Clinical synthesis based on the most recent international consensus statements: the 2025 BSI Delphi, IOC REDs 2023, Nature Reviews 2022, and the 2021-2024 JOSPT and BJSM meta-analyses.

The bone stress injury (BSI) now denotes the continuum of injury formerly fragmented into “stress reaction”, “fatigue fracture” and “complete fracture”. This consensus terminology, adopted by the Nature Reviews Disease Primers 2022 and the international Delphi 2025, better reflects the evolving nature of the condition.¹٬¹¹ 🦴

Definition, epidemiology and risk factors

A BSI results from an accumulation of bone microdamage when cyclic mechanical stresses exceed the adaptive and repair capacities of bone tissue. The spectrum runs from isolated bone marrow oedema (visible only on MRI) to a complete, displaced cortical fracture.¹

Epidemiologically, BSIs account for 15 to 20 % of musculoskeletal injuries in runners, with up to 13 % of female athletes having a prior history in some cohorts.² The sports at greatest risk are distance running, triathlon, classical ballet, basketball, and military recruits in initial training.¹

Risk factors are classically organised into two groups:

• Extrinsic factors (modifiable): training error remains the dominant trigger: rapid increase in volume, intensity or frequency, change of surface (road → track), unsuitable equipment.⁴
• Intrinsic factors: female sex, personal history of BSI (×3 risk of recurrence), low BMD, altered running biomechanics, vitamin D and calcium deficiency.¹⁰

One intrinsic factor deserves particular attention: REDs (Relative Energy Deficiency in Sport). This syndrome, redefined by the international IOC consensus in 2023, encompasses the consequences of low energy availability on bone, hormonal, metabolic and mental health, and affects both sexes.⁵

Pathophysiology and natural history of BSI

Bone is a living tissue in constant remodelling through the coupling between osteoclasts (resorption) and osteoblasts (formation).¹⁵ Under normal cyclic loading, this remodelling adapts bone tissue to the stresses applied to it (Wolff's law). Under excessive cyclic loading that is not spaced out enough, osteoclastic resorption transiently exceeds formation, creating resorption cavities that locally weaken the cortex and create increased porosity.¹

If loading continues without allowing healing, the microdamage coalesces into a cortical fracture line, up to complete fracture.

Female athletes have a risk of bone stress injury 2 to 4 times higher than that of men in the same sports.¹⁰ This over-representation is not a fatalistic biological accident: it is largely explained by a modifiable and under-diagnosed systemic factor, REDs (Relative Energy Deficiency in Sport). 🚺

Physiological specificities and a ×2-4 risk

Several mechanisms converge to explain female bone vulnerability in the sporting context:

• Lower peak bone mass that is reached earlier (around age 18-20 in women vs. 20-25 in men).¹
• Hormonal dependence of bone formation: oestradiol is protective; any disturbance of the hypothalamic-pituitary-gonadal axis directly affects BMD.²
• Bone geometry: female femoral and tibial shafts are on average narrower, with a lower section modulus.³
• Sociocultural pressures specific to sport (aesthetics, weight, dietary restrictions) and more frequent in certain at-risk disciplines (dance, gymnastics, distance running).⁴

The REDs syndrome: definition, screening, CAT2

The REDs syndrome is defined as the set of deleterious consequences of low energy availability (LEA) on health and athletic performance. It was redefined by the IOC in 2023 as a multisystem syndrome affecting both sexes.⁴ The Female Athlete Triad Coalition, which contests this model, maintains the triad and updated it in 2025: the two frameworks coexist.

The energy availability (EA) is calculated as follows:

EA = (Energy intake − Exercise energy expenditure) / Fat-free mass (kg FFM)

This calculation is hard to do outside the laboratory, and the 2025 update from the Female Athlete Triad Coalition no longer recommends a single energy availability threshold (for example 30 kcal/kg FFM/day): menstrual disturbances occur above it. When only a single measure is available, it takes as signs of a high risk of undernutrition or chronic energy deficit a weight below 85 % of expected weight, a BMI below 17.5 kg/m², an energy availability well below 45 kcal/kg FFM/day or, in adolescent girls, a BMI-for-age below the 5th percentile; markers of metabolic compensation (T3, resting metabolic rate) are a matter for the physician.¹³

The CAT2 (REDs Clinical Assessment Tool version 2, IOC 2023) proposes a 4-level stratification using an “enriched traffic light” system:⁹

Triad Cumulative Risk Assessment and clinical stratification

The Triad Cumulative Risk Assessment (CRA) aggregates 6 indicators into a score from 0 to 12 points.⁵ The 2025 update from the Female Athlete Triad Coalition revises its grids, one for adults and one for adolescent girls, and rates its predictive value for bone stress injury in female athletes at level B of its own scale (from A to D). The tool is designed for the physician; the physiotherapist uses it to identify and refer. The table reproduces, as an extract, the moderate-risk (1 point) and high-risk (2 points) categories of the 2025 adult grid, without the criteria for weight loss over one month or the treatments that act on bone; for an adolescent girl, the grid and the course of action differ.¹³

The total score categorises the risk: 0-1 = low, 2-5 = moderate, ≥6 = high. In the Tenforde 2017 cohort (239 female collegiate athletes across 16 sports), the risk of BSI was multiplied by 2.6 in the moderate category and by 3.8 in the high category.⁵

Male specificity: long neglected, REDs also affects men: the prevalence of LEA in male athletes is estimated at 49 % in a recent meta-analysis (Gallant 2025), particularly marked in endurance sports and weight-category sports.¹¹ The Triad score was modified for men and validated as predictive of BSI in male runners in 2019 (Kraus, Tenforde, Nattiv).¹²

The diagnosis rests on a tripod: targeted history, localising clinical examination, imaging suited to the level of risk. A strong clinical suspicion almost always precedes imaging confirmation. 🩺

Targeted history and red flags

The systematic history looks for a highly suggestive cluster of features:

• Insidious, localised pain, first on exertion, then increasingly early, until it is present at rest if the condition progresses.¹
• Recent change (2-6 weeks) in training load: volume, intensity, frequency, terrain, shoes.²
• Systemic intrinsic factors: amenorrhoea/oligomenorrhoea, history of eating disorders, low BMI, personal or family history of BSI or fragility fracture.³
• Vitamin D and calcium status: to be checked systematically because it is correctable.⁴

Clinical examination and differential diagnosis

The clinical examination aims to locate the painful area precisely and to reproduce the symptoms.

• Focal bone palpation: exquisite pain at a precise bony point, the most reliable sign for accessible sites (tibia, metatarsals, fibula).⁵
• Single-leg hop test: positive if the athlete cannot perform several hops without pain. Useful for lower-limb BSIs.¹
• Fulcrum test for the femur, percussion at a distance and tuning fork for the tibia: variable sensitivity, low specificity.⁵

The differential diagnosis must systematically rule out:

• The medial tibial stress syndrome (diffuse pain over > 5 cm along the posteromedial border).
• The tendinopathies (pain on tensioning of the tendon, not on bony impact).
• The chronic exertional compartment syndrome (cramping that subsides at rest).
• The radiculopathies or nerve compressions (notably for sacral BSIs that mimic sciatica).⁶

Imaging: why and how to classify?

Diagnostic performance is now well established:

Two complementary classifications guide the decision:

1. Anatomical classification (high risk vs. low risk) determines the aggressiveness of management:⁹٬¹⁰

2. Fredericson MRI classification (grades 1 to 4), correlated with the time to return to sport:⁷٬¹¹

Decision-making diagnostic algorithm

To integrate clinical suspicion, imaging and risk stratification into a coherent decision, here is the recommended summary algorithm.

The guiding principle, formalised by the 2025 international Delphi consensus, is the dynamic balance between mechanical load and repair capacity. Treatment is largely load-driven and symptom-based, not time-based.¹٬²

Hierarchy of recommended interventions

1. Stopping the activity that provokes the pain: immediate and non-negotiable. Key concept: relative rest, not absolute rest.³
2. Offloading: complete and protected with crutches for high-risk BSIs; partial, or simple activity modification, for low-risk BSIs.⁴
3. Identification and correction of causal factors: training errors, biomechanics, equipment, and above all systemic assessment (REDs, vitamin D, calcium, energy balance).⁵
4. Nutritional optimisation: aim for 1000-1300 mg/day of calcium and a serum 25(OH)D > 75 nmol/L (> 30 ng/mL), ideally > 100 nmol/L.⁶

Progressive loading and the acute-to-chronic ratio

Rehabilitation is built around optimal loading, defined by Warden et al. as “the load that produces no symptoms during, after or the day after the activity”.² 🏃

The 10 % rule (do not increase weekly volume by more than 10 %) is recommended by Gabbett on the basis of unpublished rugby data; it remains a pedagogical benchmark, with no demonstrated predictive value in runners.⁷٬¹⁵

The acute-to-chronic workload ratio (ACWR) relates the load of the current week (acute, 7 d) to the average of the previous 4 weeks (chronic, 28 d). The acute-to-chronic workload ratio (ACWR) comes from team sports: Gabbett (2016) drew his “optimal zone” of 0.8 to 1.3 and his “danger zone” from 1.5 upwards from cricket, Australian football and rugby data, while urging caution for individual sports.⁷ A 2020 systematic review (27 studies, quality scores from 48 to 64 %) found there a trend towards the lowest risk between 0.8 and 1.3, with high variability, and methodological problems to be resolved before using it.⁸ It is now contested: a reanalysis of published data finds no predictive value of its own, and the authors of that reanalysis propose abandoning it.¹² A 2025 meta-analysis finds, according to its authors, an association with injuries, to be used with caution given the heterogeneity of the studies.¹³ Another, in 2026 in team sports, judges the association small to moderate and highly heterogeneous, and concludes that the ratio should not be used alone, only as one indicator among others.¹⁴ It is therefore neither a validated prevention tool nor a return-to-sport criterion: progression is adjusted to tolerance, including next-day pain and stiffness, and no weekly progression threshold is validated. In runners, a cohort of 5,205 runners links overuse injuries neither to the acute-to-chronic ratio (the relationship there is inverse) nor to the week-to-week increase, but to the session whose distance exceeds the longest run of the previous 30 days by more than 10 % (observational study).¹⁵

In clinical practice: return after BSI is adjusted to symptoms during, after and the day after the activity, with no ratio threshold.²

Adjuvant modalities and GRADE level of evidence

Recurrence is a real issue: 20-30 % of runners will have a new BSI, with a risk about 3 times higher in women.¹٬² A structured and individualised approach is essential.

Self-management and patient education

Patient education is the best protective factor. Three priority areas:

• Understanding the concept of cumulative overload: the athlete must know that the trigger is not an acute event but an accumulation of micro-loads. Volume progression remains the no. 1 modifiable factor.³٬⁴
• Detection of REDs: screening with the REDs Clinical Assessment Tool v2 (CAT2) from the IOC 2023: questions on menstrual cycles, BMI, eating behaviour, performance, chronic fatigue.⁵ 🧠
• Keeping a training diary: loads, pain, sleep, perceived exertion (RPE), mood, in order to detect warning signs early. 📝

Performance-based return-to-sport criteria

Return to sport is not an event but a criteria-based process. The decision incorporates:

• Clinical criteria: no pain on palpation, on the hop test, or on 30 min of continuous running without symptoms, and no symptoms the next day.⁶
• Functional criteria: strength balance, neuromuscular control, quality of the running pattern (video analysis recommended for runners with recurrences).⁶
• Psychological criteria: assess kinesiophobia (using the TSK-11 scale) and confidence in one's ability to resume, both often neglected.⁷
• Contextual criteria: correction of causal factors (REDs, load, equipment): without it, recurrence is predictable.⁵

The role of imaging in allowing return remains debated. Radiological healing (resolution of the MRI bone marrow oedema) often lags several weeks/months behind the disappearance of symptoms; a normal MRI is not required to allow return in the majority of low-risk cases.⁸٬⁹ By contrast, for high-risk BSIs (femoral neck, navicular, anterior tibia), imaging verification remains prudent before returning to high-impact loading.¹⁰

Case 1: Runner, gluteal and low back pain

A healthy distance runner in his late twenties, presenting with insidious-onset low back and right gluteal pain, with no trauma, appearing after a marathon. The fracture went unrecognised for one month. MRI reveals extensive bone marrow oedema of the right sacral ala, consistent with a sacral BSI.

Key lesson: in a young, healthy runner, low back or gluteal pain without trauma, appearing after a long run, should raise the suspicion of a sacral BSI and lead to appropriate imaging (MRI).

Source: Pinheiro et al. BMJ Case Reports 2023, PMC10481734.¹

Case 2: Elite female runner, bilateral femoral neck stress fractures

A high-performance female runner presenting with chronic bilateral groin pain, dismissed as “psoas tendinopathies”. MRI reveals bilateral femoral neck BSIs on the compression side. The enquiry reveals secondary amenorrhoea (> 1 year), a low BMI, and insufficient energy intake, a picture of REDs. Multidisciplinary management: orthopaedic (prophylactic screw fixation), nutritional, psychological, endocrine.

Key lesson: any bilateral or multiple BSI requires systematic screening for REDs and a complete endocrine work-up.

Source: Jasqui-Remba et al. BMJ Case Rep 2019, PMID 31451476.²

Case 3: Series of 62 navicular fractures in a sporting cohort

Prospective series of 59 patients (62 fractures) over 11 years (impact sports + running). Key finding: mean diagnostic delay of 4 months before correct management (pain centred on the N spot dorsally), frequent need for scintigraphy or MRI (normal radiographs in the majority). Mean time to return to activity 4-6 months. Risk of recurrence significantly associated with poor adherence to initial offloading.

Key lesson: navicular BSI is under-diagnosed and at high risk of non-union. Dorsal foot pain at the N spot calls for MRI even if the radiograph is normal.

Source: Saxena et al. J Foot Ankle Surg 2017, PMID 28842101.³

• BSI is a great mimic : demand a systematic diagnostic rethink if the pain persists.
• Any multifocal or bilateral BSI calls for systematic screening for REDs.
• Sites with precarious blood supply (navicular, femoral neck, anterior tibia) call for great caution and early specialist referral.

Effective implementation of the recommendations rests on three pillars: interprofessional collaboration, shared decision-making with the patient, and systematic outcome measurement. 🧑‍⚕️

When and to which professionals should patients be referred?

Any high-risk BSI must be referred without delay to a sports physician or an orthopaedic surgeon. Advice from nutrition/endocrinology is essential in cases of multiple BSIs, a history of REDs or low BMD. Psychological support may be justified in the face of marked kinesiophobia or contributing psychosocial factors (perfectionism, low social support).¹٬²

Measuring outcomes and overcoming barriers to implementation

The integration of Patient-Reported Outcome Measures (PROMs), namely the VISA-A score modified for BSI, a numerical pain scale, a return-to-sport scale, must be accompanied by performance tests (hop test, 30 min of continuous running). Beyond statistical significance, aim for the Minimal Clinically Important Difference (MCID).³

Red flags in practice: rely on the BSI-specific red flags already listed above (night pain, weight loss, fever, history of corticosteroid therapy) rather than on a standardised grid. Verhagen's 2016 review, devoted to low back pain guidelines, moreover does not provide such a tool: on the contrary, it notes a lack of consensus between guidelines on which flags to retain and a lack of diagnostic accuracy data to support them.⁴

Shared decision-making: involving the athlete in the choices (modality, timetable for return, level of risk accepted) is part of patient-centred care. The 2018 Cochrane review by Légaré nevertheless urges caution: it remains uncertain whether interventions aimed at increasing shared decision-making are effective, the certainty of the evidence being low to very low.²

Barriers to implementation: the most frequently reported obstacles are lack of time, lack of skills in the critical appraisal of the literature, and insufficient organisational support. Facilitators include strong leadership, access to resources, and continuing education.