Neck–Shoulder Region Training for Chronic Headache in Women: A Randomized Controlled Trial

Objectives We investigated whether a specific exercise program for the neck–shoulder region reduces headache intensity, frequency, and duration, and how it influences neck disability among women with chronic headache compared to a control group. Design Two-center randomized controlled trial. Subjects 116 working-age women. Intervention The exercise group (n = 57) performed a home-based program with six progressive exercise modules, over 6 months. The control group (n = 59) underwent six placebo-dosed transcutaneous electrical nerve stimulation sessions. Both groups performed stretching exercises. Main measures The primary outcome was pain intensity of headache, assessed using the Numeric Pain Rating Scale. Secondary outcomes were frequency and duration of weekly headaches, and neck disability assessed using the Neck Disability Index. Generalized linear mixed models were used. Results Mean pain intensity at baseline was 4.7 (95% CI 4.4 to 5.0) in the exercise group and 4.8 (4.5 to 5.1) in the control group. After 6 months the decrease was slight with no between-group difference. Headache frequency decreased from 4.5 (3.9 to 5.1) to 2.4 (1.8 to 3.0) days/week in the exercise group, and from 4.4 (3.6 to 5.1) to 3.0 (2.4 to 3.6) in the control group (between-group p = 0.017). Headache duration decreased in both groups, with no between-group difference. Greater improvement in the Neck Disability Index was found in the exercise group (between-group change −1.6 [95% CI −3.1 to −0.2] points). Conclusion The progressive exercise program almost halved headache frequency. The exercise program could be recommended as one treatment option for women with chronic headache.


Hypothesis
Vy'e assume the therapeutic exercise training including stretching exercises (THEX intervention group) offers a benefit to headache-suffering working women over the effect of placebo-dose TENS treatment including stretching exercises (control group).
Commitments for data analysis l) To be deemed effective, the THEX intervention should provide a statistically significant benefit (p < 0.05) over the placebo-dose TENS treatment in the primary (average intensity per week; Visual Analogue Scale) and secondary (frequency; times per week, and duration; hours per week) outcomes of headache, assessed at0,3, and 6 months during the six-month RCT. 2) For the observed THEX effect to be deemed clinically relevant, THEX should also provide > 15 mm greater reduction in VAS intensity than TENS, and one headache attack less per week than TENS (based on the literature that is provided as supplementary material at the end of this document).
Statistical analysis plan a) For the primary and secondary outcomes, intention-to-treat (ITT) is the primary data analysis. Ancillary analyses will also be carried out. Efficacy analysis will be performed to evaluate whether there are subgroups that benefit from the program: for example, whether patients with high training adherence benefit more than patients with low training adherence, or whether some types of headachê, ê.g., migraine, either substantially benefit from the program or the program increases their headache intensity, frequency, and/or duration (a potential adverse effect of the intervention). b) Muscle strength and neck and shoulder flexibility will be also analyzed, but only as explanatory outcomes to provide support (theoretical basis and mechanisms) for possible significant findings in the analysis of the primary and secondary outcomes presented above. This means that between-group differences in the changes of disability, fear avoidance beliefs, work ability, quality of life, fatigue, depressive symptoms, ergonomics at work, and leisure time physical activity will be assessed in the interpretation of the findings. However, these analyses are used only as background information and explanatory outcomes, and to measure and interpret a potential intervention effect throughout the musculoskeletal system and in mental health in addition to social participation in society according to the International Classification of Functioning, Disability and Health (ICF), a framework for measuring health and disability. Since we utilize several outcomes in our analyses, we also use Benjamini & Hoecberg's False Discovery Rate to avoid type I statistical error. c) Ancillary analyses will be also performed using three Generclized Linear Mixed Models (GLMM) for each outcome: l) unadjusted,2) adjusted by age, type of work (office vs. other), smoking (no vs. yes), and 3) adjusted by age, type of work, smoking, use of headache medication (weekly frequency), work ability index, þpe of headache (migraine vs. cervicogenic headache), hormone therapy (no vs. yes), menstruation status (normal, irregular, or perimenopausal/postmenopausal). Data will be analyzed as longitudinal with 3 time points (baseline, 3 months and 6 months). Outcomes with only 2 measurements will be analyzed with Generalized Linear Models (GLM) with the same models as presented above.

Analysis protocol
Groups were blinded and coded as Group E and Group S by MR (who was not blinded to the data) who then delivered the data to the statistician (KT) and last author (RN). The last author prepared the first draft of the two interpretations of the blinded results based on the intention to-treat analysis approach.
Blinded review of the data: A writing committee meeting (May 27,2015) The Writing Committee of the THERAPEUTIC EXERCISE TRAINING TO REDUCE CHRONIC HEADACHE IN WORKING WOMEN trial (undersigned) discussed the two interpretations of the results on the basis of a blinded review (Group E compared to Group S) with one approach assuming that Group E was the therapeutic exercise training + stretching (THEX) group, and the other assuming that Group E was the placebo-dose TENS + stretching (TENS) group. The first author (MR) participated in the meeting as a secretary but did not participate in preparing the interpretation of the results. Based on these theoretical commitments, our interpretation of the findings are as follows: a) If the THEX group is found superior to the TENS group during the six-month intervention (clinically relevant improvement in headache pain vs. placebo, as defined above), the study is applicable to the patient group studied, but its generalizability remains unknown. b) If the THEX group is not found superior to the TENS group, the study suggests that therapeutic exercise training with six booster sessions to support home-based training during the six-month intervention period does not work for outpatients suffering from chronic cervicogenic headache and migraine in a pragmatic real-life situation. This could be due to an insufficient training amount and intensity in the THEX group or an increased leisure time physical activity amount in the TENS group, all of which will be analyzed from the explanatory outcomes. If this is the case, an efficacy design would be warranted to fully assess the efficacy of THEX (under optimal circumstances).
However, given the high external validity, such a finding (of no effectiveness) would strongly suggest that THEX does not work in ordinary ouþatient settings in primary healthcare. This assumption requires that the training adherence in the intervention group has been suffrcient to achieve a decline in headache intensity. c) If the TENS group is found superior to the THEX goup during the six-month intervention, we will conclude that the specific therapeutic exercise training (THEX) does not provide any benefit to the ouþatient goup suffering from chronic cervicogenic and tension-type headache with migraine compared to very low-dose (placebo) Transcutaneous Electrical Nerve Stimulation (TENS). Thus, the specific therapeutic exercise training program should not be recommended to the patient goup studied. Based on the blinded results, our interpretation of the findings is: Interpretation l. The THEX group was Group E and thus it was superior to the TENS Group (Group S): We found that the six-month therapeutic exercise training (THEX) decreased headache frequency clinically relevantly compared to the lowdose Transcutaneous Electrical Nerve Stimulation treatment (TENS). In detail, participants in the THEX group reduced their number of weekly headaches from 4.6 to 2.4, while the TENS group participants reduced their number of weekly headaches from 4.3 to 3.1 (p:0.017). The effect size for reduction of headache frequency was 0.53. THEX did not reduce headache intensity and duration. The training did not affect the use of headache medications and overall quality of life. Thus, because of the reduced number of headache episodes, therapeutic training (THEX) can be recommended as an effective treatment for outpatients with headaches, such as patients with cervicogenic and tension-type headache with migraine.
Interpretation 2. The TENS group was Group E and thus it was superior to THEX Group (Group S): We found that the (placebo) Transcutaneous Electrical Nerve Stimulation (TENS) did not reduce headache intensity and duration. However, it reduced headache frequency clinically relevantly when compared to the therapeutic exercise training (THEX). In detail, the placebo group reduced their number of weekly headaches from 4.6 to 2.4, while the THEX goup reduced their number of weekly headaches from 4.3 to 3.1 (p:0.017). The effect size for the reduction of headache frequency was 0.53. The placebo TENS did not affect the use of headache medications and overall quality of life. However, placebo treatment cannot be recommended as a serious therapy for clinical practice.  It has been shown that energy storage in muscle mitochondria is lower in patients suffering from chronic neck pain than in healthy individuals (Ylinen 2004). Moreover, local blood circulation inside muscles is reduced resulting in compromised oxygen transportation and overall metabolism. Also, muscle strength is lower in chronic neck pain patients compared to healthy individuals (Ylinen 2004).In addition, muscle atrophy and increased amount of fat tissue inside muscles has been observed in chronic neck patients compared to healthy individuals (Ylinen 2007).
Thus, principles of therapeutic exercise in successfully treating cervicogenic headache could also utilize therapeutic exercise and consist of low load and low intensity proprioceptive training at the early phase and specific high-load and high intensity exercises in the later pÌrase of the program, or the former or the latter throughout the treatment period (Jull 2002, Busch et al. 2008, Fricton et aL.2009, Ylinen et al. 2010). On the other hand, current level of evidence to treat headache effectively suggest that the level of evidence is low. Moreover, there is not information on owhether and which physiotherapy approach is effective' (Luedtke et al. 2015).      Lora?