Thursday, December 31, 2009

This is Spinal Compression?

It's New Year's Eve and one's thoughts turn inevitably to ageing and all that goes with it. goes with it. According to physical anthropologists, one thing than goes with it is shrinking: one's height reduces after some point in the 20's due to compression of the spine. But I am not convinced that I am any shorter that in my heydey. So using the good old NCDS I plot height at 42 against height at 23 (controlling for sex). It's pretty much on the 45 degree line, the median change is precisely 0 in fact. I rest my case. Now I just need to rest.

Chandler P.J., R.D. Bock (1991) Age changes in adult stature: trend estimation from mixed longitudinal data Annals of Human biology 18(5) 443-440

Himes J.H., W.H. Mueller (1977) Aging and secular change in adult stature in rural Columbia American Journal of Physical Anthropology 46, 275-280


Liam Delaney said...

alan has been looking at this (indeed I predict a comment in the near future). From what he was telling me, it has a big effect at older ages. I dont think any of the big data-sets would allow a clean test.

Kevin Denny said...

Yes I got the refs from Alan. The effect is supposed to have well kicked in by early 40's & I think NCDS should be plenty big for the purpose. It may be that the critical age is getting later with better nutrition or something.

Kevin Denny said...

And BTW "This is Spinal Tap" is on the telly tonight.

Alan Fernihough said...

The formula, in Mueller - I think, that corrects for 'shrinking' is a quadratic function of other anthropomorphic measurements (arm span etc.) and age. This formula assumes that max height is (roughly) reached at 30. Now if you look at heights at age 23 and 42 - the actual mean difference, in terms of measurement, is five years. So the results in the figure would not be totally inconsistent with the literature. The idea, I think, of this formula is that the effects of spinal compression, bone deterioration, etc. on height manifest in later-life.

I think you may start to see effects on this cohort once they reach 65 or so. I remember looking at birth cohort data from the HSE from 1993 to 2006. Some, quick and dirty, comparisons indicated that the mean heights, by 10-year birth cohorts, decreased for the over-60s. Of course, this wasn't longitudinal data, but certainly suggestive. There certainly needs to be more research done before anything can be concluded.

Are these data objectively measured?

Kevin Denny said...

Can't remember how they are measured: certainly early on for the cohort its objective.
The evidence seems to be all over the place about when shrinking starts. The other studies are heavily parametric whereas what I have is completely non-p.
I don't have the latest height data for the NCDS which would provide a further test. I would be skeptical about non-longitudinal data proving very robust evidence on this.

Alan Fernihough said...

Agreed – longitudinal data, containing complete life-cycle data on objective height, is the only real way to get at this. I might put together a post that attempts to look birth cohort height differences.