Big Data Reveal Actionable Findings to Prevent Disease

Ricki Lewis, PhD

May 08, 2019

A longitudinal study that used enhanced clinical assays, multiomics platforms, and wearable devices to assess health yielded 67 clinically actionable, major findings among 109 participants. The study was published online today in Nature Medicine.

The results, from Sophia Miryam Schüssler-Fiorenza Rose, MD, PhD, at Stanford University, and colleagues, have been a long time in coming. The project began with a 2012 report on monitoring a single individual with multiomics profiling for 14 months. That predates the US National Institute of Health's All of Us initiative, which grew out of President Obama's 2015 Precision Medicine Initiative, as well as efforts in the private sector from Craig Venter's Human Longevity Inc.

"The 2012 paper laid the groundwork and started this whole field. All of Us does longitudinal profiling, but only once a year, and doesn't know what assays it is doing yet," said co–senior author Michael P. Snyder, PhD, the Stanford B. Ascherman Professor, chair of genetics, and director of genomics and personalized medicine at Stanford University School of Medicine, California.

A Menu of Measurements

The report provides proof of concept regarding how multiple assessments, both old and new, can guide preventive healthcare. It offers examples of how certain findings led to further testing and diagnoses. "The goal is to use these advanced technologies, as well as profiling people longitudinally, to see if we can catch early signs of disease and make predictions," Snyder told Medscape Medical News.

For now, it is a research study, but Snyder foresees these advanced technologies entering medical practice. Plans for commercialization are underway, he noted.

The study aims to provide "a new taxonomy of disease on the basis of molecular measures," a definition of precision medicine. "We capitalized on the depth of longitudinal profiling to identify deregulated molecules and pathways associated with the transition from health to disease," the researchers write.

The big data components include genomics, proteomics, metabolomics, transcriptomics, immunomics, interactomics, and microbiome analysis. Wearable devices provided fitness measurements and continuous glucose monitoring.

Some of the clinical tests that were used were standard, such as complete blood counts and lipid assessments. Others were enhanced, such as tests for cardiovascular risk. In addition, participants responded to surveys that assessed physical activity, stress, and diet, and they underwent exercise and other physiologic testing.

Most participants were enrolled 6 years ago; the median period of participation was 2.8 years. Participants were evaluated quarterly.

The cohort was enriched for diabetes risk, and the study included a battery of assessments of insulin resistance and glucose dysregulation. The enhanced metabolic profiling considered microbiome diversity and markers of inflammation and obesity.

However, "[t]he study goes way beyond diabetes. For example, 12 people who had their genomes sequenced learned something important," Snyder said.

Actionable Findings

The longitudinal profiling yielded numerous clinically relevant findings, including the presence of diabetes or prediabetes, heart disease, stroke risk, cancer or cancer risk, hematologic abnormalities, and infection.

One clear trend was that participants tended to be unaware of loss of glucose control. Of 86 participants who reported that they had neither type 2 diabetes nor prediabetes, 43 had prediabetes, including one whose health record included the diagnosis and one whose HbA1C level was in the diabetes range.

"We caught eight people with diabetes as it was developing by continuously monitoring their glucose and insulin levels," Snyder said.

Exome sequencing showed that one participant who had been diagnosed with maturity-onset diabetes of the young (MODY) did not in fact have the causative mutation in the HNF1A gene. Another participant who had not been diagnosed with MODY had the mutation.

"The take-home messages here are that diabetes and prediabetes are highly heterogeneous states of variable glucose dysregulation, that metabolic states can be ameliorated, and that individuals have multiple different trajectories," said Gilbert S. Omenn, MD, PhD, who is the Harold T. Shapiro Distinguished University Professor, director of the Center for Computational Medicine and Bioinformatics, and professor of molecular medicine and genetics, human genetics, and public health at the University of Michigan.

These data and data from a 9-month longitudinal study Omenn was involved with show that big, personalized data can be used to guide behavioral coaching to improve clinical biomarkers, Omenn told Medscape Medical News.

Cardiovascular risk factors also emerged in the Stanford study and led to diagnoses upon follow-up. Six participants had subclinical atherosclerosis; for three participants, oxygen consumption was low; two had nocturnal supraventricular tachycardia, one owing to sleep apnea and another to atrial fibrillation; one person had dilated cardiomyopathy that was traced to a mutation in the RBM20 gene; and 18 individuals had stage II hypertension.

Five participants had strokes during the course of the study — a fact that underscores the potential value of pharmacogenetics. One woman who had been taking aspirin had a mutation that, in conjunction with taking aspirin, raised her risk for a cardiovascular event by 85%. Another metabolized clopidogrel poorly. RNA transcriptomics identified increased stroke risk in another patient. Subclinical inflammation, mostly due to cytokine activity, was detected in all five patients.

Exome sequencing detected eight patients who had genes that increased cancer risk, including APC, BRCA1, and CHEK2. Asymptomatic lymphoma was detected from microbiome changes. One patient was found to have an enlarged spleen, which was then successfully treated.

Thirty patients were anemic; for 28 of these patients, the diagnosis was new. Immunoglobulin (Ig) levels were abnormal in 10 participants, and one had smoldering myeloma. The myeloma and one of the Ig abnormalities were precancerous, enabling early diagnosis and treatment.

With longitudinal, intense monitoring, surprises can happen. That was the case for Snyder himself, who has been followed for 8 years.

"Lyme disease popped up as a shift in heart rate and blood oxygen drop. I had figured out that I was ill, and I suspected it might be Lyme disease, because I'd been in rural Massachusetts 2 weeks earlier. Different types of measurements wound up revealing the disease conditions very early," he said.

Data Led to Lifestyle Changes

For 58 participants, the researchers assessed the effect of monitoring on health-related aspects of lifestyle. Half of these participants reported better sleep, less stress, more frequent self-exams, an increase in the intake of fiber, and that they kept food diaries and attended fitness camps.

Of the 68% who discussed their findings with their physicians, 29% had follow-up assessments, including genetic testing, colonoscopy, positron-emission tomography scans, diabetes monitoring, and eye exams.

The close scrutiny that the study participants underwent was apparently not stressful for them. "While precision medicine approaches have the potential for unnecessary anxiety and over-testing, we did not observe this in our population," the researchers write.

A study limitation relates to the fact that the participants were highly educated volunteers; because of this, the generalizability of the findings may be limited.

Snyder is a cofounder of Personalis, SensOmics, Filtricine, Qbio, and Akna. Several authors report patents related to the work or other commercial interests. Omenn reports no relevant financial relationships.

Nat Med. Published online May 8, 2019. Abstract

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: