One of the commonly used statins is rosuvastatin, which has the brand name Crestor and is produced by AstraXeneca. It was evaluated in the trial known as JUPITER, which has been subject to considerable criticism (1). In a review of several key aspects, Michel de Lorgeril and colleagues concluded that the results of the trial do not support the use of statin treatment for primary prevention of cardiovascular diseases and raise troubling questions concerning the role of commercial sponsors (2). The trial tested the effects of rosuvastatin therapy (20 mg/d) in patients without cardiovascular history or established coronary heart disease (CHD).
This trial was especially controversial because it was stopped early. In fact the follow-up period was less than 2 years. Although this information is in the public demesne we do not actually know what period was originally specified as the details were not provided in the published description of the study protocol. Furthermore, we still do not know which end point was used to define them, or which level of benefit, was required to justify early termination. The early cessation was justified on the grounds that it had been approved by the Data and Safety Monitoring Board (DSMB). The Chairman of the DSBM was Sir Rory Collins of the Clinical Trial Service Unit (CTSU) at the University of Oxford. It is clearly established that the Chair and members of the DSMB should be free of any conflicts of interest. During the recent spat between Collins and the BMJ, details of the financial support for the CTSU were released. This document (3) shows that there has been a close working relationship between the Unit and AstraZeneca and the Unit. Projects involved include the following:
• £1.0 M plus drugs to ATLAS trial of Tamixofen which commenced in 1997
• £1.1 M plus drugs to CCS-2 trial of metapropol and clopidogrel (1999-2005)
• $300K to China Kadoorie Biobank which commenced in 2002
• £1.7 M to PROCARDIS genetic study (1998-2011)
• $100K to STICS trial of rosavustatin (2011-2014).
It will be evident from these contributions that there has been a close working relationship between the company and the University as far back as 1997. So although Collins was not directly involved in the JUPITER trial, he cannot be regarded as entirely disinterested in view of the major financial input made by AstraZeneca, which continues to the present.
In any case there are genuine concerns that stopping a trial early does bias the results. A thorough analysis of the results of 515 Randomized Controlled Trials (RCTs) of which 91 were truncated provides confirmation. The differences were huge. For example, if a completed RCT showed a relative risk reduction of 20%, the corresponding truncated RCTs would show one of 43%. The reality is that if there are beneficial effects, these are likely to be substantial during the early stages of a trial and then diminish. This is why it is crucial to stick to the original protocols.
The de Lorgeril critique commented that the primary end point is a composite of cardiovascular complications. Some of these, such as revascularization and hospital admissions, strictly speaking are not complications but medical decisions. Furthermore they are not objective criteria and therefore can be subject to manipulation. If the analysis is restricted to the hard end points of fatal and nonfatal heart attacks and strokes, there were only 240 events because of the early stop: another reason to question the robustness of the conclusions. From my own perspective the over-riding consideration is the impact on all-cause mortality. I cannot get enthusiastic about a treatment which may reduce my chances of dying from a heart condition but increase the risks of dying from another disease. Yet the investigators seem reluctance to focus on this aspect. Results which are limited to a single disease or condition may be biased because they are dependent on subjective judgements on the part of a physician. But there can be very little doubt if the critical measurement is death!
The graphs below from the original paper reporting on the JUPITER trial are very revealing. Those showing information on all end points except all-cause mortality diverge very nicely as the investigation proceeds (4). However in 1d, take a very careful look and it will become very clear that the values for the placebo and the treated groups are converging when the trial stopped. We can only surmise what might have happened to these lines had the study been allowed to continue as initially planned.
In the light of these results it is not surprising that great play is made about the results shown in Figures 1a, 1b and 1c. However those shown in 1d have not exactly been publicised, although I am sure this is the information most people would wish to know.
The results of this Jupiter trial play an important role in constructing the justification for the regulatory approval and the widespread usage of statins. Despite these genuine criticisms statins continue to be promoted not only for the treatment of heart disease but also as preventative measure. NICE has recently advised that those who have 10% risk of developing the disease should be considered for treatment with statins (5). Bearing mind that there can be some very nasty side-effects associated with statins, if this is typical of the evidence to demonstrate its effectiveness, I cannot be impressed. Certainly I will not be having them.
The case in support of statins continues to be eroded as more and more details emerge. The reality is that when subjected to a hard-nosed evaluation the benefits are relatively small with the strong suspicions that they have been somewhat exaggerated. Even Mark Baker of NICE had to accept that 77 people would have to be treated for 3 years for one to benefit, which is not very impressive and probably much less than most people have been led to believe (6).