What can we learn from new n-of-1 trials
Patient: “Doctor, I’ve been getting a lot of cramp in my legs and the only thing that has changed is I’m taking statins. My friend had the same problem when he started taking statins and when I googled it, the NHS website said statins can cause muscle pain.”
GP: “So you think it’s the statins causing your symptoms?”
Patient: “Yes and I want to stop taking them.”
Undoubtedly, general practitioners across the country will be familiar with this kind of conversation. Stopping statins because of perceived muscle side effects is common and has an impact on subsequent cardiovascular disease risk. But, after serious rises in muscle enzymes have been ruled out, determining whether statins are really causing a patient’s muscle symptoms is not straightforward in routine practice.
One approach would be to ask our patient to stop and start statins, while monitoring their symptoms. The key issue is that our patient knows when they are (and are not) taking a statin. This is important because patients often experience subjective side effects that they expect to experience. A simple example of this phenomenon, termed the “nocebo effect,” would be the feeling that you get when somebody close to you tells you they have nits. You immediately start to feel a bit itchier. The same may be true for statins: people expect to experience muscle symptoms, perhaps because of media reports, and therefore any small ache or pain, which might have happened anyway, is considered by the patient to be attributable to their medication.
N-of-1 trials are an excellent solution for our patient and GP. In an n-of-1 trial, our patient would undergo treatment periods of statin and placebo, in a random order, but—crucially—would be blind to what they were taking. Muscle symptoms would be monitored throughout and, after several periods of statin and placebo, our GP and patient would look at the unblinded results to determine whether symptoms were worse during statin use.
In our recent series of n-of-1 trials published in The BMJ, we recruited 200 patients from primary care who, just like our patient above, were considering discontinuing or had discontinued statins because of muscle symptoms. StatinWISE (Statin Web-based Investigation of Side Effects) was pragmatically designed alongside our active patient and public involvement committee, minimising burden on participating general practices and patients. Drugs were sent by post, and muscle symptom outcome data were reported either online, by phone, or by post, according to participant preference.
However, in running StatinWISE (Statin Web-based Investigation of Side Effects) we encountered challenges that would need to be overcome before n-of-1 trials could be incorporated as a routine element of primary care. Firstly, although statins are approved for use by drug regulators for exactly the kind of patients in our participant group, StatinWISE required approval as a CTIMP: a Clinical Trial of an Investigational Medicinal Product. CTIMPs are subject to substantial legislation and scrutiny as to how they are approved, conducted, and monitored. Secondly, the statin had to be over-encapsulated to match the placebo and maintain the blinding, which is relatively complex, and can’t be done by just any pharmacy.
Most participants completing the trial found that their symptoms were not related to statin use, and so resumed statins following completion of the trial. Most also reported finding the trial useful and valued the individualised results. Importantly, the results from individual trials can be combined to produce an answer at the population level about statin and muscle symptoms. Overall, there was no difference in the frequency or severity of muscle symptoms between statin and placebo periods.
Given the benefits of statins in cardiovascular disease prevention, and the high frequency of muscle symptom complaints among patients, it is time to incorporate research into routine practice. Rolling out a n-of-1 service into routine clinical practice would inevitably have its challenges (especially around the provision of placebos and the requirement to monitor adverse effects), but it would be a great benefit to individual patients and to the NHS. The n-of-1 design would be useful beyond the context of statins and muscle symptoms, and could be used for many reversible short term drug effects—both beneficial and harmful—where uncertainty exists.
Emily Herrett, assistant professor, Department of Non-communicable Disease Epidemiology, London School of Hygiene and Tropical Medicine
Alexander Perkins, research fellow, Department of Medical Statistics, London School of Hygiene and Tropical Medicine.
Elizabeth Williamson, professor of biostatistics and health data science, Department of Medical Statistics, London School of Hygiene and Tropical Medicine.
Danielle Beaumont, senior trial manager, Clinical Trials Unit, Department of Population Health, London School of Hygiene and Tropical Medicine
Maurice Hoffman, patient member of the trial steering committee appointed by NIHR
Liam Smeeth, professor of clinical epidemiology, Department of Non-communicable Disease Epidemiology, London School of Hygiene and Tropical Medicine
Competing interests: please see full statement on linked research paper.