Protocol, design, and randomisation

ART-TTH was a randomised multicentre trial comparing acupuncture, minimal acupuncture, and a no acupuncture waiting list condition. Minimal acupuncture served as a sham intervention; we included the additional no acupuncture waiting list control because minimal acupuncture is not a physiologically inert placebo. Patients were blinded to treatment in the acupuncture and minimal acupuncture arms of the study. Two blinded evaluators analysed headache diaries. The methods of the trial have been described in detail elsewhere.4

After a baseline phase of four weeks, we used a centralised telephone randomisation procedure (random list generated with sample size 2.0 by the statistician) to randomise patients, stratified by centre (block size 12 unknown to trial physicians), in a 2:1:1 ratio (acupuncture:minimal acupuncture:waiting list). We used the 2:1:1 ratio to facilitate recruitment and increase the compliance of trial physicians. All study participants provided written, informed consent, and the study conformed to common guidelines for clinical trials (Declaration of Helsinki, ICH-GCP, including certification by external audit).

Participants

Inclusion criteria were a diagnosis of episodic or chronic tension-type headache according to the criteria of the International Headache Society,1 at least eight days with headache a month in the previous three months and in the baseline period, age 18-65 years, duration of symptoms at least 12 months, completed baseline headache diary, and written informed consent. Main exclusion criteria were additional migraine headache, secondary headaches, start of headaches after age 50, use of analgesics on more than 10 days a month, prophylactic headache treatment with drugs during the previous four weeks, and any acupuncture treatment during the previous 12 months or at any time if done by the participating trial physician. Most participants were recruited through reports in local newspapers; a minority were patients who spontaneously contacted trial centres.

Interventions

We developed the study interventions in a consensus process with German acupuncture experts and societies.4 Physicians trained (at least 140 hours, median 500 hours) and experienced (median 10 years) in acupuncture delivered the interventions. Both the acupuncture and minimal acupuncture treatments consisted of 12 sessions of 30 minutes, given over eight weeks (preferably two sessions in each of the first four weeks, followed by one session a week in the remaining four weeks).

Acupuncture treatment was semistandardised. All patients were treated at “basic” points bilaterally unless explicit reasons for not doing so were given; additional points could be chosen individually (box 1). We instructed physicians to achieve “de qi” (an irradiating feeling considered to be indicative of effective needling) if possible and to stimulate needles manually at least once during each session. The total number of needles was limited to 25 per session.

The number, length, and frequency of the sessions in the minimal acupuncture group were the same as for the acupuncture group. In each session, physicians needled at least five out of 10 predefined distant non-acupuncture points (box 2) bilaterally (at least 10 needles) and superficially using fine needles. Physicians avoided “de qi” and manual stimulation of the needles.

Patients in the waiting list control group did not receive any prophylactic treatment for their headaches for a period of 12 weeks after randomisation. After that time, they received 12 sessions of the acupuncture treatment described above.

All patients were allowed to treat acute headaches as needed. Treatment was supposed to follow current guidelines5 and had to be documented in the headache diary.

Patients were informed with respect to acupuncture and minimal acupuncture as follows: “In this study, different types of acupuncture will be compared. One type is similar to the acupuncture treatment used in China. The other type does not follow the principles of traditional Chinese medicine, but has also been associated with positive outcomes in clinical studies.”

Outcome measurement

All patients filled in headache diaries in the four weeks before randomisation (baseline phase), the 12 weeks after randomisation, and weeks 21 to 24 after randomisation. In addition, we asked patients to fill in a pain questionnaire before treatment, after 12 weeks, and after 24 weeks.6 This included the following validated scales: the German version of the pain disability index,7 a scale for assessing sensoric and affective aspects of pain (Schmerzempfindungs-Skala SES),8 the ADS depression scale,9 and the German version of the SF-36 to assess health related quality of life.10 The primary outcome measure was the difference in number of days with headache between the four weeks before randomisation (baseline phase) and weeks nine to 12 after randomisation.

To test blinding to treatment and assess the credibility of the different treatment methods, patients filled in a credibility questionnaire after the third acupuncture session.11 At the end of the study, patients were asked whether they thought that they had received acupuncture following the principles of Chinese medicine or the other type of acupuncture.

Statistical methods

We based confirmatory testing of the primary outcome measure and all main analyses (with SPSS 11.5) on the intention to treat population and used all available data. We used SOLAS 3.0 (Statistical Solutions, Cork, Ireland) to do sensitivity analyses for the primary outcome measure, replacing missing data with baseline values or multiple imputation. We tested a priori ordered two sided null hypotheses by using Student's t test (significance level 0.05). In the first step we investigated whether acupuncture reduced the number of days with headache more than no treatment, and in the second step (only if the first null hypothesis was rejected) we investigated whether acupuncture was more efficacious than minimal acupuncture. We give exploratory analyses (analysis of covariance adjusting for baseline differences and χ² tests) for predefined secondary outcome measures. We did an additional per protocol analysis including only patients without major protocol violations until week 12.

We made the original sample size calculation for one sided testing. Under this premise we planned the study to have 80% power to detect a group difference of two days with headache assuming a standard deviation of five days (thus an effect size of 0.4) and a 20% dropout rate.4 However, we later decided to use two sided testing to comply better with common standards. Before starting the analysis, and on the basis of the recommendation of the ethical review board, we decided to exclude the data from one centre that had included 26 patients, owing to repeated severe protocol deviations and the suspicion of data manipulation in some patients. We decided to do a sensitivity analysis including this centre's data.

Participants, treatment, and blinding

Between March 2002 and January 2004, approximately 2700 patients with headache expressed an interest in participating in the study; 524 entered the four week baseline period, and 296 patients recruited in 29 outpatient centres were randomised (fig 1). As described above, we excluded from the main analysis one trial centre with 26 patients, leaving 270 patients in the intention to treat population (132 acupuncture, 63 minimal acupuncture, 75 waiting list).

View larger version:

• In a new window

• Download as PowerPoint Slide

Fig 1

Flowchart of trial (ITT=intention to treat; MOM=main outcome measure; PP=per protocol). ^*Before starting analyses the data from one trial centre were excluded from the main analysis owing to severe protocol violations and suspicion of data manipulation; the data were included in an additional sensitivity analysis (see methods section)

Groups were comparable at baseline in most respects (table 1). However, we observed some differences in previous use of acupuncture and in parts of the pain questionnaire. After three sessions, patients rated the credibility of acupuncture and minimal acupuncture very highly and very similarly (table 2). At the end of the study, patients' guesses as to their allocation status did not differ significantly between groups, but patients in the acupuncture group guessed their allocation correctly slightly more often than did patients in the minimal acupuncture group.

Effectiveness

From baseline to week 9-12, the number of days with headache decreased by 7.2 (SD 6.5) days in the acupuncture group compared with 6.6 (SD 6.0) days in the minimal acupuncture group and 1.5 (SD 3.7) days in the waiting list group (difference: acupuncture v minimal acupuncture, 0.6 days, 95% confidence interval -1.5 to 2.6 days, P = 0.58; acupuncture v waiting list, 5.7 days, 4.2 to 7.2 days, P < 0.001) (fig 2). Results were very similar in sensitivity analyses and in the per protocol analysis. The proportion of responders (at least 50% reduction in headache days, counting all patients with missing data as non-responders) was 46% in the acupuncture group, 35% in the minimal acupuncture group, and 4% in the waiting list group (exploratory P values 0.163 for acupuncture v minimal acupuncture and < 0.001 for acupuncture v waiting list).

View larger version:

• In a new window

• Download as PowerPoint Slide

Fig 2

Mean (95% confidence interval) number of days with headache. Patients in the waiting list group received acupuncture after week 12 (dotted bars)

Compared with the waiting list control group, patients receiving acupuncture or minimal acupuncture fared significantly better for most secondary outcome measures; however, we found no significant differences between the acupuncture and the minimal acupuncture group (table 3). Differences compared with waiting list became apparent in the headache diary after the first four weeks of treatment and increased until week 12 (fig 2).

The improvements seen in the acupuncture and minimal acupuncture group persisted during the follow-up period (table 4). The patients in the waiting list group who received acupuncture in weeks 13-20 also showed significant improvements after treatment, although not to the same extent as the patients who had received immediate treatment.

Safety

Within the 24 weeks after randomisation a total of three serious adverse events (two acupuncture, one waiting list) were documented. All cases were hospital stays considered unrelated to the study (two diagnostic interventions, one elective surgery). Twenty three patients in the acupuncture group reported a total of 30 side effects compared with 11 patients in the minimal acupuncture group reporting a total of 14 side effects. The most commonly reported side effects were triggering of headache or other pain, haematoma, and dizziness.

Principal finding

The acupuncture intervention tested in this study was significantly more effective than no preventive treatment but not significantly more effective than the minimal acupuncture intervention in patients with tension-type headache.

Strength and weaknesses

Compared with available studies of acupuncture for tension-type headache,12^-17 which included up to a maximum of 69 patients, our study has a much larger sample size. Other advantages include adherence to current guidelines for headache trials,18 strictly concealed central randomisation, an assessment of the credibility of interventions, blinded evaluation of diaries, interventions based on expert consensus provided by qualified and experienced medical acupuncturists, high follow-up rates, and an external audit of the quality of data.

Although the groups were comparable for sociodemographic characteristics and headache outcomes at baseline, differences existed for some scores on the pain questionnaire in spite of randomisation. The credibility of acupuncture and minimal acupuncture was rated very similarly by patients, but guesses about treatment allocation at the end of the trial differed slightly between the acupuncture and minimal acupuncture groups, which might indicate some degree of unblinding. Trial physicians could not be blinded. Therefore, the small non-significant differences between acupuncture and minimal acupuncture could be due to bias. It was not possible to blind waiting list patients, so we cannot rule out that the difference from acupuncture and minimal acupuncture is overestimated. However, several arguments exist as to why the influence of bias should be limited. A slight improvement over time occurred in the waiting list group in the first 12 weeks; this was probably due to the natural course of the disease. This improvement, however, makes it unlikely that patients in the waiting list group reported negatively biased data in their diaries. Use of analgesics was lower in both the acupuncture and minimal acupuncture groups than in the waiting list group, making an influence of effective co-interventions unlikely. Follow-up data confirmed the improvements observed after treatment. After completion of the treatment, patients had no further contact with acupuncturists; they received and sent diaries and questionnaires directly to the study centre, decreasing the likelihood of positively biased diary data.

Interpretation of findings

The lack of significant differences between acupuncture and minimal acupuncture in our study indicates that point location and other aspects considered relevant for traditional Chinese acupuncture did not make a major difference. Although our treatment strategy was consensus based, we cannot rule out the possibility that a different approach would have yielded a different result. Our findings are similar to those of three of the available trials,12 13 16 whereas two found significant effects of acupuncture over sham acupuncture.15 17 A recent large, pragmatic trial from the United Kingdom found that adding acupuncture to general practitioners' care for headache was more effective than general practitioners' care alone.19 This trial mainly included patients with migraine, however, and subgroup analyses suggested that patients with tension-type headache might benefit less.

An intriguing finding of our trial is the strong and lasting response to minimal acupuncture. The improvement over, and the differences compared with, the waiting list group are clearly clinically relevant. The minimal acupuncture intervention in our study was designed to minimise potential physiological effects by needling superficially at points distant from classical sites as well as by using fewer needles than in the acupuncture group. However, it cannot be considered completely inert. The physiological effects of superficial needling distant from classical acupuncture sites may include local alteration in circulation as well as a wide range of neurophysiological and neurochemical responses such as release of neurotransmitters or activation of segmental and heterosegmental antinociceptive systems.20

Another explanation for the improvements we observed could be that acupuncture and minimal acupuncture are associated with particularly potent placebo effects. Some evidence shows that complex medical interventions or medical devices have higher placebo effects than placebo drugs.21 Acupuncture treatment has characteristics that are considered relevant in the context of placebo effects.22 It has an “exotic” conceptual framework with emphasis on the “individual as a whole,” it is associated with frequent patient-practitioner contacts, and it includes the repeated “ritual” of needling. Finally, the high expectations of participants and our way of informing patients might have been a relevant factor.