Tuma, Rabiya S. PhD
WASHINGTON—The trend from observational studies is clear: Exercise reduces the risk of certain types of cancer, including breast, prostate, and colorectal. In the case of breast cancer, investigators hypothesize that changes in estrogen and estrogen metabolism may be a mechanistic link between physical activity and reduction in risk. However, the data thus far, presented in part here at the American Association for Cancer Research Annual Meeting and in a recent journal article, are mixed, including data from two randomized intervention trials—the Women In Steady Exercise Research (WISER) and WISER Sister studies.
WISER Study Links Metabolites & Exercise
To test the impact of regular exercise on estrogen and estrogen metabolism, Mindy S. Kurzer, PhD, Professor of Nutrition and Medicine and the Director of the Healthy Foods, Healthy Lives Institute at the University of Minnesota in St. Paul, and colleagues enrolled 391 healthy women, aged 18 to 30, in the WISER study. Participants were randomly assigned to 30 minutes of moderate to vigorous aerobic exercise five days a week or to continuation of their regular sedentary lifestyle for 16 weeks.
As described in the report in the May issue of Cancer Epide-miology, Biomarkers and Prevention (2013;22:756-764)—first author is Alma J. Smith—trainers supervised some of the exercise sessions for each of the women in the intervention arm and monitored their activity through exercise logs. The investigators measured estrogens (estrone, estradiol, and estriol) and estrogen metabolites in the women's urine collected during three consecutive 24-hour periods in the women's baseline and post-intervention menstrual cycles (follicular days 7 through 9).
As expected, significant improvements were found in aerobic fitness, lean body mass, and body fat percentage in the exercisers over the course of the study. In contrast, the participants in the control arm showed no alteration in these measures, despite a significant reduction in daily calorie intake of about 224 kcal/day.
There was no change in two of three measures of estrogen (estradiol, and estriol) between baseline and follow-up in either group. There was a significant decline in estrone in the controls over the study period, with the mean change being statistically different between the intervention and control arms. Perhaps more important, there was a significant increase in the ratio of key estrogen metabolites (2-OHE1/16α-OHE1) in the 165 participants (92%) in the exercise arm who completed the trial.
The change in the ratio from baseline to follow-up was significantly different between the intervention arms, the investigators reported. Prior research suggests that an increase in this metabolite ratio may be associated with a reduction in breast cancer risk.
“Our conclusions are that yes, there was a change in the metabolism of estrogen in the exercise group that was consistent with an association with lowered breast cancer risk,” Kurzer said in a telephone interview. “Therefore an alteration in estrogen metabolism may be one of the mechanisms by which exercise reduces breast cancer risk.”
No Link Seen in High-Risk Women
In contrast to the results reported by Kurzer et al, though, collaborators from the University of Pennsylvania Perelman School of Medicine in Philadelphia presented a study at the AACR Annual Meeting that a first-pass analysis of a similar study in high-risk women showed no association between exercise and estrogen.
In the WISER Sister study, Kathryn Schmitz, PhD, MPH, Associate Professor of Biostatistics and Epidemiology at the University of Pennsylvania, and colleagues (including Kurzer) randomly assigned 139 women at high risk of breast cancer to no exercise, to 150 minutes of moderate-to-vigorous exercise per week, or to 300 minutes of exercise per week (the abstract was one of four included in a major symposium on obesity and cancer but no details are available online).
The participants were older than those in the WISER study (median age of 33.1 to 35.1 in the three arms compared with 25 in WISER). Additionally, the WISER Sister participants had a higher than average risk of breast cancer, based on one of three measures: having a BRCA mutation or a greater than 25 percent probability of BRCA mutation based on a family member's status; Gail model risk calculator (median of 20-25 in the three arms); or Claus model risk calculator (median of 24 to 25 in the three arms).
All of the participants in the study received an in-home treadmill for the trial, though the women assigned to no exercise received theirs after completion of the study. The researchers used objective monitoring tools to ensure that those assigned to the exercise interventions adhered to the schedule, including heart rate monitors, logs, emails, and texts. “This was not a self-reported study,” Schmitz emphasized.
Adherence was good, with women in the low-dose group completing a median of 147 minutes per week and those in the high-dose group completing 244 minutes per week in the six months study. (There was a ramp-up time to allow women in the intervention groups to get to the target exercise range.)
Moreover, the team found that the women exercised in the target heart rate zone for much of their workout time. “They weren't just getting on the treadmill; they were doing it in the range we were asking them to,” she said. Overall 39 women completed the study in the high-dose group, as did 38 in the low-dose group and 47 in the control group. (Schmitz noted that even though there was a higher dropout rate in the exercise groups the completion rate was within the acceptable range for this type of intervention study.)
For the hormone analysis, the team collected first morning urine samples every day during cycles 1 and 2 prior to the exercise innervation and again during cycles 6 and 7 during the intervention.
Despite good participation and adherence, and a significant dose-dependent improvement in aerobic fitness, there was no significant alteration in estrogen: “I'm sorry to say that every single one of our analyses in estrogen and progesterone were null—every single one,” Schmitz said. “We were so disappointed.”
She said she is not sure why the results were different than expected. In a pilot study the group saw a 21 percent decrease in estrogen (area under the curve) for the total cycle between pre- and post-training study periods, with the biggest difference in the luteal phase. However, in the full study, there was a much larger standard deviation in estrogen levels than in the pilot study, and there was a lot of variability within women from cycle to cycle.
There are more analyses to be done, she said. “One hundred and thirty-nine women were randomized, and we have four cycles on 122 women, so that is a lot of data to analyze.”
Metabolic Differences Between Active and Inactive Women
Meanwhile, investigators from the National Cancer Institute reported at the AACR meeting that they find an association between overall activity and estrogen levels (estrone and estradiol) and estrogen metabolites in a large cohort of Polish women (Abstract 2519).
In the study, 540 women age 40 to 75 wore accelerometers for seven days. The researchers, led by Cher Dallal, MPH, PhD, a postdoctoral fellow in the NCI's Hormonal and Reproductive Epidemiology Branch, measured estrogens and estrogen metabolites in 12-hour urine collections.
When the women were divided into tertiles based on physical activity, there were significantly lower levels of estrone and estrogen in the most active women compared with the least active after adjusting for age and body mass index.
“We also saw that with increased activity there was increased metabolism across all three estrogen metabolism pathways,” she told OT. “We did see some association with specific metabolites, but really, in terms of what all this means it is hard to know; this is the first step and this is the first time we have looked at these metabolites.”
Estrogen May Be One Mechanism
Both Dallal and Kurzer emphasize that estrogen is likely to be only one mechanistic link between exercise and reduction in risk of cancer. Dallal and colleagues, for example, want to look at inflammatory biomarkers in the Polish cohort.
“This is extremely complicated,” Kurzer said. “We are adding a piece of information to a very, very complicated puzzle. In the end there are probably many factors that go in many directions that we would have to try to synthesize in some way to understand the overall effect, to understand how important estrogen is, relative to other pieces.”
When asked why the WISER and WISER Sister studies don't appear to line up, she said that an initial analysis of serum estrogen levels in WISER participants also showed no correlation with exercise, similar to what was just reported in the WISER Sister study. Therefore, she said, it's worth watching the WISER Sister study for further results—in particular, for the analysis of estrogen metabolites and exercise.