Large-scale studies take patience, as epidemiologist Paul Elliott at Imperial College London knows all too well. More than a decade ago, he, Jeremiah Stamler of Northwestern University in Chicago, Illinois, and an international team began collecting data for the INTERMAP project — a study on the causes of high blood pressure. This condition is a leading cause of cardiovascular disease worldwide, and a large body of evidence suggests that it is determined primarily by environmental factors such as diet. “If we can determine the key environmental stressors, there is tremendous potential to prevent heart disease and stroke,” says Elliott. To investigate just what impact different factors might have, the researchers combined information about environmental influences with metabolic profiling, an approach that Elliott describes as “a very exciting way forward”.

They began collecting data in 1996, gathering details about diet, alcohol, drug and supplement use from more than 4,600 individuals aged between 40 and 59 living in the United States, United Kingdom, Japan and China. They also measured blood pressure and body mass index, and collected urine samples. “It was a massive undertaking,” says Elliott.

Two urine samples were collected from every individual. Because urine contains the chemicals excreted by the body during metabolism, its composition is a rich source of information about the combined influences of environment, gut microbes and genetics.

Having analysed the amounts of certain urinary constituents and found links between them and a person's diet and blood pressure, Elliott and his collaborators decided to take a more systematic and detailed look at the samples. To do this, Elliott approached Jeremy Nicholson, an expert in nuclear magnetic resonance (NMR) spectrometry, also at Imperial College London. This technique analysed the entire spectrum of thousands of metabolites contained in each urine sample, a process referred to as metabolomics.

“We have shown it is possible to do metabolomics on a much larger scale than has been done before,” says Elliott. It took members of Nicholson's lab almost two years to perform NMR analysis of all 10,000 samples — including controls — collected through the INTERMAP project. It took another three years for Elliott, Stamler and Nicholson's teams to make sense of the mountains of data, as they compared metabolite profiles between different populations. “That was a huge challenge,” says Elliott. “We had to develop a very stringent statistical approach to identify discriminatory metabolites.”

They found and quantified several metabolites that were associated with high blood pressure (see page 396). If further studies confirm these associations, following the pathways that lead to these compounds' production could reveal mechanisms underlying high blood pressure and heart disease.

Elliott hopes to be able to continue these collaborations to follow up on different lines of research. “This is big science and long-term science. It requires universities and funding agencies to take a long view of research,” he says. “But the pay-off is potentially very high.”