If you don’t eat enough to support your training, you’ll trigger an array of health problems including weakened bones, which will in turn make you susceptible to stress fractures.

Everyone know that – in theory, at least. But it can be difficult to convince motivated runners that there’s a direct cause-and-effect relationship between eating enough and the success of their subsequent training and health. So some new data, published in the American Journal of Sports Medicine by a research team led by Michael Fredericson at Stanford University with lead author Adam Tenforde of the Spaulding National Running Center at Harvard Medical School, may serve as a useful wake-up call.

The study tracked 323 women athletes in 16 different sports, including 47 cross-country runners. Using questionnaires and bone-density scans, the athletes were assessed as low, moderate, or high risk based on the components of the “female athlete triad,” which consists of “low energy availability (EA) with or without disordered eating, menstrual dysfunction, and low bone mineral density.”

Then they checked back to see which athletes developed bone stress injuries, a continuum that ranges from incipient “stress reactions” to stress fractures and (in the worst cases) full fractures.

About half the runners (24 out of 47) were classified as low risk, and three of those athletes went on to develop a bone stress injury – so less than 13 percent.

Sixteen of the runners had moderate risk, and fully half of them developed bone stress injuries within an average time of a year, meaning they were about four times more likely to get injured than the low-risk group.

There were just seven high-risk athletes, and five of them developed bone stress injuries, meaning they were almost six times more like to get injured than the low-risk group.

These are pretty sobering numbers. If you’re an athlete in the moderate- or high-risk group, your chances of avoiding a season-ending injury are worse than a coin toss. Modifying behavior about eating is very tricky (and by no means something I’m qualified to offer advice about), but perhaps this information can help clarify the stakes for some people.

So how is this risk status calculated? The methods are outlined in a consensus statement published in the British Journal of Sports Medicine in 2014 (and freely available online). Figure 4 in that paper shows the risk calculation in detail, but basically it involves assigning 0, 1, or 2 points in each of six categories. A total score of 2 to 5 is moderate risk, while 6 or higher is high risk.

Here are the six categories (which had to be modified a bit in the current study based on the data available):

(1) Low energy availability or loss of body weight, as a result of past (one point) or current (two points) disordered eating.

(2) Low body mass index (BMI): one point for 17.6 to 18.4, two points for 17.5 or below.

(3) Irregular periods: one point for 6 to 9 periods in the last 12 months, two points for fewer than 6.

(4) Delayed menarche: one point for between 15 and 16 years old, two points for 16 or older.

(5) Low bone mineral density: one point for a Z-score of less than -1, two points for -2 or lower.

(6) Previous stress fractures (or stress reactions): one point for one previous fracture, two for two or more.

This risk stratification protocol came from a series of conferences in 2012 and 2013, in the hopes of developing some standardized return-to-play guidelines. As Tenforde, the study’s lead author, points out, the new results show that the risks calculated with this method are “not theoretical.”

So what do you do with this information? That’s the trickier part. Just because someone has higher risk doesn’t necessarily mean she should be stopped from training. It may be, for example, that some of the past risk factors that raise the risk score are already being addressed clinically.

Still, Tenforde offers four takeaways for coaches and medical staff:

First, anyone in the moderate- or high-risk categories needs to be assessed to ensure they’re getting enough calories to support their training, along with calcium and vitamin D.

Second, such athletes need to be monitored on an ongoing basis to ensure these areas—nutrition, menstrual function, and bone health – continue to be addressed.

Third, higher-risk athletes may want to consider altering training, for example by making low-impact cross-training a bigger part of their weekly routine. Good sleep and recovery are also important, Tenforde says.

And fourth, not all stress fractures are equal. Most of the stress fractures in low-risk athletes were in the foot – a consequence, perhaps, of biomechanics and sport choice (since many of the low-risk athletes were in other sports like netball, where jumping puts big forces on the foot).

In the higher-risk categories, on the other hand, many of the fractures were in places like the sacrum, pelvis, and femoral neck, where the bones tend to have a different and softer interior structure. These fractures may be less of a direct consequence of biomechanical forces and more the result of weakened bone.

As a result, Tenforde suggests, an injury in one of those places may be a warning sign that the athlete is at higher risk – so that should trigger an evaluation of bone health, including a bone density scan.

READ MORE ON: injury-prevention stress fracture