Eating Before Sleep: Is a High-Carbohydrate Meal Best?

Most of the sleep hygiene snippets I hear tend to focus on blue light and melatonin, but rarely nutrition strategies to promote sleep (besides tart cherry juice).

In reading through some recent nutrition-focused sleep articles, I came across a repeated concept: that a high-glycemic index (GI) meal consumed in the hours before bedtime can reduce sleep onset latency (SOL), or how long it takes you to fall asleep. Yet there was no mention of exercise—simply a high-GI meal, which felt counterintuitive from a health and blood sugar perspective. Chipping away at that high-GI idea is the topic of this month’s article.

In this article, I’ll review:

• Why carbohydrates are theorized to benefit sleep and muscle recovery.

• The difference between glycemic index and glycemic load.

• If a high-GI meal before bedtime benefits sleep, or if modified macronutrient diets do.

• How pre-bedtime nutrition can affect sleep, beyond diet composition.

• Translating the available science into athletics.

For some related reading, check out my articles on Alcohol and Sleep, Magnesium and Sleep, and Melatonin (the latter post includes sleep hygiene tips).

Carbohydrates for Sleep in Athletes

From my past article on melatonin (read here), you’ll remember that the production pathway for melatonin involves tryptophan (Trp) and occurs in the brain. Trp shares a blood-brain barrier transport protein carrier with the large neutral amino acids (LNAA), which include the branched-chain amino acids. However, when both LNAA and Trp are present and want to use the carrier to cross into the brain, LNAA has a better shot.

In the presence of dietary carbohydrates, which have a strong effect on insulin, insulin can divert enough of the LNAA towards the muscles to improve the likelihood of Trp using the carrier and accessing the brain. Melatonin production can move forward because the Trp precursor has arrived. (1)

What is the Glycemic Index (GI) and Glycemic Load (GL)?

If you’re familiar with these concepts, skip to the next section. Otherwise, the refresher will assist with your understanding of that section.

Glycemic index (GI): GI helps a user understand their potential blood sugar response if they consumed X item. The GI compares foods and beverages based on a 50-gram portion of their carbohydrate content—not a 50-gram portion of the food. For instance, equitable GI servings would be 3 tablespoons honey and 9 cups of spinach, both of which contain 50 grams of carbohydrate. The index has three categories, all relative to the comparison food given a ranking of 100, which is pure glucose (or sometimes referenced as white bread). Those categories are low-GI ≤55 (choose most often), moderate-GI 56-69, and high-GI ≥70 (choose least often). (2)

Glycemic load (GL): But we don’t often eat the GI portion. The GL accounts for the amount of carbohydrate in the portion of food we’re planning on eating:

(GI / 100) x grams of available carbohydrate in that portion (2,3)

Is a High-GI Meal Before Bedtime Beneficial for Sleep?

The crossover study consistently cited is Afaghi et al. (2007), which you can peruse here to learn the details. Their hypothesis was that if healthy adults with no known sleep concerns were provided either a high-GI or low-GI meal before bedtime, the latter would result in improved sleep quality. But that’s not what they found.

On three separate occasions, 12 participants consumed one of two isocaloric GI meals at different times: the high-GI meal either one or four hours before bedtime, or the low-GI meal four hours before bedtime. Each meal contained ~770 kcal and was prepared with 200 grams raw rice (600 grams cooked; ~3.75 cups cooked) and 200 grams steamed vegetables (~1.67 cups) in tomato puree. The high-GI meal included white rice (Jasmine aromatic long grain) with a GI rating of 109 and GL of 175. For the low-GI meal, white rice was also used (Mahatma long grain), having a GI rating of 50 and GL of 81.3. (3)

To measure sleep and sleep architecture, the gold standard of polysomnography was used. The sleep timing and duration was ad libitum, averaging out to be consistent to what the participants typically experienced. (3)

Of the sleep variables evaluated, only SOL was statistically significant at both timepoints:

• When the high-GI meal was consumed one hour before bedtime, the high GI group took 14.6 ± 9.9 minutes to fall asleep compared to the low-GI group at 17.5 ± 6.2 minutes (p = 0.01). Although significant, the difference isn't many minutes.

• When both meals were consumed four hours before bedtime, the high-GI group fell asleep in 9.0 ± 6.2 minutes versus, again, the slower low-GI’s 17.5 ± 6.2 minutes (p = 0.009). (3)

All other sleep variables, like total sleep time and rapid eye movement sleep (REM sleep or that deep recovery sleep), were not statistically different across any time scenario. Even if the high-GI meals resulted in a faster time to fall asleep, participants didn’t stay asleep longer overall. (3)

However, both meals had the same, high amount of carbohydrates: ~3.75 cups of cooked rice and 200 grams steamed broccoli contains 183 grams carbohydrates, which for a 200-pound athlete (91 kg), that's 2.0 g/kg carbohydrate in a single meal—plenty for recovery. Connected with the Trp:LNAA ratio previously mentioned, the carbohydrate values make sense in promoting melatonin production.

From Afaghi et al. (2008).

Porter & Horne (1981) measured the effects of various carbohydrate-containing meals provided to six adult males 45 minutes before bed. The meal contained either zero, 47, or 130 grams carbohydrate. The latter meal “resulted in increased REM sleep, decreased light sleep, and wakefulness … however, the caloric content of the meals was not matched in the study.” (1,4) The absence of isocaloric meals is important because the zero and 47 groups may have been hungry and that could have disturbed their ability to sleep. Yet, carbohydrate manipulation continues to show a pattern.

What About Daily Carbohydrate Intake?

A follow-up study completed again by Afaghi et al. (2008) compared two different isocaloric 2,400-kcal diets across a 48-hour period and how a very low-carbohydrate diet (VLC; 1% of total calories) versus a control diet (72% of calories from carbohydrates) affected ketosis and sleep architecture (measured by polysomnography). Within the daily dietary exposures, participants were fed a meal four hours before bedtime. The researchers found that a VLC diet significantly reduced REM and increased slow wave sleep (NREM sleep stage 4). (5) A diet reducing deep, restorative sleep is not good for athletes. However, this study provides support that carbohydrates play a role.

Lindseth et al. (2013) studied 44 adult females without sleep concerns in a crossover study. Four diets with varying macronutrient intakes were provided to each participant, having them eat each diet four days straight with a two-week washout period in between the test diets. Calories were adjusted for each participant based on indirect calorimetry and all foods and beverages were provided to the participants. Sleep was measured by actigraphy (a watch) to measure sleep efficiency, SOL, and wake episodes. Sleep architecture was not studied. (6)

The four diets had the following percentages of macronutrients, with additional columns that I calculated based on a 2,500-kcal diet for a 5’6” person weighing 154 pounds (70 kg). The study listed an average BMI of 24.8 kg/m² for the group of women, but negated listing information on heights and weights.

Two statistically significant relationships were found in comparison to the control diet:

• The high-protein diet resulted in fewer wake episodes (p = .03): 13.5 ± 7.2 versus 16.7 ± 6.4.

• The high-carbohydrate diet resulted in a shorter SOL (p < .01): 9.1 ± 7.6 minutes versus 13.9 ± 11.7 minutes. (6)

And finally a study in athletes! Falkenberg et al. (2021) evaluated food logs, food photos, sleep diaries, and activity monitors (to compare against the sleep diaries) reported by 36 elite male Australian Football League players (American soccer) over a 10-day preseason period. A summary of the significant findings related to total calories and carbohydrates are as follows:

• Average team total daily intakes were 3,344 ± 955 kcal, 3.4 ± 1.4 g/kg carbohydrate, 2.2 ± 0.8 g/kg protein, and 1.7 ± 0.7 g/kg fat.

• Regarding daily intake: For every 1 MJ (~240 kcal) increase in caloric intake, wake after sleep onset (WASO, or the amount of total time spent awake during a sleep period) increased by 2.6 minutes.

• Regarding nighttime calories consumed after 6pm:

  • Sugar intake affected multiple variables: Greater intakes resulted in less total sleep time (0.1 fewer minutes for every 1 gram of consumed sugar and 5.1 fewer minutes for every 1 g/kg), a reduction in WASO (0.01 fewer minutes for every 1 gram and 1.0 fewer minutes for every 1 g/kg), and greater sleep efficiency (0.2% more for every 1 g/kg).

  • SOL was extended by 5 minutes for every additional 1 MJ (~240 kcal) consumed. (7)

Collectively, adequate daily carbohydrate intakes seem to be important for sleep, and not simply a low-carb diet with a high dose of pre-bedtime carbohydrate.

How Pre-bedtime Nutrition Affects Digestion and Sleep

Macronutrients aside, three considerations involve why we fall asleep and how nutrition may impede sleep.

When our bodies are gearing up for sleep, one signal is a lowered core body temperature. However, eating a meal requires digestion and therefore heat production, meaning our core body temperature remains elevated beyond what’s desirable for sleep. Athletes chasing a high daily caloric target are a red flag for this, as when they skip a daytime snack or meal, they often compensate later in the night. Helping a player get ahead of their caloric needs earlier in the day can help them consume a reasonable portion later at night and managing core body temperature to improve SOL.

Additionally, gravity facilitates digestion. Yet when we eat a meal and immediately lie down, reflux is a common symptom.

Lastly, a lack of calories or long-term restriction can also affect sleep due to a decrease in leptin production, the hormone that signals fullness from eating. Without enough leptin, ghrelin has a stronger "feed me" signal. Objectively, athletes in these scenarios can have a harder time falling asleep (due to hunger) and tend to awaken in the 3-5 AM range (due to hunger). In the latter scenario, athletes have reported it to me as a different feeling than how they experience daytime hunger. However, when they do eat in the 3-5 AM range, they feel better and often return to sleep.

Taken together, giving yourself enough time between eating and falling asleep while managing the volume consumed can play a role in sleep. Oh, and eat enough overall to meet your body’s demands.

What’s Missing? Translating the Research into the Real World and Athletics

What’s lacking are athlete interventions evaluating the interaction between total daily or bedtime nutrition and how dietary composition and timing affect sleep.

Regarding the available research described above:

• Meal and diet food specifics weren’t always listed: We don’t know if fiber content played a role, if all meals were something athletes would actually eat, etc. The rice, vegetable, and tomato puree meal? It's not something I would serve athletes after competition.

• Does a high-GI meal make sense nutritionally? For baseball teams playing every night and eating in the 9:15-11 PM range, what’s the impact of a spiked blood sugar every night during the post-prandial and pre-sleep resting period?

• The Lindseth et al. (2013) study offering 2.0 g/kg/day carbohydrate in two of the diets doesn’t make sense for higher-intensity or endurance sports. Yet the highest this study went was 5.0 g/kg, which still isn't high enough for certain sports or events.

• The clinical application may not be worth it: If REM sleep or total sleep duration aren’t affected, does it matter if an athlete can fall asleep six minutes faster? Is promoting sugar at bedtime a realistic recommendation?

• However, the pattern of adequate carbohydrate intakes in healthy adults seems to improve some aspects of sleep.

Food is only one factor impacting sleep. Other important areas include stress, anxiety, and adrenaline from competition. The assumption in reviewing the literature is that all other sleep hygiene recommendations were being followed, which was beyond what the studies were designed to do. Even if the test meal composition was effective at improving sleep, the video games being played until 1 AM are likely to offset any benefit.

Key Takeaways

Collectively, sleep is a huge component for recovery and nutrition seems to have an effect in non-athlete adults, especially regarding carbohydrate intake. Here’s what you can do as the dietitian:

• Evaluate a player’s total daily calorie and macronutrient goals, coaching them to work towards those targets throughout the day—and not only hyper-fixating on a large post-competition meal close to bedtime.

• Same goes for fluid and rehydration plans: Sipping on fluids throughout the day and optimizing during-exercise fluid intake to limit sweat loss means the athlete isn’t scrambling to rehydrate at night, waking up later on to use the bathroom.

• Depending on the sport, the utility of high- and low-GI carbohydrates differ: After a 2-hour rowing practice or soccer match, oarsmen and high-minute soccer players should include high-GI options to replete muscle glycogen and the caloric deficit accrued. Whereas in baseball, a bullpen pitcher who doesn’t take the mound would benefit from low-GI options. Regardless, meeting total daily carbohydrate goals specific to a player's sport, position, and overall daily energy expenditure seems to be productive for sleep.

• Continue to harp on all aspects of sleep hygiene (read more here).

• The use of cow’s milk (tryptophan!), cherry juice (melatonin!), and/or kiwis (serotonin!) can also support SOL management. Funny that they're also all sources of carbohydrates.

References

(1) Halson, S.L. (2014). Sleep in elite athletes and nutritional interventions to enhance sleep. Sports Med,44(Suppl 1):S13-23. https://pubmed.ncbi.nlm.nih.gov/24791913/

(2) Higden, J. (2003; reviewed 2016, March). Glycemic index and glycemic load. Oregon State University Linus Pauling Institute Micronutrient Information Center. https://lpi.oregonstate.edu/mic/food-beverages/glycemic-index-glycemic-load

(3) Afaghi, A., O’Connor, H., & Chow, C.M. (2007). High-glycemic-index carbohydrate meals shorten sleep onset. Am J Clin Nutr,85(2):426-30. https://pubmed.ncbi.nlm.nih.gov/17284739/

(4) Porter, J.M., & Horne, J.A. (1981). Bed-time food supplements and sleep: effects of different carbohydrate levels. Electroencephalogr Clin Neurophysiol,51(4):426-33. https://pubmed.ncbi.nlm.nih.gov/6164541/

(5) Afaghi, A., O’Connor, H., & Chow, C.M. (2008). Acute effects of the very low carbohydrate diet on sleep indices. Nutr Neurosci,11(4):146-54. https://pubmed.ncbi.nlm.nih.gov/18681982/

(6) Lindseth, G., Lindseth, P., & Thompson, M. (2013). Nutritional effects on sleep. West J Nurs Res,35(4):497-513. https://pubmed.ncbi.nlm.nih.gov/21816963/

(7) Falkenberg, E., Aisbett, B., Lastella, M., Roberts, S., & Condo, D. (2021). Nutrient intake, meal timing and sleep in elite male Australian football players. J Sci Med Sport,24(1):7-12. https://pubmed.ncbi.nlm.nih.gov/32624442/