Outrage: New Study: Wait Seven Years to Use New Drugs

A study published in the May 1,  2002 Journal of the American Medical Association (JAMA) has resulted in a major change in the Health Research GroupÕs drug safety policy. The study, in association with physicians from the Harvard Medical School, examined the frequency and timing of the identification of new adverse drug reactions resulting in the addition of a black box warning in the drugÕs professional product labeling or its outright removal from the market.

Three of the authors have close identification with the Health Research GroupÑits director, Sidney M. Wolfe, and former HRG staffers Drs. Steffi Woolhandler and David Himmelstein. The other co-authors are affiliated with the Harvard Medical School.

As a result of this study, we have decided to modify our long-standing recommendation that you wait at least five years after a new drug is approved to use it, unless (and this is a rare exception) the new product is a drug that offers a documented therapeutic advantage over older drugs that are already available.

For safetyÕs sake, we now recommend at least a seven year precautionary wait, since, by seven years after approval about half of the black box warnings or withdrawals will have occurred.

The study analyzed the appearance of new black box warnings in the PhysiciansÕ Desk Reference from 1975 through 1999. Also, the frequency and timing of drug withdrawals over this period was calculated.

A black box warning is the strongest type of warning that the Food and Drug Administration (FDA) can require. It is generally reserved for adverse reactions that may lead to death or serious injury. Black box warnings ordinarily are based on events occurring in humans but may be required when serious animal toxicity has been seen and no information is available in humans.

Between 1975 and 1999 the FDA approved 548 new drugs never before marketed in the U.S. A new black box warning was required for 56 (10.2 percent) of these drugs. Forty-five (8.2 percent) needed more than one such warning. And during this time 16 drugs (2.9 percent) were withdrawn from the market.

The estimated probability of a new drug requiring a new black box warning or being withdrawn from the market between 1975 and 1999 was calculated to be 20 percent. In other words, the chance was one in five that a newly approved drug during this period would turn out to have a safety problem serious enough to require one of these interventions. One-half of the withdrawals occurred within two years following the drugÕs approval and one-half of the black box warnings and withdrawals happened within seven years after a drugÕs introduction.

The most common reasons for black box warnings were cardiovascular toxicity 17 times, liver toxicity 15 times, blood (bone marrow) toxicity 13 times, and risk in pregnancy 9 times. Other adverse reactions accounted for the other 36 percent of black box actions.

An editorial in the same issue of JAMA, written by physician officials from the FDAÕs Center for Drug Evaluation and Research (CDER), accompanied the study. CDER is the part of the agency responsible for reviewing new drug applications. The editorial attempted to assure readers that the safety of new drugs is getting better: ÒRecent changes in drug development should help protect against some of the most important past causes of drug withdrawal,Ó the editorial said, adding that Ò...there is reason to believe that some of the more common causes of significant toxicity will be less likely in the future.Ó

A look at the most recent history of black box warnings and drug withdrawals tells a story much different than the FDAÕs editorial message, however. Since the cut-off date for drugs included in the JAMA study an additional three new drugs have come off the market for safety reasons. These are the cholesterol-lowering drug cerivastatin (BAYCOL), rapacuronium (RAPLON) a muscle relaxant used in surgery, and alosetron (LOTRONEX) a drug that was used for treating the symptoms of irritable bowel syndrome in women. In the cases of rapacuronium and alosetron the life-threatening adverse reactions that led to their removal were seen in clinical trials conducted before these drugs were approved. CerivastatinÕs professional product label warning about a life-threatening muscle damage was strengthened four months before it was withdrawn from the market for this reason. It should have come off the market at the first sign of a problem because five other cholesterol lowering drugs in the same family already existed and are safer.

The FDA editorial specifically addressed two of the most common types of toxicity that can result in black box warnings or drug withdrawals: QT prolongation (a change in the heartÕs electrical conduction that can lead to fatal heart rhythm disturbances, and liver toxicity.

Unfortunately, drugs that were known to cause QT prolongation are continuing to be approved without black box warnings. The new antipsychotic drug ziprasidone (GEODON) was approved in February 2001. This drug was known to cause QT prolongation and was less effective than a much older antipsychotic drug. We listed ziprasidone as a Do Not Use drug in the June 2001 issue of Worst Pills, Best Pills News. In February 2002, the drugÕs existing warning about QT prolongation was strengthened, but a black box was not required.

The three newer fluoroquinolone antibioticsÑgatifloxacin (TEQUIN), moxifloxacin (AVELOX), and sparfloxacin (ZAGAM)Ñwere all known to cause QT prolongation before approval. All have warnings about this condition, but none have a black box.

We listed moxifloxacin and sparfloxacin as Do Not Use drugs in the February 2000 Worst Pills, Best Pills News. There are safer, equally or

more effective antibiotics available for the same uses as these three fluoroquinolones.

Arsenic trioxide (TRISENOX), an injectable drug used to treat a type of leukemia, was approved in September 2000 with a black box warning about QT prolongation.

The following drugs have received black box warnings about QT prolongation since the 1999 cut off date for the JAMA study: droperidol (INAPSINE); itraconazole (SPORANOX); levomethadyl (ORLAAM); mesoridazine (SERENTIL); thioridazine (MELLARIL)

If the FDA has a policy about QT prolongation warnings, it is dangerously inconsistent.

An example of a new drug with known liver toxicity before its approval is the arthritis drug leflunomide (ARAVA) that was approved in September 1998. The drug also appears to be less effective than other drugs for the treatment of rheumatoid arthritis.

Two liver toxic drugs that were removed from the market in other countries remain available in the U.S. These are the fluoroquinolone antibiotic trovafloxacin (TROVAN) and tolcapone (TASMAR) a drug used in ParkinsonÕs disease. There was simply no reason to ever approve trovafloxacin, since this drugÕs liver toxicity was known before its approval. The current warning is buried in the labeling rather than being inside a highly visible black box.

Whatever the FDAÕs policy is on liver toxic drugs, it appears to be as muddled as its QT prolongation policy.

The FDA editorial appears incoherent and far from reassuring about improvements in the safety of newly approved drugs. It implies that Òrecent changesÓ in drug development coupled with black box and other warnings in the labeling will reduce the likelihood of significant toxicity. This is plainly not true, because research has shown that the addition of warnings, even the black box kind, usually does not reduce the unsafe prescribing of drugs. Some of this research has been published by physicians and pharmacists from the FDA. In the October 2001 issue of Worst Pills, Best Pills News we reviewed the ineffectiveness of labeling changes to alter prescribing habits.

What You Can Do

Follow the Health Research GroupÕs new seven-year rule.

You should wait at least seven years from the date of release to take any new drug unless it is one of those rare Òtherapeutic advancesÓ that offers you a documented therapeutic advantage over older, proven drugs. New drugs are tested in a relatively small number of people before being released, and serious adverse effects or life threatening drug interactions may not be detected until the new drug has been taken by hundreds of thousands of people. A number of new drugs have been withdrawn within their first seven years after release. Also, black box warnings about serious new adverse reaction have been added to the labeling of a number of drugs, or new drug interactions have been detected, often within the first seven years after a drugÕs release.