Cha_633.fm

Blackwell Science, LtdOxford, UKCHACephalalgia1468-2982Blackwell Science, 20032412936Original ArticlePhysiotherapy for tension-type headacheP Torelli et al.
Physiotherapy for tension-type headache: a controlled study P Torelli, R Jensen1 & J Olesen1
Headache Centre, Unit of Neurology, Department of Neuroscience, University of Parma, Parma, Italy, and 1Danish Headache Research Centre,
Department of Neurology, University of Copenhagen, Glostrup Hospital, Glostrup, Denmark
Torelli P, Jensen R & Olesen J. Physiotherapy for tension-type headache: a con-trolled study. Cephalalgia 2004; 24:29–36. London. ISSN 0333-1024 The role of non-pharmacological therapies in the preventive treatment of tension-type headache (TTH) is still an object of debate. The primary aim of this studywas therefore to investigate the therapeutic effect of physiotherapy in properlyclassified patients with TTH in a controlled trial. Fifty patients with TTH, 26 withepisodic, frequent (ETTH) and 24 with chronic TTH (CTTH) fulfilling the Inter-national Headache Society classification criteria were included in the study. Aftera 4-week run-in period, they were randomized to either an 8-week period ofstandardized physiotherapy (group 1) or to an 8-week observation period fol-lowed by an identical course of physiotherapy (group 2); after the physiotherapyall patients were followed for a 12-week follow-up period. We then evaluated thenumber of days with headache, severity and duration of the headache attacks,and drug consumption for symptomatic treatment before and after the course ofphysiotherapy. Forty-eight patients completed the study. The average number ofdays with headache per 4-week period was reduced from 16.3 days at baseline to12.3 days in the last 4 weeks of treatment [from 14.5 days to 10.5 days (P < 0.001)in group 1 and from 18.1 days to 14.1 days (P < 0.001) in group 2]. Severity andduration of headache as well as drug consumption were unchanged throughoutthe study. Analysing the response to treatment separately in the various sub-groups, we found that the number of responders was significantly higher amongpatients with CTTH vs. patients with ETTH (P < 0.002) and in females vs. males(P < 0.02). No differences were found between patients with and without disorderof pericranial muscles. We conclude that a standardized physiotherapy pro-gramme has a good therapeutic effect, albeit on a restricted group of patients.
ᮀChronic tension-type headache, episodic tension-type headache, headache,physiotherapy, preventive treatment, tension-type headache Paola Torelli, Centro Cefalee, Sezione di Neurologia, Dipartimento di Neuroscienze, Strada del Quartiere 4, 43100 Parma, Italy. Tel. + 39 0521 23 2866, fax + 39 0521 23 4010, e-mail paolatorelli@libero.it Received 17 December 2002, accepted 22 May 2003 Introduction
now been rejected. In 1988, the InternationalHeadache Society (IHS) applied the current term Tension-type headache (TTH) is the most prevalent ‘tension-type headache’ and defined two subgroups form of headache. In population-based studies in associated and unassociated with disorder of peric- developed countries, the estimated lifetime preva- lence rates range from 35% to 78% of the adult Physiotherapy is widely used in the treatment of population (1, 2). TTH contributes a large burden TTH, but scientific evidence of any possible benefi- of disability, loss of workdays, diminished quality of cial effects is quite limited. Most of the previous life and considerable healthcare costs (3,4).
studies were conducted in improperly classified The older terms ‘muscle contraction headache’, headache groups. It is difficult to apply a placebo- ‘psychogenic headache’ or ‘tension headache’ have controlled design to non-pharmacological treatment Blackwell Publishing Ltd Cephalalgia, 2004, 24, 29–36
studies and most case reports are uncontrolled and unblinded (6) and do not fulfil present-day method-ological standards (7).
The primary aim of the present study was there- fore to investigate the therapeutic effect of physio-therapy in properly classified patients with TTH in a controlled trial. A secondary aim was to determine whether physiotherapy has a different effect onpatients with episodic (ETTH) vs. chronic tension- Figure 1 Study design.
type headache (CTTH), in males vs. females, and inpatients with TTH associated with (MUS) vs. not They were subsequently randomized to either an associated with disorders of pericranial muscles 8-week period of standardized physiotherapy (non-MUS) according to the two- and three-digit (group 1) or to an observation period of similar diagnostic levels of the IHS classification (5). In par- length with clinic visits to the neurologist (R.J.) at ticular, we aimed to characterize those patients who week 4 and week 8 (group 2, see Fig. 1). Thereafter, benefited from physiotherapy by evaluating any group 2 received the same course of physiotherapy possible factors that might predict a good therapeu- as group 1. At the end of the physiotherapy tic response. Furthermore, we aimed to evaluate the course, all patients were evaluated in a 12-week long-term effect of physiotherapy during a 12-week instructed to fill in the diary (10) at the end of eachday with headache and to record the mean sever- Materials and methods
ity of pain on a 0–3 scale, where 0 was no pain and3 was severe, debilitating pain that required bed rest. A standard dose of analgesics was defined asa dose equivalent to 1000 mg of acetylsalicylic acid Fifty patients with frequent ETTH (arbitrarily defined as ≥8 and <15 days with TTH per month)or CTTH fulfilling the IHS criteria (5) were included in the study. The patients were recruitedfrom the out-patient Headache Clinic at the Gen- Pericranial tenderness was evaluated prior to the tofte University Hospital in consecutive order and treatment or observation period by palpation of nine were given a complete physical and neurological pairs of pericranial muscles and of tendon insertions examination. Inclusion criteria were: duration of in group 1 and in group 2, respectively (11). The TTH for at least 1 year, age between 18 and manoeuvre was carried out by the same person – a 70 years, and no use of any preventive headache trained technician – throughout the study, following therapy. Exclusion criteria were: migraine attacks a standardized procedure. The trained technician more than once a month, other forms of primary was kept uninformed about the subjects’ headache headache or neuralgia, other neurological, systemic history. Tenderness was scored in each location or psychiatric disorders, and analgesic medication according to an ordinal scale from 0 to 3, and scores abuse (more than 100 tablets per month) or fre- from all sites were summated. This Total Tenderness quent use of ergotamine or triptans (more than two Score (TTS) system has proved to be reliable (12). We doses per month). Prior experience with physio- had previously demonstrated that the ideal cut-off therapy was not an exclusion criterion. Informed value for separating TTH subjects from non-head- consent was obtained from the patients and the ache subjects with respect to muscle tenderness was study was approved by the local Ethics Committee.
the 75% quartile of TTS obtained from a general This study is part of a multifaceted study of TTH, population (13). In the following studies, TTS was other parts of which have been published previ- used as the only criterion for further subdividing patients according to the IHS criteria at the three-digit level. Those with TTS > 9 (equal to the 75% quartile of TTS from healthy controls) were classifiedas having an association with muscular disorders All patients had to keep a diagnostic headache (MUS), whereas those with a TTS value of £9 were diary (10) during the 4-week run-in period to classified as unassociated with such disorders (non- ensure that they fulfilled the inclusion criteria.
Blackwell Publishing Ltd Cephalalgia, 2004, 24, 29–36
Physiotherapy for tension-type headache physiotherapy, the last 4 weeks of the observationperiod and the baseline in group 2. We have verified, The course of physiotherapy was designed to give in group 2, the modification in the same parameter patients individual treatment twice a week for between the baseline and the last 4 weeks on the 4 weeks and then physical exercise for another observation period. Responders were defined as 4 weeks in small groups together with four to six those patients with >50% reduction in headache other patients. The patients followed both a fixed days in the last 4 weeks of treatment compared with standardized programme consisting of initial mas- the run-in (baseline) period as indicated in the IHS sage, basic relaxation techniques (autogenic training guidelines for clinical trials in TTH (7). Secondary and cognitive-behavioural therapy) and smooth outcome parameters were the average duration and stretching, and a daily programme to be done at home severity of attacks and drug consumption for symp- acting on shoulder, neck and pericranial muscles.
tomatic treatment. The analysis was first carried out Self-responsibility and activity were emphasized and in the whole sample and then repeated in ETTH vs.
any form of passive treatment, such as cryotherapy, CTTH patients, in males vs. females and in MUS vs.
ultrasound and transcutaneous electrical nerve stim- ulation (TENS) were to be avoided. The physiother- The Wilcoxon signed rank sum test and the Mann– apist tried to instruct the patients on how to identify Whitney test were used to compare paired and and avoid any possible muscular stress factors in unpaired data between subjects, respectively. The their daily life. One of the purposes of this treatment Fisher’s exact test (two-sided) was used to compare was to teach patients to be active and learn how to the percentages of responders in ETTH vs. CTTH handle, control and hopefully avoid their pain. The patients, in males vs. females and in MUS vs. non- procedures were performed by an expert physiother- MUS patients. All data were analysed with SPSS apist who was blinded to the subjects’ headache his- software, version 10.0 (SPSS Inc., Chicago, IL, USA).
tory and the presence of muscular disorders.
A 5% level of significance was used.
Results are presented as mean values and standard Forty-eight patients completed the treatment period deviations. For data analysis, the clinical parameters and their demographic and clinical characteristics were evaluated in the run-in period, in the last 4 weeks of physiotherapy and in the last 4 weeks of Participant flow through the trial is displayed in follow-up in both groups, and in the last 4 weeks of Fig. 2. At baseline, group 1 and group 2 did not differ the observation period for group 2. The primary effi- significantly with respect to number of days with cacy parameter was the number of days with head- headache, severity and duration of headache, and analgesic medication used (see Table 2). At the 8- difference in the number of days with headache week evaluation, the number of drop-outs did not between the last 4 weeks of physiotherapy and the differ between the two groups – one female patient baseline in group 1, between the last 4 weeks of did not complete the treatment period because of a Table 1 Clinical characteristics of the 48 patients studied
Data are presented as number with percentage within parentheses; age, headache duration, total tenderness score and number of days with headache in the 4 weeks of run-in are presented as mean values with ranges within parentheses.
Blackwell Publishing Ltd Cephalalgia, 2004, 24, 29–36
Figure 2 Participant flow in the study.
Table 2 Clinical characteristics in Groups 1 and 2
Figure 3 Effect of physiotherapy on headache frequency.
*P < 0.001; **P < 0.001.
Data are presented as number with percentage within parentheses; number of days with headache, headache sever-ity and duration, drug consumption are presented as mean of days with headache per 4-week period was values with ranges within parentheses.
reduced from 16.3 days at baseline to 12.3 days in the Days, number of days with headache/28 days; severity, 0– last 4 weeks of treatment [from 14.5 days to 10.5 days 3 scale; duration, hours of headache/24 h; drugs, number of (P < 0.001) in group 1 and from 18.1 days to 14.1 days tablets per 28 days expressed in equivalent dose of 1000 mg (P < 0.001) in group 2]. There was no difference between the run-in period (18.1 days) and the obser-vation period (17.8 days), while the significant reduc- poor response to treatment after 4 weeks (group 1), tion in the average number of days with headache while the other patient dropped out for personal per 4-week period was confirmed between the obser- reasons unrelated to the study (group 2).
vation period (17.8 days) and after the treatment Furthermore, in the follow-up period, 11 patients (14.2 days) in group 2 (P < 0.001). The effect was main- (five males and six females) dropped out because tained throughout the 12-week follow-up period, they failed to show up regularly for the planned with an average headache frequency of 8.6 and follow-up visits. The clinical characteristics at baseline 14.2 days in group 1 and in group 2, respectively (see of the 11 drop-out patients did not differ significantly.
The severity and duration of headache were unchanged throughout the study periods in bothgroups. The severity was 1.6 at baseline and after The number of days with headache was significantly physiotherapy and 1.8 in the follow-up period in decreased after physiotherapy. The average number group 1; it was 1.7 at baseline and in the observation Blackwell Publishing Ltd Cephalalgia, 2004, 24, 29–36
Physiotherapy for tension-type headache period, 1.6 in the last 4 weeks of treatment and 1.7 respectively). Analysing the response to treatment during the follow-up period in group 2. The dura- separately in the various subgroups, we found that tion of headache varied from an average of 9.8 h at the number of responders was significantly higher baseline and 9.4 h in the follow-up in group 1 and among patients with CTTH (13/24) vs. patients with ETTH (1/24) (P < 0.002) and in females (12/33) vs.
Drug consumption, at doses equivalent to a 1000- males (2/15) (P < 0.02). In the female population, the mg dose of ASA, was significantly reduced only in proportion of CTTH responders (11/14) was signifi- the follow-up period vs. baseline for both groups cantly higher than that of ETTH responders (1/19) (P < 0.001): the average analgesic use was 19.7 tab- (P < 0.002), while in the male population the propor- lets per 28 days expressed in equivalent dose for tion of responders did not differ significantly group 1 and 23.6 for group 2 at baseline, and 14.8 between CTTH (2/10) and ETTH (0/5) (see Table 3).
for group 1 and 12.9 for group 2 during follow-up, Of the 11 female responders with CTTH, nine (82%) respectively. No significant differences were found had TTH MUS and the remaining two had TTH non- in medication use in both groups between baseline MUS. Of the two male responders with CTTH, only and the treatment period, the mean dose taken dur- one (50%) had TTH MUS (P < 0.06). In the CTTH ing physiotherapy being 18.5 for group 1 and 22.3 non-responder group, six of the eight men (75%) had TTH MUS, compared with one of the three women(33%). In the ETTH non-responder group, 13 of the 18 women (72%) had TTH MUS, compared with twoof the five men (40%). The only responder among The number of days with headache was reduced in ETTH patients was a woman who had TTH non- the last 4 weeks of physiotherapy in both groups, but MUS. No significant difference was found between significantly so only in CTTH patients (P < 0.001).
MUS patients (10/32), all of them with CTTH (nine Among ETTH patients, the number of days with female and one male), and non-MUS (4/16) patients headache was reduced from 9.5 in the run-in period (P = 0.69), three with CTTH (two female and one to 8.1 in the treatment period and to 7.5 in the follow- male) and one woman with ETTH. The duration of up period in group 1, but the reduction was not attacks (10.4 vs. 14.4 h, P < 0.01) and drug consump- significant (P = 0.57); group 2 exhibited the same pat- tion (10.7 vs. 25.4 standard doses, P < 0.01) at base- tern as group 1. Among CTTH patients, the effectwas maintained throughout the 12-week follow-up period, whereas the number of days with headache was largely unchanged in ETTH patients (P = 0.36, see Fig. 4). The other clinical parameters considered in the study, i.e. headache severity and duration anddrug consumption, did not show any significant variations within ETTH and CTTH patients.
The number of days with headache was significantly reduced after physiotherapy in both groups to the same extent (MUS, P < 0.002; non-MUS, P < 0.002) and the effect was maintained throughout the 12-week follow-up period (see Fig. 5). The analysis of the other clinical parameters did not show any dif- ferences between the periods considered within In 29% of the sample (14 patients out of 48), the number of attacks was reduced by >50% (7). The Figure 4 Effect of physiotherapy on headache frequency in
number of responders did not vary significantly episodic and chronic tension-type headache patients. , between group 1 and group 2 (9/24 and 5/24, Chronic TTH; ᭿, episodic TTH. *P < 0.001; **P < 0.001.
Blackwell Publishing Ltd Cephalalgia, 2004, 24, 29–36
line was significantly lower in responders vs. non- Among non-responders, 21% (10/48) showed a percentage reduction of attacks between 25% and50%, 25% (12/48) presented with an unchanged clin- ical course, and the remaining 25% (12/48) exhibited Discussion
Use of complementary non-pharmacological thera- pies in headache has increased considerably over the last few years. A national population-based survey conducted in the USA to determine the prevalence, costs and pattern of use of unconventional therapieshas shown that about one-third of the subjects investigated had followed some form of non- pharmacological treatment in the year prior to the survey and that the cost of such treatments was the same as the annual cost of all hospitalizations in the USA (14). In the field of preventive treatments for headache, systematic reviews and meta-analyseshave been reported, but the interventions studied Figure 5 Effect of physiotherapy on headache frequency in
were confined to pharmacological and cognitive/ tension-type headache with (MUS) and without (non-MUS) muscular disorder. , MUS+ ᭿, non-MUS. *P < 0.002; behavioural therapies and mostly to the study of migraine (15). Despite the high prevalence of TTH in Table 3 Distribution of responders (defined as patients with >50% reduction in headache days) and non-responders
Data are presented as number with percentage within parentheses.
*P < 0.02; **P < 0.002; ***P < 0.002.
Table 4 Clinical features of responders (defined as patients with >50% reduction in headache days) and non-responders
Drug consumption – standard doses/4 weeks Data are presented as mean values with ranges within parentheses.
Data are presented as mean values at baseline.
*P < 0.01; **P < 0.01.
Blackwell Publishing Ltd Cephalalgia, 2004, 24, 29–36
Physiotherapy for tension-type headache average duration of attacks and analgesic use, both assessed at baseline, were significantly lower in responders vs. non-responders. In spite of the factthat pericranial muscular disorders were slightlymore frequent in responders, there was also some response to treatment in patients without significant tenderness in the pericranial muscles. Separate dataanalysis for males and females shows that CTTH Figure 6 Percentage distribution of effect.
women with muscular disorders were thoseresponding better to treatment. CTTH males with the general population, the utilization rates of non- muscular disorders and ETTH patients, whether pharmacological therapies in TTH sufferers remain male or female, with or without muscular disorders, poorly studied. Only a few studies have assessed responded poorly to physiotherapy. These data need the effectiveness of traditional physiotherapy tech- to be confirmed by studies on larger samples that niques of heat, massage, traction, specific exercise or should consider the different parameters identified.
postural education on TTH (6). Using a controlled, Although current knowledge of the mechanisms randomized design, our study was able to demon- of action of physiotherapy and the pathophysiolog- strate a significant reduction in the number of days ical processes underlying TTH are still unclear, we with headache following physiotherapy, albeit in a cannot rule out a priori the existence of a relation small sample of patients. Our findings are in agree- between the beneficial effect of physiotherapy and ment with those of Carlsson et al. (16), who com- the pathophysiology of TTH. The most prominent pared physiotherapy treatment with acupuncture in abnormal finding in TTH is a considerable increase 60 female CTTH patients defined according to the in tenderness of the pericranial myofascial tissues criteria set by the Ad Hoc Committee on Classifica- (11, 19, 20). ETTH and CTTH may or may not share tion of Headache (17). In that trial, ‘physiotherapy’ pathogenic mechanisms and it has been suggested consisted of a variety of individualized patient- that CNS sensitization may contribute significantly initiated modalities, including relaxation techniques, to increased tenderness and to chronification of TTH stretching, TENS and ice therapy, as well as educa- (19). Furthermore, a generalized hypersensitivity to tion regarding muscle tension and how to control it pain probably also contributes to chronic pain in ‘autogenically’. Both treatments resulted in benefits these patients (21). The therapeutic efforts in phys- for the patients’ mood state and overall health func- iotherapy are primarily designed to produce muscle tion, as well as for the number of days with head- relaxation, but it may also activate central anti- ache, even though no exact figures were provided nociceptive processes. It seems therefore reasonable for the latter. The decreased severity of headache to assume that, at least in CTTH, there is some assessed on the Visual Analogue Scale in those convergence between the mechanisms activated by patients who had been treated with physiotherapy physiotherapy and the systems underlying TTH. In is different from our study results, which demon- CTTH central sensitization has been shown to play strate rather stable severity and duration of attacks a significant role. It is tempting to speculate that the greater effect in CTTH than in ETTH could be due By contrast, our observations confirm those to reduced central sensitization and/or increased reported by Hammill et al. (18), who assessed the anti-nociceptive action. On the other hand, it is not effectiveness of a treatment programme encompass- possible to rule out that the effect of physiotherapy ing ergonomic and postural education in an open is also mediated by a peripheral mechanism that acts uncontrolled study of 20 patients. A significant and by relaxing the muscles and also increasing local long-lasting decrease in headache frequency and blood flow to the affected area that could reduce the drug consumption was noted, although there was no local concentration of pain mediators (22). Muscle effect on headache severity and headache duration.
stretching is also reported to reduce the excitability Given the poor efficacy and the numerable side- of the motorneurone pool which may lead to effects of preventive pharmacological TTH treat- decreased muscle tone and pain (23).
ment, the clinical characterization of headache is Our study results are not totally unbiased, as: (i) certainly of great interest in those patients, predom- it is apparent that it is difficult to apply a placebo- inantly women, who reported >50% decrease in the controlled, double-blind design methodology in frequency of attacks following physiotherapy. Most non-pharmacological treatments such as physiother- of these responders were CTTH sufferers, and the apy, and our study is controlled but not blinded; (ii) Blackwell Publishing Ltd Cephalalgia, 2004, 24, 29–36
in TTH it is extremely important to consider the 8 Jensen R. Mechanisms of spontaneous tension-type head- placebo effect, even though the lack of response in aches: an analysis of tenderness, pain threshold and EMG.
ETTH and the absolute lack of significant variations in the clinical course during the observation period 9 Jensen R, Olesen J. Initiating mechanisms of experimen- tally induced tension-type headache. Cephalalgia 1996; compared with the run-in period seem to rule out any relevant placebo effect in therapeutic response; 10 Russell MB, Rasmussen BK, Brennum J, Iversen HK, and (iii) physiotherapy is mainly based on tradition Jensen RA, Olesen J. Presentation of a new instrument: the and experience and the applied treatment strategies diagnostic headache diary. Cephalalgia 1992; 12:369–74.
are certainly not evidence based. In our study, a 11 Langemark M, Olesen J. Pericranial tenderness in tension standardized therapeutic approach in a controlled headache. A blind controlled study. Cephalalgia 1987; design was applied for the first time, but an absolute 12 Bendtsen L, Jensen R, Jensen JK, Olesen J. Pressure- blinded, controlled study design evaluating the indi- controlled palpation: a new technique which increases the vidual treatment modalities in physiotherapy is still reliability of manual palpation. Cephalalgia 1995; 15:205–10.
13 Jensen R, Rasmussen BK. Muscular disorders in tension- type headache. Cephalalgia 1996; 16:97–103.
14 Eisenberg DM, Kessler RC, Foster C, Norlock FE, Calkins References
DR, Delbanco TL. Unconventional medicine in the UnitedStates. Prevalence, costs, and patterns of use. N Engl J Med 1 Rasmussen BK, Jensen R, Schroll M, Olesen J. Epidemiol- ogy of headache in a general population – a prevalence 15 Pryse-Phillips WEM, Dodick DW, Edmeads JD. Guide- study. J Clin Epidemiol 1991; 44:1147–57.
lines for non-pharmacological management of migraine in 2 Pryse-Phillips W, Findlay H, Tugwell P, Edmeads J, Mur- clinical practice. Can Med Assoc J 1998; 159:47–54.
ray TJ, Nelson RF. A Canadian population survey on the 16 Carlsson J, Augustinsson LE, Blomstrand C, Sullivan M.
clinical, epidemiological and societal impact of migraine Health status in patients with tension headache treated and tension-type headache. Can J Neurol Sci 1992; 19:333– with acupuncture or physiotherapy. Headache 1990; 3 Rasmussen BK, Jensen R, Olesen J. Impact of headache on 17 Ad Hoc Committee on Classification of Headache of the sickness absence and utilisation of medical services: a Danish population study. J Epidemiol Commun Health 18 Hammill JM, Thomas MC, Rosecrance JC. Effectiveness of a physical therapy regimen in the treatment of tension- 4 Edmeads J, Findlay H, Tugwell P, Pryse-Phillips W, Nel- type headache. Headache 1996; 36:149–53.
son RF, Murray TJ. Impact of migraine and tension-type 19 Jensen R. Pathophysiological mechanisms of tension-type headache on life-style, consulting behaviour, and medica- headache: a review of epidemiological and experimental tion use: a Canadian population survey. Can J Neurol Sci studies. Cephalalgia 1999; 19:602–21.
20 Jensen R, Rasmussen BK, Pedersen B, Olesen J. Cephalic 5 Headache Classification Committee of the International muscle tenderness and pressure pain thresholds in head- Headache Society. Classification and diagnostic criteria for ache: a population study. Pain 1993; 52:193–9.
headache disorders, cranial neuralgias and facial pain.
21 Bendtsen L. Central sensitization in tension-type headache Cephalalgia 1988; 8 (Suppl. 7):1–96.
– possible pathophysiological mechanisms. Cephalalgia 6 Vernon H, McDermaid M, Hagino C. Systematic review of a randomized clinical trial of complementary/alternative 22 Ernst E, Fialka V. The clinical effectiveness of massage therapies in the treatment of tension-type and cervico- therapy – a critical review. Forsch Komplementarmed genic headache. Complementary Therapies Med 1999; 23 Carlson CR, Collins FR, Nitz AJ, Sturgis ET, Rogers JL.
7 Schoenen J. Guidelines for trials of drug treatment in ten- Muscle stretching as an alternative relaxation training pro- sion-type headache. Cephalalgia 1995; 15:165–79.
cedure. J Behav Ther Exp Psychiat 1990; 21:29–38.
Blackwell Publishing Ltd Cephalalgia, 2004, 24, 29–36

Source: http://heigo.planet.ee/matu/11715900.pdf