In 2012, a sleep laboratory at Loughborough University in the English Midlands ran an experiment that anyone who has sat through a two o'clock meeting will recognise. Twelve young men, each deliberately short of sleep after a five-hour night, ate lunch and then drove for two hours on a simulator along a monotonous road. On one day the lunch was light, about 305 kilocalories. On another it was heavy, about 922, with three times the fat and twice the carbohydrate. After the heavy lunch the drivers drifted out of their lane more often, and their brainwaves showed more of the slow activity that precedes sleep [1]. It was a small study, and the men were tired to begin with. But the authors' conclusion is the part worth keeping: the heavy lunch had "exacerbated inherent sleepiness". The sleepiness was already there. The lunch made it worse.

By the second week of September, most of us have moved back indoors; we wrote about that turn of the year in The Handover. The afternoon that in July was a swim in a lake outside Riga is now a calendar block with four other people in it, and at some point around three the sentence you were reading stops going in. The popular explanation is a sugar crash: lunch spiked your blood glucose, insulin overshot, and you are running on empty. That story is not wrong, exactly. It is incomplete, and the honest picture is more useful than the simple one.

A dip that does not need lunch

The afternoon dip was in you before lunch was. Sleep researchers call it the post-lunch dip, and a 2005 review by Timothy Monk in Clinics in Sports Medicine summarised decades of work in one careful sentence: it "is a real phenomenon that can occur even when the individual has had no lunch and is unaware of the time of day" [2]. Monk was equally careful about its size. It appears, he wrote, "for some performance variables, and some individuals". It is not universal and it is not enormous. But it has "its roots in human biology", and it is "exacerbated by a high-carbohydrate lunch" rather than caused by one.

Where do those roots go? Two places. The first is the clock. Alexander Borbély's two-process model of sleep, first published in 1982 and reappraised by its authors in 2016, describes sleep pressure as the sum of two forces: a homeostatic process that builds the longer you are awake, and a circadian process, driven by the brain's master clock, that rises and falls on a roughly 24-hour schedule [3]. The molecule most closely tied to the first is adenosine, a by-product of the brain spending energy, which accumulates through the day and presses on sleep-promoting circuits until it is cleared overnight.

The second is a smaller rhythm inside the big one. In 2009, Frederik Bes and colleagues fitted a mathematical model of sleep propensity to experimental data and found that the curve has "a secondary increase, which peaks post-noon", a nap zone whose size depends on when you slept the night before [4]. Siesta cultures did not invent the siesta. They noticed it. So by mid-afternoon you have hours of accumulated adenosine, a circadian trough and a secondary rhythm nudging you towards sleep. The lunch arrives on top of that.

What the meal adds

When you eat carbohydrate, whether as bread, rice, an apple or a glass of juice, it is broken down to glucose and absorbed into the blood. The pancreas releases insulin, the hormone that tells muscle, liver and fat cells to take that glucose up. Blood glucose rises over the first half hour to an hour, peaks, and falls as insulin does its work, usually returning to baseline within two to three hours. The height and shape of that curve depend on how much carbohydrate there was, how fast it arrived, what else was in the meal, and, it turns out, who you are.

The last point was made emphatically by the PREDICT 1 study, published in Nature Medicine in 2020 by Sarah Berry, Tim Spector and colleagues [5]. They gave 1,002 people in the UK, many of them twins, identical test meals, measured their glucose continuously for two weeks through a small sensor on the arm, and repeated the protocol in 100 people in the United States. Responses to the same meal varied between people by 68 per cent for glucose and 103 per cent for blood fats. The meal's macronutrients explained about 15 per cent of the variation in glucose response, the gut microbiome about 6 per cent, and genes, even with the twins to help, under 10 per cent. Two colleagues eating the same canteen lunch will have different glucose curves, and neither can predict theirs from the menu.

The dip and the biscuit tin

What matters for the afternoon is not the peak but what follows it. Insulin does not switch off the instant glucose is back to baseline, and in some people, after some meals, the curve undershoots for a while before settling. This is the "crash" of the popular story, and the same research group tested whether it is real and whether it matters.

In 2021, Patrick Wyatt and the PREDICT team published in Nature Metabolism an analysis of 1,070 people wearing continuous glucose monitors while they ate 8,624 standardised meals and 71,715 meals of their own choosing [6]. They defined the dip as the lowest glucose reading two to three hours after eating, as a percentage below the person's baseline. Across the standardised meals the average dip ranged from about 4 to 31 per cent below baseline depending on the meal, with wide scatter between individuals. A modest undershoot, not a plunge. And yet it predicted behaviour. People in the quarter with the biggest dips felt hungrier at two to three hours, reached for their next meal about 24 minutes sooner, and ate about 75 kilocalories more in the following hour and 312 kilocalories more over the day than the quarter with the smallest dips. The dip predicted these things better than the peak did.

Two honesty notes. The correlations were real but small, between 0.16 and 0.27; the dip is one input among many. And what it predicted was hunger, not sleepiness; the study did not measure alertness, so "the sugar crash makes you sleepy" is an inference, not a finding. Put the halves together: the circadian trough would make you drowsy at three whatever you ate; a lunch that swings glucose adds a modest dip that makes you hungry; a heavy lunch, in the Loughborough drivers, made the drowsiness worse. The 3 p.m. biscuit is where those stories meet.

Why the same sugar behaves differently in an apple and in a glass

This is the part that concerns us directly, because what we make is juice, and juice is fruit sugar in liquid form. Cold-pressing keeps the flavour and body of the fruit, which is why we do it (we have written about what pressing does and does not do). It does not put the fibre back. So it is worth knowing what the fibre was doing.

The cleanest demonstration is nearly fifty years old. In 1977, Gerald Haber and Kenneth Heaton at Bristol gave ten healthy volunteers the same 60 grams of carbohydrate as whole apples, apple purée or apple juice [7]. The juice could be drunk eleven times faster than the apples could be eaten, and even at a matched pace it was significantly less satisfying. The rise in blood glucose was similar for all three. What differed was the descent: after the juice there was, in the authors' words, "a striking rebound fall" below the starting level, smaller after the purée, absent after the apples, and insulin rose higher after juice and purée than after whole fruit. The sugar was the same. The apple's structure was what slowed it down.

That is the argument against a glass of juice on its own at three o'clock, and we would rather make it ourselves. It is not an argument against juice, because the second half of the story is what happens when a sweet liquid arrives as part of a meal.

A Finnish group showed how much the rest of the plate matters in a 2011 study in the British Journal of Nutrition [8]. Eleven healthy people ate mashed potato alone or with chicken breast, rapeseed oil and salad. Alone, the potato had a glycaemic index of 108, higher than pure glucose. With the chicken, oil and salad, the same potato scored 54; calculating from its parts would have predicted 103.

Order matters as well as company. In a 2015 report in Diabetes Care, Alpana Shukla and colleagues at Weill Cornell fed eleven people with type 2 diabetes an identical 628-kilocalorie meal of bread, orange juice, grilled chicken, salad and broccoli on two occasions [9]. On one day they ate the bread and drank the juice first; on the other, last. Carbohydrate last meant blood glucose 28.6 per cent lower at 30 minutes, 36.7 per cent lower at 60 minutes and 16.8 per cent lower at two hours, with insulin over the two hours almost halved. Eleven people with diabetes is a small, specific group, and the effect in a healthy person is probably smaller. But the direction is the one every mixed-meal study points in: a glass of fruit sugar drunk after the salad and the protein is a different physiological event from the same glass drunk alone at a desk.

Ten minutes on your feet

The third lever is what you do in the half hour after lunch, and here the evidence is unusually practical.

In 2016, Andrew Reynolds and Jim Mann at the University of Otago ran a randomised crossover trial in Diabetologia with 41 people with type 2 diabetes, average age 60 [10]. For two weeks they walked 30 minutes a day whenever they liked; for another two weeks, ten minutes after each main meal. Total walking was the same. The post-meal timing lowered blood glucose after meals by 12 per cent overall and by 22 per cent after the evening meal.

It does not have to be a proper walk. A 2022 meta-analysis in Sports Medicine by Aidan Buffey and colleagues at the University of Limerick pooled seven crossover trials in which people either sat all day or broke the sitting with standing or light walking, typically two to five minutes every twenty or thirty minutes [11]. Standing produced a small reduction in post-meal glucose; light walking a moderate one, a standardised effect of about 0.7, with a similar cut in insulin. Moving muscle takes up glucose without waiting for insulin's instruction, which is why timing matters: the walk has to happen while the glucose is arriving.

One older lever keeps turning up in the folk record: vinegar. In 2004, Carol Johnston at Arizona State gave 20 grams of apple cider vinegar in water before an 87-gram-carbohydrate meal of bagel and orange juice to eleven people with insulin resistance, ten with type 2 diabetes and eight controls [12]. Insulin sensitivity for that meal improved by 34 per cent in the insulin-resistant group and by 19 per cent in the diabetic group, the latter short of significance; a 2017 meta-analysis of the trials concluded that vinegar with a meal can attenuate the glucose and insulin response [13]. Small groups, a single meal. But a dressing on the salad you eat first is not a hardship.

The coffee question

Most people solve three o'clock with caffeine, and the pharmacology is sound: it blocks the adenosine that has been building since morning. The catch is the half-life. In 2013, Christopher Drake's team at the Henry Ford sleep centre in Detroit gave twelve healthy sleepers 400 milligrams of caffeine, roughly two to three cups, at bedtime, three hours before or six hours before, on separate nights, against placebo [14]. All three timings disrupted sleep. Even the dose six hours before bed cut objectively measured sleep by more than an hour, and the participants did not notice. A three o'clock coffee for someone who sleeps at ten is inside that window. The next day's adenosine starts from a higher floor, the dip comes earlier, the coffee comes earlier, and a loop closes. We described the same loop from the stress side in The Tiger That Never Leaves; the afternoon is where the two meet.

What to do this week

None of this involves buying anything.

Accept the trough. Some of the three o'clock feeling is the clock, not the lunch [2][4]. Put the reading, writing and decisions before one o'clock, and the calls and errands after it.

Build the plate in the right order. Vegetables first, protein second, bread, potatoes and anything sweet last [9]. Make the middle course real, because protein and fat flatten the curve for the carbohydrate that follows [8]. A splash of vinegar on the first course is free [12].

Move for ten minutes after eating. Not a workout; a walk to the far end of the building, the post office, around the block. The glucose is arriving in the half hour after the meal and moving muscle takes it up directly [10][11]. Standing for the next call beats sitting; walking during it is better still.

Put the juice with the meal, not instead of it. A glass of juice on an empty stomach at three is fruit sugar with no fibre and nothing to slow it [7]. The same glass, small, 100 to 150 millilitres, drunk at the end of lunch after the salad and the protein, arrives into a meal already being digested slowly [8][9]. That is where we would put ours: Golden Blend, 82 per cent apple with ginger, turmeric and lemon, in place of a dessert; or Ruby Root, 70 per cent aronia and 30 per cent beetroot, tart rather than sweet, after a heavier meal. If what you want at three is a cold, clean liquid that is neither coffee nor sugar, Hello Aloe is 99.7 per cent aloe vera gel with a trace of citric acid, the closest thing in our range to water with a personality.

Draw the caffeine line. Count back six hours from bedtime and make that the last coffee [14]: five o'clock for a bedtime of eleven, three for a bedtime of nine. If you keep one afternoon coffee, make it the early one, straight after lunch, where it meets the trough rather than the pillow.

One line of caution: if you have diabetes, prediabetes or take medication that lowers blood glucose, the walking and vinegar findings come from trials in exactly those groups and are worth discussing with whoever manages your treatment before you change a routine.

The afternoon dip, in short

The 3 p.m. slump is partly circadian, a secondary rise in sleep pressure that happens even without lunch, and partly the meal. Continuous glucose monitoring shows post-meal dips are modest and individual but predict hunger. Eating vegetables and protein before carbohydrate, and walking ten minutes after eating, flatten the curve.

Back to the lake

The Loughborough drivers were sleepy before they ate; the lunch only decided how sleepy. That is the whole of it. The lake swim did not fix the afternoon in July because of anything in the water. It fixed it because you got up from the table, walked to the shore, moved for twenty minutes and came back to a body that had quietly handled its glucose while you were not paying attention. The meeting is not going anywhere. But the walk between the canteen and the meeting room is still yours, the salad still goes first, and the glass, if there is one, goes with the meal.

Frequently asked questions

Is the afternoon slump caused by blood sugar?

Only partly. A 2005 review found the post-lunch dip occurs even in people who have had no lunch, because sleep propensity has a secondary rise in the early afternoon [2][4]. A heavy lunch makes it worse: in a 12-person driving study, a 922-kilocalorie lunch caused more lane drifting than a 305-kilocalorie one [1].

Does a sugar crash really happen after lunch?

A dip below baseline is real but modest. In 1,070 people wearing glucose monitors, the average two-to-three-hour dip ranged from about 4 to 31 per cent below baseline depending on the meal, and the biggest dippers ate about 312 kilocalories more over the day than the smallest [6]. The study measured hunger, not sleepiness.

Does walking after eating lower blood sugar?

Yes, if it happens soon after the meal. In a two-week crossover trial of 41 people with type 2 diabetes, ten minutes of walking after each meal lowered post-meal glucose by 12 per cent compared with the same walking done at any time [10]. A 2022 meta-analysis of seven trials found even two-to-five-minute light walks helped [11].

Is fruit juice bad for blood sugar?

Alone on an empty stomach, it is the fastest form of fruit sugar: in a 1977 trial, apple juice in place of whole apples produced a rebound fall in glucose and higher insulin [7]. With a meal it behaves differently: eating vegetables and protein before carbohydrate cut the glucose rise by up to 37 per cent in one small trial [9].

References

1. Reyner LA, Wells SJ, Mortlock V, Horne JA. 'Post-lunch' sleepiness during prolonged, monotonous driving: effects of meal size. Physiol Behav. 2012;105(4):1088–1091. doi:10.1016/j.physbeh.2011.11.025
2. Monk TH. The post-lunch dip in performance. Clin Sports Med. 2005;24(2):e15–e23. doi:10.1016/j.csm.2004.12.002
3. Borbély AA, Daan S, Wirz-Justice A, Deboer T. The two-process model of sleep regulation: a reappraisal. J Sleep Res. 2016;25(2):131–143. doi:10.1111/jsr.12371
4. Bes F, Jobert M, Schulz H. Modeling napping, post-lunch dip, and other variations in human sleep propensity. Sleep. 2009;32(3):392–398. doi:10.1093/sleep/32.3.392
5. Berry SE, Valdes AM, Drew DA, et al. Human postprandial responses to food and potential for precision nutrition. Nat Med. 2020;26(6):964–973. doi:10.1038/s41591-020-0934-0
6. Wyatt P, Berry SE, Finlayson G, et al. Postprandial glycaemic dips predict appetite and energy intake in healthy individuals. Nat Metab. 2021;3(4):523–529. doi:10.1038/s42255-021-00383-x
7. Haber GB, Heaton KW, Murphy D, Burroughs LF. Depletion and disruption of dietary fibre: effects on satiety, plasma-glucose, and serum-insulin. Lancet. 1977;2(8040):679–682. doi:10.1016/S0140-6736(77)90494-9
8. Hätönen KA, Virtamo J, Eriksson JG, Sinkko HK, Sundvall JE, Valsta LM. Protein and fat modify the glycaemic and insulinaemic responses to a mashed potato-based meal. Br J Nutr. 2011;106(2):248–253. doi:10.1017/S0007114511000080
9. Shukla AP, Iliescu RG, Thomas CE, Aronne LJ. Food order has a significant impact on postprandial glucose and insulin levels. Diabetes Care. 2015;38(7):e98–e99. doi:10.2337/dc15-0429
10. Reynolds AN, Mann JI, Williams S, Venn BJ. Advice to walk after meals is more effective for lowering postprandial glycaemia in type 2 diabetes mellitus than advice that does not specify timing: a randomised crossover study. Diabetologia. 2016;59(12):2572–2578. doi:10.1007/s00125-016-4085-2
11. Buffey AJ, Herring MP, Langley CK, Donnelly AE, Carson BP. The acute effects of interrupting prolonged sitting time in adults with standing and light-intensity walking on biomarkers of cardiometabolic health in adults: a systematic review and meta-analysis. Sports Med. 2022;52(8):1765–1787. doi:10.1007/s40279-022-01649-4
12. Johnston CS, Kim CM, Buller AJ. Vinegar improves insulin sensitivity to a high-carbohydrate meal in subjects with insulin resistance or type 2 diabetes. Diabetes Care. 2004;27(1):281–282. doi:10.2337/diacare.27.1.281
13. Shishehbor F, Mansoori A, Shirani F. Vinegar consumption can attenuate postprandial glucose and insulin responses; a systematic review and meta-analysis of clinical trials. Diabetes Res Clin Pract. 2017;127:1–9. doi:10.1016/j.diabres.2017.01.021
14. Drake C, Roehrs T, Shambroom J, Roth T. Caffeine effects on sleep taken 0, 3, or 6 hours before going to bed. J Clin Sleep Med. 2013;9(11):1195–1200. doi:10.5664/jcsm.3170

This article is educational and is not medical advice. If you are pregnant, breastfeeding, taking prescribed medication or managing a diagnosed condition, speak to your doctor or pharmacist before making changes.