Most medical tests are snapshots. A single blood draw captures what’s happening in your body over a short period of time, but diseases like type 2 diabetes develop slowly, over years. What might we be missing by looking only once?
A new study in the journal Diabetes offers a striking answer. Led by Jerome I. Rotter, MD, an investigator in The Lundquist Institute’s Advanced Therapeutics and Biomedical Technologies Innovation Hub and Director of the Center for Translational Genomics, along with his colleagues Ida Chen, PhD , Xiuqing Guo, PhD, and Kent Taylor, PhD, and collaborators Zsu-Zsu Chen and Robert Gerszten in Boston, the research shows that tracking the same people over time reveals far more about diabetes than a one-time test can.
Proteins are the body’s workers, and the levels circulating in our blood can signal disease long before symptoms appear. Scientists call these signals biomarkers.
The team studied 5,322 adults from the Multi-Ethnic Study of Atherosclerosis (MESA), a long-running study of people from diverse backgrounds. Participants, who averaged about 62 years old at the start, gave blood at six visits spanning roughly 16 years, from 2000 to 2018. At three of those visits, researchers measured about 3,000 proteins.
What the study revealed
The study asked what would seem to be a simple question: which proteins predict who will develop diabetes? Dr. Rotter and colleagues answered it two ways, once using only the first blood draw, and once using all three.
Here’s what they learned:
Watching over time finds far more clues. The single snapshot turned up several hundred protein signals for diabetes (27 that were entirely new). But tracking people across all three visits identified up to four times as many protein links to the disease.
The extra clues held up. It could have meant that more findings could simply mean more false alarms. To check, the team used Mendelian randomization, a method that uses people’s genes as a natural experiment to test whether a protein may actually help cause disease, not just accompany it. A similar number of findings passed this test in both approaches. Because that share held steady, the additional proteins found by tracking over time appear just as meaningful as the ones a snapshot catches.
But remarkably, the two lists of proteins reveal different biology. Proteins found by both methods were mostly tied to metabolism, how the body breaks down and uses small molecules. The proteins found only by tracking over time pointed elsewhere: to how cells build and modify proteins, maintain their structure, and divide. These changes seem to unfold within a person as they move closer to a diabetes diagnosis, something a single test can’t see. They are also largely new discoveries. Just 42% of the tracked findings matched results from the UK Biobank, a major snapshot-based study, compared with 87% of the single-visit findings.
Learning by observing
The findings open new doors. Proteins that shift as diabetes approaches could one day help doctors spot rising risk earlier, when lifestyle changes or treatment can do the most good. Proteins that may actually drive the disease are promising starting, points for new therapies.
The study also makes a case for patience. These discoveries were possible only because thousands of people returned year after year to give blood, and because researchers sustained the work for over two decades (the study continues today). Long-term studies like MESA take time and steady support, but as this research shows, they can reveal what a single moment never will.
For Dr. Rotter and his colleagues, the lesson is clear: to understand how disease unfolds, we have to keep watching.