I. Baseline characteristics
The study population consisted of 60 obese subjects; all subjects had normal LV systolic function (i.e., LVEF>50%). There were no differences between subjects who consumed a low-carbohydrate or low-fat diet in terms of baseline characteristics (age, sex, racial composition, systolic and diastolic blood pressures [SBP and DBP, respectively], serum lipids and triglycerides, glucose, insulin, homeostasis model assessment insulin resistance or percent with metabolic syndrome). Mean age was 46±10 years, 43 (72%) female, 15 (25%) African American, BMI 37±3 kg/m2, insulin 13.7±9.9 μU/mL, and homeostasis model assessment insulin resistance of 3.1±2.5. Additional baseline characteristics of the combined group are shown in .
Serial changes in weight and hemodynamics, plasma lipids and glucose during the study.
II. Body weight response to diet intervention
The entire 24-month follow-up was completed by 47 subjects (78%). There were no differences between the two diet groups in terms of weight loss or any of the primary endpoints, namely diet-induced changes in cardiac structure/function and CIMT; therefore, the data for the two groups were combined ( and ). For the entire group, average weight loss was 7.8±4.9, 9.9±6.9, 8.4±7.6, and 4.1±8.8 kg at 3-, 6-, 12- and 24-months, respectively, representing 7.3±4.0, 9.2±5.6, 7.8±6.6 and 3.8±7.9% decrease from baseline body weight. Maximal weight loss occurred at 6-months. Although weight regain occurred at 12- and 24-months, mean body weight remained lower than baseline values. Both dietary intervention groups exhibited significant differences between baseline and some follow-up intervals in weight, heart rate, SBP, DBP, and lipid levels. Where dietary group assignment resulted in group differences, the low-carbohydrate diet showed differences in triglyceride and HDL levels whereas only HDL was significantly different in the high-carbohydrate group. Neither group demonstrated significant changes in blood glucose or in the percentage of subjects with metabolic syndrome during the follow-up interval ().
Cardiovascular structure-function response to weight loss
III. Cardiovascular structure-function response to weight loss
Compared with baseline measures, there was a significant increase (improvement) in the E’ (an index of LV diastolic relaxation) at 6, 12, and 24 months; a significant decrease (improvement) in LVM/Ht2.7 at 3, 6 and 12 months; and a significant decrease (improvement) in CIMT at 6, 12, and 24 months (). Age- and sex-adjusted mixed linear models revealed that follow-up time was significantly associated with decreases in weight (F[.05,4,58] = 42.51, p<.0001), E’ (F[.05,4,58] = 8.71, p ≤ 0.0001), LV mass (F[.05,4,58] = 5.27, p = 0.004), and carotid intima-media thickness (F[.05,4,58] = 10.18, p < 0.0001).There was a modest increase (improvement) in S’ (an index of LV contractile performance) at 24 months (F[.05,4,58] = 5.21, p = 0.001). Further post-hoc analyses divided the cohort into two equal groups by the median of S’ at baseline. In the baseline S’ < 8.0 cm/s group, a significant improvement in S’ was noted at 3, 6, 12 and 24 months compared with baseline (F[.05,4,28] = 16.75, p ≤ 0.0001); there were no significant differences over time in the baseline S’ ≥ 8.0 cm/s group.
IV. Changes in LV geometric patterns
At baseline, a majority (52%) exhibited normal geometry (); whereas at 3 months, the proportion with normal geometry and eccentric LVH both increased and the proportion with concentric remodeling and concentric LVH decreased. Further analyses show that the changes in LV geometry are related to the percentage of subjects with increased relative wall thickness (adjusted p-value = 0.03); these findings are consistent with decreased LVM/Ht2.7 at 3, 6, and 12 months.
LV Geometry and RWT at Baseline and Follow-up.