A Guest Editorial by Joel M. Kauffman, PhD, Professor of Chemistry Emeritus, University of the Sciences in Philadelphia
By now you must believe you have heard every possible objection and outrage to the Medicare Prescription Drug Benefit plan of 2003 that is now law. The loudest screams are opposites: that there are gaps in the benefit, the “doughnut hole”, which should be plugged, and the maximum benefit should be set higher, meaning that it ought to cost us taxpayers more. Others cry that drugs cost too much, that the plan is a drug company subsidy, so costs should be forced down because drugs are in the same class as a public utility. The drug companies object that this will stifle all innovation.
What you and I have not heard in the mainstream media is the naked emperor kind of truth: several of the largest classes of drugs have no real benefits and should never have been approved by the FDA. Many have side-effects that are so serious that the resulting hospitalizations and deaths cost more than the drugs. The sale of prescription medicines in the United States exceeded $73 billion in 1994. Two researchers from the University of Arizona conclude that preventable illness and death from the misuse of these medicines cost the American economy over $75 billion that year; if lost productivity is included the cost rises to between $138 billion and $182 billion annually. (http://www.drugintel.com/pharma/adverse_drug_reactions.htm#Cost) Even the American Medical Association admits to about 100,000 deaths per year due to the correct use of prescription drugs. Others think that the number is double that. Because most of these drug deaths occur in isolated individuals, mostly in hospitals, and can be concealed as other causes of death, they do not get the press coverage of 200 dead in a jetliner crash. Would we have an airline industry if it caused 100,000 deaths per year?
Thomas J. Moore, Senior Fellow in Health Policy at the George Washington University Medical Center, right there in Washington, in his 1998 book “Prescription for Disaster”, examined the top 50 selling drugs. He found that 7 can cause addiction, 18 have cancer risks, 18 are unusually toxic, and 25 have cardiac risks. The most valuable pair are insulin and thyroid hormone, for which Medicare should pay, in my opinion.
The FDA and others are easily fooled by the effect of drugs on conveniently measured symptoms, such as bone density, cholesterol levels, heart arrhythmias and blood pressure. The section below is more technical and gives evidence for the uselessness of two major classes of drugs: for blood pressure and for cholesterol. Medicare should not pay for drugs that do not lengthen life or not do so at a cost the American taxpayer can bear.
The reduction in the risk of dying in 1 year is 0.15% for people taking the anticholesterol drug Lipitor in the ASCOT trial described below. This reduction is not statistically significant. This means that your supposedly improved chance not to die in 1 year would be just 1 person of 667 taking Lipitor. The cost of a month’s supply of Lipitor (atorvastatin) at 40 mg/day is $125, or $1500/year, or $1,000,000 to prevent one death among 667 people taking this type of drug for just 1 year. Again, this was based on a trial the drug’s maker chose to report, usually the trial with the best results.
Congress should direct the U. S. Dept. of Health and Human Services to appoint a committee of financially independent, Big Pharma independent members, including ones from Canada and the UK, to review the literature and advise Medicare on which classes of drugs do not prolong life, or do not do so at a reasonable cost. Obviously the FDA is not capable of doing so, and not likely to reverse its own approval decisions in any serious manner. The 2003 law should be amended to forbid payment for such drugs. Then the cost of the Medicare Drug Benefit plan will be very reasonable and bearable, and federal deficits will be a lot lower.
Sincerely,
Joel M. Kauffman, PhD, Professor of Chemistry Emeritus,
University of the Sciences in Philadelphia
*****
Evidence from the medical literature on the non-benefits of common drugs follows. While there are many problems with this source of information, it is all we have where even some of it comes close to accuracy that is easily available to the public. The following specific examples are mostly from my forthcoming paper: “Bias in Recent Papers on Diets and Drugs in Peer-Reviewed Medical Journals”, Journal of American Physicians and Surgeons 2004;9(1): in press.
Blood Pressure
The Anti-Hypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT, 2002), in its anti-hypertensive arm, sought to compare a diuretic (chlorthalidone) with a calcium channel blocker (amlodipine) and an angiotensin converting enzyme (ACE) inhibitor (lisinopril) for the control of blood pressure (the surrogate endpoint) in high-risk patients. No placebos were used in this study that followed 33,357 subjects for a mean of 4.9 years. Primary outcomes were defined as fatal CHD or non-fatal MI, and were observed in 2,956 subjects. There was no difference in frequency found in either outcome among the three treatments. There were minor differences in stroke rates (1% absolute, lisinopril best) and in hospitalizations for “heart failure” (2% absolute, amlodipine best). Abstract context: “Anti-hypertensive therapy is well established to reduce hypertension-related morbidity and mortality, but the optimal first-step therapy is unknown.” Conclusion: “Thiazide-type diuretics are superior in preventing one or more major forms of CVD and are less expensive. They should be preferred for first-step anti-hypertensive therapy.” It should be noted that Chlorthalidone is not a thiazide. It had the greatest effect on systolic blood pressure, yet did not have the most favorable effect on any outcome (their Fig. 4). The World Health Organization does not think older adults should use this drug because the risk of serious side effects is so high (Wolfe & Sasich, 1999).
An older trial, also supported by drug companies, on 17,354 subjects with worse hypertension than in ALLHAT and with six years of follow-up, compared placebo, a beta blocker (propranolol) and a thiazide diuretic (bendrofluazide). There was no difference in the all-cause death rate and minor reductions in the rates of strokes and all cardiac events in the treatment groups (MRC, 1985). It was interesting that placebo reduced blood pressure significantly, but there was no correlation of the amount of reduction with death rates, another indication of a useless surrogate end-point.
Another trial examined 484 randomly selected hypertensive men, except that they were all aged 68 at the beginning. Many of them were taking a wide spectrum of anti-hypertensive drugs, mainly from the thiazide diuretic and beta-blocker classes. They were followed for 10 years and had a cardiac-event-free survival of 65%, while those men not on medication had an 82% survival. Those with diastolic pressure at baseline *90mm Hg had a RR = 4 with treatment, and even those with >90mm Hg fared worse with treatment. Funding was mostly from non-drug company foundations and a Swedish government agency (Merlo et al., 1986).
The older trials show that it is not certain that anti-hypertensive drugs lower morbidity or mortality. It is obvious that the ALLHAT conclusions should have been: No standard treatment with prescription anti-hypertensive agents is worthwhile. So it follows that Medicare should not be paying for them.
Cholesterol
The Myocardial Ischemia Reduction with Aggressive Cholesterol Lowering (MIRACL) study, an RCT on the effects of 80 mg/day of atorvastatin (Lipitor™) or placebo on 3,086 patients in hospital after angina or non-fatal MI and followed for 16 weeks, had the following conclusions in the abstract: “For patients with acute coronary syndrome, lipid-lowering therapy with atorvastatin, 80mg/day, reduces recurrent ischemic events in the first 16 weeks, mostly recurrent symptomatic ischemia requiring re-hospitalization.” (Schwartz et al, 2001) Actually this is true. The unmentionable findings were that there was no change in the death rate, and no significant change in either the re-infarction rate or need for resuscitation from cardiac arrest. There was a significant drop in chest pain requiring rehospitalization. The risk ratio plot was unusual in not having a vertical bar at the 1.00 point, making the outcomes hard to visualize from this figure. The discussion did not give any comparisons with alternate treatments, for example, that five weeks of aspirin would give significantly lower reinfarction and all-cause mortality rates in men (Kauffman, 2000). Lowering cholesterol levels was highlighted, despite existing knowledge that the beneficial effects of statins on CHD are independent of either the baseline or achieved levels; thus these levels were a useless surrogate endpoint (Nielsen, 2001).
The Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT) study, specifically on atorvastatin (Sever et al, 2003), did not mention in its abstract that: (1) women were worse off with treatment, the same as with aspirin (Kauffman, 2002); (2) that after 3.5 years there was no significant change in the all-cause death rate, marking atorvastatin as even less effective than pravastatin in the WOSCOPS trial; and (3) that total cardiovascular events and procedures were 2% lower with atorvastatin than with placebo after 3.5 years — instead they only gave the RR of 0.79. This is a poorer performance than that of Bufferin, for which RR = 0.31 for non-fatal MI in men (Kauffman, 2000). The control group had more previous stroke, TIA, diabetes, and other CHD, showing poor randomization.
A meta-analysis of 44 trials of atorvastatin intended to highlight its safety (Newman et al, 2003) neglected to mention in its abstract that: (1) the median treatment period was only one year; (2) that the all-cause death rate of 1% did not differ from that of placebo, and thus was not even as good as that of pravastatin in the WOSCOPS trial; (3) that 65% of the treatment group vs. 45% of controls experienced an adverse event; (4) that the total unadjusted withdrawal rate was 4% vs. 1% for placebo for all adverse events; (5) that 10% of patients suffered serious adverse events vs 8% for placebo; and (6) that reduction in treatment-associated adverse cardiovascular events (a judgmental determination) was 1% absolute (from 2% on placebo to 1% on atorvastatin, exactly as in the ASCOT trial, a poorer performance than that of Bufferin™ in men). The authors were thoughtful enough to provide the information, in the Methods section, that from April 1, 1998, the FDA allowed the exclusion of cancer and overdose from drug side-effects, both of which might have been significant with this drug based on the results of the Prospective Study of Pravastatin in the Elderly at Risk (PROSPER, Goldstein, 2003) and Cholesterol and Recurrent Events (CARE, Ravnskov, 2000) trials.
A more recent study from the LDS Hospital and University of Utah, Salt Lake City, confirmed that statin use improved the survival rate among 651 patients, 75% male, with *70% blockage in at least 1 coronary artery when infection by cytomegalovirus was present, and more so, when inflammation was severe. There was no survival benefit when both were absent. This finding strongly supports the hypothesis that cholesterol lowering was irrelevant (Horne et al., 2003). Even here, aspirin might do as well. It also means that the legitimate market for statin drugs could be about 1% of the current market. There is no reason for Medicare to pay for the other 99%.
References
ALLHAT. Major outcomes in high-risk hypertensive patients randomized to angiotensin- converting enzyme inhibitor or calcium channel blocker vs. diuretic. JAMA2002;288:2981-2997.
Goldstein MR. Letter to editor. Lancet 2003;361:427-428.
Horne BD, Muhlstein JB, Carlquist JF, et al. (2003). Statin Therapy Interacts With Cytomegalovirus Seropositivity and High C-Reactive Protein in Reducing Mortality Among Patients With Angiographically Significant Cornary Disease. Circulation, 107:1-6.
Kauffman JM. Should you take aspirin to prevent heart attack? J Sci Exploration 2000;14(4):623-641.
Kauffman JM. Aspirin study flawed. Letter to editor, J Sci Exploration 2002;16(2), 247-249.
Merlo J, Ranstam J, Liedholm H, et al. Incidence of myocardial infarction in elderly men being treated with anti-hypertensive drugs: population-based cohort study. BMJ1996;313:457- 461.
MRC. MRC trial of treatment of mild hypertension: principal results. BMJ 1985;291:97-104.
Newman CB, Palmer G, Silbershatz H, Szarek M. Safety of atorvastatin derived from analysis of 44 completed trials in 9,416 patients. Am J Cardiol 2003;92:670-6.
Nielsen JV. Serum lipid lowering and risk reduction: where is the connection? BMJ Rapid Response 2001;Nov.19, to Kmietowicz Z. Statins are the new aspirin, Oxford researchers say. BMJ 2001;323:1145.
Ravnskov, U. The Cholesterol Myths: Exposing the Fallacy that Saturated Fat and Cholesterol Cause Heart Disease. Washington, D.C: New Trends Publishing 2000.
Schwartz GG, Olsson AG, Ezekowitz MD, et al. Effects of atorvastatin on early recurrent ischemic effects in acute coronary syndromes. JAMA 2001;285:1711-1718.
Sever PS, Dahl FB, Poulter HW et al. for ASCOT. Prevention of coronary and stroke events with atorvastatin in hypertensive patients who have average or lower-than-average cholesterol concentrations, in the Anglo-Scandinavian Cardiac Outcomes Trials’ Lipid Lowering Arm (ASCOT-LLA): a multicentre randomized clinical trial. Lancet 2003;361:1149-1158.
Wolfe S, Sasich LD. Worst Pills Best Pills. New York, N.Y.: Pocket Books, 1999, p105.