Effect of coffee bean chlorogenic acid on peripheral skin temperature and skin blood flow in healthy Japanese women with cold sensitivity: a placebo-controlled, randomized, double-blind, study

Abstract

Background

Cold sensitivity is a common health problem for women in Japan and in East Asia. Cold sensitivity is characterized by a low peripheral skin temperature (SkT) in a cold environment and a slow recovery of the peripheral SkT after cold exposure. Ingestion of coffee chlorogenic acids (CGAs) has been reported to improve the recovery of peripheral SkT and skin blood flow (SkBF) after cold stress. In this study, we investigated the acute effects of a single dose of CGA on peripheral SkT and SkBF in a cold environment.

Methods

This randomized, double-blind, placebo-controlled, crossover study was conducted from February to March 2018. Twenty-three healthy Japanese women with cold sensitivity were randomly assigned to the CGA (270 mg) or the Placebo group in a crossover study with a one-week washout period (11 took CGA first and 12 took the Placebo first). Peripheral SkT (primary outcome) and SkBF (secondary outcome) were evaluated in a moderate cold environment (22 °C). The study was registered with the University Hospital Medical Information Network Clinical Trials Registry (UMIN000031184).

Results

Twenty-one participants completed all crossover periods. The crossover analysis approach showed a significant period effect for SkT (p = 0.023) and therefore, only the first period data were analyzed. The ingestion of coffee bean CGA significantly suppressed the decrease in SkT (LS mean difference 1.18 °C, 95% confidence interval 0.16 to 2.19, F (1, 277) = 5.19, p = 0.023) and SkBF (LS mean difference 0.87 a.u., 95% confidence interval 0.41 to 1.69, F (1, 278) = 4.27, p = 0.040) during cold exposure. No adverse events attributed to the test beverage were observed.

Conclusion

These results suggest that the ingestion of coffee bean CGA can suppress the decrease in SkT and SkBF in a cold environment.

Trial registration

The name of this study is ‘Examination of drink ingestion containing plant extract. NT-18002’. The study was registered with the University Hospital Medical Information Network Clinical Trials Registry. UMIN-CTR (https://center6.umin.ac.jp/cgi-open-bin/ctr_e/ctr_view.cgi?recptno=R000035601) UMIN000031184. Registered on February 7, 2018.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13063-026-09928-5.

Introduction

Cold sensitivity, called “hiesho” in Japan, is commonly seen in women in East Asia [1, 2]. Cold sensitivity is characterized by a low skin temperature (SkT) in various parts of the body, especially in the distal extremities, that feels cold throughout the seasons, and by a slow recovery from a decrease in SkT due to cold exposure. Cold sensitivity impairs a healthy life and is one possible cause of a deterioration in the Quality of Life. Therefore, research to improve cold sensitivity is important. The low SkT in cold sensitivity is caused by an insufficient skin blood flow (SkBF), which indicates that an increase in SkBF is essential to improve cold sensitivity. Many types of treatments have been reported, such as bathing [3], herbal medicine [4], and various food ingredients [5–9], all of which improve SkBF.

Coffee is one of the most popular beverages in the world and coffee beans contain an abundance of polyphenols. The major polyphenol in coffee beans is chlorogenic acid (CGA), and the major molecular species include caffeoylquinic acids (CQAs), dicaffeoylquinic acids (diCQAs), feruloylquinic acids (FQAs), and diferuloylquinic acids (diFQAs) [10]. Coffee bean CGA has been shown to have various physiological effects such as antioxidant, anti-inflammatory and antibacterial effects [11].

Previous studies have reported that a continuous intake of coffee bean CGA reduces the blood pressure of adults with grade I hypertension [12] and enhances vascular responsiveness to local warming [13]. A single dose of coffee bean CGA ameliorates postprandial vascular endothelial dysfunction [14].

Preventing the decrease of SkT in a cold environment and enhancing the decreased recovery of SkT after cold exposure through an increase in SkBF is important to improve cold sensitivity. Previous studies have shown that a single ingestion of coffee bean CGA promotes the recovery of decreased SkT after peripheral cold stress [15, 16] through an improved recovery of SkBF [15]. However, the acute effect of a single dose of coffee bean CGA to prevent a decrease of SkT in a cold environment had not been reported. We conducted a placebo-controlled, randomized, double-blind, crossover study to determine whether a single dose of coffee bean chlorogenic acids (CGAs) attenuate cold induced decreases in SkT (primary outcome) and SkBF (secondary outcome).

Materials and methods

Participants

Healthy Japanese women who met the inclusion criteria and did not meet the exclusion criteria (Table 1) were recruited via an online recruitment website targeting residents of Tokyo and the surrounding metropolitan area.

Table 1.

Inclusion criteria and exclusion criteria for the selection of study participants

Inclusion criteria
(1)	Healthy females aged 20–35 years
(2)	Those who are sensitive to cold
(3)	Those whose body mass indexes are from 18.5 to 25
(4)	Those who have a stable menstrual cycle
(5)	Those who observe both drinking up to one cup of coffee per day and drinking up to four cups of coffee per week
Exclusion criteria
(1)	Those who have a severe allergic disease such as rhinitis and/or atopic dermatitis
(2)	Participants with a vascular function disorder
(3)	Current smoker

Forty-seven participants diagnosed as healthy by a physician underwent a cold stress test as a screening test, which was performed using a modified method as previously described [17]. Both hands of each participant were immersed in cold water at 15 °C for 1 min up to the wrist joints. Thermal images of the left dorsal hand were captured using a thermograph (InfReC R500S, Nippon Avionics, Co., Ltd., Tokyo, Japan) just before (baseline) and immediately after the immersion of both hands, and 5 and 10 min later. The recovery rate of SkT was calculated as follows: Recovery rate = (T₁₀-T₀)/(T_BL-T₀) × 100(%), where T_BL = baseline SkT; T₀ = SkT immediately after cold stress; T₁₀ = SkT 10 min after cold stress. Twenty-three participants with an SkT recovery rate ≤ 50% were selected (ranked in descending order).

To standardize measurement conditions, participants were instructed to maintain their usual diet and physical activity during the study period and to refrain from using dietary supplements or functional foods (including products marketed for beauty purposes). Participants were also instructed to limit coffee (including decaffeinated coffee) and coffee-based beverages to no more than one cup per day and no more than four cups per week, and to avoid excessive consumption of spices (e.g., red pepper and ginger) and caffeine-rich foods. For the 3 days before each test day, participants were instructed to abstain from alcohol, avoid overeating and sleep deprivation, and refrain from strenuous exercise or overwork. Adherence was monitored using daily logs completed throughout the study.

Of the 23 participants enrolled, one withdrew for personal reasons and one withdrew due to a common cold. The remaining 21 participants completed the study and were included in the analysis (Fig. 1).

Fig. 1.

Flow diagram of the study. CGA: beverage with coffee bean CGA. Placebo: beverage without coffee bean CGA

Study design

This study was a randomized, double-blind, placebo-controlled, crossover study, with a 1-week washout period between the intake of the two test beverages (Fig. 2). The study was conducted by TES Holdings Co., Ltd. (Tokyo, Japan) from February 2018 to March 2018 in accordance with the Declaration of Helsinki. The data publication followed the CONSORT 2025 checklist [18, 19].

Fig. 2.

Study protocol. SkT, skin temperature; SkBF, skin blood flow

Eligible participants were randomized in a 1:1 allocation to one of two groups. Randomization was performed by a researcher using a stratified block allocation of the SkT recovery rate after cold stress (block size of 2). A study-independent manager allocated the two groups to Group A or Group B, then archived the allocation results and did not disclose them until the data were locked for analysis.

According to Mori’s guideline for a “hiesho” study [20], Period 1 and Period 2 were conducted in the same phase of the menstrual cycle of each participant. Menstrual-cycle phase was determined based on self-reported information on the first day of the last menstrual period. As a screening test, each participant underwent a cold stress test during the estimated follicular phase, based on the first day of their most recent menstruation. Period 1 of the main test was conducted early in the estimated luteal phase following this follicular phase, while Period 2 of the main test took place after a one-week washout period. The two main tests for each participant were administered at the same time on each testing day. Participants took only water after 21:00 on the day before the test. On the test day, each participant ate the provided breakfast 2 h before the start of the measurement. To avoid the effects of clothing, each participant wore a provided outfit during the screening test and the two main tests, including underwear and coats. Each participant changed into a provided half sleeve T-shirt and long pants and wore the provided slippers on her bare feet while waiting in a room maintained at 25 °C before entering the measurement room.

Test beverages

Two types of test beverages were used: a CGA beverage (100 mL) containing coffee bean CGA and a Placebo beverage (100 mL). Coffee bean CGA was extracted from green coffee beans and caffeine was removed using activated charcoal [21]. The composition of coffee bean CGA was analyzed by high-performance liquid chromatography, which identified CQAs (5-caffeoylquinic acid, 3-caffeoylquinic acid, 4-caffeoylquinic acid), FQAs (5-feruloylquinic acid, 3-feruloylquinic acid, 4-feruloylquinic acid) and diCQAs (3,4-dicaffeoylquinic acid, 3,5-dicaffeoylquinic acid, 4,5-dicaffeoylquinic acid). The CGA beverage was prepared to ensure a total amount of 270 mg of CQAs and FQAs, the same dosage used in the single-dose studies that significantly enhanced the recovery of SkT and SkBF after cold stress tests [15, 16]. The CGA beverage was blended with sweeteners, acidifiers, and flavors. The Placebo beverage, which did not contain coffee bean CGA, was formulated to have a similar taste to the CGA beverage by using the same sweeteners, acidifiers, and flavors. The test was conducted without the knowledge of either the drink provider or the participant as to which drink was ingested.

Cold exposure test

The cold exposure test was performed by modifying the method of Takumi et al. [22], which involved participants who are sensitive to cold. After the participant entered the measurement room, a probe for the SkT measurement was attached under the nail of the middle finger of the left hand, and the SkT measurement was started. After resting for 10 min in a sitting position, each participant put nitrile gloves on both their hands and placed them in a water bath at 33 °C for 10 min to make the hand SkT similar between all participants. After bathing, the gloves on both hands were removed and a probe for measuring the SkBF was attached to the lower part of the first joint of the middle finger of the left hand, and the measurement of SkBF was started. After resting for 10 min in a sitting position, each participant ingested the test beverage, and then remained in a sitting position for 70 min. During the 70 min measurement, the participant placed both hands on a table with the palm side of the left hand with the two probes facing downward. All measurements were performed at a room temperature of 22 °C ± 1 °C and a humidity of 50% ± 10%.

Measurement equipment

SkT was measured using a thermo logger (LT-200S, Gram Corporation, Saitama, Japan) and a SkT sensor (LT-2N-12, Gram Corporation, Saitama, Japan). SkBF was measured using a laser Doppler blood flowmeter (FLO-C1, Omegawave, Inc., Tokyo, Japan) and a single-point LDF probe (DS Type, Omegawave, Inc., Tokyo, Japan). All SkT and SkBF data were acquired using PowerLab (ADInstruments, Otago, New Zealand) and LabChart (ADInstruments, Otago, New Zealand) at a sampling rate of 10 Hz. SkT and SkBF were summarized as simple 5-min mean values (3000 data points per interval), and no filtering or smoothing was applied.

Statistical analysis

Sample sizes were estimated based on an unpublished preliminary study using R package TrialSize (ver. 3.4.0.) (the R Project for Statistical Computing. https://cran.r-project.org/). The required sample size was calculated as 16 participants, assuming an effect size of 1.0 °C (SD 1.5 °C) for SkT and 1.0 a.u. (SD 1.8 a.u.) for SkBF, with a two-sided α = 0.05 and 80% power (β = 0.20), which is commonly adopted in clinical and biostatistical research. However, considering potential dropouts, the target number of participants in this study was set at 24.

Statistical analysis was performed according to the Consolidated Standards of Reporting Trials statement [23] and Wellek and Blettner [24]. The primary outcome was the change in SkT and the change in SkBF was the secondary outcome. For SkT and SkBF, the average values were calculated every 5 min from 5 min before to 70 min after ingestion of the test beverage. The average values of SkT and SkBF for 5 min before ingestion of the test beverage were used as baselines. Changes in SkT and SkBF were compared for 70 min after ingestion of the test beverage.

Between-beverage differences in changes in SkT and SkBF after intake of the test beverages were assessed using a linear mixed-effects model. For the prespecified crossover analysis, fixed effects included beverage, time (measurement time point), period, sequence (order of ingestion), and the beverage-by-time interaction. Participants were modeled using a random intercept to account for between-subject heterogeneity in the overall outcome level. To model the within-participant correlation across the repeated time points, a first-order autoregressive covariance structure (AR(1)) was specified for residuals (14 time points per participant). For SkT, baseline SkT and baseline SkBF were included as covariates; for SkBF, baseline SkBF was included as a covariate. If the interaction term was not statistically significant, it was removed and the model was refitted to obtain pooled estimates.

If a significant period and/or sequence effect was observed, the prespecified analysis was restricted to Period 1 data and conducted as a parallel-group comparison. The Period 1 analysis used linear mixed-effects model with test beverage, time (measurement time point), and the beverage-by-time interaction as fixed effects and participant as a random intercept, with an AR(1) covariance for within-participant repeated measures; interaction handling and covariate adjustment followed the crossover analysis.

We also evaluated models including a participant-specific random slope for time; however, model fit was not improved based on information criteria (AIC/BIC), and estimated treatment effects were unchanged. Therefore, the random-intercept-only model was used. The significance level was less than 5% on both sides. All statistical analyses were performed using SPSS Statistics ver. 26 (IBM Corp., Armonk, NY, USA).

Results

Effect of coffee bean CGA on the SkT and SkBF decreases

Twenty-one participants completed all crossover periods. The crossover analysis approach showed a significant period effect for SkT (p = 0.023) and therefore, only Period 1 data were analyzed to compare the two test beverages (Table 2). A significant order effect was also observed for SkBF (p < 0.001) (Table 2).

Table 2.

Effects of coffee bean derived CGA on peripheral SkT and SkBF in a cold environment

Time after ingestion (min)	SkT (℃)				SkBF (a.u.)
	Period 1		Period 2		Period 1		Period 2
	CGA	Placebo	CGA	Placebo	CGA	Placebo	CGA	Placebo
5	31.4 ± 0.84	30.5 ± 0.80	30.1 ± 0.47	29.7 ± 0.49	4.27 ± 0.70	3.46 ± 0.67	3.49 ± 0.39	3.89 ± 0.41
10	31.0 ± 0.84	29.6 ± 0.80	29.1 ± 0.47	28.9 ± 0.49	4.41 ± 0.70	3.34 ± 0.67	3.75 ± 0.39	3.73 ± 0.41
15	31.0 ± 0.84	29.9 ± 0.80	29.5 ± 0.47	28.7 ± 0.49	4.52 ± 0.70	3.75 ± 0.67	4.31 ± 0.39	4.34 ± 0.41
20	30.9 ± 0.84	30.3 ± 0.80	30.1 ± 0.47	28.9 ± 0.49	4.62 ± 0.70	4.22 ± 0.67	4.11 ± 0.39	4.47 ± 0.41
25	31.0 ± 0.84	30.1 ± 0.80	29.7 ± 0.47	29.2 ± 0.49	4.63 ± 0.70	3.88 ± 0.67	3.94 ± 0.39	4.36 ± 0.41
30	30.8 ± 0.84	30.3 ± 0.80	29.5 ± 0.47	29.3 ± 0.50	4.55 ± 0.70	3.89 ± 0.67	4.50 ± 0.39	4.54 ± 0.41
35	30.7 ± 0.84	30.3 ± 0.80	29.8 ± 0.47	29.4 ± 0.50	4.42 ± 0.70	3.93 ± 0.67	3.69 ± 0.39	4.65 ± 0.41
40	30.5 ± 0.84	29.7 ± 0.80	29.3 ± 0.47	29.4 ± 0.49	4.50 ± 0.70	3.45 ± 0.67	3.64 ± 0.39	4.20 ± 0.41
45	30.6 ± 0.84	29.1 ± 0.80	29.2 ± 0.47	28.9 ± 0.49	4.43 ± 0.70	3.19 ± 0.67	3.47 ± 0.39	4.05 ± 0.41
50	30.4 ± 0.84	28.5 ± 0.80	29.0 ± 0.47	28.4 ± 0.49	4.06 ± 0.70	2.78 ± 0.67	3.40 ± 0.39	3.84 ± 0.41
55	29.7 ± 0.84	27.7 ± 0.80	28.9 ± 0.47	28.1 ± 0.49	3.80 ± 0.70	2.74 ± 0.67	2.82 ± 0.39	3.73 ± 0.41
60	29.1 ± 0.84	27.3 ± 0.80	28.3 ± 0.47	27.8 ± 0.49	3.47 ± 0.70	2.82 ± 0.67	2.82 ± 0.39	3.69 ± 0.41
65	28.5 ± 0.84	27.1 ± 0.80	28.1 ± 0.47	27.5 ± 0.49	3.29 ± 0.70	2.36 ± 0.67	2.63 ± 0.39	3.28 ± 0.41
70	27.2 ± 0.84	26.9 ± 0.80	27.7 ± 0.47	26.9 ± 0.49	3.35 ± 0.70	2.37 ± 0.67	2.22 ± 0.39	3.22 ± 0.41
Order effect	p = 0.524				p < 0.001
Period effect	p = 0.023				p = 0.072

Data are presented as LS means ± SEM; Period 1 n = 10 in the CGA group; n = 11 in the Placebo group. Period 2 n = 11 in the CGA group; n = 10 in the Placebo group. P values for order and period effects were derived from fixed-effects tests in the linear mixed-effects crossover model. SkT skin temperature, SkBF skin blood flow. CGA: beverage with coffee bean CGA. Placebo: beverage without coffee bean CGA

There was no significant difference in baseline SkT, baseline SkBF or other background data for participants between the CGA and Placebo groups in Periods 1 and 2 (Table 3).

Table 3.

Characteristics of the study participants

Characteristic	Period 1			Period 2
	CGA	Placebo	p value	CGA	Placebo	p value
Number of participants	10	11	−	11	10	−
Age (year)	25.9 ± 5.5	27.5 ± 4.8	0.473	−	−	−
Height (cm)	160.3 ± 5.9	159.2 ± 5.3	0.594	−	−	−
Weight (kg)	52.7 ± 6.2	50 ± 4.6	0.156	−	−	−
Body mass index (kg/m2)	20.5 ± 1.4	19.7 ± 0.9	0.057	−	−	−
Systolic blood pressure (mmHg)	99.1 ± 8.9	99.6 ± 9.3	0.895	101.6 ± 8.5	98.8 ± 8.0	0.443
Diastolic blood pressure (mmHg)	55.1 ± 5.6	55.2 ± 9.2	0.982	58.5 ± 8.1	55.7 ± 9.7	0.790
Heart rate (beats/min)	69.2 ± 6.2	67.7 ± 10	0.692	69.7 ± 10.4	69.4 ± 5.7	0.931
Baseline SkT (℃)	31.4 ± 1.7	31.9 ± 1.2	0.499	30.5 ± 2.6	30.8 ± 2.2	0.750
Baseline SkBF (a.u.)	4.5 ± 2.5	4.5 ± 1.3	0.950	4.5 ± 1.9	4.9 ± 3.0	0.696

Each value represents the mean ± SD. SkT skin temperature, SkBF skin blood flow. Baseline SkT and SkBF are average values at 5 min before the test beverage ingestion. CGA: beverage with coffee bean CGA. Placebo: beverage without coffee bean CGA

Figure 3A shows changes in SkT over time during cold exposure in Period 1 for the CGA and placebo groups. SkT was analyzed using a linear mixed-effects model with the beverage-by-time interaction term removed because no significant interaction was observed (p = 0.998). The decrease in SkT was significantly attenuated in the CGA group compared with the placebo group (LS mean difference, 1.18 °C; 95% CI, 0.16 to 2.19; F (1, 277) = 5.19; p = 0.023).

Fig. 3.

Effect of coffee bean CGA on SkT and SkBF in healthy women with cold sensitivity (Period 1 analysis). A SkT. P value, beverage: 0.023, time: <0.001, B SkBF. P value, beverage: 0.040, time: 0.172, CGA: beverage with coffee bean CGA. Placebo: beverage without coffee bean CGA. Data are presented as LSmeans ± SEM; n = 10 in the CGA group; n = 11 in the placebo group. Changes over time and differences between beverages were analyzed using a linear mixed-effects model

Figure 3B shows changes in SkBF during cold exposure in Period 1 for the CGA and placebo groups. SkBF was analyzed using a linear mixed-effects model with the beverage-by-time interaction term removed because no significant interaction was observed (p = 0.999). SkBF changes over time showed a trend similar to that observed for SkT, and the decrease in SkBF was significantly attenuated in the CGA group compared with the placebo group (LS mean difference, 0.87 a.u.; 95% CI, 0.41 to 1.69; F (1, 278) = 4.27; p = 0.040).

For the Period 1 analyses, within-participant correlation across repeated time points was modeled using an AR(1) covariance structure, and participant was included as a random intercept. We also compared a random-intercept-only model with a random intercept plus random slope for time model. Model fit for SkT data favored the random-intercept-only specification (Model 1: AIC = 1356, BIC = 1360; Model 2: AIC = 1360, BIC = 1371), and the estimated treatment effect (LS mean difference, 95% CI, F value, and p value) was identical across models; therefore, results from the random-intercept-only model are reported.

Safety

No adverse events attributable to the test beverage were observed.

Discussion

The purpose of this study was to evaluate the preventive effects of coffee bean CGA on the decrease of SkT and SkBF of healthy Japanese women with cold sensitivity in a cold environment. In this study, a significant period effect was observed for the primary outcome, SkT. This indicates that the results in the second period were influenced by the timing of the trials, making it impossible to evaluate the effects of the intervention. Therefore, only data from the first period were analyzed.

The results obtained from the first phase analysis showed that ingestion of the CGA beverage significantly reduced the decrease in SkT and SkBF in the cold environment compared to the Placebo beverage (Fig. 3). However, restricting the analysis to Period 1 reduced the effective sample size and statistical power (63% for SkT and 54% for SkBF). Therefore, these findings should be interpreted cautiously.

SkT is determined by the supply of heat from SkBF and the heat loss from the skin, which depends on the environmental temperature [25]. In the present study, the environmental temperature was set at 22 °C, which means that the heat loss through the skin was the same in both the CGA and the Placebo groups. Therefore, it is clear that the significant suppression of the SkT decrease in the CGA group was due to the significant suppression of SkBF reduction.

The SkBF contributes to core body temperature homeostasis by regulating heat dissipation from the body surface and a decrease in SkBF in a cold environment is caused by vasoconstriction via sympathetic nerves. On the other hand, it has been reported that a reduced SkBF in a cold environment was partially due to the reduction of endothelial nitric oxide (NO) [26] vasodilation based on the frequency analysis of SkBF during whole-body cooling [27]. In fact, it has been shown that the decrease in SkBF caused by cold exposure is suppressed by increasing NO in the blood through the intradermal administration of sodium nitroprusside [28] and by improving NO bioavailability through the subcutaneous administration of reactive oxygen species (ROS) scavengers [29]. It has also been reported that coffee bean CGA significantly improves postprandial flow-mediated dilation (FMD) by reducing oxidative stress in the blood and improving NO bioavailability through its ROS scavenging activity [14, 30]. Coffee bean CGA could be detected in the blood flow 30 min after ingestion and its Tmax (time for maximum blood concentration) is generally between 1 and 4 h, although that varies with molecular species and individual differences [31, 32]. Therefore, the suppression of the decrease in SkBF during cold exposure observed in the present study could be attributed to the improvement of NO bioavailability elicited by coffee bean CGA.

In the Placebo group, a transient decrease in SkT was observed up to 20 min after beverage consumption (Fig. 3A). This transient decrease in SkT was clearly due to a temporary decrease in SkBF (Fig. 3B). The contribution of SkBF to the homeostasis of core body temperature is one of its important roles. Therefore, when the environmental temperature is constant, as in this study, a transient decrease in SkBF indicates a temporary drop in internal body temperature. Since the test beverage in the measurement room was the same as the measurement room temperature (22 °C), it is believed that the ingestion of this test beverage caused a temporary decrease in internal body temperature. On the other hand, no transient decrease in SkBF was observed in the CGA group. Considering the transit time of coffee bean CGA into the blood, it is unlikely that the bioavailability of NO in the blood would improve immediately after ingestion of the test beverage. The mechanism underlying this between group difference remains unclear. Future studies using more frequent measurements during the early post-ingestion period (e.g., 1-min intervals rather than 5-min intervals within the first ~20 min) may help characterize short-lived SkBF dynamics.

The resting blood pressure of the participants in both the CGA and Placebo groups met the definition of hypotension, with systolic blood pressure (SBP) less than 110 mmHg and diastolic blood pressure (DBP) less than 60 mmHg. It is currently unclear whether there is a difference in the effect of coffee bean CGA on normal and hypotensive participants. However, since the proportion of hypotensive participants was similar in both the CGA and Placebo groups (approximately 40%), it is believed that the presence of hypotensive participants did not introduce bias into the results of this study. Previous work suggests that women with cold sensitivity have lower blood pressure than women without cold sensitivity, and hypotension has been considered a contributing factor [33]. Thus, the present findings may be applicable to women with cold sensitivity.

This study has several limitations. First, a significant period effect was observed for the primary outcome, which precluded valid within-participant comparisons and necessitated restriction to Period 1 data, resulting in loss of Period 2 information. Therefore, the study does not support definitive conclusions regarding treatment switching or longer-term time-dependent changes. Second, the reduced analyzable sample size decreased statistical power and widened confidence intervals. Moreover, with small sample sizes relative to the number of time points and covariates, repeated-measures models may be susceptible to overfitting and unstable estimates. Accordingly, these findings should be interpreted as exploratory.

A period effect is defined as “a change in outcome over time independent of treatment effect” [23]. Considering that the test period was 13 days from the beginning of Period 1 to the end of Period 2, it is difficult to conclude that the difference in season caused a significant period effect. However, the average temperature for the last 7 days before Period 2 was significantly higher than that before Period 1 (Appendix 1 shows this in more detail [see Additional file 1]), indicating that this difference in average temperature possibly affected the physiological response to the measurement room temperature. In fact, the baseline SkT before the ingestion of the test beverage tended to be lower in Period 2 than in Period 1 (Appendix 2). Comparing the resting DBP in Periods 1 and 2, the DBP in Period 2 tended to be higher than in Period 1 (Appendix 2 shows this in more detail [see Additional file 1]). On the other hand, in a previous study conducted in the winter [15], no period effect was observed, as there was no significant difference between the average temperature for the last 7 days before Period 1 and Period 2 (unpublished data, Appendix 3 shows this in more detail [see Additional file 1]). Additionally, there were no significant differences in baseline SkT or resting DBP between Period 1 and Period 2 (unpublished data, Appendix 4 shows this in more detail [see Additional file 1]). These data suggest that physiological responses to the cold environment (22 °C) for 30 min before ingestion of the test beverage tended to differ between Period 1 and Period 2. Therefore, a parallel group trial would be preferred for efficacy evaluation targeting changes in SkT and SkBF in a cold environment.

SkBF showed a significant sequence effect, raising the possibility of carryover. However, carryover appears unlikely because orally ingested coffee bean CGAs reportedly are no longer detectable in blood within 24 h after intake [34]. A more plausible explanation is a sequence-related imbalance in participants’ SkBF recovery characteristics following cold stress. Because SkBF dynamics were not characterized during screening and SkBF recovery was not used for stratification, differences in recovery profiles between sequence groups may have contributed to the observed sequence effect. Future crossover studies should consider assessing SkBF recovery during screening and incorporating it into randomization stratification and/or statistical adjustment.

Finally, the present findings may have implications beyond cold sensitivity in daily life. Hope et al. reported that sublingual glyceryl trinitrate improved recovery of SkT and SkBF after peripheral cold stress in individuals with non-freezing cold injury [35], and suggested that an endothelium-independent NO donor can ameliorate vasodilatory dysfunction in cold-sensitive individuals. Although coffee bean CGAs are not NO donors, their ROS-scavenging activity may enhance NO bioavailability [14, 32]. Together with prior evidence that CGAs promote recovery of SkT and SkBF after peripheral cold stress [15, 16], the present results suggest a potential benefit of CGAs for cold-related vascular dysfunction. However, because our study was conducted in healthy volunteers under moderate cold exposure, any extrapolation to non-freezing cold injury remains speculative and should be evaluated in future clinical studies.

Conclusions

In this study of healthy Japanese women with cold sensitivity, analysis restricted to Period 1 indicated that a coffee bean CGA beverage attenuated cold-induced decreases in skin blood flow and skin temperature compared with placebo under controlled indoor cold exposure. These findings suggest that coffee bean CGA may help alleviate cold sensitivity, including discomfort associated with air-conditioned environments. Given the observed period/sequence effects and the reduced effective sample size, confirmatory studies with an adequately powered design (preferably parallel-group) are warranted.

Abbreviations

CGA

Chlorogenic acid

CQAs

Caffeoylquinic acids

DBP

Diastolic blood pressure

diCQAs

Dicaffeoylquinic acids

diFQAs

Diferuloylquinic acids

FQAs

Feruloylquinic acids

NO

Nitric oxide

ROS

Reactive oxygen species

SBP

Systolic blood pressure

SkT

Skin temperature

SkBF

Skin blood flow

Authors' contributions

TN, AS, and NN contributed to the conception of the study. TN, YN, AS, and NN contributed to the design of the study. TN and KK contributed to the analysis of the data. TN, KK, AS, KH, and NN contributed to the interpretation of the data. TN and KK contributed to the original draft preparation. TN, KK, AS, KH, and NN contributed to the review and editing of the manuscript. All authors have reviewed and approved the final version of the manuscript.

Funding

 This work was funded by the Kao Corporation, which designed the study and undertook site monitoring, data collation, and analysis.

Data availability

The study protocol, statistical analysis plan, and the datasets generated and/or analyzed during the current study are not publicly available because they are considered commercially confidential under the contract with the contracted research organization; however, they are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study was reviewed and approved by the Ethics Committee for Human Studies of the Kao Corporation (ethical code number: T086-171211) and the Ethics Committee of the Oriental Ueno Health Checkup Center (ethical code number: 2018-04). Before the study began, all study participants received a written and verbal explanation of the purpose, method, and limitations of the study. Written informed consent was then obtained from each participant.

Consent for publication

Not applicable.

Competing interests

NN received research consulting fees from Kao Corporation for advice on methods to measure SkT and SkBF and for assistance with analysis and interpretation. TN, YN, KK, AS, and KH are employees of Kao Corporation. The test beverages and coffee bean chlorogenic acids (CGA) used in this study are products of Kao Corporation.