Are We Sitting too Much?
Sitting too much can be bad for you. This is the bottom-line message (no pun intended) from dozens of studies, including a new one I'll be discussing here.
If you're like me, you find these studies concerning. I spend so much of a typical day planted in a chair or sofa.
In the past, this didn't worry me, because I also hike and bike and run and, nowadays, chase my granddaughter (when she's not chasing me).
Recently I noticed something. Scientists weren't just finding that sitting too much is harmful. Some were also reporting that physical activity may not offset the risks. For instance, a 2021 study of more than 130,000 people showed that folks who sit more than 11 to 12 hours per day don't benefit from a half hour of moderate daily exercise.
If you have a desk job, it's not that hard to spend 11 hours a day sitting down. (Apart from work, there's news and social media consumption, mealtime, driving, and, of course, Netflix.) Meanwhile, not all of us have enough time or motivation for regular exercise. Broadly, the CDC estimates that only about a quarter of Americans get enough physical activity.
This leads me to a pair of straightforward questions:
1. How much sitting is too much?
2. How can we offset the health risks created by prolonged sitting?
In this newsletter I'll be discussing a new study that addresses the first question, then I'll offer some answers to both. (Fair warning: I don't see a clear answer to either question.)
Statistics are essential to understanding this topic. People who sit the most tend to have shorter life spans, for instance, but sitting is unlikely to be the main culprit. Rather, health problems shorten their lives and keep them in chairs. Complex statistical techniques are needed to tease out the influence of sedentariness per se from that of other variables such as physical health.
At the same time, for this particular topic, statistics promise a bit more than they can deliver. As you'll see, sedentariness can't be accurately measured, much less fully disentangled from other variables. There are useful messages to be gleaned from the data, but they won't be formulaic. I'm not going to suggest a 5-minute break for every half hour of sitting, for instance, even though it wouldn't hurt.
Why this is important
Here's a bit of trivia: The average body temperature for healthy adults isn't 98.6° F anymore. Studies show that the average has dropped to about 97.5°. The 98.6° figure reflects measurements taken during the 19th century; declines since then are due partly, if not mainly, to the more sedentary lifestyles we lead. Thanks to technological advances, both work and play allow us to sit more than we used to.
This historical trend is a reminder that sitting for prolonged periods is not necessarily indicative of laziness. To an extent, our activity levels have been shaped by societal change. You may have to work at a desk, for instance, and if so, it may not be feasible to use a standing desk or one of those under-the-desk fitness gadgets.
This trend is also a reminder that sedentariness may be a big deal, physiologically speaking. Scientists certainly agree: As one research team commented recently:
"Physical inactivity is associated with several chronic diseases, 3.9–5.3 million annual premature deaths globally, and a $67.5 billion per year cost to healthcare systems worldwide."
Semantic note
Although I've been using the term "sitting", as do many researchers, the ultimate focus of the research concerns physical inactivity, or sedentariness. One can be quite active while seated (e.g., when using a rowing machine), and quite inactive otherwise (e.g., when standing still). But these are relatively exceptional cases. For most people, in most situations, being sedentary and being seated amount to roughly the same thing, so I'll be using the terms "sitting", "sedentariness" and "inactivity" interchangeably.
The new study
This study, led by Dr. David Raichlen at USC, appears in the most recent issue of the Journal of the American Medical Association (JAMA). The focus of the study was on the emergence of dementia in older people.
Dementia refers to a decline in cognitive functioning, over and above what happens during normal aging. Just as "fever" is not a disease but rather a symptom that can arise from many sources, so "dementia" consists of symptoms (impairments in memory, reasoning, verbal skills, etc.) that can stem from many underlying conditions, the most common being Alzheimer's disease.
The design of Raichlen and colleagues' study was quite simple: 49,481 people over the age of 60 with no dementia were followed for roughly 6 to 7 years. During this time, dementia emerged among 414 people.
The main finding was also pretty simple: The more sedentary a person, the more likely they were to be found among that group of 414. (Later I'll explain how sedentariness was measured.)
This is not to say that highly sedentary people always develop dementia. Rather, the data suggest that these people are simply at greater risk. Other risk factors include age, health, alcohol use, etc.
The researchers used statistical methods to control for many of these risk factors. They also controlled for the presence or absence of APOE ε4, an allele that predisposes people to Alzheimer's disease and dementia later in life.
("Controlling for" these risk factors means, in effect, that Raichlen and colleagues measured how much sedentariness increases the risk of dementia, independent of the influence of age, health, and all the rest.)
How much sedentariness is risky?
At this point, you may be thinking: Should I get up and go for a walk?
Well sure, but please come back! More needs to be said about this study and others. So far I've only suggested that being sedentary increases the risk of dementia. I haven't told you how much sedentariness is too much, or how we might offset the risks.
In the ideal world, I could give you numbers. I could tell you the maximum number of hours per day you can sit without increasing your risk of health problems (and the minimum amount of physical activity per day needed to offset any risks).
Alas, it's not an ideal world, although there's some data here that seems helpful.
For instance, in Raichlen and colleagues' study, increased risk of dementia began at roughly 10 hours of sedentariness per day. In other words, once you reach that point, the more you sit, the greater your risk.
Sadly, there's nothing magical about that number. The cut-offs for increased risk vary from study to study, in some studies it matters a lot how active you are when you're not sitting, and, as I'm about to show, we can't trust the specific numbers that any of these studies generate anyway. (To be clear: I'm not saying we can't trust the studies or their findings if expressed in general terms. It's the specificity of the numerical data that's questionable.)
All of the studies on sedentariness are limited in unavoidable ways. In other words, it's the nature of the topic rather than the researchers' methodological choices that create the limitations. The data do tell us something useful, but it's just not formulaic ("get up and move around for 5 minutes every half hour").
Here I'll use Raichlen and colleagues' study to illustrate three of the most prevalent limitations in this literature, then circle back to my original questions.
Study limitations
1. Small effects.
Raichlen and colleagues found the clearest effects of sedentariness when comparing the most active 25% of the sample (12,461 people) to the least active 25% (12,460 people), after controlling for other variables such as age, education level, and health.
82 of the most active people developed dementia. That's 0.66%. In other words, a little less than one percent.
154 of the least active people developed dementia. That's 1.24%. In other words, a little more than one percent.
As you can see, dementia was a rare outcome, and being active didn't reduce the incidence all that much. There were 72 fewer cases among the most active people as compared to the most sedentary ones. That's 72 out of a total of 24,921 people, or 1 out of every 346 individuals.
If you spend a lot of time sitting down, those should be somewhat reassuring numbers.
The study tracked people for roughly 6 to 7 years, so we might ask: What happens later on? Will the incidence of dementia begin to increase more for the least active people than for the most active ones? That seems unlikely. Participants in this study were already old enough that over time, any further influence of sitting per se will diminish. The older the person, the more likely dementia will arise from health problems, unhealthy behaviors, and/or a predisposition to Alzheimer's disease, independent of how many hours per day the person spends in a chair.
In short, if you sit too much, sitting per se only slightly increases your risk of dementia. Small effects are characteristic of this literature. This is not the fault of the researchers but the nature of the beast. Sitting per se is, at most, associated with only slight increases in the risk of premature mortality, cardiovascular disease, certain cancers, and, here, dementia.
2. Measurement error.
I haven't explained yet how sedentariness was measured.
As you can imagine, it would be impossible, logistically speaking, to track the moment-to-moment activity levels of thousands of people for a period of years, much less weeks or days, so you can't blame scientists for limitations in how they end up measuring this variable. But that doesn't mean we should ignore the limitations.
In this case, participants wore a device on their wrists called an accelerometer, which records accelerating movement in any direction. Participants wore the accelerometer on their dominant wrist 24 hours per day for a 7-day period at the outset of the study. Average number of hours per day in a sedentary state were then calculated by means of a machine-learning algorithm.
I tried really, really hard to embrace the validity of this approach. I read up on accelerometers. I reached out to the lead author and got helpful feedback. I thought it over. In the end though, I couldn't escape the conclusion that as measurement procedures go, this is bit of a mess. (Again, nobody should be blamed for not coming up with something better.) Here are two of the more obvious limitations:
(a) Activity levels during a single week are unlikely to accurately reflect how active individuals are across a period of years, even though studies show some consistency when looking at group data.
(b) Accelerometers don't tell us what people are doing when they're sedentary. There's no distinction between, say, passively watching TV versus doing a crossword puzzle, yet it's clear from many other studies that cognitively stimulating activities (including puzzles) can help stave off dementia.
(I discuss a third measurement issue in the Appendix.)
3. Causal ambiguity.
As the researchers point out, it's possible that some third variable, such as an undetected health condition, was responsible for both less activity and the onset of dementia. In other words, it's not that sedentariness causes dementia. Rather, some underlying condition may have been responsible for both.
For instance, mental health was not measured in this study. Reduced physical activity is a symptom of depression, and studies have found that depression is a risk factor for the onset of dementia. Perhaps it's not the case that sedentariness per se leads to dementia. Rather, being depressed may lead to sedentariness and, separately from that, to dementia.
I'm not claiming this is what happened. I'm only saying that it can't be ruled out. As I noted earlier, the relationship between sedentariness and dementia in this study was quite small. It wouldn't require that many depressed participants to create the false appearance of such a relationship.
Conclusion
Here I will suggest some answers to my original questions:
1. How much sitting is too much?
We don't know. The measurement techniques are too imprecise, the effects are too small, and we can't rule out the possibility that other variables are responsible for those effects. I don't think anyone should watch the clock.
I think it's clear from Raichlen and colleagues' study, as well as others, that too much sitting can be harmful. What's useful about this research comes from looking at the proposed causes of harm. Why exactly would prolonged sitting undermine our health?
Researchers have implicated higher blood pressure and poorer glycemic response as culprits. These are variables we have some control over. Trying to maintain a healthy blood pressure, and eating as well as possible, seem more sensible than counting hours and saying, well, I should only sit for 10 hours per day.
2. How can we offset the health risks created by prolonged sitting?
A number of studies seem to offer guidance. For instance, a study published this May suggested that taking a 5 minute break for every 30 minutes of sitting would be sufficient to manage glycemic responses, while a 1 minute break per hour would be sufficient to manage blood pressure.
Here again, I wouldn't encourage you to watch the clock. This study relied on a tiny sample (11 people) and a short time period (less than a month). The researchers only studied four combinations of sitting times and breaks (1 or 5 minute breaks every half hour, and 1 or 5 minute breaks every hour). Meanwhile, at the other end of the spectrum, at least one study points to long-term benefits of reserving all exercise-type activity for the weekend – meaning, at least in theory, that during the week you might not need to worry about planning breaks from sitting.
So, what should we do?
Here I could punt and say that the CDC has developed evidence-based recommendations. Specifically, the CDC advises 150 minutes of moderate-intensity physical activity per week, including two days of muscle-building activities. (For more details, see here.)
This is a bit of a punt, because I've just showed you that the evidence behind evidence-based recommendations can be fallible. Plus a few small-sample studies do at least hint at independent benefits of taking breaks from sitting.
In the end, my advice is to simply take a break from sitting when you're able to (and genuinely feel like it). It certainly won't hurt. I realize this isn't very specific advice, but I don't think the data justify saying more.
Statistics has enabled deep insights into risk factors for health problems, and excessive sedentariness is clearly one of them. But the specificity of the findings promises more concrete guidance than the statistics actually deliver. Stats tell us that sitting too much can be bad for us. They don't tell us exactly when to get up and move around.
Thanks for reading. I appreciate your sharing some sedentary time with me!
Appendix: Accelerometers and activity levels
Accelerometers record movement rather than activity level in a physiological sense. If you have an accelerometer strapped to your wrist, movement will be recorded even while you're sedentary (e.g., when you point the remote at the TV, set the remote down, scratch your head, lift a can of beer, etc.) The researchers used a conventional cutoff for sedentariness: Two consecutive 30-second periods of ≤ 1.5 metabolic equivalent units (METs). An MET is the ratio of the amount of oxygen one uses when performing some activity versus how much one uses while at rest. This sounds impressively scientific, but it's not what Raichlen and colleagues measured. What they measured is acceleration or deceleration of the wrist. That's it. Then they used a "machine-learning" algorithm to infer METs from movement. (No details on that algorithm were provided.) Apart from being highly indirect, this approach cannot capture individual differences in how much oxygen people use for specific activities. This is one more reason the approach to measuring sedentariness was hopelessly imprecise. (Again, the implication is not that we should doubt the general message of the study. Sitting too much is problematic. What we need to be cautious about is the framing of that message in overly specific numerical terms.)