To explore the effects of peanut consumption on fecal energy excretion with a balanced, non-vegetarian diet.
Four arm parallel group design (that is, whole peanut (P), peanut butter (PB), peanut oil (PO) or peanut flour (PF) consumption) with one crossover (control and intervention).
In total 63 healthy men and women from Ghana, Brazil and USA (N=15–16 per group) with an average body mass index of 21.8 kg m−2.
Percent fat of fecal wet weight daily energy excretion during the control and the treatment periods.
Compared to control, the percentage of fat in the feces increased significantly for the P group (5.22±0.29%) relative to the other three groups ((PO=3.07±0.36%, PB=3.11±0.31% (P=0.001), and PF=3.75±0.40% (P=0.019)). The same findings held for kJ g−1 of feces excreted. During the P supplementation period, the energy excretion was 21.4±1.0 kJ g−1 versus 18.7±1.0 kJ g−1 for PO (P=0.034), 18.8±0.7 kJ g−1 for PB (P=0.042) and 18.5±0.8 kJ g−1 for PF (P=0.028).
Fecal fat and energy loss is greater with consumption of whole peanuts compared to peanut butter, oil or flour. This may contribute to the less than predicted change of body weight observed with peanut consumption. There were no cultural differences.
The Adventist Health Study,1 Nurses' Health Study,2 Physicians' Health Study,3 Iowa Women's Study,4 1994 CSFII (USDA) and other observational data reveal an absence or inverse association between chronic nut consumption and body weight. Numerous clinical trials primarily designed to explore associations between nut consumption and indices of cardiovascular disease risk (for example, LDL-cholesterol, triacylglycerol), also suggest nut consumption promotes little or no weight gain. This has been replicated with peanuts,5 almonds,6, 7, 8 pecans,9 walnuts,10, 11, 12 and macadamia nuts.13
To account for these observations, it has been proposed that nuts elicit dietary compensation, increased energy expenditure and/or inefficient energy absorption. The literature consistently indicates nuts have strong satiety properties14 and evoke strong energy compensation5, 6, 7, 8, 11, 12, 13, 14 (Table 1). Thus, the majority of energy provided by the inclusion of nuts in the diet is offset by spontaneous dietary adjustments. Nevertheless, most trials reveal dietary compensation is not perfect so, there is a residual positive energy balance.
Some data suggest increased energy expenditure associated with chronic nut ingestion may account for another fraction of the energy contributed by nuts. One trial reported an 11% increase of resting energy expenditure with 19 weeks of peanut consumption.5 Another recent trial noted a 5% increase of resting energy expenditure among obese, but not lean, males and females during 8 weeks of daily ingestion of milkshakes containing 300 kcal of peanut oil.15 However, others have failed to document such an effect.6 Changes of TEF and physical activity have not been observed.5, 6, 8 Thus, the evidence is presently equivocal regarding an effect of chronic nut consumption on energy expenditure.
A third mechanism that may contribute to the lack of effect of nut intake on body weight entails poor bioaccessibility of energy from nuts. Fecal analyses from intervention trials testing peanuts,16 pecans,17 and almonds18, 19, 20 document increased energy loss with nut consumption. The effect size varies markedly, as did the loads provided, hampering conclusions about the magnitude of effect. Nevertheless, some increased loss is observed consistently and is attributed to inefficient mastication coupled with resistance of the parenchymal cell walls to subsequent bacterial and enzymatic degradation.20 The only work showing increased fecal fat loss with peanuts, the most widely consumed ‘nut’ in the US (ERS/USDA, 2002), was based on a trial where the peanuts provided 95% of dietary fat intake.16
The aims of the present trial were to further explore the effects of peanut consumption on fecal energy excretion with a balanced, non-vegetarian diet using a more ecological load. Mechanistic insights were also sought. Thus, losses following ingestion of whole peanuts and peanut butter were contrasted to determine the importance of nut form. In addition, losses following ingestion of defatted peanut flour and peanut oil were compared to determine the contribution of the carbohydrate (including fiber) and lipid fractions. Further, samples collected from North Americans (West Lafayette, IN, USA), Ghanaians (Accra) and Brazilians (Viçosa) were evaluated to determine whether discrepant cultural patterns of peanut use altered their biological processing.
Materials and methods
Sixty-four (32 from Brazil, 16 from Ghana and 16 from USA) males and females (24.3±0.5 years old) satisfied the eligibility criteria of healthy, body mass index between 18.5 and 29.5 kg m−2 (21.8±0.3 kg m−2), weight stable for the past 6 months (61.9±1.23 kg), not taking medications affecting hunger or excretory function, constant level of physical activity (no deviation 30 min, more than 5 times per week), non-smoking, non-pregnant, non-lactating and no peanut allergy. One male was excluded from the study analyses due to non-compliance, resulting in a sample size of 63 subjects. The protocol was approved by each country's Institutional Review Board (IRB).
The study consisted of two sessions of controlled feeding for each participant; one served as a non-intervention control and the other was one of four treatments. All the study foods were tasted and rated as a screening step. In addition, weight, height, blood pressure, heart rate, waist and hip circumferences were measured. Participants were instructed to keep the same level of activity for the duration of the study. During the control period, they were provided a diet without peanut products for 7–9 days. The study duration was based on individual variability in bowel habits. A food dye marker was given in the form of pills (Red Carmine, and FD&C 13% blue aluminum lake, Sensient Technologies Corporation, St Louis, MI, USA) on day one of the control period, and again 3 days after the first marker was passed. Consecutive stool collections were conducted after appearance of the first food marker in feces until the second marker was passed (typically a 4-day collection). Following the baseline period, participants were randomly assigned diets containing 70 g of peanut products, as whole peanuts (Nabisco Inc., East Hanover, NJ, USA), creamy peanut butter (Kroger Company, Cincinnati, OH, USA), peanut oil (Nabisco Inc., East Hanover, NJ, USA) or peanut flour (Bio Separations, Washington, DC, USA) for a seven- to nine-day period, depending on their bowel function. The same protocol as described for the control period was applied during the treatment period. Dietary compliance was assessed by random finger stick blood glucose tests prior to meals. Additionally, upon arrival each day, participants were asked to expectorate into a small cup and were told that the sample would be used for compliance analysis. However, no such test was conducted. Blood glucose values between 3.6 and 6.1 mmol (fasting state) were indicative of subject compliance with the study protocol.
One diet was developed for the control periods, and this diet was modified to include peanut butter (PB), peanut flour (PF), peanuts (P), or peanut oil (PO) for the intervention periods. Each country included foods familiar to their respective culture. The diets provided a mean of 10 291±44 kJ per day with 55/30/15% (±3%) of energy from carbohydrate, fat and protein, respectively, as determined by local nutrient databases. Participants were served a four-day menu repeated over the length of each treatment period. They were required to eat all the foods served in the laboratory; small fat-free evening snacks were optional. Participants were instructed to return any uneaten snack the next day. Dietary energy, macronutrient and fiber intakes are presented in Table 2.
Participant weight was measured barefoot and in street clothes on a body fat analyzer (Tanita Corp. of America, Arlington Heights, IL, USA). Height was also measured with participants in bare feet in the standing position. Body mass index was computed based on these two measures (kg m−2). Waist circumference was measured midway between the lowest rib and the umbilicus. Hip circumference was measured at the point of greatest circumference around the hips and buttocks. (Anthropometric standardization reference manual, Champaign, IL, USA: Human Kinetics Books.) Blood pressure was measured with a mercury sphygmomanometer (Moore Medical Corp., New Britain, CT, USA) and heart rate was determined by radial palpation.
Fecal collection and analysis
Participants were instructed to collect all stool after appearance of the first dye marker in the feces until the second marker was passed. They were provided with plastic urine hats, lidded cups and thermal boxes to assist them with quantitative collections. In addition, a private facility with the materials needed for fecal collection was provided on campus for convenience. Samples were immediately frozen after collection.
Fecal composites from each person and treatment period were made by adding deionized water to thawed feces (water/stool, 2:1 w w−1). The mixture was blended thoroughly, and the homogenized feces were placed into seven centrifuge tubes. Four tubes were freeze-dried, while the remaining three were stored at −20 °C. All analyses were conducted in triplicate. One-gram fecal pellets were made for gross energy analyses by bomb calorimetry (Bomb calorimeter 1280, Parr instruments Moline, IL, USA). Fat content was determined by a modified method of Folch extraction (Folch 1957). Dried feces (0.5 g) were refluxed in chloroform/methanol (1:1) for 4 h at 65 °C. The homogenate was then centrifuged at 2000 r.p.m. for 10 min and the pellet was extracted three additional times. The final pellet was dried under nitrogen and the weight was recorded. The difference between the initial and final weight of feces was equal to the amount of fat in the sample. Percentage of fecal fat in wet feces was then calculated.
Descriptive statistics, including mean and s.e.m. were used to describe the distributions of all variables. Parametric repeated measures analysis of variance (ANOVA) was used to examine treatment effects of the peanut products. The criterion for statistical significance was P<0.05, two-tailed. Within treatment comparisons were conducted using paired t-tests. Statistical analyses were performed with the Statistical Package for the Social Sciences version 14.0 (SPSS Inc. Chicago, IL, USA).
Across country comparisons were performed using non-parametric tests.
There was no significant difference between groups during the control period for energy intake (10 172±24 kJ day−1). During the treatment period, the PO group received slightly, but significantly less food (9909±268 kJ day−1) than the PB (10 460±63 kJ day−1, P=0.011); the PF (10 694±22 kJ day−1, P<0.001) and the P (10 598±99 kJ day−1, P=0.002) groups.
Within group comparisons showed the PO was the only treatment that led to no significant difference between the control and treatment periods. There was an increase in energy intake when PB (P=0.002), PF (P<0.001) or P (P<0.001) were introduced into the diet compared to the control period with no peanut products.
Fecal wet weight and fecal fat weight
There was no difference between groups for fecal wet weight. However, there was a significant increase noted between the control (142.6 g d−1) and the PB (183.4 g d−1) periods (P=0.015).
No change was observed within groups for the absolute amounts of fat excreted for the PO, PB and PF groups. There was a significant increase for the P group when whole peanuts were provided (Table 3).
kJ g−1 of feces excreted
There was no between group difference during the control period for fecal energy density (Figure 1). However, there was a significant difference between the P group and the three other groups during the treatment period. During P supplementation, energy excretion was 21.4±1.0 kJ g−1 versus 18.7±1.0 kJ g−1 for PO (P=0.034), 18.8±0.7 kJ g−1 for PB (P=0.042) and 18.5±0.8 kJ g−1 for PF (P=0.028). There was no within group difference for the PO, PB and PF groups. However, the kJ g−1 excreted during the control period was smaller than during the P period (18.9±0.9 and 21.4±1.0 kJ g−1, P<0.001). GI transit times were similar between countries.
Percent fat of wet feces
During the control period, the percentage of fat of wet feces was significantly lower for the PO group (2.90±0.30%) than the PB (3.99±0.29%) and the PF group (4.27±0.47%). P-values of 0.044 and 0.013, respectively. No other difference was noted during the control period between groups (Figure 2). When the peanut products were provided, the percentage of fat in the feces increased significantly for the P group (5.22±0.59%) relative to the three others (PO=3.07±0.36%, PB=3.11±0.31% (P=0.001) and PF=3.75±0.40% (P=0.019)), (Figure 3).
Within group comparisons (Figure 4) revealed a difference between the control and the treatment periods for the PB, PF and P groups. There was a decrease in the percentage of fat excreted when PB (3.99±0.29 to 3.11±0.31% (P=0.015)) and PF (4.27±0.47 to 3.75±0.40% (P=0.036)) were ingested, and an increase when P (3.87±0.43 and 5.22±0.59% (P=0.022)) was consumed.
Absolute energy excretion
No significant difference was revealed within groups in daily fecal energy loss. For the PO group, absolute energy excretion per day rose from 626.6±35.9 to 704.7±78.0 kJ day−1. For PB it increased from 623.4±58.4 to 703.0±74.6 kJ day−1. There was an increase of 153.6 kJ day−1 (646.6±55.1 to 800.2±63.5 kJ day−1) when P was consumed. And energy excretion remained unchanged when PF was consumed (663.6±63.6 to 668.9±60.3 kJ day−1) (Table 4).
Comparisons within countries
BRAZIL: Within group comparison revealed no difference for the PO group. Percent fat of fecal wet weight was greater during the control than during the treatment period for the PB group. During the PF supplementation, the energy intake, the total kJ day−1, and the weight of dry feces was greater than during the control period. For the P group, percent fat of fecal wet weight, kJ g−1, grams of fat excreted and energy intake were greater during the treatment period than the control.
GHANA: Food intake was greater during the P, the PF and the PO supplementation periods compared to the control period. No other within group difference was noted.
USA: No within group differences were observed.
Between group comparison for the three countries
BRAZIL: Food intake for the PO group was significantly less than for the PB, the P and the PF groups during the treatment period. There was no difference in energy intake between the P and the PB groups during treatment. However, it was significantly greater when PF was consumed compared to the PB and the P supplementation periods.
During the PF feeding period the weight of dry feces and the total energy excretion were greater than during the PO and the PB feeding periods. The percent fat of fecal wet weight, the fecal energy density (kJ g−1) and the weight of fat excreted during the treatment period were greater when whole peanuts were consumed compared to PO, PB and PF. The total energy excretion was also greater during the treatment period for the P group than for the PB group.
GHANA: During the treatment period, the PO group consumed less food than the PB, the PF and the P groups. When P was provided, food intake was greater than when PB and PF were provided, and the PF group consumed more food than the PB group during the treatment period. The percent fat of fecal wet weight was greater for the PF group than the PB and the P group during the treatment period.
USA: Food intake during the treatment period was greater for the PF than the PO, the PB and the P groups. The percent fat of fecal wet weight was greater during the peanut supplementation period than the PO, PB and PF feeding periods.
This study examined the effects of 70 gram portions of different peanut products (oil, flour, butter and whole peanut) fed to humans on a daily basis as part of a balanced, non-vegetarian diet. Whole peanut consumption led to a greater fecal fat percent of wet weight and energy per gram of feces compared to the PF, PB, PO or a no peanut control diet. The mean energy losses were 153.7 kJ day−1 for P, 78.1 kJ day−1 for PO, 79.6 kJ day−1 for PB and 5.3 kJ day−1 for PF.
Comparisons across countries are not viewed as reliable. The sample size from each country was small hampering statistical analyses. In addition, although all groups followed a common protocol, there were some differences in the local foods used and this could have modified the excretion data.
Our fecal fat excretion data are, in part, consistent with the findings of Levine and Silvis,16 who used a similar protocol but fed twice as much whole peanut (153 g) and peanut butter and 76 g of peanut oil as part of a vegetarian diet. Both trials observed the greatest losses with the whole nut. However, Levine and Silvis16 reported greater percent fat losses in the stool during peanut consumption than observed here. This may be due to the different quantity of peanuts provided and higher total fiber content of the diet. Unlike our findings, they also noted a greater fat loss with peanut butter compared to peanut oil. However, they provided two times as much peanut butter as oil (153 g of PB and 76 g of PO) whereas our diets contained 70 g of PB and PO.
During the control period, the PO group excreted less fat (2.90%) than the PB (3.99%) and PF (4.27%) groups. Fecal fat excretion is variable among healthy individuals, typically ranging from 3 to 6%, but nut challenge studies have reported concentrations as low as 1.7%.18 Thus, all baseline values were generally within normal ranges. There were no differences between the PO, PB and PF groups during active treatment. In fact, there was a decrease in percent fat of fecal wet weight relative to the control period for the PB and PF interventions. During both of these interventions, the fecal wet weight increased while absolute grams of fat remained relatively constant. The fecal wet weight for the PO group increased along with the fecal fat content. Consistent with the increase in percent fecal fat loss when whole peanuts were consumed, the absolute weight of fat excreted increased significantly, while there was no significant increase in fecal wet weight.
To our knowledge there is no study besides the study of Levine and Silvis16 that investigated the effect of peanut and peanut product consumption on fecal fat or energy excretion. However, several studies with other nuts are in agreement with the current study. Zemaitis and Sabate18 measured fecal fat content after no almond, low almond, or high almond diets. Absolute amounts of almonds consumed were not reported, but they constituted 0, 10 and 20% of total energy respectively. For comparative purposes, if total energy intake is assumed to be approximately 10 460 kJ day−1, this would translate to 42 and 48 g of almonds. Fecal fat content almost doubled after the low almond diet compared to the control diet (1.7 versus 3.2%), and increased further with the high almond diet (4.1%). Haddad and Sabate17 noted 3 and 8% fat loss in the feces when subjects consumed a control diet and one containing 31–40% of total fat supplied by pecans. In the current study, we observed a 3.87% fat loss during the control period and a 5.22% fat loss when 70 g of whole peanut were consumed. This is approximately 37.6% (range 28.9–42.1%) of the total fat provided by the peanuts. The differences across studies in fecal fat excretion may be due to characteristics of the nuts as well as the duration and efficiency of mastication.
When whole peanuts were consumed, total energy loss increased by 154 kJ compared to the control period while fat energy loss increased by 84 kJ. Thus, fat loss accounts for 54.6% of total energy loss in the feces, while the nature of the remaining 70 kJ or 45% was not characterized. Prior work with almonds documented that intact nut parenchymal cells were lost in the stool.4 Considering the macronutrient composition of peanuts (about 50% fat and 45% carbohydrate and protein) it is possible that the extra 70 kJ excreted came from the protein and carbohydrate components of peanuts (Table 2).
A likely mechanism for the increased fecal fat loss associated with whole peanut consumption is reduced lipid bioaccessibility. The failure to rupture the fibrous cell walls of nuts by mechanical, enzymatic or bacterial processes results in limited release of the cellular lipid and their consequent loss in the stool.20 Studies documenting this phenomenon have only been conducted with almonds, but the consistency of results across nuts with respect to fat loss, suggests a similarly low bioaccessibility that may apply to these other nuts.
The loss of fat and energy in the stool with peanut consumption may be one of the elements protecting consumers against weight gain. The largest contributor is dietary compensation, which accounts for approximately two-thirds to three-quarters of the energy provided by nuts (Table 1). The low bioaccessibility would offset another 9–10%. A role for increased energy expenditure is less well documented,5, 6, 15 but may also be involved. The epidemiological data reveal an inverse association between frequency of nut consumption and body mass index while the clinical evidence indicates that the incorporation of nuts in the diet promotes little or no weight gain with peanuts,5 almonds,6 pecans9 or walnuts.10
In conclusion, there was a significant increase in fecal fat and energy loss in participants fed a non-vegetarian diet supplemented with 70 g of whole peanuts. This loss was not noted when the participants were supplemented with peanut oil, flour or butter. This result is consistent with the finding that there is limited bioaccessibility of the oil in nuts due to inefficient mastication and digestion. This loss of energy likely contributes to the less than predicted effect of nut consumption on body weight and complements evidence that nut consumption may reduce the risk of cardiovascular disease through favorable effects on lipid profiles.5, 6, 21, 22, 23