Category Archives: pharmacogenomics

Three clinician-researchers described their work in using individual patient information to guide the treatment of disease.

End-stage kidney disease

A golden age of anemia treatment in end-stage renal disease once existed, launched with a drug became commercially available in 1989. The drug, recombinant erythropoietin, or EPO, prevented iron overload in patients who received blood transfusions, replacing those treatments with an IV injection during dialysis. Quality of life appeared to be dramatically improved.

That golden age ended, however, with the discovery that EPO was associated with an increased risk for cardiovascular events, said Jon Klein, professor of medicine and holder of the Endowed Chair in Proteomics at the University of Louisville. Eventually, the drug was subject to a black-box warning from the FDA and a CMS mandate to use the lowest dose needed to prevent the need for a transfusion.

“Clinicians didn’t know what to do with this,” Klein said. Should they lower the dose dramatically, or consider stop giving the drug altogether? Klein and colleagues sought to provide optimal dosing information via the analysis of proteins that influence patient sensitivity to the drug.

Using a series of extractions of plasma, the researchers tracked several pools of information from fragments of proteins, intact proteins, proteins bound together and microparticles. Two proteins emerged as biomarkers of sensitivity to the drug.

Applying computational intelligence, the researchers developed an algorithm-driven dose optimizer that guided clinicians in the continued use of EPO. The model-guided treatments helped patients reach their target hemoglobin levels within 20 weeks, compared with 50 weeks using a more traditional trial-and-error clinical approach.

Heart failure

A key to keeping heart failure patients out of the hospital is regulating their intra-cardiac and pulmonary artery pressure. When it’s too high, they are at risk for arrhythmia. When it’s too low, they might pass out. Getting it right can dramatically reduce hospitalization and improve patients’ quality of life.

“If we wait until patients are symptomatic or their weight increases, the window of opportunity to intervene relatively short,” said William Abraham, chief of the division of cardiovascular medicine at Ohio State University Medical Center. “So if we can only measure the right signal and respond to it in a proactive way, we might be able to reduce hospitalizations and improve patients’ quality of life or well-being.”

Abraham has been an investigator on trials testing the safety and effectiveness of implantable devices that monitor pulmonary artery pressure. They have no batteries or moving parts, and are delivered by catheterization. Trials have tested a variety of methods of monitoring, and a clear message has emerged. When patients are guided to monitor their pressure, report the levels to a clinician and titrate their medications accordingly, hospitalizations are reduced and their quality of life improves.

“This is an example of how we are exploiting knowledge of individual patients’ physiology to develop new approaches to heart failure management,” he said.

LDL cholesterol response to statins

It is known that how patients will respond to statin therapy to lower their bad cholesterol is a strong predictor of their reduced risk for cardiovascular problems. But do genes drive that response? Researchers are trying to find out how influential genes might be, said Daniel Chasman, director of computational biology at Brigham and Women’s Hospital and Harvard Medical School.

Candidate genes identified in the past were tied to a number of effects: They might affect uptake of the drug, inhibit production of cholesterol, stop its transport or degrade it altogether.

Chasman and colleagues ran an enormous genome-wide analysis of almost 7,000 patients for whom statin treatment had been highly effective. The analysis produced 820,000 single nucleotide polymorphisms, or SNPs (pronounced snip). Each gene contains two alternative forms – called alleles  – that are identical in most people. However, when the activity level, or expression, of one allele differs from its partner allele in a single gene, that small difference is called a SNP. The SNP affects a gene’s protein-producing process.

The researchers found a total of three SNPs associated with the lowering of LDL cholesterol in the population studied. A placebo group showed no such associations. It remains unclear just how powerful these genetic variants are in influencing a patient’s response to statins, but the more that is known, the closer clinicians will come to prescribing the right dose or the right medication for the right patient on a routine basis.