REF-1392 · Barrack et al., 2014 · The American Journal of Sports Medicine · Paper · PubMed

Do missed periods or low weight raise my stress-fracture risk?

Of 259 active girls and women, 15.3% with BMI under 21 had a bone stress injury (10.8% overall); some combinations, 30-46%. Missed periods: reported only in a combination.

Main caveat: Read from the abstract only. Subgroup sizes, follow-up length, injury sites, the reference group for the odds ratios and any adjustment could not be checked.
limited evidenceMeasured on recreational to highly trained athletes

Barrack and colleagues asked whether risk factors defined by the female athlete triad — energy availability, menstrual function and bone mass, each a spectrum — predict bone stress injuries (stress reactions and stress fractures) in physically active girls and women, alone or combined.

At baseline they assessed 259 participants (mean age 18.1 ± 0.3 years) from 4 cohorts: anthropometry, eating attitudes and behaviours, menstrual function, sports participation or exercise activity, pathological weight-control behaviours, and DXA bone density of the whole body, total hip, femoral neck and lumbar spine with body composition.

They then recorded physician-confirmed injuries prospectively. Twenty-eight participants (10.8%) had a bone stress injury. Single factors significantly associated with a later injury included 12 h or more a week of purposeful exercise (14.7% had a bone stress injury), a BMI under 21.0 kg/m² (15.3%) and low bone mass, a BMD Z-score under −1.0 (21.0%).

The strongest combinations were low bone density with 12 h or more of exercise a week (29.7%; OR 5.1, 95% CI 2.2-12.1) and 12 h or more a week with a leanness sport or activity and elevated dietary restraint (46.2%; OR 8.7, 95% CI 2.7-28.3).

Of those with 12 h or more of exercise a week, a Z-score under −1.0 and three to four of BMI under 21.0, oligo- or amenorrhoea, elevated dietary restraint and a leanness sport, 46% had a bone stress injury.

The authors conclude that risk rose from about 15-20% with one significant factor to 30-50% with combined triad-related factors, so cumulative risk grows as triad factors accumulate.

Our reading: the base's red flag for menstrual and energy signs alongside bone pain rested on a cross-sectional study of 59 elite distance athletes (REF-343); this adds a prospective, physician-confirmed dose of the same idea in young women, with the caveats that the events are few and the follow-up length is not in the abstract.

Protocol

Design: prospective cohort in 4 cohorts at several sites; Level of evidence 3 (authors).
Who: n = 259 physically active girls and women, mean age 18.1 ± 0.3 years, in competitive or recreational exercise activities.
Baseline: anthropometric characteristics; eating attitudes and behaviours (dietary restraint); menstrual function (oligo- or amenorrhoea); sports participation or exercise activity, including hours of purposeful exercise and leanness-sport participation; pathological weight-control behaviours; DXA bone mineral density of whole body, total hip, femoral neck and lumbar spine, and body composition.
Outcome: a bone stress injury (stress reaction or stress fracture) confirmed by a physician during prospective follow-up; "incidence" in the paper is the proportion of participants (or of a risk-factor group) who had one during follow-up. Length of follow-up, injury sites and imaging criteria are not in the abstract.
Analysis: proportions with a bone stress injury by single and combined factors; odds ratios with 95% CIs for combinations; significance P < .05. The reference group and any adjustment are not stated in the abstract.
Funding and interests: not available from the abstract or the PubMed record; Crossref lists no funder. No device maker, federation or data platform is named.

Key Results

Group at baselineHad a bone stress injury during follow-upOdds ratio (95% CI)
Whole cohort10.8% (28 of 259)—
≥12 h/wk purposeful exercise14.7%not in the abstract
BMI <21.0 kg/m²15.3%not in the abstract
BMD Z-score <−1.021.0%not in the abstract
BMD Z-score <−1.0 + ≥12 h/wk exercise29.7%5.1 (2.2-12.1)
≥12 h/wk exercise + leanness sport/activity + elevated dietary restraint46.2%8.7 (2.7-28.3)
≥12 h/wk exercise + BMD Z-score <−1.0 + 3-4 of: BMI <21.0, oligo- or amenorrhoea, elevated dietary restraint, leanness sport46%not in the abstract
Authors' summary: about 15-20% with significant single factors, 30-50% with significant combined triad-related factors.
Oligo- or amenorrhoea alone: the abstract lists it only within the combinations; whether it was a significant single factor is not stated.
Not in the abstract: subgroup sizes, follow-up length, injury sites, whether combinations were tested against the same reference group.

Mechanisms

Measured: the proportion with a later bone stress injury rose as baseline risk factors (high exercise hours, low BMI, low bone density, dietary restraint, leanness sport, menstrual dysfunction) were combined.
Authors' interpretation: cumulative risk for bone stress injury increases as the number of female athlete triad-related risk factors accumulates; identifying them may help target those at increased risk and shape prevention guidelines.
Inference: the factors map onto the low-energy-availability pathway the base already holds — too little food for the training done turns down reproductive and bone-building hormones (PHYS-32), and bone fails when repeated loading outpaces repair (PHYS-37); high training hours add loading cycles while low energy availability slows repair. The abstract does not test this pathway; it shows the risk factors add up.

Practical Application

Do missed periods or low weight raise my stress-fracture risk? Of 259 active girls and women, 15.3% with BMI under 21 had a bone stress injury (10.8% overall); some combinations, 30-46%. Missed periods: reported only in a combination.
Bone pain plus energy or menstrual signs is a firmer red flag than either alone: in this cohort the risk factors were 12 h or more of purposeful exercise a week, a BMI under 21.0, a bone-density Z-score under −1.0, dietary restraint, a leanness sport and oligo- or amenorrhoea, and combinations of them went with about a third to a half having a bone stress injury (B, prospective cohort, from the abstract).Not from this paper; the base's own practice: bone-like pain goes to a clinician whatever else is present (SYNTH-73 red flags, whose menstrual line rests on the cross-sectional REF-343); these factors are the reason to say it more firmly. Screening questions: SYNTH-54; REDs is a physician's diagnosis (REF-341).
Count the factors, not one number: in this cohort the share with an injury roughly doubled from a single significant factor (about 15-21%) to some combinations (about 30-46%), against 10.8% in the whole cohort.
Prevention lies in the modifiable factors (inference): fuelling enough for the training and keeping the menstrual cycle are what an athlete and coach can act on; the study did not test an intervention. Not from this paper; the base's own practice: screen before advising weight loss (SYNTH-54).
What it is good for as a source: a prospective, physician-confirmed demonstration that triad-related risk factors accumulate for bone stress injury in young active women, with the sizes above. What it does not show: risk in adult women runners specifically, in men, or per season; which single factor matters most once others are accounted for; whether menstrual dysfunction alone predicted injury; that changing any factor lowers risk.

Limitations

Read from the abstract only. Subgroup sizes, follow-up length, injury sites, the reference group for the odds ratios and any adjustment could not be checked.
Few events. 28 injuries in 259 participants; the combined-factor percentages come from small subgroups, and the widest interval (OR 2.7 to 28.3) shows how imprecise they are.
Who. Young (mean 18.1), female, mixed exercise activities and competitive levels; not a runner sample and no men.
Associations, not causes. High exercise hours, low BMI and low bone density travel together; the design cannot separate their effects or show that changing one lowers risk.
Funding and interests: not available from the abstract or the PubMed record; Crossref lists no funder. No device maker, federation or data platform is named.
How this entry was checked and changed
Why this level

(from the abstract) — a prospective cohort above the rubric's 200 for B+ (measured exposure: DXA bone density, assessed eating behaviour and menstrual function; physician-confirmed outcome), held at B, as REF-1313 was for an abstract-only cohort with small strata, because the 28 injuries mean the combined-factor percentages rest on small subgroups (the abstract gives no subgroup sizes, and the OR of 8.7 has a 95% CI of 2.7 to 28.3), the follow-up length is not in the abstract, and the odds ratios' reference group and any adjustment are not stated. Read from the abstract (closed access; Europe PMC holds no full text and OpenAlex lists no open copy)