Tuesday, October 05, 2010

Sleeping the weight off

kw: news, dieting, sleep

The article is titled Dieting? More Sleep Helps Burn More Fat. It is one of about 700 news articles on this topic in the past day or so that are based on a study reported in Annals of Internal Medicine (I would link to the article, but it is embargoed until the issue goes on sale).

Ten volunteers on a low-calorie diet spent four weeks confined to a lab. For two weeks, they slept about 8.5 hours nightly, and for two they slept 5.5 hours. The key finding? During the latter two weeks the volunteers lost twice as much lean mass (muscle) and half as much fat, as compared to the first two weeks. The total loss each two-week period was the same, about 3kg. One more effect was that during the low-sleep weeks, the subjects reported feeling more hunger, and they had raised levels of ghrelin, a "hunger" hormone that also causes fat retention. It is not stated whether the study participants had the opportunity to exercise while confined to the lab.

Naturally, I find this upsetting. I have been reducing by dieting all Summer, and have so far lost about 12kg. I sleep very little, but that has long been true. A "good night's sleep" for me is 6h, and 7h is superb. I haven't slept an 8h night in thirty years.

In the article linked above, they do mention a few questions the study raises. Primarily, how does amount of sleep affect metabolism in general? But my key question is not raised, not only in this article but in a few others I perused: What was the effect of reducing sleep during the latter half of the experiment, rather than the other way around? The experiment needs to be repeated with low sleep first and more sleep later, to ferret out the effect of sleep from the effect of already having lost some weight. It is known that during a period of weight loss, the level of ghrelin rises.

It is necessary to distinguish correlation from causation. Based on this study and on what else is known, we cannot yet draw a conclusion that sleep deprivation was the primary cause of the shift from fat loss to lean loss during the study. I was very disappointed to find that the study had no control. The author of the article, Lynne Peeples, quotes the conductors of the study, who state that more and larger trials need to be performed to answer the questions this study raises. I heartily agree. A properly controlled study needs at least these four groups:
  • Sleep a lot, get lots of exercise
  • Sleep little, get lots of exercise
  • Sleep a lot, lay around
  • Sleep little, lay around
Of course, you'd need at least six volunteers in each group, to generate adequate statistics. Come to think of it, double the number of groups: half of each group needs to be people who have been exercising habitually, and half couch potatoes, and these need to be distinguished, so it is really an eight-group experiment.

Hoo, boy! Things proliferate. For each possible confounding effect you want to control for, you have to double the number of groups! For example, the same eight groups could be split according to
  1. Get plenty of sleep or get too little, all 4 weeks,
  2. Sleep lots 2 weeks then too little 2 weeks versus sleep too little the first 2 weeks, then sleep a lot, and
  3. Exercise or not.
If we add the factor of those who habitually exercise or not, that makes sixteen groups. I'll stop there. The doctors have their work cut out for them to determine what is causing what!

No comments: