San Antonio ? The first experiment ever to show that low-fat diets could help prevent a return of breast cancer now reveals, with longer follow-up, that the benefit was almost exclusively to women whose tumor growth was not driven by hormones.

That could be huge – the new results suggest but cannot prove that these women might be able to cut their risk of dying by up to 66 percent with such diets.

“That’s as great or better than any treatment intervention that we’ve given” for this type of cancer, which is notoriously hard to treat, said Dr. C. Kent Osborne of Baylor College of Medicine in Houston, who had no role in the study.

However, for women whose cancers are fueled by hormones – the vast majority of breast cancer patients – the diet change seemed to make little difference in the risk of recurrence or survival. Questions remained about whether those who did benefit truly were helped by cutting fat or by the weight loss that resulted.

“Maybe it raises as many issues as it answers,” said John Milner, chief of nutrition science research for the National Cancer Institute, which paid for the first phase of the study.

Initial findings from the study were reported at a cancer conference in 2005 and will appear in this week’s Journal of the National Cancer Institute. Updated results with longer follow-up on many of the original participants were presented Saturday at the San Antonio Breast Cancer Symposium.

The mixed results were a surprise because doctors had expected all women to benefit, said Dr. Rowan Chlebowski of the University of California at Los Angeles, who led the work.

Hormones might play such a strong role in some cancers that dietary changes have only weak impact on future risk, experts said.

The study involved 2,437 women with early stage breast cancer, average age 58, at 39 sites around the country. All had surgery followed by chemotherapy and five years of tamoxifen if their tumors were hormone-fueled.

At the start of the study, 29 percent of their calories came from fat – 10 percent to 12 percent lower than the typical American diet. Doctors told 1,462 of them to continue their normal diets. The other 975 had counseling with dietitians to cut fat to around 20 percent of daily calories.

The diet group averaged 33.3 grams of fat a day compared with 51.3 grams for the others and lost five to six pounds during the study.

Five years later, cancer had returned in 9.8 percent of the diet group and 12.4 percent of those on standard diets, which translated to a modest 24 percent lower risk for the group as a whole.

But the result barely reached statistical significance, meaning that the difference almost could have occurred by chance alone. The new results, with longer followup, put the difference for the overall group at 21 percent and even weaker statistically.

Researchers’ ability to study the women beyond the first five years was hampered by the federal grant running out. Two charities – the Breast Cancer Research Foundation and the American Institute for Cancer Research – gave money so they could resume.

Ten of the original 39 study sites have provided complete information on their participants for an additional two years, and information on deaths is available for all women in the study, Chlebowski said.

The new results: 14 percent of women on low-fat diets and 17 percent of the others have had a recurrence or second cancer. About 8 percent of dieters and 10 percent of the others have died.

However, there was a huge difference in the subgroup of women whose tumors were fueled by neither estrogen nor progesterone. Only 6 percent on low-fat diets died compared with 17 percent of the others. That translated to 66 percent lower risk of death for those who trimmed fat.