Bone health and risk of stress fracture in female endurance athletes

2020-01-09T12:50:15Z (GMT) by Rachel L. Duckham
One in two women over the age of 50 will be diagnosed with an osteoporotic fracture during their life time (Van Staa et al. 2001). Osteoporosis is a condition in which the bone mineral density (BMD) is lost causing the bone to become weak and liable to fracture. It is well established that participation in weight bearing exercise (gymnastics and running) may be beneficial to BMD, due to the high mechanical loading (Drinkwater 1994; Kannus et al. 1994a; Marcus et al. 1992; Snow 1996).The high prevalence of amenorrhoea (1-44%) (Bennell et al. 1997b) in female athletes may result in poor bone health, leading to increased risk of premature osteoporosis or stress fracture injury, disabling an athlete s present and future career (Nattiv 2000; Nattiv et al. 1997). Oestrogen deficiency in amenorrhoeic athletes may compromise the beneficial effects of exercise, leading to lower BMD (Bass 2003; Saxon and Turner 2006), but it is unknown whether this is accompanied by structural differences such as changes to section modulus (Z). There is evidence that athletes display seasonal gains and losses in BMD with changes in training (McClanahan et al. 2002; Snow et al. 2001; Winters and Snow 2000), however, in amenorrhoeic athletes, it is possible that any seasonal losses may not be recovered thus contributing to lower BMD. Studies have reported incidence rates of stress fracture to range between 8.7 -21.1 % in athletes with female endurance athletes at the greatest risk possibly due to the aforementioned high prevalence of menstrual dysfunction and a demand for thinness (Bennell et al. 1996a; Kelsey et al. 2007; Nattiv et al. 2000). However prospective monitoring of stress fracture in female endurance athletes, the gold standard for measuring incidence, is limited with conflicting evidence of incidence and risk factors of stress fracture, possibly due to varying methodology and with no standard definition of stress fracture (Snyder et al. 2006). There is also limited robust evidence to determine whether psychological traits are associated with stress fracture history. There is solid evidence to suggest that the reoccurrence rate of stress fracture in athletes is high over the first 12-months following an initial stress fracture; this could be caused by retraining when the bone is at its weakest. Studies of other musculoskeletal injuries have shown bone loss following injury up to 12-months, however there is no research to suggest how much bone may be lost following a stress fracture injury. It is important therefore, to monitor, using robust methodology, the incidence and subsequent consequences for bone loss associated with stress fracture in athletes in order to provide support for potential intervention and treatment. The main aims of this thesis are two-fold to: 1) determine prospectively the predictors of bone health and stress fracture in female endurance athletes, and 2) determine whether bone geometry and density change following a stress fracture. The specific objectives of the thesis were five-fold to: 1) determine the correlates of stress fracture history, 2) compare bone density and geometry according to menstrual function, 3) determine the incidence rates of stress fracture and identifiable risk factors, 4) quantify the seasonal variation in parameters of bone health and 5) determine the magnitude and timescale of bone loss and subsequent regain. Seventy United Kingdom based female endurance athletes (runners and triathletes) aged 18-45 years were prospectively monitored for 12-months. At each assessment (baseline, 6-month and 12-months) questionnaires were used to assess menstrual, nutritional, eating psychopathology, exercise cognition, and injury histories. BMD, bone mineral content (BMC), hip geometric parameters, and body composition were assessed using dual x-ray absorptiometry (DXA), and anthropometric measures were taken. Training and stress fracture injury were prospectively monitored. Retrospective data determined 19 (27%) of the athletes had a history of clinically diagnosed stress fracture. Athletes with a history of stress fracture had a significantly higher prevalence of current (47% vs 27%, p=0.008) and past (79% vs 53%, p=0.035) menstrual dysfunction and higher global scores on the eating disorders examination questionnaire (EDE-Q) (p=0.049) and the compulsive exercise test (CET) (p=0.016) compared to non-stress fracture athletes. Bone parameters by DXA, training duration, age, age at menarche and anthropometric measurements did not differ between groups. This study found a high prevalence of past stress fracture, and identified eating and exercise behaviour to be related to stress fracture risk independent of menstrual dysfunction. To compare bone geometry and density according to menstrual function the athletes were classified as either a/oligomenorrhoeic athletes (<9periods/year) (AA) or eumenorrhoeic athletes (>10 periods/year) (EA) and compared to 88 eumenorrhoeic sedentary controls (EC). 30 athletes were AA and 40 EA. EC were significantly older, heavier and shorter than EA and AA who did not differ significantly. Femoral neck BMD was significantly higher in EA than AA and EC (mean (SE) EA: 1.117 (0.015), AA: 1.036 (0.020) and EC: 0.999 (0.014) g/cm2 respectively; p<0.001). Section modulus (Z) was significantly higher in EA than EC (EA: 657 (20), AA: 639 (20), EC: 592 (10) cm3 p=0.004), although AA did not differ significantly from EA and EC. Lumbar spine BMD was significantly lower in AA than EC (1.141 (0.019), AA: 1.105 (0.026) EC: 1.188 (0.014) g/cm2, p=0.007). All differences persisted after adjustment for height, age, and body mass. Eumenorrhoeic athletes had significantly higher femoral neck BMD and Z than controls, consistent with previous research. Femoral neck Z and hence strength in bending was relatively maintained in athletes with menstrual dysfunction despite their lower BMD at this site, indicating possible structural adaptation. Incidence of stress fracture was determined prospectively. Following withdrawal of 9 participants, 61 female athletes were monitored prospectively for the 12-month period. Among the 61 athletes, two sustained a stress fracture, both diagnosed by MRI, giving an annual incidence rate of 3.3%. The stress fracture cases were: both 800m runners aged 19 and 22 years, training on average 14.2 hours/week, eumenorrhoeic, and with no history of amenorrhoea. BMD, energy intake and EDE-Q and CET scores were similar to the mean values in the non-stress fracture group. Thus the incidence of stress fracture in this sample of female endurance athletes is lower than previously reported, possibly due to the increased awareness of stress fracture diagnosis, risk factors and athlete management. Seasonal bone changes were determined in 61 female athletes. The greatest variation was observed in the endurance runners (n=52). The endurance runners were classified according to menstrual function (28 were EA and 24 AA). There were no significant differences at baseline or seasonal variation in height, weight, and body fat percentage. In EA, trochanter BMD increased over the summer (0.885 (0.019) to 0.947 (0.177) g/cm2, p=0.002) with no significant change over the winter (0.880 (0.018) to 0.885 (0.018) g/cm2 p=0.153). In AA femoral neck BMD decreased over the winter (1.065 (0.021) to 1.052 (0.020), g/cm2, p= 0.030) with no significant change over the summer (1.050 (0.020) to 1.052 (0.020), g/cm2, p=0.770). Minimal neck width increased in the group as a whole over the winter (28.4(0.3) to 28.7(0.3), mm p=0.039) with no significant change over the summer 28.8(0.3) 28.7(0.3), mm p=0.333). There were no significant seasonal variations in other bone parameters, and seasonal changes did not differ significantly between groups. EA increased trochanter BMD over the summer, and this was maintained over the winter. Conversely, AA lost femoral neck BMD over the winter and this was not recovered over the summer, although the increase in width of the femoral neck may have partly compensated BMD loss to maintain strength in bending. The final prospective analysis was conducted in a separate sample of female athletes who were diagnosed with a stress fracture injury. The aim of this analysis was to determine the magnitude and time scale of bone loss following a stress fracture injury and subsequent regain following retaining. A group of 4 stress fracture cases and 3 controls were followed for a period of 6-8 months following a stress fracture injury. BMD and BMC (lumbar spine, femoral neck, and trochanter) and estimations of geometric properties CSA, Z and buckling ratio) were assessed using DXA. The mean difference of bone loss and bone regain was determined by BMD, BMC and geometric parameters from baseline to 6-8 weeks and 6-8 weeks to 6-8 months respectively. No significant bone loss was found in either cases or controls from baseline to 6-8 weeks at any of the bone parameters. A significant difference at the femoral neck was found in the injured leg of the stress fracture cases from 6-8weeks to 6-8months (mean (SE) 1.042(0.102) to 1.070(0.102) g/cm2, p=0.004) with no significant change in the contra-lateral case leg 1.036 (0.102) to 1.054(0.109) g/cm2). No significant bone regain was found in the control subjects (health or injured legs ). Thus athletes do not seem to lose significant BMD during the recovery phase of training when partial weight bearing is required. Subsequent bone regain above the initial baseline value does seem to occur in the injured leg within 8 months following the stress fracture once training is resumed. In conclusion the work within this thesis has not only reinforced previous stress fracture findings, showing that a history of stress fracture is increased in athletes with a history of amenorrhoea, but has identified novel results indicating a lower incidence of stress fracture in female endurance athletes than previously reported. Exercise cogn