Old, poor and Hispanic women are all more likely to have unnecessary breast cancer surgery despite 2005 recommendations for gentler treatment, according to a US study.

Based on a California state cancer registry, researchers whose findings were published in the Archives of Surgery found that more than a third of some 18,000 women who had undergone a mastectomy for early-stage breast cancer had had lymph nodes under the armpits removed as well.

Yet for these women, their cancer had not yet spread beyond the breast.

"The women that were getting this unnecessary surgery were more likely to be of lower socioeconomic status, were older, or were Hispanic," said Sharon Lum, a breast surgeon at Loma Linda University in California, who worked on the study.

In 2005, guidelines recommended surgery that spares most of the lymph nodes, avoiding side effects such as pain, swelling and numbness later on.

In addition, since the mid-2000s, the American Society of Clinical Oncology has urged doctors to consider removing only the lymph nodes closest to the tumor to see if the cancer had spread beyond the breast.

Even if it had, a study earlier this year suggested that removing the lymph nodes does not help women live longer, as long as they are getting chemotherapy and radiation.

Lum said it was hard to figure out why the women receiving the unnecessary surgery were from those particular groups from the data. The gentler surgery isn't more expensive, but it does require more coordination between different departments, adding an extra step that some surgeons might prefer to avoid, Lum added.

"There are surgeons that choose not to do it, and there are patients that don't know it's better," she said.

"The take-home message for patients and surgeons is, you have to be educated about the downstream consequences of the surgery you choose to do."

In an editorial published along with the study, Jan Wong of East Carolina University in Greensville, North Carolina, said the study raises troubling questions.

"Do these data lend further evidence to the difficult of tranlsation evidence-based recommendations of care into clinical practice? If so, what are these reasons for this and how might this be corrected?" he wrote.

"Might these data suggest the selective management of women with early breast cancer based on the level of insurance coverage?"

He added that the study showed that health care systems need to more carefully examine the process of care and intervene with corrective action when needed.