Publication committees in clinical trials in headache

The question of control of publication of randomized controlled trials (RCTs) remains unresolved and this leads to unpublished trials and undue pharmaceutical influence on those results that are published. A lack of rules by the Food and Drug Administration (FDA) and by European Medicines Agency (EMA) leads to widely ranging practices from field to field. In the headache/migraine field, sponsor companies have almost complete control and decide on the persons who should publish and sometimes on whether to publish or not (1).

In a recent review of naratriptan, only seven RCTs out of 17 were published in full (2). The clinical investigators were most likely not involved in the decision not to publish. These RCTs were multicentre and in most cases multinational. Thus, the median number of centres participating in the naratripan RCTs was 51 (range: 10–152). The median number of patients in each RCT was 300 (range: 90–1141). The median number of patients per centre was nine (range: 1.5–16). In a more recent multinational triptan RCT, 165 clinical investigators recruited 2113 migraine patients (3). In a similar large RCT in migraine from 2007, 118 clinical centres in the USA recruited 2956 migraine patients (4). Very large multicentre, multinational (multiregional) RCTs are thus still performed with the corporate sponsor in control. We cannot know how many migraine RCTs are not published until the rest of the industry follows the excellent example of GlaxoSmithKline (www.gsk-clinicalstudyregister.com) (5) and establishes a clinical study register. Such registering should be made compulsory by regulatory authorities. It is of course the obligation of the clinical investigators to ensure publication (6), but this duty is not easy to execute as illustrated above.

In large multicentre trials there is usually little or no contact between clinical investigators. We therefore propose to increase the influence of clinical investigators by making the election of a publication committee compulsory as soon as the protocal is written. It should be stated in the paragraph on publication that a publication committee (PC) will be elected by participating investigators before the start of the study. The corporate sponsor should organize the election meeting or an election meeting by email but should not interfere with this election process.

The PC members should commit themselves to produce a first draft of the paper within a short time after receiving the clinical report. The corporate sponsor should have the opportunity to comment on this draft and the final version, each time doing so within 4weeks.