Abstract
Post-traumatic headache (PTH) is among the most prevalent of the secondary headache disorders. As currently classified by the International Classification of Headache Disorders, 2nd edition (ICHD-2), PTH must start within 7 days of a mild, moderate, or severe traumatic brain injury (TBI). PTH transitions from an acute subtype to a chronic subtype 3 months after the injury (1). No other defining clinical characteristics set PTH apart from other primary or secondary headache disorders. In fact, most PTHs appear to have a phenotype indistinguishable from primary headache disorders and may respond to the same treatments as primary headache disorders (2,3).
It seems altogether intuitive that headaches would occur acutely in the majority of individuals after a mild TBI (or concussion). More vexing, and controversial, is the link between head trauma and the subsequent development of chronic headaches (4,5). PTHs persist beyond the acute period in a significant proportion of patients. Up to 60% of TBI patients have chronic headaches persisting for up to 12 months (4,6). That chronic PTHs develop more often in mild versus moderate to severe TBI, in cases with impending litigation or in the setting of analgesic medication overuse, suggests a role for factors independent of trauma in the process of headache chronification (4,7,8). While each (or all) of these factors may be important in the genesis of chronic headaches in individual cases, there is a lack of evidence supporting them as unifying hypotheses to explain the spectrum of chronic PTHs in civilian and military patient populations. A better understanding of the processes leading to headache chronification after brain trauma is needed.
Over the past 3 years, large epidemiological studies of headaches in military, civilian, and veteran populations with head trauma have been completed (9–11), providing a valuable opportunity to better understand and classify PTHs. In this issue, Lucas et al. (10) report the findings of a multi-center, longitudinal study of headaches in a population of 378 patients with moderate to severe TBI. The subjects underwent a baseline headache evaluation shortly after TBI and were then followed by telephone interview at 3, 6, and 12 months after injury to longitudinally assess headache incidence and headache characteristics. To our knowledge, this is one of the largest longitudinal studies of headaches in TBI patients.
Many interesting findings emerge from this study, parts of which have been published previously (10, 11). The cumulative headache incidence in the study cohort was 71% over 12 months. Forty-three percent of patients had headaches at the baseline assessment shortly after TBI and this proportion was stable over the next 12 months. There were no differences in headache incidence between mild, moderate, and severe TBI patients. Eighteen percent of all patients had a history of headaches prior to TBI with 57% of these being migraine or probable migraine. Similar to previous studies, pre-traumatic headache and female sex were risk factors for reporting headaches after TBI.
Migraine and probable migraine were the most common headache types after trauma, accounting for 52% of headaches at baseline and 60% of headaches 6 months after injury. Migrainous headaches were 2–3-fold more common than tension-type headaches (TTHs; 7–21%) at all time points. The high prevalence of migraine observed by Lucas et al. (10) sharply contrasts with older studies which suggested a higher prevalence of TTH after TBI, but are similar to the findings of recent studies in military and veteran populations which found migraine in 60 to over 90% of cases (2,3,9,12). Comparing PTH studies and study populations can be difficult, but the accumulation of evidence now strongly supports migraine as the predominant PTH phenotype (9–12).
One of the clinically relevant findings is that PTHs possessing migraine features were more likely to persist beyond the acute period for up to 12 months. Moreover, migraine headaches made up the largest proportion of high-frequency headaches, occurring multiple times per week, a year after TBI. Thus, post-traumatic migraine identifies patients at high risk of developing frequent, chronic headaches. These individuals may benefit from early, aggressive headache management and closer clinical follow-up to prevent headache chronification. Another clinically important finding was that a significant minority of PTHs (30%) did not resemble migraine, TTH, or cervicogenic headache at the baseline assessment and were thereby considered unclassifiable. Selecting headache treatments can be difficult when headaches do not conform to a recognizable phenotype. The proportion of unclassifiable headaches decreased over time. Thus, following patients over time may be needed in some cases before the PTH phenotype clearly emerges.
It should be noted that the study population was comprised largely of civilian patients with moderate to severe TBI caused by a motor vehicle accident (MVA). The results may not be generalizable to mild head trauma, military populations, or trauma induced by mechanisms other than a MVA. Additionally, the study did not assess for multiple headache types in the same subject which could have resulted in an underestimation of some headache types such as TTH.
A significant proportion, 28% of patients in this study, had headaches with onset beyond the acute, post-traumatic period. According to the current classification system, these headaches would not qualify as PTHs because they began after the 7-day window following trauma. This is consistent with the findings in a large cohort of USA army soldiers of whom 19% had onset of headaches, mostly migraines, between 1 week and 1 month after a concussion (9).
When one considers that the incidence of migraine in an American male under the age of 30 is only 0.021% per month, the spontaneous onset of primary migraine cannot possibly account for the significant proportion of individuals who develop headaches more than a week after brain injury (13). While it is difficult to know if headaches with delayed onset after trauma are due to the direct or indirect effects of trauma, as discussed in the next paragraph, the initial precipitating event must be presumed to be head trauma in many cases. Thus, the 1-week latency for headache onset required by the ICHD-2 criteria for PTH needs to be re-evaluated, as the emerging clinical data support the likelihood of a longer latency in a subset of cases. A latency of 1 month for headache onset following concussion may be more appropriate. For moderate and severe TBI, where altered consciousness, post-traumatic amnesia, and other more pressing medical conditions may limit the evaluation of PTH, a period of up to 3 months could be considered. Another possibility is to divide PTHs into those that begin immediately after head trauma and those that have a delayed onset.
Whether headaches that develop beyond the acute period are attributable to TBI is a well-traveled debate. Discussion of classification and attribution of PTHs are inextricably linked. The current study and other observational clinical studies cannot stand alone as evidence that PTHs are directly attributable to the trauma itself. Lucas et al. (10) note some of the potential pathophysiological links between TBI and PTHs. Migraine as the most frequent PTH phenotype potentially links head trauma and a primary headache syndrome known to be associated with neurophysiological changes. Mounting evidence, including changes on diffusion tensor magnetic resonance imaging following even mild or blast-related TBI, suggests that the traumatic effects extend beyond what can be detected on clinical examination or routine imaging modalities (14). For some TBI patients, the aftermath of trauma includes the development of co-morbid psychiatric disorders or medication overuse secondary to the treatment of headaches, musculoskeletal, or neuropathic pain. Previous studies have noted headache syndromes similar to PTHs occurring following injuries not related to the head and neck, so-called ‘traumatic event headaches’, supporting the contention that they are stress-induced (4,15).
A ‘one size fits all’ approach to PTHs occurring across the spectrum of head injury severity, trauma mechanisms, settings, and headache phenotypes is destined for failure. Head trauma is complex and heterogeneous. For example, blast-related trauma can have primary, secondary, and tertiary effects (16). Patients with TBI are also complex. The biopsychosocial formulation of a soldier with an explosive blast-related concussion on the battlefield and a civilian with a moderate TBI after a MVA are vastly different. The debate over attribution, while important, need not completely overshadow further study of PTH. Lucas et al. (10) have provided us with compelling evidence, regardless of the attributable cause, that PTHs are frequent, often migrainous, and persistent over time in a significant proportion of civilian patients with TBI.
Clinical trials are needed to determine the most effective treatments for PTHs of all types. To facilitate future clinical studies, a re-evaluation of the current PTH classification system is necessary. A classification system that further subclassifies PTHs into clinically meaningful subgroups according to headache phenotype, delayed versus immediate onset of headache, and possibly mechanism of injury may be warranted. This will improve the quality and applicability of clinical trials needed to ultimately better treat our patients.