Favorable effects of low-fat and low-carbohydrate dietary patterns on serum leptin, but not adiponectin, among overweight and obese premenopausal women: a randomized trial

Epidemiologic studies suggest that the modern lifestyle of developed countries, characterized by high levels of physical inactivity and high caloric intake, leads to greater adiposity and potentially increases the risk of breast cancer (Calle et al. 2003; Byers et al. 2002). Proposed explanations for this association include involvement of the obesity-related adipokines, adiponectin and leptin (Siiteri 1987), although, among premenopausal women, specifically, the data supporting a breast cancer association for adiponectin (inverse) (Minatoya et al. 2013; Liu et al. 2013; Ye et al. 2014; Tian et al. 2007; Tworoger et al. 2007; Mantzoros et al. 2004; Miyoshi et al. 2003) and leptin (positive) (Wu et al. 2009; Liu et al. 2007; Mantzoros et al. 1999; Tessitore et al. 2000; Woo et al. 2006; Harris et al. 2011; Petridou et al. 2000; Falk et al. 2006) are inconsistent.

While among postmenopausal women there is strong evidence for a positive association between obesity and breast cancer risk (Carmichael and Bates 2004), evidence to support this association among premenopausal women is not currently available. Nonetheless, breast cancer outcomes are substantially worsened as a result of obesity irrespective of menopausal status (Carmichael and Bates 2004). Thus, while data suggest reduced breast cancer risk among obese premenopausal women, excess weight and weight gain during adulthood, particularly after the onset of menopause, ultimately increases their risk. Several modifiable behaviors (e.g., diet and physical activity [PA]), which could effectively prevent obesity and favorably modify circulating adipokine concentrations, may be paramount to both cancer prevention and improvement of outcomes following a breast cancer diagnosis.

The most effective dietary pattern for breast cancer prevention is a controversial topic and greatly debated among various disciplines in the scientific community. Although most studies have focused on low-fat diets (LFDs), few have investigated the optimal protein and carbohydrate intake for breast cancer prevention (Sieri et al. 2002; Toniolo et al. 1989; Martin-Moreno et al. 1994). Proponents of low carbohydrate diets (LCDs) report that high carbohydrate intake results in higher plasma insulin levels and promotes lipogenesis (Bilsborough and Crowe 2003); hence the popularity of low-carbohydrate diets (e.g., the Atkins diet) among the general population. Some short-term studies (e.g., ≤6 months) of LCDs have reported greater loss of body fat and greater maintenance of lean body mass when compared with diets high in carbohydrates (Farnsworth et al. 2003; Foster et al. 2003). These studies have been relatively short-term, and, therefore, the effect on long-term weight maintenance is unknown.

In this study, we examined the effects of two calorie-restricted diets (LFD and LCD) plus PA on serum concentrations of adiponectin, leptin and the adiponectin-to-leptin (A/L) ratio in overweight and obese premenopausal women. We hypothesized that diets low in fat and carbohydrates would favorably improve serum adipokine profiles, specifically by increasing adiponectin and decreasing leptin concentrations, as a result of diet-induced reductions in anthropometric measures (e.g., weight and body mass index [BMI]).

Premenopausal women were recruited for a 52-week randomized intervention trial of two calorie-restricted dietary patterns plus PA. The diets were: 1) LFD: 20% of total calories from fat, 20% from protein and 60% from carbohydrates); and 2) LCD: 40% of total calories from carbohydrates, 30% from protein and 30% from fat). All study participants were educated on caloric-restriction and given a personalized PA prescription. Women were randomized using stratified randomization, based on BMI (<30 kg/m² vs. ≥30 kg/m²) to ensure equal distribution of overweight and obese participants into each intervention group. This study received ethical approval from the Cancer Institutional Review Board at the Ohio State University.

Screening and recruitment occurred between May 2005 and August 2006, at primary care physician’s offices and through media advertisements. Potential participants called a designated telephone number and a staff member from the Comprehensive Cancer Center provided information about the study and conducted eligibility screening. Women who: were ≥30 years and premenopausal (confirmed by follicle stimulating hormone concentrations); had no prior cancer diagnosis; BMI 25–34 kg/m²; were residents of the Columbus, OH area during the 18 month follow-up period; and obtained medical clearance for participation in PA (from their primary care physician), were eligible to participate. Women who: were pregnant or planning to become pregnant; were enrolled in a structured weight loss program; had medical condition(s) precluding adherence to the dietary interventions; and/or had uncontrolled existing medical conditions, were ineligible from study participation. All participants provided written informed consent prior to study enrollment.

Currently, overweight and obese premenopausal women, who have an increased likelihood of sustaining higher body weight after menopause, lack effective, practical methods for reducing their breast cancer risk. Attractive strategies for risk reduction may include lifestyle modifications. In the present study, we hypothesized that LFD and LCD dietary patterns could favorably modify serum adipokine concentrations (i.e., increase adiponectin, decrease leptin, and increase the A/L ratio) through weight reduction, and therefore, promote the prevention of obesity-related breast cancer among premenopausal women. Our findings demonstrated that, overall the LFD and LCD interventions yielded significant reductions in serum leptin, while the modifications of adiponectin and the A/L ratio were statistically non-significant. Notably, among women who were adherent to the LCD intervention, adipokine profiles improved in the directions that are hypothesized to be protective.

Several dietary trials have evaluated the effects of LFD and/or LCD interventions on adipokine concentrations (de Luis et al. 2007; Bluher et al. 2012; Befort et al. 2012; Scott et al. 2013; Giannopoulou et al. 2005; Friedenreich et al. 2011; Sacks et al. 2009; Davis et al. 2009; Foster et al. 2003; Samaha et al. 2003; Brehm et al. 2003; Yancy et al. 2004; Shai et al. 2008; Bradley et al. 2009; Reed et al. 2010; Ata et al. 2010; Ong et al. 2009; Harvie et al. 2011), but few have focused on premenopausal women (Reed et al. 2010; Ata et al. 2010; Ong et al. 2009; Harvie et al. 2011). In our study, overweight and obese premenopausal women were randomized to a LFD or a LCD plus PA for a period of 52 weeks. While adherence to these interventions was fairly low, there tended to be larger decreases in serum leptin among diet-adherent participants. One study (Reed et al. 2010) examining the effects of a 4-month calorie-restricted diet of similar composition to the LFD examined herein, combined with aerobic exercise among premenopausal women, demonstrated no significant change in circulating adiponectin and a 52% decrease in leptin among normal weight to obese premenopausal women. They concluded that the substantial reduction in leptin was exercise-induced (Reed et al. 2010) and due to significant fat mass reductions over the course of their intervention trial. Our findings support this hypothesis given that we too observed significant reductions in leptin (approximately 30%), particularly among PA-adherent participants. However, the smaller reduction in leptin observed in our study could be attributable to the lower intensity of the walking-based PA to which women were prescribed, translating into smaller reductions in fat mass (which we did not examine). Another study (Ata et al. 2010), examining the effects of a 10-week LCD (with the same macronutrient composition as the LCD studied herein) plus a walking intervention among overweight and obese premenopausal women, demonstrated 10% higher adiponectin and 6% lower leptin post-intervention. A recent study (Harvie et al. 2011) examined the effects of a 6-month trial of intermittent and continuous calorie-restricted diets without exercise among premenopausal women and found that these diets yielded significant reduced leptin (approximately 40%) and increased the A/L ratio (approximately 20%). Another study (Ong et al. 2009) examined the effects of a 1-month calorie-restricted intervention (62% carbohydrates, 26% protein, and 12% fat), without exercise, on adipokine concentrations and adipokine expression profiles among overweight and obese premenopausal women. They found that post-intervention adiponectin and leptin concentrations were 3% and 60% lower, respectively (Ong et al. 2009). Interestingly, this study (Ong et al. 2009) also identified several genes involved in metabolic pathways (e.g., glycolysis and lipid synthesis) with altered gene expression in breast and adipose tissues post-intervention. These findings (Ong et al. 2009), as well as findings from our previous work (Llanos et al. 2012), highlight the importance of understanding adipokine concentrations within local breast tissues and their role in breast cancer risk and progression.

While the precise mechanism(s) involved in the potentially protective effects of caloric-restriction are unclear, several studies have concluded that dietary interventions of this type are effective for weight loss, particularly when they are combined with exercise (Ata et al. 2010; Befort et al. 2012; Reed et al. 2010; Scott et al. 2013), and therefore, would be effective for obesity-related cancer prevention (Harvie et al. 2005; Kawai et al. 2010; Trentham-Dietz et al. 2000), as well as for improving breast cancer outcomes (Scott et al. 2013; Befort et al. 2012). Our findings of favorable modification of adipokine profiles, particularly leptin, among women who were adherent to the intervention, support this hypothesis. Previous studies tend to show a positive association between circulating leptin and breast cancer (Minatoya et al. 2013; Liu et al. 2013; Ye et al. 2014; Tian et al. 2007; Tworoger et al. 2007; Mantzoros et al. 2004; Miyoshi et al. 2003), although the data for premenopausal women have varied. Additional studies have provided evidence to support a role for an inverse association between the A/L ratio and breast cancer risk and prognosis (Chen et al. 2006; Goodwin et al. 2012), suggesting the balance of adiponectin and leptin, rather than either biomarker alone, may also be clinically significant. Further study to help explain the underlying mechanisms that may be responsible for the beneficial adipokine effects of calorie-restricted interventions are necessary.

Notably, we observed stronger percent change in leptin, compared to those in adiponectin and the A/L ratio, following the intervention period. Similarly, findings from other dietary intervention trials have suggested that calorie restricted diets, particularly the LCD variety, in combination with exercise, produce rapid weight loss (Ata et al. 2010; Befort et al. 2012; Reed et al. 2010; Scott et al. 2013), which results in larger reductions in circulating leptin concentrations. This is suggestive that the effects of the macronutrient composition may be dependent on the extent of weight loss induced by the dietary pattern. Consequently, reductions in circulating leptin, with or without maintained weight loss, could elicit considerable breast cancer preventive effects. Leptin promotes cellular growth and proliferation, induces signaling pathways involved in survival of breast epithelial cells (Laud et al. 2002; Dieudonne et al. 2002), and promotes angiogenesis (Cao et al. 2001). Furthermore, assuming the reported associations between circulating leptin concentrations and breast cancer risk and adverse events in breast cancer survivors (Goodwin et al. 2012) are accurate, interventions targeting leptin may be promising for breast cancer prevention and control efforts in the future. While our findings support the hypothesis that caloric-restriction may be an attractive method for obesity-related cancer prevention through altering the leptin signaling pathway (and possibly through mechanisms not studied here), the clinical significance of the favorable changes observed herein is unclear. Future studies involving larger samples are required to confirm the effects in the general population.

There were various strengths of this study. First, the randomized nature of the dietary intervention trial and the addition of a PA component strengthened the study. In addition, most studies have examined the effects of PA and LFD and/or LCD interventions for a period of six months or shorter (de Luis et al. 2007; Befort et al. 2012; Scott et al. 2013; Giannopoulou et al. 2005; Yancy et al. 2004; Reed et al. 2010; Ata et al. 2010; Ong et al. 2009), whereas our study period was one year. Our examination of the intervention effects on weight loss as well as biomarkers previously shown to be associated with obesity and breast cancer risk was also a strength. Relatedly, our use of highly reproducible immunoassays to measure serum adipokine concentrations as well as their measurement at several time points over the course of the study was an additional strength. Furthermore, our use of mixed models for the analysis of treatment and time effects (under the assumption that missing data were random) allowed us to use all of the available data, further strengthening our findings.

This study also suffered from certain limitations that should be considered. The small sample size (41 and 38 women randomized to LFD and LCD, respectively), which included only overweight and obese premenopausal women, limited our power to observe differences in biomarker concentrations between diets as well as limited the generalizability of our findings. Secondly, the relatively high losses to follow-up and low rates of adherence to the LFD and LCD diets may have contributed to our finding small differences in adipokine concentrations between the interventions. While rates of adherence to the PA component of the intervention were higher, likely as a result of participants wearing pedometers, this too could have tempered our findings given that PA was self-monitored. Similarly, the use of self-reported dietary consumption through 7-day diet recalls may be have been a limitation given that complete data were often not collected. Additionally, it is likely that participants’ dietary patterns may have been altered and/or reported inaccurately due to the burden of recalling their consumption and possibly due to social influence. Finally, measurement of only total adiponectin, as opposed to other isoforms, which could have been modified by the interventions, may have affected our findings.

In summary, findings from this small randomized trial support LCD and LFD dietary interventions as potentially attractive methods for obesity-related breast cancer prevention, particularly through favorable modification of leptin among premenopausal women. Development of effective and practical interventions involving sustainable dietary and PA changes that have significant, favorable effects on adiponectin, leptin and the A/L ratio (and other related biomarkers) would be important next steps in this line of research.