Effects of skipping breakfast, lunch or dinner on subsequent postprandial blood glucose levels among healthy young adults

Abstract

Meal skipping has become increasingly common in modern society due to changes in lifestyle. While the adverse effects of skipping breakfast on postprandial glucose levels have been well established, less is known about the impact of skipping lunch or dinner on these levels. The aim of this study is to determine the effect of skipping breakfast, lunch, and dinner consecutively on postprandial glucose levels in healthy subjects. Thirteen healthy young adults were enrolled and instructed to consume meals freely at designated times while maintaining detailed food intake records. Participants then followed a controlled protocol in which one of the three meals was skipped for two consecutive days in a specified order, with interstitial glucose continuously monitored. Skipping lunch on both days significantly increased postprandial glucose levels at the subsequent dinner by 1.6 mmol/L compared to when lunch was consumed and breakfast was skipped (p < 0.001). In contrast, skipping dinner or breakfast did not result in significant changes in postprandial glucose levels at the following meal. We thus observed skipping each of these three meals to have different effects on subsequent postprandial blood glucose levels in the healthy subjects. Most notably, skipping lunch leads to increased postprandial blood glucose levels at dinner.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12986-025-00975-4.

Introduction

When skipping breakfast, postprandial glucose levels rise after lunch in relation to pre-lunch plasma-free fatty acid elevation [1]. Regularly skipping breakfast carries risks of obesity [2]diabetes [3–5]cancer [6]cardiometabolic disease [7, 8]and impaired cognitive function [9]. Thus, eating breakfast is generally recognized as essential for health [10].

In recent years, habitually skipping lunch or dinner, not just breakfast, has risen among young people [11]. The number of skipped lunches rose concomitantly with increasing remote work during and after the COVID-19 pandemic [12]. However, relatively little is known about the postprandial glucose effects of lunch and dinner skipping. A previous study by Kuwahara et al. indicated that skipping lunch increased the postprandial glucose incremental AUC at dinner for 2 h by 1.5 to 2.3-fold in healthy adults. However, in that study, mealtime varied among the individuals, and the test meal consisted of dietary supplements, such as jelly drinks, protein powder, and so on [13]. The effect of dinner skipping on glucose homeostasis remains unexplored. We hypothesized that skipping lunch or dinner would lead to greater postprandial glucose excursions compared to skipping breakfast. Therefore, we aimed to investigate the effects of skipping lunch or dinner on postprandial glucose fluctuations in healthy adults under daily life conditions.

Materials and methods

Participants

This study was approved by the University of Shizuoka’s Ethics Committee (approval number 1–59). Inclusion criteria were as follows: (1) Subjects aged 20 years or older; (2) Subjects without diseases such as diabetes, hypertension, dyslipidemia, regularly taking medications or supplements affecting glucose metabolism; (3) Subjects implanted medical devices were recruited employing posted advertisements; (4) No family history of diabetes; (5) Not habitually skipping meals. Exclusion criteria included: (1) Inability to provide informed consent; (2) Pregnancy; (3) Current smoking; (4) Alcohol consumption.

Study design

This was an experimental, non-randomized, open-label trial. Alcohol consumption and exercise were prohibited for 2 days before starting the study until the end. As shown in Fig. 1, participants followed the protocol below: Day 1: Participants wore intermittently scanned continuous glucose monitoring (isCGM) devices (FreeStyle Libre Pro; Abbott, Chicago, IL, USA) on their upper arm. They consumed three meals at designated times (7 a.m., 1 p.m., and 7 p.m.). Day 2: Participants ate three meals at the designated times. isCGM data from this day were used as the control for the breakfast-skipping condition. Days 3–4: Participants skipped breakfast on both days. Days 5–6: Participants ate three meals at the designated times. isCGM data from Day 6 were used as the control for the lunch-skipping condition. Days 7–8: Participants skipped lunch on both days. Days 9–10: Participants ate three meals at the designated times. isCGM data from Day 10 were used as the control for the dinner-skipping condition. Days 11–12: Participants skipped dinner on both days. Day 13: Participants ate three meals at the designated times. Day 14: Participants removed the isCGM device.

Fig. 1.

Study protocol. The participants were fitted with isCGM on day 1 and these devices were removed on day 14 (dotted arrow). The interventions were skipping breakfast on the days 3–4, skipping lunch on the days 7–8 and skipping dinner on days 11–12 (bold arrow). A washout period of 2 days (days 5–6 or 9–10) was provided between each of these interventions. All participants consumed meals for breakfast, lunch and dinner daily at the fixed times of 7:00, 13:00, and 19:00 (rectangles)

Participants recorded and photographed their meal intakes during the study. Moreover, staple and high-carbohydrate (fruits and confectionaries) foods were weighed. Participants took pictures of each meal. Trained dietitians calculated nutritional values as energy or carbohydrate amount from each photo at study completion (Additional file 1–9: Table S1-S9).

Measurements

Glucose levels were acquired as interstitial glucose concentrations every fifteen minutes using the isCGM device. Premeal, postprandial peak and the peak from premeal (Δ Glucose) glucose were analyzed. Incremental areas under the curve for postprandial glucose levels for 3 h (3 h-iAUC) were also calculated using the trapezoidal method. Data from days 2, 6, and 10 served as controls. To adjust for dietary intake, the 3 h-iAUC was divided by energy (kcal) or carbohydrate (g) intake, resulting in the adjusted 3 h-iAUC (Adj-3 h-iAUC).

Statistical analysis

Referencing past reports on skipping breakfasts for Δ Glucose [1]the required sample number for enrollment was more than 7 subjects based on statistical software using a power calculation with an effect size of 39.8 mg/dL (postprandial glucose rise to peak), a standard deviation of 25.95, an alpha of 0.05 and a power of 0.80 to detect differences. Few studies are available, and the sample sizes for the skipping lunch or dinner conditions were the same as for the skipping breakfast condition.

We analyzed the full analysis set, excluding four participants due to sudden loss of isCGM function or mechanical errors. However, there were no differences in results between those completing and not completing the study.

The groups skipping meals were analyzed using the Wilcoxon matched-pairs signed rank test for comparison with each control. Data are means with standard deviation. All analyses were conducted using EZR software (Saitama Medical Center, Jichi Medical University, Saitama, Japan) [14]. Statistical significance was set at P < 0.05 (two-tailed).

Results

Thirteen subjects provided written informed consent. Their baseline characteristics are shown in Table 1.

Table 1.

Participant characteristics

Characteristics	Total (n = 13)
Age (years)	22.0	±	1.2
Sex, male/female (n)	1/12
Height (cm)	157.3	±	4.8
Body weight (kg)	52.1	±	6.4
Body mass index (kg/m²)	21.0	±	2.1
Dietary intakes
Energy (kcal/day)	1870	±	278
per body weight (kcal/kg・day)	36.1	±	4.9
Protein (g/day)	72.7	±	14.6
% of energy	15.6	±	2.4
Fat (g/day)	66.2	±	12.7
% of energy	32.0	±	4.7
Carbohydrate (g/day)	243	±	40
% of energy	52.5	±	5.3

Data are means ± standard deviations (SD) except for sex.

Table 2; Fig. 2 showed the isCGM transitions and glucose indices for each skipped meal. First, we analyzed the effects on the first day of each skipped meal. Compared to the control group, skipping lunch significantly increased peak glucose levels at dinner (7.16 ± 0.70 vs. 9.05 ± 1.23 mmol/L, p < 0.001, Table 2; Fig. 2). ΔGlucose at dinner was also significantly increased by skipping lunch (2.63 ± 0.77 vs. 4.24 ± 1.24 mmol/L, p < 0.001). After adjusting for energy or carbohydrate intake, skipping lunch still resulted in a significant increase in ΔGlucose.

Table 2.

Comparision of glucose indices after indicated meal with (Control) or without pre-meal (Skipping)

	Lunch (n = 13)						P value		Dinner (n = 12)						P value		Breakfast (n = 9)						P value
	Control			Skipping Breakfast					Control			Skipping Lunch					Control			Skipping Dinner
First day
Preprandial plasma glucose (mmol/L)	4.47	±	0.41	4.28	±	0.43	0.042*		4.52	±	0.37	4.81	±	0.49	0.092		4.39	±	0.57	4.35	±	0.37	0.905
	(4.25 to 4.69)			(4.04 to 4.52)					(4.31 to 4.73)			(4.53 to 5.09)					(4.08 to 4.7)			(4.15 to 4.55)
Peak glucose (mmol/L)	7.91	±	1.14	8.24	±	0.95	0.455		7.16	±	0.70	9.05	±	1.23	<0.001**		6.55	±	1.02	7.02	±	1.38	0.734
	(7.29 to 8.53)			(7.73 to 8.75)					(6.76 to 7.56)			(8.36 to 9.74)					(6.00 to 7.10)			(6.27 to 7.77)
ΔGlucose (mmol/L)	3.44	±	1.07	3.95	±	1.06	0.224		2.63	±	0.77	4.24	±	1.24	<0.001**		2.15	±	0.67	2.67	±	1.22	0.734
	(2.86 to 4.02)			(3.37 to 4.53)					(2.19 to 3.07)			(3.54 to 4.94)					(1.79 to 2.51)			(2.01 to 3.33)
/Carbohydrate (×10-2 mmol/L・g)	2.21	±	0.73	4.62	±	1.65	0.244		1.84	±	0.66	4.40	±	1.20	0.027*		1.90	±	0.89	3.39	±	1.35	1.000
	(1.81 to 2.61)			(3.72 to 5.52)					(1.47 to 2.21)			(3.72 to 5.08)					(1.42 to 2.38)			(2.66 to 4.12)
/Energy (×10-3 mmol/L・kcal)	2.71	±	0.88	6.57	±	2.96	0.094		2.16	±	0.50	6.03	±	2.06	0.002**		2.76	±	1.16	4.35	±	1.75	0.250
	(2.23 to 3.19)			(4.96 to 8.18)					(1.88 to 2.44)			(4.86 to 7.20)					(2.13 to 3.39)			(3.40 to 5.30)
3 h iAUC (mmol・min/L)	331	±	106	403	±	101	0.110		255	±	97	415	±	194	0.021*		153	±	47	258	±	135	0.004**
	(273 to 389)			(348 to 458)					(200 to 310)			(305 to 524)					(127 to 179)			(184 to 331)
/Carbohydrate (mmol/L・g)	2.10	±	0.67	4.63	±	1.31	0.048*		1.70	±	0.43	4.23	±	1.80	0.027*		1.32	±	0.54	3.22	±	1.47	0.074
	(1.74 to 2.46)			(3.92 to 5.34)					(1.46 to 1.94)			(3.21 to 5.25)					(1.03 to 1.61)			(2.42 to 4.02)
/Energy (×10-1 mmol/L・kcal)	2.59	±	0.84	6.58	±	2.49	0.011*		2.07	±	0.62	5.92	±	2.83	0.021*		1.92	±	0.70	4.03	±	1.58	0.496
	(2.13 to 3.05)			(5.23 to 7.93)					(1.72 to 2.42)			(4.32 to 7.52)					(1.54 to 2.30)			(3.17 to 4.89)
Second day†
Preprandial plasma glucose (mmol/L)	4.47	±	0.41	4.44	±	0.36	0.844		4.52	±	0.37	4.75	±	0.45	0.170		4.39	±	0.57	4.28	±	0.63	0.570
	(4.25 to 4.69)			(4.24 to 4.64)					(4.32 to 4.72)			(4.49 to 5.01)					(4.08 to 4.70)			(3.94 to 4.62)
Peak glucose (mmol/L)	7.91	±	1.14	8.00	±	1.33	0.753		7.16	±	0.70	9.08	±	1.03	<0.001**		6.55	±	1.02	6.69	±	1.25	0.910
	(7.29 to 8.53)			(7.27 to 8.73)					(6.78 to 7.54)			(8.50 to 9.66)					(6.00 to 7.10)			(6.01 to 7.37)
ΔGlucose (mmol/L)	3.44	±	1.07	3.56	±	1.24	0.588		2.63	±	0.77	4.33	±	0.89	<0.001**		2.15	±	0.67	2.40	±	1.17	0.624
	(2.86 to 4.02)			(2.89 to 4.23)					(2.21 to 3.05)			(3.83 to 4.83)					(1.79 to 2.51)			(1.76 to 3.04)
/Carbohydrate (×10-2 mmol/L・g)	2.21	±	0.73	4.94	±	2.82	0.146		1.84	±	0.66	4.48	±	1.37	0.027*		1.90	±	0.89	3.81	±	1.92	0.910
	(1.81 to 2.61)			(3.41 to 6.47)					(1.48 to 2.20)			(3.71 to 5.25)					(1.42 to 2.38)			(2.76 to 4.86)
/Energy (×10-3 mmol/L・kcal)	2.71	±	0.88	5.70	±	2.94	0.305		2.16	±	0.50	5.72	±	1.73	0.005**		2.76	±	1.16	4.26	±	2.12	0.570
	(2.23 to 3.19)			(4.10 to 7.30)					(1.89 to 2.43)			(4.74 to 6.70)					(2.13 to 3.39)			(3.11 to 5.41)
3 h iAUC (mmol・min/L)	331	±	106	360	±	158	0.497		255	±	97	491	±	143	<0.001**		153	±	47	214	±	121	0.164
	(273 to 389)			(275 to 446)					(202 to 308)			(410 to 572)					(127 to 179)			(148 to 280)
/Carbohydrate (mmol/L・g)	2.10	±	0.67	4.84	±	2.61	0.216		1.70	±	0.43	5.09	±	1.92	<0.001**		1.32	±	0.54	34.23	±	21.22	0.359
	(1.74 to 2.46)			(3.42 to 6.26)					(1.47 to 1.93)			(4.00 to 6.18)					(1.03 to 1.61)			(22.70 to 45.76)
/Energy (×10-1 mmol/L・kcal)	2.59	±	0.84	5.73	±	3.34	0.305		2.07	±	0.62	6.50	±	2.56	<0.001**		1.92	±	0.70	3.79	±	2.24	0.820
	(2.13 to 3.05)			(3.92 to 7.54)					(1.73 to 2.41)			(5.05 to 7.95)					(1.54 to 2.30)			(2.57 to 5.01)

Values are means ± SD (95% confidence interval). Asterisks indicate statistical significance (*P < 0.05; **P < 0.01) of Skipping compared to Control at each meal.

iAUC; incremental area under the curve

† Pre-meals were consumed or not two days in a row

Fig. 2.

Glucose profiles on the second days. The glucose profiles for the second days of skipping breakfast (A  n = 13; black bold line), lunch (B  n = 12; gray line) or dinner (C  n = 9; light gray line) and on each subsequent meal were taken as the Control (A  n = 13, B, n = 12 C, n = 9; black line) for the 21-hour period from each timepoint of skipping a meal, and are shown for measurments taken every fifteen minutes. Data are presented as means ± SD. All subjects ate breakfast at 7:00 (black triangle), lunch at 13:00 (white triangle), dinner at 19:00 (gray triangle)

Similarly, the 3 h-iAUC at dinner was significantly elevated after skipping lunch (255 ± 97 vs. 415 ± 194 mmol·min/L, p = 0.021), and this effect remained even after adjustment for energy or carbohydrate intake (Adj-3 h-iAUC). Skipping breakfast also increased the Adj-3 h-iAUC at lunch.

Next, we analyzed the effects on the second day of each skipped meal. Skipping lunch again significantly increased peak glucose, ΔGlucose, and 3 h-iAUC at dinner, consistent with the findings on the first day (Table 2; Fig. 2). In contrast, skipping breakfast or dinner did not significantly affect any postprandial glucose parameters at lunch or dinner (Table 2; Fig. 2).

Discussion

This study showed for the first time that skipping lunch impairs postprandial glucose levels at dinner despite an identical or optimal diet and even with identical mealtimes in healthy adults (Table 2; Fig. 2B). This study was conducted under real-life meal conditions with a fixed dinner time, in contrast to previous studies involving unique dietary supplements or varying dinner times [13]. Thus, to date, the mechanism underlying this phenomenon remains unknown. However, extensive research has shown that a hyperglycemic reaction occurs when ingesting lunch after skipping breakfast in both healthy [1, 15] and diabetic individuals [16, 17].

A recent study showed that the adverse effect of breakfast skipping on postprandial glucose is limited to the first day [18]supporting the post-lunch glycemic reactions on the second day without breakfast (left lower section of Table 2). Repeatedly skipping breakfast produced glucose adaptation [18]but the details remain unknown. Interestingly, this glucose adaptation did not follow skipping lunch (center lower section of Table 2). Such differences also have not been clarified and we plan to conduct further relevant studies in the near future. We first demonstrated that skipping dinner did not affect blood glucose levels after breakfast, which was the opposite of our expectations based on several known influences of nocturnal glucose regulation [19, 20]. This mechanism might be explained by glucose homeostasis associated with eating breakfast [21–23] or by decreased glycogen storage with starvation. Future studies should focus on the mechanism whereby skipping different meals has variable physiological impacts.

This study has limitations. First, the subjects were limited sample size, young healthy people, neither elderly nor obese. Therefore, studies involving more diverse populations would be needed. Additionally, sample size was calculated based only on the skipping breakfast condition. Second, the subjects ate freely during the intervention period. Thus, dietary intakes differed, though staple foods, fruits, and confectionaries were weighed and the data were then corrected by energy or carbohydrate intake. Our trial was under normal conditions; however, identical diets would be optimal for more experimental intervention. Third, most participants skipped meals in the same sequence (breakfast → lunch → dinner). However, since the subjects in this study were young and healthy individuals without conditions such as diabetes, and a two-day washout period was implemented between each meal skipping condition, we consider the effect of the meal skipping order to be minimal.

In conclusion, we demonstrated different reactions to skipping each of the three main daily meals on subsequent postprandial blood glucose levels. Most notably, this was the first study to show that skipping lunch leads to deterioration of postprandial blood glucose levels at dinner. Thus, lunch intake was crucial for regulating glucose properly. 26.1% of US adults have skipping lunch [24]our findings provide evidence supporting the recommendation to “eat three times a day” in dietary guidance for diabetes prevention. Assuming a relative risk of 1.12 ²⁴, skipping lunch may account for approximately 3% of new diabetes cases in the U.S., equivalent to over 30,000 cases annually. The long-term effects of postprandial glucose improvement remain unknown, as we were unable to analyze the underlying molecular mechanisms. Further studies should be needed to elucidate the mechanisms related to meal skipping and its impact on blood glucose regulation and long-term health outcomes.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

isCGM

intermittently continuous glucose monitoring

ΔGlucose

postprandial peak and the peak from premeal

3h-iAUC

Incremental areas under the curve for postprandial glucose levels for 3 h

Adj-3h-iAUC

incremental area under the curve, adjusted for 3 h for energy or carbohydrate intake

Author contributions

Study design/planning: C.K., T.H.; data collection/entry: C.K., S.T., F.N., T.H.; data analysis/statistics: C.K.; data interpretation: C.K., Y.S., T.H.; preparation of the first draft version: C.K., Y.S., T.H. All authors have approved the manuscript and agreed to its submission.

Funding

This work was partially supported by JSPS KAKENHI Grant Number 24K05553.

Data availability

All data generated or analysed during this study are included in this published article and its supplementary information files.

Declarations

Ethics approval and consent to participate

Approval of the research protocol: The study protocol was approved by the Ethics Committee of the University of Shizuoka (approval no. 1–59).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Informed consent

Written informed consent was obtained from all participants in accordance with the Declaration of Helsinki.