Giving Statins or Aspirin for Primary Prevention: Would You?

Robert M. Centor, MD; Bradley P. Fox, MD; Désirée Lie, MD, MSE; R. Brian Haynes, MD, PhD; Robert W. Morrow, MD; Roy M. Poses, MD; Charles P. Vega, MD; Mark E. Williams, MD


June 28, 2011

The Roundtable Question

Since the JUPITER study was published indicating survival benefits with statins in patients with high C-reactive protein (CRP) levels but no other cardiovascular risk factors, there has been some talk about whether statins should be prescribed as primary prevention.[1] One cardiologist said that he thought statins should be put in the drinking water. Would you ever prescribe a statin for primary prevention in adult patients without cardiovascular risk factors? What about aspirin, which in fact is more likely to be used by individuals for primary prevention without their physician's knowledge?

Bradley P. Fox, MD (Family Medicine)

There are more and more data to suggest that the only non-risk factor is being a young female. Being male and being overweight can be considered risk factors for cardiac disease. The risk benefit ratio with a statin is much better than with aspirin, and pravastatin is generic, cheap, and has limited side effects. In my mind, it would be great if they could create a pill with a low dose of metformin and pravastatin together without the side effects that either of them have. I would prescribe that.

Mark E. Williams, MD (Internal Medicine)

For me the answer to prescribing statins for primary prevention is "No." The evidence is clear that the magnitude of absolute risk reduction is marginal (less than 2%). Moreover, the potential harms dilute the impact of any benefits.

Robert M. Centor, MD (Internal Medicine)

I am a skeptic. I think statins are great for secondary prevention -- that includes strong family history and type 2 diabetes -- but I am not aware of any evidence that statins are worthwhile for pure primary prevention.

Charles P. Vega, MD (Family Medicine)

Statins have been an incredible breakthrough and have helped save many lives among patients with coronary artery disease (CAD) or CAD-equivalent conditions, but I can't agree with the wider implementation of their use among younger and healthier patients. How long would treatment last? 30 years? 50? That is a tremendous commitment, both for the individual patient and for the healthcare system in general for a treatment approach with a weak background of evidence in reducing the risk for mortality.

The high-sensitivity CRP (hs-CRP) is something of an orphan test for me. It can be helpful to push the decision to treat more aggressively among patients with borderline risk factors, but that is a small percentage of patients.

Bradley P. Fox, MD

I always test in patients who are seemingly controlled but are still at risk or who have events despite being treated to goal. I do not use CRP as a screening number, but more as a fine tune result.

Robert W. Morrow, MD (Family Medicine)

My general sense is that very few physicians use the Framingham framework to decide risk, and virtually none are aware of the recent meta-analyses that show no mortality benefit. Maybe a contrarian expert could give a short intro to a roundtable on this, emphasizing the number needed to treat to prevent 1 death.

Roy M. Poses, MD (Internal Medicine)

The 2010 meta-analysis in the Archives of Internal Medicine[2] found a small reduction in all-cause mortality in the statin group: 7 per 10,000 person-years of observation (number needed to treat to prevent 1 death in 1 year, 1429) but the small risk reduction was not statistically significant, although it was close. The bottom line is: if you assume the reduction was real, you need to treat 1429 people to prevent 1 death in 1 year, so by that measure 1428 of 1429 people may not benefit. The accompanying editorial is also worth looking at.[3]

Please also note the large observational study on the adverse effects of statins, which appeared only online in 2010 in the British Medical Journal.[4] To cut to the chase, the authors calculated the number needed to harm for several adverse effects of statins in a medium-risk population (QRISK2 score > 15%). The results were:

  • Women: acute renal failure, 593; cataract, 40; liver dysfunction, 154; myopathy, 313.

  • Men: acute renal failure, 447; cataract 66; liver dysfunction, 155; myopathy, 106.

Admittedly, this was an observational study and its population did not parallel that of the meta-analysis cited above. Nonetheless, if you believe that statins prevented early death given that the difference in death rates was not significant, the number need to treat to prevent 1 death in 1 year was 1429. The numbers needed to harm for some fairly significant adverse effects were an order of magnitude less. So, assuming statins do prevent death in primary prevention, for each death prevented in 1 year, statins might cause approximately 2-3 cases of acute renal failure, 9 cases of liver dysfunction, and 14 cases of myopathy. One could potentially harm quite a few people for each death prevented. This makes it even less obvious that the benefits of statins in primary prevention outweigh their harms.

Bradley P. Fox, MD

But that is in people with no risk factors. How many people truly have no risk factors?

Roy M. Poses, MD

I don't believe the meta-analysis was restricted to trials of patients with no risk factors. Their relevant inclusion criterion was that trials of patients "without prevalent cardiovascular disease at baseline" were included. Some trials did exclude patients with diabetes, but some did not. I know the West of Scotland Coronary Prevention Study (WOSCOPS) likely included many people with risk factors.[5]

R. Brian Haynes, MD, PhD (Internal Medicine)

Recent evidence indicates no benefit from statins with no cardiovascular risk factors,[6,7,8] and some harm (that would apply to people for primary prevention with no risk factors as much as to anyone else). I can't imagine prescribing statins for someone with no risk factors -- I think it is a non-question/malpractice. The real question is at what risk level is the treatment benefit likely to outweigh the harm (including both adverse effects and cost). The Cochrane review[8] begins to address that for people with some risk factors.

Désirée Lie, MD, MSE (Family Medicine)

Here is my 2 cents. There is a big difference between aspirin and a statin. One is available over-the-counter, and patients may already have made decisions about using aspirin and are not likely to be consulting us about it, in which case we need to enquire about use and keep patients informed about relative risk vs benefits. Statins need our prescription, although the drug makers may prefer otherwise.

A discussion of aspirin can be used as a platform for talking about preventive lifestyle strategies in general and [can be] used to negotiate for exercise and weight loss (vs taking a pill with potential adverse effects) in a much broader context. The statin discussion will largely revolve around a more limited aspect of cholesterol and cardiac event prevention.

My intuition is that patients are more open to taking aspirin lifelong (or episodically depending on relative risk) because they make decisions and can get it themselves compared with statins, which are physician-centered, so that every time they want it, they have to come in to see us and pay more. In other words, if both are of similar benefit, patients would choose aspirin over statins.

Bradley P. Fox, MD

Once again, I pose the question, what is really "no risk factors"? With all the evidence for and against aspirin, I would like to know how people recommend aspirin for their patients and if they make the recommendation based on true risk/benefit ratio or not --across all ages and risks -- or if there is some other process that is used or if there is a process at all.

R. Brian Haynes, MD, PhD

The issues are basically the same: for primary prevention, at what degree of risk/benefit should one prescribe any treatment? The evidence concerning risk/benefit is about as clear for either statins or aspirin, so if you are tackling the issue, it doesn't matter which.

Charles P. Vega, MD

That's generally correct, but my impression is that primary care docs are slightly more aware of the benefit/risk balance with aspirin and are more likely to be using statins outside of the general indications because they are newer and marketed more aggressively. The approach to statin therapy reminds me of how hormone replacement therapy (HRT) became a potential wonder drug in the 90s: "Hey, they're helpful in the prevention of CVD and they might also be protective against cancer and they might reduce the risk for dementia." Findings from limited research can have a powerful influence on prescribing habits.

Bradley P. Fox, MD

I actually resent that. Marketing has nothing to do with my prescribing and I can show that in many ways. I was embracing evidence-based medicine long before it came into vogue, and I have not used a TV commercial, drug rep, or patient request to direct my prescribing.

I actually look at HRT as a benefit in a select population of people today and have not scrapped it totally. If patients are not lumped together as "diabetics" or "hypertensives" or "menopausal women," treatment can be personalized using individual studies rather than the meta-analyses that are out there. HRT is the wonder drug for some women in the first 5 years after menopause, especially those with surgically induced menopause, and while it may not cure the common cold, it is a prescription that I am not afraid to write despite the Wall Street Journal and the Women's Health Initiative.

This is why I raise the question of who is truly without risk. Is the person who has 5 relatives with an MI [myocardial infarction] before 60 but who has no issues with hypertension or lipids or diabetes or smoking truly a [candidate for] primary prevention? Is the person with a BMI [body mass index] of 27 who is not active but who has none of the above truly without risk? Where do we individualize and where do we say that the data show no help?? I am not using statins for primary prevention, and I am not willy nilly using aspirin. I have not decided that everyone with a BMI over 25 should be on low-dose metformin, but I am looking for information that might push me to consider these things.

I am reading and talking to the thought leaders and I am listening to my patients. I do talk with drug reps. I do talk with their direct managers. I would like to understand their marketing strategy from a purely intellectual standpoint, but I have not written nebivolol as a first, second, third, fourth, fifth, or sixth choice. I have gone there when all else has failed.

I do not speak for all primary care physicians, but since the response came as a response to my comments, I felt obligated to take it a little personally and to defend myself.

Charles P. Vega, MD

Brad, I certainly meant no disrespect, and my comment was not in any way aimed at you. It sounds like you're doing a great job as a patient advocate, and I agree that most guidelines do not and cannot account for every case we see. But I still believe that collectively we (to emphasize "collectively," "we") are too quick to turn to the next wonder drug without realizing the potential pitfalls in the long run.

Robert W. Morrow, MD

Medicine is swept by fashion as much as anything, and it is our job as hired curmudgeons to inspect the fabric of the king's garments. My patients want tons of stuff -- including attention and love -- and I need to separate out the drug of the day from its harms and benefits. I could go on -- and boy, have I -- about cholinergics, anticholinergics, long-acting beta-adrenoceptor agonists, and a pharmacy of others. Let me give a quick breakdown of pharmaceutical excess.

  • Everyone with cognitive decline should be on a cholinergic agent to prevent worsening of dementia, despite virtually no known clinical benefit. To compound the nuttiness of making demented folks have thick pulmonary secretions, nausea, urinary frequency, and loose stools (common side effects of cholinergics), we then use anticholinergics for irritable bladder, irritable bowel, and chronic obstructive lung disease.

  • The rage for pushing LDL [low-density lipoprotein] lower than 70 over decades for primary prevention is still the rage, despite minimal benefit, maximal cost and unclear risk.

  • The urgency to medicate with insulins to lower A1C below 6.5 killed several folks with diabetes. Type 2 diabetes is a social disease; let's not kill its victims.

  • Long-acting beta agonists (LABAs) have black box warnings and at least a 30-year history of increased mortality (for those old enough to remember their exclusion from the United States in the 1980s), and yet I frequently see people, including children, who receive an LABA before a trial of inhaled corticosteroids alone, even with moderate persistent asthma. Advertising, anyone?

  • Shall we discuss the wholesale use of HRT, the race for more and more expensive proton pump inhibitors, or shall we look at diabetes and the introduction of classes of drugs with no demonstrable endpoint outcome improvement or any clear study of long-term risk?

We are all citizens of the media fashion show, but we should first do no harm.