Post-traumatic headache: the use of the sport concussion assessment tool (SCAT-3) as a predictor of post-concussion recovery

Traumatic brain injury (TBI) is a major cause of disability. The estimated 5.3 million Americans living with TBI-related disability face numerous challenges in their efforts to return to a full and productive life [1]. Post-traumatic headache is one of the most prevalent post-concussion sequelae (up to 89% of patients suffer from it), is one of the longest lasting post-concussion symptoms, causes significant morbidity, and might be associated with slower neurocognitive recovery [2‐5].

Despite the plethora of complaints associated with traumatic brain injury and headache being the most prevalent complaint, there are no standard guidelines or recommendations for post-traumatic headache evaluation and management [6, 7]. Although patients who suffered from severe TBI might become aware of their post-traumatic headache more than seven days after the accident once their mental status improve and the post-traumatic headache severity of patients who suffered from mild TBI improves over time, the optimal timing for post-traumatic headache assessment has not been evaluated yet [8, 9]. However, there are validated tools to evaluate post-concussion signs and symptoms in athletes, which are now being implemented for use in non-athlete populations. The Sport Concussion Assessment Tool (SCAT-3) was initially developed at the Second International Conference on Concussion in Sport in Prague in 2004 by combining existing tools such as the Glasgow coma scale (GCS), modified Maddocks Score, modified Post-Concussion Symptom Scale (PCSS), mechanism of injury and background information, Standardized Assessment of Concussion (SAC), and examinations of the neck, balance, and coordination [10, 11]. The GCS is an assessment of patient’s levels of consciousness [12]. The Maddocks Score is an assessment of athletes’ orientation after traumatic brain injury [13]. The PCSS is a self-reported measure of 21 symptoms (representing four categories: somatic, cognitive, emotional, and sleep) weighted on a 7-point Likert severity scale [14]. The SAC is a rapid evaluation of patients’ orientation, concentration, and immediate and delayed memory [15]. Thus, the SCAT-3 arose from expert consensus and combines both the post-concussion symptom checklist and a quick set of examination tasks to assess some of the signs often comorbid with posttraumatic headaches such as cognitive or balance deficits. A longitudinal study of high school and college athletes indicates that the acute SCAT-3 symptom severity score is the most sensitive and significant predictor of recovery from SRC and hence of functional impairment and service utilization [16]. The King-Devick test (K-D) requires saccadic eye movements, language and other cognitive processes to perform [17]. However, a limitation of the SCAT-3 and the K-D is that they were designed for athletes, while the primary cause of TBIs in patients seen in the Emergency Department are falls, being struck by/against, motor vehicle accidents, and assaults [18]. Athletes tend to underreport their concussion and thus may not be representative of the general population [19]. Thus, additional study is required to determine the usefulness of the standardized concussion evaluation tools in non-athlete patients to screen for those who might benefit most from aggressive post-traumatic headache management to help with post-concussion syndrome recovery and prevent pain sensitization.

Since the SCAT-3 symptom severity score is the most significant predictor of recovery from SRC in athletes and since headache is the most common post-TBI symptom and is included in the SCAT-3 symptom severity score, we hypothesize that the SCAT-3 symptom severity score could be used to evaluate post-traumatic headache in athletes and non-athletes. We conducted two-tailed Student’s t-test of the SCAT-3 symptom severity scores between headache-free patients and post-traumatic headache patients with a criterion for significance of p < 0.05). Little is known about the non-athlete post-TBI recovery prognostic indicators [20]. Assessing a potential relationship between SCAT-3 scores and post-traumatic headache pain intensity and frequency (defined as daily versus not daily) might help plan post-traumatic headache management and prognostic recovery based on the SCAT-3 symptom severity scores. We sought to 1) assess the relationship between headache intensity, frequency, and concussion screening scores (SCAT-3 and King-Devick [K-D]), 2) compare these scores and post-traumatic headache of patients with and without SRC, and 3) determine whether the reported SCAT-3 concussion symptoms have a relationship to the formal assessments in our heterogeneous sample of patients.

In this heterogeneous population who presented to an outpatient multidisciplinary concussion center (heterogeneous in terms of age, gender, and mechanisms of injury), we have several findings regarding the use of SCAT-3 in post-traumatic headache patients: 1. The presence and frequency of post-traumatic headache are associated with the SCAT-3 symptom severity score, especially with the somatic SCAT-3 symptoms. 2. The SCAT-3 symptom scores might be a useful tool for neurologists and headache specialists as they incorporate balance, concentration, and memory. The self-reported symptoms of balance, concentration, and memory correlate with the formal SCAT-3 assessments and subscores in our sample.

As indicated above, our sample population is heterogeneous in terms of age, gender, and mechanism of injury. A little more than a third suffered from sport-related concussions. Although the data on the gender-specific prevalence of TBI is somewhat contradictory, it is interesting that TBI was more prevalent in women than men in our sample of non-athlete civilian patients [28, 29]. In our sample, men had a higher rate of sport-related concussion while women tended to have other mechanisms of injury. In our sample, a prior history of neither headache nor concussions is associated with the prevalence and intensity of posttraumatic headaches, but our sample size for the patients with a prior history of concussion is not large enough for adequate power. This is in contradiction with reports that found a history of mild TBI is associated with a higher prevalence of migraine-like headaches [30]. and a history of prior TBIs has also been found to correlate with post-traumatic headache prevalence [23].

Both post-traumatic headache prevalence and frequency are associated with higher SCAT-3 symptom severity and SCAT-3 symptom scores. This is important because the SCAT-3 symptom severity score is the most sensitive and significant predictor of recovery from SRC [19, 31]. Post-traumatic headache patients have higher SCAT-3 symptom scores than headache-free patients regardless of whether the “headache” symptom is counted when calculating the SCAT-3 symptom scores. Interestingly, when comparing SCAT-3 symptoms for each of the four symptom categories individually, post-traumatic headache patients have significantly higher somatic SCAT-3 symptoms, but a similar occurrence and severity of sleep, emotional, and cognitive symptoms compared to the headache-free patients. Of note, the somatic symptoms of the SCAT-3 assessment (e.g. blurred vision, photophobia, phonophobia, neck pain, nausea, vomiting) are often associated with headaches. This might explain the association between post-traumatic headache and the somatic SCAT-3 symptoms. Another explanation would be that post-traumatic headache patients have higher somatization than headache-free patients. Prior literature suggests that somatization is associated with higher SCAT-3 symptom severity scores and longer recovery [32].

When comparing the subjective and objective assessments, in our heterogeneous sample of patients with diverse mechanisms of injury who presented to the concussion center, the SCAT-3 self-reported symptoms of balance, concentration, and memory correlate with the modified BESS and SAC scores, which supports the use of these scores among our heterogeneous population of concussion center patients who are not necessarily athletes.

The SRC patients in our sample have less severe post-traumatic headache and lower SCAT-3 symptom scores than the non-sport-related concussion patients, for the same SAC scores –which is a marker of effort- [24]. The difference in headache severity and SCAT-3 symptom scores might be at least partially confounded by the significant difference in age, gender, BESS, and/or K-D scores between the two groups. Indeed, a prior study on concussion assessments in an outpatient heterogeneous population like ours showed that older age and female gender are associated with worse SCAT-3 symptom scores [23]. Interestingly, our group with the worse SCAT-3 symptom scores (the non-sport related concussion group) is older with a higher proportion of women than the SRC group. Further studies with a larger sample would be helpful to adjust for age, gender, BESS, and K-D scores.

Although we need further studies to compare post-traumatic headache and the SCAT-3 symptoms between SRC patients and patients with concussions unrelated to sports, the SCAT-3 seems to be a good tool to evaluate post-traumatic headache in the general population of post-concussion patients.

Our study had several limitations. One limitation was that we had a high number of patients with headache, thus making it hard to compare post-traumatic headache patients with headache-free patients. A second limitation is the retrospective nature of the study and its missing data. The missing data is likely multifactorial including the introduction of the different concussion tests at different time and some patients feeling too symptomatic to perform full testing. The missing data was handled with pairwise deletion. A third limitation was the wide range of time to assessment which could contribute to recall bias in that patients with mild headaches closer to the time of concussion onset may have under-reported a prior history of headaches. Very few patients were assessed within the first seven days and it is known that symptoms gradually resolve by post-concussion day 7 on average in SRCs in athletes [33]. A fourth limitation was that, if a concussion was severe, patients with a prior history of headaches may have dismissed their prior history of headaches and denied having a prior history of headaches. Having severe symptoms acutely is a risk factor for severe symptoms later on, and vice versa, having severe symptoms later on suggests having had severe symptoms early on. A fifth limitation was that our study suffers from selection bias. Our study participants willingly sought care at our tertiary referral center, so their TBI symptoms might be worse than for those who did not seek treatment. Some also might have sought care somewhere else first, which might have confounded our results. Interestingly, a large portion of our patients were evaluated at our center more than three months after their concussions and the patients who presented to our center within three months of their concussions reported less severe headaches than the patients who presented to our center more than three months after their concussions, which might illustrate a selection for patients with more severe symptoms in our sample. Patients who present earlier tend to present because they are worried they hit their heads. Patients who present on the later side are concerned about their persistent symptoms. A sixth limitation is that we do not have pre-injury scores. Concussion scores are most helpful when compared pre- and post-injury in the same patients. A seventh limitation is due to the fact that registry data was used for a report of headache following concussion (what we term “post-traumatic headache”)-not physician assessments using the International Classification of Headache Disorders (ICHD) 3 beta criteria. Therefore, we do not know if the headaches occurred within seven days of the injury or being off of medication which could have masked headaches from the injury [34]. Also, we are limited by the data from the assessments and the physician notes, and thus do not have the mean number of headache days.

It would be helpful to compare post-traumatic headache and SCAT-3 symptom scores between athletes and non-athletes adjusting for age, gender, BESS, and K-D scores. Since the SCAT-3 symptom severity score is an important prognostic indicator of post-concussion recovery and is associated with post-traumatic headache, it would be interesting to evaluate the impact of early post-traumatic headache management on post-concussion recovery and SCAT-3 symptom severity score.

Most of our post-concussion patients suffered from post-traumatic headache. Both post-traumatic headache prevalence and frequency are associated with higher SCAT-3 symptom severity and symptom scores. The association between post-traumatic headache and SCAT-3 symptom severity score stems from higher SCAT-3 somatic symptom scores in post-traumatic headache patients, which might reflect either a higher degree of somatization in post-traumatic headache patients or an overlap between the SCAT-3 somatic symptoms and symptoms often associated with headaches. SCAT-3 somatic symptoms are important in concussion management and recovery prognosis. Since the SCAT-3 symptom severity score seems to be the single most important predictor for post-concussion recovery and since SCAT-3 symptom severity score is associated with post-traumatic headache, the SCAT-3 symptom severity score seems to be a promising standardized tool to identify the post-concussion patients who might benefit from additional follow-up and management for post-traumatic headache and associated symptoms. It was shown that standardized measurement of post-concussive symptoms including headache may enhance clinical management of concussion patients. Although the SCAT-3 was designed for athletes, our study suggests that it might be helpful in the general population of concussion patients to assess for post-traumatic headache and associated symptoms.