Biochemical, Anthropometric, and Physiological Responses to Carbohydrate-Restricted Diets Versus a Low-Fat Diet in Obese Adults: A Randomized Crossover Trial

Abstract

Some research supports high-fat carbohydrate (CHO)-restricted diets for weight and fat loss and improvement of cardiovascular disease risk factors. To test this, a randomized crossover study was designed. Subjects (17 obese men and women [BMI: 30–38 kg/m²]) were fed three diets (supplying 1600 and 2200 kilocalories (kcal)/day for women and men, respectively) for 4 weeks, with each trial separated by 4-week washout periods. One CHO-restricted diet (10% CHO, 50% fat, and 40% protein content) was rich in plant foods and mushrooms, while the other CHO-restricted diet included more animal foods (10% CHO, 60% fat, and 30% protein content). The third diet was lower in fat and protein content (LF) and higher in CHOs (61% CHO, 21% fat, and 18% protein content). Body composition was assessed through hydrostatic weighing before and after each diet trial. Fasting blood samples were collected weekly for analysis of hormones and lipids. Data were analyzed through repeated measures analysis of variance with post hoc paired comparison t-tests. Weight and fat loss were similar (P > .05) among trials. Subjects lost lean mass (P < .05) during CHO-restricted trials, but not in the LF trial. Insulin concentrations decreased (P < .05) during the CHO-restricted trial and tended (P = .05) to decrease during the LF trial. Total cholesterol decreased (P < .05) for all trials; however, high-density lipoprotein cholesterol decreased (P < .05) and triacylglycerols were higher (P < .05) following the LF trial. Taken together, energy restriction regardless of diet composition promoted similar weight loss; however, CHO-restricted diets based on either plants/mushrooms or animal foods elicited a more beneficial lipid-altering effect in comparison with the LF diet.

Introduction

In recent years, renewed interest in carbohydrate (CHO)-restricted diet regimens for treatment and prevention of obesity and cardiovascular disease has emerged in the media, with the public and researchers. Studies have suggested that fat-rich diets may promote improvements in weight loss and body composition in comparison with higher CHO diets within as little as 3 weeks.¹ When subjects were asked to consume similar diets for longer periods (6 months), these diets appeared to continue to favorably affect body weight^2–4 and fat³ loss; however, after 1 year, CHO-restricted diets may be no more effective than low-fat diets.^2–5

CHO-restricted diets have been criticized due to lack of scientific information regarding their safety, especially for cardiovascular risk factors.^6,7 Coronary heart disease (CHD), the leading cause of death in the United States,⁸ is positively correlated with elevated concentrations of triacylglycerols (TGs), total cholesterol (TC), and low-density lipoprotein cholesterol (LDL-C), along with reduced concentrations of high-density lipoprotein cholesterol (HDL-C).^9–13 Multiple studies have demonstrated altered lipid profiles during weight loss for CHO-restricted diets compared with higher CHO lower-fat diets, including decreases in TGs^{1,2,4,5,14–18} and increases in both HDL-C^2,5,16–19 and LDL-C.^2,16,19 These studies have caused researchers to question whether these types of diets increase or reduce risk for CHD.

Current dietary guidelines recommend that 45%–65% of total energy intake is derived from CHOs, 20%–35% from fats, and 0.8 g/kg body weight of proteins.²⁰ However, some reports have demonstrated that higher CHO low-fat diets increase serum TG concentrations,^21–28 causing similar uncertainty regarding their contribution to CHD risk. Due to the controversy, alternative diet approaches have been considered. These include CHO-restricted diets that replace saturated fat with unsaturated fat or replace animal proteins with plant proteins. Only a fraction of the previous studies have evaluated the potential role of diet quality of fat-rich CHO-restricted diets, and few studies have compared two different CHO-restricted diets. The fat source in diets should be considered since research has previously demonstrated that foods rich in saturated fatty acids (SFAs) are more likely to raise blood lipids than those rich in unsaturated fatty acids²⁹ and animal proteins are more likely to negatively impact blood lipids than plant proteins.³⁰ Replacement of some animal foods with plant foods in CHO-restricted diets may be an effective alternative for improving the blood lipid profile.

Some research supports the use of CHO-restricted diets for short-term weight loss and improvement in blood lipid profile, yet few of the previous studies have controlled intake by providing all foods to be eaten and few have used diets that are considered very low in carbohydrate (VLC) composition (<10% of energy). Therefore, the purpose of the present study was to examine the effect of altering the macronutrient composition of the diet, types of dietary fats, and the potential role of plants and mushrooms incorporated into a 4-week VLC diet on body weight and composition, serum glucose, insulin, leptin, TGs, TC, LDL-C, and HDL-C. This was accomplished by comparing a mushroom- and plant-based VLC diet that is low in saturated fat (VLCPM) with (1) a more typical VLC diet (higher saturated fat) using primarily animal sources of protein and fat (VLCA) and (2) a diet high in CHOs and lower in fat and protein content (LF), representing usual dietary guidelines.

Materials and Methods

Subjects

Seventeen obese subjects (14 women and three men; mean [±SD] age of 43.2 ± 13.3 years; and mean [±SD] body–mass index [BMI, calculated as kg/m²] of 33.2 ± 29 kg/m²) participated in the crossover study. The criteria for inclusion were as follows: BMI between 30 and 38 kg/m² and maintenance of current weight (<10% weight shift for the past 6 months). Exclusionary criteria included smoking, pregnancy, the presence of any known medical problem resulting in metabolic disorders, use of medications known to affect lipid metabolism, and participation in >2 h of structured exercise per week. Written informed consent was obtained from all subjects, and all procedures were approved by the San Diego State University Institutional Review Board.

Experimental design

Dietary interventions

All participants received each of three study diets in a randomized, balanced crossover design for 4 weeks, each separated by 4-week washout periods between each diet regimen. Block randomization was used to randomly allocate subjects to the order of the diets. Participants were asked to return to normal eating patterns during washout periods. Each diet was intended to be equal in total energy content within each sex at appropriate levels (1600 kilocalories [kcal] for women; 2200 kcal for men) to promote a weight loss of 0.5–2.0 pounds per week. Dietary compositions were determined with Computrition (West Hills, CA, USA) nutrition software and supplemented with food labels of packed foods as needed.

Two diets were VLC, while the third was higher in CHOs and LF. One of the low-carbohydrate diets (VLCPM) was plant and mushroom based, including a daily intake of 170 g of white button mushrooms (Agaricus bisporus). This diet provided 10% of energy from CHOs, 50% from fat, and 40% from protein. Mushrooms were incorporated into typical foods (e.g., omelets, frittatas, and pasta) and as a replacement for some meat. The other low-carbohydrate diet (VLCA) was devoid of mushrooms and included more animal products, which yielded a higher saturated fat content than the VLCPM. This diet provided 10% of energy from CHOs, 60% from fat, and 30% from protein. The third diet was a low-fat diet, which was rich in CHOs (61% of energy) and lower in fat (21% of energy) and protein (18% of energy).

A 5-day rotating meal pattern was utilized for each of the dietary trials. All meals were prepared and packaged in the food preparation laboratory at San Diego State University following sanitary food-handling procedures as described by the California Department of Health. Foods were distributed and/or delivered to participants at 7-day intervals. Participants were furnished with daily printed menu checklists that included each food item to be consumed to monitor adherence to the diets. Participants provided a check next to each item that was consumed. In the event that an item or portion of an item was not eaten, the participant was instructed to record the amount not consumed. Participants were asked to adhere strictly to the diet; however, they were asked to record details of all additional food eaten. Adherence to the diet was enhanced through frequent contact by the investigators. Contact occurred by e-mail and/or telephone one to two times per week, highlighting the importance of the research. Furthermore, all participants visited the laboratory weekly, which provided opportunities for personal contact with investigators.

Measurements

Anthropometrics

Height was determined at the beginning of the study using a wall-mounted stadiometer. Weight was recorded at baseline and then once every week throughout the study with subjects in the exact clothing as the previous visit and without shoes. Body composition was assessed at the beginning and end of each 4-week diet period by hydrodensitometry. Body water content was tested using bioelectrical impedance with a Bodystat 1500 (Detroit, MI, USA).

Biochemical assays

Blood samples were collected following an overnight fast (12 h) at baseline and every week during the trial periods. Subjects were asked to refrain from caffeine and alcohol during the 24 h before testing, and exercise was restricted during the 32 h before each blood collection. Blood was drawn into 13-mL vials containing the gel separator and clot activator. Blood was allowed to clot and then centrifuged at 1200 g for 15 min at 2°C–8°C. Serum was stored at −80°C for batch analysis. Serum concentrations of glucose, TC, TGs, HDL-C, and β-hydroxybutyrate were assessed at baseline and week 4 using enzymatic colorimetric kits from Stanbio Labs, Inc., (Boerne, TX, USA). LDL-C was calculated using the Friedewald equation.³¹ Serum concentrations of insulin and leptin were assessed at baseline and after 4 weeks using radioimmunoassay kits obtained from Linco Research, Inc., (St. Charles, MO, USA). Leptin was measured only in a subset of 13 randomly selected subjects. Insulin sensitivity was calculated using the QUICKI method, as previously described.³²

Dietary analysis

All food records were analyzed by computer using My Diet Analysis 2.0 powered by ESHA Research Nutrition Software and Database.

Statistical analyses

Data were analyzed using the Statistical Program for Social Sciences (SPSS, version 11.5.0, 2002, SPSS Inc., Lead Technologies, Inc., Chicago, IL, USA) software package. Baseline characteristics were assessed for potential differences using paired samples t-tests to ensure similarities at the beginning of each dietary trial. Possible diet-induced differences between the trials from baseline to the fourth week were assessed using a 3 (diets) × 5 (time points) repeated measures ANOVA design for body weight and a 3 (diets) × 2 (time points) repeated measures ANOVA for all other variables. Mauchly's test was used to examine the sphericity for each analysis. If the assumption of sphericity was violated, the Greenhouse–Geisser adjustment was utilized to protect against type I errors. Significant effects were followed up with paired comparison t-tests. Significance was set at an alpha level of P < .05. Data are expressed as mean ± SD.

Results

Energy and nutrient intake

Table 1 provides details of nutrient intake during the three experimental trials. Female subjects tended to consume less energy during the VLCPM trial than the LF (P < .05) and VLCA trials (P = .07). Energy consumed (for females) was also lower on the VLCA diet compared with the LF diet (P < .05). Subjects (male and female) consumed more (P < .05) CHOs and sugar and less fat during the LF diet trial compared with the low-CHO trials, as intended. During the VLCPM trial, subjects (male and female) consumed more (P < .05) fiber than during the VLCA and LF trials. For all subjects (male and female), saturated fat intake was greater (P < .05) during the VLCA trial compared with the VLCPM trial, as intended. Subjects (male and female) consumed more (P < .05) saturated fat during both low-CHO trials than the LF diet.

Table 1.

Energy and Nutrients as Provided and as Consumed by Participants During Each Trial

	Dietary trial
	LF		VLCA		VLCPM
Variable	Provided	Consumed	Provided	Consumed	Provided	Consumed
Energy (kcal)
Women	1600	1555 ± 76	1600	1410 ± 98*	1600	1330 ± 103*
Men	2200	1880 ± 225	2200	1833 ± 174	2200	1865 ± 141
Fat (g)
Women	38	51 ± 23	91	92 ± 22	64	98 ± 24*#
Men	54	61 ± 13	128	130 ± 20	87	169 ± 51*#
Saturated fat (g)
Women	15	15 ± 1	41	38 ± 8^	19	17 ± 2
Men	19	16 ± 2	61	50 ± 5^	27	24 ± 4
Sodium (mg)
Women	2750	2680 ± 150	3641	3140 ± 270	4101	3580 ± 250
Men	3760	3280 ± 51	4953	4070 ± 560	5335	4690 ± 900
Total CHO (g)
Women	221	212 ± 14	77	79 ± 46	71	64 ± 7*#
Men	292	245 ± 31	83	76 ± 9	97	89 ± 25*#
Fiber (g)
Women	19	18 ± 1	24	16 ± 4	23	19 ± 3*#
Men	25	20 ± 3	21	19 ± 6	39	31 ± 9*#
Sugar (g)
Women	95	91 ± 7	16	21 ± 24	13	12 ± 2*#
Men	118	95 ± 15	19	17 ± 3	17	16 ± 4*#
Protein (g)
Women	67	67 ± 3	110	93 ± 11	151	131 ± 10
Men	92	80 ± 11	139	116 ± 16	226	186 ± 26

Data are presented as mean ± SD. ^*P < .05 VLCA or VLCM versus LF; ^#P < .05 VLCM versus VLCA; ^{^}P < .05 VLCA versus LF or VLCM.

CHO, carbohydrate; LF, lower in fat and protein content; VLC, very low in carbohydrate; VLCA, VLC with animal sources of protein and fat; VLCPM, mushroom- and plant-based VLC.

Body weight

Subjects lost a significant amount of weight from baseline to week 4 for all dietary trials (P < .05). Weight loss during the VLCA (3.3 ± 0.6 kg) diet did not differ significantly from the LF (2.8 ± 1.1 kg) or the VLCPM (3.7 ± 1.1 kg) diet. Significantly greater weight was lost (P < .05) for subjects during the VLCPM trial in comparison with the LF trial (Fig. 1A). On average, subjects consumed 1272 more kcal while on the LF diet than during the VLCPM diet, accounting for the additional weight loss during the VLCPM diet. Total body water did not differ significantly among the three dietary trials or from baseline to week 4 (Fig. 1C). Therefore, differences in changes of body water are not a likely explanation for changes in body weight.

FIG. 1.

(A) Body weight, (B) body composition (lean and fat mass) loss, and (C) total body water from baseline through week 4. *P < .05 between VLCPM and LF. n = 17/group/time point. Data are expressed as mean ± SD. LF, lower in fat and protein content; VLC, very low in carbohydrate; VLCA, VLC with animal sources of protein and fat; VLCPM, mushroom- and plant-based VLC.

Body composition

As depicted in Figure 1B, participants lost (P < .05) fat mass during each dietary trial. Fat loss from baseline to the final measure was 2.21 ± 2.66 kg for the LF dietary trial, 1.98 ± 1.32 kg for the VLCPM dietary trial, and 1.77 ± 1.56 kg for the VLCA dietary trial. However, when expressed as a percentage, only the VLCPM diet decreased (P < .05) fatness. VLCPM and VLCA diets decreased (P < .05) lean body mass (1.79 ± 1.23 kg and 1.46 ± 1.64 kg, respectively), while no statistically significant effect of the LF diet (0.73 ± 2.69 kg) was detected.

Biochemical profiles

Concentrations of β-hydroxybutyrate increased (P < .05) during both CHO-restricted trials, but not during LF (Fig. 2). The VLCPM diet decreased (P < .05) fasting serum glucose concentrations (Fig. 3A), but similar effects were not observed during the VLCA or the LF trial. The VLCPM and VLCA diets decreased fasting serum insulin concentrations (P < .05) (Fig. 3B), and a trend (P = .05) toward decreases was detected after consuming the LF diet. Insulin sensitivity improved (P < .05) from baseline to the end of the VLCPM (0.017 ± 0.001) and VLCA dietary trials (0.013 ± 0.002) (Fig. 3C). A trend (P = .08) of increased insulin sensitivity was noted from baseline to final assessment during the LF dietary trial (0.010 ± 0.003). Fasting serum leptin concentrations were decreased (P < .05) after each dietary trial (Fig. 4). Leptin concentrations during the VLCPM (9.9 ± 3.8 ng/mL) and VLCA (12.2 ± 5.2 ng/mL) diets were decreased to greater (P < .05) degrees than during the LF trial (6.4 ± 2.3 ng/mL).

FIG. 2.

Fasting serum β-hydroxybutyrate concentrations 1 and 4 weeks following each dietary trial. ^‡P < .05 between the LF and low-carbohydrate diets; ^P < .05 versus baseline in VLCA; ^#P < .05 versus baseline in VLCPM. n = 17/group/time point. Data are expressed as mean ± SD.

FIG. 3.

Fasting (A) serum glucose, (B) serum insulin concentrations, and (C) insulin sensitivity measured at baseline and week 4 for each dietary trial. *P < .05, **P < .01. n = 17/group/time point. Data are expressed as mean ± SD.

FIG. 4.

Fasting serum leptin concentration results at baseline and at week 4 for each dietary trial. *P < .05. n = 17/group/time point. Data are expressed as mean ± SD.

Fasting serum TG concentrations were higher (P < .05) following 4 weeks of consuming the LF diet (118.2 ± 44.8 mg/dL) compared with the VLCA (93.1 ± 31.38 mg/dL) and VLCPM (92.1 ± 37.0 mg/dL) diets (Fig. 5A). During the LF trial, serum TG concentrations tended to remain stable, while tending to decrease during the VLCPM (P = .05) and VLCA (P = .08) trials.

FIG. 5.

Fasting serum (A) TG, (B) TC, (C) HDL-C, and (D) LDL-C concentrations for the three dietary trials. *P < .05 for LF versus the VLCPM and VLCA diets; **P < .05 for LF versus VLCA; ^#P < .05 versus baseline in LF diet; ^P < .05 versus baseline in VLCA; ^‡P < .05 versus baseline in VLCPM. n = 17/group/time point. Data are expressed as mean ± SD. HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TC, total cholesterol; TGs, triacylglycerols.

All three diets lowered (P < .05) serum TC concentrations from baseline to 4 weeks (Fig. 5B). No significant differences were detected between the diets at any time point, including the fourth week time point at which the values were as follows: LF = 197.7 ± 47.0 mg/dL; VLCA = 188.7 ± 33.7 mg/dL; and VLCPM = 184.9 ± 64.5 mg/dL. Serum HDL-C concentrations decreased (P < .05) after 1 week following consumption of the LF diet and remained lower than baseline throughout the feeding period (Fig. 5C). No significant changes were detected for the VLCPM and VLCA diets.

Fasting serum LDL-C concentrations increased (P < .05) after 1 week of the LF diet, then tended to decline thereafter (Fig. 5D). LDL-C concentrations were significantly (P < .05) lower following weeks 2, 3, and 4 when consuming the VLCPM diet, but were lower (P < .05) than baseline levels for the VLCA only at week 3. Of note, a significant difference was observed for the LF (127.6 ± 38.0 mg/dL) compared with the VLCPM (145.1 ± 35.98 mg/dL) diets at baseline, which may have contributed to the observed effects and confounding interpretation.

Discussion

Significant weight loss was achieved during all three dietary trials. In line with previous literature, the observed weight loss was relative to the number of kilocalories consumed.^{2,5,14,15,18,33} While previous literature demonstrates weight loss attributed to low-CHO diets as partially due to decreases in body water,^34,35 no differences in total body water were detected in the present study.

Despite instruction to consume all provided foods, subjects voluntarily limited intake of the provided food by amounts that exceeded 96 kcal/day during both of the very low-CHO diet periods. This may have been due to increased satiety on the very low-CHO diets. Stubbs et al.³⁶ suggest that protein-rich meals are more satiating than CHO-rich meals. The lower energy intake was most pronounced during the VLCPM diet, during which subjects consumed 190 fewer kcal than during the LF dietary trial. This lower energy intake may be partially explained by the satiating effects of higher protein content combined with the satiating effect of increased fiber provided in the VLCPM diet compared with the VLCA and LF diets. Previous literature has demonstrated that subjects instructed to consume low-CHO diets voluntarily decreased energy intake more than subjects on lower fat diets.^18,33

Low-CHO diet studies have yielded varied results regarding preservation of lean body mass throughout the diet period. Similar to Meckling et al.,¹⁵ results of the present study show that the LF diet preserved lean body mass to a greater extent than the two low-CHO diets, which may have, in part, been due to differences in skeletal muscle glycogen content.³⁷ These results contrast those of Layman et al.¹⁴ who observed a tendency for lean body mass loss during a high-CHO dietary trial compared with a low-CHO dietary trial. Farnsworth et al.³⁸ also suggested that women lost more lean body mass during a high-CHO dietary trial compared with a low-CHO dietary trial, while no differences were observed in men. These studies all provided a higher percentage of CHOs compared with the VLC diets that were provided in the current study. It is important to note that while all diets produced significant fat loss and that fat loss was most pronounced in the LF diet, the VLCPM diet was the only diet that demonstrated a significant decrease in percent body fat.

The two CHO-restricted dietary trials produced the greatest reductions in fasting serum insulin concentrations and the greatest improvements in insulin sensitivity. The VLCPM trial also significantly decreased fasting blood glucose concentrations. These results are similar to other studies that detected greater reductions in fasting insulin and improved insulin sensitivity with low-CHO diets compared with higher CHO diets.^4,15,38 Volek et al.³⁹ observed that decreased insulin levels were related to fat loss. The improvements in insulin sensitivity may suggest that over time all three diets could have the potential to reduce risk for development of type 2 diabetes mellitus, particularly low-CHO diets, but follow-up trials should aim to confirm this. Additionally, higher blood insulin concentrations are associated with reduced satiety,⁴⁰ increased food intake, and increased body weight,⁴¹ so a decreased fasting insulin concentration may be a benefit of decreasing CHO intake during weight loss.

Leptin concentrations were significantly decreased from baseline to week 4 for all dietary trials, but leptin concentrations were significantly lower following the low-CHO diets versus the low-fat diet. These results are similar to previous low-CHO studies performed by Boden et al.⁴² and Koutsari et al.⁴³ who also demonstrated that low-CHO diets decrease leptin concentration to a greater extent than high-CHO diets.^33,43,44 Positive correlations have also been noted between the CHO content of a diet and blood leptin concentrations.⁴³ Lower fasting concentrations of leptin may be beneficial as they have been linked to higher resting energy expenditure.^44,45 Decreased leptin has also been theorized to increase hunger; however, a study by Mars et al.⁴⁶ demonstrated that subjects do not compensate for decreased leptin levels by increasing food intake.

Four weeks of consuming the energy-restricted VLC diets produced generally favorable changes in the fasting serum lipid profiles of the obese men and women in the current study. Although weight loss also occurred in subjects while consuming the low-fat diet, this diet produced fewer positive responses in serum lipid profiles. Many other reports in the literature indicate that lipid profiles are improved during diet-induced weight loss.^{1–5,14,15,17,18,33,36,46–49} Although some improvements were observed for each of the weight loss diets, several different responses that were detected among the diets suggest that diet composition is a potentially important variable to consider even during weight loss.

Fasting serum TG concentrations were lower following consumption of both VLC diets compared with the LF diet, which may have been partially due to the increase in TGs, following week 1, observed during LF consumption. It is possible that the higher CHO content of the LF diet induced the initial rise in TGs. This is consistent with other studies that have shown a decrease in fasting serum TG concentrations with the consumption of CHO-restricted diets when compared with higher CHO lower fat diets.^{1–5,14,15,17,18} Research has shown that during weight maintenance, increasing CHOs while lowering the amount of fat in a diet can increase serum TG concentrations.^{21–23,25–28} Furthermore, weight loss without major changes in diet composition has been demonstrated to reduce serum TG concentrations.⁵⁰

The decreases in serum TC detected at the end of 4 weeks of the dietary trials were 6.2%, 13.0%, and 16.7% for the LF, VLCPM, and VLCA diets, respectively. Research has been inconsistent in regard to the influence of CHO-restricted diets on serum TC. Some studies have shown an increase in TC,^16,47 while Boden et al.⁴² demonstrated a decrease in TC during CHO-restricted diets. More consistently, higher CHO lower fat diets have been demonstrated by numerous studies to decrease TC^{14,15,49,51,52} and LDL.^{14,15,49,51,53} During the present study, there was no significant difference among the diets at the end of 4 weeks. This is consistent with studies done by Foster et al.² and Stern et al.⁵ and may be more related to weight loss than diet composition.

HDL-C decreased by ∼10% in subjects after consuming the LF diet for 1 week, and the concentration remained lower than baseline throughout the feeding period. A similar decrease in HDL-C concentration has also been demonstrated by numerous other researchers feeding lower fat CHO-rich diets.^{21,49,53–55} By the end of 4 weeks of feeding, however, there were no significant differences in HDL-C levels between subjects, which is consistent with the research by Layman et al.¹⁴ Although a decrease in HDL-C is considered an adverse cardiovascular disease risk factor,^54,55 the concentrations between groups were not different at the end of 4 weeks, which may have been due to a concurrent loss of body weight. Furthermore, research has not shown that the physiological reduction of serum concentrations of HDL-C that occasionally occurs with the consumption of higher CHO lower fat diets increases cardiovascular disease risk.^54,56 However, further research in this area is necessary to draw a more decisive conclusion.

The VLCPM diet caused a significant reduction in concentration of LDL-C over the 4-week study period; however, the concentrations of LDL-C after 4 weeks were not significantly different from the other diet groups. This finding is inconsistent with previous research that has shown that LDL-C increases with the consumption of CHO-restricted diets.^16,18,47 This could be due to the relatively low levels of SFAs fed during the VLCPM trial, considering consumption of SFAs can raise LDL-C.⁵⁷ The female subjects consumed only 17 ± 2 g (males 24 ± 4 g) of the SFA VLCPM diet, while the female subjects on the other CHO-restricted diet trial consumed 38 ± 8 g and the men consumed 50 ± 5 g. LDL-C tends to increase with higher levels of SFAs. Replacing SFAs with monounsaturated fatty acids or polyunsaturated fatty acids has also been shown to have beneficial results on cardiovascular risk factors by decreasing LDL-C and increasing HDL-C.^11,29 The VLCPM trial contained more unsaturated fatty acids than the other trials. Weight loss could also independently contribute to the observed decrease in LDL-C.⁵⁰ During the LF trial, the LDL-C concentrations rose after 1 week and then tended to decrease between 1 and 4 weeks. Last, consumption of plant proteins³⁰ as well as mushrooms⁵⁸ can reduce serum LDL-C concentration, which may explain the different responses after consuming the VLCPM diet versus the VLCA diet.

All three diets produced positive results with regard to total serum cholesterol concentrations, but the CHO-restricted diets produced more favorable changes in the blood lipid profiles overall. Consumption of the VLCPM diet appeared to be the most beneficial of the three during weight loss since lower concentrations of TGs, TC, and LDL-C were detected with no adverse change in HDL-C. Serum TGs and TC were also lowered after consuming the VLCA diet, but the lipoprotein responses were not of sufficient magnitude to detect statistically significant improvements. After a transient increase in serum TG concentrations during the LF trial, no significant effects on TGs were detected by the end of the fourth week, although a hypocholesterolemic effect was induced. The VLCPM diet may have produced the most beneficial effects during weight loss due to a variety of differences, including higher unsaturated versus SFAs, a higher intake of plant proteins relative to animal proteins, and/or the incorporation of 170 g of white button mushrooms daily. Furthermore, a slightly greater loss of body weight was detected during the VLCPM trial, which may, in part, explain the differing effects.

There are several possible limitations in the current study: (1) we did not provide food based on energy requirement; (2) food was provided for consumption at home rather than in a metabolic ward; (3) although not significantly different, there were differences in actual energy intake between the diet trials; (4) a majority of the subjects were women; and (5) given that a majority of the subjects were women, it is possible that differences in sex hormones could have influenced our results.

Taken together, these data suggest that an energy-restricted VLC diet, including mushrooms and additional plant foods, produces weight loss at least equal to an animal-based low-carbohydrate diet and perhaps greater than a lower fat diet by decreasing energy intake. Furthermore, although each diet had favorable effects on metabolism, including enhanced insulin sensitivity, lower serum leptin concentrations, and improvements in some serum lipids, low-carbohydrate diets may produce more favorable effects on serum concentrations of TGs, HDL-C, and LDL-C. Moreover, a low-carbohydrate diet, including mushrooms and more plant foods, may produce additional advantages over an animal-based low-carbohydrate diet such as greater improvements in serum glucose homeostasis.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

The work was supported by a grant from The Mushroom Council (M.K.).