3500 cals per pound is "wrong" but it's good enough

Why do we use consider that 3500 calories equals the energy deficit necessary to lose 1 lb of fat? Is this right? Why am I not losing a pound of fat for a 3500 calorie deficit? Will my deficit vs lbs lost change over time? (I'll get to these, promise)

First, consider:

A) A pound corresponds to 453.59 g.

There are about nine calories per gram of fat (this is an estimate of the energy available in a gram of ingested fat, it's wrong but we'll go with it).

That corresponds to 4086 cals per pound of fat. So if the fat lost during weight-loss was just dense lipids we could consider that energy deficit - But dietary fat isn't exactly the same as body fat. Physiological fat is held in cells mostly and needs some connective tissue and vessels, etc. to support it.

C) Human fat tissue is estimated to be 87% lipid.

Now that gives us 3555 cals per lb! Pretty close!

Now this 87% number comes from where? Hmmm. A little research into the archives and we find that it references the calculations of Wishnofsky (1952) from the work of one Bozenraad (1911). I find it interesting to note that this number, now taken as fact, was being challenged in 1959.



(from Techniques for Measuring Body Composition, pg 176 - available in Google scholar)

Now that doesn't mean that 3500 cals/lb is outright wrong, just that it's an estimate and might be quite a bit off.

But, let's take this further and look at how much an energy deficit corresponds to a pound of weight loss in more recent literature. What happens when scientists actually observe as the required energy deficit per weight loss?

This is the important part that might affect your weight loss...

http://www.nature.com/ijo/journal/v32/n3/full/0803720a.html



In these graph, the author looks at data from various studies and shows the deficit needed for a specific weight loss versus starting fat mass or body weight. Unfortunately he uses metric KJ rather than calories but that red line corresponds to the rule of thumb of 3500 calories/lb lost. This estimate seems reasonable at 50kg of intial body fat. But for lean individuals the calorie deficit observed to lose might be much lower.

Wait! What?

Why is it then so hard to lose the last 10lbs? Of course, that energy deficit is harder to achieve in a leaner individual if they are already eating less and getting less of a "burn" from exercise (because smaller and lighter and more effcient), because adherence might have waivered or due to, but only in part, the effects of adaptive thermogenesis.

So I should stop because it's all inexact?


You might feel that this is confusing and might as well give up?

But the key message is that the calorie deficit of 3500 / lb of weight loss is an estimate. And it varies not-so-slight not only from individual to individual but also over the weight loss period. There is no need to focus on the the exact number of pounds that will be lost, or the exact calories down to the single digits.

At the end of the day 3500 might be 2600 or 4000. The burn you consider 400 might be 350 or 450. But it doesn't matter as long as we work with a general net deficit, observe and adjust. Adherence to logging as accurately as possible helps to offset if not reduce some of the inherent variability in these estimates.

It's physiology not physics.

tl;dr -- 3500 cals / lb of fat lost is an estimate. Weight loss for leaner individuals is reported to occur at a higher rate in some literature. Use the number as an estimate and live happily ever after.