The COSMOS trial demonstrated that the daily intake of monomeric and polymeric flavanols from cocoa in older adults reduces risk for cardiovascular morbidity and mortality. Although the COSMOS study utilized a flavanol supplement compared to a whole food, it is a case study to support the need for larger trials with clinical outcomes based on the use of multi-site data of surrogate outcomes from dietary interventions that use randomized, double-blind controlled trials in crossover or parallel-arm study designs for studies of nuts or berries. A common study design for whole foods is the replacement of the test food with a nutritionally matched, isocaloric substitute. However, matching nutritional content can be a challenge because food processing, such as blending berries and roasting nuts, causes a disruption to the nutrient matrix, potentially changing the bio-availability of key nutrients. For nuts, controls often include the complete omission of the nut of interest. For berry research, a number of considerations exist that are alternative to consuming the whole food. One is the use of freeze-dried berry powders as the test product, controlled with an isocaloric powder either lower or devoid of potential bio-actives. Attempts have been made to mask the control powders, but issues such as product color, texture, scent, and mouth feel are challenging to completely match. Although this approach is similar to a classical pharmaceutical trial design, blinding study personnel and participants is challenging, thus creating both performance and detection bias. Additionally, freeze-dried berry powders can have a different food matrix compared to the whole food,drainage collection pot which could influence outcome measures as well as limit generalizability to the whole fruit.

A second approach for berry research is the encapsulation of test and control powders. This can aid in participant masking, but the total amount of test product provided can be limiting, and large intakes of control gelatin capsules have resulted in adverse effects. A third option can be examining 2 or more intake levels, with or without a true control group. Finally, the use of macro- and micronutrient matched gummies with similar amounts of calories, sugars, and fiber, but devoid of other bio-actives, is a novel option for use as a comparative control. In all of these approaches, the potential bioactivity of the control itself must be considered. For example, isocaloric control powders that are lowin polyphenols may still have a considerable amount of fiber in order to obtain similar mouth feel and texture, but the fiber content may have effects on lipid metabolism and the microbiome, which could influence outcome measures. Multiple cultivars of berries exist, some of which have differences in the content of bio-active ingredients, thus limiting comparison and extrapolation of results . For nuts, walnuts contain a variety of phenolic acids, catechins, and flavonoids, most of which have been reported to possess bioactivity. Significant differences in the concentration of 16 phenolic compounds were identified when comparing black and English walnuts. More than 50 cultivars of strawberries exist in the United States. To help reduce the potential experimental variability created with the use of different cultivars, the California Strawberry Commission has produced a freeze-dried test material that utilizes a composite of genotypes to produce a powder that is characterized for its macro- and micronutrients and bio-active components. The US high bush Blueberry Council also provides a powder that is a 50/50 mixture of 2 cultivars.

A limitation of this approach is that the standardized mixture may contain varieties with reduced or low bio-activity. However, the advantage of this approach is that the composite represents the “market basket” available to consumers and allows comparison of results from studies conducted among different research groups and generalizability of results to a broader berry application actually used by consumers. In addition to cultivar differences, factors such as climate and seasonal differences due to heat, sunlight, and rainfall can contribute additional variability. New analytical equipment and techniques have increased the precision of food composition compared to analyses performed decades ago. Current advances in the development of nutrition databases have been reviewed elsewhere. For example, databases such as that from the USDA FoodCentral could be strengthened if the date of the analyses was included, along with the protocols used and the number of samples analyzed. Linking resources from repositories detailing data, such as chemical composition and bio-activity, will help both plant scientists and health professionals to make accurate and timely recommendations and guide future research.Free-living populations have differences in background diets that can influence their responses to the intake of test foods, potentially creating significant variation in baseline measurements. This variability presents a challenge when elucidating clinically relevant effects, especially if unknown a priori, where statistical significance can be masked by combining and analyzing groups together. Inter individual variability may be mitigated by increasing sample size as well as using a crossover design, but challenges in recruitment, retention, and budget constraints exist. One way to help minimize experimental variability is through a run-in period to identify participants who may be differentially metabolizing bio-active phenolics or with the goal of minimizing or removing potentially confounding metabolites from circulation prior to the intervention. However, study designs that employ highly controlled settings, strict inclusion and exclusion criteria, extended washout periods that alter background diets, and ask participants to follow an atypical consumption pattern does not reflect “normal” life and may have limited applicability to the general population.

Another useful model that also has limitations is the provision of nuts or berries in amounts and duration that are greater than normally consumed. Feeding relatively high amounts of nuts or berries for a limited period of time has been employed to demonstrate proof-of-concept and provide a basis for further exploration for changes in physiology, cognitive performance, and gut microbiome profiles. Subsequent study designs must be realistic, guided by the USDA FoodCentral database for portion size. These trial designs should also use a duration that is realistically achievable by consumers, whose food purchasing behavior can be influenced by cost, access, and seasonal availability of the food. Studies using average daily portion sizes typically require intervention periods of months, which present challenges regarding participant compliance and retention and cost of the study. In a review of 231 reports on berries and health, approximately 70% of studies used interventions of less than 3 mo or contained less than 50participants. Meeting the challenge of conducting long-term studies using amounts of foods in a typical diet, with a representative sample of participants, requires a significant commitment of resources. The health and functional levels of participants are other factors that influence study designs and outcomes. For example, studies on cognitive performance with both nuts and berries have assessed effects among those both with and without cognitive impairments. In such studies, short-term interventions may show little or no response after the addition of nuts or berries to the diet. Although the net change may not be statistically significant, this model does not address the ability of the food to prevent decline, which would require long-term testing. Further, an individual with cognitive impairments might demonstrate favorable responses compared to baseline measures following nut or berry intake but may still not reach the level of performance of a healthy individual. In both instances, neither change from baseline,square plastic pot nor absolute values of performance, fully captures the beneficial cognitive response. Dietary interventions require the incorporation of foods into an individual’s eating pattern, which may present a number of challenges. One is the creation of boredom with eating the same food on a regular basis. Second is that the caloric load of the test nut or berry may displace the intake of other nutrient-dense foods. These factors may make compliance for the entire study duration an issue, particularly if the intervention is weeks or months in duration. A third challenge involves compliance. In berry research studies, compliance is often not reported, or the reported range of intake is so variable that it is hard to discern the significance of the results. The use of food intake metabolite markers is an emerging tool that can help verify compliance. In addition to compliance, dietary patterns are an important consideration needed for the interpretation of results because individuals do not eat a single food in the absence of other foods. Background or habitual intake is often not addressed in nutritional trials. The potential variability in habitual dietary intake of participants is often a confounding factor in nutrition research. Dietary assessment methods, with 24-h recalls, 3-d food records, and food frequency questionnaires, all have limitations.

These subjective measures may also not accurately capture the potential for nutrient-nutrient interactions that may alter polyphenolic or other bio-active components attributed to nut and berry consumption. Further complicating this issue is the observation that study designs utilizing longer-term interventions or that require the intake of a large amount of the test food are more likely to result in over reporting food intake due to fear that participants may be dismissed from the intervention. Innovations in dietary assessment methodology using “smart” eyeglasses or other image-based technologies have been proposed to address this issue. Assessing the relationship between the intake of nutrients and bio-actives from a whole food product to physiologic responses is difficult, as a multitude of processes are affected, including regulation of vascular function, provision of oxidant defense, and changes in gut microbiome profiles and subsequent output of secondary metabolites. Additionally, bio-actives from nuts and berries can interact with each other as well as other dietary components to alter bio-availability and health-promoting properties. For example, intake of dietary fats in conjunction with berries has been demonstrated to increase carotenoid bio-availability .Results could also be confounded by dietary changes made by participants in addition to incorporation of the test nut or berry. Habitual dietary intake is often measured through food frequency questionnaires or repeated 24-h dietary recalls. However, these subjective measures may not accurately capture the potential for nutrient-nutrient interactions that may alter polyphenolic or other bio-active components attributed to nut and berry consumption. Further complicating this issue is the observation that study designs utilizing longer-term interventions or that require the intake of a large amount of the test food are more likely to result in over reporting food intake due to fear that participants may be dismissed from the intervention. Expanding the scope of populations to be studied is another key area for future research. Most clinical trials using nuts and berries have been conducted in middle-aged or older Caucasian adults with one or more cardiometabolic risk factors. Whether these results extend to other population groups is either inferred or unknown. Future research would benefit from extending the study populations to include those from other racial and ethnic groups. This is particularly important in order to address the current NIH research initiative in precision nutrition and health, the “Nutrition for Precision Health powered by the All of Us Research Program” . The inclusion of biological females in clinical nutrition trials is imperative, yet the current literature includes predominantly male participants. Because many studies on nuts and berries focus on cardiometabolic outcomes, the unique aspects of female physiology must be considered. For example, vascular function fluctuates with the phase of the menstrual cycle, which has largely been ignored in most past studies. More studies are also needed in young children as well as in young adults up to about the age of 40. A pilot study reported a correlation between blueberry supplementation and acute positive effects on memory and executive function in 7- to 10-y old children. A large study among pregnant women-infant dyads reported positive protective neuropsychological effects on long-term cognitive development in children at 1, 5, and 8 y of age when nuts were consumed during gestation. Finally, translation of research results is challenging when considering socioeconomic status , particularly when food items are not accessible or affordable. Barriers to participation in clinical research studies among those of low SES include a low interest in clinical trials, inefficient or inadequate explanation of the study in culturally appropriate terms, participants’ distrust of biomedical research, and participant burden, including lack of transportation or the inability to prioritize participation in research over work obligations .Like many other dietary studies, research on nuts and berry studies often use acute studies evaluating postprandial effects. However, either a lack of or successful demonstration of benefits does not necessarily predict a similar outcome over extended periods of intake.