Despite headlines, role of statins in cancer prevention far from proven

The potential role of statins in cancer prevention has received tremendous attention from the medical research community and the media.

Results from several recent studies suggested that statins may reduce risk for and mortality from many types of malignancies. Bold headlines — such as “Super statins beat cancer” and “Statins: The next miracle drug?” — appeared on websites and front pages around the world.

When it comes to complex scientific investigations, however, the headlines do not always tell the entire story.

“It shows up on the news that statins are protective of cancer,” Dale R. Shepard, MD, PhD, staff physician in the department of solid tumor oncology at Cleveland Clinic’s Taussig Cancer Institute, told HemOnc Today. “This might be a disservice because it has not been studied properly. A number of trials are investigating how statins may prevent cancer or recurrence. For the most part, though, these tend to be observational studies, which obviously do not carry the weight that prospective trials do. The general public does not get this message.”

In observational studies, researchers may ask questions related to lifestyle choices and medication use, but the inquiries are not specific to statin use and cancer.

“These studies are not necessarily investigating or demonstrating cause and effect,” Shepard said.

Frank de Vries, PhD, PharmD, assistant professor at Maastricht University Medical Centre and Utrecht University in the Netherlands, remains unconvinced.

“There have been various meta-analyses of randomized clinical trials that compared statin use with placebo,” de Vries told HemOnc Today. “The risk of cancer was not increased or decreased. Statins do not impact cancer risk.”

Although Shepard and others agree that any potential associations are far from proven, they acknowledge some data sets culled from large sample sizes contain encouraging findings that may provide clues to statins’ mechanisms of action or the downstream benefits they may offer.

Those results have revitalized a debate Shepard called “a roller coaster ride.”

“If you look in the literature in late 1990s and early 2000s, you will see a great deal of information on statins being like chemotherapy and having direct antitumor effects,” he said. “That excitement has died, but recent research has put the argument in the forefront again.”

Downstream effects

Nearly one in four Americans aged older than 45 years take a statin, according to the National Center for Health Statistics.

Statins help treat and prevent heart disease by blocking the enzyme HMG-CoA reductase, which the body requires to make cholesterol.

There may be additional downstream benefits that could be beneficial in the prevention of first cancers or recurrence, according to Marco Mielcarek, MD, associate member of the clinical research division at Fred Hutchinson Cancer Research Center and associate professor of medicine at the University of Washington. These include inhibition of cell signaling, induction of apoptosis and dampening of inflammation.

However, some reports from observational studies correlate decrements in cholesterol levels in statin-treated patients with increments in cancer risk, Mielcarek said.

“Some investigators suggested that the reported inverse relationship between cancer risk and cholesterol reduction may be related to statin-induced increases of regulatory T-cell numbers in the blood,” Mielcarek told HemOnc Today. “Some investigators believe that regulatory T cells play a role in reducing our immune system’s ability to eradicate early cancerous cells.”

Investigations in animal models have shown that when cholesterol is lowered, testosterone levels in tumors are lowered, said Stephen J. Freedland, MD, of the departments of surgery and urology at Duke University.

“Blood serum testosterone levels do not change on a statin, but tumor testosterone does,” Freedland said. “Hormonally sensitive tumors, like prostate and breast tumors, might be impacted.”

Reduced mortality

In November, Nielsen and colleagues published a population-based study in The New England Journal of Medicine that showed a reduction in cancer-related mortality among statin users in Denmark.

The study involved participants aged younger than 40 years who were diagnosed with cancer between 1995 and 2007. The analysis included 18,721 statin users and 277,204 participants who had never used statins.

The findings suggested that statin use was associated with reduced all-cause mortality (HR=0.85; 95% CI, 0.83-0.87) and cancer-related mortality (HR=0.85; 95% CI, 0.82-0.87). The investigators observed the association across all 13 cancer subtypes studied.

This paper had the “greatest impact” of any of the recent studies related to statins and cancer, Mielcarek said.

“Statin use among 300,000 patients with hypercholesterolemia lowered risks of overall mortality and cancer specific-mortality with no apparent dose-response effect,” he said. “Even the smallest dose had a protective effect.”

Shepard raised questions about the dose response.

“Typically, we like to think that if it is good for prevention, twice as much is twice as good,” he said. “In this study, patients who had higher doses didn’t have the benefit. Patients with less statin did better. It does not make sense logically that a higher dose wouldn’t be as protective. A randomized controlled trial may provide some answers.”

Mielcarek noted a final concern with the Nielsen findings.

“Population studies tend to have a healthy user bias,” he said. “The patients who fared better may have been more conscious of health anyway. It is difficult to avoid this in this kind of study.”

Freedland agreed.

“In general, we see two types of people on statins,” he said. “One type is someone who has high cholesterol, which may imply that they are a smoker or have poor diet or exercise habits, so they have a high cancer risk anyway. The other is someone who has become conscious about their health and they are changing dramatically for the positive, which includes improving their habits.”

The challenge in studying these two cohorts is fairly straightforward, 
Freedland said.

“Is it the statin that is impacting the cancer risk, or is it everything else that being on a statin tells us about a person?” he said.

Effects on HCC

Singh and colleagues conducted a systematic review and meta-analysis to determine the effect of statins on risk for hepatocellular carcinoma (HCC). The analysis — results of which were published earlier this year in Gastroenterology — included 4,298 cases of HCC among 1,459,417 patients eligible for review.

The adjusted OR for developing HCC among statin users was 0.63 (95% CI, 0.52-0.76), researchers wrote.

HCC is an increasing problem, according to William Sanchez, MD, of the departments of gastroenterology and hepatology at Mayo Clinic in Rochester, Minn.

“As baby boomers age, we are seeing a big population who have hepatitis C virus,” Sanchez, a senior researcher on the Singh paper, told HemOnc Today. “As they age, they are getting cirrhosis. As a result, we are seeing increased incidence in this malignancy.”

Studies such as these have shown there is benefit to statin therapy in this population, Sanchez said. However, the mechanism of action is not yet clear.

“There are a lot of potential mechanisms, but the data are inconclusive,” he said.

Still, because statins are taken up by and enriched in the liver, it may be “a good organ to start with when you look at cancer chemo-prevention,” Mielcarek said.

The trials included in the Singh study were retrospective so, although those who took statins had a lower cancer risk, cause and effect are difficult to prove, Sanchez said.

Study location may have caused heterogeneity within the results, according to Singh and colleagues.

The OR in the four studies conducted in Asia was 0.52 (95% CI, 0.42-0.64) compared with an OR of 0.67 (95% CI, 0.53-0.85) among patients included in six studies conducted in Western countries. Study design also may have been a factor, as the adjusted OR for the seven observational studies was 0.60 (95% CI, 0.49-0.73) and the adjusted OR for three clinical trials was 0.95 (95% CI, 0.62-1.45).

Tsan and colleagues reported similar results in a study published in April in the Journal of Clinical Oncology.

The investigators conducted a population-based cohort study of 260,864 patients with HCV enrolled in the Taiwan National Health Insurance Research Database between 1999 and 2010.

The researchers reported 1,378 cases of HCC among 35,023 statin users and 26,505 cases of HCC among 225,841 non-statin users. The researchers reported a dose-response relationship between statin use and HCC risk. Adjusted HRs were 0.66 (95% CI, 0.59-0.74) for patients with 28 to 89 cumulative defined daily doses (cDDD) per year, 0.47 (95% CI, 0.40-0.56) for patients with 90 to 180 cDDD per year, and 0.33 (95% CI, 0.25-0.42) for patients with more than 180 cDDD per year relative to nonusers.

“The risk reduction was very pronounced in Asian populations,” Sanchez said. “We know that the main driver for HCC in Asia is hepatitis B virus, which is different than what we see in the West. This may be a hypothesis we can work from or a population we can target.”

Role in breast cancer

Brewer and colleagues presented a retrospective cohort study at the 2012 San Antonio Breast Cancer Symposium. Results of the analysis, which included 724 patients with primary breast cancer, showed hydrophilic and lipophilic statin use were associated with improved DFS and OS times across the entire cohort.

Results of a multivariate analysis adjusting for radiation therapy, as well as ER, PR and HER-2 expression, indicated that hydrophilic statin use was associated with improved DFS compared with no statin use (HR=0.49; 95% CI, 0.28-0.84). Researchers also observed a nonsignificant link between hydrophilic statin use and time to mortality.

“Findings from this retrospective study confirm that women with inflammatory breast cancer have a poor prognosis,” Mielcarek said. “The results here are consistent with the other papers with regard to statin use, and they are hypothesis-generating. However, they are by no means convincing.”

The data contain no statistical significance, and there may be larger issues to contend with in patients with a poor prognosis, Shepard said.

“Again, the problem is that there is a little bit of bias in terms of why people may have been taking statins,” he said. “We are not trying to prevent heart disease in 5 years in patients with poor-prognosis cancers, so it is not likely that they will be put on statins. If they have liver disease, they won’t be put on statins. You could be self-selecting healthier patients.”

Randomized trials

Several studies presented at the 2013 AACR Annual Meeting further propelled the argument that statins may offer a protective benefit against cancer.

Murtola and colleagues conducted a population-based study of 31,236 breast cancer patients with a new diagnosis in Finland between 1995 and 2003. Statin use after diagnosis was associated with a 67% reduction in breast cancer mortality among women with localized disease and a 48% decrease for women with metastatic disease.

Amsler and colleagues reported an 83% reduction in recurrence among 60 women with ovarian cancer who were taking statins to control diabetes.

McGlynn and colleagues conducted a study that included 75 patients with liver cancer and 373 controls. Univariate analysis indicated statins were associated with a 40% reduction in liver cancer risk, and a nonsignificant 52% reduction was observed in multivariate analysis.

Despite the conclusions of recent studies, Mielcarek said he is not convinced that statins are linked to improved prognosis.

“Statins have known effects on cell-cycling and signaling and may therefore slow down the growth of certain tumors, but the mechanisms need to be investigated more thoroughly,” Mielcarek said. “Therefore, randomized controlled trials in selected high-risk populations would be most informative.”

De Vries said he was less optimistic.

“It is probably a waste of resources and — given the current evidence — possibly not even ethical to conduct,” he said.

Researchers who conduct such investigations would face tremendous challenges, including the time it would take to show benefit in terms of preventing malignancy.

“If you lower cholesterol, you lower heart disease risk immediately,” Freedland said. “Cancer processes occur over a much longer period of time. Smoking, ionizing radiation, these things take 10 to 15 years to increase cancer risk. You might see a benefit on cardiovascular outcomes in a year or two, but you won’t see a cancer benefit.”

Sanchez put the problem in financial terms.

“It is impractical, particularly here in the United States, to do a trial like this because of the funding,” he said. “It may be better to do it in a place with a single-payer system, where there is room for more prospective research to prove cause and effect.”

Mielcarek agreed.

“How many patients would have to take statins to prevent one case of cancer per year?” he said. “It could be more than 5,000. This raises red flags.”

Targeted investigations

A randomized controlled trial might be effective if investigators carefully selected a high-risk population, Mielcarek said.

“We may get some answers if we look at Asians at a high risk for liver cancer, for example, or more in general at secondary prevention in patients in remission who have a high relapse risk,” he said.

Shepard said there has not been enough evidence of antitumor effects to show statins as a primary antitumor agent, adding that an adjuvant trial would be a better option.

“It will require a long time to get the data,” Shepard said. “We will need two arms — one with statins and one without — and we should look at which one has recurrence faster [and] which one has better OS.”

An investigation of patients with tumors that have been resected may be helpful, Sanchez said.

“Those patients are already at a high risk,” he said. “This type of secondary prevention might be easier to study than primary prevention.”

Shepard said a trial designed to minimize risk for recurrence in patients who have had cancer or resection may be feasible.

“You might be able to recruit patients with an increased likelihood of recurrent tumor who would be more likely to follow along with the study,” he said. “More of these people will get cancer than the general population.”

A study in the metastatic setting also may be effective, Shepard said.

“As the tumor metastasizes, we could observe how quickly — or if — the tumor responds,” he said. “But this will present clinical challenges because you will be simultaneously trying to treat the tumor, prevent progression and study the drug.”

Statins work at several different checkpoints of cell proliferation and dampen inflammatory response, which may make research difficult, Sanchez said.

“As we have seen, the mechanism might be different in an Asian man vs. a white woman with diabetes,” he said. “If we identify the mechanism in each patient group, we might see where they can provide benefit, but that will present challenges.”

The right message

Reporting complex trial results to the general public is difficult, and regardless of how future trials are designed, increased efforts must be made to ensure the results are kept in perspective, Shepard said.

Failure to do so can be a disservice to patients and the clinicians who treat them.

“Cancer is a bad disease that impacts a lot of people,” Shepard said. “We often don’t have great therapies. When something comes along that looks encouraging, it gathers attention. It might be semantics, but people hear what they want to hear when these headlines appear.”

He used the study by Singh and colleagues — an observational study in which patients on statins had less likelihood of developing liver cancer — as an example.

“The headline reads, ‘Statins linked to reduced liver cancer,’” Shepard said. “If you look at the study, it could be true — there are fewer liver cancers in those patients — but it is a big jump to make. The nuances are lost. ‘If you use statins, you get less cancer’ becomes the final message, even if it is not the right message.”

Information fatigue also is a factor, he said.

“Every day, we get an overload of observational trials,” Shepard said. “After a while, you start either believing them all or believing none of them. Patients come into clinic confused about what they should or shouldn’t eat, what they should or shouldn’t do, what causes prevention or what doesn’t. We have to be more careful about how we present this information.” – by Rob Volansky

References:

Amsler ME. Abstract #LB-18. Presented at: American Association for Cancer Research Annual Meeting; April 6-10, 2013; Washington, D.C.

Brewer TM. Abstract #PD03-08. Presented at: San Antonio Breast Cancer Symposium; Dec. 4-8, 2012; San Antonio.

McGlynn KA. Abstract #4829. Presented at: American Association for Cancer Research Annual Meeting; April 6-10, 2013; Washington, D.C.

Murtola TJ. Abstract #136. Presented at: American Association for Cancer Research Annual Meeting; April 6-10, 2013; Washington, D.C.

Nielsen SF. N Engl J Med. 2012;367:1792-1802.

Singh S. Gastroenterology. 2013;144:323-332.

Tsan YT. J Clin Oncol. 2013;31:1514-1521.

For more information:

Frank de Vries, PhD, PharmD, can be reached at Division of Pharmacoepidemiology and Clinical Pharmacology, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, P.O. Box 80125, 3508 TC Utrecht, the Netherlands; email: f.devries@uu.nl.

Stephen J. Freedland, MD, can be reached at DUMC 2626, Durham, NC 27710.

Marco Mielcarek, MD, can be reached at Fred Hutchinson Cancer Research Center, 1100 Fairview Ave. N., P.O. Box 19024, Seattle, WA 98109-1024; email: mmielcar@fhcrc.org.

William Sanchez, MD, can be reached at Mayo Clinic, 200 First St. SW #W4, Rochester, MN 55905.

Dale R. Shepard, MD, PhD, can be reached at Department of Solid Tumor Oncology, Cleveland Clinic, Mail Code R35, 9500 Euclid Ave., Cleveland, OH 44195.

Disclosure: Sanchez reports grant support from Esai Pharmaceuticals and Phillips. De Vries, Freedland, Mielcarek and Shepard report no relevant financial disclosures.

 

Does evidence support the hypothesis that statins may be linked to increased cancer incidence?

Yes.

In addition to lowering blood cholesterol, statins have been shown to increase the numbers and functionality of peripheral regulatory T cells and decrease natural killer cell cytotoxicity. Although these immunosuppressive properties are beneficial in stabilizing atherosclerotic plaque and reducing transplant rejection, they may impair host anti-tumor immune responses, resulting in cancer promotion.

Age is the leading risk factor for prevalent cancer, and older individuals commonly have pre-clinical microscopic foci of cancer cells maintained in dormancy in part by intact immune surveillance. So, is there evidence suggesting that statins may be linked to increased cancer incidence? In short, yes.

Skin cancers appear to have been increased as a result of statin therapy. Immunosuppression is a well-known risk factor for nonmelanoma skin cancer (NMSC), and early randomized statin trials have shown a significantly increased risk of NMSC in statin-treated patients. It is noteworthy that NMSC has been increasing in the United States over the past 2 decades, paralleling the increased use of statins. Disturbingly, recent observational data have demonstrated a significantly increased incidence of Merkel cell carcinoma (MCC) among younger statin users. This highly malignant cutaneous neuroendocrine cancer is typically associated with immunosuppression.

Randomized trial evidence suggests statin therapy may be associated with other cancers, as well. One early statin trial was notable for a striking and highly significant increase in breast cancer (some were recurrences) among women randomly assigned to pravastatin compared with placebo (The Cholesterol and Recurrent Events Trial Investigators. N Engl J Med. 1996;335:1001-1009). Since that trial, subjects with a cancer history have been excluded from statin trials.

Interestingly, the only randomized statin trial specifically involving elderly subjects (mean age at trial entry was 75) demonstrated a significant 25% increase in cancer incidence over the 3.2-year trial among the subjects randomly assigned to statins (The Prosper Study Group. Lancet. 2002;360:1623-1630). This resulted in an increase in cancer mortality that equaled in magnitude the decrease in cardiovascular mortality, leaving overall mortality unchanged.

Therefore, the immunosuppressive actions of statin therapy may increase the incidence of some dermatologic malignancies and allow the promotion of dormant pre-clinical cancers, particularly in the elderly.

Randomized statin trials have been of relatively short duration and have largely excluded subjects with prevalent or remote cancer. This is problematic because statins are prescribed without regard to cancer history and potentially given to diverse age groups for decades. Moreover, the post-market surveillance of drug safety has been poor.

Studies are desperately needed to fully access the long-term safety of the immunosuppressive actions of statin therapy in various segments of the population. Finally, the ultimate goal of statin therapy is to decrease cardiovascular disease and not trade one morbidity or mortality for another.

Mark R. Goldstein, MD, FACP, is a staff physician with NCH Physician Group in Naples, Fla. He can be reached at NCH Physician Group, 1845 Veterans Park Drive, Suite 110, Naples, FL 34109. Luca Mascitelli, MD, is a physician with the medical service at Comando Brigata Alpina "Julia" in Udine, Italy. He can be reached at Medical Service, Comando Brigata Alpina "Julia", 8 Via S. Agostino, Udine 33100, Italy. Disclosure: Goldstein and Mascitelli report no relevant financial disclosures.

No.

The current evidence does not support that statin use would be linked with increased cancer risk.

In fact, there is currently enough evidence to show that use of statin drugs does not affect cancer risk adversely.

This has been demonstrated by several clinical trials [Cholesterol Treatment Trialists’ (CTT) Collaborators. PLoS One. 2012;7:e29849].

Clinical trials are widely considered to provide most reliable and bias-free information. The CTT Collaborators performed a meta-analysis of data from 27 randomized trials that included a combined 175,000 people. The researchers determined the use of statins for about 5 years to lower LDL cholesterol had no effect on newly diagnosed cancer or death from cancer.

Additionally, the data showed no association between statin therapy and cancer incidence/mortality with increased duration of treatment, with any one statin or in any patient subgroup.

The possible protective effect of statins against cancer is a more controversial but promising topic. Although statins may not affect the risk of getting cancer, there is some preliminary evidence suggesting that statins could limit disease progression and ultimately cancer death (Nielsen SF. N Engl J Med. 2013;368:576-577). However, this has not yet been proven in clinical trials on cancer patients. Therefore the time is not yet ready to recommend statins for cancer indications.

Teemu Murtola, MD, PhD, is a post doc researcher in the department of epidemiology at Johns Hopkins Bloomberg School of Public Health. He can be reached at tmurtola@jhsph.edu. Disclosure: Murtola reports stock ownership in Orion Pharma, as well as participation in scientific conferences at the expense of Ferring Pharmaceuticals and Astellas.