OBJECTIVE: To evaluate the effect of an almond-enriched (high monounsaturated fat, MUFA) or complex carbohydrate-enriched (high carbohydrate) formula-based low-calorie diet (LCD) on anthropometric, body composition and metabolic parameters in a weight reduction program.
DESIGN: A randomized, prospective 24-week trial in a free-living population evaluating two distinct macronutrient interventions on obesity and metabolic syndrome-related parameters during weight reduction.
SUBJECTS: In total, 65 overweight and obese adults (age: 27–79 y, body mass index (BMI): 27–55 kg/m2).
INTERVENTION: A formula-based LCD enriched with 84 g/day of almonds (almond-LCD; 39% total fat, 25% MUFA and 32% carbohydrate as percent of dietary energy) or self-selected complex carbohydrates (CHO-LCD; 18% total fat, 5% MUFA and 53% carbohydrate as percent of dietary energy) featuring equivalent calories and protein.
MAIN OUTCOME MEASUREMENTS: Various anthropometric, body composition and metabolic parameters at baseline, during and after 24 weeks of dietary intervention.
RESULTS: LCD supplementation with almonds, in contrast to complex carbohydrates, was associated with greater reductions in weight/BMI (−18 vs −11%), waist circumference (WC) (−14 vs −9%), fat mass (FM) (−30 vs −20%), total body water (−8 vs −1%) and systolic blood pressure (−11 vs 0%), P=0.0001–0.05. A 62% greater reduction in weight/BMI, 50% greater reduction in WC and 56% greater reduction in FM were observed in the almond-LCD as compared to the CHO-LCD intervention. Ketone levels increased only in the almond-LCD group (+260 vs 0%, P<0.02). High-density lipoprotein cholesterol (HDL-C) increased in the CHO-LCD group and decreased in the almond-LCD group (+15 vs −6%, P=0.05). Glucose, insulin, diastolic blood pressure, total cholesterol, triglycerides, low-density lipoprotein cholesterol (LDL-C) and LDL-C to HDL-C ratio decreased significantly to a similar extent in both dietary interventions. Homeostasis model analysis of insulin resistance (HOMA-IR) decreased in both study groups over time (almond-LCD: −66% and CHO-LCD: −35%, P<0.0001). Among subjects with type 1 diabetes, diabetes medication reductions were sustained or further reduced in a greater proportion of almond-LCD as compared to CHO-LCD subjects (96 vs 50%, respectively).
CONCLUSION: Our findings suggest that an almond-enriched LCD improves a preponderance of the abnormalities associated with the metabolic syndrome. Both dietary interventions were effective in decreasing body weight beyond the weight loss observed during long-term pharmacological interventions; however, the almond-LCD group experienced a sustained and greater weight reduction for the duration of the 24-week intervention. Almond supplementation of a formula-based LCD is a novel alternative to self-selected complex carbohydrates and has a potential role in reducing the public health implications of obesity.
The epidemic of overweight and obesity among affluent nations is fueling the increasing prevalence of the metabolic syndrome and type II diabetes, which has serious implications for healthcare systems and the economy. Metabolic syndrome encompasses a cluster of abnormalities including insulin resistance, central obesity, hypertension and dyslipidemia.1 Both type 2 diabetes and the insulin resistance syndrome are associated with a marked increased risk for cardiovascular disease (CVD)2 and have become a significant clinical challenge for practitioners. Contributors to these conditions are improper nutrition and inadequate physical activity; thus weight management is a cornerstone of treatment and prevention.
The effects of dietary fat on adiposity have been intensively debated in recent years. Obesity experts have argued that habitual consumption of fatty diets contributes to the development and maintenance of excess body weight.3,4,5,6,7,8 Others have noted that diets high in fat do not appear to be the primary cause of the high prevalence of excess body fat within the United States (US).9 Emphasis on dietary fat reduction within the US public health dietary guidelines has yielded an increased consumption of processed grains, reduced-fat and fat-free products.10 Failure to consider portion sizes and total energy intake during the past two decades has contributed towards an increased risk for obesity, type 2 diabetes and CVD.11,12
According to a recent meta-analysis from the Cochrane Collaboration, low-fat diets (20% of energy or less) have poor clinical effectiveness in the outpatient treatment of obesity.13 Obese individuals enrolled in long-term weight reduction programs frequently maintain a steady state after 12 weeks of dietary intervention,14 possibly due to waning compliance and diminished energy requirements.15,16 An ideal weight reduction dietary regimen would create an optimal hormonal milieu to improve the obese individual's metabolic abnormalities,17 mobilize abdominal adipose tissue and promote compliance. Thus, studies examining the effects of other dietary macronutrient compositions and modifications to the traditional formula-based LCD are of considerable scientific and public health interest.
Dietary fatty acids (saturated, mono- and poly-unsaturated) are unequal in their ability to improve metabolic parameters associated with the metabolic syndrome and coronary heart disease (CHD).18 Previous studies featuring isocaloric diets have shown that plasma concentrations of total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C) typically improve when monounsaturated fat (MUFA) replaces saturated fat in dietary patterns.19,20 Incorporating a preponderance of MUFA under hypocaloric conditions may further improve the metabolic syndrome through multiple mechanisms, for example, decreased triacylglycerol concentrations,21 decreased LDL-C to HDL-C ratio14 and attenuated day-long hyperinsulinemia.22,23 In an effort to reduce the incidence of CHD, the current National Cholesterol Education Program guidelines24 emphasize a more liberal approach in total and MUFA consumption, similar to Mediterranean food patterns (total fat=40–45% of energy, rich in MUFA (specifically oleic acid)). In addition to their high levels of MUFA, and to a lesser degree polyunsaturated fatty acids (PUFA), nuts have been shown through epidemiological25,26 and interventional studies27,28,29,30,31,32,33 to confer beneficial effects on reducing cardiovascular risk. The possible association between high PUFA diets and carcinogenesis,34,35 in addition to potential palatability and compliance issues, pre-empted consideration of enriching a formula-based low-calorie diet (LCD) with large amounts of a high PUFA oil. In light of their high levels of oleic acid, as well as textural characteristics, whole unblanched unsalted almonds were chosen to evaluate their influence on anthropometric, body composition and metabolic parameters in one of two different enriched formula-based LCDs (complex carbohydrate or almond) during a 24-week weight reduction intervention.
Subjects were recruited from the pool of outpatients entering into the 24-week Diabetes and Cardiovascular Risk Reduction Program (D & CVRRP) for medically supervised weight reduction at City of Hope (COH) National Medical Center in Duarte, California. Acceptance into the D & CVRRP requires a medical diagnosis that can be ameliorated or improved by weight reduction, age ≥18 y, and body mass index (BMI) ≥25 kg/m2. Patients on lipid-lowering medications and women receiving hormone replacement therapy were excluded from study participation. The study was conducted from February 2000 to May 2002. The study protocol was approved by the Institutional Review Boards of COH and Loma Linda University and all participants gave written informed consent.
Subjects were randomized to consume almonds (almond-LCD) or self-selected complex carbohydrates (CHO-LCD) using computer generated Random Number Generation software (RNDGEN, Stanford, USA) with stratification according to gender and physician documented presence or absence of type 2 diabetes. Study recruitment led to a population of male and female subjects with BMI ranging from 27 to 55 kg/m2 and a median age of 55 y (range 27–79 y). In total, 65 subjects were randomized into the two intervention groups (Figure 1) and 52 patients completed the 24-week study. The number of withdrawals from the two groups was similar, eight from the almond-LCD group and five from the CHO-LCD group. Primary reasons for withdrawal included work and time conflicts. No significant difference in baseline demographic, anthropometric or metabolic characteristics were found in the two intervention groups after randomization (Table 1).
The total dietary intervention period for each subject was 24 weeks, preceded by a 2-week run-in with no intervention other than a multivitamin/mineral supplement. Subjects consumed a formula-based LCD supplemented with 84 g/day almonds or self-selected complex carbohydrates and safflower oil (see below). The two study groups had distinctly different levels of total and MUFA; however, both groups were equally balanced on total calories, protein, cholesterol and saturated fat (Table 2). Health Management Resources (HMR) 70 Plus, a protein-sparing formulation prescribed during LCDs to ameliorate the loss of lean body mass, was prescribed according to D&CVRRP guidelines. Subjects were instructed to mix the HMR powder according to package instructions and to consume two HMR vitamin/mineral supplements and D&CVRRP ‘protocol’ salad daily. Almond-LCD subjects were given weekly allotments of prepackaged whole unblanched unsalted almonds. The amount of almonds was selected based on published data reporting favorable changes in blood lipids incorporating this amount of almonds under isocaloric conditions. Almond-LCD subjects were advised to consume the almonds at the time of day most convenient to their lifestyle. CHO-LCD subjects were given explicit instructions on how to self-select a combination of complex carbohydrates daily from a food list that were equivalent in calories to 84 g almonds. The food list featured a wide range of glycemic index complex carbohydrate-containing foods (peas, corn, potato, pasta, rice, etc). Additionally, CHO-LCD subjects were instructed to consume two teaspoons of safflower oil daily to meet essential fatty acid requirements. Energy and nutrient composition of the diets were computed using Food Processor (Version 7.11, ESHA Research, Salem, OR, USA).
Subjects completed detailed daily food and exercise records in their weekly D & CVRRP logbooks. The Program Dietitian performed weekly face-to-face reviews of the records and suggestions to enhance compliance were provided. Subjects were advised to refrain from exercise during the first 4 weeks of the D&CVRRP to allow time for adaptation to the LCD, and then subsequently were encouraged to walk for 20–30 min three to five times per week. Excessive deviations were operationally defined as failure to stay within 25% of the weekly prescribed calories and minutes of activity, which prompted additional individualized sessions with the Program Psychologist and/or Program Dietitian. Both groups had equivalent levels of noncompliance necessitating the aforementioned sessions during the 24-week intervention.
Clinical assessments were made at the 24 consecutive weekly visits during normal D & CVRRP outpatient clinic operations. All subjects attended weekly clinic visits with an endocrinologist, followed by nutrition and behavior modification classes. Weight, blood pressure (BP) and heart rate were measured weekly using calibrated office instruments, and ketone levels were taken using a Hemocue™ monitor. Body weight was determined to the nearest 0.1 kg. Plasma lipids, insulin and glucose were measured at baseline (week 0) and at weeks 8, 16 and 24. Additionally, subjects were asked to subjectively evaluate and record the acceptability of their prestudy diet and their assigned study intervention at weeks 8, 16 and 24, in terms of satiety, palatability and texture using a 0–10-point semantic scale (0, not satisfied at all; 5, neutral; and 10, very satisfied).
Venous blood was collected after an overnight fast at the General Clinical Research Center (GCRC) unit at COH. The University of Southern California Core Lipid Laboratory performed the lipid panel and Lipoprotein Quantification (LPQ) analysis. Plasma was analyzed using the Centers for Disease Control certified Lipid Research Clinics Protocol36 for TC, TG and HDL-C after dextran sulfate–magnesium chloride preparation. LPQ was performed for direct evaluation of HDL-C and LDL-C using a COBAS MIRAS analyzer (Roche). Insulin concentrations were determined by human specific radioimmunoassay (Linco St. Charles, Missouri) methodology and glucose concentrations were measured by a Stat Glucose/Lactate Analyzer Model 2000 (YSI, Yellow Springs, OH, USA).
Additional anthropometric assessments were performed at baseline and at week 24, including bioelectrical impedance analysis using the Tanita® TBF-300 body composition analyzer/scale and waist measurements. The Tanita® analyzer/scale utilizes ‘foot-to-foot’ pressure contact electrode technology to determine internal body composition,37 which others have validated in the obese population.38 A proprietary algorithm derives the percent body fat by combining impedance and weight measurements with height, gender, age and physical activity level. BMI was calculated as weight(kg)/height(m2). WC positively correlates with metabolic syndrome-related atherogenic lipid abnormalities among obese individuals.39 WC measurements were made to the nearest 0.1 cm, midway between the last rib and the ileac crest, to estimate the change in visceral fat in the abdominal region.
Homeostasis model analysis40 (HOMA) was utilized to estimate the change in insulin resistance (HOMA-IR) from fasting glucose and insulin concentrations. HOMA-IR was calculated using the formula [insulin(pM)·glucose(mM)]/22.5.40
Sample size and power calculations were performed utilizing UnifyPower Macro for SAS. Data were entered into a JMP database and statistical analysis was performed using SAS software (JMP Version 4.0.5, and, SAS Version 8.2; SAS Institute, Cary, NC, USA). Two-sided unpaired t-tests were performed on all subject baseline characteristics using a probability value of 0.05. In light of the longitudinal structure of the data, a mixed model with an autoregressive covariance structure of lag 1 was used to test all hypotheses, including those involving multiple covariates. All percent change values presented are least-squares means estimated from mixed models. The effect of each treatment on each end point over time was plotted as the least-squares mean and 95% CI by treatment group over time for each end point. Models were adjusted for baseline measurements, and all time points (0, 8, 16, and 24 weeks) were included in the analysis, with the exception of weight and BMI which included data from all 24 weeks. Both an intent-to-treat and as-treated analysis were performed and produced similar findings. Thus, the intent-to-treat data, which is the least biased analysis, are presented within this paper. The assumption used in the intent-to-treat model with regard to unmeasured end points for the dropouts was that they were missing at random.
Results are presented based on percent change in least-squares means (Table 3). The study sample size did not allow for evaluation of the ethnic diversity of this cohort.
Anthropometrics and body composition
Figure 2 shows the weekly change in body weight during the 24-week intervention. Almond consumption was associated with greater reduction in weight/BMI (−18 vs −11%, P<0.0001), WC (−14 vs −9%, P<0.05), fat mass (FM) (−30 vs −20%, P<0.05) and total body water (TBW) (−8 vs 1%, P<0.05). Neither gender nor chronic disease (type 2 diabetes mellitus, hypertension) were independently responsible for the weight difference found between the two groups. A divergence in weight loss between groups occurred at week 16 of the intervention, at which time the CHO-LCD group experienced a ‘plateau’. In the CHO-LCD group, 92% of the total weight loss was seen in the first 16 weeks of the intervention in contrast to the almond-LCD group which experienced 77% of the total weight loss by week 16. Both interventions were associated with a decline in fat-free mass (FFM) over the study period (P<0.0001); however, no difference was found between the study groups.
Both study interventions were associated with similar declines in mean fasting insulin and glucose concentrations (insulin: P<0.0001; and, glucose P<0.001) and were associated with decreases in HOMA-IR (P<0.0001). Diastolic BP decreased similarly by 8% in both interventions. Almond consumption was associated with a decrease in systolic BP as compared to no observed change within the CHO-LCD group (−11 vs 0%, P<0.01). Ketone levels increased significantly in the almond-LCD intervention compared to the CHO-LCD intervention (+260 vs 0%, P<0.02). TC and TG levels decreased over time in both groups (P<0.001 and 0.0001, respectively), but did not show significant variations between the two interventions. LDL-C decreased significantly in the almond-LCD and CHO-LCD interventions over time (−15 vs −10%, P<0.0001, respectively). Although HDL-C increased only in the CHO-LCD group compared to the almond-LCD group (+15 vs −6%, P<0.05), the LDL-C to HDL-C ratio decreased (P<0.01) equivalently in both groups (P=NS).
Additional testing on hematological and biochemical laboratory parameters was performed at weeks 0, 12 and 24 according to D & CVRRP guidelines. No clinically significant changes in electrocardiogram, complete blood count, basic metabolic panel, uric acid, liver enzymes, total protein, albumin, calcium or total bilirubin occurred during the study.
Self-reported satiety and acceptability
Subjects did not differ in their self-reported evaluation of the acceptability of their assigned dietary intervention in terms of satiety, palatability and texture at weeks 0, 8, 16 and 24. There were no significant differences between or within the groups over time. Hence, both almonds and complex carbohydrates were reported to be as satiating and satisfying as the subject's respective baseline diets.
The present study was designed to evaluate two distinct macronutrient approaches in the context of a formula-based LCD and their subsequent impact on anthropometric, body composition and metabolic parameters during a 24-week weight reduction program in overweight and obese adults. Both nutritional interventions were effective in decreasing body weight beyond the weight loss observed during long-term interventions with sibutramine hydrochloride.41 The CHO-LCD group reached a plateau at week 16, similar to the onset of plateaus shown while using intermittent and continuous sibutramine therapy.41 However, the almond-LCD group experienced a sustained and greater weight reduction for the duration of the 24-week intervention. A 62% greater reduction in weight/BMI, 50% greater reduction in WC, and 56% greater reduction in FM were observed in the almond-LCD as compared to the CHO-LCD intervention. The decreases in FFM between the two interventions were similar, whereas TBW decreased significantly more in the almond-LCD group.
The difference in weight loss was unexpected, given the study design featuring a matched prescribed total calorie intake and equivalent levels of self-reported physical activity between the groups. Neither gender, presence of type 2 diabetes mellitus nor hypertension influenced the magnitude of the difference in weight change. McManus et al42 have shown that higher fat diets may be more satiating than lower fat diets containing high glycemic index foods. Others have noted that unanticipated weight loss occurs in controlled feeding trials featuring the inclusion of nuts under isocaloric conditions.43,44 The fiber matrix of the nut may have compromised the absorption of the fat from the almonds yielding an imbalance of energy sources between the groups. These data suggest that the greater weight loss observed in almond consumers may have been secondary to greater satiety and the lower bioavailability of calories from nuts.
A greater improvement of fasting glucose and insulin using a hypocaloric low carbohydrate diet enhanced with MUFA has been reported.45 Our MUFA-enriched almond intervention showed an overall 54% reduction in fasting insulin, as compared to a 32% reduction in the carbohydrate intervention, in the context of an unequal magnitude of weight loss. Comparison of a subject's fasting glucose and insulin concentrations within the HOMA-IR prediction model provides a quantitative assessment of the contribution of insulin resistance to the fasting metabolic state.40 A 31% difference in the magnitude of HOMA-IR reduction occurred between the two study groups (almond-LCD: −66% and CHO-LCD: −35%). Almond consumption was associated with improved insulin sensitivity, which might have produced a compensatory reduced load on the pancreas. Also, the high oleic acid content in the almonds may have contributed to improved β-cell efficiency through enhanced intestinal secretion of glucagon-like-peptide-1 (GLP-1).46 Enhancing the β-cell secretory response by dietary measures will improve the regulation of postprandial glucose disposal and insulin sensitivity,47 which has the potential for improving the cluster of abnormalities linked to the metabolic syndrome.48 Further, among subjects with type 2 diabetes, diabetes medication reductions were sustained or further reduced in a greater proportion of almond-LCD as compared to CHO-LCD subjects (96 vs 50%, respectively).
Hypertensive patients in both groups experienced reductions in their antihypertensive medications during the 24-week study. However, there were more patients in the CHO-LCD group (62%) than in the almond-LCD group (50%) with documented hypertension at baseline. Thus, it is possible that the absence of a reduction in systolic BP within the CHO-LCD group may have been due to better BP control at baseline secondary to antihypertensive therapy.
The change in the hormonal milieu that occurs in the presence of high-fat diets may have favorably improved the utilization of adipose fat stores as an energy source in the almond intervention. The higher ketone levels in the almond-LCD group were anticipated due to the higher proportion of energy from total fat (almond-LCD: total fat=39% of energy, CHO-LCD: total fat=18% of energy). The increased production of ketones within the almond-LCD group may reflect a higher rate of fat breakdown that exceeded the level at which the tissues could oxidize the fat intermediates for energy production and some calories may have been lost through ketonuria. Also, a small proportion of the difference in the magnitude of weight loss in the almond-LCD group may have been due to ketosis-induced diuresis in the context of the significantly greater loss of FM.
Similar decreases in TC levels were observed in the context of an unequal magnitude of weight loss between the two interventions. TG levels decreased in a sustainable manner in both groups over the 24-week intervention in contrast to others who have noted a transient increase in TG levels in subjects treated with hypocaloric diets.49 The almond-LCD group experienced a reduction in LDL-C from baseline by 15% as compared to a 10% reduction in the CHO-LCD group. Under isocaloric conditions, other investigators have found similar dose responses using almonds to reduce LDL-C, for example, 1% for every 7,27,30, 831,50 and 10 g/day.51
HDL-C concentrations have been shown to vary dependent on whether subjects are studied during an acute weight-loss phase as compared to a reduced and stable weight.52 During acute weight reduction tissue concentrations of lipoprotein lipase decrease by 50–80%49 resulting in reduced TG-rich lipoprotein synthesis, impaired very-LDL-C catabolism, and diminished transfer of lipids to HDL-C, and, therefore reduced HDL-C concentrations. Although HDL-C did not increase in the almond-LCD group, the LDL-C to HDL-C ratio, an important predictor of cardiovascular risk,53 decreased equivalently in both groups. Jenkins et al27 found a 12% reduction in the LDL-C to HDL-C ratio using 73 g/day almonds under isocaloric conditions in hypercholesterolemic subjects (BMI: 20–32 kg/m2), whereas our study found a 10% reduction using 84 g/day under hypocaloric conditions in overweight and obese subjects (BMI: 27–55 kg/m2). These studies complement each other as they have produced diminished cardiovascular risk using liberal amounts of total fat (specifically MUFA) among different high-risk patient populations, and have produced findings that are consistent with epidemiologic data that has found nut consumption to be associated with reduced CHD risk.54,55
Zambon et al14 found that LDL-C decreased significantly in a group of obese normolipemic premenopausal women consuming hypocaloric diets enriched in olive oil (MUFA) and complex carbohydrates. However, with an equal magnitude of weight loss (−11% in both groups), HDL-C increased significantly in the high MUFA group and decreased in the complex carbohydrate group.14 Our study is unique from others in that it addressed the effect of almonds as a source of MUFA (specifically oleic acid) compared to complex carbohydrate in the context of a formula-based LCD, and featured equivalent levels of calories, protein, saturated fat and PUFA in the two interventions. In light of the wide range of self-selected glycemic index foods, we are unable to explain the rise in HDL-C in the CHO-LCD group. The difference in HDL-C between groups may be partially attributed to the difference in daily fiber intake (almond-LCD=20 g, CHO-LCD=32 g).
The present study demonstrates that the use of almonds in the context of a formula-based LCD is a feasible option for consideration and has a potential role in the public health implications of obesity. A primary complaint among overweight and obese subjects participating in formula-based medically supervised weight reduction programs is the lack of texture variability and satiety. If clinicians are capable of sustaining motivation and compliance using nuts, a successful formula-based LCD could contain larger proportions of MUFA. However, careful patient monitoring is required to ensure that there is improvement in metabolic parameters in parallel to weight reduction. Our findings call for additional studies in larger numbers of subjects to allow evaluation of different fat to CHO ratios and whether consumption of almonds (or other nuts) as the delivery vehicle were responsible for the disparity in weight loss. The use of almonds to ameliorate the recidivism commonly observed after active weight reduction is also worthy of exploration.
We thank Zaida Cordero-MacIntyre, PhD for her assistance in study development, Joanne Shifflett, RN for clinical services, Frieda Brandt, RN for administrative support, Stephen Scott for laboratory technical support, and, Paul Frankel, PhD and David Smith, PhD for statistical support and data analysis. This research was supported in part by a General Clinical Research Center Grant NIH (M01RR00043) awarded to the City of Hope National Medical Center, a satellite center of the University of Southern California, and, the Almond Board of California.