A large study using wearable data from nearly 100,000 people suggests that lower activity levels do not cause Parkinson’s but are an early sign of the disease – raising the possibility of detecting it years earlier and intervening sooner.
Parkinson’s disease is a progressive neurodegenerative disorder that disrupts how the brain coordinates movement: communication between nerves breaks down, leading to tremor, stutter steps and stiffness. Rates have more than doubled worldwide since 2000, causing serious disability and death for millions, according to the World Health Organization.
Since the early 1990s, scientists have been aware of a relationship between physical activity and Parkinson’s – people who report being more active are less likely to develop Parkinson’s. The possibility that exercise – a universally available, free intervention – could prevent a devastating disease drew strong interest from public health officials.
But it has always been a “chicken and egg situation,” says Aiden Doherty, a professor of Biomedical Informatics at the University of Oxford in the United Kingdom. Does exercise reduce the risk of developing Parkinson’s, or does underlying Parkinson’s reduce activity levels years before diagnosis?
A new study, published recently in npj Parkinson’s, used wearable step-count data – an objective measure of daily movement – combined with long-term follow-up to untangle this cause and effect.
After tracking the step counts of nearly 95,000 people – 407 of whom later developed Parkinson’s – and linking these data to diagnoses more than a decade later, the researchers conclude that lower activity levels most likely reflect early disease processes rather than acting as a root cause.
Taken together, the findings suggest that higher step counts do not have a protective effect – consistent with evidence that subtle motor and behavioural changes can precede diagnosis by many years.
The finding is “disappointing”, says co-author Aiden Doherty. “Exercise may not be a panacea that prevents Parkinson’s,” he says, but the findings also point to a different opportunity. Low activity levels could be used to identify people who are developing Parkinson’s before more obvious symptoms develop, which could lead to earlier intervention and more effective treatments.
A dataset built for hindsight
To tease apart the relationship between exercise and Parkinson’s, the researchers turned to the UK Biobank – a decades-long project tracking health data from more than 500,000 people.
Until recently, researchers relied on participants to estimate their own activity levels – something we are all famously bad at. But between 2013 and 2015, Aiden Doherty and his team deployed wearable accelerometers – essentially early smartwatches – to 100,000 UK Biobank participants to measure movement continuously and objectively in daily life.
These volunteers are part of the “worried well,” Aiden Doherty explains – middle aged people between 40 and 70 “generally interested in their health status, a bit more healthy than the general UK population” and overwhelmingly white, similar to the general British population, meaning that the findings will need to be confirmed in more diverse populations.
“The UK Biobank is uniquely suited to this kind of question,” Aiden Doherty says. “It is one of the world’s greatest biomedical datasets.”
The volunteers wore the step counters for about a week. Seven days was enough to capture a stable picture of each person’s typical activity level.
“People are pretty much creatures of habit,” Aiden Doherty says. The team confirmed this in a subset of 3,000 participants tracked over longer periods.
The real advantage of this study is hindsight. More than a decade later, lead author Aidan Acquah, a health data scientist at the University of Oxford, compared the activity profiles of the 407 people who later developed Parkinson’s with those of the 95,000 participants who did not.
When the apparent protection disappears
At first, the data seemed to indicate that exercise protects against Parkinson’s, the authors say. People who logged more than 10,000 steps per day were about half as likely to be diagnosed with Parkinson’s within the next 8 years compared with people who walked fewer than 5,000 steps a day.
Even small boosts in activity seemed to move the needle – each additional 1,000 steps a day corresponded to an 8% lower risk.
But that difference fell apart when the researchers accounted for how far participants were from their eventual diagnosis. These very low step counts were “driven by people who are imminently about to be diagnosed with Parkinson’s” within a couple years of wearing the step counter, Aiden Doherty says – pointing to activity as a marker of disease already underway rather than a factor shaping risk.
“Parkinson’s Disease takes many years to develop before symptoms appear and a diagnosis is made,” meaning that the underlying disease process was already well established by the time participants wore the devices.
At that stage, exercise is unlikely to tip the scales either way, he says.
Instead, the scientists focused on the period “when physical activity might influence disease risk rather than simply reflect underlying disease,” Aiden Doherty says. When they looked at people at least six years away from their Parkinson’s diagnosis, “that apparent protection diminishes and may disappear,” he says.
Steps as an early signal
This was not the outcome the researchers had hoped for from a public health perspective.
“Our data do not suggest that exercise will prevent Parkinson’s disease,” Aiden Doherty says. “But it does not mean people should stop being physically active – exercise has many benefits, from cardiovascular to mental health.”
What stands out is that these low step counts appeared years before diagnosis. This could enable healthcare providers to intervene earlier, when emerging treatments may have a greater chance of slowing progression.
“They are now trialling various drugs and treatments to look at potential approaches to slow down the progression of the disease,” Aidan Acquah says.
Acquah says screening step counts could give researchers access to the crucial prodromal window – the period when the disease is developing in the brain before clear symptoms appear.
The rapid uptake of wearable fitness trackers – used by an estimated one in three people in the United States in 2020 – means that similar analyses can now be carried out across much larger and more diverse populations.
The All of Us study, a United States programme that aims to enrol 1 million participants, enables volunteers to upload data from their Apple Watches and Fitbits.
“In five or 10 years, we will be able to do similar analyses in a United States population with a broader range of ethnicities, ages and income levels,” Aidan Doherty concludes.
