Benefits of Statins Only Exceed Harms at Higher 10-Year Risk Thresholds, Study Suggests

Current 10-year risk thresholds used to determine when to start statin therapy for the primary prevention of cardiovascular disease are too low, a new modeling study suggests. 

When the full range of potential harms from statins are taken into account, including the risk of developing type 2 diabetes, cataracts, and myopathy, among others, the net benefit of treatment is observed in men with a 10-year risk of CVD between 14% and 21%, according to researchers. For women, treatment yields a net benefit only when used in those with a 10-year risk of 17% to 22%.

To TCTMD, senior investigator Milo Puhan, MD, PhD (University of Zurich, Switzerland), said there has been extensive debate about statins’ place in primary prevention, particularly the risk thresholds for initiating therapy. Explaining the rationale for their modeling study, which was published December 3, 2018, in the Annals of Internal Medicine, he said the full gamut of adverse events from statin therapy are not fully considered in the guidelines.

Moreover, “putting together the entirety of evidence, which is far greater than the randomized controlled trials, is not that easy,” said Puhan. “The guideline developers—and it doesn’t just apply to the cholesterol guidelines—don’t use explicit methods to fit all the pieces together. This is why we did it. We have a method to put it all together in a transparent way, and we wondered if we could figure out the threshold where the benefits exceed the harms of statins at the population level.”

Their newly identified risk thresholds far exceed the current American College of Cardiology/American Heart Association (ACC/AHA) 10-year risk threshold for initiating statins in individuals without existing CVD. For adults aged 40 to 75 years without diabetes and LDL cholesterol levels ≥ 70 mg/dL, the newest cholesterol guidelines recommend a moderate-intensity statin in patients with a 10-year atherosclerotic CVD risk of 7.5% or greater following a favorable physician-patient discussion of treatment pros and cons (class I).  

Sometimes we have blind spots to these things in the medical community, where it’s all about benefit and we ignore harm, and not just harm but also disutility. Andrew Foy

Andrew Foy, MD (Penn State Health, Hershey, PA), who was not involved in the analysis, said that while there are many patients potentially eligible for starting statins, and there is ample evidence to show that statins reduce the risk of coronary heart disease, the benefit these drugs in primary prevention is likely small. “[That said] I think the potential benefit of any primary prevention [therapy] is small,” he said. “It’s a matter of degrees, but at the population level, it does lead to a reduction in events.”

Regarding this new analysis, Foy called the net-benefit calculations extremely “complex,” noting that it would require a PhD in biostatistics to fully understand the methodology and how the model was developed. Outside of those complexities, though, Foy agreed that harm from medical therapy, along with patient preference, should be taken into account.

“Sometimes we have blind spots to these things in the medical community, where it’s all about benefit and we ignore harm, and not just harm but also disutility,” he told TCTMD. “That disutility might just involve taking a daily pill, things that we don’t really consider but do affect the equation.”

Risk-Benefit Model Factoring in Patient Preference

Using a model created at the National Cancer Institute, Puhan and colleagues sought to find the 10-year CVD risk threshold where the benefits of statins would outweigh the risk in men and women aged 40 to 75 years. They evaluated the benefits and risks of statins overall, as well as individual therapy with atorvastatin, rosuvastatin, simvastatin, and pravastatin, across 5-year age groups.

With data from the randomized trials, they estimated the number of fatal and nonfatal cardiovascular events prevented with statins, and also estimated the harms attributable to statins from not only the randomized trials but also observational studies. The model also accounted for competing mortality. Additionally, they factored in patient preferences for statin therapy using surveys of primary-prevention patients in Switzerland and Ethiopia and included preferences of events important to patients that they wished to avoid.

In doing so, they developed a weighted benefit-harm index to identify thresholds where benefit exceed risk.

In men aged 40 to 44 years, statin therapy was deemed beneficial for those with a 10-year CVD risk of 14% or greater. The risk threshold increased as patients aged, with a favorable risk-benefit profile observed in men aged 50 to 54, 60 to 64, and 70 to 75 years with a 10-year CVD risk of 15%, 18%, and 21%, respectively. In women aged 40 to 44 years, statin therapy was beneficial in those with a 10-year risk of 17% or greater. Again, the threshold increased with age such that risks outweighed the benefits in women 50 to 54, 60 to 64, and 70 to 75 years with a 10-year risk of 18%, 20% and 22%, respectively.

Individually, atorvastatin had the most favorable benefit-harm balance, followed by rosuvastatin. For example, among men aged 45 to 49 years, a net benefit of atorvastatin was observed in subjects with a 10-year CVD risk of 15%. With rosuvastatin, pravastatin, or simvastatin, the risk threshold was 18%, 19%, and 21%, respectively.

‘One Size Doesn’t Fit All’

To TCTMD, Puhan said that while the guidelines recommend a conversation with patients about risks and benefits of statin therapy, such a discussion often doesn’t happen. Most importantly, on a population level, the 10-year risk thresholds regarding risk and benefit don’t incorporate patient disutility: the adverse or harmful effects of carrying out a behavior, such as taking medicine, over a long period of time.

“One size doesn’t fit all,” said Puhan. “We showed that not only is the threshold [for initiating statins] substantially higher, but the benefit-harm balance depends very much on age, sex, and the type of statin used.” He added that he would like to see future guidelines incorporate more explicit harm-modeling with respect to medical therapy. Physicians should “look at cardiovascular risk and also look at sex and age, and select the most favorable statin with the best benefit-harm balance,” said Puhan. “Think about higher thresholds to initiate statins and don’t go as low as 7.5%.”

Foy questioned some of the harms identified and incorporated into the risk-benefit model, which included adverse event outcomes such as myopathy, renal dysfunction, hemorrhagic stroke, hepatic dysfunction, type 2 diabetes, any cancer, cataracts, and headache/nausea. Many of the harms are not from randomized clinical trials but instead from observational data, which makes the results challenging to interpret. Moreover, the risk of certain adverse events, such as cataracts, are small and might be given too much weight in the modeling, suggested Foy.  

“The whole point of their message, which I do agree with, is that guidelines really only look at benefit when making considerations for when to start statin therapy,” said Foy. For some primary-prevention patients, the small theoretical reduction in CVD risk with statins might not be as important as avoiding exposure to potential harm. “If that’s true, then yes, it does change things a lot,” said Foy, “but I think if shared decision-making is done properly, it will account for a lot of this.”   

Overall, Foy had reservations about the researchers’ conclusion to raise the 10-year CVD risk threshold for starting statins in primary prevention. “And I’m someone who doesn’t think we should be putting statins in the drinking water,” he said. “I’m kind of concerned about the strength of their conclusions based on this complex modeling and methodology.”

‘Uncertainty Lurks Beneath Every Choice’

In an editorial, Ilana Richman, MD, and Joseph Ross, MD (Yale University School of Medicine, New Haven), point out that the cut point of 7.5% has become “instantly recognizable” to primary care physicians and cardiologists, although the threshold for starting statins in primary prevention has long been controversial because many older patients achieve this degree of risk with age alone. The 2018 cholesterol guidelines keep the 7.5% threshold in place but also emphasize patient preference and the use of coronary artery calcium screening to help guide decision-making.

In contrast, the United States Preventive Services Task Force (USPSTF) recommends statins for primary prevention in adults with one clinical risk factor and a 10-year CVD risk of 10% or higher (grade B recommendation). Even the USPSTF suggests statins may be reasonable in lower-risk patients, those with a 10-year risk of 7.5% to 10%, depending on the circumstances.        

Richman and Ross state that the new study findings “paint a nuanced—if less optimistic—picture of the net benefits of statins, particularly in older adults who may not live long enough to benefit.” Like Foy, they question some of the adverse events included in the model, pointing out that hemorrhagic stroke, acute kidney injury, and hepatic dysfunction have largely been dismissed by the ACC/AHA and USPSTF. Regardless, Richman and Ross believe the study can help physicians inform patients who might be more concerned about side effects than potential benefits.

“[The] primary prevention of CVD must be patient-centered, because healthy patients are asked to assume risk, benefits are experienced only as the absence of disease, and uncertainty lurks beneath every choice,” they write.