Bone Density and Stress Fractures in Runners: The Evidence

What actually predicts a stress fracture in runners?

The most rigorous synthesis is the systematic review with meta-analysis by Wright et al. (2015), which pooled running-specific studies of lower-extremity stress fractures. Two factors stood out. A previous stress fracture carried a pooled odds ratio of 4.99, meaning runners with a past injury were roughly five times more likely to sustain another. Female sex carried a pooled odds ratio of 2.31.

What is striking is how little else reached that level of confidence. Many popular candidates, from arch height to leg-length differences, had limited or inconclusive evidence in this analysis. That does not prove they are irrelevant, but it does mean the data do not yet support treating them as reliable predictors.

For a practical reader, the message is to take history seriously. If you have broken down before, your individual risk is elevated regardless of how your training feels, and that warrants more conservative load progression and closer attention to the modifiable factors covered below.

Does low bone density cause stress fractures, or is it more complicated?

Bone density is intuitively appealing as an explanation, but the evidence is mixed and depends on how you frame the question. In the prospective cohort of collegiate cross-country runners studied by Kliethermes et al. (2021), bone mineral density was not a statistically significant predictor of bone stress injury in the final model. That is a useful corrective: a single whole-body BMD number does not neatly sort who gets injured.

Yet density clearly matters at the severe end. Nattiv et al. (2013) found that lower BMD was associated with higher-grade injuries on MRI, and Tenforde et al. (2024) reported that female runners with high-risk fractures had significantly lower lumbar-spine BMD z-scores than those with low-risk injuries.

So the fairer summary is that low BMD is less a universal cause and more a marker that raises the odds of the worst fractures, especially when it travels with under-fuelling and poor recovery.

How does the female athlete triad raise the risk?

The female athlete triad describes the interplay of low energy availability, menstrual dysfunction and impaired bone health. The prospective multisite study by Barrack et al. (2014) followed 259 physically active girls and women and found a clear dose-response pattern: bone stress injury incidence rose steeply as the number of triad risk factors increased.

Overall, 10.8% sustained an injury, but among those with the highest cumulative risk, defined by high training volume, low BMD and several of low BMI, oligo- or amenorrhoea, dietary restraint or leanness-sport participation, 46% were injured. That is a large gradient driven by cumulative exposure rather than any single variable.

Tenforde et al. (2017) confirmed this prospectively using a Cumulative Risk Assessment score: moderate-risk athletes were about twice as likely, and high-risk athletes nearly four times as likely, to develop a bone stress injury than low-risk athletes. Cross-country runners accounted for most of those injuries.

Can nutrition protect bone in runners?

One of the few interventions tested in a randomised controlled trial is calcium and vitamin D supplementation. Lappe et al. (2008) studied 5,201 female Navy recruits during eight weeks of basic training, a period of sharply increased mechanical load not unlike a heavy running block.

Recruits taking 2,000 mg of calcium and 800 IU of vitamin D daily had a 20% lower incidence of stress fractures than those on placebo (5.3% versus 6.6% by intention-to-treat, and 21% lower in the per-protocol analysis). This is a modest but genuine protective effect from a cheap, low-risk measure.

Two caveats matter. The population was military recruits, not distance runners, so generalisation is imperfect. And supplementation is not a substitute for adequate overall energy and nutrient intake. Still, ensuring sufficient calcium and vitamin D is a reasonable, evidence-supported step, particularly for runners ramping load or with limited sun exposure.

Does running form matter for stress fractures?

Two recent papers sharpen the biomechanical picture, and they push in different directions than folk wisdom expects. Kliethermes et al. (2021) prospectively followed collegiate cross-country runners and found that bone stress injury risk fell by roughly 5% for each one step-per-minute increase in cadence. Lower step rate emerged as an independent, modifiable biomechanical risk factor.

Impact magnitude, by contrast, looks weaker than assumed. The systematic review and meta-analysis by Milner et al. (2023), pooling 14 studies of tibial stress fracture, found that vertical impact peak, vertical active peak and peak braking force were not significantly different between injured runners and controls.

Taken together, the gross size of the ground-reaction force may distinguish injured runners less reliably than how that force is applied. The authors point instead toward loading rate and other factors for further study. For runners, nudging cadence slightly higher is a low-cost adjustment with some prospective support.

Are all stress fractures equally serious?

No, and the distinction has real consequences for recovery. Nattiv et al. (2013), in a five-year prospective study of collegiate track and field athletes, found that higher MRI-graded injury severity correlated with longer time to return to play. Female sex, lower BMD and prior injury were each associated with higher-grade fractures.

Location matters too. Tenforde et al. (2024) compared female runners with high-risk fractures, at the pelvis, sacrum and femoral neck, against those with low-risk injuries. The high-risk group had lower lumbar-spine BMD z-scores, higher eating and body-shape concern scores, and markedly worse sleep: 80% slept under seven hours on weeknights, versus 33% of controls.

This supports stratifying injuries by site rather than treating them as one category. A metatarsal reaction and a femoral-neck injury are not equivalent, and the higher-risk sites warrant closer screening of bone health and energy availability.

Where is the evidence still weak or uncertain?

It is worth being candid about the limits. Much of the strongest data comes from military recruits (Lappe et al., 2008) and collegiate athletes (Barrack et al., 2014; Tenforde et al., 2017), predominantly female, so extrapolation to recreational and male runners is imperfect. Several studies here are cross-sectional or observational, which shows association rather than proof of cause.

The Wright et al. (2015) review also reminds us that many proposed risk factors simply lack good evidence either way, which is different from being ruled out. And while Milner et al. (2023) challenge impact magnitude as a marker, they explicitly call for more work on loading rate rather than closing the question.

The reasonable position is therefore probabilistic. Address what is well supported, prior injury awareness, adequate energy and nutrient intake, and cadence, while treating single measurements like one BMD scan as informative but not decisive on their own.

What is the single biggest risk factor for a stress fracture in runners?

According to the meta-analysis by Wright et al. (2015), a previous stress fracture is the strongest predictor, with a pooled odds ratio of about 4.99, meaning affected runners are roughly five times more likely to sustain another. Female sex was the second strongly supported factor, with an odds ratio around 2.31. Most other proposed risk factors had limited or inconclusive evidence in that review.

Does low bone density mean I will get a stress fracture?

Not directly. In the prospective study by Kliethermes et al. (2021), bone mineral density was not a statistically significant predictor on its own. However, lower BMD is linked to more severe, higher-grade injuries (Nattiv et al., 2013) and to high-risk fractures of the pelvis, sacrum and femoral neck (Tenforde et al., 2024). It is best seen as one contributing marker, not a sole cause.

Do calcium and vitamin D supplements actually reduce stress fractures?

There is randomised-trial evidence that they can. Lappe et al. (2008) gave 5,201 female Navy recruits 2,000 mg calcium and 800 IU vitamin D daily during eight weeks of basic training and found a 20% lower stress-fracture incidence than placebo. The effect is modest and the population was military recruits, but it supports ensuring adequate intake, particularly during heavy training blocks.

Can changing my running cadence lower stress-fracture risk?

There is some prospective support. Kliethermes et al. (2021) found that bone stress injury risk fell by roughly 5% for each one step-per-minute increase in cadence among collegiate cross-country runners, making low step rate an independent, modifiable factor. Notably, Milner et al. (2023) found impact-force magnitude did not reliably distinguish injured runners, so how force is applied may matter more than its size.

What is the female athlete triad and why does it matter for bone?

The triad is the interplay of low energy availability, menstrual dysfunction and impaired bone health. Barrack et al. (2014) showed injury risk rises steeply with the number of triad factors present: 46% of the highest-risk group were injured versus 10.8% overall. Tenforde et al. (2017) found high-risk athletes were nearly four times as likely to sustain a bone stress injury, underscoring the importance of adequate fuelling.

Are some stress fractures more serious than others?

Yes. Nattiv et al. (2013) found that higher MRI-graded severity correlated with longer return to play. Location also matters: Tenforde et al. (2024) distinguished high-risk sites (pelvis, sacrum, femoral neck) from low-risk ones, with high-risk cases showing lower lumbar BMD, more eating and shape concern, and worse sleep. High-risk sites warrant earlier assessment and closer bone-health screening.