Very Low-Carbohydrate versus Isocaloric High-Carbohydrate Diet in Dietary Obese Rats

Authors

  • Kathleen V. Axen,

    Corresponding author
    1. Department of Health and Nutrition Sciences, Brooklyn College of City University of New York, Brooklyn, New York
      Department of Health and Nutrition Sciences, Brooklyn College of C.U.N.Y., 2900 Bedford Avenue, Brooklyn, NY 11210. E-mail: kaxen@brooklyn.cuny.edu
    Search for more papers by this author
  • Kenneth Axen

    Corresponding author
    1. Department of Health and Nutrition Sciences, Brooklyn College of City University of New York, Brooklyn, New York
      Department of Health and Nutrition Sciences, Brooklyn College of C.U.N.Y., 2900 Bedford Avenue, Brooklyn, NY 11210. E-mail: kaxen@brooklyn.cuny.edu
    Search for more papers by this author

Department of Health and Nutrition Sciences, Brooklyn College of C.U.N.Y., 2900 Bedford Avenue, Brooklyn, NY 11210. E-mail: kaxen@brooklyn.cuny.edu

Abstract

Objective: The effects of a very low-carbohydrate (VLC), high-fat (HF) dietary regimen on metabolic syndrome were compared with those of an isocaloric high-carbohydrate (HC), low-fat (LF) regimen in dietary obese rats.

Research Methods and Procedures: Male Sprague-Dawley rats, made obese by 8 weeks ad libitum consumption of an HF diet, developed features of the metabolic syndrome vs. lean control (C) rats, including greater visceral, subcutaneous, and hepatic fat masses, elevated plasma cholesterol levels, impaired glucose tolerance, and fasting and post-load insulin resistance. Half of the obese rats (VLC) were then fed a popular VLC-HF diet (Weeks 9 and 10 at 5% and Weeks 11 to 14 at 15% carbohydrate), and one-half (HC) were pair-fed an HC-LF diet (Weeks 9 to 14 at 60% carbohydrate).

Results: Energy intakes of pair-fed VLC and HC rats were less than C rats throughout Weeks 9 to 14. Compared with HC rats, VLC rats exhibited impaired insulin and glycemic responses to an intraperitoneal glucose load at Week 10 and lower plasma triacylglycerol levels but retarded loss of hepatic, retroperitoneal, and total body fat at Week 14. VLC, HC, and C rats no longer differed in body weight, plasma cholesterol, glucose tolerance, or fasting insulin resistance at Week 14. Progressive decreases in fasting insulin resistance in obese groups paralleled concomitant reductions in hepatic, retroperitoneal, and total body fat.

Discussion: When energy intake was matched, the VLC-HF diet provided no advantage in weight loss or in improving those components of the metabolic syndrome induced by dietary obesity and may delay loss of hepatic and visceral fat as compared with an HC-LF diet.

Introduction

Very low-carbohydrate (VLC)1 diets are advocated to the public for weight loss and reduction of risk of cardiovascular disease and type 2 diabetes (1). The high fat (HF) content (50% to 65% of energy) typically supplied by such diets (2) is contrary to the recommendations of professional organizations regarding these chronic diseases (3, 4). In some studies of obese humans, short-term (1 to 6 months) use of VLC-HF vs. high-carbohydrate (HC), low-fat (LF) diets produced a greater reduction in body weight (5, 6, 7), body fat (8), and plasma triacylglycerol levels (5, 6, 7, 9) and a greater rise in plasma high-density lipoprotein (HDL)-cholesterol levels (9, 10). In other studies, some of these effects did not occur (5, 7, 10) or disappeared by 12 months (5, 11). Because most human studies of VLC-HF diets rely on the subjects’ self-reports of food consumption, differences in energy intakes between groups may go undetected; therefore, the relative contributions of lower carbohydrate intake vs. lower energy intake in producing these outcomes remain uncertain.

When energy intake is held constant, LC (15% to 35%), HF diets have been shown in rats to produce greater adiposity (12), impaired glucose tolerance (13, 14), and higher hepatic fat mass (13) than HC-LF diets. Because the premise underlying the effectiveness of VLC diets is a marked reduction in plasma insulin levels due to lack of stimulation by dietary carbohydrate (1), the aforementioned studies may not have employed a sufficiently low carbohydrate content to achieve the conditions sought in VLC regimens. Although an extremely long-term (16 months) VLC-HF diet (10% carbohydrate) has been shown to induce some features of the metabolic syndrome in lean rats (15), the effects of a VLC-HF dietary regimen on obese rats has not been reported.

The present study was, therefore, designed to compare the effect of a VLC-HF diet with that of an isocaloric HC-LF diet on body weight and a number of the components of the metabolic syndrome, a group of risk factors for cardiovascular disease and type 2 diabetes. Rats were first rendered obese through ad libitum consumption of an HF diet previously shown to induce hyperphagia (13). The present protocol utilized daily pair-feeding to equalize energy intakes of the VLC-HF and HC-LF diets in two weight-matched groups. The VLC-HF diet followed a popular two-phase regimen consisting of an extremely low carbohydrate (5%) intake, followed by a somewhat higher (15%) intake (1); it had a similar macronutrient distribution as reported for adherents of the popular diet at-large (2) and in clinical trials (7, 8, 9, 10); and it was nutritionally complete and provided a nutritionally favorable distribution of fatty acids.

Research Methods and Procedures

Animals and Diets: Induction of Obesity (Weeks 1 to 8)

Male Sprague-Dawley rats (N = 36, age ∼8 weeks; Charles River Laboratories, Wilmington, MA) were individually housed in mesh-bottomed cages at 20 °C to 22 °C, with a 12-hour light/dark cycle. Rats were separated into two weight-matched groups fed ad libitum; 26 rats (HF) received an HF diet (Table 1), which we have found to be effective in producing obesity in rats (13), and a lean control (C) group of 10 rats received Purina 5001 pellets (Chow; PMI Feeds, Inc., St. Louis, MO). The HF diet included hydrogenated vegetable fat (Proctor & Gamble, Cincinnati, OH) and was supplemented with protein and micronutrients (Bio-serv, Frenchtown, NJ). Water was consumed ad libitum throughout the experiment. At Week 8, 4 C and 6 HF rats received glucose tolerance tests (GTTs) and were used in a terminal procedure for tissue collection.

Table 1.  Diet composition
 
Ingredients (g/kg)*ChowHFHC-LFVLC-HF1VLC-HF2
  • HF, high fat; HC-LF, high carbohydrate-low fat; VLC-HF, very low-carbohydrate, high fat.

  • *

    Values, except for energy density, are rounded to units.

PMI 50011000411   
Cellulose  505250
Cornstarch  59372204
Casein 232244495336
Methionine 3364
Hydrogenated fat 329   
Lard  48277263
Corn oil  179591
Canola oil  2109
AIN-93 vitamin mix 20343534
AIN-93 mineral mix 6101010
Carbohydrate (% energy)601560515
Fat (% energy)1260156060
Protein (% energy)2825253525
Fiber (% weight)146555
Energy (kJ/g)13.822.618.325.125.1

VLC-HF vs. HC-LF Diets (Weeks 9 to 14)

The remaining 20 HF rats were separated into two weight-matched groups (each N = 10). The VLC group received a two-phase VLC-HF diet, designed according to Atkins’ description (1) and matching the macronutrient distribution reported for its adherents (2). The VLC-HF1 diet provided 5% of energy as carbohydrate during Phase 1 (Weeks 9 and 10), and the VLC-HF2 diet provided 15% of energy as carbohydrate in Phase 2 (Weeks 11 to 14). The HC group received an HC-LF diet for 6 weeks (Table 1). Note that no diet contained sucrose; all diets in Phases 1 and 2 were high in protein and had a ratio of saturated:monounsaturated:polyunsaturated fat of ∼1:1.5:2.

VLC and HC rats were pair-fed throughout the 6-week study period to match energy intakes of the two groups. Obese rats spontaneously consumed more of the HC-LF diet during Phase 1 (Weeks 9 and 10) but more of the VLC-HF diet during Phase 2 (Weeks 11 to 14). The daily energy intakes of VLC and HC rats were, therefore, matched as follows: HC rats received the amount of energy (kilojoules) consumed ad libitum by VLC rats the previous day during Phase 1, whereas VLC rats received the amount of energy consumed ad libitum by HC rats the previous day during Phase 2. The remaining six lean C rats received the HC-LF diet ad libitum. The protocol was approved by the Brooklyn College Institutional Animal Care and Use Committee.

In Vivo Measurements

Food intakes, corrected for spillage, were measured at least twice a week (daily during pair-feeding); body weights were recorded once a week. To assess ketosis at the end of Phase 1 (Week 10), tail blood was obtained in the morning from non-food-deprived VLC and HC rats for measurement of β-hydroxybutyrate.

At the end of the obesity induction period (Week 8), four C and six HF rats were deprived of food for 16 hours overnight and were given an intraperitoneal (IP) injection of glucose (1 g/kg body weight, 50% solution); tail blood was obtained at pre-injection (t = 0) and 10, 20, and 75 minutes post-injection for determination of plasma glucose and insulin levels. The same protocol was utilized during Week 10 (end of Phase 1) for four VLC and four HC rats and during Week 14 (end of Phase 2) for six VLC, six HC, and six C rats. At Weeks 8, 10, and 14, the same rats that had undergone the GTT were anesthetized in the morning in the fed state by IP injection with a mixture of ketamine (63 mg/kg) and xylazine (9.4 mg/kg) (Butler, Columbus, OH). Blood was obtained from the aorta for determination of plasma, triacylglycerol, total cholesterol, and HDL-cholesterol levels. Anesthetized rats were killed by exsanguination. Livers were excised, and samples were stored at −80 °C for later lipid measurement. Fat pads from three visceral fat regions (epididymal, retroperitoneal + perirenal, and mesenteric + omental) were dissected from the rats and weighed. The shaved coat with attached subcutaneous fat was removed. Composition analysis (16) was done on the coat and on the carcass, defined here as the body minus the coat, liver, intestine, pancreas, and visceral fat pads.

Analyses

Plasma insulin was measured using a double-antibody radioimmunoassay kit specific for rat insulin (Linco, St. Charles, MO). Assay kits were utilized for measurement of plasma concentrations of the following: β-hydroxybutyrate (Stanbio, Boerne, TX), triacylglycerol (GPO method; Stanbio), total cholesterol (cholesterol-E; Wako, Richmond, VA), and HDL-cholesterol (HDL-cholesterol-E; Wako). Plasma glucose levels were measured using a YSI Biochemistry Analyzer (YSI, Yellow Springs, OH), and liver lipid was extracted with chloroform-methanol (16). Statistical analyses were performed by ANOVA and Newman-Keuls post hoc test (Crunch 4, Crunch Software Corp.); data are presented as mean ± standard error (SE), and differences are considered to be significant at p < 0.05.

Results

Induction of Obesity (Weeks 1 to 8)

Body Weight and Body Composition

During the 8-week period of ad libitum feeding, rats on the HF diet had a greater energy intake (p < 0.001, Figure 1) and attained higher body weights than did rats on the C diet (p < 0.001, Figure 2). When compared with C rats at Week 8, HF rats had greater visceral fat mass, including heavier mesenteric plus omental (p < 0.001), epididymal (p < 0.001), and retroperitoneal plus perirenal (p < 0.01) fat pads; greater subcutaneous (p < 0.02) and carcass (p < 0.05) fat masses; and higher hepatic lipid content (p < 0.02, Table 2). Total fat represented a greater percentage of body weight in the HF group (19.4 ± 1.0%, mean ± SE) than the C group (12.2 ± 1.1%, p < 0.001).

Figure 1.

Energy intake vs. time. Plotted symbols depict mean ± SE of data from groups of C, HF (Weeks 1 to 8), VLC, and HC rats. Vertical lines at Weeks 8 and 10 indicate beginning of Phases 1 and 2, respectively. * HF > C, p < 0.05; # C > VLC ∼ HC, p < 0.05. Note that spontaneous energy intake of C rats exceeds that of pair-fed VLC and HC rats during Phases 1 and 2.

Figure 2.

Body weight vs. time. Plotted symbols depict mean ± SE of data from groups of C, HF (Weeks 1 to 8), VLC, and HC rats. Vertical lines at Weeks 8 and 10 indicate beginning of Phases 1 and 2, respectively. * HF > C, p < 0.05; # C < VLC ∼ HC, p < 0.05. Note that differences in body weight among groups disappear at Week 12.

Table 2.  Body composition
 Fat pad (g)Lipid extract (g)
 EpididymalRetro*MesentericSubQCarcass§LiverTotal
  • Retro, retroperitoneal; SubQ, subcutaneous; C, control; HF, high-fat; VLC, very low-carbohydrate; HC, high-carbohydrate.

  • *

    Retroperitoneal + perirenal fat pads.

  • Mesenteric + omental fat pads.

  • Subcutaneous fat extracted from dissected coat.

  • §

    Carcass = body − (coat + dissected fat pads + liver + intestine + pancreas).

  • Different from C rats, p < 0.05; mean ± SE.

  • Different from HC rats, p < 0.05.

Week 8       
 C7.8 ± 1.08.3 ± 1.46.6 ± 1.117.4 ± 3.823.2 ± 3.80.9 ± 0.164.2 ± 8.2
 HF17.5 ± 2.224.1 ± 1.913.1 ± 1.940.4 ± 4.431.1 ± 1.8 2.2 ± 0.4128.3 ± 10.7
Week 10       
 VLC19.1 ± 0.421.4 ± 1.516.0 ± 1.951.8 ± 7.932.7 ± 2.31.9 ± 0.2142.8 ± 12.8
 HC16.2 ± 4.022.2 ± 4.013.9 ± 2.949.5 ± 11.329.0 ± 4.62.0 ± 0.4132.7 ± 25.5
Week 14       
 C14.6 ± 2.020.4 ± 2.113.1 ± 2.037.5 ± 3.930.1 ± 2.31.0 ± 0.2116.7 ± 11.0
 VLC17.4 ± 1.127.1 ± 1.9 14.6 ± 1.250.8 ± 6.032.7 ± 2.31.7 ± 0.2144.4 ± 10.9
 HC15.0 ± 1.421.3 ± 1.611.6 ± 1.342.6 ± 3.028.5 ± 1.51.3 ± 0.2120.3 ± 8.3

Plasma Lipid Concentrations

Total plasma cholesterol levels were higher in HF than C rats (p < 0.02), but there were no significant differences between groups in plasma levels of HDL-cholesterol or triacylglycerol in the fed state (Table 3).

Table 3.  Plasma lipid concentrations
 Triacylglycerol (mM)*HDL-cholesterol (mM)Total cholesterol (mM)
  • HDL, high-density lipoprotein; C, control; HF, high-fat; VLC, very low-carbohydrate; HC, high-carbohydrate.

  • *

    Based on the weight of triolein.

  • Different from C at Week 8, p < 0.05.

  • Different from VLC at Week 14, p < 0.01.

Week 8   
 C1.11 ± 0.190.96 ± 0.211.19 ± 0.13
 HF1.33 ± 0.261.25 ± 0.091.86 ± 0.18
Week 10   
 VLC2.08 ± 0.781.08 ± 0.041.95 ± 0.17
 HC1.86 ± 0.691.00 ± 0.111.78 ± 0.21
Week 14   
 C1.57 ± 0.341.40 ± 0.451.85 ± 0.17
 VLC1.04 ± 0.211.41 ± 0.152.18 ± 0.28
 HC1.90 ± 0.181.17 ± 0.191.94 ± 0.20

Glycemic Control

Although fasting plasma glucose levels (G0) and fasting plasma insulin levels (I0) did not differ significantly between the HF and C groups, calculated values of fasting I × G [I0 × G0, a measure of insulin resistance analogous to the homeostasis model assessment index (17)] were considerably higher (∼190% of mean C values) in the HF (4.27 ± 0.63 nM × mM) than C group at Week 8 (2.24 ± 0.32 nM × mM, p < 0.05). This result indicates that obese HF rats exhibited fasting insulin resistance.

During the GTT, plasma glucose levels 10 minutes after an IP injection of glucose were higher in HF than C rats (p < 0.001, Figure 3A), demonstrating impaired glucose tolerance in obese HF rats. These higher levels of plasma glucose in HF rats were accompanied by higher levels of plasma insulin at 20 minutes post-injection (I20) (HF > C, p < 0.05, Table 4), providing evidence that obese HF rats also exhibited post-load insulin resistance. Comparison of I20 with I0 demonstrated that both the obese HF group and the lean C group produced a significant rise in plasma insulin (I20 > I0) in response to an IP glucose load (p < 0.05).

Figure 3.

Plasma glucose responses to an IP glucose load. Data collected at Weeks 8 (A), 10 (end of Phase 1; B), and 14 (end of Phase 2; C). Plotted symbols depict mean ± SE of data from C, HF, VLC, and HC groups obtained at 0 (pre-injection), 10, 20, and 75 minutes post-injection. * p < 0.05.

Table 4.  Plasma concentrations of insulin
 Fasting (pM)§20 minutes post-load (pM)
  • C, control; HF, high-fat; VLC, very low-carbohydrate; HC, high-carbohydrate.

  • Blood was drawn in the morning from rats food-deprived for 16 to 18 hours.

  • Blood was drawn 20 minutes after an intraperitoneal injection of 50% glucose (1 g/kg body weight).

  • §

    1 ng/mL ∼ 175 pM for rat insulin.

  • Greater than fasting insulin concentration, p < 0.05.

  • Greater than values in C rats, p < 0.05.

Week 8  
 C324 ± 38679 ± 72
 HF572 ± 75998 ± 66
Week 10  
 VLC297 ± 168598 ± 250
 HC385 ± 119913 ± 51
Week 14  
 C320 ± 80677 ± 142
 VLC259 ± 119464 ± 140
 HC178 ± 68338 ± 175

Glucose clearance between 20 and 75 minutes of the GTT (ΔG%) was quantified in each rat by dividing the observed reduction in plasma glucose during this period (G20 − G75) by the amount by which plasma glucose at 20 minutes exceeded its fasting baseline level (G20 − G0) and expressing the ratio as a percentage, i.e., ΔG% = (G20 − G75) × 100/(G20 − G0), where G0, G20, and G75 represent plasma glucose levels at 0 (baseline), 20, and 75 minutes after injection, respectively. As expected from the similar slopes of the lines depicting the decline in plasma glucose between 20 and 75 minutes of the GTT (Figure 3A), plasma glucose clearance during this latter 55 minutes period did not differ significantly between the HF (59.8 ± 11.1%) and C (71.0 ± 2.9%) groups. This finding of similar glucose clearance in the face of higher plasma insulin provides corroborating evidence that obese HF rats exhibited post-load insulin resistance.

Comparison of Isocaloric VLC-HF and HC-LF Diets

Body Weight and Body Composition

Spontaneous energy intakes of C rats exceeded those of pair-fed VLC and HC rats throughout Phases 1 and 2 (C > VLC ∼ HC, p < 0.001, Figure 1); the VLC-HF and HC-LF diets were, therefore, hypocaloric under the present experimental conditions. At the end of Phase 1, plasma levels of β-hydroxybutyrate were elevated (>0.25 mM) in five of 10 rats on the VLC-HF1 diet and in four of 10 rats on the HC-LF diet, indicating that both hypocaloric diets were also ketogenic. Averaged values of β-hydroxybutyrate were the same in the VLC (0.292 ± 0.065 mM) and HC (0.285 ± 0.062 mM) groups.

Despite lower energy intakes (Figure 1), body weights of both VLC and HC rats still exceeded those of C rats throughout Phase 1 (VLC ∼ HC > C, p < 0.01). However, during Phase 2, C rats continued to gain weight, as expected from normal growth, whereas VLC and HC rats maintained essentially constant body weights throughout this latter 4-week period. As a result, group differences in body weight disappeared by the end of Phase 2 (VLC ∼ HC ∼ C).

At the end of Phase 1, no significant differences were found between the VLC and HC groups in visceral fat pad and carcass fat masses or hepatic lipid contents (Table 2). At the end of Phase 2, however, retroperitoneal fat pads of VLC rats were heavier than those of both C (p < 0.05) and HC (p < 0.02) rats (Table 2); total body fat was greater in VLC than C rats (p < 0.05); hepatic lipid contents of VLC rats, but not HC rats, exceeded those of C rats (p < 0.01); and hepatic lipid contents of HC rats, but not VLC rats, were lower than those of HF rats at Week 8 (p < 0.05). Therefore, when compared with a 6-week isocaloric HC-LF diet, the VLC-HF diet retarded the loss of hepatic, retroperitoneal, and total body fat in previously obese rats.

Plasma Lipid Concentrations

Plasma levels of total cholesterol and HDL-cholesterol did not differ between the VLC and HC groups at Weeks 10 or 14 (Table 3). Non-fasting plasma levels of triacylglycerol of VLC rats at Week 14 were similar to those of HC rats at Week 10 and C rats at Week 14 but were lower than those of HC rats at Week 14 (HC > VLC ∼ C, p < 0.01) (Table 4). This result shows that the HC-LF diet was associated with higher plasma levels of triacylglycerol than the VLC-HF diet at the end of Phase 2.

Glycemic Control (Phase 1)

At Week 10, G0 and I0 were similar in the VLC and HC groups. Fasting I × G values from the VLC and HC groups at Week 10 were similar to those of the C group at Week 8, indicating that 2 weeks of energy restriction by either diet eliminated the fasting insulin resistance observed in obese HF rats at Week 8.

When compared with HC rats at Week 10, VLC rats showed higher plasma glucose levels 10 minutes after an IP glucose load (p < 0.005, Figure 3B) and greater areas under the glucose vs. time curve (p < 0.05). When compared with HF rats at Week 8, VLC rats also showed higher plasma glucose levels at 10 and 20 minutes post-injection (p < 0.01, Figure 3A). In contrast, plasma glucose responses from HC rats at Week 10 were similar to those of HF rats at Week 8, demonstrating that 2 weeks of the VLC-HF1 diet, but not the HC-LF diet, worsened glucose tolerance in obese rats.

At Week 10, the HC group, but not the VLC group, exhibited a significant rise in plasma insulin (I20 > I0, Table 4, p < 0.05) in response to the IP glucose load. This impaired insulin response in VLC rats helps explain the higher levels of plasma glucose observed in this group during the GTT (Figure 3B).

Glycemic Control (Phase 2)

At Week 14, G0 and I0 (Table 4), fasting I × G values, and plasma glucose levels throughout the GTT (Figure 3C) no longer differed among the three groups (VLC ∼ HC ∼ C), indicating that 6 weeks of either hypocaloric diet restored normal glycemic control in previously obese rats. Analysis of plasma insulin responses to an IP glucose load at Week 14 (Table 4) revealed that the C group (p < 0.02), but not the VLC group or the HC group, produced a significant rise in plasma insulin (I20>I0), and the I20 levels of both VLC and HC groups were lower than those observed in HF rats at Week 8 (HF>VLC∼HC, p < 0.001). These findings demonstrate that 6 weeks of either hypocaloric diet attenuated the insulin response to an IP glucose load in previously obese rats.

Despite these differences in insulin response, however, glucose clearance between 20 and 75 minutes of the GTT (ΔG%) in VLC (58.3 ± 8.0%) and HC (62.6 ± 6.1%) rats at Week 14 were similar to those observed in HF rats at Week 8 (59.7 ± 12.1%) and C rats at Weeks 8 (70.9 ± 3.2%) and 14 (63.1 ± 11.7%). This finding of similar glucose clearance in the face of subnormal insulin responses suggests that the restoration of normal glycemic control in VLC and HC rats at the end of Phase 2 (Week 14) was mediated, in part, by corresponding increases in post-load insulin sensitivity. Consistent with this view, the fasting I × G values of HF rats at Week 8 (4.27 ± 0.63 nM × mM) greatly exceeded those of VLC (1.58 ± 0.63 nM × mM, p < 0.02) and HC (1.16 ± 0.47 nM × mM, p < 0.001) rats at Week 14, indicating corresponding reductions in fasting insulin resistance in both the VLC and HC groups.

Relationship between Fasting Insulin Resistance and Body Composition

Linear regression analysis revealed significant correlations between individual fasting I × G values and concomitant values for hepatic lipid content (r = 0.49, p < 0.02), retroperitoneal fat (r = 0.44, p < 0.05), and total body fat (r = 0.41, p < 0.05) when data from obese HF rats at Week 8, VLC rats at Weeks 10 and 14, and HC rats at Weeks 10 and 14 were analyzed (Figure 4). These results show that the progressive decrease in fasting insulin resistance over the 6-week period of the VLC-HF and HC-LF diets paralleled the concomitant reductions in hepatic, retroperitoneal, and total body fat induced by these diets. When data from C rats at Weeks 8 and 14 were incorporated into the analysis, however, the relationship between fasting I × G values and hepatic lipid contents became weaker (r = 0.38, p < 0.05), and the relationships between fasting I × G values and retroperitoneal and total body fat were no longer statistically significant (p > 0.05).

Figure 4.

Fasting I × G value (an index of insulin resistance) vs. hepatic lipid content. Plotted symbols depict mean of data from groups of HF and C rats at Week 8, VLC and HC rats at Week 10, and VLC, HC, and C rats at Week 14. Note that data from C rats do not fall on the line formed by averaged data from HF, VLC, and HC rats.

Discussion

As described previously (13), the initial HF dietary protocol (Weeks 1 to 8) produced hyperphagia (Figure 2), obesity, and features of the metabolic syndrome. These features included increased visceral and hepatic fat mass (Table 2), elevated plasma total cholesterol levels (Table 3), impaired glucose tolerance (Figure 3A), fasting insulin resistance (based on elevated fasting I × G values), and post-load insulin resistance (based on concomitantly higher glucose and insulin levels during an IP GTT) when compared with C rats.

Obese HF rats were separated into two weight-matched, pair-fed groups, with VLC rats receiving the two-phase VLC-HF diet and the HC rats receiving the HC-LF diet. Because VLC and HC rats had lower energy intakes than C rats during Phases 1 and 2 (Weeks 9 to 14, Figure 1), both VLC-HF and HC-LF dietary regimens were hypocaloric as compared with that of lean C rats feeding ad libitum. The effects of these diets on obese rats were analyzed by partitioning results into those due to energy restriction (similarities between VLC and HC groups) and those due to diet composition (differences between VLC and HC groups).

Energy restriction and not diet composition was responsible for the observed ketosis at the end of Phase 1 (Week 10) because equal numbers of rats on the VLC-HF and HC-LF diets showed elevated plasma β-hydroxybutyrate levels in the fed state. The failure of the VLC-HF1 diet to be more ketogenic than the HC-LF diet, which was equally hypocaloric but much higher in carbohydrate (5% vs. 60%), may be due to the high protein content of the VLC diet (35%, Table 1). In our pilot study (data not shown) and in other studies (18, 19), ketosis, along with marked hypophagia, was consistently produced using higher fat but much lower protein contents (5% to 13%). The results suggest that the dietary conditions needed to induce ketosis may differ between humans and rats.

Diet composition did not influence body weight (Figure 2) or epididymal, mesenteric, subcutaneous, or carcass fat masses (Table 2) at the end of Phases 1 or 2, given that values from the VLC and HC groups did not differ. In contrast, diet composition was responsible for the higher retroperitoneal fat pad weights, elevated hepatic lipid content, and total fat mass in VLC, but not HC, rats when compared with that of C rats at the end of Phase 2 (Table 2). These results agree with other studies in rats in which HF feeding produced elevated visceral fat mass (12, 13, 20) and hepatic lipid content (13, 21) as compared with isocaloric HC feeding. Taken together, these findings do not support the proposition that dietary carbohydrate and its resultant elevation of insulin levels are responsible for fat deposition in adipose tissue and liver and that these effects are mitigated by a low-carbohydrate, high-fat diet.

Energy restriction and not diet composition accounted for the reduction in total plasma cholesterol levels from that seen at Week 8 in HF rats after 2 or 6 weeks on either hypocaloric diet (Table 3). These results agree with studies showing that energy restriction decreases plasma cholesterol levels in rats (22) and humans (23).

A more complex pattern was seen for plasma triacylglycerol levels in the present study. Levels rose during Phase 1 (Week 10) in several obese rats on either diet, but at the end of Phase 2 (Week 14), triacylglycerol levels in the HC group were significantly higher than those of the C or VLC groups. The finding of higher plasma triacylglycerol in HC rats agrees with results in rats fed HC (63%) vs. LC (11%), sucrose-free diets (24). Lower plasma triacylglycerol levels have also been reported for humans instructed to follow VLC vs. HC diets (5, 6, 7, 8, 9). Triacylglycerol-lowering effects of low-carbohydrate diets have been ascribed to decreased secretion of very low-density lipoprotein-triglyceride (24, 25, 26, 27) and to the more rapid clearance of chylomicron-triglyceride (derived from dietary fat) than of very low-density lipoprotein-triglyceride (derived from carbohydrate) (26, 27).

In contrast to the metabolic syndrome, elevated plasma triacylglycerol levels in HC rats in the present study were not accompanied by fasting or post-load insulin resistance. In agreement with this result, a similar dissociation between elevated triacylglycerol levels and insulin resistance was also found in a study comparing non-restricted HC and LC diets in rats (24). Because it has been proposed that elevation in plasma triacylglycerol levels may not be a cardiovascular risk factor when insulin resistance is not present (28, 29), the relevance of this effect of diet composition to disease risk remains uncertain.

On the other hand, lower plasma levels of triacylglycerol in VLC rats in the present study were accompanied by higher hepatic lipid content, a component of the metabolic syndrome. Lower plasma triacylglycerol levels coupled with higher hepatic lipid content were also observed in our previous study of lean rats exhibiting normal growth on an energy-restricted diet similar to VLC-HF2 (13). In contrast, elevated plasma levels of triacylglycerol were associated with higher hepatic lipid content in rats rendered obese by alternating HF (36% carbohydrate) and LF diets (30). These findings collectively suggest that: 1) HF diets promote lipid storage in liver but that elevation of plasma triacylglycerol levels might require a higher carbohydrate intake than that typically provided by VLC diets, and/or 2) lower plasma triacylglycerol levels observed in energy-restricted VLC-HF diets might be at the expense of greater retention of hepatic and visceral lipid stores (Table 2).

Diet composition affected glucose tolerance in Phase 1, when the VLC but not the HC group showed a diminished insulin response coupled with a further deterioration in glycemic control from Week 8 (Figure 3B). These results agree with reports of decreased insulin secretory response and impaired glucose tolerance in humans (31, 32) and in rats (33) with very low carbohydrate intakes for a few days or for more than a year (15). Impaired ability to dispose of a carbohydrate load could be of functional significance when a higher carbohydrate diet is resumed or deviation from the VLC regimen occurs.

Energy restriction and not diet composition was responsible for the progressive abatement of fasting insulin resistance from Week 8, as evidenced by the similar reductions in fasting I × G values in both VLC and HC groups throughout the 6-week dietary regimen. These reductions in fasting insulin resistance paralleled the concomitant reductions in hepatic lipid content (Figure 4), retroperitoneal fat mass, and total fat mass. These findings agree with previous studies showing a similar relationship between hepatic lipid content and insulin resistance in rats (30), mice (34), and humans (35).

Energy restriction was also responsible for the improvement in glucose tolerance after a glucose load because glucose values in both VLC and HC groups no longer differed from those in the C group by the end of Phase 2 (Figure 3C). This restoration of normal glucose tolerance occurred despite markedly attenuated insulin responses when compared with obese HF rats at Week 8 or C rats at Weeks 8 and 14 (Table 4). These results indicate that VLC and HC rats differed metabolically from C rats at Week 14 despite similar body weights, and that the improvement in glycemic control in VLC and HC rats was mediated, in part, by corresponding increases in post-load insulin sensitivity. These findings collectively suggest that VLC and HC rats exhibited a similar enhancement of both fasting and post-load insulin sensitivity despite widely differing carbohydrate and fat intakes; such an effect can be attributed to the chronic energy restriction of both groups, possibly through mobilization of lipid stores in muscle (34, 36) and liver (35).

In summary, when energy intake was matched, the VLC-HF diet provided no advantage over the HC-LF diet in loss of body fat, decrease in plasma cholesterol levels, improvement in fasting or post-load insulin resistance, or restoration of glycemic control in dietary obese rats. These results underscore the need for strict control of energy intake in studies of the effects of VLC diets on obesity and the metabolic syndrome. Although rats on the VLC-HF diet did show lower plasma triacylglycerol levels than did rats on the HC-LF diet by the end of Phase 2, this reduction was accompanied by retarded loss of hepatic, retroperitoneal, and total body fat mass. These latter findings support the view that a VLC-HF diet may promote storage of fat by the liver, visceral fat pads, and other sites as compared with an HC-LF diet.

Acknowledgments

We thank Amit Khaneja for contributions in data collection. This work was supported, in part, by Professional Staff Congress-City University of New York Research Award 65274-00 34.

Footnotes

  • 1

    Nonstandard abbreviations: VLC, very low-carbohydrate; HF, high-fat; HC, high-carbohydrate; LF, low-fat; HDL, high-density lipoprotein; C, control; GTT, glucose tolerance test; IP, intraperitoneal; SE, standard error; G0, fasting plasma glucose level; I0, fasting plasma insulin level; I20, plasma insulin at 20 minutes post-injection.

  • The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Ancillary