STATIN: Web-based Investigation of Side Effects

1. INTRODUCTION Statins reduce cardiovascular disease (CVD) risk and are recommended as part of the treatment strategy for primary and secondary prevention of CVD. Although statins are the most commonly prescribed treatment in the UK, there is uncertainty about adverse effects. Severe statin adverse effects are rare but there is widespread reporting of less well-defined statin-related symptoms in the media, notably muscle pain and weakness that significantly affect statin users. These reports have been prompted by non-randomised, non-blinded observational studies but have not been confirmed in blinded randomised controlled trials (RCTs). A major limitation of observational studies is lack of blinding: patients taking a medication expect to experience adverse effects and therefore report high levels of symptoms vs. statin-free population. This phenomenon, the "nocebo" effect, leads to bias. Many patients believe their muscle symptoms are statin related, leading to therapeutic discontinuation. GPs face challenging decision making when patients present statin related symptoms and there is no diagnostic tool to evaluate statin symptom burden. StatinWISE is a N-of-1 trial and offers patients individual study results. Patients are their own control, and therefore optimal treatment can be established. StatinWISE will address some of the criticisms of previous evidence.
2. TRIAL DESIGN i) Randomised, double blind, placebo controlled N-of-1 trial ii) Patients who have stopped or are considering discontinuation of their statin due to muscle symptoms iii) Once-daily oral administration of Atorvastatin (20mg) or placebo iv) Study treatment is 12-months v) IMP in 2-month treatment periods vi) Quantify the occurrence of self-reported muscle symptoms vii) 200 patients will be recruited. 2.1 RECRUITMENT OF PARTICIPANTS Participants will be recruited directly from GP Practices or by advertising to the public. Participating practices will recruit eligible patients from two groups as follows:
3. Patients who are considering discontinuation of their statin due to muscle symptoms: These patients will be invited to take part in the trial when they visit the GP to report muscle symptoms believed to be associated with statins and where the patient/GP is considering stopping statins because of the muscle symptoms. The GP or Research Nurse will approach the patient and give the patient information sheet. If interested, patients will be able to consent and complete the screening visit with the GP or the Research Nurse during this appointment or it can be arranged for another suitable time. ii) Patients who have stopped taking a statin in the last 3 years due to muscle symptoms: A search of the practice electronic records will be performed by the Research Nurse on a two-monthly basis for one year (or until recruitment targets are reached) to identify potentially eligible patients. All screened patients will be documented on a screening log. The list will be reviewed by the GP to confirm clinical eligibility before patients are invited to take part. A letter inviting them to attend a screening visit, accompanied with the patient information sheet for the patient to consider, will be sent by the trial team from their GP practice. Contact details of the Research Nurse will be provided should the patient have any questions. A reply slip will be enclosed for the patient to complete if they wish to attend the screening visit, which will be returned to the Clinical Trials Unit (CTU), during which the trial will be explained, and they will have the opportunity to ask questions. Patients will be sent a letter of invitation to consider participation up to a maximum of three times. iii) Patients who contact the CTU from advertising: Patients who contact the CTU in response to advertising material will be sent a letter to request their GP details on a reply slip. Following receipt of these documents the CTU will contact their GP with their consent. The GP will be asked to confirm that the patient is potentially suitable for the trial and to provide brief clinical information to allow eligibility to be assessed. This information will then be provided to the GP surgery responsible for recruiting the patient. 2.2 DRUG MANUFACTURE, BLINDING AND SUPPLY OF TRIAL MEDICATION Atorvastatin will be purchased on the open market. Placebo will be manufactured specially to match the IMP. Capsules and packaging will be identical in appearance for both IMP and placebo. DBcaps® capsules have a unique locking mechanism to help with assuring the integrity of the blind will be used for over encapsulation of both active and placebo treatments. The blinding process will involve encapsulating the active tablet, complete removal of the original manufacturer's label and replacement with the clinical trial label bearing the randomisation number which will be used as the pack identification. Outer pack labelling will be identical for IMP and placebo and will be in compliance with regulations. 2.3 DATA COLLECTION Baseline data will be collected at each GP practice and will be entered directly online to the trial database provided by the LSHTM CTU. Follow up data will be collected directly from each patient at the end of each two-month period. Patients will choose their most suitable method of data collection:
4. Bespoke mobile app which will require patients to use their own smartphone ii) On-line database using a computer, phone or tablet iii) Paper forms which they will receive by post at the same time with their trial treatment and which they can complete iv) Contact by phone. Trial staff will telephone the patient on each data collection day and complete the questionnaire based on the patient answers. Only data outlined on the baseline, follow up, end of trial and adverse events data forms will be collected as part of this trial database. END OF TRIAL DATA Patients will receive their individual results at the beginning of month 14 and have a telephone or face-to-face appointment to discuss these results. At month 15, trial staff will contact the patient to document their decision on statin use and whether their results helped reach this decision. This will be the last data collection point of the trial.
5. OUTCOME MEASURES Primary outcome: Self-reported 'muscle symptoms', defined as pain, weakness, tenderness, stiffness or cramp to the body of any intensity. The primary outcome will be assessed by the mean difference in VAS scores (range 0 to 100) between statin and placebo treatment periods, estimated via a linear mixed model. Secondary outcomes:

• Participant belief about the cause of their muscle symptoms, the site of muscle symptoms, how the muscle symptoms affected the participant and information about any other symptoms.
• Adherence to medication
• Participant's decision about statin treatment following the trial
• Whether they found their own trial result helpful.

1. ANALYSIS

Individual N-of-1 trials:

The purpose of these is to inform individual patients of the effect of the IMP on their muscle symptom score. The analysis and presentation of individual level results will be developed in collaboration with Patient and Public Involvement (PPI) groups and will include a range of graphical summaries and statistical analyses to identify the most informative presentation of individual results.

Combined analysis of N-of-1 trials

Primary analysis:

To estimate the population level estimate of the trial treatment in VAS muscle symptom score, data from each N-of-1 trial will be aggregated to form a powerful dataset, using an intention-to-treat approach.

Patients who enter data on muscle symptoms at least once during a treatment period with the IMP and at least once during a treatment period with placebo will be included in the primary analysis.

The primary analysis will be a linear mixed model for VAS muscle symptom score with random effects for participant and treatment. Residual errors will be modelled using a first-order auto-regressive error structure within each treatment period to account for correlation between the 7 daily measurements, with robust standard errors to account for non-normality of the VAS scores. Although VAS muscle symptom scores are unlikely to be exactly normally distributed, analysing such data using normal-based methods is likely to be a sufficiently robust approach.

All tests will be two-sided. P<0.05 will be considered statistically significant.

Secondary analyses:

Secondary outcomes will be analysed in a similar manner to the primary outcome, omitting the auto-regressive correlation structure since these secondary outcomes are measured once per treatment period.

Descriptive statistics will be used to summarise adherence to randomised treatment, and their relationship to the IMP and placebo periods.

The adherence to randomised treatment will underpin an efficacy analysis based around an instrumental variables approach. Because these analyses require much stronger assumptions than the intention-to-treat analysis above, the results of the efficacy analysis will be presented and interpreted as a secondary analysis.

The secondary outcomes include a single binary measure of whether the participant reports having muscle symptoms during that treatment period. This will be combined with the follow-up question pertaining to attribution, to obtain a single binary measure of whether the participant reports having muscle symptoms that they attribute to the study medication. These two binary outcome measures will be assessed using a logistic mixed model with random participant and treatment effects.

The investigators will relate the patients' decision regarding future statin use, and whether or not the participant found their own result helpful in making their subsequent treatment decisions, to their individual estimated effect of the IMP.

Subgroup analyses:

There are no priori subgroup analyses planned.