Hype is promotion of an intensive extravagant nature that is often deceptive and frequently meant to stimulate the recipient to do or not do something. There is both positive hype (meant to stimulate action) and negative hype (to discourage action). Hype can be differentiated from hyperbole (and what in advertising is called “puffery”), an exaggerated statement or claim not meant to be taken literally. Complicating matters is the imprecision of language; ambiguity and vagueness abound.
We should separate hype from lies, although in many cases extreme hype tends toward lying. But to define lies we must first define truth. Truth applies to statements (propositions). A statement is true if:
- 1) The statement corresponds to reality;
- 2) The statement is coherent with other facts;
- 3) The statement has pragmatic implications, e.g., predictions based on the statement have not been falsified.
Lies are statements that are not true and may be told with or without intent to deceive.
In America and much of the developed world, hype is present everywhere: on the Internet and in advertisements, television, newspapers, and books. Most of the time, educated persons can separate hype and lies from the truth, but there are situations where hype is often “effective” and harmful or can cause great uncertainty, e.g., in medicine, government, or the law. Before I discuss some medical examples of hype, some important background information is in order.
Physicians, including pediatricians, radiologists, and surgeons, have five main objectives:
- 1) Whenever possible, physicians try to prevent disease (e.g., with vaccines);
- 2) If disease occurs, they try to cure it with medicine, surgery, or radiation;
- 3) If a cure is not possible, they try to prolong and improve the quality of life;
- 4) Physicians often give patients support to normalize troublesome symptoms (e.g., remedies for insomnia and constipation);
- 5) Physicians try to achieve these objectives at reasonable cost (value).
To accomplish these goals, physicians must often deal with variabilities among individuals, especially in prognosis and their response to therapy. It is true that in some cases (e.g., vaccination against hepatitis A and B or papilloma virus) there is a 100 percent response and complete protection against the viruses in the vaccine. However, in many situations, the responses to drugs and vaccines are variable. To overcome this variability in drug response and the inability to identify in advance which patients will respond, regulatory bodies (e.g., the U.S. Food and Drug Administration [FDA]) have generally required the application of scientific methods, which include replicative randomized controlled trials (often blinded) comparing the new treatment versus a placebo or appropriate control for licensure and sale of prescription drugs. If there are enough patients in the trials and they correspond to the patient population to be treated, these trials allow a reasonable estimate of efficacy (e.g., whether the patients will live or die) and toxicity (although rare severe toxicity may occur post-marketing when large numbers of patients are treated).
Some cancer patients cannot be cured by surgery, radiation, or drugs. In such cases, there is general agreement that their treatments:
1) Should prolong life;
2) Should improve the “quality” of remaining life; and
3) Should have a cost consistent with its value in 1) and 2) above (Workman et al. 2017).
It is often difficult in controlled trials (new drug versus placebo or standard regimen) for technical and ethical reasons to find the average prolongation of life in cancer therapy, so median survival is often employed (i.e., the time for half the patients to die). This is a practical endpoint and is widely accepted by the FDA and public as meaningful (Hsu and Shaw 2017); there is no ambiguity about whether the patients are alive or dead.
In terms of the epidemiology of cancer in the United States, for the first time, cancer deaths in the United States in 2017 will probably exceed heart disease as the number one cause of death (over 600,000) (Spector 2016). In 1950, heart disease killed many times more patients than cancer, and stroke was the second leading cause of death (Spector 2016; Mozaffarian et al. 2016). Now stroke is the fifth leading cause of death in the United States with about 130,000 deaths annually. Since 1950, there has been a remarkable decline in deaths in age and sex matched populations in the United Sates due to heart disease and stroke (Mozaffarian et al. 2016). This is largely due to medications such as statins and aspirin, blood pressure control (with medications), and lifestyle (risk factor) modifications including diet, weight loss, exercise, and cessation of smoking. Some argue that cancer is a disease of the elderly and therefore has increased, but so is heart disease and stroke.
Although there have been multiple waves of hype about cancer therapy in the United States since 1950, there indeed has been striking progress in treating some cancer patients with drug and radiation therapy; for example, choriocarcinoma, testicular cancer, Hodgkin’s disease, and several relatively uncommon childhood cancers can in most cases be cured. Moreover, the number of smokers has declined appreciably, resulting in a significant decline in lung cancer. There has also been a largely unexplained decline in stomach cancer. Most impressive is the fact that two vaccines can prevent cancer—hepatitis B vaccine can prevent liver cancer and human papilloma vaccine can prevent most cases of cancer of the cervix and other rare cancers. However, the common cancers—e.g., lung, breast, prostate, stomach, pancreas, brain, and others—still kill more than 600,000 each year in the United States, and the number is rising. Later I will discuss the reasons for the widespread failure of cancer therapy. Now to examples of hyped drugs.
Avastin (bevacizumab) and Opdivo (nivolumab)
Over ten years ago, there was in the press, on the Internet, and on the television tremendous hype about the possibility of interfering with vascular endothelial growth factor (VEGF) with an antibody bevacizumab (Avastin). The idea was bevacizumab could stop blood vessels growing into the tumors by blocking VEGF action and hence “starve the tumor” of nutrients and oxygen; this mechanism of action would slow or stop tumor growth and spread. For a decade, Avastin has been widely used, and about $6 billion per year has been spent purchasing the drug (Tucker 2011). But according to the FDA-approved product insert for Avastin (accessed June 2017), Avastin does not increase median survival (MS) compared to controls in breast, brain, or kidney cancer. In fact, in a recent definitive controlled study of Avastin and brain cancer, not only did Avastin not increase MS, improve the quality of life, or decrease steroid useage, Avastin greatly increased serious side effects. Avastin has no place in the therapy of brain cancer (Wick et al. 2017). In lung cancer, one study showed a two-month increase in median survival; a second study showed no increase. Finally, in the initial study in metastatic colon cancer, there was a four-month increase in medium survival (twenty versus sixteen months). However, a second study (in the product insert) showed a two-month increase in MS (thirteen versus eleven months). Subsequent studies showed less than two months.
In 2007, the FDA gave accelerated approval for Avastin for breast cancer because the drug delayed cancer progression (a surrogate end point) as measured by X-rays, MRI, and other techniques (Tucker 2011). The price of Avastin was set at about $10,000/month (Tucker 2011; Workman et al. 2017). However, the FDA properly demanded follow-up breast cancer studies to confirm the value of the surrogate end point (delayed cancer progression), and in these controlled studies there was no increase in median survival. Moreover, in subsequent studies, there was also no improvement in median survival and no convincing data showing improvement in the quality of life (Tucker 2011). Therefore, with an overwhelming vote by the FDA advisory committee (against approving Avastin for breast cancer), the FDA rescinded its approval in 2011 based on lack of efficacy and the significant toxicity of Avastin. The Roche Company appealed the FDA decision and fought vigorously to retain its indication for breast cancer but lost (Tucker 2011). In its appeal of the revocation of licensure of Avastin for breast cancer, Tucker (2011) commented that the Roche Company employed patient and oncologist anecdotes. But Tucker (2011) pointed out that “anecdote is not science,” and that “testimonials” conceal “the many patients who have been treated with Avastin but are not here to tell their stories. Serious progress will not be the result of polemics, lobbying or marketing.”
So I would ask after ten years and the above data why Avastin is still used if it doesn’t increase median survival (except possibly by two months or less in colon cancer), has substantial and serious side effects, and for its cost, has relatively little value (Workman et al. 2017).
I offer the following explanations based on the literature:
- 1) The original hype concerning the notion of “starving” tumors of nutrients is still bandied about;
- 2) Oncologists have an incentive to use this and other intravenous drugs because they supplement their incomes by infusing drugs (Smith and Hillner 2011);
- 3) Avastin in some cases causes shrinkage of tumors on X-rays, MRI, or CAT scans. This shrinkage can be very impressive to the patient and family. However, this often does not translate into improved median survival or improved quality of life except as noted above;
- 4) Hype often is translated (consciously or unconsciously) into hope; and
- 5) Patients, their families, and interest groups often press hard to do something—anything.
Unfortunately, Avastin for breast cancer is not the only example of a cancer drug approved initially based on surrogate end points (e.g., delay of time to progression of the cancer or tumor shrinkage by CAT or MRI scans) that did not turn out to save lives. From 2008 to 2012, thirty-six cancer drugs were approved based on surrogate endpoints (Kim and Prasad 2015). However, only five were subsequently proven to prolong life; eighteen did not, and for the remainder, any potential survival benefit remains unproven (Kim and Prasad 2015). Kim and Prasad (2015) state that “most cancer drug approvals have not been shown to, or do not, improve clinically relevant end points,” (e.g., death). They conclude that the FDA has been approving “many costly, toxic drugs that do not improve overall survival. Enforcement of post-marketing studies is therefore of critical importance.” Yet the FDA “has never exercised its authority to remove” the eighteen worthless products from the market as in the case of Avastin for breast cancer (Kim and Prasad 2015). Two years later, Bauer and Redberg (2017) emphasized that the FDA is not demanding rapid withdrawal of these worthless drugs as “outlined in the Federal Food, Drug, and Cosmetic Act.”
A second recent instance of hype is exemplified by one of the newer biological drugs (antibodies) designed to stimulate the immune system to kill cancer cells. These drugs have been the subject of intense interest on television and the Internet and in print media. One example is nivolumab (Opdivo), which must be given (like Avastin) by intravenous infusion.
In non-small-cell lung cancer (NSCLC) patients, the manufacturer’s advertisement states (June 17, 2017; Time magazine) that: “This (Opdivo) is big. A chance to live longer.” But what are the facts? In the first trial, described in smaller print in the advertisement, in patients who failed platinum-based chemotherapy, patients on Opdivo had nine-month median survival versus six months in controls; a net improvement of three months. In a second trial, lung cancer patients on Opdivo had a twelve-month median survival versus nine months for controls: again, a net increase of three months in median survival on Opdivo. So, in both trials, one half or more of the patients on Opdivo were dead within one year. Moreover, there are significant side effects due to Opdivo in these trials including pneumonitis, colitis, hepatitis, encephalitis, and hormonal unbalances apparently due to excessive stimulation of the immune system. The cost of this drug is about $10,000 per month (Workman et al. 2017).
Is this really “big” in lung cancer, taking into account the entire data set and cost? The company also presents no data supporting an improvement in the quality of life. Moreover, there are no data showing what happens to the patients when three-fourths of the control patients have died. Do the Opdivo patients still show an advantage, or, as has happened frequently with other drugs, is the median survival advantage lost? Are there any long-term lung cancer survivors on Opdivo versus controls? We don’t know as of this writing.
The reader may ask, “Why is this not just a private issue between the patient and his or her oncologist?” The answer is in part because the government (Medicare, Medicaid, the VA system) pays for many of these drugs, so we taxpayers foot the bill (Smith and Hillner 2011). Therefore, all citizens have an interest in the “value” aspect of treatment (Workman et al. 2017). If Avastin and Opdivo are really marginal, why should the government support such expensive drug therapy? In a large controlled trial, hospice care led to “happier” cancer patients and better results than those subjected to heroic and expensive end-of-life chemotherapy (Temel et al. 2010).
The reader may also ask why many drugs and biologicals such as Avastin and Opdivo that initially, in some cases, spectacularly shrink tumors ultimately and spectacularly fail. Tentative answers have recently emerged. First, multiple investigators have shown that, even in the original tumor, the cells from different parts of the tumor are different, i.e., the genetic and/or epigenetic make-up and the packaging of the genome in chromosomes are different depending upon where the sample is taken (Jamal-Hanjani et al. 2017; Robles and Zenklusen 2017). (When I was a medical student at Yale in 1966, my preceptor [Professor Manuelidis] showed me there were different numbers of chromosomes in cells from the same tumor.) So although most of the cells are sometimes sensitive to the drug or biological, a few are not, and these then take over in a Darwinian sense and grow with vigor and abandon (Jamal-Hanjani et al. 2017). Metastatic lesions are even less stable (Jamal-Hanjani et al. 2017). As Hsu and Shaw (2017) say in their commentary “Lung Cancer: A Wily Genetic Opponent,” “Therapeutic interventions that can limit tumor heterogeneity and reshape tumor evolution hold the key to improving cancer outcomes, including the most important outcome of survival.” This, of course, is a huge challenge.
Thus, the notion often hyped (in so-called “personalized medicine”) is that if you take a biopsy from one segment of a tumor and study it and then use drugs “effective” against the growth-promoting genes and their products in the biopsy, you can rationalize the therapeutic process. But as Hsu and Shaw (2017) and Robles and Zenklusen (2017) point out, even in different parts of the original tumor there are often profound differences in genetic, epigenetic, and chromosomal characteristics. Thus, a single biopsy and analysis will often be misleading.
A second even more recent finding in cancer cells is the discovery of circular, non-chromosomal rings of DNA, now called the “circulome” (Pennisi 2017). These rings often contain multiple copies of genes (e.g., oncogenes) that drive cellular division and spread. The finding of the circulome “basically opens up a new field and a new way of thinking about how dynamic” DNA is in cells, especially cancer cells (Pennisi 2017). The nature and origins of the circulome are under intense investigation, but we do know that viruses in humans can cause cancer (e.g., cervical and liver cancer) amenable to protection by vaccines as noted earlier.
There are also other characteristics of some cancer cells, such as their ability to engulf and employ the genes in stem cells, but these complex activities are only now beginning to be understood.
In sum, notwithstanding the hype, there has been relatively little progress in treating lung cancer and the other major cancer killers in the United States in the past twenty years relative to heart disease and stroke; about 600,000 patients in the United States died of cancer in 2016 (Spector 2013; 2016; Mazaffarian et al. 2016).
An excellent example of negative hype is the widely misreported side effects (especially affecting muscle) of statin drugs. Dr. Harriet Hall (2017) has discussed this negative hype at length in her recent article in Skeptical Inquirer, “Statin Denialism.” I will not repeat what she correctly points out but will make a few additional comments. The statin drugs are unique; they are effective in saving lives from heart attacks and stroke with beneficial effects that increase over time. (Full disclosure: I was in charge of development at Merck when for the first time in 1993 we showed that statins did not slow down coronary atherosclerosis by quantitative coronary angiography [Blankenhorn et al. 1993].) However, simvastatin did decrease deaths by about one-third (and also heart attacks and strokes comparably) in the blinded, randomized, placebo-controlled 4S trial in high-risk patients over five years (Scandinavian Simvastatin Survival Study Group 1994). It is now generally agreed—as promulgated in national guidelines (Spector 2013; 2016; Mazaffarian et al. 2016)—that all high-risk patients (e.g., those with a history of a heart attack, stroke, angina, or diabetes) should be treated with a statin (and aspirin) irrespective of their serum cholesterol. Statins benefit high-risk patients irrespective of the serum cholesterol. The notion of treatment to a cholesterol goal was a theory—now known to be incorrect. Current evidence suggests that the evaluation of risk, not serum cholesterol, is the important variable. (See Mazaffarian et al. 2016; Krumholz and Hayword 2010; Takagi and Umemoto 2013; Spector and Snapinn 2011; Spector 2013; and Spector 2016 for further detailed discussion of these points.) Simvastatin, now generic and very inexpensive, is the safest widely used statin (Naci et al. 2013) in the 20–40 mg dose range, and because of its proven safety and efficacy it is the drug of choice. (See below).
The primary mechanism of action of statins is to stabilize nonobstructing atheromatous wall plaques in open arteries so clots do not form on the plaques and obstruct blood flow to downstream tissue, thus preventing some heart attacks, strokes, and death. Aspirin also prevents arterial clotting although less effectively than statins. In patients with signs and symptoms of a heart attack or stroke, as quickly as possible, heroic efforts are made to find the “culprit” artery (i.e., the artery with the clot) and open it with either “clot busters” such as tissue plasminogen activator or mechanical means. If this is done quickly, cardiac or brain tissue distal to the clotted artery can be saved by reperfusion of the endangered tissue. It is worth noting that lowering serum cholesterol profoundly did not save lives in a two-year trial with a non-statin drug (Sabatine et al. 2017). In fact, there were slightly more deaths on that drug than on placebo although there were fewer heart attacks (Sabatine et al. 2017). This is consistent with previous data showing that lowering LDL cholesterol with very high doses of statins does not reduce mortality more than moderate doses (e.g., 20–40 mg simvastatin) (Spector and Snapinn 2011; Takagi and Umemoto 2013).
In summary, in controlled trials, doses of 40 mg or less of inexpensive generic simvastatin or atorvastatin show side effects comparable to placebo (Tobert and Newman 2016; Hall 2017). A very recent trial—the ASCOT-LLA trial—showed in the randomized double-blind portion no difference in muscle side effects of atorvastatin versus placebo (Gupta et al. 2017). But in the open extension (when the patients knew they were on drug), the patients voiced complaints about many “side effects” that they ascribed to the statin. The authors conclude:
These analyses illustrate the so-called nocebo effect, with an excess rate of muscle-related AE reports only when patients and their doctors were aware that statin therapy was used and not when its use was blinded. These results will help assure both physicians and patients that most adverse effects associated with statins are not causally related to use of the drug and should help counter the adverse effect on public health of exaggerated claims about statin related side effects. (Gupta et al. 2017)
These new data further strengthen the arguments made by Hall (2017) and Tobert and Newman (2016) that there are minimal side effects associated with statin therapy at moderate doses. However, at doses of simvastatin and atorvastatin higher than 40 mg, there are in fact significant side effects, and such doses should only rarely be employed except in special situations (Spector and Snapinn 2011; FDA labelling of statins).
The reader might ask: Why has the varying inter-individual responses to various drugs (e.g., the statins discussed here) not been a major problem? This is because the drug industry, the regulatory bodies, and the public have bought into a scientific approach to drug development, registration, and sale—although in some cases the FDA has been very slow in carrying out its mandate and the law as noted earlier. The science is based on large randomized (often blinded) controlled trials employing statistics and logic. The public has recognized that it is often difficult or impossible to identify which patients might respond to a treatment; therefore, the public accepts many patients may need to be treated for some to obtain the beneficial effect, especially if the drug has minimal side effects and is reasonably priced. With this approach, the FDA can assert with great confidence that Avastin does not prolong median survival in breast cancer patients and that simvastatin and atorvastatin do save lives and the life-saving effects increase over time. Moreover, Avastin has significant side effects whereas simvastatin and atorvastatin in the 20–40 mg daily dose range do not (Tobert and Newman 2016; Hall 2017). Unfortunately, as I have noted, there are still many oncology drugs on the market that have not been subjected to rigorous testing or that have not yet been removed from the market due to lack of clinically relevant efficacy (Kim and Prasad 2015; Bauer and Redberg 2017).
I have provided three significant examples of medical hype that matters. In some cases, hyping of new therapies is of no or minimal consequence. But in some situations, it greatly matters, and we must distinguish between true and hyped claims (e.g., the putative benefits of Avastin in breast cancer). The medical community, pharmaceutical industry, the regulatory bodies such as the FDA, and the public have brought the use of the scientific method into the development and registration of drugs. However, with some drugs such as statins, it took multi-year controlled trials to establish their safety and efficacy in preventing deaths (and heart attacks and strokes) and the proper way to employ them, i.e., use based on risk not serum cholesterol levels. However, although hype has been minimized in advertising prescription drugs, it still occurs. Is a two- or three-month increase in median survival in cancer therapy really “big” when taking into account the side effects and cost of the drug (Workman et al. 2017)? We are again back to definitional problems. In my view, we need a précising (Copi and Cohen 2009) definition of “big” in this situation. From my perspective, prolongation of life for years or a cure would be “big.” A three-month or less increase in median survival with many serious side effects at great monetary cost is “small” in my view and that of others (Workman et al. 2017). “Big” in this context is hype.
I thank Michiko Spector for her many valuable points and her aid in preparing the manuscript.
- Bauer, S.R., and R.F. Redberg. 2017. Setting the record straight on FDA approvals in oncology. JAMA Internal Medicine 177: 1223.
- Blankenhorn, D.H., S.P. Azen, D.M. Mack, et al. 1993. MARS research group: The monitored atherosclerosis regression study. Coronary angiographic changes with lovastatin therapy. Annals of Internal Medicine 119: 969–976.
- Copi I.M., and C. Cohen. 2009. Introduction to Logic. Pearson Prentice Hall: New Jersey.
- Gupta, A., D. Thompson, A. Whitehouse, et al. 2017. Adverse events associated with unblinded, but not with blinded, statin therapy in the Angio-Scandinavian Cardiac outcomes Trial—Lipid-Lowering Arm (ASCOT-LLA): A randomized double-blind placebo-controlled trial and its non-randomized non-blind extension phase. The Lancet Online. Available online at http://dx.doi.org/10.1016/S0140-6736(17)31075-9.
- Hall, H. 2017. Statin denialism. Skeptical Inquirer 41(3) (May/June): 41–43.
- Hsu, P., and A.T. Shaw. 2017. Lung cancer: A wily genetic opponent. Cell 169: 777–779.
- Jamal-Hanjani M., G.A.Wilson, N. McGranahan, et al. 2017. Tracking the evolution of non-small-cell lung cancer. New England Journal of Medicine 376(22): 2109–2121.
- Kim, C., and V. Prasad. 2015. Cancer drugs approved on the basis of a surrogate end point and subsequent overall survival. JAMA Internal Med 175: 1992–1994.
- Krumholz, H., and R. Hayward. 2010. Shifting views on lipid lowering. BMJ 341: 332–333.
- Mozaffarian, D., E.J. Benjamin, A.S. Go, et al. 2016. American Heart Association Statistics Committee and Stroke Statistics Subcommittee: Heart disease and stroke statistics—2016 update: A report from the American Heart Association. Circulation 133: e38–e360.
- Naci, H., J. Brugts, and T. Ades. 2013. The comparative tolerability and harms of individual statins: A study-level network meta-analysis of 246, 955 participants from 135 randomized controlled trials. Circulation Cardiovascular Quality and Outcomes 6: 390–399.
- Pennisi, E. 2017. Circular DNA throws biologists for a loop: Studies hint at health and disease roles of unusual DNAs. Science 356: 996.
- Robles, A.I., and J.C. Zenklusen. 2017. Seeing the forest through the phylogenetic tree. New England Journal of Medicine 376: 2190–2191.
- Sabatine, M.S., M.P.H. Giugliano, and A.C. Keech, et al. 2017. Evolocumab and clinical outcomes in patients with cardiovascular disease. New England Journal of Medicine doi: 10.1056/NEJMoa1615664.
- Scandinavian Simvastatin Survival Study Group. 1994. Randomized trial of cholesterol lowering in 4444 patients with coronary heart disease; the Scandinavian Simvastatin Survival Study (4S) Lancet 344: 1383–1389.
- Smith, T.J., and B.E. Hillner. 2011. Bending the cost curve in cancer care. New England Journal of Medicine 364: 2060–65.
- Spector, R. 2013. Efficacy and safety of statins: Focus on atorvastatin and simvastatin. Journal of Symptom Signs 2: 241–245.
- ———. 2016. New insight into the dietary cause of atherosclerosis: Implications for pharmacology. Journal of Pharmacology and Experimental Therapy 358: 103–108.
- Spector, R., and S.M. Snapinn. 2011. Statins for secondary prevention of cardiovascular disease: The right dose. Pharmacology 87: 63–69.
- Takagi, H., and I. Umemoto. 2013. Limits to benefits of large reductions in low-density cholesterol levels. JAMA Internal Medicine 173: 1028–1029.
- Temel, J.S., J.A. Greer, A. Muzikansky, et al. 2010. Early palliative care for patients with metastatic non-small-cell lung cancer. New England Journal of Medicine 363: 733–742.
- Tobert, J.A., and C.B. Newman. 2016. Statin tolerability: In defense of placebo-controlled trials. European Journal of Preventive Cardiology 23: 891-896.
- Tucker, F.C. 2011. Drugs and profits. New York Times (May 25).
- Wick W., T. Gorlia, M. Bendszus et al. 2017. Lomustine and bevacizumab in progrssive glioblastoma New England Journal of Medicine 377(20): 1954–1963.
- Workman, P., G.F. Draetta, J.H.M. Schellens, et al. 2017. How much longer will we put up with $100,000 cancer drugs? Cell 168: 579–583.