Dyslipidemia characterized by elevated TG, reduced HDL cholesterol and small dense LDL (pattern B) is closely linked to obesity (17
), and weight loss is associated with improvements in these lipid abnormalities (20
). Previous studies have demonstrated the ability to substantially reverse pattern B with weight loss in overweight and obese men (BMI from 26 to 35) (10
). The goal of the current study was to determine the extent to which reversal of pattern B could be achieved through the normalization of BMI <25 and percentage body fat <20%. In the majority of pattern B men, weight loss of ~8kg was associated with conversion to pattern A, and the reversal of all components of the ALP. Weight loss improved insulin sensitivity to a similar extent irrespective of LDL subclass pattern conversion status, indicating that conversion did not appear to be mediated by improved insulin sensitivity. Notably, men who were pattern B at baseline were less likely to achieve study weight loss goals compared to men who were pattern A at baseline. This disparity was not due to differences in physical activity as measured by steps per day. Because the current protocol did not enable us to determine whether the decreased propensity towards weight loss in pattern B vs. pattern A men was due to lack of compliance or underlying metabolic differences, further investigation will be required to test the latter possibility.
Nonetheless, this study supports the concept that despite evidence for genetic influences on atherogenic dyslipidemia and LDL subclass pattern B (18
), these traits can be reversed in a high proportion of individuals by weight loss and normalization of adiposity in the context of moderate levels of carbohydrate in the diet. Furthermore, pattern B men had a differential response to weight loss with regard to LDL subclass distribution compared to pattern A men. With weight loss, the reversal of ALP in pattern B men was associated with a greater decrease in small, dense LDL-IIIb and a concomitant increase in larger, more buoyant LDL-Iia compared to pattern A men.
Changes in LDL subclass distributions with dietary intervention are strongly correlated with changes in both triglyceride-rich lipoproteins and HDL. In the present study, regression analyses indicated that both reductions in triglyceride and increases in HDL-C associated with increases in LDL peak diameter with weight loss independently of weight, percentage body fat and plasma insulin. Moreover, increases in HDL-C were associated with conversion from pattern B to A and changes in triglyceride were associated with changes in levels of LDL subclasses.
This study further demonstrated that weight loss in pattern B men resulted in lipoprotein and biochemical measurements that were comparable to those of pattern A men at baseline who did not undergo weight loss. Notably, the BMI required to elicit LDL subclass pattern A was significantly lower in pattern B men. Similarly, lower percentages of body fat were required to achieve pattern A in men who were pattern B at baseline. These data suggest a fundamental metabolic disparity, such that pattern B men may be more resistant to normalizing their plasma concentrations of lipids and lipoproteins at a given body weight and level of adiposity. However, BMI is only an indirect measure of what are thought to be important fat depots that can influence ALP (22
), and total body fat measurements do not provide enough detail to discern these metabolically active tissues. Despite the fact that the pattern A men in the control group were heavier and had increased total body fat relative to pattern B men who underwent weight loss, they may be metabolically more lean, i.e., they may have lower visceral or hepatic fat mass. Differences in such metabolically active tissues may explain their relative protection from ALP (23
During the weight-stable period following weight loss, the men were still in caloric deficit relative to their average consumption at the beginning of the study (estimated reductions of 895 ± 457 and 1,029 ± 427 kcal/day for pattern A and pattern B men, respectively). Interestingly, achieved weight loss may lead to lipoprotein profiles that are improved relative to persons at comparable BMIs who have not lost weight. There was an 81% conversion of pattern B men to pattern A in those men who reached the BMI target of <25 in this study. Assuming a pre-weight loss population prevalence of pattern B of 29% in men with BMIs between 25 and 28 (based on cross-sectional data in 229 men; R.M. Krauss, unpublished data), we would predict that the prevalence of pattern B after weight loss would be ~6%, an estimate that is much lower than the prevalence of pattern B in nonweight reduced men (~18%; R.M. Krauss, unpublished data). Improved lipoprotein profiles in the weight-reduced compared to the nonweight reduced state are not without precedent. Williams has previously shown higher HDL cholesterol concentrations in male runners who had lost weight than in runners with similar BMIs who had not lost weight (25
). These findings speak to the potential metabolic benefits of weight reduction over baseline weight status.
Some pattern B men at baseline did not convert to pattern A, and as this may have been due to decreased compliance (which was not directly measured in this study), the inability to convert may be related to underlying metabolic or physiological characteristics. Logistic regression models suggested that changes in HDL cholesterol significantly predicted conversion patterns. Furthermore, pattern B men who remained pattern B after weight loss had greater body fat and trunk fat, higher TG and smaller LDL particle diameters at baseline, with expected shifts in LDL subclass distribution (data not shown). When only the subset of men who reached the BMI<25 goal was considered, the one distinguishing characteristic that differentiated the B→BRx from the B→ARx group was baseline particle distribution with a significantly larger LDL peak particle diameter in the converters (data not shown). These data suggest that peak LDL particle size may be predictive of adaptability to environmental modulators of body weight.
Our previous studies have shown that dietary carbohydrate plays a more important role than dietary fat in influencing the expression of pattern B (10
). Other investigators have also shown an improvement in markers of cardiovascular disease risk with low to moderate intakes vs. high-carbohydrate intakes (27
). In the present study, dietary carbohydrate was maintained at 40% of energy (with relatively high concentrations of total and saturated fat) with parallel measurements in the control and weight loss groups to control for variability in response to the experimental diet. Pattern B men in the control group had unexpected, but significant increases in TG over the course of the study. However, all men exhibited a significant decrease in TG between the screening and baseline visits (data not shown), and by the end of the 4-week weight-stable period in the control arm, TG concentrations had increased in all study participants, with a greater increase observed in pattern B men whose plasma TG reached pre-study concentrations. The changes in plasma TG in the control arm were, therefore, probably related to deviations from the participants' usual (pre-study) diets. There were several caveats to this study. Because of its relatively short-term nature, the results may not be applicable to longer-term dietary interventions. A related issue was the shorter intervention time in the control relative to the weight loss group (4 vs. 9 weeks), which was a design feature based on the potential difficulty of maintaining stable weight over long periods of time. Furthermore, because this study was conducted in a population of predominantly middle-aged white men, extrapolation of the results to women and other ethnic groups may be limited. Women have a significantly lower prevalence of pattern B, even at higher BMIs (30
) so weight loss may not be as important a contributor to ALP in women. Furthermore, other ethnic groups, particularly Asians, have been shown to exhibit dyslipidemia at BMI values lower than would be expected among Whites (31
Nonetheless, the findings of this study indicate that the achievement of body mass indices in the currently recommended “normal” range is sufficient to reverse ALP in a substantial proportion of the population and further reduction may be advised to ensure maximal benefit to this group of individuals.