Different clinical phenotypes of persistent post-traumatic headache exhibit distinct sensory profiles

Abstract

Introduction:

Persistent post-traumatic headache remains a poorly understood clinical entity. Although there are currently no accepted therapies for persistent post-traumatic headache, its clinical symptoms, which primarily resemble those of migraine or tension-type headache, often serve to guide treatment. However, evidence-based justification for this treatment approach remains lacking given the paucity of knowledge regarding the characteristics of these two major persistent post-traumatic headache phenotypes and their etiology.

Methods:

We compared clinical features and quantitative sensory testing profiles between two distinct cohorts of persistent post-traumatic headache subjects that exhibited symptoms resembling either migraine (n=15) or tension-type headache (n=13), as well as to headache-free subjects that had suffered traumatic brain injury (n=19), and to healthy controls (n=10). We aimed to determine whether the two persistent post-traumatic headache subgroups could be discriminated based on additional clinical features, distinct quantitative sensory testing profiles, or the interaction of pain severity with the level of post-traumatic stress disorder.

Results:

Persistent post-traumatic headache subjects with migraine-like symptoms reported that bright light and focused attention aggravated their pain, while stress and nervousness were reported to aggravate the headache in subjects with tension-type headache-like symptoms. Quietness was better in alleviating migraine-like persistent post-traumatic headache, while anti-inflammatory medications provided better relief in tension-type headache-like persistent post-traumatic headache. The two persistent post-traumatic headache subgroups exhibited distinct quantitative sensory testing profiles with subjects exhibiting tension-type headache-like persistent post-traumatic headache displaying a more pronounced cephalic and extracephalic thermal hypoalgesia that was accompanied by cephalic mechanical hyperalgesia. While both persistent post-traumatic headache subgroups had high levels of post-traumatic stress disorder, there was a positive correlation with pain severity in subjects with tension-type headache-like symptoms, but a negative correlation in subjects with migraine-like symptoms.

Conclusions:

Distinct persistent post-traumatic headache symptoms and quantitative sensory testing profiles may be linked to different etiologies, potentially involving various levels of neuropathic and inflammatory pain, and if confirmed in a larger cohort, could be used to further characterize and differentiate between persistent post-traumatic headache subgroups in studies aimed to improve treatment.

Introduction

Traumatic brain injury (TBI) gives rise to numerous long-term physical and psychological outcomes. One of the most commonly occurring, and often most debilitating, outcomes is post-traumatic headache (1,2). Post-traumatic headache often resolves within 1–3 months following the head trauma, but in many cases can persist beyond this time frame. Although reported incidences of persistent post-traumatic headache (PPTH) vary, recent longitudinal studies point to cumulative incidence of 70-90% during the first year after the injury (3,4).

PPTH has significant impact on quality of life (5,6), in large part due to limited responsiveness to available treatments (7,8). One major hurdle for improving treatment of PPTH is the poor knowledge of its precise characteristics and underlying etiology. Attempts made to classify PPTH using symptom-based clinical features suggest resemblance mainly to migraine and tension-type headache (TTH) (8). While these nosological entities do not necessarily implicate shared mechanisms between PPTH and primary headaches (8,9), the presence of two major PPTH clinical phenotypes raises the question of whether they are etiologically different. A better understanding of such differences thus could be essential for guiding treatment, and for the design of clinical trials to test novel therapies.

PPTH pathophysiology could involve damage to the cephalic tissue and its sensory innervation, as well as to the central neural pathways that transmit and modulate pain. Assessment of changes in the sensory profile of subjects with chronic pain has been useful in delineating potential underlying peripheral and central mechanisms (e.g. (10,11)). We discovered previously, using systematic quantitative sensory testing (QST), major alterations in the sensory profiles and pain modulation in a mixed cohort of individuals with PPTH when compared to headache-free TBI subjects, and to healthy controls (12,13). In an effort to shed further light on this disabling condition, the objective of the current study was to examine potential differences between subjects with PPTH that exhibit clinical symptoms reminiscent of migraine or TTH. Here, we present data showing that these two PPTH subgroups involve additional different clinical features, distinct sensory profiles, as well opposing correlations with the level of post-traumatic stress disorder (PTSD).

Methods

Subjects

Given the higher incidence of head injuries and TBI in males (14), and that the bulk of clinical data related to PTH primarily reflect finding in males (particularly in military populations), the current study included only male subjects. Data was collected from 70 individuals, comprising three major groups: a) Individuals after a single TBI with PPTH (n=31; mean age 39.2±11 years), b) headache-free TBI subjects (n=19; 36.7±10 years), and c) pain-free, healthy individuals (n=20; 37.1±7 years). TBI subjects were recruited from the Department of Head Injury, and the Headache Unit of the Department of Neurology at Sheba Medical Center, Tel-Hashomer, as well as from the Neuropsychological Unit for Treatment and Rehabilitation under Tel-Aviv Sourasky Medical Center. Healthy controls were recruited among the workers of the two Medical Centers and the Tel-Aviv University. Recruitment of TBI subjects was initiated by screening the computerized lists of patients for those who met the inclusion criteria (see below). Then, letters were sent randomly to every third person on these lists explaining the aims of the study. The letters were followed by phone calls three weeks later explaining the study in more detail, and further verifying the inclusion/exclusion criteria. The eligible patients who agreed to participate in the study were invited to arrive for a single testing session at their own convenience. Recruitment of healthy controls was initiated by advertisements posted in the campuses, and followed by phone calls to those who responded, in which the study was explained and inclusion criteria were verified. The eligible control subjects who agreed to participate in the study were invited to arrive for a single testing session at their own convenience. Inclusion criteria for all subjects with TBI were: a) A diagnosis of TBI as determined by a clinical evaluation and CT and/or MRI scans; b) a minimum duration of 12 months since injury; c) no evidence of neurological and systemic diseases (e.g. diabetes, Parkinson’s disease); d) no pre-existing headache condition, including medication-overuse headache; e) no prior psychiatric illnesses; f) no current involvement in litigation; g) normal communication and understanding capabilities, as assessed by clinical examination and by the routine neuropsychological tests that were administered at the time of admission to each unit. All subjects with TBI were living in households in the community and none were using wheelchairs or walking aids. The majority of subjects with PPTH included in the study sustained mild TBI (Table 1). PPTH was diagnosed by a headache specialist according to the International Classification of Headache Disorders 3rd edition, beta version (ICHD-3 beta) criteria (15). Clinical and imaging examinations excluded secondary disorders such as haematoma, cerebral vein thrombosis, cerebral hemorrhage, and epilepsy, as well as medication-overuse headache. TBI subjects with diagnosis of a whiplash injury were also excluded. Data regarding diagnoses and collateral anamneses were retrieved from the subjects’ medical records available at the units. The study was approved by the Sheba Medical Center, the Tel-Aviv Sourasky Medical Center and the Tel-Aviv University institutional review boards. Informed consent was obtained from all subjects according to the Declaration of Helsinki after they received full explanation of the goals and protocols of the study.

Table 1.

Characteristics of the TBI subgroups.

	Migraine-like PPTH	TTH-like PPTH	Headache-free TBI
Number of patients	15	13	19
Age (years, mean, SD)	42.3 (13)	40.1 (10)	36.7 (10)
Time after injury (years, mean, SD)	13.2 (10)	12.6 (9)	11.3 (12)
Cause of injury (n, %)
MVA	10 (66.6)	7 (53.8)	11 (57.9)
GS	1 (6.6)	2 (15.4)	4 (21.1)
FOH	1 (6.6)	3 (23.1)	2 (10.5)
HIT	2 (13.3)	1 (7.7)	2 (10.5)
Severity of injury (n, %)
Mild	11 (73.3)	12 (92.3)	12 (63.2)
Moderate	2 (13.3)	–	2 (10.5)
Severe	2 (13.3)	1 (7.7)	5 (56.3)
Past involvement in lawsuits (y) (n, %)	7 (46.6)	9 (69.2)	10 (52.6)
Neck injury (yes) (n, %)	5 (33.3)	4 (30.8)	2 (10.5)
PTSD symptomatology (mean, SD)	55.6 (10)	53.9 (10)	36.4 (11) **

PPTH: Persistent post traumatic headache; MVA: motor vehicle accidents; GS: gun shot; FOH: fall of height; HIT: direct hit to the head; PTSD: posttraumatic stress disorder;

^**p<0.01 unpaired t-test between headache-free TBI patients and PPTH subjects.

PPTH classification

PPTH subjects were divided into subgroups of headache types according to the following clinical features: Intensity, location in the head, sidedness (unilateral/bilateral), quality, temporal properties and aggravating and alleviating factors. The decision for allocating subjects to a subgroup was based on the number of features associated with each headache type. The classification yielded three subgroups. Subjects diagnosed as suffering from PPTH with migraine-like symptoms (n=15) had attacks lasting several hours of unilateral pain of moderate to severe intensity, throbbing/pounding, drilling and piercing in quality, which was aggravated by physical activity, bright light and/or loud noise, or strong odors. Among these subjects, seven reported experiencing an aura before the onset of an acute exacerbation of headache/migraine-like symptoms. Subjects diagnosed as suffering from PPTH with tension-type headache like symptoms (n=13) had bilateral pain of mild to moderate intensity, not pounding but pressing and dull in quality, which is aggravated by emotional stress and tension and without photophobia. The remaining three individuals presented with either mixed symptoms of migraine and TTH (n=2) or with cluster headache-like symptoms (n=1) and were not included in the statistical analysis.

Quantitative sensory testing equipment

Thermal stimulator: Heat stimuli were delivered using a Peltier-based computerized thermal stimulator (TSA II, Medoc Ltd., Ramat-Ishai, Israel), with a 3×3 cm contact probe. A passage of current through the Peltier element produces temperature changes at rates determined by an active feedback system. As soon as the target temperature is attained, probe temperature actively reverts to a preset adaptation temperature by passage of an inverse current. The probe was attached to the testing site by means of a Velcro band.

Pressure algometer: Pressure stimuli were delivered using a hand-held pressure algometer (Somedic Sales AB, Algometer type II, Sösdala, Sweden). The algometer has a built-in pressure transducer, electronics recording and display unit, power supply and subject-activated push button connected via a cable to the instrument. The principle of operation of the algometer is the exertion of a constantly increasing rate of pressure, the latter being monitored by a cursor presented on the display. The tip of the algometer that is pressed against the skin was 1 cm². The algometer was calibrated before each measuring day.

Von Frey filaments: Mechanical stimuli were applied with Semmes-Weinstein Monofilaments (Touch-Test^™, North Coast Medical, Inc, Morgan Hill, California, USA). The kit comprises 20 calibrated monofilaments, with sizes ranging between 1.65–6.65 units. Each filament is attached to a plastic holder. Vertical pressure applied with the handle induces a calibrated force ranging between 0.008–300 g, respectively.

Quantitative sensory testing protocols

Sensory testing was conducted at two distinct sites, a seemingly affected (cephalic) region and a seemingly unaffected, remote (extracephalic) region. For cephalic measurements, testing was performed over the temples and forehead in the site denoted by the subjects as being most painful. For extracephalic measurements, stimuli were administered to the volar aspect of the forearm. Testing of control subjects was performed at comparable regions.

Thermal thresholds: Thermal thresholds were measured with the thermal stimulator, using the Method of-Limits (16). For warmth and cold sensation thresholds (WST and CST, respectively) determination, subjects received four successive, increasing or decreasing stimuli, starting from an adaptation temperature of 32°C, at a rate of 2°C/sec with inter-stimulus intervals of 15 sec. The subject was asked to press a switch at the first warmth or cold sensation perceived, and the temperature at that point was recorded by the computer. WST and CST were the average reading of four successive stimuli of increasing or decreasing temperature. For heat-pain threshold (HPT) determination, subjects received four successive stimuli starting from an adaptation temperature of 32°C at a rate of 2°C/sec. A minimal interval of 30 sec was kept between successive stimuli in the same session, in order to avoid possible sensitization or desensitization of the cutaneous receptors. The subject was asked to press a switch at the first pain sensation perceived. HPT was the average reading of four successive stimuli of increasing temperature.

Mechanical thresholds: As opposed to thermal stimuli, which only activate cutaneous receptors, pressure stimuli also activate deep tissue mechanoreceptors and therefore have the advantage of revealing changes in deep tissue pain sensitivity (17,18). Pressure pain threshold (PPT) was measured with the pressure algometer using the modified method of limits. The tip of the pressure algometer probe was placed perpendicular over the skin. Gradual pressure was applied from a baseline intensity of 0 kPa at a rate of 30 kPa/second, with an inter-stimulus interval of 30 seconds. Each application was administered at a different location on the skin within a pre-determined area of 3×3 cm in order to avoid changes in the sensitivity of the skin. Subjects were instructed to press the switch at the first pain sensation perceived, thus ‘‘freezing’’ the display with the corresponding pressure reading and recording it. PPT was the averaged reading of three successive stimuli of increasing pressure (19).

Allodynia: Static mechanical allodynia was measured to assess centrally mediated pain hypersensitivity (20) and was determined using a single application of a Von Frey filament (filament no. 4.74; 6 g/force). The subject was asked to report the quality of sensation evoked by the stimulus. If the stimulus elicited painful sensation the subject was asked to rank its intensity on a visual analog scale (VAS) (21).

Questionnaires and additional data collection

Characteristics of the TBI: All subjects with TBI were interviewed regarding their injury, including the date and age at the time of injury, causes and circumstances, additional injuries incurred along with the TBI, as well as co-morbidities. Information obtained at these interviews was cross-checked with subjects’ medical records. The severity of injury was determined according to the Glasgow Coma Scale (GCS (22)).

Characteristics of PPTH: Subjects with PPTH were interviewed about its onset, duration, frequency, quality, cranial location, dynamic characteristics, use of medications, and alleviating and aggravating factors, as indicated above. In addition, subjects were asked to rate the intensity of their headache pain on the VAS (0, “no pain” to 10 “the most intense pain sensation imaginable”). All subjects with PPTH also completed the McGill pain questionnaire (MPQ) (23), which provides a quantitative evaluation of the patient’s pain experience and is also used to assess headache (24,25). Four quantitative parameters were derived: a) The pain rating index (PRI) – based on summing the values of the words chosen by the subject from a list of 64 pain descriptors, b) the number of words chosen (NWC) from that list, c) pain intensity at its least (Pleast), and pain intensity when it becomes worse (Pmost). The latter two parameters are scaled on a five word/number scale (1–5).

Post-Traumatic Stress Disorder (PTSD) scale: PTSD coexists with TBI, and is associated with chronic pain among individuals after TBI (6,26). The PTSD Inventory was used (27). This is a 17-item self-report scale based on the DSM-IV criteria for PTSD, which evaluates post-traumatic stress symptomatology. PTSD severity is calculated as the number of symptoms endorsed. Internal consistency calculated was high (α=0.88 in the present sample).

Testing procedures

Subjects were invited to a single testing session lasting about 3 hours. The experiments took place in a quiet room with an ambient temperature of 22±2°C. The subjects sat in a comfortable armchair with the forearm supported on a holder. Prior to sensory testing the subjects were interviewed with regard to their demographics, TBI, and their related PPTH symptoms and current pain level, and then completed the MPQ and the PTSD inventory. Before initiating the sensory testing, we confirmed that the subjects were not experiencing a headache attack, and their headache level allowed them to properly perform in the sensory testing. The subjects were then trained for the sensory testing, and underwent the actual testing subsequently, which included the measurement of cold threshold, warm threshold, heat-pain threshold, pressure-pain threshold, and allodynia. The different testing procedures were conducted in a random fashion. Note that at the time of sensory testing, the experimenter was blinded with regard to the PPTH sub-classification, which was created only during the data analysis phase.

Data analyses

Data were analyzed with the IBM SPSS statistics software (version 21) and R (version 3.5.1). The sample size was estimated a priori based on our previous experience, the expected difference between control and PPTH subjects in the mean values of HPT (a major outcome measure) and considering standard deviations. For α=0.05, and statistical power of 80%, the calculation yielded a sample size of 13 subjects. Average values of continuous variables are described by means and SD. Categorical variables are described by frequencies. One-way analyses of variance (ANOVA), followed by Fisher’s LSD post-hoc tests, were used to evaluate differences between the groups’ means in the continuous variables. Type I errors were addressed by controlling the false discovery rate (FDR) using the procedure of Benjamini and Hochberg (28), with q value set as 0.05. Group differences in categorical variables were analyzed using Chi-square. Pearson correlation examined associations between variables. To assess whether different elements of the sensory profile observed in PTH subjects could be used to classify subtypes, we employed a logistic regression analysis using the Logit regression package in R. The algorithm-generated models were ranked using the Akaike information criterion with the two top variables used to generate the final model. A receiver-operating characteristic curve was generated for the top-ranking model. The null hypothesis was rejected at p<0.05. FDR-adjusted p values (q-values)<0.05 were considered significant.

Results

General characteristics

Headache-free TBI subjects and those subjects with PPTH did not differ in age, the time elapsed since injury, its cause, the presence of additional neck injury, as well as past involvement in a TBI-related lawsuit (Table 1). Subjects with and without PPTH reported taking medications for various conditions including sleep disorders, depression, anxiety, blood pressure, GI problems, and attention deficit disorder, with no significant difference between the two groups. However, the relative number of subjects that reported taking analgesic medications on a regular basis, being mainly non-steroidal anti-inflammatory drugs (NSAIDs) and antidepressants within the PPTH group (21/31; 67.7%) was higher than in the pain-free group (6/19; 31.6%; χ²=4.7; p=0.03).

TBI characteristics

Table 1 describes the TBI characteristics in the three following groups: a) PPTH subjects with migraine-like symptoms, b) PPTH subjects with TTH-like symptoms, c) headache-free TBI subjects. While the groups did not differ in any of the TBI characteristics, in particular the incidence of moderate and severe TBI, as well as the time elapsed since the injury, PPTH subjects had a significantly higher PTSD level when compared to headache-free TBI subjects (p<0.001). PTSD level, however, was not different between the two PPTH subgroups (p=0.97).

PPTH characteristics

Table 2 describes the headache characteristics among PPTH subjects exhibiting migraine- or TTH-like symptoms. The two subgroups had a similar latency to headache onset after the TBI with the majority reporting that headache started immediately after the trauma. In addition, the frequency of headache was similar between the two groups, with the majority of subjects reporting daily headache, or headache several times a week. The headache intensity was also not different between the two subgroups. However, we discovered group differences in some of the aggravating factors. Individuals with TTH-like symptoms were more likely to report that stress/nervousness aggravated their headache (12/13 vs. 4/15; χ²=12.24; p=0.0005) whereas individuals with migraine-like symptoms were more likely to report that bright light and focused attention aggravated their pain (10/15 vs. 2/13, χ²=7.48; p=0.006; 4/15 vs. 0/13, χ²=4.04; p=0.044, respectively). The groups also differ in some of the alleviating factors: Subjects with migraine-like symptoms were more likely to report that quietness was better in alleviating their headache (9/15 vs. 3/13; χ²=3.88; p=0.049) whereas individuals with TTH-like symptoms were more likely to report that analgesics alleviated their pain (10/13 vs. 5/15; χ²=5.32; p=0.021). Note, however, that a similar proportion of individuals in the two PPTH subgroups used antidepressants (five and four, for migraine-like and TTH-like PPTH respectively) or anti-epileptics (four and three, respectively). Acute triptan use was limited in both PPTH subgroups (two in the migraine-like PPTH, and one in the TTH-like PPTH).

Table 2.

PPTH characteristics.

	Migraine-type PPTH n = 15	TTH-type PPTH n = 13
Onset of PTH after the TBI (days, mean, SD)	1.6 (6.2)	0.7 (1.6)
PPTH intensity
VAS (mean, SD)	7.8 (1.8)	7.3 (1.3)
PRI (mean, SD)	33.1 (16)	34.0 (11)
NWC (mean, SD)	13.3 (6.1)	14.2 (5.1)
Pleast (median)	2	2
Pmost (median)	5	4
Frequency of PPTH (n, valid %)^
Daily	5 (38.5)	7 (58.3)
1–6 times/week	5 (38.5)	4 (33.3)
1–4 times/month	2 (15.4)	0 (0)
Once a month	1 (7.7)	1 (8.3)
Pain aggravating factors (n, %)
Stress/nervousness	4 (26.6)	12 (92.3)***
Bright light	10 (66.6)**	2 (15.4)
Loud noise	9 (60)	5 (38.5)
Physical exercise	5 (33.3)	3 (23.1)
Smells	4 (26.6)	1 (7.7)
Sleeplessness/tiredness	4 (26.6)	4 (30.8)
Concentrating	4 (26.6)*	0 (0)
Cold/hot surroundings	3 (20)	2 (15.4)
Pain alleviating factors (n, %)
Darkness	9 (60)	4 (30.8)
Quietness	9 (60)*	3 (23.1)
Rest/sleep	9 (60)	6 (46.2)
Analgesics (NSAIDs, paracetamol)	5 (33.3)	10 (76.9)*
Massage/relaxation	3 (20)	6 (46.2)
Pain in other body regions (n, %)	10 (66.3)	7 (53.8)

PPTH=Persistent post traumatic headache, TTH=tension-type headache, VAS=visual analog scale, PRI=pain rating index, NWC=number of words chosen, Pleast=pain at its least, Pmost=pain at its most;

^*p<0.05,

^**p<0.01,

^***p<0.001, χ² test between PPTH subjects with migraine- and TTH-like symptoms.

^{^}missing data on two participants in the PPTH with migraine-like symptoms, and one participant in the PPTH with TTH-like symptoms.

Bold indicates significant values.

While there was no difference in the location of the headache between the two PPTH subgroups (Table 3), subjects with migraine-like symptoms were more likely to report a sensation of “spreading pain”, with some describing the pain originating at the base of the skull and “moving” towards the forehead or temple or in a reverse-wise pattern (7/15 vs. 1/13; χ²=5.66; p=0.017). Furthermore, subjects with TTH-like PTH were more likely to report neck pain when compared to those with migraine-like symptoms (8/13 vs. 3/15; χ²=5.04; p=0.025).

Table 3.

Location of Headache and Other Pain Complaints.

	Migraine-like PPTH n=15	TTH-like PPTH n=13
Location in the head (n, %):
Forehead	4 (26.6)	8 (61.5) ^
Temple	8 (53.3)	7 (53.8)
Back of head	5 (33.3)	5 (38.5)
Occipital region	5 (33.3)	5 (38.5)
Vertex	3 (20)	2 (15.4)
Around/behind the eyes	2 (13.3)	2 (15.4)
Side of head	2 (13.3)	–
The pain spreads from one head location to another	7 (46.6) *	1 (7.7)
Neck pain	3 (20)	8 (61.5) *
Location in other body regions:
Back	3 (20)	2 (15.4)
Lower extremity	3 (20)	1 (7.7)
Upper extremity	2 (13.3)	3 (23.1)

^*p<0.05;

^{^}p=0.06; χ² test.

Bold indicates significant values.

Sensory profile

Cephalic measurements: Analyses of variance revealed significant group effects for CST (F(3,62)=5.3, p=0.002), WST (F(3,64)=5.1, p=0.003), HPT (F(3,62)=8.3, p=0.0001), and PPT (F(3,64)=4.1, p=0.001). Figure 1 presents the sensory thresholds of the three distinct TBI groups and healthy controls. Post-hoc comparisons of the CST values (Figure 1(a)) revealed decreased cold sensation (i.e. hyposensitivity) in the two PPTH subpopulations when compared to headache-free TBI subjects (q=0.008 for both) and to healthy controls (q=0.004 for both). The two PPTH subgroups had similar CST values (q=0.99). CST values in headache-free TBI subjects were similar to those observed in healthy controls (q=0.83).

Figure 1.

Cephalic sensory thresholds of TBI subjects with PPTH exhibiting migraine-symptoms, TTH-like symptoms, or no headache (TBI/nPTH), and of healthy controls. Thresholds for cold sensation (a), warmth sensation (b), heat pain (c), and pressure pain (d). (*q<0.05; **q<0.01; ***q<0.0001; FDR corrected values following Fisher’s LSD post-hoc test). Bars denote mean±SD. (e) QST profiles of the four different subject groups. Positive Z scores indicate positive sensory signs (hyperalgesia). Negative values indicate negative sensory signs (hypoalgesia). (f) Receiver Operating Characteristic (ROC) curve depicting the performance of the algorithm to identify the heat pain threshold and pressure pain threshold values as predictors of the probability of having PPTH with TTH-like symptoms. The Area Under the Curve (AUC) is 93.8%, which is significantly different from random chance. CST: cold sensory threshold; WST: warm sensory threshold; HPT: heat pain threshold; PPT: pressure pain threshold; QST: quantitative sensory testing.

Analyses of the WST data (Figure 1(b)) revealed that subjects with TTH-like symptoms had higher values (i.e. warmth sensation hyposensitivity) when compared to subjects with migraine-like symptoms (p)=0.001) and to healthy subjects (p=0.0003), as well as a trend towards higher values when compared to subjects with TBI but no PPTH (p=0.058). The latter group also displayed thermal hyposensitivity when compared to healthy subjects (q=0.04). WST values in subjects with migraine-like symptoms were similar to those observed in headache-free TBI subjects (q=0.09) and in healthy controls (q=0.87).

Analysis of the HPT data (Figure 1(c)) revealed that subjects with TTH-like symptoms had higher HPT values (heat-pain hyposensitivity) when compared to all other subject groups (q=0.0006 vs. migraine-like symptoms; q=0.002 vs. headache-free TBI; q=0.0001 vs. healthy controls). Subjects with migraine-like symptoms exhibited HPT values that were similar to those observed in headache-free TBI subjects (q=0.63) and in healthy subjects (q=0.32).

Analysis of the PPT data (Figure 1(d)) revealed lower PPT thresholds (i.e. cephalic mechanical pain hypersensitivity) in subjects with TTH-like symptoms when compared to all other groups (q=0.03 vs. migraine-like symptoms; q=0.009 vs. headache-free TBI; q=0.002 vs. healthy controls). As with the WST and HPT, subjects with migraine-like symptoms had PPT values that were similar to those observed in headache-free TBI subjects (q=0.67) and in healthy subjects (q=0.21). A summary of Z-scored QST parameters obtained for all tested groups showing the relative increase or decrease in cephalic sensory thresholds is shown in Figure 1(e).

Static allodynia was present only in TBI subjects. In total, 13/31 (41.9%) TBI subjects with PPTH and 3/19 (15.8%) headache-free TBI subjects reported pain and intense unpleasantness following light mechanical stimulation (χ²=3.7; p=0.027). There were no significant differences in the frequency of allodynia between subjects with migraine- and TTH-like symptoms (7/15 vs. 5/13; (χ²=0.19; p=0.66).

Differentiation between PPTH subjects with migraine- and TTH-like symptoms with cephalic QST measurements: Because we identified significant differences in cephalic WST, HPT and PPT between the two PPTH subgroups, we sought to determine whether these distinct sensory features could serve as classifiers of the two different clinical symptom profiles. Using logistic regression, we identified HPT (p=0.017) and PPT (p=0.03) as the best predictors. The fitted model was:

$$logit\left(p\right)=0.42*\text{HPT}-0.07*\text{PPT}-8.98$$

with p being the probability of having PPTH with TTH-like symptoms. The model had a Briar score of 0.0105, and the area under the receiver-operating characteristic curve indicated excellent discrimination (AUC=0.938, Figure 1(f)).

Extracephalic measurements: The distinct sensory abnormalities observed at cephalic sites in PPTH subjects with TTH-like and migraine-like symptoms were partly mirrored by the extracephalic sensory changes. Analyses of variance revealed significant group effects only for WST (F(3,46)=8.1, p=0.0002) and CST (F(3,45)=3.00, p=0.04). Figure 2 presents the sensory thresholds of the three distinct TBI groups and healthy controls. Post-hoc comparisons of the WST values revealed that subjects with migraine- or TTH-like symptoms had reduced warm sensation (i.e. higher thresholds) only when compared to healthy controls (q=0.0008 and q=0.0007 respectively). Headache-free TBI subjects had also reduced warm sensitivity when compared to healthy individuals (q=0.04). WST values, however, were not significantly different between the two PPTH subgroups (q=0.57). PPTH subjects with TTH-like symptoms had lower cold sensitivity only when compared to healthy individuals (q=0.03). Finally, PPTH subjects exhibiting migraine-like symptoms, headache-free TBI subjects, as well as healthy individuals had similar CST values.

Figure 2.

Sensory thresholds at a non-painful extracephalic site (arm) of TBI subjects with PPTH exhibiting migraine-like symptoms, TTH-like symptoms, or no headache (TBI/nPTH), and of healthy controls. Thresholds for cold sensation (a), warmth sensation (b), heat pain (c), and pressure pain (d). (*q<0.05; ***q<0.0001; FDR corrected values following Fisher’s LSD post-hoc test). Bars denote mean±SD.

Correlations between sensory profile and headache intensity

Overall, there were no significant correlations between the headache intensity and the different sensory parameters measured in the two PPTH subgroups. One exception, however, was a significant positive correlation between PRI and WST in subjects with TTH-like symptoms (r=0.62; p=0.032). Namely, the stronger the headache severity was, the higher the cephalic warmth sensation thresholds were (i.e. thermal hyposensitivity.

Correlations between headache pain magnitude and PTSD

Analyses of variance revealed a significant group effect for PTSD scores (F(3,43)=25.62, p<0.0001). As Figure 3(a) depicts, when compared to healthy subjects, all TBI subjects had higher PTSD scores regardless of the presence of PPTH (q<0.0001 for all comparisons, post-hoc test). However, PTSD scores in the two PPTH subgroups were higher than those recorded in the headache-free TBI group (q<0.0001 for both, post-hoc test). Nonetheless, the two PPTH subgroups had similar PTSD scores (q=0.17, post-hoc test). As Figures 3(b) and (c) depict, significant correlations were also detected between PTSD levels and the MPQ (PRI index) recorded for the PPTH subjects. However, the two PPTH subgroups displayed opposing directions of correlation. In subjects with migraine-like symptoms there was a negative correlation between the PRI and PTSD scores (r=−0.68, p=0.02; Figure 3(b)), while subjects with TTH-like symptoms exhibited a positive correlation (r=0.65, p=0.016, Figure 3(c)).

Figure 3.

Relationship between PTSD scores and PPTH. (a) PTSD scores of subjects with PPTH exhibiting migraine-like symptoms, TTH-like symptoms, TBI without headache, and healthy controls. (***q<0.0001; FDR corrected values following Fisher’s LSD post-hoc test). Bars denote mean±SD. A significant negative correlation was found between PTSD scores and pain rating index (PRI) in subjects with PPTH and migraine-like symptoms (b). A significant positive correlation was found between PTSD score and the PRI in subjects with PPTH with TTH-like symptoms (c). p<0.05 for both.

Discussion

Here we studied a cohort of male subjects with de novo PPTH that exhibit migraine- or TTH-like symptoms and were able to characterize a novel subset of differentiating clinical characteristics. We also observed distinct sensory profiles that supported this symptoms-based classification of PPTH. While some of the symptoms we documented are reminiscent of migraine or TTH, many were not congruent with previous studies of subjects with primary headache conditions, suggesting differences between traumatic and atraumatic headache. For example, we found that stress/nervousness and bright light serve as aggravating factors in PPTH resembling TTH and migraine respectively. Of interest is a recent pilot data showing that bright light can decrease extracephalic thermal pain thresholds in PPTH with a migraine phenotype (29). Stress is thought to contribute to TTH (30), while bright light often exacerbates the headache in primary migraine (31). However, while our subjects also reported that darkness was effective in alleviating both PPTH subtypes, it is primarily considered to alleviate migraine pain. While PPTH subjects with migraine-like symptoms reported that quietness was more effective in alleviating their headache, loudness was considered an aggravating factor in both PPTH subgroups. PPTH subjects with TTH-like symptoms were more likely to involve localized neck pain – a key feature in primary TTH. It is, however, also a factor associated with migraine (32). Subjects with PPTH resembling migraine also reported poor headache localization, as in chronic migraine (32). However, TTH pain is also poorly localized. Finally, we identified subjects with PPTH and TTH-like symptoms as more responsive to analgesic medication, in particular NSAIDs. While primary TTH is often responsive to NSAIDs, so is migraine (33).

Our QST findings suggest that the two PPTH subgroups exhibit distinct cephalic sensory thresholds for warm sensation, heat pain, and mechanical pressure pain. Importantly, our data suggest that PPTH subjects with TTH-like symptoms exhibit abnormal thermal and mechanical sensory thresholds, while those with migraine-like symptoms exhibit normal warmth, heat pain and mechanical pain thresholds. If confirmed in a larger cohort, the logistic regression model suggests that having a higher cephalic heat pain threshold and a lower cephalic mechanical pain threshold could be used to distinguish between these two major PPTH subgroups.

The finding of cephalic mechanical hyperalgesia in subjects with primary TTH (34) point to the possibility of a shared mechanism with TTH-like PPTH. However, the finding that these PPTH subjects exhibited thermal hypoalgesia, while chronic TTH is associated with thermal hyperalgesia (35–37) suggest that the two conditions involve different or additional mechanisms. We also propose that PPTH with migraine-like symptoms involves different mechanisms than those underlying primary migraine. This notion is supported by the finding of increased cold detection thresholds and normal pressure pain thresholds in PPTH subjects exhibiting migraine-like symptoms, while migraine is associated with reduced thresholds for heat and pressure pain, and normal cold detection thresholds (38,39). It should be noted that recent imaging studies also support the view that PPTH and primary migraine involve different mechanisms (40–43). Nonetheless, in these studies the distinction between the different subsets of PPTH subjects was not addressed.

The presence of cephalic and extracephalic thermal hyposensitivity in the two PPTH subgroups points to potential damage to central relay pathways that convey sensory input from the whole body. It should be noted that a recent pilot study reported reduced cutaneous thermal pain threshold in subjects with PPTH exhibiting a migraine phenotype (29); the exact methodology used to capture this hypersensitivity was not indicated, however. Damage to the temperature and/or the pain system is a key feature of central neuropathic pain (16,44–46). The deficits in thermal sensitivity in cranial painful regions together with the extracephalic warmth hypoalgesia found in PPTH subjects exhibiting symptoms resembling migraine or TTH suggest that PPTH, regardless of its phenotype, may be a subtype of central neuropathic pain. However, given the minor cephalic changes observed in PPTH subjects with migraine-like symptoms, such a mechanism is likely to play a larger role in mediating TTH-like PPTH.

In addition to exhibiting more extensive sensory deficits, PPTH subjects with TTH-like symptoms also uniquely displayed cephalic mechanical hyperalgesia. While damage to the trigemino-thalamic pathway could explain the deficits in cephalic warmth and heat pain sensations observed in PPTH subjects with TTH-like symptoms, it seems at odds with the finding of cephalic mechanical hyperalgesia in this subpopulation. The mechanism underlying this modality-specific pain hypersensitivity could involve peripheral and/or central mechanisms. Damage to peripheral sensory nerve endings in deep cephalic tissues, or persistent peripheral inflammation in injured cephalic structures, could initiate a chain of events leading to peripheral and central sensitization with ensuing mechanical hyperalgesia. It should be emphasized that the local mechanical hyperalgesia revealed herein, using the pressure pain test, is indicative mainly of deep rather than cutaneous tissue sensitization. It is therefore possible that despite partial damage to the nociceptive central pathways in PPTH subjects with TTH-like symptoms, chronic sensitization of primary afferents that innervate deep cephalic tissue may be sufficient to drive the mechanical hyperalgesia in this PPTH subgroup. Interestingly, while cranial mechanical hyperalgesia was specific to this PPTH subgroup, we observed cranial static mechanical allodynia in the two PPTH subgroups. The notion that static allodynia is a manifestation of central sensitization (20) points to the possible contribution of this mechanism, at least in part, to the pathogenesis of PPTH regardless of its subtype.

Another characteristic shared by the two PPTH subgroups is high levels of PTSD when compared to headache-free TBI subjects, despite a similar incidence of moderate and severe TBI in these distinct groups. PTSD is a common characteristic among subjects with PPTH and is often reported to correlate with headache severity and disability (47–49). A novel finding of the current study is that the interaction between PTSD level and the magnitude of the headache in the two PPTH subtypes followed opposite directions. We identified a positive correlation between headache intensity and PTSD symptomatology among subjects with TTH-like symptoms. These subjects also reported aggravation of the headache by tension and emotional stress. These findings, which are in accordance with previous reports on PPTH populations in general, point towards the potential contribution of a psychogenic factor in this PPTH subpopulation. Longitudinal studies suggest that PTSD may be a risk factor for other types of chronic post-traumatic pain (50). PTSD symptoms may have preceded the PPTH in this subgroup, thus contributing to its maintenance (50). PTSD and PPTH may also reinforce one another. PTSD may contribute to PPTH via increased levels of anxiety, distress and depression, and PPTH could perpetuate emotional problems, which further exacerbate the pain. The finding of negative correlation between PTSD and headache intensity in the PPTH subjects with migraine-like symptoms is puzzling given the postulated contribution of PTSD symptoms to chronic pain and points to a unique interaction between TBI and PTSD in this PPTH subpopulation. While a clear causation between these two parameters is uncertain, a possible mechanism linking heightened PTSD to reduced chronic PTH levels in these subjects may involve circuits underlying stress-induced analgesia (51). Such a mechanism, however, is unlikely to contribute to the emergence of the migraine-like characteristics, such as throbbing pain, or photophobia (49). The notion that subjects with primary migraine have increased risk of developing PTSD (52) also serves as another indicator that the mechanisms underlying migraine-like PPTH are distinct from those underlying primary migraine.

The findings of this study should be viewed in the context of several limitations. Although the majority of our PPTH subjects suffered mild TBI, our cohort also included subjects that suffered moderate or severe TBI. It should be noted, however, that while PPTH is considered to be more common after mild TBI than after moderate to severe TBI (53), our cohort of headache-free TBI subjects had a similar incidence of mild and more severe TBI. The cross-sectional design of the current study restricts data interpretation regarding the underlying mechanisms of PPTH; in particular, our ability to determine whether the positive signs (i.e. mechanical hypersensitivity and allodynia) observed in PPTH subjects with TTH-like symptoms preceded the headache or rather developed as a consequence. Although our relatively small sample size allowed us to detect statistical differences, we acknowledge some heterogeneity even within the two PPTH subgroups. Larger cohorts will be required to validate the differences we observed between the two PPTH subgroups. Inclusion of data regarding family history of primary headache disorders in future studies may also be helpful to identify potential genetic factors. Inclusion of females in future studies will also be required to assess potential sexual dimorphism in the difference between the two PPTH subtypes. It should be emphasized that symptoms reported by TBI subjects may be, consciously or unconsciously, influenced by factors motivated by litigation. Although PPTH may, at times, reflect malingering, our subjects were tested long after litigation processes and compensation claims had been resolved, therefore malingering seems less likely. While none of the subjects reported taking preventive medications to treat their headache, some indicated the use of pain medications. As we did not specifically assess the use of analgesics just prior to the QST evaluation session, such treatment may have affected the finding. The use of antidepressant and anti-epileptic drugs may also have influenced the results. However, since the relative number of subjects that reported using these drugs in the two PPTH subgroups was similar, it was unlikely to play a role in mediating the differences noted.

In summary, we observed distinct clinical characteristics and QST profiles in PPTH with symptoms resembling TTH or migraine as well as different interactions between the headache and PTSD levels in these subgroups. Our findings point to different etiologies of these two PPTH subtypes, which involve a combination of injuries to both peripheral and central structures that may lead to peripheral hyperalgesia, particularly in TTH-like PPTH. A better, evidence-based classification of chronic PTH linked to mechanism could potentially be used for targeted therapy to improve the management of this chronic pain syndrome.

Key findings.

• Post-traumatic headache resembling migraine exhibits different clinical and quantitative sensory testing profiles than post-traumatic headache with tension-type headache-like symptoms.

• Distinct clinical symptoms and sensory profiles may be linked to different etiologies of post-traumatic headache and differentiate between subgroups in studies aimed at improving treatment.