|Home | About | Journals | Submit | Contact Us | Français|
The aim of the study was to determine what effect weight loss had on intra-abdominal adipose tissue (IAAT) and cardiovascular disease (CVD) risk in 135 premenopausal overweight African-American (AA) and European-American (EA) women matched for BMI. Blood lipids, systolic blood pressure (SBP), diastolic BP (DBP), and IAAT (computed tomography determined) were examined prior to and after an 800 kcal/day diet producing 12 kg-weight loss. Significant decreases in IAAT (~38%), total cholesterol (TC; 3%), low-density lipoproteins (LDLs: 6%), triglycerides (TGs: 27%), cholesterol/high-density lipoprotein ratio (C/HDL ratio: 18%), SBP (3%), and DBP (3%) occurred while HDL increased (16%), following weight loss and 1 month energy balance. Significant interactions between time and race showed that AA women decreased TG and increased HDL proportionately less than EA women. After adjusting for ΔIAAT, none of the CVD variables significantly changed after weight loss with the exception of HDL and C/HDL ratio. After adjusting for ΔLF (leg fat), ΔTC, ΔTG, ΔLDL, and ΔC/HDL ratio were significantly different. Multiple regression showed that independent of each other, ΔIAAT was significantly and positively related to ΔTC (adjusted β = 0.24) and ΔTG (adjusted β = 0.47), and ΔLF was negatively related to ΔTC (adjusted β = −0.19) and ΔTG (adjusted β = −0.18). Overweight and premenopausal AA and EA women benefitted from weight loss by decreasing IAAT and improving CVD risk. The changes in IAAT were significantly related to blood lipids, but loss of LF seems to be related to reduced improvement in TC and TG. Based on these results, interventions should focus on changes on IAAT.
According to the 2008 American Heart Association's Heart and Stroke Statistical Update, cardiovascular disease (CVD) remains the top reason for deaths in the United States (1). Obesity is known to increase CVD risk (2). Unfortunately, despite its known deleterious impact on health, overweight and obesity continue to increase in the western society.
It is now known that not all body fat regions affect risk of disease equally (3, 4). Intra-abdominal adipose tissue (IAAT) confers much more risk for CVD than fat in other parts of the body (3–8). In fact, a number of studies have shown that while IAAT is positively related to increased blood pressure (BP) and an impaired blood lipid profile (3,8), increased leg fat (LF) is negatively related to CVD risk if differences in IAAT are accounted for (3,5,9). Although caution must be used in hypothesizing cause and effect in correlational studies (i.e., elevated LF somehow causes a decreased risk of CVD) these data do strongly suggest that large amounts of IAAT may be harmful, whereas large amounts of LF, in the absence of large amounts of IAAT, are relatively benign (5).
Both overweight/obesity and CVD are higher in African-American (AA) women compared to European-American (EA) women (1,2,10). Conversely, AA women have ~30% less IAAT than EA women of similar BMI (11,12). In addition, AA women have better blood lipid profiles compared to EA women of a similar BMI (1,13). For example, AA women have triglyceride (TG) levels that are over 35% lower than EA women with a similar BMI (11).
When compared to men (6,14) and postmenopausal women (15), premenopausal women have relatively low IAAT. For example, premenopausal slightly overweight EA women (~26% body fat) reportedly have IAAT levels that are ~50% below that for postmenopausal women of a similar BMI, and these IAAT amounts were ~40% below the established IAAT cut-points for CVD risk (16). However, it is not known how much of an improvement in CVD risk will occur following weight loss for moderately overweight premenopausal women, especially AA women.
Aerobic exercise training has been shown to be associated with loss of body fat and improvement in CVD risk (17,18). It is difficult to separate the independent effects of exercise on blood lipids from the effects of caloric restriction on blood lipids, because many studies combine diet and exercise intervention to promote weight loss (19,20). It is well-established that resistance training is a successful mode of exercise for increasing strength and muscle mass (21,22), but it is not clear what effects it has on CVD risk.
Few studies have compared the losses in IAAT and improvements in CVD risk between AA and EA women following diet-induced weight loss (11,12,23). The few that have been conducted have either been on a relatively small sample size (23), studied obese women, or have not included exercise training during diet-induced weight loss (11,12). Therefore, the purpose of this study is to determine what impact combined exercise training (either aerobic or resistance) and diet-induced weight loss (800 kcal/day) have on fat mass, IAAT, and CVD risk in AA and EA premenopausal overweight women. A second objective is to determine whether losses in IAAT or LF best explain the improved CVD risk.
Participants were 135 sedentary, healthy (no previous metabolic or CVD) premenopausal women (69 AA, 66 EA) aged 20–41 years with a BMI between 27 and 30 kg/m2. Women were matched for age and BMI. Requirements for entry into study included a family history of obesity in at least one first-degree relative (BMI > 27 kg/m2), non-smoker, and not taking medications that may interfere with energy expenditure, blood insulin, heart rate, and thyroid function. Subjects were screened for normal glucose tolerance as determined by a 2-h serum glucose of <140 mg/dl following 75 g oral glucose load (24). Participants showed a normal resting and stress electrocardiogram. Both parents/grandparents were of the reported race of the participant. Women were recruited through newspaper advertisements and campus flyers. The Institutional Review Board for Human Use at The University of Alabama at Birmingham (UAB) approved the study, and informed consent was obtained from all participants.
To ensure weight stability before evaluations (<1% variation), women entered a 4-week weight maintenance phase before and after weight loss. During weeks 3 and 4 of the first weight maintenance phase, participants were provided meals (20–22% energy fat, 18–22% energy protein, and 58–62% energy carbohydrate in both weight maintenance and weight loss phases) through the General Clinical Research Center at UAB and continued consuming only these meals for the remainder of the study. Subjects received meals throughout the second 4-week weight maintenance phase. Women were weighed three times during the first 2 weeks and five times during the last 2 weeks of weight maintenance, and energy content was adjusted accordingly to achieve a stable weight. Upon completion of the weight maintenance phases, women were admitted as in-patients for 4-day evaluations during the follicular phase of their menstrual cycle. Metabolic testing took place only during these evaluations. After being discharged from the first evaluation, women began the weight loss phase and were provided an 800 kcal/day diet designed to reduce body weight by >10 kg and to reach a target of ideal weight (BMI < 25 kg/m2).
Body composition (total fat and LF) was evaluated using dual-energy X-ray absorptiometry (Lunar DPX-L densitometer; LUNAR Radiation, Madison WI) at UAB's Nutrition Sciences building using Adult Software version 1.33. IAAT and subcutaneous abdominal adipose tissue (SAAT) was measured at L4–L5 using computed tomography scans in the University Hospital Radiology Department (GE HiLight/Advantage scanner, Milwaukee, WI). All scans were analyzed by the same investigator. Waist and hip circumferences were measured according to the procedures of the Arlie Circumference on the standardization of anthropometric measures (25).
BP was measured with automatic auscultation while lying in the supine position. Readings were taken in the morning after a 12-h fast, and were reported as an average of three successive mornings.
Blood lipids were drawn in the morning after a 12-h fast and were analyzed using the Ektachem DT II system (total cholesterol (TC), high-density lipoprotein (HDL), TG). Low-density lipoprotein (LDL) were estimated using the Friedewald Formula (26).
Aerobic training (40 min, 3×/week, n = 45), resistance training (80% 1-repetition maximum (1-RM), 2 × 10, 3×/week, n = 58), and controls (no exercise intervention, n = 32). Training sessions were supervised at a training facility dedicated to research on UAB's campus and lasted ~50 min. Exercisers began with a 3-min warm up followed by 3–5 min of stretching. Workouts persisted throughout the weight loss phase as well as during the second energy balance phase and immediately before the 4-day hospital admission for evaluation.
By the 15th week of training, aerobic trainers were exercising for 40 min at 75–80% of their maximum heart rate. Th ey continued to train at that intensity and duration through the end of the study. Exercise modalities included cycle ergometry, stair stepping, walking, and running.
Resistance trainers performed 2 sets of 10 repetitions of resistance training exercises at 80% of their 1-RM. The exercises were squats, leg extension, leg curl, elbow flexion, triceps extension, lateral pull-down, bench press, military press, lower back extension, and bent leg sit-ups. Every 3 weeks, the women's 1-RM was evaluated, and adjustments for workouts were made accordingly. Strength was evaluated in all exercises every 3 weeks using methods described previously (27,28). Briefly after a warm up of four to five repetitions with a light weight, progressively heavier lifts were attempted until two consecutive failures occurred. One-minute rest was allowed between attempts and the heaviest weight lifted was considered the 1-RM. The R2 for repeat 1-RM testing for these exercises varies from 0.96 to 0.98 in our lab.
A 2 (time) by 2 (race) ANOVA with repeated measures on time was used to observe the effect of time (weight loss) and race on all variables of interest. Risk factors were also evaluated with two separate analyses of covariance with one using ΔIAAT as an adjusting variable, and a second using ΔLF as an adjusting variable. Pearson product correlations were used to observe relationships between ΔCVD risk factors and ΔIAAT, ΔSAAT, and ΔLF. Multiple regression was used to determine whether ΔIAAT and ΔLF were independently related to the two CVD risk factors that were significantly correlated with both ΔIAAT and ΔLF in the simple correlations (ΔTC and ΔTG).
No significant difference was observed between the adherence to the training for the aerobic trainers 81.1 ± 17.2% and the resistance trainers 77.9 ± 16.0%. A 2 (race) × 3 (group) × 2 (time) ANOVA with repeated measures on time was used to compare CVD risk before and after weight loss. As there were no differences between the exercisers and nonexercisers, the groups were collapsed and all subsequent analyses were done using a time-by-race ANOVA with repeated measures on time. Descriptive values are contained in Table 1. There were no significant differences in age, weight, and BMI at baseline. Despite similar weight loss for AA and EA women (−11.7 kg vs. −12.6 kg, respectively), EA women lost slightly more weight. Despite comparable BMIs, EA women had more total fat at baseline compared to AA women, but both races lost the same amount of fat. EA women had significantly higher IAAT at baseline and lost significantly more after the intervention (−35.3 cm2 vs. −25.2 cm2 for EA and AA women, respectively). No significant racial difference was observed for any variable in Table 1 when percent change between before and after weight loss was observed (Table 1).
Blood lipids and BP (Table 2) changed significantly with weight loss. No race effect or time-by-race interaction was found for either TC or LDL. EA women had significantly higher TG, and the significant time-by-race interaction shows that they lost significantly more TG after weight loss than the AA women. HDL were significantly higher in AA women at baseline and after weight loss, but the time-by-race interaction indicates that the EA women increased HDL significantly more than AA women. Cholesterol/HDL (C/HDL) ratio was higher in EA women at baseline with the significant time-by-race interaction showing that the EA women decreased C/HDL ratio more than the AA women. Systolic BP (SBP) and diastolic BP (DBP) were higher in AA women but both races decreased BP similarly. No significant racial difference was observed for any variable in Table 2 when percent change between before and after weight loss was observed (Table 2).
After adjusting for changes in IAAT, the time effect was no longer significant for TC, TG, and LDL, suggesting the weight loss effect on these variables was mediated by changes in IAAT (Table 2). When ΔLF was used as a covariate, the time effect for HDL was no longer significant (Table 2). The ΔSBP and ΔDBP was no longer significant after adjusting for both IAAT and LF changes, suggesting that both fat depots may be contributing to the decrease in BP, following weight loss.
ΔIAAT was positively related to all ΔCVD risk factors except ΔHDL and ΔDBP, which were unrelated to ΔIAAT (Table 3). ΔSAAT was not related to any ΔCVD risk factor. ΔLF was negatively related to ΔTC and ΔTG but was unrelated to the other ΔCVD risk factors (Table 3). As SAAT was unrelated to any ΔCVD risk factor it was not included in any multiple regression modeling. Multiple regression results show that both ΔIAAT and ΔLF were independently related to ΔTC and ΔTG after adjusting for each other as well as age and race (Table 4).
Our results indicate that moderately overweight AA and EA women benefitted in terms of CVD risk from a moderate weight loss. Both races significantly decreased overall fat and IAAT as well as improved all CVD risk factors measured. This was particularly surprising in the AA women who had IAAT well below the 110 cm2 set point previously proposed for identification of CVD risk (16) and had a relatively benign blood lipid profile prior to weight loss. Consistent with cross-sectional studies in which IAAT is clearly related to CVD risk (3–8), IAAT changes with weight loss explained reductions in TC, TG, and LDL. However, it is not clear which fat depot change mediated changes in other CVD risk factors.
EA women had higher absolute amounts IAAT, and lost more IAAT than AA women. However, proportional decreases in IAAT were similar between the races (AAs = −39.3%, EAs = −37.4%). In addition, the losses in IAAT in the two races fit recently developed models for predicting the well known preferential loss of IAAT that occurs during interventions, which induce weight loss (29,30). It appears that the amount of IAAT lost was appropriate for the amount of IAAT that was present, prior to weight loss in both races.
It is well-established that IAAT is related to increased CVD risk, whereas LF is not (3,5,9). In fact, LF has been associated with decreased CVD risk, at least in studies that have had individuals with relatively similar percent body fat or studies in which statistical adjustments have been made for either total fat or IAAT (5,9). In the present study, IAAT changes were positively related to changes to all CVD risk variables with the exception of ΔHDL and ΔDBP. LF changes correlated negatively with changes in ΔTC and ΔTG. Multiple regression modeling shows that the respective positive and negative relationships for change in ΔIAAT and ΔLF with change in ΔTC and ΔTG are independent of each other. Consistent with the cross-sectional data the results of this study suggest that losses in IAAT may be beneficial to CVD risk. The findings that changes in LF are negatively associated with measured metabolic variables, and that changes in IAAT are positively associated with these variables may indicate that different fat depots have distinct biological activity. Thus, it could be envisioned that accumulation of subcutaneous fat is acting as a healthy way to accumulate energy surplus, whereas lipid partition to the IAAT depot promotes key metabolic perturbations such as high TG. These metabolic perturbations may originate from the biological activity of these fat depots. However, we do not find any significant relationships between blood lipids and BP with SAAT, suggesting that biological activity is different between subcutaneous fat in the abdomen and fat in the legs, which is also primarily subcutaneous. Other explanations are also possible, however. For example variables that may cause individuals to preferentially deposit fat in the viscera, such as parity (31), cytokines (32), catecholamines, cortisol (33), estrogens, or some other hormones may be the mediating factor in reducing CVD risk rather than the LF per se. This would mean of course that this variable or variables would have to be in the decreasing range during weight loss, mediating the preferential reduction in IAAT over LF while also mediating a larger decrease in CVD risk. Obviously, further research is needed in addressing this important conundrum.
Consistent with other studies (34–36), exercise appeared to offer no enhanced benefit to CVD risk factors over diet weight loss; it seems plausible to conclude that loss of fat, especially IAAT, was the driving force for metabolic profile-improvements in this study. Confusion exists whether exercise training has an effect on causing a preferential loss of IAAT with some studies suggesting preferential loss of IAAT (17,18,21,22) and others showing no preferential loss in IAAT (37–39). The effects of weight loss from restricting calories may be masking the independent effects of exercise on fat distribution and CVD risk factors. It is possible that the exercise stimulus was insufficient to induce an additional improvement in blood lipid change over weight loss due to diet. We feel that it would be difficult to achieve a higher intensity of training with nonathletes as the aerobic exercisers trained for 40 min at 80% of maximum heart rate and the resistance exercisers did 2 sets of 10 repetitions for 10 exercises at 80% of 1-RM. Although both the aerobic and resistance exercisers were scheduled to train three times each week, adherence was only 78% for the resistance trainers and 80% for the aerobic trainers. This resulted in the two groups averaging just over 2.3 exercise sessions/week. It certainly can not be ruled out that more frequent exercise training may have resulted in greater improvements in the blood lipid profile.
It is interesting that a number of different reports on obese polycystic ovary syndrome patients have reported similar findings to those found in more normal subjects. They have shown that visceral fat is associated with cardiovascular risk (40) and that exercise training is associated with improved fertility (41), improved cardiopulmonary function, decreased C-reactive protein, and improved insulin sensitivity indexes (42) but not TGs, TCs, or HDLs. However in a different study, the authors report significant improvements in blood lipids after 12 weeks of aerobic training that totally disappeared after 12 weeks of detraining (43). Irrespective of whether a higher exercise frequency would have added to the improvement in blood lipids, these results should not be interpreted to mean that exercise has no benefit during diet-induced weight loss. Exercise has been previously shown to have positive effects on other factors, such as insulin sensitivity (38), energy expenditure and maintenance of muscle (39), and bone density (44) during diet-induced weight loss. The average amount of time for exercise having taken place before metabolic testing was 60 h in the present study, and there may have been acute effects of exercise confounding the blood analyses, because some studies have reported temporary changes in blood lipids after 24–48 h (45–47). The present study did not control for the acute effects of exercise, and it may have been a limitation in our results. It should also be mentioned that even though we did not find an independent effect of ΔIAAT or ΔLF on the BP, other studies (48) have documented that decreases in waist circumference were related to changes in BP.
Both overweight and premenopausal AA and EA women benefitted from weight loss by decreasing IAAT and improving CVD risk. This finding occurred even in AA women who had little IAAT and relatively low CVD risk before weight loss. The changes in IAAT were significantly related to blood lipids. A loss of LF seems to be related to reduced improvement in TC and TG. Based on these results, interventions should focus on changes on IAAT.
We thank everyone who worked on this project, including program coordinator Paul A. Zuckerman, research assistants David Bryan, Amy Thomas, Paul McCarthy, the wonderful graduate assistants involved, and participants who volunteered, for their contribution to this project. This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases grants R01 DK-49779 and R01 DK-51684, DRR General Clinical Research Center grant M01 RR-00032 from the National Center for Research Resources, and Clinical Nutrition Research Unit grant P30-DK 56336. Stouffer's Lean Cuisine entrees, Nestle Food, Solon, OH and Weight Watchers Smart Ones, HJ Heinz Foods, Pittsburgh, PA kindly provided food for dietary control.
DISCLOSURE The authors declared no conflict of interest.