Burden and risk factors for Achilles tendon rupture in the military population from 2006 to 2015: A retrospective cohort study

INTRODUCTION

The Achilles tendon, also known as the calcaneal tendon, has an important role in propulsion, force attenuation, and energy return during walking, running, and jumping during function. Achilles tendon rupture (ATR) is a serious musculoskeletal injury that results in substantial functional decline, especially in highly physically demanding occupations such as service in the military.¹ In the general public, ATR affects an estimated 32 906 individuals in the US per year, at an overall rate of 2.1 per 100 000 person-years.² Some of the risk factors associated with ATR include previous musculoskeletal injury, chronic Achilles tendon inflammation and tendinopathy, regular participation in athletic activity, spring season, genetic factors, chronic systemic conditions (diabetes, hyperparathyroidism, and renal failure), and medication use (steroids, quinolones, and oral bisphosphonate).³ ATR management includes both conservative and surgical approaches, with return to sport occurring in 80% of affected athletes⁴ and return to duty in 90-100% of affected military members.^{1, 5} Among military members, rehabilitation takes an average of 8.2 months in nonoperative cases¹ and 5.4 to 6.7 months following surgery^{1, 5} before they are cleared for return to full, unrestricted duty.

In an earlier study of the ATR burden in US military members, 1441 ATR cases were identified between 1998 and 2001, occurring at a rate of 30.9 per 100 000 person-years.⁶ The 15-fold higher burden in the military may in part be due to the unique exposures incurred during military service. In a more recent study of care episodes for Achilles tendinopathy in the military, the prevalence was 17.65 per 1000 in officers and 12.22 per 1000 in enlisted members, with male sex, older age, senior rank, military occupation, and service branch found to be salient associated factors.⁷ It is highly plausible that occupation may also be a salient factor for ATR, given that tendinopathy may be a prodromal sign of future tendon failure. Due to the changes in operational requirements, training, engagement in overseas contingency operations, and force composition that have occurred over the past 22 years, an updated assessment of burden and the associated risk factors of ATR is warranted. Therefore, the purpose of this study was to determine the burden of ATR in the military population with consideration of military occupation and identify risk factors that may contribute to this condition.

METHODS

A population-based epidemiological retrospective cohort study of all service members in the US Armed Forces was performed, assessing sex, military occupation, rank, branch, and year on ATR prevalence from 2006 to 2015. The Defense Medical Epidemiology Database (DMED; Defense Health Agency, Falls Church, VA, https://bit.ly/DHADMED5) was utilized to identify relevant healthcare encounters. This database provides aggregated data for International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes and de-identified patient characteristics, including sex, categories of military occupations, and branch of service for all active duty and reserve military service members. The database is HIPAA compliant, does not include any personally identifiable or personal health information, and has been used previously for the epidemiological study of ankle-foot injury in the military.^{8, 9} This study was approved as non–human-subjects research by the Institutional Review Board . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines¹⁰ were used to guide reporting.

The database was queried for the number of distinct care episodes for a primary diagnosis of ATR (ICD-9-CM codes 727.67, nontraumatic rupture of the Achilles tendon, and 845.09, other sprains and strains of ankle) from 2006 to 2015. Prevalence of ATR care episodes was calculated for male and female military members, enlisted and officers, in each service branch (Army, Navy, Marine Corps, and Air Force) and occupational category.^{8, 9} Prevalence ratio (PR) point estimates and 95% confidence intervals (CIs), risk difference point estimates, attributable risk, number needed to harm, and chi-square statistics were calculated in the assessment of sex and occupation category using Microsoft Excel for Mac 2016 (Microsoft Corp., Redmond, WA) and a custom epidemiological spreadsheet.¹¹ In the unadjusted assessment of sex and occupation on ATR risk, male service members, enlisted infantry, and ground/naval gunfire officer groups were used as the reference categories.

Due to overdispersion of the data, a negative binomial model was employed over a Poisson model for the adjusted assessment of female sex (male reference), branch (US Army reference), officer rank (enlisted reference), and year (2006 reference) on the risk of ATR.¹² Since there were excess zeros noted in the data, with many demographic categories not having any reported outcomes of interest (especially among the smaller subpopulations), assessment of convergence between the unadjusted and adjusted (hurdle) negative binomial regression models were performed.⁷ In the hurdle negative binomial regression model, a logistic link is employed to remove demographic categories that contributed no episodes of ATR from the final count model. By doing so, excess zeros that contribute to skewed point estimates, standard errors, and overdispersion are parsed in the truncated regression model.¹² Results of hurdle negative binomial regression are reported using calculations of the predictors regressed on count data (count model assessing the rate of the outcome) as well as a linked logistic regression (zero model that assessed the probability within demographic categories of not having the outcome).

Both unadjusted and adjusted models are reported in the supplemental material. The regression analyses were performed using the PSCL (version 1.5) and MASS (version 7.3-58.1) packages on R (version 3.5.1, The R Foundation for Statistical Computing, Vienna, Austria). The level of significance was p < 0.05 for all analyses. PR point estimates were considered statistically significant if CIs did not cross the 1.00 threshold. Convergence of p values, effect sizes, and 95% CIs were considered when evaluating significant findings.

RESULTS

Supplemental tables 1 and 2 detail the counts and prevalence of ATR in officer and enlisted personnel. In the study epoch, officers incurred 15 978 episodes of ATR at a prevalence of 7.43 per 1000 (male officers: 8.11 per 1000; female officers: 3.89 per 1000). Among enlisted personnel, 59 242 ATR episodes occurred at a prevalence of 6.23 episodes per 1000 (male enlisted: 6.49 per 1000; female enlisted: 4.48 per 1000). Table 1 details the univariate factor of sex within occupational specialty on ATR episodes. Apart from enlisted aviation specialists (with no significant difference in risk was between men and women), both female officers and enlisted service members had significantly lower risk of ATR compared with their male counterparts in all occupations (PR: 0.26-0.73).

In the assessment of occupational exposure, aviation and service officers demonstrated significantly lower risk of ATR episodes (PR: 0.87-0.91) and administration, operations, intelligence, and logistic officers demonstrated increased risk (PR: 1.16-1.31) compared with ground and naval gunfire officers (Table 2). There was no significant difference in PR between engineering and maintenance officers and the reference group. Among enlisted specialties, all but mechanized/armor and combat engineers had significantly higher ATR risk compared with infantry (PR: 1.14-2.13), with the highest risk observed in the administration, intelligence, and communication fields.

Table 3 details the results of the hurdle negative binomial regression model (adjusted for zero inflation) assessing sex, age range, rank, branch, and year on ATR risk. When controlling for other factors, female sex and senior officers demonstrated significantly lower PR of ATR than the male and junior enlisted reference groups, respectively. Older age (>30 years) and service in the US Marine Corps and Air Force were salient associated factors for increased risk of ATR (PR: 1.17-2.50). Supplemental table 3 compares the negative binomial models with and without adjustment for zero inflation. While the results converged between the models for most of the variables assessed, the hurdle model outperformed the standard negative binomial model.

DISCUSSION

This study found that ATRs were common in the US military and identified multiple salient associated risk factors, including male sex, older age, enlisted or junior officer rank, military occupation, and service in the US Marine Corps or Air Force. Among military occupations, administration, operations, intelligence, and logistics specialties had the greatest risk for ATR in both officer and enlisted personnel. Integrating our understanding from the literature on civilian populations with the present study presents unique opportunities to interrogate specific aspects of mechanisms underlying ATR risk, including the contribution of exposure to activities with high Achilles tendon loads, sex, and age.

We found the burden of ATR more closely aligns with those previously described in the civilian population,² contrasted with the high rates previously reported in the US military.⁶ The annual burden was consistent throughout the study epoch, despite the changes in exposure and tempo of military operations. This finding diverges from studies in the civilian population describing a progressive increased burden of ATR beginning in the 1990s and continuing to the present.¹³ It may be that an increasingly sedentary civilian population engaged in irregular bouts of activity (“weekend warriors”) has led to increased burden of ATR over time compared with the more consistently physically active military population.

Findings from this study underscore the importance of activity exposure on ATR risk. Interestingly, this study supports the framework that both overloading and underloading with bouts of high loads can be detrimental to Achilles tendon health. It may seem contradictory that service in the US Marine Corps (routinely encountering high Achilles tendon loading activities) and Air Force (routinely involved in low Achilles tendon loading activities) both had increased risk of Achilles tendon rupture compared with the US Army. We posit that these two branches represent different occupational risk exposures. Individuals in the US Marine Corps are routinely engaged in high-load running and jumping tasks and may represent an “overload” mechanism contribution to ATR risk. Alternatively, individuals in the Air Force are more routinely involved in technical occupations with longer bouts of sitting interspersed with less-frequent bouts of high Achilles loading activities. This group may represent a more traditional, “underload with intermittent high loading” mechanism contributing to ATR risk. Similarly, there is a lower risk of ATR in senior officers (when accounting for age), supporting the idea that individuals with high administrative loads, generally associated with longer bouts of sedentary activities without intermittent high loading bouts, have less occupational exposure to the high eccentric loads, leading to ATR.

A higher burden of ATR in men than in women has been well established in the civilian population, with injuries occurring 3-7 times more frequently in men.¹⁴ The mechanisms underlying this sexual dimorphism is a point of ongoing study. Studies of preclinical models of the Achilles tendon have found regulation of female hormones to be an important contributor to tendon mechanical behavior in the context of maintaining homeostasis and healing.^{15, 16} Furthermore, women have been found to have greater passive dorsiflexion motion compared with men, a function of triceps surae extensibility.¹⁷ Recent studies in human tendon have identified the presence of estrogen receptors in tendon tissue and estrogen effects on tendon biomechanical function, supporting the notion of sex hormone-mediated signaling in tendon.^18–20 Studies investigating sex-mediated differences in patient outcomes after ATR have found conflicting results, though the propensity for male patients to incur this injury more than female patients is highly consistent across studies.^21–23

The influence of sex on tendon homeostasis and healing in humans, though, is potentially confounded by a host of factors outside of direct hormonal influence. For example, it is possible that behavioral differences between sexes relating to engaging in recreational sport activity²⁴ could partially account for differences in ATR burden, given that rupture tends to be associated with forceful eccentric contractions encountered during team sports like basketball and football. The present study informs this conversation by demonstrating that burden of ATR is higher in men than women when occupational exposure is comparable (for every occupational specialty included in this study, except for enlisted aviation). However, the relative risk encountered by women compared with men was lowest in specialties in which exposures are generally considered to have lower physical demands and are more sedentary (e.g., officer specialties, particularly administrative and logistics roles). The findings of this study add support to the notion that there is a mechanism by which sex influences ATR risk, and this risk is also influenced by repeated exposure to activities in which the Achilles tendon is likely to experience higher loads.

Service members above the age of 30 were found to be at increased risk of ATR. This finding aligns well with prior studies of ATR in civilian and elite athlete populations, where the highest incidence occurs in individuals aged 30-50 years.^{13, 14, 25, 26} Individuals in middle and older age groups have also consistently been found to have less optimal functional outcomes compared with younger individuals,^{21, 23} though individuals over the age of 65 have been reported to have greater improvements in self-reported postinjury outcomes compared with younger age groups.²² Identifying mechanisms underlying the isolated factor of age on ATR risk is challenging because it is often difficult to disentangle behavioral changes (increased sedentary behavior) and tendon maturation (age-related mechanical and physiological changes).²⁷ That said, it does seem that aging can have an effect on tendon cellular function and matrix composition, though regular activity mitigates many of these changes.²⁷ Similarly, aging has inconsistently been associated with reduced muscle cross-sectional area and altered Achilles tendon mechanical behavior, but these characteristics also seem to be moderated by physical activity level.^{27, 28} Reductions in muscle cross-sectional area and Achilles tendon extensibility, findings typically associated with aging, were not observed when participants were matched for height, weight, and activity level.²⁹ The findings from the present study further support that the risk of ATR does not likely reflect aging alone but rather the combined effects of both age and exposure.

A substantial strength of this study was that using DMED allowed for assessment of population-level differences in ATR across several strata, including sex, occupation, military rank, branch of service, and year. This study also had some limitations. Using diagnostic codes has inherent weaknesses, especially in military populations where many determinants can influence care seeking.³⁰ The data used in this study only represent individuals who sought treatment for their injuries and do not include individuals who self-managed their condition, though ATR presents with substantial functional deficits that would be difficult to self-manage, mitigating concerns regarding care-seeking behavior. This study also included each occurrence of ATR as a separate care episode. Individuals with a history of ATR have been found to have increased burden of rupture on the contralateral limb compared with the general population,³¹ so it is possible that a single person may have contributed more than one episode of care in the present study. Lastly, although personnel were categorized by their military occupation, a detailed task analysis to determine the specific duties and exposures that may be associated with these injuries was not possible.

CONCLUSION

ATR was ubiquitous in the US military, with multiple risk factors identified, including male sex, older age, rank, military occupation, and service branch. Based on the burden of ATR in the US military observed in this study, these findings highlight both the need for prophylactic interventions and identification of the populations who can most greatly benefit from preventive screening and care.