Strawberries Improve Insulin Resistance and Related Cardiometabolic Markers in Adults with Prediabetes: A Randomized Controlled Crossover Trial

Abstract

Background

Prediabetes, as a risk factor for type 2 diabetes, as well as cardiovascular disease, is a burgeoning public health concern in the United States and worldwide. Dietary supplementation of polyphenol-rich berries has been demonstrated to be a feasible nutritional intervention in improving multiple cardiometabolic risk factors in adults. However, reported clinical trials are quite heterogeneous in study findings, and focus on prediabetes is lacking.

Objectives

We examined the effects of a feasible dietary dose of strawberries (32 g freeze-dried strawberries ∼2.5 servings fresh strawberries) on glycemic control (primary) and cardiometabolic markers in adults with prediabetes in a 28-wk randomized controlled (no strawberry) crossover single-blinded study (12 wk/period).

Methods

A total of 25 adults were recruited in each period of the study, and anthropometric, clinical, and dietary data and blood samples were collected at baseline, 6 wk, 12 wk, 16 wk (washout), 22 wk, and 28 wk of the crossover study. A mixed-model analysis of variance was used to examine treatment effects accounting for the fixed effects of treatment, time, order of randomization, age, sex, ethnicity, body mass index (in kg/m²), dietary calories, and physical activity over time, as well as the baseline value for each outcome.

Results

Strawberry period significantly improved glycemic control (serum insulin, insulin resistance, fasting glucose and glycated hemoglobin) and serum total cholesterol in an adjusted model compared with control; [adjusted mean difference 95% confidence interval (CI): –6.9 μIU/mL (–3.2, –10.7) μIU/mL, –2.3 (–1.3, –3.4), –8.9 mg/dL (–4.7, –13.2) mg/dL, and –0.2% (–0.1, –0.3)%, and –7.0 mg/dL (–2.0, –12.0) mg/dL, respectively, all P < 0.05]. Strawberry period also decreased body weight, high sensitivity C-reactive protein, and interleukin-6 in the adjusted model (all P < 0.05).

Conclusions

These findings show that a 2.5-serving dose of strawberries consumed daily for 12 wk can improve prediabetes status and overall cardiometabolic profile in adults.

Introduction

Type 2 diabetes (T2D) continues to plague the public health burden in the United States and worldwide and is accompanied by increased risks of vascular complications, advanced cardiovascular events, and loss of productivity [[1], [2], [3]]. Insulin resistance sets the early clinical course of the disease, and together with increasing adiposity, poor dietary intakes, and environmental factors, it can quickly progress to diminished pancreatic β cell capacity to produce insulin and subsequent chronic hyperglycemia [4,5]. Thus, prediabetes which affects 34.5% United States adults and defined by the American Diabetes Association (ADA) as fasting blood glucose between 100 mg/dL and 125 mg/dL or glycated hemoglobin (HbA1c) between 5.7% and 6.4% or oral glucose tolerance test-based 2-h plasma glucose between 140 g and 199 g after 75 g of oral glucose [3,6], is a logical target of treatment to prevent the onset and progression of T2D. Among the dietary and lifestyle strategies, caloric restriction and weight loss have been shown to produce remission of diabetes in adults with obesity and T2D [7,8]; this strategy lacks sustainability, and weight regain can quickly offset the benefits. Prediabetes itself has also been shown to increase risk of cardiovascular disease (CVD) and all-cause mortality in the general population and in adults with existing CVD [9,10]. Thus, prediabetes as an outcome deserves urgent attention in dietary studies beyond weight loss and general cardiometabolic health.

Functional foods have been defined as natural foods, as well as nutraceuticals, containing bioactive compounds, such as polyphenols, and live microorganisms, that have been shown to provide benefits beyond basic nutrition and reduce risk of chronic diseases, including T2D and CVD [11,12]. Among these functional foods, berries and their bioactive constituents have been shown to improve insulin resistance and lower T2D risk in epidemiological studies and clinical trials [13,14]. Dietary berries represent a low-calorie, nutrient-dense food with an excellent combination of fiber, nutrients, and dietary bioactive polyphenols [13]. Among the popularly consumed berries, blueberries, cranberries, and strawberries have been studied for their effects on T2D risk management. Although the results are promising in adults with different types of cardiometabolic risk or T2D [[15], [16], [17]], there is a dearth of data from intervention studies targeting adults with prediabetes and insulin resistance using dietary achievable doses of these berries. Azari et al. [16] (2022) reported data from 21 clinical studies using blueberries, among which no significant overall effects were noted in blood glucose, insulin resistance, and HbA1c, but blueberry supplementation significantly lowered blood insulin concentrations in these studies [16]. In another meta-analysis report of 22 clinical trials comparing the effects of blueberries or cranberries in adults with or without T2D, this berry intervention was shown to reduce fasting glucose and HbA1c only in adults with T2D (Delpino et al. [17] 2022). When examining the effects of strawberries, Gao et al. [18] (2020) reported data from 11 clinical trials in which strawberry (as whole fruit or freeze-dried powder) feeding showed no effects on glycemic control, including insulin concentration, but improved lipid parameters in adults with overweight and obesity and elevated blood lipids [18]. Anthocyanins constitute an important category of berry polyphenols, but such studies are lacking in prediabetes. A meta-analysis of 13 clinical trials in T2D using anthocyanins from fruits or individual supplements revealed significant improvements in HbA1c and fasting blood glucose but with no effects in insulin on these adults with diabetes [19].

Overall, when reviewed individually, the reported clinical studies using berries demonstrate large heterogeneity in selected participants who range from experiencing overweight or obesity to varying concentrations of elevated blood lipids, glucose, insulin, and/or blood pressure. Based on population data, combinations of different metabolic syndrome components predict T2D risk differently, and impaired blood glucose, insulin, and lipid concentrations were shown to predict the strongest risk of T2D when compared to blood pressure and body weight [[20], [21], [22], [23]]. Consequently, there is an urgent need to intervene in adults targeting these markers. In cognizance of these gaps in reported literature on the role and dosage of dietary berries in glycemic control, our group previously reported a dose-response study of 2 different dietary feasible doses of strawberries (1 and two-and-a-half servings per day) in adults with cardiometabolic risk and insulin resistance [24]. We found significant improvements in fasting insulin and insulin resistance in adults with the two-and-a-half servings of strawberries daily supplementation within a timeframe of 4 wk [24]. Based on these findings, we subsequently examined the effects of this dose specifically in adults with prediabetes as defined by the ADA in a randomized controlled crossover trial of 2 12-wk periods. We examined the effects of strawberries on insulin resistance (based on fasting serum glucose and insulin) as a primary variable and HbA1c, serum lipids, and biomarkers of inflammation as secondary variables, compared to a no strawberry control period in adults with prediabetes.

Methods

Participants

This study was conducted at the University of Nevada Las Vegas (UNLV) School of Medicine Section of Endocrinology outpatient clinic following approval of the UNLV ethics committee. All participants provided written informed consent, and the study was registered at clinicaltrials.gov as NCT05362968. Participants were mainly recruited from the UNLV Endocrinology patient registry and through physician referrals between 2022 and 2024. Inclusion criteria involved adults with prediabetes as defined by the ADA (fasting blood glucose between 100 mg/dL and 125 mg/dL or HbA1c between 5.7% and 6.4% or oral glucose tolerance test-based 2-h plasma glucose between 140 g and 199 g after 75 g of oral glucose) [3,6], as well as HOMA-IR >2.0, and enlarged waist circumference (>35 inches for females and >40 inches for males) [25,26]. Participants were excluded if they were taking medications for blood glucose control or were diagnosed with diabetes, taking herbal/polyphenol supplements, were unable to provide informed consent, were pregnant or lactating, smoking, allergic to strawberries, were on weight loss diets or intended to do so in the near future, were taking medications that would affect lipid metabolism or had recent hospitalization.

Study design and intervention

This was a 28-wk randomized controlled crossover trial, in which each participant meeting the study criteria was randomly assigned to the “strawberry” or “no strawberry” period, each for 12 wk, separated by a 4 wk washout period. Randomization was performed using a SAS-generated randomization order of 2 intervention codes produced by the study statistician, and the research coordinator implemented the process. Following an initial screen visit to determine qualifications, each participant made the following visits for questionnaire-based data and serum collection: baseline, 6 wk, 12 wk, 16 wk (washout), 22 wk, and 28 wk of the crossover study. Height, weight, waist circumference, blood pressure, serum samples, and dietary data were collected at each of these visits (except serum only at the end of the washout). Participants consumed a total of 32 g freeze-dried strawberries (equivalent to two-and-a-half servings of fresh strawberries) divided into 2 doses (16 g powder each provided in sealed packs) in the morning and evening for 12 wk. The freeze-dried strawberry powder was provided by the California Strawberry Commission, and the nutrient and phytochemical composition is presented in Table 1. Participants were instructed to reconstitute the strawberry powder in drinking water right before consumption and not to combine the powder with any other beverage. In addition, participants were also instructed not to consume the strawberry beverage with any other meal or snack and to store the provided powder under refrigerated conditions. At each visit, participants were attended to by the designated medical assistant and a registered dietitian at the clinic. The participants were instructed to maintain their habitual diet and lifestyle throughout the study, including the “no strawberry” control period, and to not consume any other dietary berries throughout the study except those provided in the strawberry period.

TABLE 1.

Composition of strawberry powder consumed per day.

Variable	Strawberry powder
Weight, g	32
Calories, kcal	125
Total carbohydrates, g	30
Ash, g	2.0
Dietary fiber, g	5.5
Vitamin C, mg	65
Total polyphenols, mg1	784
Total anthocyanins, mg2	72
Total hydroxycinnamic acid, mg	19
Total ellagic acid, mg	30
Total flavonols, mg	166

¹Expressed as mg gallic acid equivalents.

²Expressed as mg cyanidin-3-glucoside equivalents.

Anthropometrics and blood pressure

Participant body weight, height, waist circumference, and systolic and diastolic blood pressure were measured at baseline and at the end of 6-wk and 12-wk periods of the strawberry and control treatments. Body weight and height were measured using a digital scale (Accustat Genentech) in light clothing and no shoes. Waist circumference was measured at the superior iliac crest using the Gulick II tape measure (Vital Signs). Systolic and diastolic blood pressure was measured in mmHg using Spot Vital Signs Device (Welch Allyn). At each visit, participants were asked to lie down and relax for ∼8–10 min, following which 3 blood pressure measurements were recorded at intervals of 5–8 min and mean values were recorded.

Biochemical markers

At each visit, freshly drawn serum samples were sent to Quest Diagnostics for analyses of the complete metabolic panel, serum glucose, lipids, insulin, and high-sensitivity C-reactive protein (hs-CRP). Whole blood drawn at 12 wk of each period was analyzed for HbA1c. Inflammatory biomarkers of IL-6, TNF-α, and adiponectin were determined using ELISA kits based on the manufacturer’s protocol [Catalog numbers: D6050B (IL-6), HSTA00E (hs TNF-α), DRP300 (adiponectin) Quantikine, R&D Systems). Each sample was measured in triplicates, and the intra-assay coefficient of variation was ∼ 5% for IL-6, 6% for hs TNF-α, and 4% for adiponectin. The serum concentration of ellagic acid was determined at baseline and 12 wk of each period, as well as also at the washout period for compliance using previously published procedures [27].

Compliance and side effects

Participants were instructed to return unused strawberry powder for compliance, as well as record the timing of strawberry beverage consumption each day. Serum ellagic acid as a biomarker of strawberry consumption was also measured for compliance. Side effects were monitored by asking participants to comment on any symptom, especially gastrointestinal, that they experienced during the strawberry period and were manually recorded on case report forms.

Habitual diet and physical activity

Participants maintained their usual diet and levels of physical activity throughout the study and turned in food records and minutes of moderate and/or vigorous physical activity at the scheduled visits. Each participant maintained a weekly 3-d food record at baseline, 6 wk, and 12 wk of each period of the study. Dietary analyses were conducted by the study registered dietitian using the ESHA’s Food Processor nutrition analysis software for energy, nutrients, and food group intake for each participant. Self-reported physical activity participation was quantified by the international physical activity questionnaire short form, which assesses physical activity levels during the last 7 d, and participant responses (yes/no), as well as actual minutes, were recorded for meeting exercise recommendations (≥150 min of moderate and/or ≥90 min vigorous exercise per week) [28].

Statistical analysis

Our primary study variable is insulin resistance (based on fasting serum glucose and insulin), and HbA1c, serum lipids, and biomarkers of inflammation are secondary variables. Based on our previously published randomized controlled crossover study and using estimated mean changes from baseline in control and strawberry high dose groups (prepost change) [24], we needed a total of 16 adults/treatment to detect significant differences in serum insulin (effect size 0.75) for a 2-tailed paired samples test, with ≥80% power, and α of 0.05. Accounting for 30% attrition, we estimated a sample size of 21 adults/treatment for significant differences between strawberry and no strawberry (control) periods. For statistical analysis, data were graphed and assessed for outliers, and descriptive statistics were calculated for each variable. Data are presented as means ± SD for continuous variables. Outcomes were assessed for normality based on the calculated skewness of each model’s residuals with a skewness threshold of 2 and –2, thresholds identified by prior work from Hair et al. [29] (2010). However, more recent work has demonstrated that parameter estimates of mixed-effects models are robust to residual skewness as extreme as 3 or –3 [30]. None of the models had residuals that were >2 or <–2 in this analysis. For each outcome variable, differences between strawberry and no-strawberry periods were determined using a mixed-model analysis of variance accounting for the fixed effects of treatment, period, order of randomization, age, sex, BMI, ethnicity, dietary calories, and physical activity over time, as well as the baseline value for each outcome. The participant was modeled as a random effect and with unstructured variance for treatment/period. Treatment-by-period interactions was included to test for potential carryover effects between periods [31]. If significant, the treatment-by-period interaction was retained in the model, and lower-order effects of treatment and period were not interpreted. When period and treatment-by-period interactions were nonsignificant, they were removed from the model. When period effects were significant, they were retained in the final model of treatment effects. Adjusted mean differences between strawberry and control periods and 95% confidence intervals (CIs) were also generated to estimate the change in each outcome variable by treatment in adjusted analyses. P values were adjusted for multiple hypotheses testing (false discovery rate), which are presented in the main tables. All P values were 2-tailed, and treatment effects were considered significant if <0.05. Analyses were performed using SAS (version 9.4; SAS Institute Inc.).

Results

Baseline features and compliance

As shown in Figure 1, a total of 37 participants qualified and were enrolled in the study; 12 withdrew in the first 2 wk of the study due to time constraints and being unable to make the first follow-up visit, and 25 participants completed all visits in the crossover study. Based on the return of unused strawberry powder, food logs, and plasma ellagic acid, we observed >85% compliance among participants who completed the study. Plasma ellagic acid was detectable at 6- and 12-wk strawberry period, at concentrations of 27.6 ± 9.1 ng/mL and 30.2 ± 8.8 ng/mL, respectively, and was not detectable at baseline, washout, and 6- and 12-wk control periods. Table 2 shows the baseline features of the participants who completed the study. The majority of participants were of self-reported Hispanic origin, females, and had a BMI in the obesity category. Baseline features and clinical variables did not differ by sequence of randomization.

FIGURE 1.

Study design.

TABLE 2.

Baseline and clinical characteristics of study participants by sequence of randomization.

Variable	Entire sample (n = 25)	Strawberry to control sequence (n = 12)	Control to strawberry sequence (n = 13)
Age, (years)	52 ± 14	54 ± 16	52 ± 13
M/F	6/19	3/9	3/10
Ethnicity (Hispanic %)	56	50	61
BMI, (kg/m2)	32 ± 2	32 ± 2	32 ± 2
Systolic blood pressure, mmHg	127 ± 7	127 ± 6	125 ± 9
Diastolic blood pressure, mmHg	79 ± 5	78 ± 5	79 ± 5
Blood pressure medication use, n (%)	8 (32)	4 (33)	4 (31)
Antidepressant use, n (%)	6 (24)	2 (16)	4 (31)
Multivitamin use, n (%)	17 (68)	9 (75)	8 (61)
Serum AST, U/L	14.6 ± 3.0	15.4 ± 3.0	13.7 ± 3.0
Serum ALT, U/L	14.5 ± 5.2	16.2 ± 3.6	12.6 ± 5.8
Serum fasting glucose, mg/dL	107 ± 15	112 ± 14	103 ± 13
Blood HbA1c, %	6.0 ± 0.3	6.0 ± 0.2	6.0 ± 0.3
Serum Insulin, μIU/mL	23.7 ± 14.7	25.5 ± 17.8	21.7 ± 10.8
HOMA-IR	6.4 ± 4.3	7.1 ± 4.9	5.8 ± 3.5
Serum total cholesterol, mg/dL	188 ± 34	191 ± 36	186 ± 31
Serum LDL cholesterol, mg/dL	113 ± 27	118 ± 30	108 ± 22
Serum HDL cholesterol, mg/dL	49 ± 9	48 ± 9	50 ± 10
Serum triglycerides, mg/dL	139 ± 67	142 ± 61	134 ± 72
Serum hs-CRP, mg/L	3.6 ± 1.7	3.2 ± 2.0	4.4 ± 1.1
Serum adiponectin, ng/mL	4.3 ± 2.0	4.3 ± 1.6	4.4 ± 2.3
Serum IL-6, pg/mL	31.4 ± 8.1	30.6 ± 9.4	33.0 ± 6.5
Serum hs-TNF-α, pg/mL	2.7 ± 1.3	2.8 ± 1.4	2.6 ± 1.1

Data are presented as mean and SD. Count data presented as n (%).

Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; F, female; HbA1c, glycated hemoglobin; HDL cholesterol, high-density lipoprotein cholesterol; HOMA-IR, homeostatic model of insulin resistance; hs-CRP, high sensitivity-C reactive protein; hs-TNF-α, high sensitivity tumor necrosis factor α; IL, interleukin; LDL cholesterol, low-density lipoprotein cholesterol; M, male; SD, standard deviation.

Anthropometrics, blood pressure, and metabolic variables

The strawberry period showed significant improvements in glycemic and lipid variables compared with the control (no strawberry) period (Table 3). In the model adjusted for time, order of randomization, age, ethnicity, sex, BMI, dietary calories and physical activity over 12 wk, and baseline value for each outcome, the adjusted mean differences and 95% CIs of fasting glucose, insulin, HOMA-IR, HbA1c, and serum total cholesterol, were –8.9 mg/dL (–4.7, –13.2) mg/dL, –6.9 μIU/mL (–3.2, –10.7) μIU/mL, –2.3 (–1.3, –3.4), –0.2% (–0.1, –0.3)%, and –7.0 mg/dL (–2.0, –12.0) mg/dL, respectively, in the strawberry compared with control period (all P < 0.05). No significant differences were noted in serum LDL cholesterol, HDL cholesterol, and triglycerides following the strawberry period. The adjusted mean difference for systolic blood pressure was also significantly lower following the strawberry period compared with the control. Finally, the strawberry period also showed significant improvements in body weight and waist circumference compared with control [–3.4 (–1.9, –4.8) lb, and –0.4 (–0.2, –0.7) inches, respectively)] in adjusted analysis (Table 3). No carryover effects were noted for our primary variable of serum insulin based on examining group differences at weeks 6 and 22 of the study (P = 0.27 and P = 0.46, respectively).

TABLE 3.

Anthropometric, blood pressure, glucose, and lipid profiles in adults with prediabetes following each treatment period in the 28-wk randomized crossover study.

Variable	Baseline	Control(6-wk)	Control(12-wk)	Strawberry(6-wk)	Strawberry(12-wk)	P-treatment(FDR-adjusted)1
Body weight, lb	197 ± 21	198 ± 22	198 ± 21	196 ± 20	195 ± 21	<0.001
BMI, kg/m2	32.0 ± 2.1	32.2 ± 2.1	32.3 ± 2.0	31.7 ± 1.8	31.6 ± 2.0	<0.001
Waist circumference, inches	41.4 ± 3.8	41.0 ± 4.0	41.4 ± 3.9	41.0 ± 3.7	40.6 ± 3.2	0.004
Systolic blood pressure, mmHg	127 ± 7	127 ± 7	127 ± 8	122 ± 8	125 ± 7	<0.001
Diastolic blood pressure, mmHg	79 ± 5	78 ± 5	79 ± 5	78 ± 4	78 ± 4	0.052
Serum total cholesterol, mg/dL	188 ± 34	186 ± 35	191 ± 30	183 ± 29	177 ± 30	0.005
Serum LDL cholesterol, mg/dL	113 ± 27	111 ± 29	116 ± 22	110 ± 24	106 ± 25	0.120
Serum HDL cholesterol, mg/dL	49 ± 9	49 ± 11	49 ± 10	52 ± 9	49 ± 9	0.084
Serum triglycerides, mg/dL	139 ± 67	140 ± 67	144 ± 49	130 ± 40	128 ± 41	0.060
Serum fasting glucose, mg/dL	107 ± 15	102 ± 16	109 ± 10	97 ± 12	97 ± 12	<0.001
Blood HbA1c, %	6.0 ± 0.3	—	6.1 ± 0.2	—	5.9 ± 0.2	<0.001
Serum insulin, μIU/mL	23.7 ± 14.7	21.8 ± 15.0	28.4 ± 12.2	18.0 ± 7.1	17.3 ± 7.0	0.001
HOMA-IR	6.4 ± 4.3	6.2 ± 4.5	7.4 ± 3.0	4.5 ± 1.8	4.1 ± 2.2	<0.001

Data presented as means ± SD, n = 25/period.

Abbreviations: BMI, body mass index; FDR, false discovery rate; HbA1c, glycated hemoglobin; HDL cholesterol, high-density lipoprotein cholesterol; HOMA-IR, homeostatic model of insulin resistance; LDL cholesterol, low-density lipoprotein cholesterol; SD, standard deviation.

¹P for the effect of treatment from the MIXED procedure (SAS version 9.4; SAS Institute Inc), adjusted for the period, order of randomization, age, ethnicity, sex, BMI, dietary calories, and physical activity over 12 wk, and baseline value for each outcome. P < 0.05 in bold font.

Serum biomarkers of inflammation

The strawberry period showed significant changes in serum biomarkers of inflammation compared with the control (no strawberry) period (Table 4). In the model adjusted for time, order of randomization, age, ethnicity, sex, BMI, dietary calories and physical activity over 12 wk, and baseline value for each outcome, the adjusted mean differences and 95% CIs of serum hs-CRP and IL-6 were –0.8 mg/L (–0.3, –1.3) mg/L, and –10.8 ng/mL (–7.8, –13.7) ng/mL, respectively. Strawberry treatment did not affect serum TNF-α and adiponectin concentrations. A significant period effect was noted for only TNF-α (mean difference = 0.3 pg/mL, adjusted P < 0.001). This period effect was adjusted in the analysis.

TABLE 4.

Serum biomarkers of inflammation in adults with prediabetes following each treatment period in the 28-wk randomized crossover study.

Variable	Baseline	Control(6-wk)	Control(12-wk)	Strawberry(6-wk)	Strawberry(12-wk)	P-treatment(FDR-adjusted)1
Serum hs-CRP, mg/L	3.6 ± 1.7	5.4 ± 0.9	5.0 ± 1.5	4.7 ± 1.6	4.3 ± 1.5	0.001
Serum adiponectin, ng/mL	4.3 ± 2.0	4.0 ± 1.6	4.4 ± 2.2	4.3 ± 1.9	4.0 ± 2.0	0.748
Serum IL-6, pg/mL	31.4 ± 8.1	33.3 ± 5.5	35.6 ± 8.1	25.0 ± 10.2	22.0 ± 7.2	<0.001
Serum hs-TNF-α, pg/mL	2.7 ± 1.3	3.0 ± 1.3	3.6 ± 1.6	4.3 ± 2.3	4.4 ± 2.1	0.852

Data presented as means ± SD, n = 25/period.

Abbreviations: BMI, body mass index; FDR, false discovery rate; hs-CRP, high sensitivity-C reactive protein; hs-TNF-α, high sensitivity tumor necrosis factor-α; IL, interleukin; SD, standard deviation.

¹P for the effect of the treatment from the MIXED procedure (SAS version 9.4; SAS Institute Inc), adjusted for the period, order of randomization, age, ethnicity, sex, BMI, dietary calories, and physical activity over 12 wk, and baseline value for each outcome. P < 0.05 in bold font.

Habitual dietary intake and physical activity

Daily total caloric intake based on habitual diet showed decreases in the strawberry period compared with control in the adjusted analysis [–127 (–70, –184) kcal]. These caloric differences were mainly driven by small changes in calories derived from dietary carbohydrates. Between essential food groups, the strawberry period revealed an overall decrease in fruit and vegetable consumption that met daily recommendations. A significant period effect was noted for only vitamin C (mean difference = 0.7 mg, adjusted P < 0.001). This period effect was adjusted in the analysis. Moderate and/or vigorous physical activity also showed a significant increase in the strawberry period, though of a small magnitude [6.2 (3.2, 9.2) min/wk] (all P < 0.05, Table 5).

TABLE 5.

Background daily dietary intakes and minutes of physical activity in adults with prediabetes following each treatment period in the 28-wk randomized crossover study.

Variable	Baseline	Control(6-wk)	Control(12-wk)	Strawberry(6-wk)	Strawberry(12-wk)	P-treatment(FDR-adjusted)1
Total calories, kcal	2130 ± 277	2137 ± 269	2151 ± 251	2060 ± 228	1988 ± 158	<0.001
Carbohydrates, % kcal	48 ± 6	48 ± 6	48 ± 6	50 ± 5	50 ± 5	<0.001
Fats, % kcal	31 ± 6	30 ± 6	30 ± 6	32 ± 7	30 ± 7	0.212
Proteins, % kcal	22 ± 5	22 ± 5	22 ± 5	20 ± 4	20 ± 5	0.121
Fiber, g	19 ± 3	19 ± 3	20 ± 3	18 ± 4	19 ± 4	0.055
Vitamin C, mg	34 ± 10	34 ± 10	40 ± 10	30 ± 10	34 ± 10	0.131
Vitamin E, mg	9 ± 3	8 ± 3	9 ± 4	8 ± 3	9 ± 4	0.059
Fruits, % recommended intake2	41 ± 9	41 ± 9	41 ± 9	31 ± 12	35 ± 11	<0.001
Vegetables, % recommended intake	56 ± 11	57 ± 11	57 ± 11	54 ± 9	54 ± 7	<0.001
Grains, % recommended intake	84 ± 5	85 ± 5	85 ± 5	88 ± 5	87 ± 4	0.095
Dairy, % recommended intake	56 ± 10	55 ± 10	56 ± 11	58 ± 10	55 ± 10	0.073
MVPA, min/wk	92 ± 17	93 ± 17	96 ± 17	99 ± 14	100 ± 17	<0.001

Data presented as means ± SD, n = 25/period.

Abbreviations: FDR, false discovery rate; MVPA, moderate and/or vigorous physical activity; SD, standard deviation.

P < 0.05 in bold font.

¹P for the effect of the treatment from the MIXED procedure (SAS version 9.4; SAS Institute Inc), adjusted for the period, order of randomization, age, ethnicity, sex, and baseline value for each outcome.

²Excluding the strawberry dose provided in the study.

Side effects

Overall, participants revealed excellent compliance with few side effects. Two participants experienced gastrointestinal disturbance during the strawberry period, and another participant experienced headaches. These side effects were anticipated, and participants were made aware of them during the consenting process.

Discussion

To the best of our knowledge, this is the first 12-wk randomized controlled crossover study using a dietary feasible dose of strawberry fruit supplementation showing improved glycemic and lipid control and selected biomarkers of inflammation in adults with obesity and prediabetes when compared to the control period. These findings remain significant when adjusted for background dietary calories and physical activity, among other covariates. Strawberries are a popular fruit that received overall high compliance among participants in the study, and the dose administered (two-and-a-half servings per day) is within the USDA dietary guidelines recommendations for daily intake of fruits and vegetables, as well as falls within the observed habitual doses in epidemiological studies of diabetes risk [32,33]. Our study in primarily Hispanic adults also addresses the role of functional food or polyphenol-based dietary strategies in mitigating health disparities, especially keeping in view the rising prevalence of T2D among Hispanic adults in the United States in recent years [34,35].

Our primary variable of insulin resistance was significantly improved in the strawberry compared with the control period, thereby leading to improved glycemic control as reflected in HbA1c in our study participants; the magnitude of differences detected are of clinical relevance as each 0.1–1% increase in HbA1c has been predictive of T2D and cardiovascular events in adults [[36], [37], [38]]. Insulin resistance is a critical factor in the development and progression of T2D but has shown to be unaffected in most of the previous studies of whole-berry fruits (fresh or freeze-dried) in adults with metabolic syndrome or prediabetes [16,18]. Our group has previously reported clinical studies using freeze-dried strawberries in adults with different features of the metabolic syndrome; findings revealed that 25 g or 50 g freeze-dried strawberries improved total and LDL cholesterol but did not affect blood glucose, insulin, and HOMA-IR in normoglycemic adults [39,40]. On the contrary, when we examined a daily dose of 32 g freeze-dried strawberries (equivalent to two-and-a-half servings of fresh strawberries per day) in adults with above optimal LDL cholesterol and insulin resistance, we observed significant improvements in fasting insulin and HOMA-IR in a 14-wk randomized crossover trial [24]. Similar improvements in postprandial insulin responses were also reported with strawberry supplementation in adults who were overweight and consumed a high-fat, high-carbohydrate meal with or without strawberry beverage (10 g freeze-dried strawberries) [41]. Findings from this study show significantly lower postprandial insulin concentrations following a meal with strawberries compared with without. The serum anthocyanin concentration of pelargonidin and its metabolites, indicative of strawberry consumption, also increased following strawberry consumption with meal, thereby explaining their potential role in the improvement of postprandial insulin responses and improving insulin sensitivity compared with control [41]. When examining the effects of other commonly consumed berries in the United States diet, such as blueberries, results also remain inconsistent on glycemic outcomes. Stull et al. [42] reported significant improvements in insulin sensitivity following a 6-wk blueberry beverage consumption (22.5 g blueberry bioactive) compared with control beverages in adults with insulin resistance and obesity [42]. However, no such improvements were observed in our previous report when 50 g freeze-dried blueberries were supplemented for 8 wk in normoglycemic adults with other features of the metabolic syndrome [43]. In a longer study of 6 mo, measures of glycemic control, including serum glucose, insulin, HbA1c, HOMA-IR, and the quantitative insulin sensitivity index, were not significantly affected in individuals with metabolic syndrome who consumed either 13 g or 26 g freeze-dried blueberries (the equivalent of half or 1 cup fresh blueberries per day, respectively) [44]. These variations in results may be attributed to different treatment durations, participant baseline characteristics, and/or nutritional and phenolic compounds present in strawberries compared with blueberries, such as the lower sugar, total carbohydrate, and glycemic index of strawberries when compared to blueberries per serving [45]. Mechanistic studies explain our clinical observations, as anthocyanins and ellagic acid, being 2 of the main categories of bioactive compounds in strawberries, have been shown to preserve pancreatic function and reduce apoptosis, activate insulin receptor phosphorylation, decrease oxidative stress and inflammation, and thereby insulin resistance in experimental diabetes [[46], [47], [48]]. Prediabetes is characterized by elevated hepatic lipids and circulating free fatty acids, as well as oxidative stress and inflammation, which together promote insulin resistance and impaired glucose tolerance [49,50]. Treatment of experimental rats with whole strawberry powder and/or with blueberries demonstrates a decrease in visceral fat adiposity, body weight gain, inflammatory biomarkers, and insulin resistance [[51], [52], [53]]. These mechanistic data largely explain our findings of lower body weight and HOMA-IR in the strawberry period after adjusting for caloric intake and physical activity in the study.

Strawberry treatment also improved our secondary variables of lipids and biomarkers of inflammation. Our study also demonstrated a dietarily achievable dose of strawberries yielding a significant improvement in serum total cholesterol within 12 wk. In contrast to our previous studies, LDL cholesterol was not significantly reduced, which may be attributed to prior interventions being tested in subjects with abdominal adiposity and elevated serum LDL cholesterol [39] or 1 or more features of metabolic syndrome, which may include the aforementioned criteria [24]. One possible explanation of this cholesterol-lowering effect could be due to anthocyanins, the main polyphenol in strawberries, and their effects on the lipid profile; although a prior study demonstrated anthocyanins from berry intake significantly improved HDL cholesterol and LDL cholesterol in dyslipidemia in adults, with no significant effects on total cholesterol, their finding of significantly lower plasma cholesteryl ester transfer protein after treatment may explain a potential mechanism explaining our findings [54]. In addition to the direct cholesterol-lowering effects of anthocyanins, our observed effect can also be explained by improved insulin resistance in these adults. Since insulin resistance has been associated with a shift in cholesterol metabolism, thereby increasing cholesterol synthesis and serum cholesterol concentrations [55,56], lowering insulin resistance may also explain a concomitant decrease in serum cholesterol in the strawberry period. Our current study also demonstrated significant improvements in selected serum biomarkers of inflammation in prediabetes following strawberry intake, specifically a reduction of hs-CRP and IL-6. Our previous study showed no significant effects on C reactive protein after 12 wk of 25 g or 50 g freeze-dried strawberry supplementation in obese individuals with elevated serum lipids[39], and this may be explained by the overall normoglycemic status of adults, as well as interindividual variabilities in our previous parallel study. Regarding IL-6, a pro-inflammatory cytokine, our study findings conform to a previous study on the postprandial effects of strawberry intake after a high-fat, high-carbohydrate meal, demonstrating strawberry intake before the meal blunted the increase in IL-6 than placebo (though not significant) [57], and another study showing significant reduction of serum IL-6 over 10 h postprandial following strawberry dose [58]. Our result of a decrease in hs-CRP also conforms to a previous observational study finding in which females consuming >3 servings per week had a reduced risk of elevated C reactive protein [59]. Considering inflammation is a key factor in the progression from prediabetes to diabetes, usually exacerbated by hyperglycemia and insulin resistance, reduction in inflammatory biomarkers suggests strawberry intake can significantly improve T2D risk attributable to inflammation [60]. Multiple mechanistic studies support our clinical observations in which strawberries bioactive, such as ellagic acid and pelargonidin-3-O-glucoside, have shown anti-inflammatory effects in lowering the transcription of inflammatory molecules, including IL-6 and C reactive protein [61,62]. Further, IL-6 and C reactive proteins have been shown to induce insulin resistance, mainly in the setting of chronic obesity [63,64], and thus, improvements in our primary variable of insulin resistance may also explain the observed decreases in these inflammatory biomarkers in the strawberry period.

Our study design has several strengths, including a randomized crossover study to limit the confounding factor of interindividual variability, as well as a 4-wk washout phase to reduce potential carryover effects. Furthermore, considering the strawberry and no strawberry phases were each 12 wk, the longer duration of treatment compared to previous studies allows for greater detection of changes over time. Regarding practical relevance, our study utilized a dietary achievable dose of strawberry intake rather than a high dose as in previous studies, and the significant results show the potential effectiveness of food-based intervention in free-living adults. Furthermore, we recorded background dietary intakes and physical activity levels over the entire study duration and observed differences in caloric intake in the strawberry period, showing a decreased intake compared with the control period. The daily mean energy intake observed in our study participants with obesity is quite similar to reported data from United States adults with obesity [65,66]. Thus, our models were adjusted for caloric intake and physical activity in the final analyses, thereby attributing the observed outcome differences to strawberry intake in the trial.

Some limitations of our study include the lack of participant blinding, which led to unintentional bias, as participants were not given a placebo drink during the no-strawberry period. Our study was single-blinded, as none of the team members were aware of the treatment period except the study coordinator. Designing a placebo for a whole food-based intervention is quite challenging due to the synergistic effects of several bioactive compounds and nutrients in the whole fruit that explain the observed health effects [67,68]. To address this challenge, the concept of a “usual diet” as a comparison group has also been identified in feeding studies [69,70]. Although our previous strawberry studies used a calorie-matched powder with sensory characteristics similar to freeze-dried strawberries [24] or a calorie- and fiber-matched placebo drink [39], in this trial, we aimed to study how a “usual diet and lifestyle” as a control may affect cardiometabolic outcomes when compared to the strawberry treatment. Further, the concept of adding “empty calories” as a placebo agent for a 12-wk study in adults with prediabetes could pose a health risk and must be ethically evaluated in study designs. Thus, the approach of a “usual diet and lifestyle” control with no strawberries is more realistic and effective from a public health perspective, which emphasizes whole fruit addition to diets instead of examining the individual fruit compounds on metabolic outcomes using a “fruit-like” placebo comparator. Dose-wise, we only administered a single dose of strawberries and not a dose-response design, including lower doses of 1 or half serving daily doses. Also, since participant drop-outs occurred at the beginning of the study, we were unable to perform intent-to-treat and completed analyses for this study. We also used a single baseline value for the 2 crossover treatment periods and did not examine differences at the end of the 4-wk washout; other studies on berries also report using a similar model [[71], [72], [73]]. We acknowledge that not having a new baseline value after the washout could introduce a potential bias in analysis, but we did not anticipate clinically meaningful changes in primary variables within 4 wk of washout and from values measured at the end of the first treatment period. All participants with prediabetes were under medical supervision, and no abnormal glycemic events were reported within the washout phase and throughout the 28-wk study. Finally, we designed the washout to examine whether ellagic acid as a biomarker of strawberry intake and compliance was detectable in plasma. These limitations can be addressed in future studies using a continuous glucose/insulin monitoring system to capture detailed changes over time. Further, although excluding participants who were on medications for blood glucose control or diets for weight loss helps reduce confounding factors, this also limits generalizability as a significant portion of individuals with prediabetes begin medical interventions to prevent disease progression. Future studies are warranted in patients with T2D who are on medications to determine if strawberry intake is an effective adjunct therapy for blood glucose control.

In conclusion, our findings from the current study support the benefits of strawberry intake at a dose of 2.5 servings daily for 12 wk, and results remain significant in improving prediabetes status accounting for dietary calories and physical activity in the study. Thus, dietary strawberries can be recommended in medical nutritional therapy as a nonpharmacological, feasible, and sustainable intervention in prediabetes management in adults.

Author contributions

The authors’ responsibilities were as follows– AB, KI, JLE, RHS: designed the study; AB, SG, AH: completed the data coding, conducted the statistical analysis, and wrote the first draft of the manuscript; AB, AC, KI: supervised the study; PD: recruited and followed up participants in the trial; SG, JLE, RHS: contributed to interpretation of results; AB: had primary responsibility for final content; and all authors: read and approved the final manuscript.

Data availability

Data described in the manuscript can be made available upon request and following approval of the ethics committee.

Funding

This study was supported by an award from the California Strawberry Commission, Watsonville, CA (to AB). Additional support was provided by an award from the National Institute of General Medical Sciences, NIH under grant number U54 GM104944 (to AB and JLE) and the NIH grant U54GM104938 (Oklahoma Shared Clinical and Translational Resource) at the University of Oklahoma Health Sciences Center (to RHS).

Conflict of interest

AB serves on the editorial board of the Journal of Nutrition. All other authors report no conflicts of interest.