Fraser Forum

Using our Heads on Head-to-Head Trials

R

by John R. Graham

emember the Pepsi™ Taste Test1? Launched in 1975, the test used a host, who stood in a public place with unmarked glasses of Pepsi and its competitor, and who invited passersby to taste both of them. Lo and behold, they always preferred Pepsi (at least on TV). Suppose that, instead, the host stood on one street corner offering Pepsi and coloured water, while his competitor conducted the same test with her own brand of cola and coloured water on another corner. We would find it difficult to know which cola was superior, because the two colas would not have been tested against each other. The Pepsi Taste Test is similar to a head-to-head trial in medical research: two supposedly similar products are used on similar people, with the goal of learning the different effects of the two products.2 Most clinical trials look like the second taste test scenario, where a medicine is tested against a placebo (dummy). This makes it difficult to judge how one drug differs from another.

The value of head-to-head trials was recently brought into focus by the widely reported results of the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT), which tested medicines used to reduce blood pressure. ALLHAT tested an inexpensive diuretic, or “water pill,” versus an angiotensin-covertingenzyme (ACE) inhibitor and a dihydropyridine calcium-channelblocker (DHP-CCB), both of which are much more expensive than water pills. As a result of the trial’s outcomes, its investigators recommended that water pills be prescribed first, rather than the more expensive drugs. They estimated that the US health care system would have saved US$3.1 billion dollars between 1982 and 1992, had this recommendation been followed (ALLHAT Investigators, 2002, p. 2994). The New York Times used ALLHAT to blame out-of-control marketing by drug companies for the fact that doctors were prescribing more ACE inhibitors and DHP-CCBs instead of diuretics, and proposed that the government should either force manufacturers of new drugs to conduct head-to-head trials, or conduct them itself (2002).

John R. Graham ([email protected]) is Director of Pharmaceutical Policy Research at The Fraser Institute. He earned his BA (Hons) in Economics and Commerce at the Royal Military College of Canada and his MBA at the London Business School, University of London.

6 |

How soon they forget. Just two years previously, the same newspaper reported on the hot competition for market share between two statins, which are used for cholesterol: Bristol-Myers Squibb’s Pravachol® (pravastatin) and Pfizer’s Lipitor® (atorvastatin). In response to Lipitor’s success, Bristol-Myers Squibb committed tens of millions of dollars to a head-to-head trial of the two drugs, named PROVE IT. In the huge statin market, companies are investing hundreds of millions of dollars in clinical trials to demonstrate the unique qualities of each of their drugs (Stolberg and Gerth, 2000). However, generally requiring headto-head trials before allowing drugs to be sold is irresponsible and absurd. The latest estimate for the average cost of researching and developing a new prescription drug is US$802 million, spread over 10 to 15 years (Tufts CSDD, 2001). The ALLHAT trial cost US$125 million (Hensley, 2003) and studied only one DHP-CCB and one ACE inhibitor. Commenting on ALLHAT, one cardiologist stated that all ACE inhibitors and DHP-CCBs are the same (Appel, 2002, p. 3,040). However, at least one other trial has shown different outcomes for different ACE inhibitors, and (in British Columbia, at least) physicians have changed their prescribing practices within each class as new ACE inhibitors and DHP-CCBs were introduced (Bourgeault et al., 1999; Graham, 2002, p. 26-28). British Columbia’s provincial Pharmacare lists 8 ACE inhibitors and 5 DHP-CCBs on its list of drugs reimbursed. Given the costs of ALLHAT, head-to-head trials for each of them and one diuretic would cost over $800 million: doubling the already high cost of clinical trials! Head-to-head trials are more expensive than placebo trials because they require

Pharmaceutical Policy

many more subjects. This gives the trial more statistical power, which is required to observe the differences when the two drugs have similar effects. According to US Food and Drug Administration regulations, Phase III trials (the final stage before approval) must have at least 1,000 to 3,000 people. ALLHAT had over 33,000 people! Because of the large numbers of enrollees, it took a long time to recruit them and run the trial. ALLHAT started recruiting in February 1994 and closed on March 31, 2002. A blanket regulation that adds another 8 years to the time required to approve a drug for sale is hardly good policy. However, there must be cases that warrant head-to-head trials when placebo trials indicate that more information would be valuable. If they are too expensive for private companies to undertake, government sponsorship is an option. However, it is unlikely that government agencies face the right incentives to determine when a headto-head trial is warranted. Indeed, the National Institutes of Health, one of whose institutes was the primary sponsor of ALLHAT, had previously been a loud cheerleader for ACE inhibitors (which appear to have fared poorly in ALLHAT), noting proudly that NIH researchers had discovered the angiotensin converting enzyme, and that the corporate researchers who developed the drugs had cited much NIH research (NIH, 2000). As the statin case described above shows, we can expect competitive pressures to motivate manufacturers to sponsor head-to-head trials in some circumstances. Furthermore, the private sector has the resources to do so: in 2001, the US pharmaceutical industry invested US$30 billion in R&D, whereas the National Institutes of Health invested US$20 billion (PhRMA, 2002, p. 14).

As well, Pfizer, manufacturer of Norvasc® (amlodipine), the CCB tested in the trial, voluntarily donated one third of ALLHAT’s $125 million cost (Hensley, 2003). The ALLHAT investigators concluded that previous fears that Norvasc might increase the risk of cancer, gastrointestinal bleeding, and other problems, were unfounded (ALLHAT Investigators, 2002, p. 2989). We should expect Pfizer to trumpet these findings. Furthermore, we should expect ALLHAT to result in a number of further head-to-head trials, as manufacturers of competing ACE inhibitors

A ... regulation that adds another 8 years to the time required to approve a drug for sale is hardly good policy.

References ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group (2002). “Major Outcomes in High-Risk Hypertensive Patients Randomized to Angiotensin-Converting Enzyme Inhibitor or Calcium Channel Blocker vs Diuretic: The Antihypertensive and Lipid-Lowering Treatment to Prevent Health Attack Trial (ALLHAT).” Journal of the American Medical Association 288, 23 (December 18): 2981-2997. Appel, Lawrence J. (2002). “The Verdict from ALLHAT—Thiazide Diuretics Are the Preferred Initial Therapy for Hypertension.” Journal of the American Medical Association 288, 23 (December 18): 3039-3041. Bourgault, Chantal et al. (1999). “Reference-based Pricing of Prescription Drugs: Exploring the Equivalence of Angiotensin-converting-enzyme Inhibitors.” Canadian Medical Association Journal 168, 3 (August 10): 255-260. Graham, John R. (2002). The Fantasy of Reference Pricing and the Promise of Choice in BC’s Pharmacare. Public Policy Source No. 66. Vancouver, BC: The Fraser Institute. Hensley, Scott (2003). “More Drug-Comparison Studies Could Aid Patients’ Pocketbooks.” Wall Street Journal Online (January 8). New York Times (2002). “When Cheaper Is Also Better.” Editorial. New York Times (December 19): A38.

and CCBs respond to the challenge of ALLHAT, and the questions of millions of patients who suffer from hypertension. Do marketing concerns shape pharmaceutical research? Let’s hope so.

NIH (2000). NIH Contributions to Pharmacuetical Development. Bethesda, MD: National Institutes of Health (February).

Notes

Stolberg, Sheryl Gay and Jeff Gerth (2000). “Drug Makers Design Studies With Eye to Competitive Edge.” New York Times (December 23): A1.

1

Pepsi™, Pravachol®, Lipitor®, and Norvasc® are registered trademarks and are the property of their respective owners. 2

The tests are not quite the same as each other because subjects in a clinical trial do not take both medicines. Rather, they are divided into two (hopefully) homogenous groups.

PhRMA (2002). PhRMA Industry Profile 2002. Washington, DC: Pharmacuetical Research and Manufacturers Association.

Tufts CSDD (2001). “Tufts Center for the Study of Drug Development Pegs Cost of a New Prescription Medicine at $802 Million.” Press Release (November 30). Boston, MA: Tufts Center for the Study of Drug Development. &

February

2003

| 7