To the Editor:

We read with interest the article by Mullins et al¹ reporting recommendations for clinical trials of off-label drugs used to treat advanced-stage cancer. The 14 statements summarizing their recommendations are absolutely shareable and actually represent an excellent checklist for improving quality of clinical trials, no matter the field to which they are applied. However, we doubt that they can significantly affect the off-label phenomenon, driven by the awareness of safety (at least in the setting in which the treatment is approved) and the existence of true unmet medical needs.

In 2006, we conducted a survey in 15 Italian centers,² collecting data on treatment of 644 patients corresponding to 1,053 prescriptions. Overall, 199 prescriptions (18.9%) were off label because of one of four reasons: drug used in a cancer different than that authorized (n = 92; 46.2%), drug used within an inappropriate combination (n = 41; 20.6%), drug used in a line of treatment different than the registered one (n = 39; 19.6%), and drug used in a rare tumor (n = 27; 13.6%). All the prescriptions of the first group had supporting scientific evidence (assessed by physicians different from those who wrote the prescription) from one or more randomized trials or several positive phase II studies published in major oncology journals. Therefore, clinical trials in this case had been performed but had not produced registration. In the second group, the only drug used in combination but approved as a single agent (at that time) was capecitabine (20 prescriptions; 10%). The remaining were drugs approved for use in combination that were instead used as single agents, possibly because of toxicity and consideration of patients' conditions. From the perspective of Mullins et al,¹ the latter should probably be tested in clinical trials dedicated to unfit patients. The third group mostly (28 v 11 patient cases) included drugs that were used for lines of treatment subsequent to those for which they were registered, in common cancers such as breast and colorectal. It is hard to believe that proper randomized trials can be planned involving patients receiving their fourth or fifth line of treatment, primarily because of the difficulty of identifying a control arm that should probably be placebo or best supportive care. Finally, the fourth group included rare neoplasms, which represent a well-known field of debate because of the impossibility of applying common rules of clinical trial methodology.³

In our opinion, the Italian data underline two important concepts. First, when scientific evidence is available to justify the off-label use of anticancer drugs (half of off-label prescriptions in the Italian survey), the Mullins et al¹ recommendations can improve the quality of clinical trials, but it is even and especially more necessary that regulatory agencies define guidelines to allow the registration of new indications. In fact, until now, performing one or more clinical trials is not enough to start a registration process unless the story is managed by the pharmaceutical industry. When the latter actor is active, the off-label phenomenon may not exist, because registration arrives roughly together with scientific evidence. However, producing even convincing evidence within academic clinical trials of drugs not intended for registration can result in the unwanted adverse effect of instigating off-label use, and this requires management by regulatory agencies.

Second, the other half of the off-label prescriptions cannot be modified by the suggestions of Mullins et al.¹ In fact, in such cases, the off-label prescription is the consequence of an extremely fragmented therapeutic landscape, in which it is really hard to find out the assumptions that are needed to properly plan clinical trials. Observational studies evaluating the impact in clinical practice of these off-label prescriptions could probably help us understand this phenomenon, which is hard to regulate.

Therefore, we propose that the suggestions of Mullins et al¹ be integrated with other available experiences. For example, in Italy, the appropriate use of many cancer drugs is regulated by the so-called Cancer Drugs' Register (http://antineoplastici.agenziafarmaco.it), managed by the Italian Medicines Agency, where authorized treatments are registered and monitored; in some cases, this register has also been used to collect data on off-label drug use, under the coverage of a specific Italian law that allows reimbursement for off-label indications supported by phase II clinical trials (Law No. 648/1996). This more comprehensive and conservative approach would be not too expensive and would provide some important evidence and outcomes. Pharmaceutical industries should provide the drugs at no cost in exchange for the openness of agencies in discussing approval in selected patient populations, among whom the drugs would have a positive impact on clinical outcomes.

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest.