neurocognitive testing, and performance metrics, they still rely heavily on an athlete’s subjective reporting of symptoms. This leaves these protocols vulnerable to various external variables that could affect recovery time. Further studies are needed to understand this difference.
In terms of postinjury factors, the study results are consistent with previous research showing that the total number of symptoms and total symptom score were correlated with prolonged RTP times in collegiate athletes.7,8,14,18 This study also identified that the number and severity of symptoms were additive in a dose-dependent fashion. Each additional symptom resulted in a mean of 0.7 days added to the recovery time, and each additional symptom severity point resulted in a mean of 0.2 days added to the recovery time. Thus, if a patient endorsed 20 symptoms, his or her RTP time would be an estimated 7 days longer than that of someone who endorsed 10 symptoms on one’s initial postinjury SCAT5. A patient with a symptom severity score of 30 would be predicted to return to play 2 days later than someone with a symptom severity score of 20. This dose-dependent relationship has the potential to improve clinical care by enabling clinicians to more accurately estimate differences in RTP times.
This study was able to identify 6 specific symptoms that were correlated with prolonged RTP times. Of these symptoms, difficulty concentrating was the most statistically significant symptom in terms of correlation with prolonged recovery. However, balance problems had the largest magnitude of effect on RTP times. Contrary to previous studies, we did not find symptoms such as headaches, dizziness, and mood dysphoria to be correlated with delayed recovery in this cohort.3,7,8,14
Although the VOMS has been widely adopted as a diagnostic tool for concussions, very little research has investigated whether it can be used as a predictor of recovery time.6,11,17 This study showed that if VOMS findings were abnormal, the RTP time increased by an estimated 11.3 days (P = .002). Furthermore, 6 elements of the VOMS were significantly correlated with prolonged RTP times. These findings suggest that the VOMS should be included in the diagnosis and management of all concussions, as it not only adds to diagnostic accuracy but also provides a more objective prognostic measure of which patients are at risk for prolonged RTP times.
The final aim of this study was to quantify which of the above risk factors were most predictive of prolonged RTP times, as previous studies have not compared such a comprehensive list of preinjury and postinjury risk factors. By comparing these factors in a multivariable model, this study identified 3 factors that were most predictive of prolonged RTP times: (1) a history of concussions, (2) vertical saccades on the VOMS, and (3) participation in an individual sport versus team sport. Of these 3 factors, participation in an individual sport had the largest effect on recovery time (19.7 days longer compared with participation in team sports). Although the importance of previous concussions has been well documented in the literature, the difference in RTP times between team and individual sports is novel and warrants further studies.
The results of this study can help clinicians by enabling them to provide more accurate estimates of RTP times to athletes. This not only helps to guide expectations for the patient, family, coaching staff, and medical care team but can also allow the medical team to direct additional resources toward athletes who are at an increased risk for a prolonged RTP time.
There were some limitations to this study that may have affected the results. The study was performed as a retrospective chart review, and as a result, there was a small amount of missing data for some athletes. Although many preinjury and postinjury factors were investigated, this study could not account for all potential confounding variables, such as socioeconomic status. It should also be noted that a general limitation of studying concussion metrics is that symptoms and several aspects of concussion evaluations are self-reported. Further, although a standardized protocol was used for all athletes, there may have been some variation among different examiners and clinicians when conducting the VOMS or making RTP clearance decisions. Moreover, certain subgroup analyses may have been underpowered because of small sample sizes, such as those involving patients with a history of migraines. Last, although overall this was a large cohort that was studied, data were collected from a very specific population (collegiate athletes) with access to many resources (eg, neuro-optometrists, neuropsychiatrists, and dedicated athletic trainers) and thus may not be representative of other populations.
Conclusion
This study helps to advance our understanding of how different preinjury and postinjury factors may influence recovery times in collegiate athletes who sustain a concussion. By comparing a large range of risk factors in a single cohort, we have identified which risk factors are most strongly predictive of prolonged RTP times and which may be additive. Future research should seek to replicate the results of this study by evaluating the effects of multiple preinjury and postinjury risk factors on RTP times. Specific areas for future studies include the effect of sport type on recovery time. Additionally, future studies could investigate whether early targeted interventions are effective at reducing prolonged RTP times associated with the risk factors identified in the present study.
Footnotes
Submitted July 6, 2021; accepted January 13, 2022.
One or more of the authors has declared the following potential conflict of interest or source of funding: G.D.A. has received consulting fees from Cytonics, Fidia Pharma, RubiconMD, and Sideline Sports Doc; other financial or material support from Arthrex and Stryker; and royalties from Orthofix Medical. He is also an unpaid consultant for TeachAids and holds stock or stock options in AxGen and Cytonics. AOSSM checks author disclosures against the Open Payments Database (OPD). AOSSM has not conducted an independent investigation on the OPD and disclaims any liability or responsibility relating thereto.
