On the plus side, the data is from [link|http://www.webmd.com/content/article/122/114782.htm|68,183] nurses studied since 1976:

In 1986, the women reported how much sleep they usually got per night. They reported their weight every two years for the next 16 years.

Women who reported sleeping five or six hours per night gained more weight than those who got seven hours of nightly sleep, the study shows. The researchers included Sanjay Patel, MD, of Cleveland's Case Western Reserve University.

Other studies have also linked a good night's sleep to lighter bodies. But it's not always clear which comes first -- healthy sleep habits or weight gain -- or if people who get plenty of sleep at night have other factors that keep their weight lower.

Bumping Up Weight Gain

Data came from 68,183 healthy women who took part in the Nurses' Health Study. The Nurses' Health study started in 1976, when women were 30-55 years old.

In 1986, the women reported how much sleep they usually got per night. They reported their weight every two years for the next 16 years.

[...]

Over 16 years, women who reported sleeping five hours or less per night gained about 2.3 more pounds than those who slept seven hours per night. During the same period, women who got six hours of nightly sleep gained 1.5 more pounds than those who slept an extra hour per night.

Major weight gain -- about 33 pounds, in this case -- was most common among women who reported sleeping five hours per night. Next came the six-hours-per-night group.

What kept those women awake? Did their sleep habits change over time? How do sleep shortfalls affect women's weight? Do the results apply to men?

The study doesn't answer those questions. But the researchers adjusted for factors including the women's age, BMI (body mass index) in 1986, physical activity, and dietary behaviors.

Based on the info here, I find it very hard to believe that average weight gains of 2.3 pounds or 1.5 pounds are significant from a health perspective. (Yes, it may be statistically significant, but that's another issue.) My weight can easily vary +/- 3 pounds in a week even when I'm careful to weigh myself at the same time and the same way every day. I would hope the test subjects were told to weigh themselves the same way every time (same time of day, always before or always after a meal, etc.). The lighter sleepers started off 5 pounds heavier than the 7 hour sleepers, and they say that the lighter sleepers ate less than the longer sleepers, but there's more to eating than raw calories.

I personally would think that someone who always gets 5 hours of sleep a night, always at night, would be better adjusted than someone who has shifts that change over time (e.g. in some factory work it's not uncommon for people to work 1st shift for 3 weeks, then 2nd shift for 3 weeks then 3rd shift for 3 weeks). Or than someone who has sleep apnea or allergies that wake them up every few hours, etc. Or than someone who gets 10 hours of sleep on the weekend and 5 hours during the week. I think many nurses at hospitals have to change shifts frequently, at least these days, while nurses in Dr.s offices wouldn't have that issue. How were things like this accounted for in the self-reported numbers?

I know it's difficult to get information otherwise, but I'm skeptical of self-reported information in health studies. Perhaps the subjects keep careful notes for the 20+ years they're part of this study, but too often these things are reported as if people just write down what they recall about what they've eaten, etc. I would have great difficulty telling anyone, with any confidence, what I ate or how much I slept over the last week.

There's no doubt that sleep is very important, and I wouldn't be surprised if it had some effect on weight gain over time. I just don't think that an average gain of 1.5 or 2.3 pounds over 16 years is a meaningful difference between the groups, and more than that, I am very skeptical that there would be any health effects that could be tied to that weight difference if other important factors were identical.

I think the strongest evidence to support what the press summaries present as the major findings would be: If a particular person's sleep time changed from 5 hours a night to 7 hours a night, did their weight drop (or did they stop gaining weight)? Vice versa? How many people fit into those 2 categories? How much of their reported results can be explained by the fact that it seems to be easier to gain weight once you're overweight?

My $0.02.

Cheers,
Scott.