Cedars-Sinai Blog

Surprising Study Finds PCOS Does Not Cause Diabetes, Heart Disease or Stroke

A nurse comforting a patient with a PCOS diagnosis.

The medical community takes for granted that polycystic ovary syndrome (PCOS) is a causal factor for diabetes, due to epidemiological studies that draw a link between the two diseases. A new genetic study from Cedars-Sinai, however, strongly suggests the opposite.

Rethinking previous causal assertions

Mark Goodarzi, MD, PhD

"We were quite surprised when we found no relationship—PCOS is not a causal factor for diabetes, heart disease or stroke," says Mark Goodarzi, MD, PhD, the Eris M. Field Chair in Diabetes Research and lead investigator and director, Division of Endocrinology, Diabetes and Metabolism, Department of Medicine. "We presented this data to an annual society meeting in 2019, and when I unveiled the findings, the room was thunderstruck. After a few moments of silence, disbelief set in. Physicians asked various technical questions about the research. They couldn't believe that the data was accurate. But it is. And it's a very exciting negative result: It has important implications for the way we care for and counsel patients." 

PCOS, widely recognized as the most common hormonal disorder in women of reproductive age, has often been linked by epidemiologic studies to increased risk of diabetes, heart disease and stroke. It is true that women with PCOS tend to have risk factors such as abnormal cholesterol levels or are overweight or obese, and it has become accepted in the field of endocrinology that a diagnosis of PCOS means a higher risk of diabetes. In fact, it is standard care for providers to counsel PCOS patients about the increased risk of diabetes and preventive measures patients should take. Though not as strong, data linking PCOS and heart disease or PCOS and stroke has also been a concern.

"The problem is, when you have two associated conditions, it's just an association," explains Goodarzi. "You don't know whether it's actually a causal association." Goodarzi embarked on this study in early 2019 to determine, definitively, whether the relationship between the two diseases was a causal one. The study is now in press. 

Genetic technique offers high level of accuracy

Goodarzi and his team did a causality analysis, examining the relationship between PCOS and diabetes using the Mendelian randomization (MR) technique to adjust data for genetic factors. MR is a technique for using genetic variants to analyze causality and disease. 

"When I first learned about the MR technique, as a geneticist I found it quite exciting," says Goodarzi. "MR gives the investigator a means to use genetics to learn about biology even if the way the genetic variants work in the cell is unclear. If the researcher knows that the variant is a strong risk factor for the exposure, the technique works." 

In a classic randomized clinical trial, the investigator randomizes half of the subjects to intervention—a drug therapy or otherwise—and half of the subjects to placebo, and then follows those subjects for a couple of years to see whether the intervention effected any change. In MR, the randomization reflects the random assortment of genetic variants an individual receives from their parents. The key benefit is that this exposure, the genetic variants, is lifelong, as opposed to a clinical trial which may last four or five years. In Goodarzi's study, for example, PCOS is the exposure and diabetes, heart disease and stroke are the outcomes. Using MR, Goodarzi assigned the particular variants for PCOS as the exposure, obtained data from large, previously published studies and then used analysis to obtain a causal estimate for each outcome.

This type of genetic analysis has been done on causal relationships in many different fields. For example, obesity factors into both diabetes and heart disease—and has been proven to have a causal relationship by this type of study with MR. The medical community now takes that causal relationship for granted.

As it turns out, there are key features of diabetes that are also present in patients with PCOS. The first is obesity. Data indicates that between 30% and 80% of women with PCOS are also overweight or obese. Obesity is well established, with epidemiologic and genetic evidence, as a causal factor for diabetes and heart disease. 

Prior to this research, Goodarzi and his team had published a paper using MR to study the causal relationship between PCOS and obesity. The research found that obesity—measured by BMI—was a causal factor for PCOS. However, the reverse is not true: PCOS does not cause obesity. This represents an important finding for the field of endocrinology, which often grapples with obesity and its causal relationships with various other conditions. 

"If BMI causes heart disease and diabetes, and BMI causes PCOS, that explains the association between PCOS and diabetes observed by physicians," explains Goodarzi.  

Additionally, there is published genetic evidence that indicates that elevated testosterone in women is a causal factor for both PCOS and diabetes. The same is true for low levels of a blood protein called SHBG—low levels of SHBG is a causal factor for both diseases. 

Surprising findings improve standard of care

The exciting implication of this information, and in general with research like Goodarzi's, is the role it will play in tailoring medical counseling and treatment to individual patients: precision medicine. The most powerful measure for diabetes prevention is lifestyle modification— exercising and losing weight has shown a 60% risk reduction in individuals with prediabetes. Currently, it is considered best practice to counsel all PCOS patients on lifestyle changes to prevent diabetes regardless of presence or absence of these other individual risk factors. 

"We now know that several external factors mediate the observed relationship. This has important clinical implications. Previously, we'd counsel any woman who has PCOS on the risk factors and prevention measures for diabetes. That's general medicine," says Goodarzi. "Now, we don't need to frighten all patients with diabetes counseling. We can tailor our conversation to women who are obese or have elevated testosterone or low SHBG. That's precision medicine."

Physicians can now direct lifestyle modification counseling to patients with a high BMI or other diabetes risk factors. They can choose and prescribe specific therapies—like a prescription for metformin, for example—to patients who will see a measurable benefit from those therapies. And diabetes prevention in PCOS patients is just the beginning.

Goodarzi and his team are now looking at additional genetic studies to identify other risks for patients with PCOS. Breast cancer is an important one—a causal relationship, like diabetes, in which the epidemiologic studies don't align with the genetics research. Genetic analysis shows an increased risk for breast cancer in patients with PCOS. 

"This has important implications—right now we don't even discuss breast cancer with our PCOS patients," says Goodarzi. "We need to address this."