BioWorld International Correspondent

LONDON - Population screening for cardiovascular disease could be available within just a few years, if the promise of a new test holds true. The test, which involves reading the nuclear magnetic resonance (NMR) spectrum of a serum sample, would allow clinicians to target treatment with lipid-lowering drugs, such as statins, to those who would benefit most.

A large trial to evaluate the test in clinical practice has already begun. Simultaneously, numerous pilot studies are under way to apply the technology to the diagnosis of a range of other conditions, from cancer and Alzheimer's disease to osteoporosis and osteoarthritis.

David Grainger, Royal Society university research fellow in the Department of Medicine at Addenbrooke's Hospital in Cambridge, UK, told BioWorld International, "The principle behind this test is applicable to every disease. This technology has been available for some years, but this is the first time it has been successfully applied to the diagnosis of a human disease."

He predicted that in 10 years this type of test would be one of the most important diagnostic technologies in use, with, probably, a machine carrying out such analysis on every general practitioner's desk.

When a serum sample is put in a magnetic field, all the molecules within it resonate. Each type of molecule resonates at a unique frequency. The result is a spectrum containing a series of peaks, each of which varies in intensity according to the concentration of the different molecules present in the serum.

Grainger described that as a "metabolic fingerprint" that contains an enormous amount of information about the metabolism of the individual. "The clever bit," he said, "is the computerized analysis of these spectra, which allows us to detect minute but reproducible pattern differences that are associated with particular diseases."

A spin-off company of Imperial College, London, named Metabometrix Ltd., holds the patent on the technology, which has been called NMR-based metabonomics. It in turn has agreed to license the technology to Fingerprint Diagnostics Ltd., of Cambridge, UK, which is funding the large trial currently in progress.

The study on cardiovascular disease is reported in the Nov. 25, 2002, online Nature Medicine in a paper titled "Rapid and non-invasive diagnosis of the presence and severity of coronary heart disease using 1H-NMR-based metabonomics." The paper shows that the technique allowed more than 90 percent of people with narrowing of the three major coronary vessels to be distinguished from those with normal coronary arteries.

Grainger said the study came about when his wife, Elaine McKilligin, who works in the Medicines Research Centre at GlaxoSmithKline in Stevenage, UK, came across a paper reporting that it was possible to distinguish black mice from white mice through computerized analysis of the NMR spectra of serum from the animals. "She knew that the aim of the laboratory that I worked in was to find a way to diagnose cardiovascular disease cheaply and simply, and she suggested that it would be useful to apply this method to the diagnosis of heart disease," Grainger said.

He duly contacted the research team of Jeremy Nicholson at Imperial College of Science, Technology and Medicine in London, who had pioneered the technology, and who had carried out the study on mice, and their collaboration began.

Working also with clinicians at Papworth Hospital near Cambridge, under the leadership of Peter Schofield, Grainger and collaborators collected serum samples from their patients, all of whom had undergone coronary angiography, and were known to have either narrowing of all three major coronary vessels, or normal coronary arteries.

"We wanted to know," Grainger said, "whether the computer could identify patterns in the profile of small metabolites in the serum of these people. Was there a pattern that was always present in people with heart disease, but never present in people who are healthy? The answer, we found, was yes.'" The team also was able to predict which patients had heart disease when they used the method to test blinded samples.

They found that the test was both sensitive and specific, and could, therefore, be suitable for use in population screening. A study of 3,650 patients, called MaGiCAD, has already begun at Papworth Hospital. Grainger hopes the results of this trial will be published during early 2004.

He agreed that the method also could probably also be used "in reverse" to allow researchers to find out what genetic defects underlie particular diseases. "You get a more dynamic view of the individual from looking at their metabolites. This technique tells you everything about their metabolic tendencies, including features that may have been set when they were in utero, so in a sense this brings you nearer to the sharp end' of the organism than their genes do."

Grainger said the NMR machines needed for the serum analysis, while extremely expensive at about GBP 0.5 million, could carry out tens of thousands of tests a year, making the cost of each test quite low.

He and his colleagues are also working on ways of improving the already good sensitivity and specificity of the test further. "We only used very crude one-dimensional NMR spectra but it is possible to get much more informative spectra than this," he said.