In early 1985 the National Heart, Lung and Blood Institute halted a study of two heart-attack therapies that it had begun less than six months before. In that short time, NHLBI's researchers found that one drug, tissue plasminogen activator (TPA), was so superior in dissolving coronary blood clots that they could no longer ethically administer the other drug.

Few Americans realize it but, thanks to the Food and Drug Administration, the entire country has now been inducted back into that study. Worse yet, of those stricken by heart attacks who make it to a hospital, practically no one will receive the drug of choice.

What brought this all about was the May 29 decision of an FDA advisory committee recommending against TPA but favoring approval of the other drug: intravenously administered streptokinase. In an Italian study of over 10,000 heart-attack patients, streptokinase had reduced in-hospital mortality by 18 percent. For patients treated within the first hour of symptoms, the reduction was an astounding 50 percent.

Good as streptokinase was, the NHLBI found TPA to be even better -- nearly twice as effective in dissolving the coronary blood clots that lead to heart attacks. This, however, was not enough for the advisory panel. The NHLBI researchers had assumed that coronary thrombolysis (clot dissolution) can save lives, based largely on the extensive streptokinase data. So they did not attempt to assemble the several thousand patients necessary for meaningful mortality data on TPA; instead, they focused on determining which drug unblocked arteries faster.

But the advisory committee rejected this approach. It is possible, the panel argued, that streptokinase saves lives not by thrombolysis but by some unknown means. Its clot-dissolving action might thus be an insignificant side effect that indicates nothing about whether other thrombolytic agents, such as TPA, will also reduce mortality. To this day, TPA is unavailable except in clinical tests aimed at producing more data.

To countless physicians who believe that rapid clot dissolution can spell the difference between life and death for heart-attack victims, this is like banning an improved life jacket because its manufacturer hasn't done a controlled study of drowning.

Were this simply a scientific dispute, as it should be, doctors could choose their theories and patients could choose their doctors. FDA's approval power over new drugs, however, creates an ethical dilemma for those physicians who are convinced of TPA's efficacy. The World Medical Association's Helsinki Declaration on Human Research states that a "doctor must be free to use a new . . . therapeutic measure, if in his or her judgment it offers hope of saving life." As for subordinating patient welfare to the gathering of scientific data, it states that "concern for the interests of the subject must always prevail over the interest of science and society." Emergency rooms and coronary care units may not be hotbeds of political activism, but some well-considered civil disobedience by the medical profession could be therapeutic in more ways than one.

More important is how the TPA issue demonstrates the insidiousness of "drug lag" -- the fact that, as a result of FDA regulation, new drugs are frequently available abroad long before their introduction in the United States. The FDA will nearly always err on the side of delay rather than risk releasing a drug that later proves to be bad. This is because the consequences of the latter are highly visible -- witness thalidomide. But the results of delaying a good drug, while just as real, are practically invisible. When did you last see a drug-lag victim's grieving family on the nightly news? How many lives are being lost daily if TPA turns out to be as effective as it appears to be? We can get some idea by applying the Italian streptokinase data to U.S. statistics. Of the 700,000 Americans hospitalized annually with acute myocardial infarction, about 9 percent die in-hospital. If streptokinase produces the same 18 percent mortality reduction here as in the Italian study, it would prevent over 11,000 deaths yearly.

TPA reopened arteries twice as fast as streptokinase in the NHLBI trial. If this leads to a proportionate reduction in mortality, then TPA would save 11,000 more lives annually than streptokinase, or 30 lives a day. (This is just for starters. As thrombolytic therapy becomes widespread, it will be used outside hospitals by paramedics and others, reaching many of the over 400,000 heart-attack victims who each year die before reaching a hospital.)

This figure may be lower if TPA causes more complications than streptokinase. While the NHLBI trial showed no real difference in this respect between the two drugs (with TPA at 100 milligram dosage), earlier experience with higher TPA dosages indicated a significant risk of intracranial bleeding. If one assumes, very pessimistically, that intracranial bleeding from lower-dose TPA occurs in 0.5 percent of cases, that it is always fatal and that it is totally in excess of any streptokinase-induced bleeding, then TPA would still save 20 more lives a day.

This is too high a price for regulatory delay on one drug; worse yet, it is a price that will probably never be publicly disclosed. Charges that an approved drug has caused 20 or so deaths over a period of years will send reporters and congressional oversight panels into high gear. When the culprit is not drugs but drug lag, however, the level of investigative energy seems to vanish into a conceptual black hole.

In June, eight former agency officials declared that FDA Commissioner Frank Young "needs a little time out of the glare of publicity to review" the TPA issue. He is evidently taking that time, but I hope he realizes its cost. By my calculation, he's at 3,160 lives . . . and counting. Sam Kazman heads the Free Market Legal Program of the Competitive Enterprise Institute, a nonprofit policy organization in Washington. Second Opinion is a forum for points of view on health policy issues.