Eating Before Bed: What Pre-Sleep Protein Research Shows

A small flexible tube under the skin of the belly looks, in Dr. Mike Ormsbee's own description, "almost like a belly piercing" (06:40). Ormsbee, a researcher at Florida State University, has spent two decades asking what happens inside fat tissue while people sleep, and the tube is how he finds out. Push a thin cannula into the fat just under the skin, leave it there overnight, and measure what leaks out. That technique, called microdialysis, is the quiet mechanism behind one of the more contested questions in nutrition: should anyone eat before bed?

The sound bite most people know is "don't eat before bed." Ormsbee's answer starts with a question of his own: don't eat what? His lab defines pre-sleep feeding narrowly. It means a meal at least two hours after dinner, then a small, protein-centered snack within 30 minutes of getting into bed (03:33). It does not mean a second dinner. A large, mixed meal of carbohydrate, protein, and fat late at night can disrupt heart rate variability and sleep in some labs, he says, and that is not what his studies test.

"What we define pre-sleep feeding as is as two hours after dinner and within 30 minutes before going to bed." (Dr. Mike Ormsbee, 03:33)

Within that narrow window, the protein source turns out not to matter much. For years the field assumed casein, the slow-digesting milk protein, was the obvious choice overnight, since it trickles amino acids into the blood for hours. But newer work from Ormsbee's lab and from researcher Luke Van Loon's group found casein and whey produce no difference in overnight muscle protein synthesis (08:39). Stranger still, Ormsbee suspects the slow-release property of casein may not survive the night at all. Lying flat changes stomach pH slightly, he says, which may interfere with how casein clots and digests (76:31). If that is right, the daytime distinction between a fast protein and a slow one mostly disappears once someone is horizontal.

Exercise as the bigger lever

The more striking finding in Ormsbee's research is not about protein type. It is about what cancels any downside of eating late. In an early study, his lab gave pre-sleep shakes, casein, whey, or carbohydrate, to young, sedentary women with obesity. The next morning, insulin and glucose were elevated no matter which drink they had taken (09:51). Then the same women completed four weeks of exercise training. The negative response disappeared entirely.

"The exercise itself was just such a blunt instrument to control metabolism that the pre-sleep feeding no longer had any sort of deleterious effect at all." (Dr. Mike Ormsbee, 09:51)

Other data point the same way. A 2013 study by researcher James Betts's group restricted people to fewer than 4,000 steps a day. The team then overfed both an exercising and a non-exercising group by 50 percent (11:01). Both groups gained weight. But insulin roughly doubled in the sedentary group, while it stayed close to normal in the group that ran 45 minutes a day. In a study from Ormsbee's doctoral student Brett Cross, just three nights of 30 percent sleep deprivation produced measurable changes in vascular health markers, changes that exercise during that window blunted (12:37). Ormsbee has a name for this pattern: "sometimes I like to refer to it as, uh, sort of like our metabolic apology to our poor lifestyle choices" (13:15).

That apology seems to run deeper than calorie burn. Resistance training triggers an acute rise in growth hormone, which Ormsbee describes as both anabolic, meaning it helps build muscle, and lipolytic, meaning it helps break down fat for fuel (33:53). An older study comparing a Nautilus weight-training session to walking at four miles an hour found significantly higher growth hormone and catecholamine responses from the resistance training (16:00). Ormsbee's working theory, built on years of this kind of evidence, is that muscle and fat communicate during training. "We now know that the muscle actually talks to the fat, tells the fat that that should be going down while muscle mass is increasing," he says (19:28).

When calories get extreme

The clearest demonstration of that idea came from a study Ormsbee ran with people preparing for bariatric surgery. The medically prescribed diet was 800 calories a day, so low that his research team had to petition their ethics board to add protein, since patients were getting only about 0.3 to 0.4 grams per kilogram of body weight, far below what protects muscle (41:11). With added protein, calories landed around 1,100 to 1,200. One group did standard bariatric care, essentially supervised movement in a conference room. The other trained with weights.

"The group that just did standard of care lost a ton of weight, but almost 25 percent of it was from their lean tissue." (Dr. Mike Ormsbee, 42:19)

In the resistance-trained group, lean tissue made up only 4 percent of the weight lost (42:19). The study was small, fewer than fourteen people finished it, but Ormsbee considers it one of the more important findings of his career, especially now that GLP-1 medications are producing rapid, large-scale weight loss in millions of people. The muscle-loss problem, he argues, has not gone away just because the tool for losing weight has changed.

Not every thread in Ormsbee's research resolves so neatly. His group has been testing alpha lactalbumin, a whey-derived protein rich in tryptophan, for its effect on sleep. Published studies using a 93 percent pure form of the protein found improvements in how quickly people fell asleep and how long they slept (52:16). When Ormsbee's team unblinded their own supplement, they discovered it was only about 44 percent pure, which may explain why they saw no benefit at all (52:16). It is a small, almost mundane detail, but it says something about supplement research generally: a named ingredient on a label does not guarantee the dose that actually worked in the trial that made it famous.

Ormsbee is also direct about a myth he hears from clinicians, not just the public. At a scientific conference in Boston, physicians untrained in nutrition told Gabrielle Lyon, the show's host, that excess protein is converted to and stored as body fat. Ormsbee disagrees, citing work from researcher Jose Antonio showing that very high protein intakes, in the range of 3.3 to 4.4 grams per kilogram, produced no long-term fat gain (60:55). "Additional protein does not get stored primarily as fat," he says. "This is not the primary disposal place of protein" (60:12).

One supplement did change his mind after years of skepticism. Collagen, a protein low in the amino acids that build muscle, struck Ormsbee as nutritionally incomplete when he first encountered it. Then his lab ran what may be the longest collagen trial on record, nine consecutive months comparing no collagen, 10 grams a day, and 20 grams a day (67:50). Oddly, the 10-gram dose outperformed the 20-gram dose on joint pain, and the effect was strongest in people who already exercised more than 180 minutes a week, a pattern Ormsbee partly attributes to blood flow that carries amino acids to the joints (69:29). He now takes 10 grams of collagen daily himself.

What runs under all of it, the pre-sleep snack, the bariatric trial, the alpha lactalbumin purity problem, the collagen reversal, is the same habit of mind: ask what, how much, and compared to what, before accepting a rule. Ormsbee's next project tests exogenous ketones on decision-making in SWAT team members during a live-fire exercise under exhaustion (73:07), approached the same way as everything else. "I like to run these studies and then figure out what I think once we run it," he says, and it is close to a thesis statement for the whole conversation.