Thoracic spine thrust manipulation for individuals with cervicogenic headache: a crossover randomized clinical trial

ABSTRACT

Objective

To determine if thoracic spine manipulation (TSM) improves pain and disability in individuals with cervicogenic headache (CeH).

Methods:

A randomized controlled crossover trial was conducted on 48 participants (mean age: 34.4 years) with CeH symptoms. Participants were randomized to 6 sessions of TSM or no treatment (Hold) and after 4-weeks, groups crossed over. Outcomes were collected at 4, 8 and 12 weeks and included: headache disability inventory (HDI), neck disability index (NDI), and the global rating of change (GRC). Outcomes were analyzed using a linear mixed-effects model with Bonferroni correction. Odds of achieving the minimal clinically important difference (MCID) on the GRC of +4 or greater were also calculated. Scores at 4 weeks represent the only timepoint where 1 group is fully treated and other group has not received any treatment.

Results:

Comparing hold to active treatment, HDI were not significantly different between groups (mean difference = 7.39, 95 CI: −4.39 to 19.18; P = 0.214) at any timepoint; the NDI was significant (mean difference = 6.90, 95 CI: 0.05 to 13.75; P = 0.048) at 4 weeks. Odds of achieving the +4 MCID on the GRC (OR = 38.0, 95 CI: 6.6 to 220.0; p < 0.001) favored TSM at 4 weeks.

Conclusion:

TSM had no effect on headache-related disability but resulted in significant improvements in neck-related disability and participant reported perceived improvement. Future studies are needed to examine the long-term impact of TSM in this population.

Introduction

The worldwide lifetime prevalence of headache is reported to be 96% [1], and an estimated 46% of adults worldwide currently suffer from an active headache disorder [2]. Due to the high levels of disability and pain created by these disorders, the World Health Organization has listed headaches as 1 of the 10 most disabling conditions in the world for both men and women [2]. Although general population prevalence estimates vary highly (0.17%-4.6%) due to changing diagnostic criteria [3], it is estimated that cervicogenic headaches (CeH) comprise up to 17.8% of those individuals with headache complaints [4,5]. CeH results in pain and levels of disability similar to that of both migraines and episodic tension-type headaches, leading to a significant burden to society [6].

Cervicogenic headache is defined as a secondary disorder in which pain is referred from the neck and perceived in one or more areas of the head and or face [7,8] (Figure 1). Multiple studies have identified general cervical features indicative of CeH, including links to restricted and painful upper cervical joints [9–12]. Zito and Jull [9] demonstrated a significant reduction in cervical flexion/extension range of motion as well as a significantly greater incidence of upper quarter muscle tightness in individuals with CeH compared to those with migraine and tension-type headaches. According to Ahn and colleagues [13], 50% of patients with persistent neck pain will develop headache symptoms.

Figure 1.

Anatomic artist interpretation of cervicogenic headache presentation in which pain is referred from the neck and perceived in one or more areas of the head and or face [68].

Cervical manipulation is commonly used in the treatment of individuals with CeH and cervical disorders [14,15]; however, there are perceived risks, including cervical artery dissection [16,17]. Due to its proximity to the neck, the thoracic spine can be a contributor to cervical spine symptoms [18–23]. Thoracic spine manipulation (TSM) has been shown to increase cervical range of motion, decrease pain, and lead to improved self-reported function in individuals with neck pain [24–27]. It is speculated that CeH arises secondary to cervical dysfunction [28]. The concept of regional interdependence suggests that ‘ … a patient’s primary musculoskeletal symptoms may be directly or indirectly related or influenced by impairments from various body regions and systems regardless of proximity to the primary symptoms’ [29]. Regional interdependence is a paradigm frequently considered in the treatment of musculoskeletal disorders [30–33], but has not been well studied in relation to CeH. Despite the recommendation in clinical practice guidelines [34,35] for the use of manual therapy to the thoracic spine for patients with chronic neck pain with headache, no clinical trials have actually assessed the effects of TSM on individuals with CeH. Therefore, the purpose of this study was to determine if individuals with CeH who received thoracic manipulation demonstrate a greater reduction in headaches (frequency, intensity, pain and overall disability) when compared to no treatment. Neck disability and overall pain levels were also tracked in order to more accurately assess the treatment model of regional interdependence with regard to cervical and CeH treatment.

Methods

Study design

A repeated-measures, randomized crossover trial design was utilizedi. In this design all participants receive both treatments, but the order of the treatment is randomized with one group getting an active treatment first and the other group waiting to start treatment (labeled ‘hold’) during the same initial time period. Individuals then switch groups, essentially serving as their own control group (Figure 2).

Figure 2.

CONSORT diagram of participant flow through the study, from participant recruitment to retention and follow-up.

Participants

Consecutive individuals seeking care for a primary complaint of headache symptoms at the University of Colorado Anschutz Medical Campus, Aurora, CO, between August 2016 and October 2019 were screened for eligibility. Inclusion criteria consisted of individuals, ages 18 to 65 with a primary complaint of headache that was exacerbated by neck movements or sustained postures. To make a clinical diagnosis of ‘probable’ CeH [36] and to be included in the study, the presence of the following was confirmed by a physical therapist: 1) unilateral or unilateral dominant side-consistent headache, associated with neck pain and aggravated by neck postures or movement [7] 2) joint tenderness and peripheralization of symptoms into the head in at least one of the three upper cervical joints (C0-C3) as assessed by manual palpation [9] 3) headache frequency of at least one per week over the past 2 months. Exclusion criteria consisted of any contraindications to manipulative therapy or serious pathology (tumor, fracture, rheumatoid arthritis, osteoporosis, or history of prolonged steroid use) identified during the physical examination or on the individuals’ medical intake form. Additionally, individuals were excluded if they had any of the following: 1) recent and significant trauma to the cervical or thoracic region which preceded the onset of headache 2) history of whiplash within the past 6 weeks 3) acute fracture in the thoracic region 4) prior surgery to the cervical or thoracic spine 5) evidence of central nervous system involvement (as evidenced by hyperreflexia or upper motor neuron signs such as Hoffman’s test, clonus, or Babinski) 6) two or more positive neurologic signs consistent with nerve root compression 7) received chiropractic or physical therapy treatment for their headaches within the past 6-months 8) a diagnosis of migraine headaches with or without aura 9) receiving workers compensation or had pending legal action regarding their headaches 10) inability to comply with treatment and follow-up schedule 11) did not want to receive spinal manipulation treatment.

Ethics approval was granted by the Colorado Multiple Institutional Review Board (15–0983). Informed consent was obtained from all participants prior to en-rollment, and all rights were protected. This trial was registered a priori (Clinicaltrials.gov NCT02708953).

Clinicians

Three physical therapists, with a mean of 17.33 years of clinical experience (10, 18, and 24 years) participated in the recruitment, examination and treatment of all participants in this study. All three physical therapists were board certified in orthopedics by the American Board of Physical Therapy Specialties, two had fellowship training in manual therapy and one completed an orthopedic residency program. All physical therapists underwent a standardized training regimen which included studying a manual of standard procedures with the operational definitions of each examination and treatment procedure and a 2-hour hands-on training session provided by the principal investigator to ensure that treatment was performed in a standardized fashion. All physical therapists were trained to perform both roles (examination and treatment). To assure fidelity to the study procedures the primary investigator met with clinicians involved in recruitment and treatment on a monthly basis. The treating therapists were blinded to the results of the examination. Due to the nature of the interventions used in this study, treating physical therapists could not be blinded to intervention techniques.

Outcomes

Participants completed several instruments to assess pain, disability and headache behavior at baseline, 4, 8, and 12 weeks including the Headache Disability Inventory (HDI), the Neck Disability Index (NDI), and the Numeric Pain Rating Scale (NPRS). The HDI was selected as the primary outcome measure; however, since the HDI has not been studied in individuals with CeH, we also chose to assess scores from the NDI, which has been recommended as an outcome measure for CeH in previous studies [37,38].

The HDI is a highly reliable (r = 0.83–0.95) [39,40] 25-item questionnaire that assesses perceived impact of the headache on daily life, scored as ‘yes’ (4 points), ‘sometimes’ (2 points), or ‘no’ (0 points) with a score out of 100 points (higher scores equal higher disability) [39]. The minimal detectable change (MDC) for the HDI has been reported between 16 and 29 points in individuals with headaches [39,40].

The NDI is widely used for individuals with neck pain [26] and consists of 10 items scored from 0 to 5 (doubled for a score out of 100), with higher scores indicating greater disability. Young et al., reported that the NDI is well suited as a short-term self-report measure for patients with CeH[41]. The NDI is a valid and reliable measure for patients with neck pain and CeH with an MDC ranging from 10 to 14 points for neck pain and 5.9 for CeH [42–45]. The NDI exhibited excellent reliability (ICC = 0.92), and adequate responsiveness and construct validity in individuals with CeH with a minimal clinically important difference (MCID) of 5.5-points [45].

The NPRS is an 11-point scale (0 = no pain; ‘10’ = worst imaginable pain) shown to be valid and reliable for measuring pain intensity [46–49], with an MDC and MCID of 2.1 and 1.3 points respectively [50]. Participants rated their current, worst, and least level of headache pain in the previous 24 hours. Worst pain scores were utilized in the analyses as CeH symptoms tend to be episodic with periods of no pain [7]. At 4, 8, and 12-weeks patient perception of improvement in their primary symptom of headache was measured with the Global Rating of Change (GRC), which has been shown to have good-to-excellent reliability in patients with cervical disorders [51]. The 15-point scale ranges from – 7 (a very great deal worse) to zero (about the same) to +7 (a very great deal better). GRC scores are moderately correlated with improved pain and disability [51], and clinically relevant thresholds of change range from +2 to +4 [52–54]. We used a conservative minimum of +4 (‘moderately better’) to identify those with clinically important change, which has been used in other neck pain studies [54,55].

Additionally, at each treatment session participants were asked if they experienced any side effects. Side effects were considered mild if the symptom intensity was rated 1 or 2 (1–4 intensity scale) and symptoms resolved within 48 hours. Longer duration or higher intensity of symptoms were classified as 3 (moderate) or 4 (severe) accordingly.

History and physical examination

The historical examination included demographic information, past medical history, and assessment of the nature and onset of symptoms. The physical examination assessed strength, endurance, flexibility, range of motion (ROM), and joint mobility and other common clinical tests used to determine if HVLA techniques are safe to perform.

Randomization

Concealed allocation was performed prior to the start of recruitment by a research assistant not involved in data collection using a computer-generated randomized table of numbers created for the participating site prior to the beginning of the study. Individual, sequentially numbered index cards with the random assignment were prepared by the research assistant; index cards were folded and placed in sealed opaque envelopes. Once the baseline examination was completed by the examination therapist, the randomization envelope was handed to the treating therapist who was blinded to the baseline examination results. Participants were randomly assigned to one of two groups: 1) TSM plus a mobility exercise program (Group 1) or 2) Hold (no treatment) Group (Group 2).

Sequence of treatment

The sequence of treatment was different for each group. Participants in Group 1 were assigned to receive the active TSM intervention during the first 4 weeks; immediately upon enrollment in the study. Participants in Group 2 were assigned to a hold group for the first 4 weeks, receiving no treatment during this time; this served as the control. At the end of 4 weeks, Group 2 which had been holding with no treatment transitioned to receive the active TSM intervention for the next 4 weeks (weeks 4 to 8) while Group 1 which had just received the active TSM intervention transitioned to a 4-week hold period. At 4 weeks, the scores represent differences between one group that had received the full treatment and the other group that had not. At the end of 8 weeks, scores represent both groups having received the active TSM intervention. All participants were followed for an additional 4 weeks to assess the short-term stability of the treatment, with a total study participation time of 12 weeks for every participant.

Intervention procedures

Participants in both groups attended physical therapy sessions 1–2 times weekly for up to 4 weeks (each treatment session was approximately 15 minutes in duration) for a maximum of 6 sessions in an outpatient research designated facility at the University of Colorado, Anschutz Medical Campus. Individuals in both groups received prescriptive manual therapy to the cervicothoracic junction and the thoracic spine (Appendix A) plus a thoracic mobility exercise (Appendix B, Figure B1). Participants received 5 thoracic spine high-velocity low amplitude (HVLA) techniques targeting the upper, middle and lower thoracic spine. Details regarding the TSM interventions are described in Appendix A using the model proposed by Mintken et al [56]. All of the techniques described in Appendix A were repeated at the subsequent treatment visits. Following the TSM the physical therapist instructed each participant in a home exercise mobility program. The exercise was a general thoracic mobility exercise performed in supine over a towel or foam roller[57]. Participants were instructed to perform 8–10 repetitions of the exercise in up to 3 areas, 3–4 times per day while receiving intervention in the study. Additionally, participants were instructed to continue activities that do not increase symptoms, and avoid activities, which aggravate symptoms.

The first treatment session was always performed on the day of the initial examination, and the participant was scheduled for a follow-up visit within 2 to 4 business days. After the initial TSM group received 4 weeks of treatment, they crossed over into the hold group. Participants in both groups were also advised to continue usual activities that did not increase symptoms and to avoid any activities that exacerbated their symptoms.

Statistical analysis

Sample size calculation

An a priori sample-size calculation determined that 48 participants were required to achieve 90% power, accounting for a dropout rate of 15%. The HDI has had very little use in trials for cervicogenic headaches, providing minimal data for power analyses. The minimal detectable change (MDC) for the HDI has been reported between 16 and 29 points in other headache populations [39, 40]. We used the threshold of a 16-point MDC, with a beta of 0.9, two-sided significance level of 0.05 and expecting a within patient standard deviation of 22 points as seen in other trials of manual therapy for headaches [58]. This provided us with a sample size estimation of 42 total patients (21 per group) and we added 6 additional (15%) to account for potential loss to follow-up yielding a total enrollment goal of 48 participants. Power analysis calculated with use of the Massachusetts General Hospital Biostatistics Center cross-over trial design sample size calculator [59].

Data analysis

Baseline and descriptive statistics were reported for each group. Primary and secondary outcomes were analyzed using a linear mixed effects model, which is robust for handling unbalanced and missing data [60], and particularly appropriate for repeated measures designs [61,62]. We modeled time as the repeated effect, and sequence of treatment as the random effect using an autoregressive covariance structure. We reported estimated marginal means with 95% confidence intervals and ran Bonferroni correction for multiple comparisons. The time period to best isolate treatment effects would be at the 4-week mark, as this differentiates the outcomes between the hold and the active treatment group. By 8 weeks both groups had received the active treatment. Additionally, we compared the odds of achieving the minimal clinically significant difference on the GRC of 4+ or higher between groups. Data analysis was performed with the SPSS, version 26.0.

Missing Data: Our primary method to account for missing data was the use of the linear mixed effects model, which is flexible to account for missing data [60], and particularly missing data in cross-over trials [62]. For the GRC data, we imputed missing values (Markov chain Monte Carlo with 20 imputations) [63].

Changes to study from protocol registration

We had included the Pain Self-Efficacy Questionnaire (PSEQ) as an outcome in our initial trial registration; however, it was not meant to support any of our primary or secondary aims. Our focus was on changes in pain and disability in individuals with CeH, and not self-efficacy. We reported the baseline scores for this measure but did not report it as an outcome in our trial.

Results

Of the 94 participants screened for this study 48 (mean age of 34.4 years, 72.9% female [N = 35], and a mean duration of symptoms of 62.1 months) satisfied eligibility criteria and agreed to participate (Table 1). Approximately 10.3% of the outcome values were missing. Little’s Missing Completely at Random (MCAR) test was not significant (p = 0.306) confirming a specific pattern of missingness was not likely.

Table 1.

Demographics and outcome measures at baseline

	Group 1: TSM First 4 Weeks (Active Treatment) (N = 24)	Group 2: Hold First 4 Weeks (No Treatment) (N = 24)	Total Group 1+ Group 2 (n = 48)
Age	34.96 (±9.38)	33.88 (±11.79)	34.42 (±10.45)
Sex	18 (female)	17 (female)	35 (female)
Race	White = 21 Black = 1 Hispanic = 1 Asian = 1	White = 21 Black = 0 Hispanic = 3 Asian = 0	White = 42 Black = 1 Hispanic = 4 Asian = 1
Duration of symptoms (days) Mean (median)	2185.46 (1153.00)	1543.5 (997.00)	1864.48 (1002.00)
HDI (0–100)a	28.75 (±13.39)	31.33 (±22.24)	30.04 (±18.02)
NDI (0–50)a	13.67 (±6.54)	11.63 (±5.61)	12.65 (±6.05)
NPRS (0–10)a (current) (worst in 24 hours)	2.76 (±2.15) 4.54 (±1.93)	3.75 (±2.36) 4.83 (±2.08)	3.31 (±2.24) 4.69 (±1.97)
PSEQ (0–60)b	48.88 (±9.80)	51.29 (±9.70)	50.08 (±9.62)
HA duration (years) Mean (median)	4/24 = 6 months or less 0/24 = 6 months-1 yr 7/24 = 1–5 yrs 9/24 = 5–10 yrs 4/24 = 10 yrs+ Mean = 3.38 (4.0), 1–5 yrs (5–10 yrs)	1/24 = 6 months or less 3/24 = 6 months-1 yr 11/24 = 1–5 yrs 8/24 = 5–10 yrs 1/24 = 10 yrs+ Mean = 3.08 (3.0), 1–5 yrs	5/48 = 6 months or less 3/48 = 6 months-1 yr 18/48 = 1–5 yrs 17/48 = 5–10 yrs 5/48 = 10 yrs+ Mean = = 3.29 (3.0), 1–5 yrs
HA duration (hours) Mean (median)	6/24 = 1–3 hrs 7/24 = 3–6 hrs 5/24 = 6–12 hrs 3/24 = 12–24 hrs 3/24 = other Mean = 2.58 (2.0), 6–12 hr	5/24 = 1–3 hrs 9/24 = 3–6 hrs 6/24 = 6–12 hrs 2/24 = 12–24 hrs 2/24 = other Mean = 2.46 (2.0), 3–6 hr	11/48 = 1–3 hrs 16/48 = 3–6 hrs 11/48 = 6–12 hrs 5/48 = 12–24 hrs 5/48 = other Mean = 2.52 (2.0), 6–12 hr
HA frequency (days/week) Mean (median)	2.67 (2.50)	3.50 (3.00)	3.08 (3.00)
HA intensity (0–10)a Mean (median)	5.29 (2.50)	5.17 (5.00)	5.23 (5.00)

Values are mean ± SD unless otherwise indicated.

Abbreviations: HA, headache; NPRS, Numeric Pain Rating Scale; PSEQ, Pain Self-Efficacy Questionnaire; NDO, Neck Disability Index; HDI, Headache Disability Index; HA, Headache; Years, yrs; Year, yr; Hour, hr; Hours, hrs;

^aLower Scores are better.

^bHigher Scores are better.

At four weeks, the only time point where one group had received the full active treatment and the other had not, HDI scores favored Group 1 (active TSM first), but were not significant (mean difference = 7.39, 95 CI: −4.39 to 19.18; P = 0.214; Table 2 and Figure 3). The difference for NDI scores was significant (mean difference = 6.90, 95 CI: 0.05 to 13.75; P = 0.048) and for NPRS (mean difference = 2.2, 95 CI: 0.7 to 3.8; P = 0.006) favoring Group 1 (active TSM) (Figures 4 and 5 respectively). Participants who received the active treatment had greater odds of reporting a clinically meaningful change on the GRC of +4 (OR = 17.5, 95 CI 3.3 to 92.9; p < 0.001).

Table 2.

Primary and secondary outcomes

	Group 1: 0–4 Weeks: TSM 5–8 Weeks: Hold 9–12 Weeks: No Tx N = 24	Group 2: 4 Weeks: Hold 5–8 Weeks: TSM 9–12 Weeks: No Tx N = 24	Between Group Difference	P Value
Headache Disability Inventory (0 to 100)
Baseline	28.75 (20.03, 37.47)	31.33 (22.61, 40.06)	2.58 (−9.75 to 14.92)	0.677
4 weeks†	24.17 (15.87, 32.46)	31.56 (23.19, 39.93)	7.39 (−4.39 to 19.18)	0.214
8 weeks††	24.26 (16.02, 32.51)	24.27 (16.08, 32.45)	0.002 (−11.62 to 11.62)	1.000
12 weeks	24.27 (16.03, 32.50)	23.58 (15.35, 31.81)	0.69 (−12.33 to 10.96)	0.906
Neck Disability Index (0 to 100)
Baseline	27.50 (22.46, 32.54)	23.58 (18.55, 28.62)	3.92 (−3.21 to 11.04)	0.276
4 weeks†	16.42 (11.65, 21.19)	23.32 (18.40, 28.23)	6.90 (0.05 to 13.75)	0.048
8 weeks††	16.60 (11.96, 21.23)	15.00 (10.49, 19.51)	1.60 (−4.87 to 8.06)	0.621
12 weeks	17.55 (12.21, 22.90)	16.32 (11.00, 21.65)	1.23 (−6.32 to 8.78)	0.746
Numeric Pain Rating Scale (0 to 10)
Baseline	4.5 (3.7, 5.4)	4.8 (4.0, 5.7)	0.3 (−0.9 to 1.5)	0.620
4 weeks†	2.7 (1.7, 3.8)	5.0 (3.8, 6.1)	2.2 (0.7 to 3.8)	0.006
8 weeks††	2.7 (1.6, 3.7)	3.3 (2.3, 4.3)	0.6 (−0.8 to 2.1)	0.387
12 weeks	2.8 (1.7, 3.9)	3.5 (2.4, 4.6)	0.7 (−0.8 to 2.2)	0.361

Values represent estimated marginal means (95% confidence intervals); TSM = Thoracic Spine Manipulation, No Tx = No Treatment; † Active treatment (TSM) for Group 1 ends at 4 weeks, while Group 2 has been getting no treatment. †† Active treatment (TSM) complete for Group 2 while Group 1 received no additional treatment after their active treatment (TSM) ended 4 weeks prior (both groups have received TSM by this time point and through 12 weeks).

Figure 3.

Differences in scores on the headache disability inventory.

Figure 4.

Differences in scores on the neck disability index.

Figure 5.

Differences in scores on the numeric pain rating scale (worst pain in last 24 hours).

At eight weeks, once both groups had received the active intervention, none of the outcomes were different between groups (Table 2). The mean HDI mean difference was = 0.002 (95 CI −11.62 to 11.62; P = 1.000), the NDI mean difference was 1.60 (95 CI −4.87 to 8.06; P = 0.621), and the NPRS mean difference was 0.6 (95 CI −0.8 to 2.1 p P = 0.387).

Differences between groups at 8 weeks and beyond, after all participants had received the active treatment, were not significantly different (Table 2). Participants that had received the active treatment had greater odds of reporting a clinically meaningful change on the GRC of +4 (OR = 17.5, 95 CI 3.3 to 92.9; p < 0.001) at 4 weeks compared to the Hold group with no treatment. At 4 weeks only, 2 (8.3%) patients from Group 2 (hold first) reported a + 4 or higher GRC score compared to 15 (62.5%) in the active TSM group. By 8 weeks, after all participants had received the active treatment, and for the remainder of the study follow-up period, the number of responders was similar in both groups (Figure 6). No participants reported experiencing any side effects with the interventions.

Figure 6.

Differences in scores on the global rating of change.

Sensitivity Analysis: We ran the same analysis for the primary and secondary outcomes removing any individuals without at least 2 time points of data (N = 47), and without both the 4 and the 8-week assessments as these represent the cross-over treatment periods (N = 40) and the outcomes were unchanged.

Discussion

Despite the recommendation in clinical practice guidelines (CPGs) for the use of manual therapy to the thoracic spine for patients with chronic neck pain with headache, [34,35] no clinical trials have actually assessed the effectiveness of TSM for CeH. One study comparing the combined effectiveness of cervical and thoracic manipulation to mobilization and exercise in individuals with CeH found that upper cervical and upper thoracic manipulation were shown to be more effective than mobilization and exercise out to 3 months [38]. The Dunning et al. [38] study suggested that including thoracic manipulation may facilitate improved outcomes in individuals with CeH, however they included manual therapy focused on both the cervical and thoracic spine into one treatment package, making it impossible to extrapolate the isolated effects of cervical versus thoracic interventions. To our knowledge, this is the first randomized trial to investigate the effects of treatment of the thoracic spine on pain and disability in individuals with CeH. This trial provides preliminary evidence that TSM paired with general thoracic mobility exercises may be effective at reducing pain and neck related disability, but not headache related disability, in individuals with CeH. The results suggest that 6 sessions of TSM had a limited effect on headache disability (as measured by the HDI). The results also suggest that TSM resulted in greater improvements in neck disability (as measured by the NDI) and pain compared to a control group receiving no treatment. At 4 weeks more individuals in Group 1 (TSM) reported a clinically meaningful improvement in their primary complaint of headache on the GRC (n = 14, Figure 6) compared to those that had been holding with no treatment (n = 2). Group 2 (hold first), which served as a ‘control,’ remained the same during a 4-week period where no treatment was received, suggesting that symptoms were stable and not changing due to the chronicity of the participants’ symptoms (mean duration of symptoms >5 years). A similar linear reduction in neck disability and pain occurred in Group 2 (hold first) once they ‘crossed over’ and treatment with TSM was initiated. By 8 weeks, after all participants had received the active treatment, and for the remainder of the study follow-up period, the proportion of participants who reported a clinically meaningful change on the GRC of +4 or higher was similar in both groups

We a priori chose the HDI as the primary outcome measure in our study [39]. In the original psychometric analysis of the HDI, participants with all different types of headaches were included, thus it is not specific to CeH-related disability. The HDI has never been investigated solely in individuals with CeH so it is possible that this instrument may not be as useful for this specific headache population. The MDC for the HDI has been reported between 16 and 29 points in other headache populations [39,40]. The 29-point threshold is equal to the mean baseline HDI scores in our cohort, making this level of improvement nearly impossible for most individuals in our study to achieve [40]. As the definition of CeH includes impairments in the cervical spine, it would appear from our results that the Neck Disability Index, which also captures headache-related symptoms (question 5), may be a more suitable instrument to capture the constructs of pain and disability in individuals with CeH. Had this been chosen as our primary outcome, the differences in treatment groups would have been significantly different, favoring the TSM. More research is needed to identify the optimum outcome measure to use for individuals with CeH.

At 4 weeks, the between-group difference on the NDI was 6.9 points favoring the intervention group, which exceeds the MCID of 5.5-points for individuals with CeH [41]. Additionally, at 4 weeks, the participants in Group 1 (active TSM first) exhibited a within-group mean change on the NDI of 11.08 points, which exceeds the MCID of 5.5 points and the MDC of 5.9 points at 4 weeks for individuals with CeH [41]. Similarly, Group 2 (hold first), after receiving TSM also exhibited a within-group mean reduction in the NDI of 8.32 points.

One of the clinical implications for this study is that clinicians uncomfortable with utilizing cervical manipulation or patients uncomfortable with the idea of receiving cervical manipulation could still expect to receive clinical benefit with thoracic manipulation. No significant adverse events were reported in this study, and the risk of adverse events with TSM appears to be low in other studies as well [64]. The underlying mechanisms driving reductions in pain and neck-related disability with TSM are unclear, but Bialosky et al. [65,66] suggests potential neurophysiological and biomechanical effects, as well as possibly placebo [67]. Our current results suggest that clinicians should consider utilizing TSM in the management of patients with CeH. This treatment has been recommended in 2 separate CPGs [34,35], despite the lack of strong evidence to support it.

Limitations

It is not known if the observed benefits would be sustained long-term. The manual therapy TSM package only utilized high velocity low amplitude interventions, thus these results cannot be generalized to those seen with the use of other manual therapy techniques. Another limitation is our initial selection of the HDI as our primary outcome measure, which may not have been the optimal tool to assess CeH disability. Our results suggest that the HDI may not be specific enough to identify symptoms relevant to CeH, and in future trials, the NDI is likely a better choice for primary outcome. Finally, our interventions were applied in a very prescriptive fashion which may not reflect the pragmatic nature of manual therapy in clinical practice.

Conclusion

This study found that the use of TSM in individuals with chronic CeH did not significantly change headache disability as measured by the HDI, but it did provide meaningful improvements in pain intensity and neck disability as measured by the NDI. Additionally, TSM resulted in a significantly greater likelihood of patient perceived improvement compared to no treatment. We believe our findings suggest that TSM may be useful in a plan of care for individuals with chronic CeH for reducing pain and neck-related disability.

Biographies

Dr. Amy W. McDevitt is an Associate Professor in the Physical Medicine and Rehabilitation Department at the University of Colorado, Anschutz Medical Campus where she teaches entry-level physical therapist students. Clinically, she practices at the University of Colorado Health, CU Sports Physical Therapy and Rehabilitation. She is a board-certified Orthopaedic Clinical Specialist and a Fellow in the American Academy of Orthopaedic Manual Physical Therapists. She is currently completing her PhD in Physiotherapy through the University of Newcastle, Australia. Dr. McDevitt has an active research agenda including publications and national presentations in the areas of shoulder pain, regional interdependence, dry needling and the assessment of clinical reasoning in DPT students. She is a past recipient of the Rose Excellence in Research Award from the Academy of Orthopaedic Physical Therapy of the American Physical Therapy Association.

Facebook: https://www.facebook.com/awmcdevitt

LinkedIn: https://www.linkedin.com/in/amy-mcdevitt-258711109/

Twitter: @awmcdevittDPT

Dr. Joshua A. Cleland is the Director of Research and Faculty Development in the DPT program at Tufts University. He is actively involved in numerous clinical research studies investigating the effectiveness of manual physical therapy and exercise in the management of spine and extremities disorders. He has published over 280 manuscripts in peer-reviewed journals. He is an Editor for the Journal of Orthopaedic and Sports Physical Therapy. He is currently an author/editor on 4 textbooks. Dr Cleland is a well-known speaker at both the national and international level. He is the recipient of the 2008 Jack Walker Award, the 2009 Eugene Michels New Investigator Award, the 2011 Chattanooga Research Award, the 2015 Rothstein Golden Pen Award for Scientific Writing all from the American Physical Therapy Association. He also received the 2013, 2014 and 2015 Rose Excellence in Research Award from the Academy of Orthopaedic Physical Therapy. Additionally, he received both the Dorothy Baethke-Eleanor J. Carlin Award for Excellence in Academic Teaching and was also selected as a Catherine Worthingham Fellow of the American Physical Therapy Association in 2018.

Twitter: JoshClelandPT

Dr. Daniel I. Rhon is a clinician and active health services researcher. He is an Associate Professor in the Department of Rehab Medicine, Uniformed Services University of Health Sciences and an Assistant professor in the Army-Baylor University, Doctoral Program in Physical Therapy. He completed a postdoc research fellowship through the University of Utah and has a strong research interest in the effectiveness of clinical care pathways for musculoskeletal disease, both at primary and specialty care levels, and the intersection of these two. He is the Primary Investigator on several CDMRP and NIH-funded multi-site trials focused on these clinical problems. He is a past recipient of the Rose Excellence in Research Award from the Orthopaedic Section of the American Physical Therapy Association and the US Army COL Mary Lipscomb Hamerick Lifetime Research Award.

LinkedIn: linkedin.com/in/danrhon

Twitter: @danrhon

Dr. Rebecca A.K. Altic completed her Doctor of Physical Therapy degree at the University of Colorado Anschutz Medical Campus in 2019. She has been involved in orthopedic physical therapy research at the University since 2017. She practices orthopedic and pelvic health physical therapy with UCHealth in Denver, Colorado.

Facebook: Rebecca Kretschmer Altic

LinkedIn: Rebecca Altic

Dr. Drew J. Courtney earned his Doctor in Physical Therapy from the University of Colorado Anschutz Medical Campus and his Bachelor’s degree in Athletic Training from Colorado Mesa University. He is now the head physical therapist at DBC Fitness San Diego, a sports performance and rehabilitation facility in San Diego, California.

Facebook: Drew Courtney

LinkedIn: Drew Courtney

Twitter: @drewcourtneydpt

Dr. Paul E. Glynn has over 25 years experience in the PT field to include acting as the Director of Orthopedic Surgery and Rehabilitation services at Newton-Wellesley Hospital, Director of Physical Therapy for gWell Health inc., and most recently opening a private practice in Lexington, MA. He holds academic appointments at Evidence in Motion as well as the MGH Institute of Health Profession in Boston, MA. He has published research in numerous peer-reviewed journals, has authored multiple textbook chapters, and has co-authored a textbook and application. He is an active researcher and national presenter in the field, a manuscript reviewer for numerous PT journals, an item writer for the PT National Examination, and an associate member of the PT Federation Board. Dr. Glynn has served as the Chair of the Practice Committee for the Massachusetts American Physical Therapy Association and he remains active in multiple local educational groups.

Facebook: https://www.facebook.com/glynnphysicaltherapy/

Twitter: @pauleglynn

Dr. Paul E. Mintken is a Professor in the Physical Medicine and Rehabilitation Department at the University of Colorado, Anschutz Medical Campus. He maintains an active research agenda investigating conservative care for musculoskeletal disorders. He has over 50 peer-reviewed publications and has co-authored 3 eBooks and 7 book chapters. His awards include the Gould Excellence in Teaching Orthopaedic Physical Therapy Award from the American Physical Therapy Association (APTA), the Baethke-Carlin Award for Excellence in Academic Teaching from the APTA, the Rose Excellence in Research Award, The Journal of Orthopaedic & Sports Physical Therapy Excellence in Research Award, the Chattanooga Research Award from the Physical Therapy Journal, and the Outstanding Physical Therapist Award for the State of Colorado from the Colorado chapter of the APTA.

Facebook: Paul Mintken

Twitter: @PMintkenDPT

LinkedIn: Paul Mintken

Appendix A. Thoracic Spine Manipulation Techniques

Technique 1.

Seated Mid-Thoracic Technique: A high-velocity, mid-to-end-range, traction force to the mid-thoracic spine on the lower thoracic spine in a sitting position, in slight flexion with the patient’s arms crossed. The therapist placed his or her upper chest at the level of the patient’s middle thoracic spine and grasped the patient’s elbows. A high-velocity traction force was performed in an upward direction.

Technique 2.

Supine Mid-Thoracic Technique: A high-velocity, end-range, anterior-posterior force applied through the elbows to the flexed middle thoracic spine on the lower thoracic spine in a supine position with patient’s arms crossed. The therapist used his or her hand to stabilize the inferior vertebra of the motion segment targeted and used his or her body to push down through the patient’s arms to perform a high-velocity, low- amplitude thrust directed in the direction of the arrow toward T5 through T8.

Technique 3.

Supine Upper Thoracic Technique: A high-velocity, end-range, anterior-posterior force applied through the elbows to the flexed upper thoracic spine on the middle thoracic spine in a supine position with patient’s arms crossed. The therapist used his or her hand to stabilize the inferior vertebra of the motion segment targeted and used his or her body to push down through the patient’s arms to perform a high-velocity low amplitude force directed in the direction of the arrow toward T1 through T4.

Technique 4a.

Supine Cervicothoracic Junction Technique: A high-velocity, end-range, anterior-posterior force through the elbows to the cervicothoracic junction on the upper thoracic spine in a supine bridged position. The therapist used his or her hand to stabilize the T1 segment and used his or her body to perform a high-velocity, low-amplitude force in the direction of the arrow through the patient’s arms.

Technique 4b.

Seated Cervicothoracic Junction Technique: A high-velocity, end-range, traction force to the cervicothoracic junction on the upper thoracic spine in a sitting position with the patient’s hands interlaced behind their neck. The individual interlocks their fingers at the base of their neck. The therapist weaves his or her arms through the patient’s relaxed arms and places fingers at C7. The therapist supports the patient with compression of their forearms. The therapist leans the individual back until cervicothoracic junction is perpendicular to the floor. The therapist produces a high velocity low amplitude vertical traction force in an upward direction.

Technique 5.

Prone Mid-Lower Thoracic Technique: A high-velocity, mid to end-range, posterior-to-anterior force to the mid-thoracic spine on the upper thoracic spine in a prone position. The therapist achieves a ‘skin lock’ with the pisiforms of each hand over the transverse processes of the target vertebra pushing caudal with one hand and cephalad with the other. The therapist then uses their body to push down through the arms to perform a high-velocity low amplitude posterior-to-anterior force.

Appendix B. Thoracic Mobility Exercise

Figure B1.

a and b. Supine thoracic extension over a towel performed 3 sets of 8–10 repetitions.

Funding Statement

This work was supported by The American Academy of Orthopedic Manual Physical Therapists (AAOMPT) Cardon Rehabilitation Grant

Acknowledgments

The authors would like to acknowledge individuals who contributed to the study including Paige Williams PT, DPT, and Lauren Hinrichs, PT, DPT. We would also like to thank students from the University of Colorado Physical Therapy Program who contributed to this study including, Diana Edwards, Simone Addison and anatomic illustrator Mariah Donofrio.

Disclosure statement

Drs. Cleland, Mintken and Glynn receive honorariums for teaching continuing education courses which include techniques used in this study. All other author(s) do not have conflicts of interest to report.

Grant support

This work was supported by The American Academy of Orthopaedic Manual Physical Therapists (AAOMPT) under a grant from Cardon Rehabilitation (Ontario, Canada). Neither AAOMPT nor the funding agency had a role in the study design, analysis, interpretation, or decisions about publication.

Author contributions

Drs. McDevitt, Cleland, Glynn and Mintken were involved in the conception and design of the study. Drs. McDevitt, Mintken were responsible for recruitment and data collection. Drs. McDevitt, Courtney and Altic managed the data. Drs. McDevitt, Rhon, Cleland and Mintken analyzed and interpreted the data. All authors contributed to the writing and revision of the manuscript and approved the final version of the manuscript.

Data sharing

Individual participant data that underlie the results reported in this article (text, tables and figures) are available, after deidentification, for researchers who provide a methodologically sound proposal. Proposals should be directed to the corresponding author.

Patient and public involvement

The patients and the public were not involved as research partners.