Cashew nut (Anacardium occidentale L.) and cashew nut oil were produced in Brazil, coming from donation of the Brazilian Agricultural Research Corporation (Embrapa), Agroindústria Tropical, Fortaleza (Brazil).

All procedures described next were carried out at the Embrapa. Oil samples were extracted by centrifugation. For sample preparation, the cashew nut was roasted at 110°C for 15 min; cashew nut was ground in a food processor; adding water to the cashew nut (4,1 cashew nut, water w/w) and the mixture was homogenized in a processor at 90°C for 10 min. This mixture was centrifuged for 1 h at 4,500 rpm at room temperature. After centrifugation, the oil was heated in an oven at 105°C for 1 h (14). The raw material was obtained from the same crop, and its microbiological quality was analyzed and assured via reports by the supplier company until they were delivered to the Laboratory of Energy Metabolism and Body Composition of the Universidade Federal de Viçosa (LAMECC/UFV).

For the intervention, cashew nuts were portioned into laminated and vacuum-sealed packages (30 g), while cashew nut oil was fractionated and stored in 250 mL amber glass bottles. Both foods were stored in a freezer at −20°C until distribution to participants to avoid nutrient oxidation, sensory changes, and microbiological contamination. All material for consumption was handled following hygienic-sanitary standards, including the use of clean lab coats, caps, masks, and disposable gloves.

Regarding nutrients and bioactive compounds of cashew nut, moisture, ash, protein, lipids, carbohydrates, dietary fibers, amino acids, and in vitro digestibility were evaluated. The moisture, ash, and protein contents were performed according to the methodology indicated by the AOAC (16), the last one was obtained by combustion in the Nitrogen/Protein Analyzer equipment. Carbohydrate content was calculated by the difference of 100 and the sum of the values obtained for moisture, ash, proteins, and lipids. The energy value per 100 g of each product was calculated using the Atwater system: Caloric value = (g of protein × 4) + (g of lipids × 9) + (g of carbohydrates × 4). Total dietary fiber (soluble and insoluble fiber) was determined by the gravimetric non-enzymatic method, using the commercial kit (Total dietary fiber assay kit, Sigma^®, San Luis, Missouri, EUA) (16). The amino acid contents (aspartic acid, glutamic acid, serine, glycine, histidine, taurine, arginine, threonine, alanine, proline, tyrosine, valine, methionine, cystine, isoleucine, leucine, phenylalanine, lysine, hydroxyproline, tryptophan and the sum of total amino acids) were performed based on the MA-009 R0 method (17, 18), and tryptophan by the MA-010 R.1 method (19). In vitro digestibility was analyzed by the previously reported method (20).

Both cashew nut and cashew nut oil underwent analysis for minerals, vitamin E and its derivatives, total phenolics, and antioxidant capacity. Mineral analyzes (phosphorus, potassium, calcium, magnesium, selenium, sodium, copper, iron, zinc, and manganese) were performed according to the methodology of the Food and Drug Administration (FDA) (21). The preparation and analysis of the vitamin E isomers (α-, β-, γ-, δ-tocopherols and tocotrienols) were extracted according to Pinheiro-Sant’Ana et al. (2011), and performed in five replicates by High-Performance Liquid Chromatography (HPLC). During analysis, the samples were protected from sunlight and artificial light using amber glassware, aluminum foil, and blackout curtains, and protected from oxygen by using lids and environments with nitrogen gas in glass bottles. The total phenolic compound content was obtained from reading of absorbance in a spectrophotometer (Thermo Scientific, Evolution 606, United States) at 765 nm. Analytical curve of gallic acid (0.005–0.10 mg/mL) was used to quantify the compounds. The results were expressed in mg of gallic acid equivalents/g of cashew nut (mg GAE/g). The antioxidant activity was determined by the sequestering capacity of free radical DPPH (2,2-diphenyl-1-picryl-hydrazil) as described before (22).

Additionally, we also performed analysis of fatty acids, acidity level, and peroxide index in cashew nut oil. Lipids were obtained using the high-pressure, high-temperature extraction system in Ankom XT-15 equipment according to the American Oil Chemists’ Society (23). The fatty acid profile was determined using the procedure described by Hartman and Lago (1973). The determinations of acidity and peroxides were performed according to AOCS (2003).

This is an 8-wk randomized controlled three-arm dietary intervention, in which subjects were assigned to receive control (CT), cashew nut (CN) or cashew nut oil (OL) plus an energy-restricted diet. This study was conducted at the Department of Nutrition and Health of the Universidade Federal de Viçosa (UFV), Brazil, between January 2022 and July 2022, according to the guidelines laid down in the Declaration of Helsinki. All procedures involving human subjects were approved by the Ethics Committee in Research with Human Experimentation of the Universidade Federal de Viçosa (No. 4.543.541/CEPH). Written informed consent was obtained from all subjects/patients. The study is registered at the Brazilian Registry of Clinical Trials (ReBEC) with ID number RBR-8xzkyp2.

During the intervention, participants attended on three occasions at the LAMECC/UFV: initial and final days for blood sample collection, anthropometry, body composition evaluation and fill out questionnaires about physical activity practice and food record, and in the fourth week (30 days) for a face-to-face monitoring visit and anthropometric measurements. Between face-to-face visits, participants received online monitoring (Figure 1).

Compliance was evaluated during the monitoring visit to determine whether participants were adhering to the intervention. We tracked the consumption of cashew nut and cashew nut oil from the initial day of the intervention until their return. If participants in the intervention group still possessed cashew nuts or oil from the beginning of the study upon returning, they were required to surrender them for assessment, verifying whether they had indeed consumed the prescribed quantity. Additionally, during the application of the return questionnaire, we evaluated whether participants had commenced any new medication or developed any illnesses. The participants who started taking any drug or developed any disease listed in the non-inclusion criteria was excluded from the study. Furthermore, compliance was gauged at the study’s conclusion by monitoring weight gain. Since the intervention aimed to body fat loss by energy restriction, participants were expected to experience weight loss. Consequently, individuals who exhibited weight gain were excluded from the study due to non-compliance.

Study participants were recruited in Viçosa, Minas Gerais, Brazil, via radio announcements, social media, and the UFV network platform. An online form assessed individuals’ eligibility, with eligible candidates undergoing a face-to-face selection questionnaire to confirm eligibility. This questionnaire covered clinical, dietary, sociodemographic, and anthropometric data, along with body composition, blood pressure, and recent biochemical test results. Participants received a booklet containing guidelines about the study and were instructed to report any changes in medication or health status.

The inclusion criteria for participants in the study consisted of men or women (20–55 y); with overweight (27–29.9 kg/m²), waist circumference (WC) ≥80 cm for women and ≥90 cm for men and with body fat percentage >30% for women and >20% for men associated with at least another component of metabolic syndrome (MS): triglycerides (TG) ≥150 mg/dL; blood pressure ≥130/85 mmHg or fasting blood glucose ≥100 mg/dL or who uses medication to control these markers; or men or women with obesity (BMI ≥ 30 kg/m²), WC ≥80 cm for women and ≥90 cm for men, and body fat percentage >30% for women and >20% for men with or without metabolic complications.

The non-inclusion criteria included pregnant, lactating, or menopausal women; athletes; vegans; or have a diagnosis of insulin-dependent diabetes; diagnosis of HIV, digestive, hepatic, renal, cardiovascular, thyroid, cancer, inflammatory diseases and eating disorders; history of drug and/or alcohol abuse; have an aversion or allergy to nuts; present infection in the last month; habitually consume nuts above 30 g/day; use drugs such as anti-inflammatories, corticosteroids, and antibiotics, capable of biochemical alterations; chewing difficult; weight instability (5% of usual weight) in the last 3 months; alcohol consumption >21 units (≈168 g) per week; and intake of vitamin, mineral, and omega 3 supplements.

One week before intervention, the participants participated in a run-in period. During run-in the subjects were instructed to consume their habitual diets without nuts, dried fruits like berries (cranberry, blueberry, goji berry and raisins), açaí, cocoa, cinnamon, olive oil and alcoholic beverages, and to maintain their usual activities. Following the run-in period, individuals whose body weight fluctuated beyond ±1 kg or who consumed prohibited foods or beverages were categorized as “poor responders” and excluded from the study.

All participants received an energy-restricted (−500 kcal/d) diet. In addition, the cashew nut group received 30 g/d of vacuum-sealed cashew nut to be consumed daily, and the cashew nut oil group received 250 mL amber glass bottles of oil along with a measuring cup to standardize the amount to 30 mL/d of cashew nut oil. All dietary advice was individualized and provided by dietitians. At the beginning, five energy-restricted diet options for all groups were designed and divided into five meals: breakfast, morning snack, lunch, afternoon snack, and dinner. All menus were calculated in an Excel spreadsheet using the Brazilian Institute of Geography and Statistics (IBGE) table. Energy requirements were calculated according to the Mifflin’s formula (24). For everyone, 500 kcal were reduced from the total calculated energy requirement, considering the level of physical activity of each participant. For the interventional groups, a daily cashew nut (30 g/d) or cashew nut oil (30 mL/d) was added to the individual meal plans, and the percentage of energy from total fat was around 27% for cashew nut group, 32% for the cashew nut oil group, while the control group had around 21% (Table 1). This amount of cashew nut was based on previous studies that have used similar amounts, the PREvención con DIeta MEDiterránea (PREDIMED), which demonstrated beneficial effects in the improvement of blood pressure, lipid profile, lipoprotein particles, inflammation, oxidative stress, and carotid atherosclerosis (25, 26). For the cashew nut oil group, 30 mL/d was calculated to reach similar amounts of lipid between the two intervention groups. Moreover, most guidelines recommend a dietary intake ranging from 10 to 25% for monounsaturated fats (MUFA) and from 6 to 11% for polyunsaturated fats (PUFA) (27). As shown in Table 1 of our paper, the prescribed amounts for cashew nut group were 17.38 ± 2.71% for MUFA and 6.41 ± 1.31% for PUFA, and 14.52 ± 2.32% for MUFA and 6.01 ± 1.3% for PUFA for cashew nut oil group, including 30 g/day of cashew nuts and 30 mL/day of cashew nut oil, respectively. These values are in close alignment with the recommended doses, and people can easily consume on a daily basis. There was no statistical difference in energy calculated between the groups (p = 0.959).

Participants were instructed to incorporate cashew nut as a mid-morning snack, while those assigned to the cashew nut oil group were provided with recipes for incorporating the oil into shakes and salad dressings. Members of the cashew nut and cashew nut oil groups (intervention) were explicitly directed not to use the cashew nut or their oil for cooking, roasting, or frying purposes. Additionally, they were advised against consuming olive oil, avocado, or any other nuts aside from the allocated quantity of cashew nut, as well as any other foods with high unsaturated fat content. Control group participants were similarly instructed to refrain from consuming any type of nuts, olive oil, avocado, or other foods high in unsaturated fat.

The primary outcome of the trial was a change in body fat. Secondary outcomes were changes in the values of body weight, BMI, waist, hip and neck circumferences, waist-to-hip ratio (WHR), waist-to-height ratio (WHtR), cardiometabolic (TG, total cholesterol, LDL-c, HDL-c, VLDL-c, ApoA1, ApoB, cortisol, total cholesterol:HDL-c, LDL-c:HDL-c) and liver function markers (AST, ALT, GGT, alkaline phosphatase) after 8 weeks of follow-up.

Body composition was assessed by dual-energy X-ray absorptiometry (Lunar Prodigy Advance DXA System, GE Lunar) and provided fat mass (FM), fat-free mass (FFM), lean mass (LM), and total mass were obtained from the total body and regions, such as trunk, android, and gynoid. The android area is between the ribs and the pelvis, while the gynoid region includes the hips and upper thighs and overlaps the leg and truncal regions. The body composition in percentages was calculated in relation to total body measurements. Body weight was assessed by a bioelectrical impedance analysis device (Inbody 230, Biospace Corp.). Height (meters), waist, hip, and neck circumferences (centimeters) were measured according to standard protocols. BMI was calculated as weight divided by squared height (kg/m²) according to World Health Organization (WHO) (28). WHR was calculated as waist divided by hip circumference, and WHtR was calculated as waist divided by height.

Fasting (10–12 h) venous whole blood samples were collected by a registered nurse at baseline and the end of the study (8-wk) into vacuum tubes containing EDTA as an anticoagulant. Then, blood samples were centrifugated (3,500 r.p.m., 10 min, 4°C), separated in aliquots and stored until analysis. The biochemical determinations were performed by the Hemolab clinical analysis laboratory (Viçosa-MG, Brazil). Trained nursing technicians, specifically employed for this project, conducted the blood collection, obtaining samples ranging from 20 to 30 mL via vacuum. Samples were collected for evaluation of cardiometabolic risk as TG (≥150 mg/dL), total cholesterol (≥240 mg/dL), LDL-c (≥160 mg/dL), HDL-c (<40 or <50 mg/dL for men and women, respectively), and VLDL-c (≥30 mg/dL). Also, apolipoprotein-A-1 (APO-A-1), apolipoprotein-B (APO-B), liver markers such as AST transaminase, gamma GT, ALT transaminase, and alkaline phosphatase were compared as mean and standard deviation between groups. Besides, the atherogenic indices, total cholesterol:HDL-c and LDL-c:HDL-c proposed by Castelli (1988) were calculated.

At baseline and the end of the study, we applied a 24 h recall (24HR) to monitor food consumption during the intervention. The reported intake was analyzed using the 24HR-ERICA software, adapted for the Brazilian population, and the IBGE table (29, 30).

The sample size and study power were determined using the G*Power 3.1 program. For this calculation, a total of 57 volunteers were determined, based on an average estimated effect size derived from clinical studies (0.30), considering statistical analyses for three groups, two intervention points (baseline and endpoint), an alpha value set at 0.05, and a power of 0.80. By adding 20% as a result of losses during follow-up, the total sample size was determined to be 68 participants (Supplementary Figure S1).

For the power of the study, the effect size of 0.28 was calculated from the Eta squared (0.074) based on the values of body fat from our database, an α of 0.05 was used, three groups, two intervention points (baseline and endpoint), and the total sample size of 68 individuals, whom we have information on body fat data. The calculation revealed a study power of 0.94 (Supplementary Figure S2).

To initiate the intervention, after the run-in period, researchers performed the randomization using MinimPy 0.3 program (31). This was achieved through the stratified minimization method, accounting for sex, age, and BMI, with three levels per factor. This approach ensured a well-balanced distribution of potential factors that could interfere with the outcome variables.

Statistical analysis was conducted using SPSS version 22.0 (SPSS, Inc.), and figures displaying statistical analysis were produced using Microsoft Excel. A p-value <0.05 was considered statistically significant. The Shapiro–Wilk test was performed to check the normality of variables. Data are expressed as mean values and standard deviation. Among groups, variable changes were compared by one-way ANOVA followed by Tukey’s post hoc test or using the non-parametric Kruskal–Wallis test followed by Dunn’s post hoc test. To compare differences between baseline and post-intervention within the groups, pairwise tests were performed (paired t-test or Wilcoxon). McNemar’s test was employed to analyze paired nominal data.

Regarding minerals, the content of calcium (CN: 0.37 g/kg vs. OL: 0.01 g/kg) and iron (CN: 64.00 mg/kg vs. OL: 6.10 mg/kg) was higher in cashew nut compared to the oil. Other minerals were not detected in the oil. The oil demonstrated elevated amounts of vitamin E (OL: 2225.93 μg/100 g vs. CN: 1334.02 μg/100 g) and γ tocopherol (OL: 2055.12 μg/100 g vs. CN: 1334.02 μg/100 g) compared to cashew nut. Additionally, β tocopherol, γ tocotrienol, and δ tocotrienol, which were not present in cashew nut, were found in the oil. Conversely, cashew nut exhibited higher levels of total phenolics (CN: 60.45 vs. OL: 2.25 mg GAE (gallic acid equivalent)/100 g) and antioxidant capacity (CN: 15.99 vs. OL: 9.18 uM TE/g sample) in comparison to their oil (Supplementary Tables S1, S2).

Among the participants initially assessed for study eligibility (n = 166), 98 were included and randomly assigned to the following groups: CT (n = 32), CN (n = 32) and OL (n = 34). Of these, 74 participants completed the 8-wk intervention, allocated as follows: CT (n = 20), CN (n = 25) and OL (n = 29). Of these 74 participants, six participants were subsequently excluded due to non-compliance with the prescribed diet as they gained weight. Since all participants were on a low-energy diet, weight loss was expected. Consequently, those who concluded the study with weight gain were excluded due to non-compliance, resulting in the following numbers for analysis in each group: CT (n = 19), CN (n = 24) and OL (n = 25) (Figure 2).

The study population predominantly comprised females (n = 40), individuals with a completed college education/incomplete postgraduate (n = 27), white (n = 33), self-reported single marital status (n = 39), and a family income between 2 and 3 minimum wages (n = 23). Regarding lifestyle habits, the majority did not smoke (n = 65) and did not engage in regular physical activity (n = 43) (Table 2). Concerning cardiometabolic risk at baseline, 19 (79.2%), 15 (78.9%), and 23 (92%) individuals had obesity, while 13 (54.2%), 11 (57.9%), and 13 (52%) had dyslipidemia in the CN, CT, and OL groups, respectively (data not shown).

Following the 8-wk intervention, all participants in the study demonstrated a reduction in energy intake (−205 kcal; p = 0.026), indicating adherence to energy-restriction (data not shown). However, this reduction was not as substantial as expected (−500 kcal). When examining the groups individually, the control group exhibited a reduction in the intake of saturated fat (SFA), while the cashew nut group experienced a decrease in polyunsaturated fat (PUFA) and α-linolenic acid (Table 3).

After 8-wk of intervention, all groups presented significant reduction in body fat (kg), with consequent weight-loss (CT: −4.4%; CN: −4.1%; OL: −3.5%). The participants also had significant reduction in other adiposity indicators as: weight-loss, WC (cm), hip circumference (HC) (cm), and WHtR. Concerning WHR and body fat (%), significant losses were observed only in the control and cashew nut groups. Both intervention groups (cashew nut and oil) exhibited a significant reduction in neck circumference. No differences were found between groups, except for android fat in the endpoint between CT and OL groups (CT: 9.5 ± 1.3%; OL: 8.1 ± 1.5%) (Table 4). Additionally, after the intervention, there was a decrease in the number of individuals with obesity among those who consumed cashew nut (19 (27.94%) vs. 13 (19.12%); p = 0.032) (Supplementary Figure S3).

After 8-wk intervention, all groups reduced total cholesterol and GGT. In the intervention groups, both the cashew and oil groups had reductions in apo B, while those consuming only the oil experienced reductions in LDL-c and the atherogenic index (total cholesterol/HDL-c). The control and cashew nut groups observed reductions in TG and VLDL-c. In terms of liver enzymes, the cashew nut group demonstrated reductions in AST and ALT. No differences were found between groups (Table 5).