In a related question, Are all cells of the human body completely replaced every seven to ten years?, it is explained that there are cells in the human body, like neurons, that are never replaced.

However, are the atoms within those neurons replaced?

Some examples of people claiming they are:

  • Quora

    Every ~16 days nearly 100% of the water is exchanged in a healthy body. Heavy elements like carbon, sodium and potassium take occupancy far longer perhaps 8 months - 11 months.

    In about a year every atom in your body would have been exchanged. Not a single atom in your body resides there forever and there is a 100% chance that 1000s of other humans through history held some of the same atoms that you currently hold in your body. -

  • Time Magazine

    There was a study done at Oak Ridge Lab. by Paul C. Aebersold in 1953 that found that 98 percent of all the atoms in a person’s body change out every year, and that within five years all the atoms had changed.

  • Reddit

    Studies at the Oak Ridge Atomic Research Center have revealed that about 98 percent of all the atoms in a human body are replaced every year. You get a new suit of skin every month and a new liver every six weeks. The lining of your stomach lasts only five days before it’s replaced. Even your bones are not the solid, stable, concrete-like things you might have thought them to be: They are undergoing constant change. The bones you have today are different from the bones you had a year ago. Experts in this area of research have concluded that there is a complete, 100 percent turnover of atoms in the body at least every five years. In other words, not one single atom present in your body today was there five years ago.

  • 2
    @jinawee: 1. They're not identical. 2. Replace some of them with isotopes you can track, and extrapolate for the rest, apparently.
    – endolith
    Nov 18, 2013 at 20:39
  • 1
    For this claim to be true, either the cells that never get replaced don't have DNA (of which there is only one: red blood cell, and these are replaced a lot), or the DNA must get completely rebuilt within those cells out of fresh atoms. Dunno, but I'm skeptical. At the very least, this seems like a waste of energy and an extra risk of mutations, so should be selected against.
    – RomanSt
    Nov 18, 2013 at 20:49
  • 3
    Not posting answers since these are my own thoughts without sources, but a) 98% per year means that even if 98% of atoms are replaced randomly then it will take ~16 years for every last one to have been replaced. Secondly, it is likely that the 2% that didn't get replaced are exactly those residing in places that see a much lower atom turnover rate, thus making it less likely that every last atom got replaced during a person's lifetime.
    – RomanSt
    Nov 18, 2013 at 20:58
  • 1
    It is possible that "100%" is due to rounding, and that while it might not be all atoms, the percent may be 99.999... to several decimal places. Rounding up to 100% would not make such a claim false in that case. Nov 26, 2013 at 3:47
  • 4
    @trav1s That could be the case, but the claims also use the word "complete", "all the atoms", "not a single atom". Those are definitive, and not susceptible to a rounding interpretation.
    – user5582
    Nov 26, 2013 at 6:46

1 Answer 1


Atoms in tooth enamel are not replaced. As a consequence, the C14 levels found in a person's tooth enamel can be used to estimate the year at which their teeth formed, due to the fact that atmospheric levels of C14 have decreased steadily since the cessation of above-ground nuclear bomb testing in 1964.

Details can be found at the link below, along with information on the relative rates of turnover for other tissues.


Figure 9 (page 34) shows that some tissues turn over more slowly than others, and therefore retain more of the C14 from the person's early years. This demonstrates that the rate of turnover varies by tissue, and that there is not a single uniform rate of turnover that can be used to judge how quickly atoms are replaced.

In total nine tissue types were selected for measurement from the 36 donors [who died in 2006].... The tissues with the lowest radiocarbon values (ie closest to atmospheric levels in 2006) and the lowest standard deviations within the sample set were blood, hair and nails. Skin and bone lipid values have slightly higher radiocarbon values, but relatively low scatter. Surprisingly, skin collagen radiocarbon content was quite variable, suggesting it is not a good candidate for Year-of-Death determination.

That paper also cites other studies reporting that there is no turnover of neuronal DNA and eye lens crystaline protiens.

Work by Spalding et al. (2005b) and Bhardwaj et al. (2006) on neural tissue DNA, and by Lynnerup et al. (2008) on eye lens crystalline proteins have confirmed that other tissue components in the body besides tooth enamel, once formed, do not turnover during life.

  • 2
    I didn't find "information on the rates of turnover for other tissues" (what pages is it on?), but I did see an assertion on page that "non-turning over tissues" include "tooth enamel, neural cell DNA, and eye lens crystallines", with references such as, Lynnerup N, Kjeldsen H, Heegaard S, Jacobsen C Heinemeier J (2008) Radiocarbon Dating of the Human Eye Lens Crystallines Reveal Proteins without Carbon Turnover throughout Life. PLoS One (www.plosone.org) 1: 1-3.
    – ChrisW
    Nov 22, 2013 at 19:43
  • 7
    I guess that's enough to refute the claim that all atoms are replaced.
    – user5582
    Nov 22, 2013 at 19:50
  • Perhaps "rate of turnover" was is not a technically correct term, since they never performed that exact calculation. However, you can easily get a rough sense of the relative rates of turnover from their data. They describe the soft-tissue measurements in paragraph three of the abstract. You can find details by doing a text search for those tissues samples (e.g. skin lipid). If you can think of a more accurate way of describing that data, please edit the answer. And thank you for pointing out the references.
    – adam.r
    Nov 22, 2013 at 20:10
  • 2
    Articuno, more generally, I think that this answer demonstrates the logical error in some of the initial assertions. The human body is not a well-mixed bag of chemicals -- so you can't simply measure that there is 98% turnover in a year (i.e. 2% remaining) and then extrapolate the fraction that has not turned over as 0.02 ^ t (where t is the number of years).
    – adam.r
    Nov 22, 2013 at 20:15

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .