The authors state that they have no conflicts of interest.
Calcium and Vitamin D Supplementation Decreases Incidence of Stress Fractures in Female Navy Recruits†
Article first published online: 4 FEB 2008
Copyright © 2008 ASBMR
Journal of Bone and Mineral Research
Volume 23, Issue 5, pages 741–749, May 2008
How to Cite
Lappe, J., Cullen, D., Haynatzki, G., Recker, R., Ahlf, R. and Thompson, K. (2008), Calcium and Vitamin D Supplementation Decreases Incidence of Stress Fractures in Female Navy Recruits. J Bone Miner Res, 23: 741–749. doi: 10.1359/jbmr.080102
- Issue published online: 4 DEC 2009
- Article first published online: 4 FEB 2008
- Manuscript Accepted: 28 JAN 2008
- Manuscript Revised: 1 JAN 2008
- Manuscript Received: 23 MAY 2007
- physical training;
- military training;
- fracture prevention;
- young adult
Introduction: Stress fractures (SFx) are one of the most common and debilitating overuse injuries seen in military recruits, and they are also problematic for nonmilitary athletic populations. The goal of this randomized double-blind, placebo-controlled study was to determine whether a calcium and vitamin D intervention could reduce the incidence of SFx in female recruits during basic training.
Materials and Methods: We recruited 5201 female Navy recruit volunteers and randomized them to 2000 mg calcium and 800 IU vitamin D/d or placebo. SFx were ascertained when recruits reported to the Great Lakes clinic with symptoms. All SFx were confirmed with radiography or technetium scan according to the usual Navy protocol.
Results: A total of 309 subjects were diagnosed with a SFx resulting in an incidence of 5.9% per 8 wk. Using intention-to-treat analysis by including all enrolled subjects, we found that the calcium and vitamin D group had a 20% lower incidence of SFx than the control group (5.3% versus 6.6%, respectively, p = 0.0026 for Fisher's exact test). The per protocol analysis, including only the 3700 recruits who completed the study, found a 21% lower incidence of fractures in the supplemented versus the control group (6.8% versus 8.6%, respectively, p = 0.02 for Fisher's exact test).
Conclusions: Generalizing the findings to the population of 14,416 women who entered basic training at the Great Lakes during the 24 mo of recruitment, calcium and vitamin D supplementation for the entire cohort would have prevented ∼187 persons from fracturing. Such a decrease in SFx would be associated with a significant decrease in morbidity and financial costs.
Physical training and exercise during early adulthood is known to increase bone strength.[1-4] Therefore, intense training for young athletes or military recruits provides an opportunity to maximize bone strength or resistance to fracture. Paradoxically, prolonged intensive training can result in fatigue damage and stress fractures. Stress fractures are among the most common and debilitating overuse injuries seen in military recruits, with fracture rates ranging from 0.2% to 5.2% of male recruits and 1.6% to 21.0% of female recruits.[5-10] Stress fracture rates for female recruits are consistently higher than for men. Stress fractures are also problematic for nonmilitary athletic populations. The highest incidence rates, ranging from 10% to 31%, are seen in members of track and field teams.[8, 11-13]
Stress fractures occur when bones are repetitively loaded over short periods without sufficient time for repair and are seen most often among persons who are involved in physical activity to which they are not adapted. Stress fractures are a major concern for the armed forces because they result in morbidity ranging from pain to permanent disability. In addition, these injuries incur considerable expense for the military. At one U.S. Army training base, recent estimates were that over the span of 1 yr, $26 million was lost in training costs for the 749 soldiers who were discharged from training, an average of more than $34,000 per soldier. In the U.S. Army, 40% of men and 60% of women trainees with stress fracture do not complete basic training. Thus, stress fractures also negatively affect military readiness.
In recent years, the U.S. Military Services have decreased the incidence of stress fractures in recruits by making changes to basic training. At the Great Lakes Naval Station, these changes included different types of boots and training shoes, strict guidelines for physical training (PT), and shorter stride lengths while marching in formation. Although these training guidelines have reduced stress fractures, a significant number of fractures still occur. The remaining fractures may be caused by high risk at entry associated with genetic or lifestyle factors or poor skeletal adaptation to physical training.
Numerous studies have found genetic or lifestyle factors that significantly predict stress fracture.[7, 8, 15-17] However, for the most part, the predictor variables have low sensitivity and specificity. Although these and other risk factors may be useful for identification of modifiable variables, they are not useful for identifying individual recruits at highest risk for stress fracture. Thus, the military has no accurate method of selecting individual recruits to target for stress fracture prevention. Because prevention programs add cost to training, the military services are reluctant to implement such programs unless they can be highly targeted. Alternatively, interventions that can safely and effectively be applied to all recruits are desirable.
Strong rationale exists for a calcium and vitamin D intervention in Navy recruits and other similar groups. Calcium is essential for bone mineralization, and a positive calcium balance is necessary for maximal bone adaptation to mechanical loading. There are five primary reasons that calcium and vitamin D supplementation may be important for young women during basic military training: (1) women <30 yr of age have not achieved peak bone mass and have the potential to gain bone, which requires that they be in positive calcium balance[1, 18]; (2) intense training stimulates bone formation increasing calcium demands[19-21]; (3) microfracture repair through remodeling requires calcium for bone formation[22, 23]; (4) substantial cutaneous calcium losses can occur during training[24-26]; and (5) calcium and vitamin D nutrition of young women is typically less than optimal.[27, 28]
Calcium and vitamin D supplementation is widely used to prevent osteoporotic fractures in older individuals, and it is safe and generally well tolerated. Thus, the goal of this study was to determine whether a calcium and vitamin D intervention could reduce the incidence of stress fractures in female recruits during basic training.
MATERIALS AND METHODS
We conducted a double-blind, placebo-controlled, randomized clinical trial in female Naval basic trainees. During ∼24 mo of study recruitment between May 2001 and March 2006, 14,416 women entered basic training at the Great Lakes Naval Training Center, Great Lakes, IL, USA. All of the U.S. Navy enlisted personnel receive their basic training at this location. The recruits were approached during one of the first days of processing and invited to participate in the research project. Subjects included 5201 female recruits who volunteered for the study (Fig. 1). All subjects were healthy based on entry physical examinations. The project was approved by three Institutional Review Boards: Creighton University, Naval Institute for Dental and Biomedical Research, and the U.S. Army Medical research and Materiel Command. All study participants gave written informed consent.
At baseline, participants completed a risk factor questionnaire, and during the study, they maintained a record of menstrual periods to ascertain amenorrhea and use of contraceptives. The risk factor assessment was done to determine presence of factors other than properties of bone that are known to contribute to fracture risk: dairy food consumption, previous fractures, family history of osteoporosis, current or past smoking, regular weight-bearing exercise, and use of corticosteroid medication and contraceptives. Dairy food consumption was described as having one or more servings per day of dairy foods. Contraceptive use was defined as having ever used any type of contraceptive (oral, patch, ring, implant, or injection) for >3 mo. Weight-bearing exercise was defined as having participated in activity such as walking or running at least 3 times/wk. Subjects who reported smoking were asked to record the number of years they had smoked and the number of packs per day. Similarly, those who reported daily dairy food consumption, regular weight-bearing exercise, or use of contraceptives were asked to record the length of time. Last, the recruits were asked to estimate the number of alcoholic drinks per week.
Height and weight were self-reported on a subset of the recruits. Body mass index (BMI) was calculated by dividing body weight in kilograms by the square of body height in meters. Fitness at entry to basic training was measured as a 1.5-mi entry run time. During the first week of basic training, each recruit is required to complete a 1.5-mi run in running shoes, and the time to complete that run is recorded. This is done to obtain an estimate of the physical fitness of the recruit. We were able to obtain that value for a subset of the recruits.
As individuals consented to the study, they were randomly assigned to the two intervention groups by block randomization. That is, the first two participants enrolled were randomly assigned to one of the two groups, then the next two, and so on. The randomization scheme was known only to the study statistician until the study end. The statistician was responsible for randomization of treatment by subject number and was the only person unblinded.
The study groups were as follows: treatment (2000 mg calcium and 800 IU vitamin D/d) or control (identical placebo). During the 8 wk of training, subjects were asked to take four supplement pills per day made available in the galley line at breakfast and dinner. To monitor pill taking, project staff observed the galley food lines, visited recruits in their quarters, and conducted an exit interview.
Because recruits within training divisions were randomly assigned to treatment group, both groups were subject to the same training conditions. Thus, even though some training changes occurred during the course of the study, such as a change in boots, subjects in each treatment group experienced any given change in approximately the same numbers. Recruit schedules during basic training are tightly regulated and include 1 h of physical training a day, 6 d/wk. In addition, they walked on base to and from classes and the food galleys. During the seventh week of training, recruits' knowledge, skills, strength, and endurance are tested during “Battle Stations,” in a rigorous training exercise that simulates a combat situation.
The Great Lakes is located at latitude 41°N, and at that latitude very little cutaneous conversion of vitamin D takes place between the months of September and March. Unfortunately, serum 25 hydroxyvitamin D was not measured in our study. We attempted to include that into the study design, but drawing blood was not feasible because of our limited access to the recruits. However, during each season, subjects were randomly assigned to each treatment group.
Stress fracture diagnosis
Stress fractures were ascertained when recruits reported to the Great Lakes clinic with symptoms. All stress fractures were confirmed with radiography or technetium scan according to the usual Navy protocol.
Subjects were characterized by medians and ranges of individual characteristics. Fisher's exact test and the Cochran-Mantel-Haenszel (C-M-H) test were used to test and estimate differences in fracture incidence between the two treatment groups. For the main intention-to-treat (ITT) analysis, relative risk (RR) and 95% CIs were estimated with the C-M-H approach using fracture status (yes versus no) as the response variable and treatment group (calcium + vitamin D versus placebo) as the factor.
We also performed a number of secondary analyses: (1) per-protocol (PP) with the Fisher exact and the C-M-H tests; (2) several ITT univariate analyses to determine the effect of individual risk factors; and (3) ITT with a single logistic regression model to determine the effect of categorical covariates while adjusting for treatment, including first-order interaction effects. The ORs generated by the logistic regression model are good approximations of the respective RRs because fracture is a rare event in our study population (i.e., fracture incidence < 10%). The categorical covariates adjusted for in the analyses were: history of exercise sessions per week (high, ≥3 versus low, <3), use of depomedroxyprogesterone (Depo; yes versus no), menses during basic training (yes versus no), smoking ever (yes versus no), age group (<25 versus ≥25 yr), and run time per 1.5 mi at entry of basic training (fast, <15 min versus slow, ≥15 min).
A Forest plot of the RR or OR with 95% CI was generated to show (1) the effect of treatment with calcium and vitamin D from the main analysis; (2) the RR of each individual covariate adjusted for treatment; and (3) the effect of all covariates adjusted for treatment in a single logistic regression analysis. A level of significance of p < 0.05 and the statistical package SAS 9.1 (SAS Institute, Cary, NC, USA) were used in all analyses.
Subject baseline characteristics are presented in Table 1. There were no significant differences between treatment groups in any of these characteristics. The two progesterone preparations are contraceptives that suppress ovulation and estrogen production by the ovary. During basic training, 2786 subjects reported having no menstrual periods (1391 in the calcium and vitamin D group and 1395 in the placebo group). The remainder reported having one or more periods. The median dairy intake was <1 serving/d, which provides ∼300 mg of calcium.
The racial/ethnic distribution of the sample was as follows: American Indian, 3%; Asian, 4%; non-Hispanic black, 18%; Hispanic, 13%; non-Hispanic white, 54%; other, 4%. The racial/ethnic distribution did not differ between treatment groups.
Of the 5201 subjects enrolled, 365 (7%) were discharged from the Navy before the end of training, and an additional 1136 (21.8%) withdrew from the study (Fig. 1). Therefore, 3700 recruits completed 8 wk of study. Adverse events include any reported symptoms associated with withdrawal from the study. The primary events were gastrointestinal disruption such as constipation, diarrhea, upset stomach (4%), and musculoskeletal soreness (0.9%). Roughly 16% of the subjects quit because they changed their mind, forgot, it was too difficult, or for no reported reason. A small number of subjects withdrew because they were prescribed calcium by Great Lakes medical personnel (1.5%). All enrolled subjects were included in the ITT analyses. We were able to determine fracture status in subject who withdrew from study but remained on base. We were not able to ascertain fracture status from subjects after they left the military. Subjects who withdrew from study for any reason were excluded from the per protocol analyses.
Three hundred nine subjects were diagnosed with a stress fracture resulting in an incidence of 5.9% per 8 wk. Using ITT analysis by including all enrolled subjects, we found that the calcium and vitamin D group had a 20% lower incidence of stress fractures than the control group (5.3% versus 6.6%, respectively; p = 0.026 by Fisher's exact test), with RR = 0.80, 95% CI = 0.64–0.99 from the C-M-H test (Fig. 2A). In a per protocol analysis (including only the 3700 recruits who completed the study), there were 21% fewer fractures in the supplemented versus the control group (6.8% versus 8.6%, respectively; p = 0.020 for Fisher's exact test).
The 309 fracturing individuals sustained 496 fractures (Table 3). Most of these fractures were in the tibia or fibula, but 53 were major fractures of the femur or pelvis. The largest relative differences in fracture numbers between treatment and placebo groups were at the tibia/fibula and pelvic sites, but, individually, these differences were not statistically significant. Also, there were no statistically significant differences in fracture incidence among the racial/ethnic groups. The median time to fracture for the entire cohort was 36 days, for the placebo group, it was 34 days, and for the calcium and vitamin D group, it was 37 days.
Other risk factors for stress fracture
We performed secondary analyses aimed at determining factors that might increase the risk of fracture in this cohort of female military recruits. The specific factors were selected because they are known to contribute to stress fracture risk. Univariate analyses showed the following factors to be statistically significant predictors of fracture: age, amenorrhea during basic training, entry run time, and history of smoking, regular weight-bearing exercise, and Depo use. A Forest plot is used to show (1) the effect of treatment with calcium and vitamin D from the main analysis; (2) the RR of each individual covariate adjusted for treatment; and (3) the effect of all covariates adjusted for treatment in a single logistic regression analysis (Fig. 2).
The risk of fracture in those having amenorrhea was 91% higher than those with one or more menstrual periods during training (RR, 1.91; 95% CI = 1.47–2.47; p < 0.0001). The risk remained significant when controlled for Depo use. When adjusted for treatment, the risk of fracture in those having amenorrhea was decreased to 83% (p = 0.0035; Fig. 2).
In a univariate analysis using age as a continuous variable, every year of age increased the risk of fracture by 5% (RR = 1.05; 95% CI = 1.011–1.090; p = 0.012). Older recruits (age > 25 yr) were at a 60% higher risk of fracture than their younger counterparts when treatment was included in the model (p = 0.013; Fig. 2B).
The univariate analysis showed that Depo users had a 48% greater risk of fracture than nonusers (RR, 1.48; 95% CI = 1.18–1.97; p = 0.006). Those who had history of longer use had greater odds for fracture per year of Depo use than those with shorter use (RR = 1.063; 95% CI = 1.01–1.119; p = 0.0188). There was no significant difference in fracture incidence between those who used other contraceptive types and nonusers. Adjusting for treatment, those who had a history of Depo use had a 45% higher risk of stress fracture than nonusers (p = 0.0057; Fig. 2B).
In the univariate analysis, the risk of fracture in recruits with a history of smoking was higher than for those who had not smoked (RR = 1.44; 95% CI = 1.096–1.695; p = 0.009). Those who reported a greater number of pack years of smoking had a greater risk than those with fewer pack years (per pack year, RR = 1.044; 95% CI = 1.004–1.087; p = 0.03). Adjusting for treatment, recruits with history of smoking had a 41% higher risk of fracture than those who had not smoked (p = 0.0075; Fig. 2B).
The median run time of the recruits in this study was 15.5 min. The slowest recruit's run time was 23 min. In a univariate analysis, the RR of fracture was 1.35 for each minute above the median time (95% CI = 1.25–1.45; p = 0.001). Calculations indicate that the risk of fracture in the slowest recruit was almost 10 times the risk in the persons with an average run time of 15.5 min. When controlled for the effect of treatment, the fast runners (<15 min/1.5 mi) had a 22% smaller risk of incident fracture than the slow runners (p = 0.0373; Fig. 2B).
In a univariate subanalysis within the placebo group, those with a history of high exercise had a 34% lower risk of stress fracture than the low exercise group (5.13% versus 7.75%; RR = 0.66; 95% CI = 0.49–0.90; p = 0.008). This exercise history effect was not seen in the supplemented group. Adjusting for treatment, subjects who were in the high exercise (≥3 times/wk) group had a 30% lower risk of stress fracture than those who reported less activity (p = 0.004; Fig. 2B).
Dairy food consumption, alcohol use, previous fractures, race/ethnicity, and family history of osteoporosis were not significantly predictive of fracture in this cohort. Also, height, weight, and BMI were not associated with fracture risk, but the sample sizes were small because height and weight were not available for all of the subjects.
In a single logistic regression model, we tested the treatment effect and the effects of the categorical covariates and their first-order interactions (Fig. 2C). None of the first-order interactions were even marginally statistically significant (p > 0.15; i.e., none of the covariates, such as amenorrhea, interacted with each other). Thus, we do not report them here. In this model, the factors that continued to be significant predictors of increased fracture risk were amenorrhea during basic military training (OR = 1.862; 95% CI = 1.416–2.450; p < 0.0001), with 86% increase of incident fracture risk, and age > 25 yr (OR = 1.658; 95% CI = 1.072), with 66% increase in fracture risk. The only factor that was significantly protective for fracture was history of exercise. Subjects who were in the high exercise (≥3 times/wk) group had a 34% lower risk of stress fracture than those who reported less activity (OR = 0.662; 95% CI = 0.510–0.859; p = 0.002).
The rest of the factors in the combined model were marginally significant but pronounced. Thus, Depo users had a 24% higher risk of stress fracture (OR = 1.241; 95% CI = 0.911–1.689; p = 0.1713), the fast runners (<15 min/1.5 mi) had 20% smaller risk (OR = 0.811; 95% CI = 0.623–1.056; p = 0.119), smokers had a 32% higher risk (OR = 1.319; 95% CI = 0.993–1.751; p = 0.0561), and those on treatment had 20% smaller fracture risk (OR = 0.789; 95% CI = 0.616–1.01; p = 0.0589). Because these effects do not interact, and the ORs are good approximations of the respective RRs, we could use these results to calculate an approximate RR of fracture for the worst case (amenorrhea, age > 25 yr, no exercise, Depo user, slow runner, smoker, no treatment) compared with the best case:
This means that the worst-case scenario has 12 times larger fracture risk than the best-case scenario. The relative risk for any other two scenarios can be calculated similarly.
To our knowledge, this is the first randomized controlled trial of the efficacy of calcium and vitamin D supplementation that has shown a decreased incidence of stress fracture. Schwellnus and Jordaan studied the effect of calcium supplementation in preventing stress fractures in male military recruits in South Africa but found no statistically significant effect. That study was underpowered because the sample size included only 250 recruits given supplementation, and the calcium supplementation was low (500 mg/d). Only 14 stress fractures were diagnosed over the 9-wk training period (incidence of 1.4% in the control group and 0.64% in the supplement group).
It is well established that adequate calcium nutrition is essential for skeletal strength. When serum calcium levels fall because of insufficient intake or excessive loss of calcium, PTH production is increased to stimulate bone resorption and liberate calcium from the skeleton to restore serum calcium concentration.[30, 31] This has been called nontargeted (stochastic) remodeling[22, 23] and can persist over extended periods of time, as long as the stress to plasma calcium persists. The bone response to the increased PTH is to increase the rate of activation of new remodeling sites, thus increasing the remodeling space. This results in an imbalance of bone remodeling in which more bone is lost from the skeleton than is replaced, and skeletal strength is compromised. Under conditions of intense mechanical loading, such as occurs during basic military training, microfractures develop and lead to targeted remodeling to repair the microfractures.[22, 23, 32] In this situation, optimal levels of circulating calcium are needed to provide substrate for repair of microdamage and to inhibit an increase of nontargeted remodeling to maintain serum calcium concentration.
High levels of physical activity in the presence of low (or even moderately plentiful) calcium intake can cause additional stress on the skeleton because of the need to offset the substantial cutaneous calcium loss in the sweat. Under conditions of heavy sweating and insufficient calcium intake, calcium is drawn from the bone reservoir under the influence of elevated levels of PTH. In fact, acute bouts of exercise increase PTH levels proportional to exercise intensity.[30, 33-36] The cutaneous calcium losses during heavy physical activity can be substantial,[24-26] and the resultant secondary hyperparathyroidism can weaken the skeleton even over short periods. In that regard, Thorsen et al. reported that young women showed increased bone turnover, decreased serum ionized calcium, and increased serum PTH after a single bout of moderate endurance exercise. This calcium stress may act to limit skeletal adaptation and repair mechanisms in both military recruits and athletes during strenuous activity.
Klesges et al. measured cutaneous calcium loss in 11 members of the University of Memphis' men's basketball team during a 10-day training period. Cutaneous calcium loss averaged 422 mg per each 2-h training session. Average total body BMC decreased by 3.8% from preseason to midseason (p = 0.02), a period of 4 mo, whereas leg bone mass decreased 6% (p = 0.01). From preseason to late summer, the players lost 6.1% of their total BMC and 10.5% of their leg bone mass (p = 0.001 and p < 0.001, respectively). Individual players lost as much as 20.2% of their leg bone mass during this 10.5-mo interval. During year 2 of the study, the athletes were supplemented with calcium and vitamin D at doses based on each individual's BMC loss in year 1. Supplementation not only stopped losses in BMC, but increased total BMC by almost 2% and leg BMC by 3% by midseason of year 2 (p = 0.04 and 0.05, respectively). This study showed that sweat loss contributes to a considerable loss of BMC in young men who should not be losing bone. Furthermore, sufficient dietary calcium can offset cutaneous losses and allow adequate bone adaptation. The bone loss and subsequent gain with supplementation in the study of Kleges et al. were seen over a relatively short period of time, ∼4 mo. In our study at the Great Lakes, supplementation prevented stress fracture over ∼2 mo.
Although no studies were found using vitamin D supplementation to decrease stress fracture incidence, Ruohola et al. found that low baseline 25 hydroxyvitamin D [25(OH)D] predicted stress fracture in Finnish male military recruits. In that study, 800 randomly selected recruits (mean age, 19 yr) were followed prospectively over 90 days, which included 8 wk of basic training. Twenty-two recruits with stress fracture were identified (2.9%). In the final multivariate analysis, a significant risk factor for stress fracture was serum 25(OH)D below the median level of 75.8 nM (OR = 3.9; 95% CI = 1.2–11.1; p = 0.002). It is well established that a small reduction in vitamin D status can contribute to mild increase in PTH concentration because vitamin D, in the form of 1,25 dihydroxyvitamin D, regulates the active transport mechanism of calcium absorption from the gut.[38, 39] Bone turnover increases in response to the elevated serum PTH levels. An earlier study of Finnish male recruits found that high serum PTH levels were associated with stress fracture. These reports support our finding that supplementation with vitamin D can decrease stress fracture incidence.
Supplementation with calcium and vitamin D significantly reduced the risk of stress fracture overall despite the negative effects of several lifestyle factors. The supplementation was well tolerated. In fact, the percent of subjects withdrawing from study because of adverse events was 4% in each of the groups (placebo and treatment). These findings lend confidence that supplementation with calcium and vitamin D is a viable option to for preventing stress fractures, even in populations with factors that significantly increase the risk of fracture.
It is important to note that several factors increase the risk of stress fracture in these female recruits despite treatment: amenorrhea during training, age > 25 yr, and history of smoking, low levels of physical activity, and Depo use. Also, those who were less physically fit at the beginning of training had a higher risk of fracture than their more fit counterparts.
Poor physical fitness is widely reported to be associated with stress fracture.[8, 9, 40-42] In our earlier study of female Army recruits, we found that nonexercisers in the lowest quintile of quantitative ultrasound speed of sound (SOS) had nearly a nine times greater risk of fracture than did nonexercisers in the highest quintile. Thus, strong evidence supports the importance of pretraining activities to improve the fitness of young women gradually before initiating intense physical activity programs. In our Naval study, it is interesting to note that, within the placebo group, the high exercise group had a 34% lower risk of stress fracture than the low exercise group, whereas this exercise history effect was not seen in the supplemented group. Numerous studies of military recruits have found that a history of regular physical activity decreases risk of stress fracture.[7-9, 16] Our findings suggest that calcium/D supplementation can somewhat compensate for a history of low physical activity, which is prevalent in U.S. youth,[43, 44] the group entering the military services.
Although supplementation lowered the risk somewhat, recruits with a history of smoking still had a 41% higher risk of fracture than those who had never smoked. Other researchers have reported that smoking increases risk of stress fracture.[15, 45, 46] Furthermore, numerous investigators have reported an inverse relationship between BMD and smoking.[47-51] In our previous study, we found that female recruits in the lower quintile of SOS who did not exercise and smoked had a risk of fracture nearly 12 times greater than those with higher SOS values who exercised and did not smoke. Krall and Dawson-Hughes reported that smokers lost bone more rapidly than nonsmokers and had significantly lower calcium absorption, suggesting that poor absorption of calcium may contribute to the faster rate of bone loss. In a meta-analysis of the effects of cigarette smoking on bone, Ward and Kleges found that smokers had significantly lower bone mass than nonsmokers at all skeletal sites, and the difference was dose dependent. Thus, history of smoking in female recruits should serve as a “red flag” indicating high risk of sustaining a stress fracture during basic training.
As we found in an earlier study at a U.S. Army training base, women who had a history of using the long-acting progesterone contraceptive, Depo, had a higher risk of fracture than those who had not used Depo. Similar to other stress fracture studies,[8, 9, 46] women at the Great Lakes who reported no menses during basic training had a significantly greater risk of stress fracture than those with one or more menstrual periods. This risk remained significant when controlled for Depo use and treatment group. Absence of menses, whether because of Depo use or other causes, is a reflection of very low circulating estrogen levels, which are associated with lower bone mass[54, 55] and higher risk of fracture.[56, 57]
On average, the Navy recruits in our study had suboptimal calcium intake: 300 mg/d compared with the recommended 1000 mg/d for young women. Thus, it is not surprising that calcium supplementation reduced their risk of stress fracture. As would be expected, our study showed that young women enter basic military training with a variety of risk factors for stress fractures in addition to low calcium intake, such as smoking history, use of progesterone contraceptives, poor physical fitness, etc. The effect of calcium and vitamin D supplementation was strong even when adjusting for these factors. Thus, it seems prudent to provide supplementation to all female recruits.
The incidence of stress fracture in our cohort of young women was 5.9%. This is lower than other reports and may reflect the successful efforts of the Navy to change the skeletal demands of basic training. Our study shows that supplementation with calcium and vitamin D can provide additional benefits to the successful training changes that have already been implemented by the armed forces.
Although there were only 11 fractures of the pelvis and pubis, there were more than twice as many in the placebo group as in the calcium and vitamin D group. The difference between treatment groups was not statistically significant because of the small numbers of fractures. However, the difference is medically significant considering the morbidity and potential disability, as well as the medical treatment costs, associated with fractures at these sites.
One limitation of this study was that 1136 (21.8%) of the subjects withdrew from the study. However, ∼7% of those withdrew so that they could take calcium and vitamin D supplementation provided by Great Lakes medical caregivers and not risk being on placebo. Discounting that group, the attrition is 20%, which is congruent with many clinical trials.
In conclusion, calcium and vitamin D substantially reduced the incidence of stress fractures by 20% in female Naval recruits in this study. Generalizing the findings to the population of 14,416 women who entered basic training at the Great Lakes during the study period, calcium and vitamin D supplementation for the entire cohort would have prevented ∼187 persons from fracturing. Such a decrease in stress fracture would be associated with a significant decrease in morbidity and financial costs. Supplementation with calcium and vitamin D provides a safe, easy, and inexpensive intervention that does not interfere with training goals.
This sutdy was funded by Department of Defense Bone Health and Military Readiness Grant DAMD-17-01-1-0807. The calcium and vitamin D supplement was provided by GSK Consumer Healthcare, Parsippany, NJ, USA. We thank our project staff members, Denise Berry, Vikki Morales, Sharon Osler, and Mary Dierks, and the Navy recruits, Recruit Training Instructors, and officers at the Naval Institute for Dental and Biomedical Research, without whom we could not have conducted this study.
- 92005 Predictors of stress fracture susceptibility in young female recruits. Am J Sports Med 16: 1–8., , , ,
- 14U.S. Army Research Institute of Environmental Medicine Bone Health and Military Medical Readiness. 2006. U.S. Army Research Institute of Environmental Medicine, Natick, MA, USA.
- 271994 Energy and macronutrient intakes of persons ages 2 months and over in the United States: Third National Health and Nutrition Examination Survey, Phase 1, 1988–91. National Center for Health Statistics, Hyattsville, MD, USA., , , , , , ,
- 321960 Presence of microscopic cracks in vivo in bone. Bull Henry Ford Hosp 8: 25–35.
- 381985 Calcium, bone metabolism, and structural failure. Triangle 24: 91–100.,
- 391999 Preclinical safety profile of alendronate. Int J Clin Pract Suppl 100: 3–8.,
- 43CDC 1996 Youth risk behavior surveillance-United States. MMWR Morb Mortal Wkly Rep 45: SS-4.
- 441990 Exercise and Children's Health. Human Kinetics, Champaign, IL, USA.