What is ‘personalized medicine?’
I’ve become a bit skeptical at the apparent creation of new fields of human endeavor that start narrowly and then spread by force of their own personality. Medicine remains an art to be practiced. Given that no two doctors and no two patients are the same in all respects, it has always been personalized.
If we narrow the definition of medicine to be a synonym of pharmaceutical, by “personalized,” we mean the right drug(s), at the right dose, in the right formulation, in the right patient at the right time. There are six distinct pediatric populations for each of two sexes. There are geriatric populations as well. This begs the question: How do we know what is right—or more right—than something else?
That begets more jargon, including “pharmacogenomics,” “theragnostics” (or “theranostics”), “targeted (or “tailored”) therapy,” “companion diagnostics,” “P4 medicine” and—most frightening of all—“comparative effectiveness.” How have we all been personalized or exclusively customized?
(1) The disease(s) that we have: It is no longer enough to be depressed, infected with a virus, or attacked by epilepsy, prostate cancer or schizophrenia. The variances among like diseases are manifold. Whether it is tuberculosis or prostate cancer, it is a class of several diseases.
(2) The target for drug pharmacology: As chemical warfare agents, drugs have their targets, and it helps to be sure the right target is in our sights. This justifies theragnostics.
(3) Targets for toxicology: Some of us have expressed unintended targets for the drug that bring adverse consequences. Let’s check first to avoid friendly fire incidents.
(4) Drug metabolizing enzymes: These are everywhere and important, converting the drug to something good, bad or useless. Unfortunately, humans don’t express these enzymes uniformly. Even more confounding is that many drugs induce metabolizing enzymes over time. Who said this would be easy?
(5) Drug transporters: There are too many of these for me to master, but they impact whether a drug will get to its target in sufficient concentration or be discarded as a threat.
(6) Drug-drug interactions: Quite often, drug metabolizing enzymes and drug transporters result in nonlinear effects between drugs, where the bioavailability of one is enhanced or diminished by the presence of the other. For drugs with a narrow therapeutic index, this can be hazardous to your health. Dietary interactions add to the challenge, and only a few have been elucidated in detail.
(7-10) Genes, age, lifestyle and sex: Yesterday, it was our family history. Now, risk factors are increasingly revealed by genotyping that suggests predisposition to disease. This has been disappointing as the data are imprecise and often not actionable. As for predicting the weather, I suspect we will get better at forecasting disease over time.
The above list suggests a daunting number of parameters. What can bioanalytical chemistry do to define our unique status?
The $1,000 genome seems just around the corner. Metabolomics and other searches for biomarkers continue unabated. Searching doesn’t guarantee finding, and there are multiple challenges. Statistics in biology most often involve very broad standard deviations that suggest very little to help the doctor-patient arrive at a decision. Of all the measurable parameters, few have diagnostic value, and even fewer have definitive diagnostic value at reasonable cost. As a contrarian, it seems to me if you really need 10,000 patients to see if a drug or biomarker is viable, you already know it is not!
The concept of comparative effectiveness, whether with drugs or diagnostics, is thus fraught with problems. We allow for off-label use with good justification. There are too many things to determine at once at a cost that now exceeds the GNP of France. Job one is to narrow the possibilities and link them in test panels.
Therapeutic drug monitoring (TDM) has been an important part of the IVD market since late 1970s, yet most prescribed drugs are never monitored in a patient. I suspect this is true for >99 percent, even >90 percent in the hospital setting. Thus, it is only a rough guess that the circulating concentration of a prescribed drug is adequate, but not too high.
Given that no two individuals are exactly alike and there are many phenotypes that influence drug absorption, metabolism and drug-drug interactions, it is a reasonable hypothesis that more TDM would be a very good thing if the technology/cost allowed for it. TDM advanced rapidly in the 1980s based on enzyme immunoassays, but those tests have typically not been dynamic because there was no serial blood sampling and the economics were unfavorable. Data was for a single point in time, and typically was reported hours after the blood sample was drawn. Linking serial data to rapid decision-making is becoming more feasible for critical-care markers, disease markers and drugs.
A lot of attention has been given to the field of pharmacogenomics in an effort to optimize drug selection and dose based on patient genetics. Genes do not tell the whole story. They interact in largely unknown ways, and protein expression is impacted by environmental and life style considerations. TDM is a blue-collar pragmatic approach that does not require reductionist biology.
If the circulating concentration of drug/s is determined in an individual subject versus time, these data capture ALL of the influences of genetics and disease on transporters, enzymes, etc., and can tightly link dose to an individual response. While TDM has been broadly available since the late 1970s, in most cases, the optimum response in an individual is still guessed from the statistical average of a carefully selected group in a trial.
One caveat to keep in mind is that TDM covers all the influences of transporters, diet, fluid flows†and metabolism (including D-D interactions) on drugs in circulation, but it does not say anything about variations in the site of drug action. Thus, while much patient variation is covered by TDM, it also makes sense to have theragnostic tests that validate the pharmacological target for the patient, the tumor or the infecting microbe.
Peter T. Kissinger is chairman emeritus of BASi, CEO of Prosolia in Indianapolis and a professor of chemistry at Purdue University.