More on intermittent fasting

I would like to thank George Albaugh for calling my attention to two studies from Heilbronn et al. ("Alternate-day fasting in nonobese subjects: effects on body weight, body composition, and energy metabolism" Am J Clin Nutr. 2005, 81:69-73.[PubMed]. and "Glucose tolerance and skeletal muscle gene expression in response to alternate day fasting;" Obes Res. 2005, 13:574-81 [PubMed]). These papers describe the effect of a single brief period (3 weeks) of intermittent fasting. I was heartened to see that SIRT1 expression (measured by RT-PCR on muscle biopsies) does indeed increase (p = 0.01), although the effect is relatively small. But I was also concerned to see a possible adverse effect on glucose tolerance.

However, there are reasons why this study may not be directly applicable to my case. They describe a "24 hour" fast, from midnight to midnight, with no mention of how this rule was followed. If I were to follow those instructions I would almost certainly eat at least some food right around midnight each day (just before or just after, depending on whether I was beginning or ending a fast)! This would be similar to the diet I follow now, being a 24 h. fast, except that my large meal is at dinnertime and my average length of fast is probably about 23 hours rather than a full 24. Two things about this concern me. One is that food eaten immediately prior to sleeping may not be digested as well. The other is that the food I eat at midnight, if I do eat at midnight, is Häagen-Dazs Vanilla Swiss Almond or Chocolate Chocolate Chip ice cream. Everyone is different, but social and practical factors might lead those eating at midnight to include a higher proportion of prepared foods in their diets.

The alternative way of following the midnight-to-midnight rule would be to fast from going to bed to waking up. This is much different. A 32 hour fast is only a third longer than a 24 hour fast (perhaps as much as 40% longer than what I'm doing), but probably doubles or triples the time spent hungry, which is presumably the relevant variable for inducing physiological responses. I would find this diet, which is closer to what George is doing, to be significantly harder to follow (especially the bit where I have to go to sleep hungry). In fact, the article in the American Journal of Clinical Nutrition concludes that "Hunger on fasting days did not habituate over the course of the study, which perhaps indicates the unlikelihood of subjects continuing on this diet for extended periods of time." In contrast, what I'm doing now is something that I know I can keep up indefinitely, and that is a big advantage.

So, my concern is that this study had people following two different protocols. My intuition is that men are more likely than women to follow the midnight snack protocol, and more likely to eat high-calorie foods when they do. This could explain some of the differences observed. My advice to the authors of these studies would be to test one protocol or the other. If a one-day diet like the one I'm on is desired, then the instruction could be to fast for at least 22 hours from dinner to dinner (leaving it up to the individual to choose meal time). On the other hand, if the full-day diet is desired, then the instruction could be no food from 4 am to 4 am and to fast for at least 30 hours. Apparently, that was their intent. I have exchanged email with Eric Ravussin (the senior author), who says that most subjects had fast times of about 30 hours, but I wonder if they have interview data that would allow them to sort out what people actually did. The effectiveness of a shorter fast is of obvious interest to me, and I wonder if their existing data bears on this.

In any case, it's exciting that this research is being done, and heartening that they are seeing an effect on SIRT1 expression.