Custom-designing drug doses to fit the genes.

Custom-Designing Drug Doses to Fit the Genes

Recently, 108 children suffering from leukemia in a hospital in Tennessee were given an anti-cancer drug called methotrexate. In an effort to pinpoint the optimum therapeutic dose of the drug, as well as to cure the cancer, a team of physician-researchers at the hospital gave each child infusions of 1,000 milligrams of methotrexate per square meter of body surface, in a complex treatment regimen. They found, as reported in the Feb. 20 NEW ENGLAND JOURNAL OF MEDICINE, that though the doses were standardized, the drug concentration in each child's bloodstream varied as much as three-fold. And chances of survival turned out to depend in large part on the child having a high level of circulating drug.

Doctors use drugs as long-distance tools against diseases hidden inside bone and blood; medicine is a "practice," an inexact art. Perhaps nothing says more about the complexity of prescribing than that the researchers in this case aren't sure why there were such variations in drug level.

This isn't the first time doctors have been faced with the problem of unresponsiveness to a drug dose that is usually effective, or with the related problem of a catastrophic reaction to a generally well-tolerated drug. There are a host of elements that determine a patient's response to a drug, according to William Evans of the University of Tennessee in Memphis, who led the methotrexate study. Age, sex, diet, kidney efficiency and fat distribution -- and the interactions of these and many other factors -- can all affect drug metabolism. Increasingly, researchers like those in the methotrexate group are looking to the role of genes in drug metabolism to fill in some of the unknowns of prescribing.

"Many people take drugs on a certain prescribed basis, maybe two or three times a day," says pharmacologist Robert Smith of St. Mary's Hospital School of Medicine in London, England. "The assumption underneath it all is they're all going to handle the drug equally, and in many cases they don't." Pharmacogenetics pioneer Elliott Vesell, of Pennsylvania State College of Medicine in Hershey, adds, "Physicians are becoming more attuned to the necessity for individualizing drug doses. That's part of the art of being a good physician. . . . [Otherwise] they'll kill their patients."

Though the study of pharmacogenetics got its start in the '50s and '60s, a discovery by Smith about a decade ago dramatized the need to reckon with genes when prescribing even common drugs. Smith's investigation was prompted by his own collapse after taking a normal dose of a blood pressure medication he was studying.

"My blood pressure fell to 70 over 50. My colleagues were really concerned; they thought I was on the way out," Smith says. "Then we came round to analyzing what had happened, and we found that I was not metabolizing the drug."

Nearly untouched by the enzyme that normally degrades it into harmless parts, the drug, called debrisoquine, had reached levels in Smith's blood that "became very unpleasant, very quickly." The St. Mary's group tested the responses to small doses of debrisoquine in a group of 94 medical students and found two more "poor metabolizers," as Smith dubbed them. Says Smith, "When we tested the three families [his and the two students'], we knew we had a genetically determined deficiency."

What was going on in these debrisoquine hyperresponders? Studies by the St. Mary's group and others located the problem in a liver enzyme, one of a system of enzymes called the cytochrome p450s. According to Vesell, whose research on twins provided early evidence of genetic control of drug metabolism, there is still disagreement over whether the poor metabolizers have a less efficient form of the enzyme or lack it altogether, but it is clear that the enzyme's ability to oxidize chemicals is impaired in these people. And because oxidation is "the most common pathway of metabolism," Smith says, such a defect puts poor metabolizers at risk of adverse reactions to a number of drugs. A group at Sweden's Karolinska Institute, for instance, reported in the Feb. 6, 1982 LANCET that poor metabolizers show exaggerated responses to several betablockers that are used to decrease heart rate.

Poor metabolizers may experience exaggerated responses to drugs left in a potent, unmetabolized form, or therapeutic failures with drugs that must be metabolized to become active. Often, these are effects that physicians can easily avoid, says Smith. In the case of a drug like debrisoquine, for instance, a precipitous drop in blood pressure signals the physician that the dose had better be adjusted.

But when drugs are not so easily "read" they can kill, as they did in England in the mid-1970s when doctors began prescribing a drug called perhexiline for chronic chest pain. In that case, there were no readily measurable signs to indicate whether the drug was behaving normally. It caused severe nerve damage in about 400 patients, and fatal liver damage in a few, before scientists realized what was going on. "There were no real overt signs of what was happening," Smith says, "until these people had accumulated 50 or 60 grams of drug in their tissues and the weakest links started to break."

The perhexiline episode pointed out a major weakness in drug development, according to Werner Kalow, one of the pioneers in pharmacogenetics, at the University of Toronto. Though drugs are tested in animals before they become available for clinical use, experimental animals are usually highly inbred to avoid the very kind of genetic variations that caused problems in humans taking the drug. Scientists at that time "didn't think of this kind of variation," Kalow says. "This is the main problem that we have. [And] drugs have been tested on too few people to find" genetic variations that occur infrequently.

Population studies done by the St. Mary's group showed that the debrisoquine defect is a recessive trait, caused by a single gene -- only people who have inherited the gene from both parents will show its effects. Even so, millions of people in the world have this genetic impairment in their ability to oxidize chemicals. According to Kalow, the defect appears in about 10 percent of the Caucasian populations of the United Kingdom and Canada (and probably the United States as well), about 3 percent of Oriental populations and about 1 percent of Semitic populations.

There are other types of problem metabolizers as well. Even before the debrisoquine discovery, scientists at the University of Cincinnati reported that another metabolic pathway, acetylation, showed genetically controlled variability. In this pathway, an acetyl group is added to a foreign chemical as part of the process of degradation. Research continues at Cornell University on the possible connection between slow acetylation and drug-induced or spontaneous systemic lupus erythematosus, a multisystem inflammatory disorder of uncertain origin. Researchers have found many other enzymatic variants, and, Vesell says, "we think there's much more genetic variation than we've discovered."

Though these genetic quirks in metabolism are clearly a potential problem for large numbers of people, when it comes to prescribing drugs some doctors may not be fully aware of the dangers, according to Smith. "There's still quite a long way to go in terms of education in these things. You have to bear in mind that the average practicing physician was trained 15, 20 years ago, and all these [discoveries] have happened in the last decade."

The problem metabolizer carries no distinguishing marks as he or she walks through a physician's door. If a needed drug isn't widely known to be subject to variable metabolism, and if it accumulates without immediate and noticeable effect, complications may go unrecognized; or they may be ascribed to the disease, instead of to patient-drug interactions. Even the physician primed to recognize the adverse reactions of a problem metabolizer may not get the chance, since patients don't always come back after receiving a prescription.

The situation is exacerbated by the narrow therapeutic window of many of the newest, most powerful drugs. With some, like the anticonvulsant phenytoin, the effect may go from therapeutic to toxic with an increase of just 10 micrograms per milliliter blood concentration. In a poor metabolizer, a drug stays potent so long that it is as though a higher dose has been given; with these patients, a doctor may need to reduce the dose to one-tenth to get into the therapeutic window.

There are some new technologies that can help. Rapid, though expensive, blood tests are available that can tell the doctor how much of the drug has been metabolized. A growing number of hospitals and physicians use computer programs to flag drugs prone to variable metabolism, and help adjust doses for a problem metabolizer. These "pharmacokinetic" programs can also take into account other factors that may increase or decrease metabolic difficulties. For instance, cigarette smoking, according to pharmacologist Daniel Robinson of the University of Florida in Gainesville, who designed a pharmacokinetic computer program, "induces [patients'] enzymes in the liver to metabolize the hydrocarbons in the smoke, and just as a by-product of that they happen to metabolize certain drugs faster."

But the tool most widely used by physicians is the package insert, instructions provided by the drug company that note contraindications or other possible problems. The pharmaceutical industry has adapted rapidly to the last decade's discoveries in the field. Early in research, companies now investigate the pathways by which new drug candidates are metabolized, to see if the drug is prone to variable metabolism.

"The industry is interested to try to identify whether [genetic variability] affects their drug, so they can give the physician more guidance about attentiveness to prescribing," says Smith. "That's really the bottom-line message with this -- we now know one of the discrete reasons why you need to individualize dosage."

Good physicians, of course, are already sensitive to the possibility of idiosyncratic responses with each patient. "The astute physician has always individualized dose; the astute physician has always looked at the patient to see the effect of the medication," Vesell says. "Each time, it's a guesstimate of the right dose.

"I think we have to recognize that giving drugs is nowhere near as precise as we would like it to be. In the future, it may be that everyone will be typed [for metabolic characteristics], when we get down to the DNA for these genetic defects, and will carry a card like they do for the blood groups. But until then physicians have to be very, very careful to get into the therapeutic window--out of the toxic and into the therapeutic range."