In the LIFE study under conditions of weight maintenance, the legume-rich, high fermentable fiber, low GI diet (LG) led to greater reductions in fasting serum TC and LDL-C concentrations. It is possible that the higher fiber and reduced cholesterol consumption acted to decrease fat absorption and lower hepatic synthesis of cholesterol contributing to lower circulating lipids. Another possibility is that the reduced GI and GL of the legume diet played a role in favorably altering lipid concentrations. Low GI and GL diets favor insulin sensitivity. Insulin inhibits the mobilization of free fatty acids from adipose tissue, thus lowering hepatic production of very low density lipoprotein (VLDL) and maintaining low levels of TC and LDL-C. Possibly, these changes accounted for greater reductions in TC and LDL-C concentrations on the LG diet.
Many but not all [23
] studies have shown hypo-cholesterolemic effects of dietary fiber [11
]; however, studies specifically on legume consumption are still limited. Anderson et al. [4
] reported that incorporating 100 g of dried beans into a Western diet for 28 days decreased TC (18.7%) and LDL-C (23.1%) in men (n
= 6). In a feeding study of 24 hyperlipidemic men who consumed 120 or 162 g beans with tomato sauce for 21 days [28
], both serum TC and LDL-C were significantly decreased (10.4 and 8.4%, respectively). Pittaway et al. [29
] reported that an addition of 104 g of chickpeas into ad libitum diet for 12 weeks led to improvements in both TC (7.7 mg/dL) and LDL-C (7.3 mg/dL) among 13 pre-and 19 postmenopausal women and 13 men at high risk for CVD. Collectively, these studies demonstrate a hypocholesterolemic effect of legume consumption; yet, they were not designed as controlled feeding experiments. Among these clinical studies, weight changes either were not mentioned [14
] or weight slightly decreased [26
]. In our study, subject body weights were measured daily and their calorie intake adjusted for weight maintenance; therefore, changes in lipids were independent of weight loss.
The present study showed that both the LG diet and the HA diet significantly lowered fasting TAG concentrations. Our observations supported several [28
] but not all [13
] earlier feeding studies in terms of TAG-lowering influence. We observed that the HA diet lowered fasting TAG similarly to the LG diet. We compared the total and added sugars of the two experimental diets with that of the subjects’ own diet; however, we did not find significant differences. It is possible that the TAG-lowering effects might not result from fiber consumption or GI. The LG diet provided increased plant protein in the diet. To match protein intake and maintain a lower cholesterol intake, we added more chicken, milk, and fish to the HA diet. Differences among both test diets compared to pre-study diets (e.g. protein consumption) may have contributed to the reductions in TAG observed with both diets [32
Fruit and vegetables are good sources of many vitamins, minerals and bioactive compounds, which have been shown to possess cardioprotective effects. Antioxidants in fruit and vegetables such as vitamin C, carotenoids, vitamin E, and flavonoids may reduce CVD by reducing lipid oxidation and inflammation in the artery wall. The B-complex vitamins such as folate and B6
may reduce CVD by lowering circulation homocysteine. A number of epidemiological studies demonstrate the inverse association between fruit and vegetable consumption and CVD risk [33
]. In the National Heart, Lung, and Blood Institute Family Heart Study of 4,466 men and women aged 25 and older followed from 1993 to 1995 [38
], persons in the higher range of fruit and vegetable consumption (mean of 5.4 servings/day for men and 5.5 servings/day for women) had lower fasting LDL-C concentrations (P
for trend <0.0001 for each), compared to those in the lower range (mean of 1.4 servings for both). In the present study, we observed that both the LG and HA diets provided higher amounts of fruit and vegetables compared with subjects’ usual diets, although the differences were only significant between the LG diet and subjects’ usual diets.
Several cross-sectional studies have reported that high dietary GI was inversely associated with HDL-C [39
] and positively associated with LDL-C [43
] and TAG concentrations [40
]. High dietary GL was inversely associated with HDL-C [39
] and was positively associated with TAG concentrations [40
]. Ma et al. [41
] reported an inverse associations between GL, TC and LDL-C concentrations; however, Du et al. [42
] reported non-significant associations between GL and lipid and lipoprotein concentrations. Randomized intervention trials showed that low GL diets [45
] favorably improved HDL-C and/or TAG concentrations. In the present study, we observed that a high legume low GI/GL diet significantly reduced fasting serum TC and LDL-C concentrations; the results are consistent with the previous investigations. We also observed that fasting HDL-C concentrations decreased after the 4-week low GI/GL dietary intervention, which is not consistent with the aforementioned studies. The difference may be due to different study designs and target populations. In addition, unlike these studies, the present study controlled subjects’ body weight and overall food intake.
The LIFE study was designed to evaluate the effects of fiber consumption on lipid profiles; however, other components such as phytosterols may also consider a mechanism for lipid reduction. Stanols and sterols, are primarily present in nuts, vegetable oils, seeds, cereals, and legumes [47
]. Plant stanols and sterols have similar chemical structures to cholesterol; they may reduce cholesterol absorption by enterocytes [48
] and the esterification rate of cholesterol in the enterocytes [49
]. Decreased cholesterol absorption stimulates cholesterol synthesis [50
] as well as the expression of LDL receptor mRNA [51
] increasing LDL clearance and lowering LDL production resulting in lower circulating total cholesterol levels.
Metabolic syndrome causes dyslipidemia partially as the result of insulin resistance [52
]. Our results demonstrate that insulin resistance may blunt response to a legume-enriched, high fiber diet. Among IS subjects, the LG diet led to significant reductions of all lipid profiles as well as the three ratios; however, among IR subjects, the same diet only had strong effects on the reductions of TC, LDL-C, and HDL-C. The HA diet led to significant reductions in TC and HDL-C among IS subjects; the same diet also had strong effects on TC, TAG, and TAG/HDL-C among IR subjects. Lefevre et al. [53
] reported that subjects who had higher BMIs and waist circumferences, greater percentages of body fat and higher fasting insulin concentrations had smaller reductions in TC, LDL-C, TC/HDL-C after a Step II (low fat, low saturated fat) diet. Our results are consistent with this study. We observed that subjects who had higher HOMA-IR values at study entry had smaller reductions in TAG, TC/HDL-C, LDL-C/HDL-C, and TAG/HDL-C after the LG diet, though the reduction was only statistically significant for TC/HDL-C. Different from Lefevre’s report, we found that subjects’ BMIs at study entry did not predict changes in any lipid profiles or HDL-C related ratios.
Many human clinical studies have shown that a reduced intake of saturated fat may lower the risk for cardiovascular disease by decreasing TC and LDL-C concentrations [22
]. The effects of these types of diets on lowering TAG levels have been variable [22
]. In the present study, we observed that moderate total (34%) and saturated fat (12%) with high fermentable fiber consumption also lowered TC and LDL-C concentrations, whereas the HA diet only lowered TC. The differences observed between test diets was significant for TC and LDL-C demonstrating an effect of total and soluble fiber, and possibly GL. Changes of this magnitude suggest that high fermentable fiber consumption may be another approach to lower risk for CVD [55
There are several strengths of our study. First, we implemented a randomized, cross-over controlled feeding study design. In addition, we matched the two test diets so that they were isocaloric and provided similar percentages of energy from total fat, saturated fat, carbohydrate, and protein under conditions of weight maintenance. Second, though fiber consumption was higher than daily recommendations, the LG diet was well-tolerated.
There are some limitations in the present study. The present study defined insulin resistance by the HOMA-IR index, which has been widely used for clinical evaluation of insulin resistance. The cutoff value used in our study and others is somewhat subjective; however, the decision was based on two recent human clinical trials [20
]. Additionally, studies show that the classification of insulin resistance according to the HOMA-IR index is highly correlated to the ones based on the euglycemic insulin clamp technique, the gold standard for evaluation of insulin sensitivity/resistance [56
]. The LG diet included less dietary cholesterol than the HA diet; therefore, we are not able to rule out the possibility that dietary cholesterol consumption contributed to the observed effects on serum cholesterol levels. However, our analysis showed that the difference in dietary cholesterol between the LG and HA diet was not statistically significant. In addition, a 20 mg/day increment in dietary cholesterol intake only results in a very small change of serum cholesterol according to a meta-analysis of 27 studies [58
]. Next, we only measured the changes in selected lipid/lipoprotein biomarkers. Recent studies indicate that small dense LDL-C particle (sd-LDL) level is a better predictor for CHD [59
] than LDL-C. Apolipoprotein B (apoB) is the primary apolipo-protein in LDL-C; measurement of apo B may provide additional information on changes in LDL particle size. Similarly, apo A-I is the primary apolipoprotein in HDL-C; measuring the change in apo A-I may help to track the change in HDL particle size.
In conclusion, this study adds to the growing evidence that incorporating legumes in the moderate-fat diet improves lipid profiles, thus potentially lowering CVD risk. However, a cautionary note must be added since insulin resistance is highly prevalent and increasing, and individuals in this study with insulin resistance responded less favorably to the legume enriched, high fermentable fiber, low glycemic index diet than those without insulin resistance.