The JUPITER Trial

5
ByCrossFitMarch 17, 2019

This highly influential 2008 trial considered the impact of rosuvastatin, commonly known as Crestor, on heart disease risk.

The JUPITER trial differed from most other statin trials in two key areas. First, it was a primary prevention trial. Prior to JUPITER, statins were indicated only for hyperlipidemics and some others at high risk of heart disease. JUPITER deliberately chose a group of subjects who had, by traditional metrics, much lower risk of heart disease: men with LDL cholesterol below 130 mg/dL. The motivation was to determine if Crestor could be indicated for men who were not at high risk of a cardiovascular event. The second way in which JUPITER differed from other statin trials was that it sought to establish a new risk marker in the place of LDL—i.e., C-reactive protein—as the basis for prescribing the drug.

The trial design was relatively simple: 17,802 men and women between the ages of 50 and 60 with an LDL-C less than 130 mg/dL and high C-reactive protein (> 2.0 mg/L) were randomized to receive 20 mg of rosuvastatin (Crestor). Recruiting these subjects was a massive endeavor, drawing from 1,315 sites across 26 countries. Subjects were expected to be followed for five years. The primary outcome was an unusual compound measure: “first major cardiovascular event, defined as nonfatal myocardial infarction, nonfatal stroke, hospitalization for unstable angina, an arterial revascularization procedure, or confirmed death from cardiovascular causes.” The trial was scheduled to end once 520 of these cardiovascular events were observed, based on a preliminary power analysis. Randomization was excellent with few differences between treatment and placebo groups at baseline.

The trial was terminated early at a median follow-up of 1.9 years after 142 “first major cardiovascular events” had been observed in the treatment group and 251 in the control (393 total). The rates of primary endpoint per 100 person-years were 0.77 and 1.36, respectively, providing an HR of 0.56 for the treatment group with regard to the primary endpoint. Within the (approximately) two-year duration of the study, the number that needed to be treated to prevent one major cardiovascular event (NNT) was 95. All results were statistically significant.

The treatment group showed similar improvements in individual cardiovascular event measures including heart attacks (HR = 0.46) and strokes (0.52) (both significant). Additionally, there were 1.00 and 1.25 deaths per 100 person-years in the treatment and placebo groups, respectively, providing a significant HR = 0.80 for all-cause mortality for rosuvastatin. A subgroup analysis showed rosuvastatin led to statistically significant improvements in the primary endpoint in all groups, such as those divided according to sex, age, race, and other characteristics.

Cumulative Incidence of Cardiovascular Events According to Study Group. Figure 1 from Rosuvastatin to Prevent Vascular Events in Men and Women with Elevated C-Reactive Protein. 

The rosuvastatin group showed a dramatic decrease in LDL cholesterol (to 55 mg/dL, compared to 110 mg/dL in placebo) and C-reactive protein (2.2 mg/L vs. 3.5 mg/L in placebo). The one significant adverse effect was a 25-percent increase in newly diagnosed diabetes in the treatment group (270 cases vs. 216 in controls).

Despite these impressive results, some elements of the trial were unusual, if not questionable. The first and most visible issue was the decision to stop the trial early based on the relative effectiveness of rosuvastatin. While this decision was made by an independent review board, this board also showed some conflicts of interest. This is particularly concerning in light of previous research suggesting trials that are stopped early tend to show artificially high effect sizes(1). Equally important, the short observation period limits our ability to understand the long-term impact of rosuvastatin. This is critical, given that JUPITER was a primary prevention trial in low-risk individuals and thus designed specifically to justify providing the drug to large populations for long periods even when the risk of an immediate cardiovascular event is low.

Additionally, there were major conflicts of interest. As expected, the trial was entirely funded by AstraZeneca, and many of the individual authors had pharma-related conflicts of interest. Additionally, the PI, Paul Ridker, was a co-inventor of a patent related to testing C-reactive protein.

Collectively, these and other errors have called into question the results of the JUPITER trial.


References

  1. Gordon H Guyatt professor, Matthias Briel assistant professor, Paul Glasziou professor, Dirk Bassler professor, Victor M Montori professor. Problems of stopping trials early. BMJ  344(2012): 1-3.

Comments on The JUPITER Trial

5 Comments

Comment thread URL copied!
Back to 190318
Michael Eades
March 19th, 2019 at 5:02 am
Commented on: The JUPITER Trial

In my view, the JUPITER study is problematic from the very beginning, starting with its name. When scientists plan a study, they start with an hypothesis. They then do their best to scrub all bias from from their research methodology. In the case of the JUPITER study, the bias was right there front and center starting with the name, which is an acronym for Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin. In my opinion, that sort of tells the tale. The authors of the study, all statinators to their cores, want to justify the use of their favorite drugs.


Aside from the title, there are multiple other issues at play with this study. One of the problems with statin studies in general is that although these drugs do seem to minimally reduce the rates of death from heart disease, they haven't been shown to reduce the all-important all-cause mortality. In other words, they haven't reduced deaths overall. Subjects taking statins die less frequently from heart attacks, but more frequently from other diseases. So, in essence, taking a statin does not prolong life, but merely changes the agent of death. The results of the JUPITER study did reduce all-cause mortality in those who took Crestor, which is why the community of statinators was over the moon with the results. Finally, proof that a statin could prolong life.


Much though the statinating community would love to have us all taking statins daily, the JUPITER results are pretty limited. They apply only to men over 50 or women over 60 who have normal LDL-colesterol levels (>130 mg/dl) AND elevated c-reactive protein. These were the study subjects, and results from these subjects can't rationally be extrapolated to the population as a whole. In fact, these parameters are unusual to say the least. There were 17,802 subjects in the study gathered from 1315 sites spread across 26 countries. Some quick math shows that each study site had an average of about 13 subjects. These subjects weren't your average subject, and the results can't be assumed to hold for everyone everywhere.


When you look at the data from the study, you see that there were about 50 more deaths in the placebo arm of the study than from the arm in which Crestor was given. A small, yet significant, difference. But when you look at how the randomization shook out, you find some intriguing differences. Although the researchers were careful to randomly distributed smokers between the two groups, they weren't so careful with other risk factors. There were 51 more subjects with a family history of premature cardiovascular disease (probably the strongest risk factor for developing heart disease) in the placebo group. And there were 71 more people with Metabolic Syndrome in the placebo group. Just like the difference in 50 deaths between the two groups, these number are small, but it doesn't take but a few differences in deaths to whittle that 50 subject difference down to something that is no longer statistically significant. Having 51 more people with bad family history and 71 more with Metabolic Syndrome (another major risk factor for heart disease) might be enough to radially alter the outcome of this study statistically speaking.


A strange finding in this study was that those taking the Crestor - a potent statin - had no more reported side effects than those taking a placebo. Very strange indeed. Muscle aches are one of the most common side effects experienced by those taking statins. In the JUPITER study, there were only 10 subjects out of around 9,000 taking Crestor who reported muscle pain as compared to 9 subjects in the control group. This is extremely difficult for me to reconcile with the reality of the statistics on statin users and muscle pain.


One of the downsides of the study from the Crestor perspective was that subjects taking Crestor developed diabetes at significantly higher rates than those taking placebo.


The most bizarre part of this study was that the authors terminated it right in the middle. The reason given was that to continue the study would have been unethical as so many people in the placebo group were dying. Sometimes trials are stopped prematurely when it is discovered that subjects on the study medication are dying at unacceptable rates, but this is the only one I can think of that was stopped because subjects on the placebo had a minimally greater rate of death.


Since the subjects on Crestor were developing diabetes at greater rates than those on placebo, and a point had been reached at which the small differences in death rates had reached statistical significance, it does make me wonder if either of those issues drove the termination. Since the JUPITER study was published, we've learned a lot about Big Pharma fiddling with studies in one way or another. The strange way in which this study was handled certainly gives me pause.

Comment URL copied!
Matthieu Dubreucq
December 8th, 2019 at 12:37 pm

I was wondering also about the increase in diabetes and the potential link with stopping the study before it showed a link between diabetes and statins.

Comment URL copied!
Mary Dan Eades
March 19th, 2019 at 4:09 am
Commented on: The JUPITER Trial


Prior to the 'vaunted' Jupiter trial, Crestor had not been shown to do diddly in preventing MIs or cardiac deaths. And that was a problem for them, but they dutifully spun their way around that sticky wicket:


Language embedded in their own ad in AARP magazine in September and October 2005, admitted: 'Crestor is prescribed along with diet for lowering high cholesterol and has not been determined to prevent heart disease, heart attacks, or strokes.'


Of course, those same hard endpoints were the specific outcome measures laid out at the start of 2008 Jupiter trial. The name of the trial gives you a clear clue to the underlying bias of the investigators (and possibly even some on the IRB) from the get go. Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin. When you start a study by naming it 'Justification for (fill in the blank)' that sure shows your open-minded posture.


So off they went on their Justification Quest, but it stopped short. Why?

You stop if you're trying to 'justify' selling a drug whatever the outcome. Especially when, coincidentally, there was a slight, slight, ever so slightly increased benefit in the treatment group at that point, an effect that might well have vanished were the study to have gone to its conclusion. To be fair the benefit might have gotten stronger or stayed the same, too. We don't know because what began as a 5-year study was truncated after 2 years 'because the drug's overwhelming efficacy had been proven'. Well that and the inescapable specter of significantly higher rates of developing diabetes in those who took it compared to those on placebo. And who knows what that tendency would have looked like carried out to the full original 5-year length?


To be sure, a drug's showing minimal actual efficacy (albeit in a narrowly defined group of men over 50 and women over 60 who had normal LDL cholesterol AND elevated C-reactive protein but no identifiable systemic inflammatory disorders) while possibly causing diabetes in the treatment group is an excellent reason to interrupt a trial less than half way to the end.


Bringing out the drug as safe and beneficial to reduce the risk of heart attack, stroke, or death for all people of all ages anywhere in the world anyway is just flatly wrong. How many doctors have prescribed it and how many 'healthy' patients have taken it secure in their minds that it would safely reduce the risk of a heart attack? Millions, I would imagine.

Comment URL copied!
Nathan Jenkins
March 18th, 2019 at 5:29 pm
Commented on: The JUPITER Trial

A 2010 critical reappraisal of JUPITER noted several of these significant concerns.


https://www.ncbi.nlm.nih.gov/pubmed/20585068


From the abstract,


"The trial was flawed. It was discontinued (according to prespecified rules) after fewer than 2 years of follow-up, with no differences between the 2 groups on the most objective criteria. Clinical data showed a major discrepancy between significant reduction of nonfatal stroke and myocardial infarction but no effect on mortality from stroke and myocardial infarction. Cardiovascular mortality was surprisingly low compared with total mortality-between 5% and 18%-whereas the expected rate would have been close to 40%. Finally, there was a very low case-fatality rate of myocardial infarction, far from the expected number of close to 50%. The possibility that bias entered the trial is particularly concerning because of the strong commercial interest in the study.


CONCLUSION:

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."


What is interesting is that the prevalence of statin use has nevertheless only continued to increase, with approximately 1/3 of the U.S. adult population currently on a statin drug :/

Comment URL copied!
Katina Thornton
March 18th, 2019 at 2:12 pm
Commented on: The JUPITER Trial

The open poll of 2553 respondents reveals that this study does not represent clear and convincing evidence. And, once again, let's not ignore that the addition of nutritional changes and exercise in low risk patients should preceed the addition of an exogenous medication that de facto includes unwanted side effects.

Comment URL copied!