Study suggests that unintentional weight loss is a signal to see a doctor

Unintentional weight loss is associated with an increase in the risk of a cancer diagnosis within the coming year, according to a study from Dana-Farber Cancer Institute.

"If you are losing weight and you aren't trying to lose weight by making changes in your exercise routine or diet, people should see their doctor to consider possible causes," says lead investigator Brian Wolpin, MD, MPH, Director of the Gastrointestinal Cancer Center at Dana-Farber and Director of the Hale Family Center for Pancreatic Cancer Research.

"There are many conditions that can result in unexpected weight loss. Your doctor can determine if there is something that needs evaluation."

The findings were published in the JAMA on January 23, 2024.

Compared with participants who did not lose weight, recent weight loss was associated with significantly increased risk for several types of cancer, including upper gastrointestinal tract (including esophageal, stomach, liver, biliary tract, and pancreatic cancer), hematological (including non-Hodgkin lymphoma, multiple myeloma, and leukemia), colorectal, and lung cancers. However, recent weight loss wasn't found to be associated with increased risk for other cancer types, such as breast cancer, genitourinary cancer, brain cancer, or melanoma.

"Unexpected weight loss can come from cancer or many other conditions," says Wolpin. "Sometimes weight loss is due to more exercise or a healthier diet, and this can be beneficial to people's health. However, when a patient experiences unintentional weight loss not due to healthier behaviors, seeing your primary care doctor is appropriate, so they can determine whether additional evaluation is necessary for other causes of weight loss, including cancer."

The study assessed 157,474 participants in two large longitudinal studies: The Nurses' Health Study, which enrolled nurses aged 30 to 55 starting in 1976, and the Health Professionals Follow-Up Study, which enrolled male health professionals aged 40 to 75 years starting in 1986. Participants were followed until 2016.

Weight was reported by participants every other year in a biennial questionnaire that also included questions about physical activity. The questionnaire requested responses about dietary changes every four years. This information enabled Wolpin and colleagues to assess each participant's level of weight loss promoting behaviors. Weight loss promoting behaviors were classified into "high" for those making both dietary improvements and increases to physical activity, "medium" if they made only one change, and "low" if they made no changes to diet and exercise.

"We wanted to differentiate healthy weight loss from unhealthy weight loss," says Qiaoli Wang, MD, Ph.D., a research fellow at Dana-Farber and the manuscript's first author. "Healthy weight loss can come from dietary changes or increased exercise. But unhealthy weight loss that occurs unexpectedly can be due to an underlying cancer."

Patients with advanced cancer often lose weight, but weight loss is often not thought to occur with early-stage disease. This study found that similar levels of weight loss occurred before diagnosis of both early and late-stage disease. This is important because unintentional weight loss could be a sign of a developing cancer that could help diagnose the cancer earlier when there's a chance for more effective treatment.

The mechanisms by which cancer results in weight loss vary depending on the type of cancer. This study strengthens findings from past research that connected unexpected weight loss with an increased cancer risk.

In previous research, weight data was collected by doctors from patients potentially seeking out care for an illness. In this study, weight data was collected prospectively and regularly for decades and was not dependent on doctor visits to identify weight changes. This study also considered all types of cancers. However, the two studies analyzed were focused on health professionals, which is not a group that is fully representative of the U.S. population.